Annotation of gforth/doc/gforth.ds, revision 1.75
1.1 anton 1: \input texinfo @c -*-texinfo-*-
2: @comment The source is gforth.ds, from which gforth.texi is generated
1.28 crook 3:
1.21 crook 4: @comment TODO: nac29jan99 - a list of things to add in the next edit:
1.28 crook 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.
1.36 anton 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
1.29 crook 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:
1.66 anton 17: @c POSTPONE, COMPILE, [COMPILE], LITERAL should have their own section
1.28 crook 18:
1.1 anton 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
1.49 anton 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:
1.1 anton 33: @comment @setchapternewpage odd
1.29 crook 34: @comment TODO this gets left in by HTML converter
1.12 anton 35: @macro progstyle {}
36: Programming style note:
1.3 anton 37: @end macro
1.48 anton 38:
39: @macro assignment {}
40: @table @i
41: @item Assignment:
42: @end macro
43: @macro endassignment {}
44: @end table
45: @end macro
46:
1.1 anton 47: @comment %**end of header (This is for running Texinfo on a region.)
48:
1.29 crook 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:
1.10 anton 71: @include version.texi
72:
1.49 anton 73: @ifnottex
1.11 anton 74: This file documents Gforth @value{VERSION}
1.1 anton 75:
1.62 crook 76: Copyright @copyright{} 1995--2000 Free Software Foundation, Inc.
1.1 anton 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.
1.49 anton 101: @end ifnottex
1.1 anton 102:
103: @finalout
104: @titlepage
105: @sp 10
106: @center @titlefont{Gforth Manual}
107: @sp 2
1.11 anton 108: @center for version @value{VERSION}
1.1 anton 109: @sp 2
1.34 anton 110: @center Neal Crook
1.1 anton 111: @center Anton Ertl
1.6 pazsan 112: @center Bernd Paysan
1.5 anton 113: @center Jens Wilke
1.1 anton 114: @sp 3
1.47 crook 115: @center This manual is permanently under construction and was last updated on 15-Mar-2000
1.1 anton 116:
117: @comment The following two commands start the copyright page.
118: @page
119: @vskip 0pt plus 1filll
1.62 crook 120: Copyright @copyright{} 1995--2000 Free Software Foundation, Inc.
1.1 anton 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)
1.49 anton 143: @ifnottex
1.1 anton 144: Gforth is a free implementation of ANS Forth available on many
1.11 anton 145: personal machines. This manual corresponds to version @value{VERSION}.
1.49 anton 146: @end ifnottex
1.1 anton 147:
148: @menu
1.21 crook 149: * License:: The GPL
1.26 crook 150: * Goals:: About the Gforth Project
1.29 crook 151: * Gforth Environment:: Starting (and exiting) Gforth
1.48 anton 152: * Tutorial:: Hands-on Forth Tutorial
1.21 crook 153: * Introduction:: An introduction to ANS Forth
1.1 anton 154: * Words:: Forth words available in Gforth
1.24 anton 155: * Error messages:: How to interpret them
1.1 anton 156: * Tools:: Programming tools
157: * ANS conformance:: Implementation-defined options etc.
1.65 anton 158: * Standard vs Extensions:: Should I use extensions?
1.1 anton 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
1.24 anton 164: * Binding to System Library::
1.13 pazsan 165: * Cross Compiler:: The Cross Compiler
1.1 anton 166: * Bugs:: How to report them
167: * Origin:: Authors and ancestors of Gforth
1.21 crook 168: * Forth-related information:: Books and places to look on the WWW
1.1 anton 169: * Word Index:: An item for each Forth word
1.41 anton 170: * Name Index:: Forth words, only names listed
1.1 anton 171: * Concept Index:: A menu covering many topics
1.12 anton 172:
1.48 anton 173: @detailmenu --- The Detailed Node Listing ---
1.12 anton 174:
1.29 crook 175: Gforth Environment
176:
1.32 anton 177: * Invoking Gforth:: Getting in
178: * Leaving Gforth:: Getting out
179: * Command-line editing::
1.48 anton 180: * Environment variables:: that affect how Gforth starts up
1.32 anton 181: * Gforth Files:: What gets installed and where
1.48 anton 182: * Startup speed:: When 35ms is not fast enough ...
183:
184: Forth Tutorial
185:
186: * Starting Gforth Tutorial::
187: * Syntax Tutorial::
188: * Crash Course Tutorial::
189: * Stack Tutorial::
190: * Arithmetics Tutorial::
191: * Stack Manipulation Tutorial::
192: * Using files for Forth code Tutorial::
193: * Comments Tutorial::
194: * Colon Definitions Tutorial::
195: * Decompilation Tutorial::
196: * Stack-Effect Comments Tutorial::
197: * Types Tutorial::
198: * Factoring Tutorial::
199: * Designing the stack effect Tutorial::
200: * Local Variables Tutorial::
201: * Conditional execution Tutorial::
202: * Flags and Comparisons Tutorial::
203: * General Loops Tutorial::
204: * Counted loops Tutorial::
205: * Recursion Tutorial::
206: * Leaving definitions or loops Tutorial::
207: * Return Stack Tutorial::
208: * Memory Tutorial::
209: * Characters and Strings Tutorial::
210: * Alignment Tutorial::
211: * Interpretation and Compilation Semantics and Immediacy Tutorial::
212: * Execution Tokens Tutorial::
213: * Exceptions Tutorial::
214: * Defining Words Tutorial::
215: * Arrays and Records Tutorial::
216: * POSTPONE Tutorial::
217: * Literal Tutorial::
218: * Advanced macros Tutorial::
219: * Compilation Tokens Tutorial::
220: * Wordlists and Search Order Tutorial::
1.29 crook 221:
1.24 anton 222: An Introduction to ANS Forth
223:
1.67 anton 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::
1.24 anton 232:
1.12 anton 233: Forth Words
234:
235: * Notation::
1.65 anton 236: * Case insensitivity::
237: * Comments::
238: * Boolean Flags::
1.12 anton 239: * Arithmetic::
240: * Stack Manipulation::
241: * Memory::
242: * Control Structures::
243: * Defining Words::
1.65 anton 244: * Interpretation and Compilation Semantics::
1.47 crook 245: * Tokens for Words::
1.65 anton 246: * The Text Interpreter::
247: * Word Lists::
248: * Environmental Queries::
1.12 anton 249: * Files::
250: * Blocks::
251: * Other I/O::
252: * Programming Tools::
253: * Assembler and Code Words::
254: * Threading Words::
1.26 crook 255: * Locals::
256: * Structures::
257: * Object-oriented Forth::
1.65 anton 258: * Passing Commands to the OS::
259: * Keeping track of Time::
260: * Miscellaneous Words::
1.12 anton 261:
262: Arithmetic
263:
264: * Single precision::
1.67 anton 265: * Double precision:: Double-cell integer arithmetic
1.12 anton 266: * Bitwise operations::
1.67 anton 267: * Numeric comparison::
1.32 anton 268: * Mixed precision:: Operations with single and double-cell integers
1.12 anton 269: * Floating Point::
270:
271: Stack Manipulation
272:
273: * Data stack::
274: * Floating point stack::
275: * Return stack::
276: * Locals stack::
277: * Stack pointer manipulation::
278:
279: Memory
280:
1.32 anton 281: * Memory model::
282: * Dictionary allocation::
283: * Heap Allocation::
284: * Memory Access::
285: * Address arithmetic::
286: * Memory Blocks::
1.12 anton 287:
288: Control Structures
289:
1.41 anton 290: * Selection:: IF ... ELSE ... ENDIF
291: * Simple Loops:: BEGIN ...
1.32 anton 292: * Counted Loops:: DO
1.67 anton 293: * Arbitrary control structures::
294: * Calls and returns::
1.12 anton 295: * Exception Handling::
296:
297: Defining Words
298:
1.67 anton 299: * CREATE::
1.44 crook 300: * Variables:: Variables and user variables
1.67 anton 301: * Constants::
1.44 crook 302: * Values:: Initialised variables
1.67 anton 303: * Colon Definitions::
1.44 crook 304: * Anonymous Definitions:: Definitions without names
1.71 anton 305: * Supplying names:: Passing definition names as strings
1.67 anton 306: * User-defined Defining Words::
1.44 crook 307: * Deferred words:: Allow forward references
1.67 anton 308: * Aliases::
1.47 crook 309:
1.63 anton 310: User-defined Defining Words
311:
312: * CREATE..DOES> applications::
313: * CREATE..DOES> details::
314: * Advanced does> usage example::
315:
1.47 crook 316: Interpretation and Compilation Semantics
317:
1.67 anton 318: * Combined words::
1.12 anton 319:
1.71 anton 320: Tokens for Words
321:
322: * Execution token:: represents execution/interpretation semantics
323: * Compilation token:: represents compilation semantics
324: * Name token:: represents named words
325:
1.21 crook 326: The Text Interpreter
327:
1.67 anton 328: * Input Sources::
329: * Number Conversion::
330: * Interpret/Compile states::
331: * Literals::
332: * Interpreter Directives::
1.21 crook 333:
1.26 crook 334: Word Lists
335:
1.75 ! anton 336: * Vocabularies::
1.67 anton 337: * Why use word lists?::
1.75 ! anton 338: * Word list example::
1.26 crook 339:
340: Files
341:
1.48 anton 342: * Forth source files::
343: * General files::
344: * Search Paths::
345:
346: Search Paths
347:
1.75 ! anton 348: * Source Search Paths::
1.26 crook 349: * General Search Paths::
350:
351: Other I/O
352:
1.32 anton 353: * Simple numeric output:: Predefined formats
354: * Formatted numeric output:: Formatted (pictured) output
355: * String Formats:: How Forth stores strings in memory
1.67 anton 356: * Displaying characters and strings:: Other stuff
1.32 anton 357: * Input:: Input
1.26 crook 358:
359: Programming Tools
360:
361: * Debugging:: Simple and quick.
362: * Assertions:: Making your programs self-checking.
1.46 pazsan 363: * Singlestep Debugger:: Executing your program word by word.
1.26 crook 364:
1.63 anton 365: Assembler and Code Words
366:
367: * Code and ;code::
368: * Common Assembler:: Assembler Syntax
369: * Common Disassembler::
370: * 386 Assembler:: Deviations and special cases
371: * Alpha Assembler:: Deviations and special cases
372: * MIPS assembler:: Deviations and special cases
373: * Other assemblers:: How to write them
374:
1.26 crook 375: Locals
376:
377: * Gforth locals::
378: * ANS Forth locals::
379:
380: Gforth locals
381:
382: * Where are locals visible by name?::
383: * How long do locals live?::
384: * Programming Style::
385: * Implementation::
386:
1.12 anton 387: Structures
388:
389: * Why explicit structure support?::
390: * Structure Usage::
391: * Structure Naming Convention::
392: * Structure Implementation::
393: * Structure Glossary::
394:
395: Object-oriented Forth
396:
1.48 anton 397: * Why object-oriented programming?::
398: * Object-Oriented Terminology::
399: * Objects::
400: * OOF::
401: * Mini-OOF::
1.23 crook 402: * Comparison with other object models::
1.12 anton 403:
1.24 anton 404: The @file{objects.fs} model
1.12 anton 405:
406: * Properties of the Objects model::
407: * Basic Objects Usage::
1.41 anton 408: * The Objects base class::
1.12 anton 409: * Creating objects::
410: * Object-Oriented Programming Style::
411: * Class Binding::
412: * Method conveniences::
413: * Classes and Scoping::
1.41 anton 414: * Dividing classes::
1.12 anton 415: * Object Interfaces::
416: * Objects Implementation::
417: * Objects Glossary::
418:
1.24 anton 419: The @file{oof.fs} model
1.12 anton 420:
1.67 anton 421: * Properties of the OOF model::
422: * Basic OOF Usage::
423: * The OOF base class::
424: * Class Declaration::
425: * Class Implementation::
1.12 anton 426:
1.24 anton 427: The @file{mini-oof.fs} model
1.23 crook 428:
1.48 anton 429: * Basic Mini-OOF Usage::
430: * Mini-OOF Example::
431: * Mini-OOF Implementation::
1.23 crook 432:
1.12 anton 433: Tools
434:
435: * ANS Report:: Report the words used, sorted by wordset.
436:
437: ANS conformance
438:
439: * The Core Words::
440: * The optional Block word set::
441: * The optional Double Number word set::
442: * The optional Exception word set::
443: * The optional Facility word set::
444: * The optional File-Access word set::
445: * The optional Floating-Point word set::
446: * The optional Locals word set::
447: * The optional Memory-Allocation word set::
448: * The optional Programming-Tools word set::
449: * The optional Search-Order word set::
450:
451: The Core Words
452:
453: * core-idef:: Implementation Defined Options
454: * core-ambcond:: Ambiguous Conditions
455: * core-other:: Other System Documentation
456:
457: The optional Block word set
458:
459: * block-idef:: Implementation Defined Options
460: * block-ambcond:: Ambiguous Conditions
461: * block-other:: Other System Documentation
462:
463: The optional Double Number word set
464:
465: * double-ambcond:: Ambiguous Conditions
466:
467: The optional Exception word set
468:
469: * exception-idef:: Implementation Defined Options
470:
471: The optional Facility word set
472:
473: * facility-idef:: Implementation Defined Options
474: * facility-ambcond:: Ambiguous Conditions
475:
476: The optional File-Access word set
477:
478: * file-idef:: Implementation Defined Options
479: * file-ambcond:: Ambiguous Conditions
480:
481: The optional Floating-Point word set
482:
483: * floating-idef:: Implementation Defined Options
484: * floating-ambcond:: Ambiguous Conditions
485:
486: The optional Locals word set
487:
488: * locals-idef:: Implementation Defined Options
489: * locals-ambcond:: Ambiguous Conditions
490:
491: The optional Memory-Allocation word set
492:
493: * memory-idef:: Implementation Defined Options
494:
495: The optional Programming-Tools word set
496:
497: * programming-idef:: Implementation Defined Options
498: * programming-ambcond:: Ambiguous Conditions
499:
500: The optional Search-Order word set
501:
502: * search-idef:: Implementation Defined Options
503: * search-ambcond:: Ambiguous Conditions
504:
505: Image Files
506:
1.24 anton 507: * Image Licensing Issues:: Distribution terms for images.
508: * Image File Background:: Why have image files?
1.67 anton 509: * Non-Relocatable Image Files:: don't always work.
1.24 anton 510: * Data-Relocatable Image Files:: are better.
1.67 anton 511: * Fully Relocatable Image Files:: better yet.
1.24 anton 512: * Stack and Dictionary Sizes:: Setting the default sizes for an image.
1.32 anton 513: * Running Image Files:: @code{gforth -i @i{file}} or @i{file}.
1.24 anton 514: * Modifying the Startup Sequence:: and turnkey applications.
1.12 anton 515:
516: Fully Relocatable Image Files
517:
1.27 crook 518: * gforthmi:: The normal way
1.12 anton 519: * cross.fs:: The hard way
520:
521: Engine
522:
523: * Portability::
524: * Threading::
525: * Primitives::
526: * Performance::
527:
528: Threading
529:
530: * Scheduling::
531: * Direct or Indirect Threaded?::
532: * DOES>::
533:
534: Primitives
535:
536: * Automatic Generation::
537: * TOS Optimization::
538: * Produced code::
1.13 pazsan 539:
540: Cross Compiler
541:
1.67 anton 542: * Using the Cross Compiler::
543: * How the Cross Compiler Works::
1.13 pazsan 544:
1.24 anton 545: Other Forth-related information
1.21 crook 546:
1.67 anton 547: * Internet resources::
548: * Books::
549: * The Forth Interest Group::
550: * Conferences::
1.21 crook 551:
1.24 anton 552: @end detailmenu
1.1 anton 553: @end menu
554:
1.26 crook 555: @node License, Goals, Top, Top
1.1 anton 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: 675 Mass Ave, Cambridge, MA 02139, 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
1.49 anton 620: @ifnottex
1.1 anton 621: @center TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
1.49 anton 622: @end ifnottex
1.1 anton 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
1.49 anton 845: @ifnottex
1.1 anton 846: @center NO WARRANTY
1.49 anton 847: @end ifnottex
1.1 anton 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
1.49 anton 875: @ifnottex
1.1 anton 876: @center END OF TERMS AND CONDITIONS
1.49 anton 877: @end ifnottex
1.1 anton 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., 675 Mass Ave, Cambridge, MA 02139, 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
1.21 crook 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.
1.1 anton 954: @end iftex
955:
1.28 crook 956: @comment TODO much more blurb here.
1.26 crook 957:
958: @c ******************************************************************
1.29 crook 959: @node Goals, Gforth Environment, License, Top
1.26 crook 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
1.65 anton 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.
1.29 crook 1003:
1.26 crook 1004: @c ******************************************************************
1.48 anton 1005: @node Gforth Environment, Tutorial, Goals, Top
1.29 crook 1006: @chapter Gforth Environment
1007: @cindex Gforth environment
1.21 crook 1008:
1.45 crook 1009: Note: ultimately, the Gforth man page will be auto-generated from the
1.29 crook 1010: material in this chapter.
1.21 crook 1011:
1012: @menu
1.29 crook 1013: * Invoking Gforth:: Getting in
1014: * Leaving Gforth:: Getting out
1015: * Command-line editing::
1.48 anton 1016: * Environment variables:: that affect how Gforth starts up
1.29 crook 1017: * Gforth Files:: What gets installed and where
1.48 anton 1018: * Startup speed:: When 35ms is not fast enough ...
1.21 crook 1019: @end menu
1020:
1.49 anton 1021: For related information about the creation of images see @ref{Image Files}.
1.29 crook 1022:
1.21 crook 1023: @comment ----------------------------------------------
1.48 anton 1024: @node Invoking Gforth, Leaving Gforth, Gforth Environment, Gforth Environment
1.29 crook 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
1.21 crook 1031:
1.30 anton 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:
1.21 crook 1037: @example
1.30 anton 1038: gforth [file | -e forth-code] ...
1.21 crook 1039: @end example
1.29 crook 1040: @noindent
1041: This interprets the contents of the files and the Forth code in the order they
1042: are given.
1.21 crook 1043:
1.30 anton 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}).
1047:
1.29 crook 1048: In general, the command line looks like this:
1.21 crook 1049:
1050: @example
1.30 anton 1051: gforth[-fast] [engine options] [image options]
1.21 crook 1052: @end example
1053:
1.30 anton 1054: The engine options must come before the rest of the command
1.29 crook 1055: line. They are:
1.26 crook 1056:
1.29 crook 1057: @table @code
1058: @cindex -i, command-line option
1059: @cindex --image-file, command-line option
1060: @item --image-file @i{file}
1061: @itemx -i @i{file}
1062: Loads the Forth image @i{file} instead of the default
1063: @file{gforth.fi} (@pxref{Image Files}).
1.21 crook 1064:
1.39 anton 1065: @cindex --appl-image, command-line option
1066: @item --appl-image @i{file}
1067: Loads the image @i{file} and leaves all further command-line arguments
1.65 anton 1068: to the image (instead of processing them as engine options). This is
1069: useful for building executable application images on Unix, built with
1.39 anton 1070: @code{gforthmi --application ...}.
1071:
1.29 crook 1072: @cindex --path, command-line option
1073: @cindex -p, command-line option
1074: @item --path @i{path}
1075: @itemx -p @i{path}
1076: Uses @i{path} for searching the image file and Forth source code files
1077: instead of the default in the environment variable @code{GFORTHPATH} or
1078: the path specified at installation time (e.g.,
1079: @file{/usr/local/share/gforth/0.2.0:.}). A path is given as a list of
1080: directories, separated by @samp{:} (on Unix) or @samp{;} (on other OSs).
1.21 crook 1081:
1.29 crook 1082: @cindex --dictionary-size, command-line option
1083: @cindex -m, command-line option
1084: @cindex @i{size} parameters for command-line options
1085: @cindex size of the dictionary and the stacks
1086: @item --dictionary-size @i{size}
1087: @itemx -m @i{size}
1088: Allocate @i{size} space for the Forth dictionary space instead of
1089: using the default specified in the image (typically 256K). The
1090: @i{size} specification for this and subsequent options consists of
1091: an integer and a unit (e.g.,
1092: @code{4M}). The unit can be one of @code{b} (bytes), @code{e} (element
1093: size, in this case Cells), @code{k} (kilobytes), @code{M} (Megabytes),
1094: @code{G} (Gigabytes), and @code{T} (Terabytes). If no unit is specified,
1095: @code{e} is used.
1.21 crook 1096:
1.29 crook 1097: @cindex --data-stack-size, command-line option
1098: @cindex -d, command-line option
1099: @item --data-stack-size @i{size}
1100: @itemx -d @i{size}
1101: Allocate @i{size} space for the data stack instead of using the
1102: default specified in the image (typically 16K).
1.21 crook 1103:
1.29 crook 1104: @cindex --return-stack-size, command-line option
1105: @cindex -r, command-line option
1106: @item --return-stack-size @i{size}
1107: @itemx -r @i{size}
1108: Allocate @i{size} space for the return stack instead of using the
1109: default specified in the image (typically 15K).
1.21 crook 1110:
1.29 crook 1111: @cindex --fp-stack-size, command-line option
1112: @cindex -f, command-line option
1113: @item --fp-stack-size @i{size}
1114: @itemx -f @i{size}
1115: Allocate @i{size} space for the floating point stack instead of
1116: using the default specified in the image (typically 15.5K). In this case
1117: the unit specifier @code{e} refers to floating point numbers.
1.21 crook 1118:
1.48 anton 1119: @cindex --locals-stack-size, command-line option
1120: @cindex -l, command-line option
1121: @item --locals-stack-size @i{size}
1122: @itemx -l @i{size}
1123: Allocate @i{size} space for the locals stack instead of using the
1124: default specified in the image (typically 14.5K).
1125:
1126: @cindex -h, command-line option
1127: @cindex --help, command-line option
1128: @item --help
1129: @itemx -h
1130: Print a message about the command-line options
1131:
1132: @cindex -v, command-line option
1133: @cindex --version, command-line option
1134: @item --version
1135: @itemx -v
1136: Print version and exit
1137:
1138: @cindex --debug, command-line option
1139: @item --debug
1140: Print some information useful for debugging on startup.
1141:
1142: @cindex --offset-image, command-line option
1143: @item --offset-image
1144: Start the dictionary at a slightly different position than would be used
1145: otherwise (useful for creating data-relocatable images,
1146: @pxref{Data-Relocatable Image Files}).
1147:
1148: @cindex --no-offset-im, command-line option
1149: @item --no-offset-im
1150: Start the dictionary at the normal position.
1151:
1152: @cindex --clear-dictionary, command-line option
1153: @item --clear-dictionary
1154: Initialize all bytes in the dictionary to 0 before loading the image
1155: (@pxref{Data-Relocatable Image Files}).
1156:
1157: @cindex --die-on-signal, command-line-option
1158: @item --die-on-signal
1159: Normally Gforth handles most signals (e.g., the user interrupt SIGINT,
1160: or the segmentation violation SIGSEGV) by translating it into a Forth
1161: @code{THROW}. With this option, Gforth exits if it receives such a
1162: signal. This option is useful when the engine and/or the image might be
1163: severely broken (such that it causes another signal before recovering
1164: from the first); this option avoids endless loops in such cases.
1165: @end table
1166:
1167: @cindex loading files at startup
1168: @cindex executing code on startup
1169: @cindex batch processing with Gforth
1170: As explained above, the image-specific command-line arguments for the
1171: default image @file{gforth.fi} consist of a sequence of filenames and
1172: @code{-e @var{forth-code}} options that are interpreted in the sequence
1173: in which they are given. The @code{-e @var{forth-code}} or
1174: @code{--evaluate @var{forth-code}} option evaluates the Forth
1175: code. This option takes only one argument; if you want to evaluate more
1176: Forth words, you have to quote them or use @code{-e} several times. To exit
1177: after processing the command line (instead of entering interactive mode)
1178: append @code{-e bye} to the command line.
1179:
1180: @cindex versions, invoking other versions of Gforth
1181: If you have several versions of Gforth installed, @code{gforth} will
1182: invoke the version that was installed last. @code{gforth-@i{version}}
1183: invokes a specific version. If your environment contains the variable
1184: @code{GFORTHPATH}, you may want to override it by using the
1185: @code{--path} option.
1186:
1187: Not yet implemented:
1188: On startup the system first executes the system initialization file
1189: (unless the option @code{--no-init-file} is given; note that the system
1190: resulting from using this option may not be ANS Forth conformant). Then
1191: the user initialization file @file{.gforth.fs} is executed, unless the
1.62 crook 1192: option @code{--no-rc} is given; this file is searched for in @file{.},
1.48 anton 1193: then in @file{~}, then in the normal path (see above).
1194:
1195:
1196:
1197: @comment ----------------------------------------------
1198: @node Leaving Gforth, Command-line editing, Invoking Gforth, Gforth Environment
1199: @section Leaving Gforth
1200: @cindex Gforth - leaving
1201: @cindex leaving Gforth
1202:
1203: You can leave Gforth by typing @code{bye} or @kbd{Ctrl-d} (at the start
1204: of a line) or (if you invoked Gforth with the @code{--die-on-signal}
1205: option) @kbd{Ctrl-c}. When you leave Gforth, all of your definitions and
1.49 anton 1206: data are discarded. For ways of saving the state of the system before
1207: leaving Gforth see @ref{Image Files}.
1.48 anton 1208:
1209: doc-bye
1210:
1211:
1212: @comment ----------------------------------------------
1.65 anton 1213: @node Command-line editing, Environment variables, Leaving Gforth, Gforth Environment
1.48 anton 1214: @section Command-line editing
1215: @cindex command-line editing
1216:
1217: Gforth maintains a history file that records every line that you type to
1218: the text interpreter. This file is preserved between sessions, and is
1219: used to provide a command-line recall facility; if you type @kbd{Ctrl-P}
1220: repeatedly you can recall successively older commands from this (or
1221: previous) session(s). The full list of command-line editing facilities is:
1222:
1223: @itemize @bullet
1224: @item
1225: @kbd{Ctrl-p} (``previous'') (or up-arrow) to recall successively older
1226: commands from the history buffer.
1227: @item
1228: @kbd{Ctrl-n} (``next'') (or down-arrow) to recall successively newer commands
1229: from the history buffer.
1230: @item
1231: @kbd{Ctrl-f} (or right-arrow) to move the cursor right, non-destructively.
1232: @item
1233: @kbd{Ctrl-b} (or left-arrow) to move the cursor left, non-destructively.
1234: @item
1235: @kbd{Ctrl-h} (backspace) to delete the character to the left of the cursor,
1236: closing up the line.
1237: @item
1238: @kbd{Ctrl-k} to delete (``kill'') from the cursor to the end of the line.
1239: @item
1240: @kbd{Ctrl-a} to move the cursor to the start of the line.
1241: @item
1242: @kbd{Ctrl-e} to move the cursor to the end of the line.
1243: @item
1244: @key{RET} (@kbd{Ctrl-m}) or @key{LFD} (@kbd{Ctrl-j}) to submit the current
1245: line.
1246: @item
1247: @key{TAB} to step through all possible full-word completions of the word
1248: currently being typed.
1249: @item
1.65 anton 1250: @kbd{Ctrl-d} on an empty line line to terminate Gforth (gracefully,
1251: using @code{bye}).
1252: @item
1253: @kbd{Ctrl-x} (or @code{Ctrl-d} on a non-empty line) to delete the
1254: character under the cursor.
1.48 anton 1255: @end itemize
1256:
1257: When editing, displayable characters are inserted to the left of the
1258: cursor position; the line is always in ``insert'' (as opposed to
1259: ``overstrike'') mode.
1260:
1261: @cindex history file
1262: @cindex @file{.gforth-history}
1263: On Unix systems, the history file is @file{~/.gforth-history} by
1264: default@footnote{i.e. it is stored in the user's home directory.}. You
1265: can find out the name and location of your history file using:
1266:
1267: @example
1268: history-file type \ Unix-class systems
1269:
1270: history-file type \ Other systems
1271: history-dir type
1272: @end example
1273:
1274: If you enter long definitions by hand, you can use a text editor to
1275: paste them out of the history file into a Forth source file for reuse at
1276: a later time.
1277:
1278: Gforth never trims the size of the history file, so you should do this
1279: periodically, if necessary.
1280:
1281: @comment this is all defined in history.fs
1282: @comment NAC TODO the ctrl-D behaviour can either do a bye or a beep.. how is that option
1283: @comment chosen?
1284:
1285:
1286: @comment ----------------------------------------------
1.65 anton 1287: @node Environment variables, Gforth Files, Command-line editing, Gforth Environment
1.48 anton 1288: @section Environment variables
1289: @cindex environment variables
1290:
1291: Gforth uses these environment variables:
1292:
1293: @itemize @bullet
1294: @item
1295: @cindex @code{GFORTHHIST} -- environment variable
1296: @code{GFORTHHIST} -- (Unix systems only) specifies the directory in which to
1297: open/create the history file, @file{.gforth-history}. Default:
1298: @code{$HOME}.
1299:
1300: @item
1301: @cindex @code{GFORTHPATH} -- environment variable
1302: @code{GFORTHPATH} -- specifies the path used when searching for the gforth image file and
1303: for Forth source-code files.
1304:
1305: @item
1306: @cindex @code{GFORTH} -- environment variable
1.49 anton 1307: @code{GFORTH} -- used by @file{gforthmi}, @xref{gforthmi}.
1.48 anton 1308:
1309: @item
1310: @cindex @code{GFORTHD} -- environment variable
1.62 crook 1311: @code{GFORTHD} -- used by @file{gforthmi}, @xref{gforthmi}.
1.48 anton 1312:
1313: @item
1314: @cindex @code{TMP}, @code{TEMP} - environment variable
1315: @code{TMP}, @code{TEMP} - (non-Unix systems only) used as a potential
1316: location for the history file.
1317: @end itemize
1318:
1319: @comment also POSIXELY_CORRECT LINES COLUMNS HOME but no interest in
1320: @comment mentioning these.
1321:
1322: All the Gforth environment variables default to sensible values if they
1323: are not set.
1324:
1325:
1326: @comment ----------------------------------------------
1327: @node Gforth Files, Startup speed, Environment variables, Gforth Environment
1328: @section Gforth files
1329: @cindex Gforth files
1330:
1331: When you install Gforth on a Unix system, it installs files in these
1332: locations by default:
1333:
1334: @itemize @bullet
1335: @item
1336: @file{/usr/local/bin/gforth}
1337: @item
1338: @file{/usr/local/bin/gforthmi}
1339: @item
1340: @file{/usr/local/man/man1/gforth.1} - man page.
1341: @item
1342: @file{/usr/local/info} - the Info version of this manual.
1343: @item
1344: @file{/usr/local/lib/gforth/<version>/...} - Gforth @file{.fi} files.
1345: @item
1346: @file{/usr/local/share/gforth/<version>/TAGS} - Emacs TAGS file.
1347: @item
1348: @file{/usr/local/share/gforth/<version>/...} - Gforth source files.
1349: @item
1350: @file{.../emacs/site-lisp/gforth.el} - Emacs gforth mode.
1351: @end itemize
1352:
1353: You can select different places for installation by using
1354: @code{configure} options (listed with @code{configure --help}).
1355:
1356: @comment ----------------------------------------------
1357: @node Startup speed, , Gforth Files, Gforth Environment
1358: @section Startup speed
1359: @cindex Startup speed
1360: @cindex speed, startup
1361:
1362: If Gforth is used for CGI scripts or in shell scripts, its startup
1363: speed may become a problem. On a 300MHz 21064a under Linux-2.2.13 with
1364: glibc-2.0.7, @code{gforth -e bye} takes about 24.6ms user and 11.3ms
1365: system time.
1366:
1367: If startup speed is a problem, you may consider the following ways to
1368: improve it; or you may consider ways to reduce the number of startups
1.62 crook 1369: (for example, by using Fast-CGI).
1.48 anton 1370:
1371: The first step to improve startup speed is to statically link Gforth, by
1372: building it with @code{XLDFLAGS=-static}. This requires more memory for
1373: the code and will therefore slow down the first invocation, but
1374: subsequent invocations avoid the dynamic linking overhead. Another
1375: disadvantage is that Gforth won't profit from library upgrades. As a
1376: result, @code{gforth-static -e bye} takes about 17.1ms user and
1377: 8.2ms system time.
1378:
1379: The next step to improve startup speed is to use a non-relocatable image
1.65 anton 1380: (@pxref{Non-Relocatable Image Files}). You can create this image with
1.48 anton 1381: @code{gforth -e "savesystem gforthnr.fi bye"} and later use it with
1382: @code{gforth -i gforthnr.fi ...}. This avoids the relocation overhead
1383: and a part of the copy-on-write overhead. The disadvantage is that the
1.62 crook 1384: non-relocatable image does not work if the OS gives Gforth a different
1.48 anton 1385: address for the dictionary, for whatever reason; so you better provide a
1386: fallback on a relocatable image. @code{gforth-static -i gforthnr.fi -e
1387: bye} takes about 15.3ms user and 7.5ms system time.
1388:
1389: The final step is to disable dictionary hashing in Gforth. Gforth
1390: builds the hash table on startup, which takes much of the startup
1391: overhead. You can do this by commenting out the @code{include hash.fs}
1392: in @file{startup.fs} and everything that requires @file{hash.fs} (at the
1393: moment @file{table.fs} and @file{ekey.fs}) and then doing @code{make}.
1394: The disadvantages are that functionality like @code{table} and
1395: @code{ekey} is missing and that text interpretation (e.g., compiling)
1396: now takes much longer. So, you should only use this method if there is
1397: no significant text interpretation to perform (the script should be
1.62 crook 1398: compiled into the image, amongst other things). @code{gforth-static -i
1.48 anton 1399: gforthnrnh.fi -e bye} takes about 2.1ms user and 6.1ms system time.
1400:
1401: @c ******************************************************************
1402: @node Tutorial, Introduction, Gforth Environment, Top
1403: @chapter Forth Tutorial
1404: @cindex Tutorial
1405: @cindex Forth Tutorial
1406:
1.67 anton 1407: @c Topics from nac's Introduction that could be mentioned:
1408: @c press <ret> after each line
1409: @c Prompt
1410: @c numbers vs. words in dictionary on text interpretation
1411: @c what happens on redefinition
1412: @c parsing words (in particular, defining words)
1413:
1.62 crook 1414: This tutorial can be used with any ANS-compliant Forth; any
1415: Gforth-specific features are marked as such and you can skip them if you
1416: work with another Forth. This tutorial does not explain all features of
1417: Forth, just enough to get you started and give you some ideas about the
1418: facilities available in Forth. Read the rest of the manual and the
1419: standard when you are through this.
1.48 anton 1420:
1421: The intended way to use this tutorial is that you work through it while
1422: sitting in front of the console, take a look at the examples and predict
1423: what they will do, then try them out; if the outcome is not as expected,
1424: find out why (e.g., by trying out variations of the example), so you
1425: understand what's going on. There are also some assignments that you
1426: should solve.
1427:
1428: This tutorial assumes that you have programmed before and know what,
1429: e.g., a loop is.
1430:
1431: @c !! explain compat library
1432:
1433: @menu
1434: * Starting Gforth Tutorial::
1435: * Syntax Tutorial::
1436: * Crash Course Tutorial::
1437: * Stack Tutorial::
1438: * Arithmetics Tutorial::
1439: * Stack Manipulation Tutorial::
1440: * Using files for Forth code Tutorial::
1441: * Comments Tutorial::
1442: * Colon Definitions Tutorial::
1443: * Decompilation Tutorial::
1444: * Stack-Effect Comments Tutorial::
1445: * Types Tutorial::
1446: * Factoring Tutorial::
1447: * Designing the stack effect Tutorial::
1448: * Local Variables Tutorial::
1449: * Conditional execution Tutorial::
1450: * Flags and Comparisons Tutorial::
1451: * General Loops Tutorial::
1452: * Counted loops Tutorial::
1453: * Recursion Tutorial::
1454: * Leaving definitions or loops Tutorial::
1455: * Return Stack Tutorial::
1456: * Memory Tutorial::
1457: * Characters and Strings Tutorial::
1458: * Alignment Tutorial::
1459: * Interpretation and Compilation Semantics and Immediacy Tutorial::
1460: * Execution Tokens Tutorial::
1461: * Exceptions Tutorial::
1462: * Defining Words Tutorial::
1463: * Arrays and Records Tutorial::
1464: * POSTPONE Tutorial::
1465: * Literal Tutorial::
1466: * Advanced macros Tutorial::
1467: * Compilation Tokens Tutorial::
1468: * Wordlists and Search Order Tutorial::
1469: @end menu
1470:
1471: @node Starting Gforth Tutorial, Syntax Tutorial, Tutorial, Tutorial
1472: @section Starting Gforth
1.66 anton 1473: @cindex starting Gforth tutorial
1.48 anton 1474: You can start Gforth by typing its name:
1475:
1476: @example
1477: gforth
1478: @end example
1479:
1480: That puts you into interactive mode; you can leave Gforth by typing
1481: @code{bye}. While in Gforth, you can edit the command line and access
1482: the command line history with cursor keys, similar to bash.
1483:
1484:
1485: @node Syntax Tutorial, Crash Course Tutorial, Starting Gforth Tutorial, Tutorial
1486: @section Syntax
1.66 anton 1487: @cindex syntax tutorial
1.48 anton 1488:
1489: A @dfn{word} is a sequence of arbitrary characters (expcept white
1490: space). Words are separated by white space. E.g., each of the
1491: following lines contains exactly one word:
1492:
1493: @example
1494: word
1495: !@@#$%^&*()
1496: 1234567890
1497: 5!a
1498: @end example
1499:
1500: A frequent beginner's error is to leave away necessary white space,
1501: resulting in an error like @samp{Undefined word}; so if you see such an
1502: error, check if you have put spaces wherever necessary.
1503:
1504: @example
1505: ." hello, world" \ correct
1506: ."hello, world" \ gives an "Undefined word" error
1507: @end example
1508:
1.65 anton 1509: Gforth and most other Forth systems ignore differences in case (they are
1.48 anton 1510: case-insensitive), i.e., @samp{word} is the same as @samp{Word}. If
1511: your system is case-sensitive, you may have to type all the examples
1512: given here in upper case.
1513:
1514:
1515: @node Crash Course Tutorial, Stack Tutorial, Syntax Tutorial, Tutorial
1516: @section Crash Course
1517:
1518: Type
1519:
1520: @example
1521: 0 0 !
1522: here execute
1523: ' catch >body 20 erase abort
1524: ' (quit) >body 20 erase
1525: @end example
1526:
1527: The last two examples are guaranteed to destroy parts of Gforth (and
1528: most other systems), so you better leave Gforth afterwards (if it has
1529: not finished by itself). On some systems you may have to kill gforth
1530: from outside (e.g., in Unix with @code{kill}).
1531:
1532: Now that you know how to produce crashes (and that there's not much to
1533: them), let's learn how to produce meaningful programs.
1534:
1535:
1536: @node Stack Tutorial, Arithmetics Tutorial, Crash Course Tutorial, Tutorial
1537: @section Stack
1.66 anton 1538: @cindex stack tutorial
1.48 anton 1539:
1540: The most obvious feature of Forth is the stack. When you type in a
1541: number, it is pushed on the stack. You can display the content of the
1542: stack with @code{.s}.
1543:
1544: @example
1545: 1 2 .s
1546: 3 .s
1547: @end example
1548:
1549: @code{.s} displays the top-of-stack to the right, i.e., the numbers
1550: appear in @code{.s} output as they appeared in the input.
1551:
1552: You can print the top of stack element with @code{.}.
1553:
1554: @example
1555: 1 2 3 . . .
1556: @end example
1557:
1558: In general, words consume their stack arguments (@code{.s} is an
1559: exception).
1560:
1561: @assignment
1562: What does the stack contain after @code{5 6 7 .}?
1563: @endassignment
1564:
1565:
1566: @node Arithmetics Tutorial, Stack Manipulation Tutorial, Stack Tutorial, Tutorial
1567: @section Arithmetics
1.66 anton 1568: @cindex arithmetics tutorial
1.48 anton 1569:
1570: The words @code{+}, @code{-}, @code{*}, @code{/}, and @code{mod} always
1571: operate on the top two stack items:
1572:
1573: @example
1.67 anton 1574: 2 2 .s
1575: + .s
1576: .
1.48 anton 1577: 2 1 - .
1578: 7 3 mod .
1579: @end example
1580:
1581: The operands of @code{-}, @code{/}, and @code{mod} are in the same order
1582: as in the corresponding infix expression (this is generally the case in
1583: Forth).
1584:
1585: Parentheses are superfluous (and not available), because the order of
1586: the words unambiguously determines the order of evaluation and the
1587: operands:
1588:
1589: @example
1590: 3 4 + 5 * .
1591: 3 4 5 * + .
1592: @end example
1593:
1594: @assignment
1595: What are the infix expressions corresponding to the Forth code above?
1596: Write @code{6-7*8+9} in Forth notation@footnote{This notation is also
1597: known as Postfix or RPN (Reverse Polish Notation).}.
1598: @endassignment
1599:
1600: To change the sign, use @code{negate}:
1601:
1602: @example
1603: 2 negate .
1604: @end example
1605:
1606: @assignment
1607: Convert -(-3)*4-5 to Forth.
1608: @endassignment
1609:
1610: @code{/mod} performs both @code{/} and @code{mod}.
1611:
1612: @example
1613: 7 3 /mod . .
1614: @end example
1615:
1.66 anton 1616: Reference: @ref{Arithmetic}.
1617:
1618:
1.48 anton 1619: @node Stack Manipulation Tutorial, Using files for Forth code Tutorial, Arithmetics Tutorial, Tutorial
1620: @section Stack Manipulation
1.66 anton 1621: @cindex stack manipulation tutorial
1.48 anton 1622:
1623: Stack manipulation words rearrange the data on the stack.
1624:
1625: @example
1626: 1 .s drop .s
1627: 1 .s dup .s drop drop .s
1628: 1 2 .s over .s drop drop drop
1629: 1 2 .s swap .s drop drop
1630: 1 2 3 .s rot .s drop drop drop
1631: @end example
1632:
1633: These are the most important stack manipulation words. There are also
1634: variants that manipulate twice as many stack items:
1635:
1636: @example
1637: 1 2 3 4 .s 2swap .s 2drop 2drop
1638: @end example
1639:
1640: Two more stack manipulation words are:
1641:
1642: @example
1643: 1 2 .s nip .s drop
1644: 1 2 .s tuck .s 2drop drop
1645: @end example
1646:
1647: @assignment
1648: Replace @code{nip} and @code{tuck} with combinations of other stack
1649: manipulation words.
1650:
1651: @example
1652: Given: How do you get:
1653: 1 2 3 3 2 1
1654: 1 2 3 1 2 3 2
1655: 1 2 3 1 2 3 3
1656: 1 2 3 1 3 3
1657: 1 2 3 2 1 3
1658: 1 2 3 4 4 3 2 1
1659: 1 2 3 1 2 3 1 2 3
1660: 1 2 3 4 1 2 3 4 1 2
1661: 1 2 3
1662: 1 2 3 1 2 3 4
1663: 1 2 3 1 3
1664: @end example
1665: @endassignment
1666:
1667: @example
1668: 5 dup * .
1669: @end example
1670:
1671: @assignment
1672: Write 17^3 and 17^4 in Forth, without writing @code{17} more than once.
1673: Write a piece of Forth code that expects two numbers on the stack
1674: (@var{a} and @var{b}, with @var{b} on top) and computes
1675: @code{(a-b)(a+1)}.
1676: @endassignment
1677:
1.66 anton 1678: Reference: @ref{Stack Manipulation}.
1679:
1680:
1.48 anton 1681: @node Using files for Forth code Tutorial, Comments Tutorial, Stack Manipulation Tutorial, Tutorial
1682: @section Using files for Forth code
1.66 anton 1683: @cindex loading Forth code, tutorial
1684: @cindex files containing Forth code, tutorial
1.48 anton 1685:
1686: While working at the Forth command line is convenient for one-line
1687: examples and short one-off code, you probably want to store your source
1688: code in files for convenient editing and persistence. You can use your
1689: favourite editor (Gforth includes Emacs support, @pxref{Emacs and
1690: Gforth}) to create @var{file} and use
1691:
1692: @example
1693: s" @var{file}" included
1694: @end example
1695:
1696: to load it into your Forth system. The file name extension I use for
1697: Forth files is @samp{.fs}.
1698:
1699: You can easily start Gforth with some files loaded like this:
1700:
1701: @example
1702: gforth @var{file1} @var{file2}
1703: @end example
1704:
1705: If an error occurs during loading these files, Gforth terminates,
1706: whereas an error during @code{INCLUDED} within Gforth usually gives you
1707: a Gforth command line. Starting the Forth system every time gives you a
1708: clean start every time, without interference from the results of earlier
1709: tries.
1710:
1711: I often put all the tests in a file, then load the code and run the
1712: tests with
1713:
1714: @example
1715: gforth @var{code} @var{tests} -e bye
1716: @end example
1717:
1718: (often by performing this command with @kbd{C-x C-e} in Emacs). The
1719: @code{-e bye} ensures that Gforth terminates afterwards so that I can
1720: restart this command without ado.
1721:
1722: The advantage of this approach is that the tests can be repeated easily
1723: every time the program ist changed, making it easy to catch bugs
1724: introduced by the change.
1725:
1.66 anton 1726: Reference: @ref{Forth source files}.
1727:
1.48 anton 1728:
1729: @node Comments Tutorial, Colon Definitions Tutorial, Using files for Forth code Tutorial, Tutorial
1730: @section Comments
1.66 anton 1731: @cindex comments tutorial
1.48 anton 1732:
1733: @example
1734: \ That's a comment; it ends at the end of the line
1735: ( Another comment; it ends here: ) .s
1736: @end example
1737:
1738: @code{\} and @code{(} are ordinary Forth words and therefore have to be
1739: separated with white space from the following text.
1740:
1741: @example
1742: \This gives an "Undefined word" error
1743: @end example
1744:
1745: The first @code{)} ends a comment started with @code{(}, so you cannot
1746: nest @code{(}-comments; and you cannot comment out text containing a
1747: @code{)} with @code{( ... )}@footnote{therefore it's a good idea to
1748: avoid @code{)} in word names.}.
1749:
1750: I use @code{\}-comments for descriptive text and for commenting out code
1751: of one or more line; I use @code{(}-comments for describing the stack
1752: effect, the stack contents, or for commenting out sub-line pieces of
1753: code.
1754:
1755: The Emacs mode @file{gforth.el} (@pxref{Emacs and Gforth}) supports
1756: these uses by commenting out a region with @kbd{C-x \}, uncommenting a
1757: region with @kbd{C-u C-x \}, and filling a @code{\}-commented region
1758: with @kbd{M-q}.
1759:
1.66 anton 1760: Reference: @ref{Comments}.
1761:
1.48 anton 1762:
1763: @node Colon Definitions Tutorial, Decompilation Tutorial, Comments Tutorial, Tutorial
1764: @section Colon Definitions
1.66 anton 1765: @cindex colon definitions, tutorial
1766: @cindex definitions, tutorial
1767: @cindex procedures, tutorial
1768: @cindex functions, tutorial
1.48 anton 1769:
1770: are similar to procedures and functions in other programming languages.
1771:
1772: @example
1773: : squared ( n -- n^2 )
1774: dup * ;
1775: 5 squared .
1776: 7 squared .
1777: @end example
1778:
1779: @code{:} starts the colon definition; its name is @code{squared}. The
1780: following comment describes its stack effect. The words @code{dup *}
1781: are not executed, but compiled into the definition. @code{;} ends the
1782: colon definition.
1783:
1784: The newly-defined word can be used like any other word, including using
1785: it in other definitions:
1786:
1787: @example
1788: : cubed ( n -- n^3 )
1789: dup squared * ;
1790: -5 cubed .
1791: : fourth-power ( n -- n^4 )
1792: squared squared ;
1793: 3 fourth-power .
1794: @end example
1795:
1796: @assignment
1797: Write colon definitions for @code{nip}, @code{tuck}, @code{negate}, and
1798: @code{/mod} in terms of other Forth words, and check if they work (hint:
1799: test your tests on the originals first). Don't let the
1800: @samp{redefined}-Messages spook you, they are just warnings.
1801: @endassignment
1802:
1.66 anton 1803: Reference: @ref{Colon Definitions}.
1804:
1.48 anton 1805:
1806: @node Decompilation Tutorial, Stack-Effect Comments Tutorial, Colon Definitions Tutorial, Tutorial
1807: @section Decompilation
1.66 anton 1808: @cindex decompilation tutorial
1809: @cindex see tutorial
1.48 anton 1810:
1811: You can decompile colon definitions with @code{see}:
1812:
1813: @example
1814: see squared
1815: see cubed
1816: @end example
1817:
1818: In Gforth @code{see} shows you a reconstruction of the source code from
1819: the executable code. Informations that were present in the source, but
1820: not in the executable code, are lost (e.g., comments).
1821:
1.65 anton 1822: You can also decompile the predefined words:
1823:
1824: @example
1825: see .
1826: see +
1827: @end example
1828:
1829:
1.48 anton 1830: @node Stack-Effect Comments Tutorial, Types Tutorial, Decompilation Tutorial, Tutorial
1831: @section Stack-Effect Comments
1.66 anton 1832: @cindex stack-effect comments, tutorial
1833: @cindex --, tutorial
1.48 anton 1834: By convention the comment after the name of a definition describes the
1835: stack effect: The part in from of the @samp{--} describes the state of
1836: the stack before the execution of the definition, i.e., the parameters
1837: that are passed into the colon definition; the part behind the @samp{--}
1838: is the state of the stack after the execution of the definition, i.e.,
1839: the results of the definition. The stack comment only shows the top
1840: stack items that the definition accesses and/or changes.
1841:
1842: You should put a correct stack effect on every definition, even if it is
1843: just @code{( -- )}. You should also add some descriptive comment to
1844: more complicated words (I usually do this in the lines following
1845: @code{:}). If you don't do this, your code becomes unreadable (because
1846: you have to work through every definition before you can undertsand
1847: any).
1848:
1849: @assignment
1850: The stack effect of @code{swap} can be written like this: @code{x1 x2 --
1851: x2 x1}. Describe the stack effect of @code{-}, @code{drop}, @code{dup},
1852: @code{over}, @code{rot}, @code{nip}, and @code{tuck}. Hint: When you
1.65 anton 1853: are done, you can compare your stack effects to those in this manual
1.48 anton 1854: (@pxref{Word Index}).
1855: @endassignment
1856:
1857: Sometimes programmers put comments at various places in colon
1858: definitions that describe the contents of the stack at that place (stack
1859: comments); i.e., they are like the first part of a stack-effect
1860: comment. E.g.,
1861:
1862: @example
1863: : cubed ( n -- n^3 )
1864: dup squared ( n n^2 ) * ;
1865: @end example
1866:
1867: In this case the stack comment is pretty superfluous, because the word
1868: is simple enough. If you think it would be a good idea to add such a
1869: comment to increase readability, you should also consider factoring the
1870: word into several simpler words (@pxref{Factoring Tutorial,,
1.60 anton 1871: Factoring}), which typically eliminates the need for the stack comment;
1.48 anton 1872: however, if you decide not to refactor it, then having such a comment is
1873: better than not having it.
1874:
1875: The names of the stack items in stack-effect and stack comments in the
1876: standard, in this manual, and in many programs specify the type through
1877: a type prefix, similar to Fortran and Hungarian notation. The most
1878: frequent prefixes are:
1879:
1880: @table @code
1881: @item n
1882: signed integer
1883: @item u
1884: unsigned integer
1885: @item c
1886: character
1887: @item f
1888: Boolean flags, i.e. @code{false} or @code{true}.
1889: @item a-addr,a-
1890: Cell-aligned address
1891: @item c-addr,c-
1892: Char-aligned address (note that a Char may have two bytes in Windows NT)
1893: @item xt
1894: Execution token, same size as Cell
1895: @item w,x
1896: Cell, can contain an integer or an address. It usually takes 32, 64 or
1897: 16 bits (depending on your platform and Forth system). A cell is more
1898: commonly known as machine word, but the term @emph{word} already means
1899: something different in Forth.
1900: @item d
1901: signed double-cell integer
1902: @item ud
1903: unsigned double-cell integer
1904: @item r
1905: Float (on the FP stack)
1906: @end table
1907:
1908: You can find a more complete list in @ref{Notation}.
1909:
1910: @assignment
1911: Write stack-effect comments for all definitions you have written up to
1912: now.
1913: @endassignment
1914:
1915:
1916: @node Types Tutorial, Factoring Tutorial, Stack-Effect Comments Tutorial, Tutorial
1917: @section Types
1.66 anton 1918: @cindex types tutorial
1.48 anton 1919:
1920: In Forth the names of the operations are not overloaded; so similar
1921: operations on different types need different names; e.g., @code{+} adds
1922: integers, and you have to use @code{f+} to add floating-point numbers.
1923: The following prefixes are often used for related operations on
1924: different types:
1925:
1926: @table @code
1927: @item (none)
1928: signed integer
1929: @item u
1930: unsigned integer
1931: @item c
1932: character
1933: @item d
1934: signed double-cell integer
1935: @item ud, du
1936: unsigned double-cell integer
1937: @item 2
1938: two cells (not-necessarily double-cell numbers)
1939: @item m, um
1940: mixed single-cell and double-cell operations
1941: @item f
1942: floating-point (note that in stack comments @samp{f} represents flags,
1.66 anton 1943: and @samp{r} represents FP numbers).
1.48 anton 1944: @end table
1945:
1946: If there are no differences between the signed and the unsigned variant
1947: (e.g., for @code{+}), there is only the prefix-less variant.
1948:
1949: Forth does not perform type checking, neither at compile time, nor at
1950: run time. If you use the wrong oeration, the data are interpreted
1951: incorrectly:
1952:
1953: @example
1954: -1 u.
1955: @end example
1956:
1957: If you have only experience with type-checked languages until now, and
1958: have heard how important type-checking is, don't panic! In my
1959: experience (and that of other Forthers), type errors in Forth code are
1960: usually easy to find (once you get used to it), the increased vigilance
1961: of the programmer tends to catch some harder errors in addition to most
1962: type errors, and you never have to work around the type system, so in
1963: most situations the lack of type-checking seems to be a win (projects to
1964: add type checking to Forth have not caught on).
1965:
1966:
1967: @node Factoring Tutorial, Designing the stack effect Tutorial, Types Tutorial, Tutorial
1968: @section Factoring
1.66 anton 1969: @cindex factoring tutorial
1.48 anton 1970:
1971: If you try to write longer definitions, you will soon find it hard to
1972: keep track of the stack contents. Therefore, good Forth programmers
1973: tend to write only short definitions (e.g., three lines). The art of
1974: finding meaningful short definitions is known as factoring (as in
1975: factoring polynomials).
1976:
1977: Well-factored programs offer additional advantages: smaller, more
1978: general words, are easier to test and debug and can be reused more and
1979: better than larger, specialized words.
1980:
1981: So, if you run into difficulties with stack management, when writing
1982: code, try to define meaningful factors for the word, and define the word
1983: in terms of those. Even if a factor contains only two words, it is
1984: often helpful.
1985:
1.65 anton 1986: Good factoring is not easy, and it takes some practice to get the knack
1987: for it; but even experienced Forth programmers often don't find the
1988: right solution right away, but only when rewriting the program. So, if
1989: you don't come up with a good solution immediately, keep trying, don't
1990: despair.
1.48 anton 1991:
1992: @c example !!
1993:
1994:
1995: @node Designing the stack effect Tutorial, Local Variables Tutorial, Factoring Tutorial, Tutorial
1996: @section Designing the stack effect
1.66 anton 1997: @cindex Stack effect design, tutorial
1998: @cindex design of stack effects, tutorial
1.48 anton 1999:
2000: In other languages you can use an arbitrary order of parameters for a
1.65 anton 2001: function; and since there is only one result, you don't have to deal with
1.48 anton 2002: the order of results, either.
2003:
2004: In Forth (and other stack-based languages, e.g., Postscript) the
2005: parameter and result order of a definition is important and should be
2006: designed well. The general guideline is to design the stack effect such
2007: that the word is simple to use in most cases, even if that complicates
2008: the implementation of the word. Some concrete rules are:
2009:
2010: @itemize @bullet
2011:
2012: @item
2013: Words consume all of their parameters (e.g., @code{.}).
2014:
2015: @item
2016: If there is a convention on the order of parameters (e.g., from
2017: mathematics or another programming language), stick with it (e.g.,
2018: @code{-}).
2019:
2020: @item
2021: If one parameter usually requires only a short computation (e.g., it is
2022: a constant), pass it on the top of the stack. Conversely, parameters
2023: that usually require a long sequence of code to compute should be passed
2024: as the bottom (i.e., first) parameter. This makes the code easier to
2025: read, because reader does not need to keep track of the bottom item
2026: through a long sequence of code (or, alternatively, through stack
1.49 anton 2027: manipulations). E.g., @code{!} (store, @pxref{Memory}) expects the
1.48 anton 2028: address on top of the stack because it is usually simpler to compute
2029: than the stored value (often the address is just a variable).
2030:
2031: @item
2032: Similarly, results that are usually consumed quickly should be returned
2033: on the top of stack, whereas a result that is often used in long
2034: computations should be passed as bottom result. E.g., the file words
2035: like @code{open-file} return the error code on the top of stack, because
2036: it is usually consumed quickly by @code{throw}; moreover, the error code
2037: has to be checked before doing anything with the other results.
2038:
2039: @end itemize
2040:
2041: These rules are just general guidelines, don't lose sight of the overall
2042: goal to make the words easy to use. E.g., if the convention rule
2043: conflicts with the computation-length rule, you might decide in favour
2044: of the convention if the word will be used rarely, and in favour of the
2045: computation-length rule if the word will be used frequently (because
2046: with frequent use the cost of breaking the computation-length rule would
2047: be quite high, and frequent use makes it easier to remember an
2048: unconventional order).
2049:
2050: @c example !! structure package
2051:
1.65 anton 2052:
1.48 anton 2053: @node Local Variables Tutorial, Conditional execution Tutorial, Designing the stack effect Tutorial, Tutorial
2054: @section Local Variables
1.66 anton 2055: @cindex local variables, tutorial
1.48 anton 2056:
2057: You can define local variables (@emph{locals}) in a colon definition:
2058:
2059: @example
2060: : swap @{ a b -- b a @}
2061: b a ;
2062: 1 2 swap .s 2drop
2063: @end example
2064:
2065: (If your Forth system does not support this syntax, include
2066: @file{compat/anslocals.fs} first).
2067:
2068: In this example @code{@{ a b -- b a @}} is the locals definition; it
2069: takes two cells from the stack, puts the top of stack in @code{b} and
2070: the next stack element in @code{a}. @code{--} starts a comment ending
2071: with @code{@}}. After the locals definition, using the name of the
2072: local will push its value on the stack. You can leave the comment
2073: part (@code{-- b a}) away:
2074:
2075: @example
2076: : swap ( x1 x2 -- x2 x1 )
2077: @{ a b @} b a ;
2078: @end example
2079:
2080: In Gforth you can have several locals definitions, anywhere in a colon
2081: definition; in contrast, in a standard program you can have only one
2082: locals definition per colon definition, and that locals definition must
2083: be outside any controll structure.
2084:
2085: With locals you can write slightly longer definitions without running
2086: into stack trouble. However, I recommend trying to write colon
2087: definitions without locals for exercise purposes to help you gain the
2088: essential factoring skills.
2089:
2090: @assignment
2091: Rewrite your definitions until now with locals
2092: @endassignment
2093:
1.66 anton 2094: Reference: @ref{Locals}.
2095:
1.48 anton 2096:
2097: @node Conditional execution Tutorial, Flags and Comparisons Tutorial, Local Variables Tutorial, Tutorial
2098: @section Conditional execution
1.66 anton 2099: @cindex conditionals, tutorial
2100: @cindex if, tutorial
1.48 anton 2101:
2102: In Forth you can use control structures only inside colon definitions.
2103: An @code{if}-structure looks like this:
2104:
2105: @example
2106: : abs ( n1 -- +n2 )
2107: dup 0 < if
2108: negate
2109: endif ;
2110: 5 abs .
2111: -5 abs .
2112: @end example
2113:
2114: @code{if} takes a flag from the stack. If the flag is non-zero (true),
2115: the following code is performed, otherwise execution continues after the
1.51 pazsan 2116: @code{endif} (or @code{else}). @code{<} compares the top two stack
1.48 anton 2117: elements and prioduces a flag:
2118:
2119: @example
2120: 1 2 < .
2121: 2 1 < .
2122: 1 1 < .
2123: @end example
2124:
2125: Actually the standard name for @code{endif} is @code{then}. This
2126: tutorial presents the examples using @code{endif}, because this is often
2127: less confusing for people familiar with other programming languages
2128: where @code{then} has a different meaning. If your system does not have
2129: @code{endif}, define it with
2130:
2131: @example
2132: : endif postpone then ; immediate
2133: @end example
2134:
2135: You can optionally use an @code{else}-part:
2136:
2137: @example
2138: : min ( n1 n2 -- n )
2139: 2dup < if
2140: drop
2141: else
2142: nip
2143: endif ;
2144: 2 3 min .
2145: 3 2 min .
2146: @end example
2147:
2148: @assignment
2149: Write @code{min} without @code{else}-part (hint: what's the definition
2150: of @code{nip}?).
2151: @endassignment
2152:
1.66 anton 2153: Reference: @ref{Selection}.
2154:
1.48 anton 2155:
2156: @node Flags and Comparisons Tutorial, General Loops Tutorial, Conditional execution Tutorial, Tutorial
2157: @section Flags and Comparisons
1.66 anton 2158: @cindex flags tutorial
2159: @cindex comparison tutorial
1.48 anton 2160:
2161: In a false-flag all bits are clear (0 when interpreted as integer). In
2162: a canonical true-flag all bits are set (-1 as a twos-complement signed
2163: integer); in many contexts (e.g., @code{if}) any non-zero value is
2164: treated as true flag.
2165:
2166: @example
2167: false .
2168: true .
2169: true hex u. decimal
2170: @end example
2171:
2172: Comparison words produce canonical flags:
2173:
2174: @example
2175: 1 1 = .
2176: 1 0= .
2177: 0 1 < .
2178: 0 0 < .
2179: -1 1 u< . \ type error, u< interprets -1 as large unsigned number
2180: -1 1 < .
2181: @end example
2182:
1.66 anton 2183: Gforth supports all combinations of the prefixes @code{0 u d d0 du f f0}
2184: (or none) and the comparisons @code{= <> < > <= >=}. Only a part of
2185: these combinations are standard (for details see the standard,
2186: @ref{Numeric comparison}, @ref{Floating Point} or @ref{Word Index}).
1.48 anton 2187:
2188: You can use @code{and or xor invert} can be used as operations on
2189: canonical flags. Actually they are bitwise operations:
2190:
2191: @example
2192: 1 2 and .
2193: 1 2 or .
2194: 1 3 xor .
2195: 1 invert .
2196: @end example
2197:
2198: You can convert a zero/non-zero flag into a canonical flag with
2199: @code{0<>} (and complement it on the way with @code{0=}).
2200:
2201: @example
2202: 1 0= .
2203: 1 0<> .
2204: @end example
2205:
1.65 anton 2206: You can use the all-bits-set feature of canonical flags and the bitwise
1.48 anton 2207: operation of the Boolean operations to avoid @code{if}s:
2208:
2209: @example
2210: : foo ( n1 -- n2 )
2211: 0= if
2212: 14
2213: else
2214: 0
2215: endif ;
2216: 0 foo .
2217: 1 foo .
2218:
2219: : foo ( n1 -- n2 )
2220: 0= 14 and ;
2221: 0 foo .
2222: 1 foo .
2223: @end example
2224:
2225: @assignment
2226: Write @code{min} without @code{if}.
2227: @endassignment
2228:
1.66 anton 2229: For reference, see @ref{Boolean Flags}, @ref{Numeric comparison}, and
2230: @ref{Bitwise operations}.
2231:
1.48 anton 2232:
2233: @node General Loops Tutorial, Counted loops Tutorial, Flags and Comparisons Tutorial, Tutorial
2234: @section General Loops
1.66 anton 2235: @cindex loops, indefinite, tutorial
1.48 anton 2236:
2237: The endless loop is the most simple one:
2238:
2239: @example
2240: : endless ( -- )
2241: 0 begin
2242: dup . 1+
2243: again ;
2244: endless
2245: @end example
2246:
2247: Terminate this loop by pressing @kbd{Ctrl-C} (in Gforth). @code{begin}
2248: does nothing at run-time, @code{again} jumps back to @code{begin}.
2249:
2250: A loop with one exit at any place looks like this:
2251:
2252: @example
2253: : log2 ( +n1 -- n2 )
2254: \ logarithmus dualis of n1>0, rounded down to the next integer
2255: assert( dup 0> )
2256: 2/ 0 begin
2257: over 0> while
2258: 1+ swap 2/ swap
2259: repeat
2260: nip ;
2261: 7 log2 .
2262: 8 log2 .
2263: @end example
2264:
2265: At run-time @code{while} consumes a flag; if it is 0, execution
1.51 pazsan 2266: continues behind the @code{repeat}; if the flag is non-zero, execution
1.48 anton 2267: continues behind the @code{while}. @code{Repeat} jumps back to
2268: @code{begin}, just like @code{again}.
2269:
2270: In Forth there are many combinations/abbreviations, like @code{1+}.
2271: However, @code{2/} is not one of them; it shifts it's argument right by
2272: one bit (arithmetic shift right):
2273:
2274: @example
2275: -5 2 / .
2276: -5 2/ .
2277: @end example
2278:
2279: @code{assert(} is no standard word, but you can get it on systems other
2280: then Gforth by including @file{compat/assert.fs}. You can see what it
2281: does by trying
2282:
2283: @example
2284: 0 log2 .
2285: @end example
2286:
2287: Here's a loop with an exit at the end:
2288:
2289: @example
2290: : log2 ( +n1 -- n2 )
2291: \ logarithmus dualis of n1>0, rounded down to the next integer
2292: assert( dup 0 > )
2293: -1 begin
2294: 1+ swap 2/ swap
2295: over 0 <=
2296: until
2297: nip ;
2298: @end example
2299:
2300: @code{Until} consumes a flag; if it is non-zero, execution continues at
2301: the @code{begin}, otherwise after the @code{until}.
2302:
2303: @assignment
2304: Write a definition for computing the greatest common divisor.
2305: @endassignment
2306:
1.66 anton 2307: Reference: @ref{Simple Loops}.
2308:
1.48 anton 2309:
2310: @node Counted loops Tutorial, Recursion Tutorial, General Loops Tutorial, Tutorial
2311: @section Counted loops
1.66 anton 2312: @cindex loops, counted, tutorial
1.48 anton 2313:
2314: @example
2315: : ^ ( n1 u -- n )
2316: \ n = the uth power of u1
2317: 1 swap 0 u+do
2318: over *
2319: loop
2320: nip ;
2321: 3 2 ^ .
2322: 4 3 ^ .
2323: @end example
2324:
2325: @code{U+do} (from @file{compat/loops.fs}, if your Forth system doesn't
2326: have it) takes two numbers of the stack @code{( u3 u4 -- )}, and then
2327: performs the code between @code{u+do} and @code{loop} for @code{u3-u4}
2328: times (or not at all, if @code{u3-u4<0}).
2329:
2330: You can see the stack effect design rules at work in the stack effect of
2331: the loop start words: Since the start value of the loop is more
2332: frequently constant than the end value, the start value is passed on
2333: the top-of-stack.
2334:
2335: You can access the counter of a counted loop with @code{i}:
2336:
2337: @example
2338: : fac ( u -- u! )
2339: 1 swap 1+ 1 u+do
2340: i *
2341: loop ;
2342: 5 fac .
2343: 7 fac .
2344: @end example
2345:
2346: There is also @code{+do}, which expects signed numbers (important for
2347: deciding whether to enter the loop).
2348:
2349: @assignment
2350: Write a definition for computing the nth Fibonacci number.
2351: @endassignment
2352:
1.65 anton 2353: You can also use increments other than 1:
2354:
2355: @example
2356: : up2 ( n1 n2 -- )
2357: +do
2358: i .
2359: 2 +loop ;
2360: 10 0 up2
2361:
2362: : down2 ( n1 n2 -- )
2363: -do
2364: i .
2365: 2 -loop ;
2366: 0 10 down2
2367: @end example
1.48 anton 2368:
1.66 anton 2369: Reference: @ref{Counted Loops}.
2370:
1.48 anton 2371:
2372: @node Recursion Tutorial, Leaving definitions or loops Tutorial, Counted loops Tutorial, Tutorial
2373: @section Recursion
1.66 anton 2374: @cindex recursion tutorial
1.48 anton 2375:
2376: Usually the name of a definition is not visible in the definition; but
2377: earlier definitions are usually visible:
2378:
2379: @example
2380: 1 0 / . \ "Floating-point unidentified fault" in Gforth on most platforms
2381: : / ( n1 n2 -- n )
2382: dup 0= if
2383: -10 throw \ report division by zero
2384: endif
2385: / \ old version
2386: ;
2387: 1 0 /
2388: @end example
2389:
2390: For recursive definitions you can use @code{recursive} (non-standard) or
2391: @code{recurse}:
2392:
2393: @example
2394: : fac1 ( n -- n! ) recursive
2395: dup 0> if
2396: dup 1- fac1 *
2397: else
2398: drop 1
2399: endif ;
2400: 7 fac1 .
2401:
2402: : fac2 ( n -- n! )
2403: dup 0> if
2404: dup 1- recurse *
2405: else
2406: drop 1
2407: endif ;
2408: 8 fac2 .
2409: @end example
2410:
2411: @assignment
2412: Write a recursive definition for computing the nth Fibonacci number.
2413: @endassignment
2414:
1.66 anton 2415: Reference (including indirect recursion): @xref{Calls and returns}.
2416:
1.48 anton 2417:
2418: @node Leaving definitions or loops Tutorial, Return Stack Tutorial, Recursion Tutorial, Tutorial
2419: @section Leaving definitions or loops
1.66 anton 2420: @cindex leaving definitions, tutorial
2421: @cindex leaving loops, tutorial
1.48 anton 2422:
2423: @code{EXIT} exits the current definition right away. For every counted
2424: loop that is left in this way, an @code{UNLOOP} has to be performed
2425: before the @code{EXIT}:
2426:
2427: @c !! real examples
2428: @example
2429: : ...
2430: ... u+do
2431: ... if
2432: ... unloop exit
2433: endif
2434: ...
2435: loop
2436: ... ;
2437: @end example
2438:
2439: @code{LEAVE} leaves the innermost counted loop right away:
2440:
2441: @example
2442: : ...
2443: ... u+do
2444: ... if
2445: ... leave
2446: endif
2447: ...
2448: loop
2449: ... ;
2450: @end example
2451:
1.65 anton 2452: @c !! example
1.48 anton 2453:
1.66 anton 2454: Reference: @ref{Calls and returns}, @ref{Counted Loops}.
2455:
2456:
1.48 anton 2457: @node Return Stack Tutorial, Memory Tutorial, Leaving definitions or loops Tutorial, Tutorial
2458: @section Return Stack
1.66 anton 2459: @cindex return stack tutorial
1.48 anton 2460:
2461: In addition to the data stack Forth also has a second stack, the return
2462: stack; most Forth systems store the return addresses of procedure calls
2463: there (thus its name). Programmers can also use this stack:
2464:
2465: @example
2466: : foo ( n1 n2 -- )
2467: .s
2468: >r .s
1.50 anton 2469: r@@ .
1.48 anton 2470: >r .s
1.50 anton 2471: r@@ .
1.48 anton 2472: r> .
1.50 anton 2473: r@@ .
1.48 anton 2474: r> . ;
2475: 1 2 foo
2476: @end example
2477:
2478: @code{>r} takes an element from the data stack and pushes it onto the
2479: return stack; conversely, @code{r>} moves an elementm from the return to
2480: the data stack; @code{r@@} pushes a copy of the top of the return stack
2481: on the return stack.
2482:
2483: Forth programmers usually use the return stack for storing data
2484: temporarily, if using the data stack alone would be too complex, and
2485: factoring and locals are not an option:
2486:
2487: @example
2488: : 2swap ( x1 x2 x3 x4 -- x3 x4 x1 x2 )
2489: rot >r rot r> ;
2490: @end example
2491:
2492: The return address of the definition and the loop control parameters of
2493: counted loops usually reside on the return stack, so you have to take
2494: all items, that you have pushed on the return stack in a colon
2495: definition or counted loop, from the return stack before the definition
2496: or loop ends. You cannot access items that you pushed on the return
2497: stack outside some definition or loop within the definition of loop.
2498:
2499: If you miscount the return stack items, this usually ends in a crash:
2500:
2501: @example
2502: : crash ( n -- )
2503: >r ;
2504: 5 crash
2505: @end example
2506:
2507: You cannot mix using locals and using the return stack (according to the
2508: standard; Gforth has no problem). However, they solve the same
2509: problems, so this shouldn't be an issue.
2510:
2511: @assignment
2512: Can you rewrite any of the definitions you wrote until now in a better
2513: way using the return stack?
2514: @endassignment
2515:
1.66 anton 2516: Reference: @ref{Return stack}.
2517:
1.48 anton 2518:
2519: @node Memory Tutorial, Characters and Strings Tutorial, Return Stack Tutorial, Tutorial
2520: @section Memory
1.66 anton 2521: @cindex memory access/allocation tutorial
1.48 anton 2522:
2523: You can create a global variable @code{v} with
2524:
2525: @example
2526: variable v ( -- addr )
2527: @end example
2528:
2529: @code{v} pushes the address of a cell in memory on the stack. This cell
2530: was reserved by @code{variable}. You can use @code{!} (store) to store
2531: values into this cell and @code{@@} (fetch) to load the value from the
2532: stack into memory:
2533:
2534: @example
2535: v .
2536: 5 v ! .s
1.50 anton 2537: v @@ .
1.48 anton 2538: @end example
2539:
1.65 anton 2540: You can see a raw dump of memory with @code{dump}:
2541:
2542: @example
2543: v 1 cells .s dump
2544: @end example
2545:
2546: @code{Cells ( n1 -- n2 )} gives you the number of bytes (or, more
2547: generally, address units (aus)) that @code{n1 cells} occupy. You can
2548: also reserve more memory:
1.48 anton 2549:
2550: @example
2551: create v2 20 cells allot
1.65 anton 2552: v2 20 cells dump
1.48 anton 2553: @end example
2554:
1.65 anton 2555: creates a word @code{v2} and reserves 20 uninitialized cells; the
2556: address pushed by @code{v2} points to the start of these 20 cells. You
2557: can use address arithmetic to access these cells:
1.48 anton 2558:
2559: @example
2560: 3 v2 5 cells + !
1.65 anton 2561: v2 20 cells dump
1.48 anton 2562: @end example
2563:
2564: You can reserve and initialize memory with @code{,}:
2565:
2566: @example
2567: create v3
2568: 5 , 4 , 3 , 2 , 1 ,
1.50 anton 2569: v3 @@ .
2570: v3 cell+ @@ .
2571: v3 2 cells + @@ .
1.65 anton 2572: v3 5 cells dump
1.48 anton 2573: @end example
2574:
2575: @assignment
2576: Write a definition @code{vsum ( addr u -- n )} that computes the sum of
2577: @code{u} cells, with the first of these cells at @code{addr}, the next
2578: one at @code{addr cell+} etc.
2579: @endassignment
2580:
2581: You can also reserve memory without creating a new word:
2582:
2583: @example
1.60 anton 2584: here 10 cells allot .
2585: here .
1.48 anton 2586: @end example
2587:
2588: @code{Here} pushes the start address of the memory area. You should
2589: store it somewhere, or you will have a hard time finding the memory area
2590: again.
2591:
2592: @code{Allot} manages dictionary memory. The dictionary memory contains
2593: the system's data structures for words etc. on Gforth and most other
2594: Forth systems. It is managed like a stack: You can free the memory that
2595: you have just @code{allot}ed with
2596:
2597: @example
2598: -10 cells allot
1.60 anton 2599: here .
1.48 anton 2600: @end example
2601:
2602: Note that you cannot do this if you have created a new word in the
2603: meantime (because then your @code{allot}ed memory is no longer on the
2604: top of the dictionary ``stack'').
2605:
2606: Alternatively, you can use @code{allocate} and @code{free} which allow
2607: freeing memory in any order:
2608:
2609: @example
2610: 10 cells allocate throw .s
2611: 20 cells allocate throw .s
2612: swap
2613: free throw
2614: free throw
2615: @end example
2616:
2617: The @code{throw}s deal with errors (e.g., out of memory).
2618:
1.65 anton 2619: And there is also a
2620: @uref{http://www.complang.tuwien.ac.at/forth/garbage-collection.zip,
2621: garbage collector}, which eliminates the need to @code{free} memory
2622: explicitly.
1.48 anton 2623:
1.66 anton 2624: Reference: @ref{Memory}.
2625:
1.48 anton 2626:
2627: @node Characters and Strings Tutorial, Alignment Tutorial, Memory Tutorial, Tutorial
2628: @section Characters and Strings
1.66 anton 2629: @cindex strings tutorial
2630: @cindex characters tutorial
1.48 anton 2631:
2632: On the stack characters take up a cell, like numbers. In memory they
2633: have their own size (one 8-bit byte on most systems), and therefore
2634: require their own words for memory access:
2635:
2636: @example
2637: create v4
2638: 104 c, 97 c, 108 c, 108 c, 111 c,
1.50 anton 2639: v4 4 chars + c@@ .
1.65 anton 2640: v4 5 chars dump
1.48 anton 2641: @end example
2642:
2643: The preferred representation of strings on the stack is @code{addr
2644: u-count}, where @code{addr} is the address of the first character and
2645: @code{u-count} is the number of characters in the string.
2646:
2647: @example
2648: v4 5 type
2649: @end example
2650:
2651: You get a string constant with
2652:
2653: @example
2654: s" hello, world" .s
2655: type
2656: @end example
2657:
2658: Make sure you have a space between @code{s"} and the string; @code{s"}
2659: is a normal Forth word and must be delimited with white space (try what
2660: happens when you remove the space).
2661:
2662: However, this interpretive use of @code{s"} is quite restricted: the
2663: string exists only until the next call of @code{s"} (some Forth systems
2664: keep more than one of these strings, but usually they still have a
1.62 crook 2665: limited lifetime).
1.48 anton 2666:
2667: @example
2668: s" hello," s" world" .s
2669: type
2670: type
2671: @end example
2672:
1.62 crook 2673: You can also use @code{s"} in a definition, and the resulting
2674: strings then live forever (well, for as long as the definition):
1.48 anton 2675:
2676: @example
2677: : foo s" hello," s" world" ;
2678: foo .s
2679: type
2680: type
2681: @end example
2682:
2683: @assignment
2684: @code{Emit ( c -- )} types @code{c} as character (not a number).
2685: Implement @code{type ( addr u -- )}.
2686: @endassignment
2687:
1.66 anton 2688: Reference: @ref{Memory Blocks}.
2689:
2690:
1.48 anton 2691: @node Alignment Tutorial, Interpretation and Compilation Semantics and Immediacy Tutorial, Characters and Strings Tutorial, Tutorial
2692: @section Alignment
1.66 anton 2693: @cindex alignment tutorial
2694: @cindex memory alignment tutorial
1.48 anton 2695:
2696: On many processors cells have to be aligned in memory, if you want to
2697: access them with @code{@@} and @code{!} (and even if the processor does
1.62 crook 2698: not require alignment, access to aligned cells is faster).
1.48 anton 2699:
2700: @code{Create} aligns @code{here} (i.e., the place where the next
2701: allocation will occur, and that the @code{create}d word points to).
2702: Likewise, the memory produced by @code{allocate} starts at an aligned
2703: address. Adding a number of @code{cells} to an aligned address produces
2704: another aligned address.
2705:
2706: However, address arithmetic involving @code{char+} and @code{chars} can
2707: create an address that is not cell-aligned. @code{Aligned ( addr --
2708: a-addr )} produces the next aligned address:
2709:
2710: @example
1.50 anton 2711: v3 char+ aligned .s @@ .
2712: v3 char+ .s @@ .
1.48 anton 2713: @end example
2714:
2715: Similarly, @code{align} advances @code{here} to the next aligned
2716: address:
2717:
2718: @example
2719: create v5 97 c,
2720: here .
2721: align here .
2722: 1000 ,
2723: @end example
2724:
2725: Note that you should use aligned addresses even if your processor does
2726: not require them, if you want your program to be portable.
2727:
1.66 anton 2728: Reference: @ref{Address arithmetic}.
2729:
1.48 anton 2730:
2731: @node Interpretation and Compilation Semantics and Immediacy Tutorial, Execution Tokens Tutorial, Alignment Tutorial, Tutorial
2732: @section Interpretation and Compilation Semantics and Immediacy
1.66 anton 2733: @cindex semantics tutorial
2734: @cindex interpretation semantics tutorial
2735: @cindex compilation semantics tutorial
2736: @cindex immediate, tutorial
1.48 anton 2737:
2738: When a word is compiled, it behaves differently from being interpreted.
2739: E.g., consider @code{+}:
2740:
2741: @example
2742: 1 2 + .
2743: : foo + ;
2744: @end example
2745:
2746: These two behaviours are known as compilation and interpretation
2747: semantics. For normal words (e.g., @code{+}), the compilation semantics
2748: is to append the interpretation semantics to the currently defined word
2749: (@code{foo} in the example above). I.e., when @code{foo} is executed
2750: later, the interpretation semantics of @code{+} (i.e., adding two
2751: numbers) will be performed.
2752:
2753: However, there are words with non-default compilation semantics, e.g.,
2754: the control-flow words like @code{if}. You can use @code{immediate} to
2755: change the compilation semantics of the last defined word to be equal to
2756: the interpretation semantics:
2757:
2758: @example
2759: : [FOO] ( -- )
2760: 5 . ; immediate
2761:
2762: [FOO]
2763: : bar ( -- )
2764: [FOO] ;
2765: bar
2766: see bar
2767: @end example
2768:
2769: Two conventions to mark words with non-default compilation semnatics are
2770: names with brackets (more frequently used) and to write them all in
2771: upper case (less frequently used).
2772:
2773: In Gforth (and many other systems) you can also remove the
2774: interpretation semantics with @code{compile-only} (the compilation
2775: semantics is derived from the original interpretation semantics):
2776:
2777: @example
2778: : flip ( -- )
2779: 6 . ; compile-only \ but not immediate
2780: flip
2781:
2782: : flop ( -- )
2783: flip ;
2784: flop
2785: @end example
2786:
2787: In this example the interpretation semantics of @code{flop} is equal to
2788: the original interpretation semantics of @code{flip}.
2789:
2790: The text interpreter has two states: in interpret state, it performs the
2791: interpretation semantics of words it encounters; in compile state, it
2792: performs the compilation semantics of these words.
2793:
2794: Among other things, @code{:} switches into compile state, and @code{;}
2795: switches back to interpret state. They contain the factors @code{]}
2796: (switch to compile state) and @code{[} (switch to interpret state), that
2797: do nothing but switch the state.
2798:
2799: @example
2800: : xxx ( -- )
2801: [ 5 . ]
2802: ;
2803:
2804: xxx
2805: see xxx
2806: @end example
2807:
2808: These brackets are also the source of the naming convention mentioned
2809: above.
2810:
1.66 anton 2811: Reference: @ref{Interpretation and Compilation Semantics}.
2812:
1.48 anton 2813:
2814: @node Execution Tokens Tutorial, Exceptions Tutorial, Interpretation and Compilation Semantics and Immediacy Tutorial, Tutorial
2815: @section Execution Tokens
1.66 anton 2816: @cindex execution tokens tutorial
2817: @cindex XT tutorial
1.48 anton 2818:
2819: @code{' word} gives you the execution token (XT) of a word. The XT is a
2820: cell representing the interpretation semantics of a word. You can
2821: execute this semantics with @code{execute}:
2822:
2823: @example
2824: ' + .s
2825: 1 2 rot execute .
2826: @end example
2827:
2828: The XT is similar to a function pointer in C. However, parameter
2829: passing through the stack makes it a little more flexible:
2830:
2831: @example
2832: : map-array ( ... addr u xt -- ... )
1.50 anton 2833: \ executes xt ( ... x -- ... ) for every element of the array starting
2834: \ at addr and containing u elements
1.48 anton 2835: @{ xt @}
2836: cells over + swap ?do
1.50 anton 2837: i @@ xt execute
1.48 anton 2838: 1 cells +loop ;
2839:
2840: create a 3 , 4 , 2 , -1 , 4 ,
2841: a 5 ' . map-array .s
2842: 0 a 5 ' + map-array .
2843: s" max-n" environment? drop .s
2844: a 5 ' min map-array .
2845: @end example
2846:
2847: You can use map-array with the XTs of words that consume one element
2848: more than they produce. In theory you can also use it with other XTs,
2849: but the stack effect then depends on the size of the array, which is
2850: hard to understand.
2851:
1.51 pazsan 2852: Since XTs are cell-sized, you can store them in memory and manipulate
2853: them on the stack like other cells. You can also compile the XT into a
1.48 anton 2854: word with @code{compile,}:
2855:
2856: @example
2857: : foo1 ( n1 n2 -- n )
2858: [ ' + compile, ] ;
2859: see foo
2860: @end example
2861:
2862: This is non-standard, because @code{compile,} has no compilation
2863: semantics in the standard, but it works in good Forth systems. For the
2864: broken ones, use
2865:
2866: @example
2867: : [compile,] compile, ; immediate
2868:
2869: : foo1 ( n1 n2 -- n )
2870: [ ' + ] [compile,] ;
2871: see foo
2872: @end example
2873:
2874: @code{'} is a word with default compilation semantics; it parses the
2875: next word when its interpretation semantics are executed, not during
2876: compilation:
2877:
2878: @example
2879: : foo ( -- xt )
2880: ' ;
2881: see foo
2882: : bar ( ... "word" -- ... )
2883: ' execute ;
2884: see bar
1.60 anton 2885: 1 2 bar + .
1.48 anton 2886: @end example
2887:
2888: You often want to parse a word during compilation and compile its XT so
2889: it will be pushed on the stack at run-time. @code{[']} does this:
2890:
2891: @example
2892: : xt-+ ( -- xt )
2893: ['] + ;
2894: see xt-+
2895: 1 2 xt-+ execute .
2896: @end example
2897:
2898: Many programmers tend to see @code{'} and the word it parses as one
2899: unit, and expect it to behave like @code{[']} when compiled, and are
2900: confused by the actual behaviour. If you are, just remember that the
2901: Forth system just takes @code{'} as one unit and has no idea that it is
2902: a parsing word (attempts to convenience programmers in this issue have
2903: usually resulted in even worse pitfalls, see
1.66 anton 2904: @uref{http://www.complang.tuwien.ac.at/papers/ertl98.ps.gz,
2905: @code{State}-smartness---Why it is evil and How to Exorcise it}).
1.48 anton 2906:
2907: Note that the state of the interpreter does not come into play when
1.51 pazsan 2908: creating and executing XTs. I.e., even when you execute @code{'} in
1.48 anton 2909: compile state, it still gives you the interpretation semantics. And
2910: whatever that state is, @code{execute} performs the semantics
1.66 anton 2911: represented by the XT (i.e., for XTs produced with @code{'} the
2912: interpretation semantics).
2913:
2914: Reference: @ref{Tokens for Words}.
1.48 anton 2915:
2916:
2917: @node Exceptions Tutorial, Defining Words Tutorial, Execution Tokens Tutorial, Tutorial
2918: @section Exceptions
1.66 anton 2919: @cindex exceptions tutorial
1.48 anton 2920:
2921: @code{throw ( n -- )} causes an exception unless n is zero.
2922:
2923: @example
2924: 100 throw .s
2925: 0 throw .s
2926: @end example
2927:
2928: @code{catch ( ... xt -- ... n )} behaves similar to @code{execute}, but
2929: it catches exceptions and pushes the number of the exception on the
2930: stack (or 0, if the xt executed without exception). If there was an
2931: exception, the stacks have the same depth as when entering @code{catch}:
2932:
2933: @example
2934: .s
2935: 3 0 ' / catch .s
2936: 3 2 ' / catch .s
2937: @end example
2938:
2939: @assignment
2940: Try the same with @code{execute} instead of @code{catch}.
2941: @endassignment
2942:
2943: @code{Throw} always jumps to the dynamically next enclosing
2944: @code{catch}, even if it has to leave several call levels to achieve
2945: this:
2946:
2947: @example
2948: : foo 100 throw ;
2949: : foo1 foo ." after foo" ;
1.51 pazsan 2950: : bar ['] foo1 catch ;
1.60 anton 2951: bar .
1.48 anton 2952: @end example
2953:
2954: It is often important to restore a value upon leaving a definition, even
2955: if the definition is left through an exception. You can ensure this
2956: like this:
2957:
2958: @example
2959: : ...
2960: save-x
1.51 pazsan 2961: ['] word-changing-x catch ( ... n )
1.48 anton 2962: restore-x
2963: ( ... n ) throw ;
2964: @end example
2965:
1.55 anton 2966: Gforth provides an alternative syntax in addition to @code{catch}:
1.48 anton 2967: @code{try ... recover ... endtry}. If the code between @code{try} and
2968: @code{recover} has an exception, the stack depths are restored, the
2969: exception number is pushed on the stack, and the code between
2970: @code{recover} and @code{endtry} is performed. E.g., the definition for
2971: @code{catch} is
2972:
2973: @example
2974: : catch ( x1 .. xn xt -- y1 .. ym 0 / z1 .. zn error ) \ exception
2975: try
2976: execute 0
2977: recover
2978: nip
2979: endtry ;
2980: @end example
2981:
2982: The equivalent to the restoration code above is
2983:
2984: @example
2985: : ...
2986: save-x
2987: try
2988: word-changing-x
2989: end-try
2990: restore-x
2991: throw ;
2992: @end example
2993:
2994: As you can see, the @code{recover} part is optional.
2995:
1.66 anton 2996: Reference: @ref{Exception Handling}.
2997:
1.48 anton 2998:
2999: @node Defining Words Tutorial, Arrays and Records Tutorial, Exceptions Tutorial, Tutorial
3000: @section Defining Words
1.66 anton 3001: @cindex defining words tutorial
3002: @cindex does> tutorial
3003: @cindex create...does> tutorial
3004:
3005: @c before semantics?
1.48 anton 3006:
3007: @code{:}, @code{create}, and @code{variable} are definition words: They
3008: define other words. @code{Constant} is another definition word:
3009:
3010: @example
3011: 5 constant foo
3012: foo .
3013: @end example
3014:
3015: You can also use the prefixes @code{2} (double-cell) and @code{f}
3016: (floating point) with @code{variable} and @code{constant}.
3017:
3018: You can also define your own defining words. E.g.:
3019:
3020: @example
3021: : variable ( "name" -- )
3022: create 0 , ;
3023: @end example
3024:
3025: You can also define defining words that create words that do something
3026: other than just producing their address:
3027:
3028: @example
3029: : constant ( n "name" -- )
3030: create ,
3031: does> ( -- n )
1.50 anton 3032: ( addr ) @@ ;
1.48 anton 3033:
3034: 5 constant foo
3035: foo .
3036: @end example
3037:
3038: The definition of @code{constant} above ends at the @code{does>}; i.e.,
3039: @code{does>} replaces @code{;}, but it also does something else: It
3040: changes the last defined word such that it pushes the address of the
3041: body of the word and then performs the code after the @code{does>}
3042: whenever it is called.
3043:
3044: In the example above, @code{constant} uses @code{,} to store 5 into the
3045: body of @code{foo}. When @code{foo} executes, it pushes the address of
3046: the body onto the stack, then (in the code after the @code{does>})
3047: fetches the 5 from there.
3048:
3049: The stack comment near the @code{does>} reflects the stack effect of the
3050: defined word, not the stack effect of the code after the @code{does>}
3051: (the difference is that the code expects the address of the body that
3052: the stack comment does not show).
3053:
3054: You can use these definition words to do factoring in cases that involve
3055: (other) definition words. E.g., a field offset is always added to an
3056: address. Instead of defining
3057:
3058: @example
3059: 2 cells constant offset-field1
3060: @end example
3061:
3062: and using this like
3063:
3064: @example
3065: ( addr ) offset-field1 +
3066: @end example
3067:
3068: you can define a definition word
3069:
3070: @example
3071: : simple-field ( n "name" -- )
3072: create ,
3073: does> ( n1 -- n1+n )
1.50 anton 3074: ( addr ) @@ + ;
1.48 anton 3075: @end example
1.21 crook 3076:
1.48 anton 3077: Definition and use of field offsets now look like this:
1.21 crook 3078:
1.48 anton 3079: @example
3080: 2 cells simple-field field1
1.60 anton 3081: create mystruct 4 cells allot
3082: mystruct .s field1 .s drop
1.48 anton 3083: @end example
1.21 crook 3084:
1.48 anton 3085: If you want to do something with the word without performing the code
3086: after the @code{does>}, you can access the body of a @code{create}d word
3087: with @code{>body ( xt -- addr )}:
1.21 crook 3088:
1.48 anton 3089: @example
3090: : value ( n "name" -- )
3091: create ,
3092: does> ( -- n1 )
1.50 anton 3093: @@ ;
1.48 anton 3094: : to ( n "name" -- )
3095: ' >body ! ;
1.21 crook 3096:
1.48 anton 3097: 5 value foo
3098: foo .
3099: 7 to foo
3100: foo .
3101: @end example
1.21 crook 3102:
1.48 anton 3103: @assignment
3104: Define @code{defer ( "name" -- )}, which creates a word that stores an
3105: XT (at the start the XT of @code{abort}), and upon execution
3106: @code{execute}s the XT. Define @code{is ( xt "name" -- )} that stores
3107: @code{xt} into @code{name}, a word defined with @code{defer}. Indirect
3108: recursion is one application of @code{defer}.
3109: @endassignment
1.29 crook 3110:
1.66 anton 3111: Reference: @ref{User-defined Defining Words}.
3112:
3113:
1.48 anton 3114: @node Arrays and Records Tutorial, POSTPONE Tutorial, Defining Words Tutorial, Tutorial
3115: @section Arrays and Records
1.66 anton 3116: @cindex arrays tutorial
3117: @cindex records tutorial
3118: @cindex structs tutorial
1.29 crook 3119:
1.48 anton 3120: Forth has no standard words for defining data structures such as arrays
3121: and records (structs in C terminology), but you can build them yourself
3122: based on address arithmetic. You can also define words for defining
3123: arrays and records (@pxref{Defining Words Tutorial,, Defining Words}).
1.29 crook 3124:
1.48 anton 3125: One of the first projects a Forth newcomer sets out upon when learning
3126: about defining words is an array defining word (possibly for
3127: n-dimensional arrays). Go ahead and do it, I did it, too; you will
3128: learn something from it. However, don't be disappointed when you later
3129: learn that you have little use for these words (inappropriate use would
3130: be even worse). I have not yet found a set of useful array words yet;
3131: the needs are just too diverse, and named, global arrays (the result of
3132: naive use of defining words) are often not flexible enough (e.g.,
1.66 anton 3133: consider how to pass them as parameters). Another such project is a set
3134: of words to help dealing with strings.
1.29 crook 3135:
1.48 anton 3136: On the other hand, there is a useful set of record words, and it has
3137: been defined in @file{compat/struct.fs}; these words are predefined in
3138: Gforth. They are explained in depth elsewhere in this manual (see
3139: @pxref{Structures}). The @code{simple-field} example above is
3140: simplified variant of fields in this package.
1.21 crook 3141:
3142:
1.48 anton 3143: @node POSTPONE Tutorial, Literal Tutorial, Arrays and Records Tutorial, Tutorial
3144: @section @code{POSTPONE}
1.66 anton 3145: @cindex postpone tutorial
1.21 crook 3146:
1.48 anton 3147: You can compile the compilation semantics (instead of compiling the
3148: interpretation semantics) of a word with @code{POSTPONE}:
1.21 crook 3149:
1.48 anton 3150: @example
3151: : MY-+ ( Compilation: -- ; Run-time of compiled code: n1 n2 -- n )
1.51 pazsan 3152: POSTPONE + ; immediate
1.48 anton 3153: : foo ( n1 n2 -- n )
3154: MY-+ ;
3155: 1 2 foo .
3156: see foo
3157: @end example
1.21 crook 3158:
1.48 anton 3159: During the definition of @code{foo} the text interpreter performs the
3160: compilation semantics of @code{MY-+}, which performs the compilation
3161: semantics of @code{+}, i.e., it compiles @code{+} into @code{foo}.
3162:
3163: This example also displays separate stack comments for the compilation
3164: semantics and for the stack effect of the compiled code. For words with
3165: default compilation semantics these stack effects are usually not
3166: displayed; the stack effect of the compilation semantics is always
3167: @code{( -- )} for these words, the stack effect for the compiled code is
3168: the stack effect of the interpretation semantics.
3169:
3170: Note that the state of the interpreter does not come into play when
3171: performing the compilation semantics in this way. You can also perform
3172: it interpretively, e.g.:
3173:
3174: @example
3175: : foo2 ( n1 n2 -- n )
3176: [ MY-+ ] ;
3177: 1 2 foo .
3178: see foo
3179: @end example
1.21 crook 3180:
1.48 anton 3181: However, there are some broken Forth systems where this does not always
1.62 crook 3182: work, and therefore this practice was been declared non-standard in
1.48 anton 3183: 1999.
3184: @c !! repair.fs
3185:
3186: Here is another example for using @code{POSTPONE}:
1.44 crook 3187:
1.48 anton 3188: @example
3189: : MY-- ( Compilation: -- ; Run-time of compiled code: n1 n2 -- n )
3190: POSTPONE negate POSTPONE + ; immediate compile-only
3191: : bar ( n1 n2 -- n )
3192: MY-- ;
3193: 2 1 bar .
3194: see bar
3195: @end example
1.21 crook 3196:
1.48 anton 3197: You can define @code{ENDIF} in this way:
1.21 crook 3198:
1.48 anton 3199: @example
3200: : ENDIF ( Compilation: orig -- )
3201: POSTPONE then ; immediate
3202: @end example
1.21 crook 3203:
1.48 anton 3204: @assignment
3205: Write @code{MY-2DUP} that has compilation semantics equivalent to
3206: @code{2dup}, but compiles @code{over over}.
3207: @endassignment
1.29 crook 3208:
1.66 anton 3209: @c !! @xref{Macros} for reference
3210:
3211:
1.48 anton 3212: @node Literal Tutorial, Advanced macros Tutorial, POSTPONE Tutorial, Tutorial
3213: @section @code{Literal}
1.66 anton 3214: @cindex literal tutorial
1.29 crook 3215:
1.48 anton 3216: You cannot @code{POSTPONE} numbers:
1.21 crook 3217:
1.48 anton 3218: @example
3219: : [FOO] POSTPONE 500 ; immediate
1.21 crook 3220: @end example
3221:
1.48 anton 3222: Instead, you can use @code{LITERAL (compilation: n --; run-time: -- n )}:
1.29 crook 3223:
1.48 anton 3224: @example
3225: : [FOO] ( compilation: --; run-time: -- n )
3226: 500 POSTPONE literal ; immediate
1.29 crook 3227:
1.60 anton 3228: : flip [FOO] ;
1.48 anton 3229: flip .
3230: see flip
3231: @end example
1.29 crook 3232:
1.48 anton 3233: @code{LITERAL} consumes a number at compile-time (when it's compilation
3234: semantics are executed) and pushes it at run-time (when the code it
3235: compiled is executed). A frequent use of @code{LITERAL} is to compile a
3236: number computed at compile time into the current word:
1.29 crook 3237:
1.48 anton 3238: @example
3239: : bar ( -- n )
3240: [ 2 2 + ] literal ;
3241: see bar
3242: @end example
1.29 crook 3243:
1.48 anton 3244: @assignment
3245: Write @code{]L} which allows writing the example above as @code{: bar (
3246: -- n ) [ 2 2 + ]L ;}
3247: @endassignment
3248:
1.66 anton 3249: @c !! @xref{Macros} for reference
3250:
1.48 anton 3251:
3252: @node Advanced macros Tutorial, Compilation Tokens Tutorial, Literal Tutorial, Tutorial
3253: @section Advanced macros
1.66 anton 3254: @cindex macros, advanced tutorial
3255: @cindex run-time code generation, tutorial
1.48 anton 3256:
1.66 anton 3257: Reconsider @code{map-array} from @ref{Execution Tokens Tutorial,,
3258: Execution Tokens}. It frequently performs @code{execute}, a relatively
3259: expensive operation in some Forth implementations. You can use
1.48 anton 3260: @code{compile,} and @code{POSTPONE} to eliminate these @code{execute}s
3261: and produce a word that contains the word to be performed directly:
3262:
3263: @c use ]] ... [[
3264: @example
3265: : compile-map-array ( compilation: xt -- ; run-time: ... addr u -- ... )
3266: \ at run-time, execute xt ( ... x -- ... ) for each element of the
3267: \ array beginning at addr and containing u elements
3268: @{ xt @}
3269: POSTPONE cells POSTPONE over POSTPONE + POSTPONE swap POSTPONE ?do
1.50 anton 3270: POSTPONE i POSTPONE @@ xt compile,
1.48 anton 3271: 1 cells POSTPONE literal POSTPONE +loop ;
3272:
3273: : sum-array ( addr u -- n )
3274: 0 rot rot [ ' + compile-map-array ] ;
3275: see sum-array
3276: a 5 sum-array .
3277: @end example
3278:
3279: You can use the full power of Forth for generating the code; here's an
3280: example where the code is generated in a loop:
3281:
3282: @example
3283: : compile-vmul-step ( compilation: n --; run-time: n1 addr1 -- n2 addr2 )
3284: \ n2=n1+(addr1)*n, addr2=addr1+cell
1.50 anton 3285: POSTPONE tuck POSTPONE @@
1.48 anton 3286: POSTPONE literal POSTPONE * POSTPONE +
3287: POSTPONE swap POSTPONE cell+ ;
3288:
3289: : compile-vmul ( compilation: addr1 u -- ; run-time: addr2 -- n )
1.51 pazsan 3290: \ n=v1*v2 (inner product), where the v_i are represented as addr_i u
1.48 anton 3291: 0 postpone literal postpone swap
3292: [ ' compile-vmul-step compile-map-array ]
3293: postpone drop ;
3294: see compile-vmul
3295:
3296: : a-vmul ( addr -- n )
1.51 pazsan 3297: \ n=a*v, where v is a vector that's as long as a and starts at addr
1.48 anton 3298: [ a 5 compile-vmul ] ;
3299: see a-vmul
3300: a a-vmul .
3301: @end example
3302:
3303: This example uses @code{compile-map-array} to show off, but you could
1.66 anton 3304: also use @code{map-array} instead (try it now!).
1.48 anton 3305:
3306: You can use this technique for efficient multiplication of large
3307: matrices. In matrix multiplication, you multiply every line of one
3308: matrix with every column of the other matrix. You can generate the code
3309: for one line once, and use it for every column. The only downside of
3310: this technique is that it is cumbersome to recover the memory consumed
3311: by the generated code when you are done (and in more complicated cases
3312: it is not possible portably).
3313:
1.66 anton 3314: @c !! @xref{Macros} for reference
3315:
3316:
1.48 anton 3317: @node Compilation Tokens Tutorial, Wordlists and Search Order Tutorial, Advanced macros Tutorial, Tutorial
3318: @section Compilation Tokens
1.66 anton 3319: @cindex compilation tokens, tutorial
3320: @cindex CT, tutorial
1.48 anton 3321:
3322: This section is Gforth-specific. You can skip it.
3323:
3324: @code{' word compile,} compiles the interpretation semantics. For words
3325: with default compilation semantics this is the same as performing the
3326: compilation semantics. To represent the compilation semantics of other
3327: words (e.g., words like @code{if} that have no interpretation
3328: semantics), Gforth has the concept of a compilation token (CT,
3329: consisting of two cells), and words @code{comp'} and @code{[comp']}.
3330: You can perform the compilation semantics represented by a CT with
3331: @code{execute}:
1.29 crook 3332:
1.48 anton 3333: @example
3334: : foo2 ( n1 n2 -- n )
3335: [ comp' + execute ] ;
3336: see foo
3337: @end example
1.29 crook 3338:
1.48 anton 3339: You can compile the compilation semantics represented by a CT with
3340: @code{postpone,}:
1.30 anton 3341:
1.48 anton 3342: @example
3343: : foo3 ( -- )
3344: [ comp' + postpone, ] ;
3345: see foo3
3346: @end example
1.30 anton 3347:
1.51 pazsan 3348: @code{[ comp' word postpone, ]} is equivalent to @code{POSTPONE word}.
1.48 anton 3349: @code{comp'} is particularly useful for words that have no
3350: interpretation semantics:
1.29 crook 3351:
1.30 anton 3352: @example
1.48 anton 3353: ' if
1.60 anton 3354: comp' if .s 2drop
1.30 anton 3355: @end example
3356:
1.66 anton 3357: Reference: @ref{Tokens for Words}.
3358:
1.29 crook 3359:
1.48 anton 3360: @node Wordlists and Search Order Tutorial, , Compilation Tokens Tutorial, Tutorial
3361: @section Wordlists and Search Order
1.66 anton 3362: @cindex wordlists tutorial
3363: @cindex search order, tutorial
1.48 anton 3364:
3365: The dictionary is not just a memory area that allows you to allocate
3366: memory with @code{allot}, it also contains the Forth words, arranged in
3367: several wordlists. When searching for a word in a wordlist,
3368: conceptually you start searching at the youngest and proceed towards
3369: older words (in reality most systems nowadays use hash-tables); i.e., if
3370: you define a word with the same name as an older word, the new word
3371: shadows the older word.
3372:
3373: Which wordlists are searched in which order is determined by the search
3374: order. You can display the search order with @code{order}. It displays
3375: first the search order, starting with the wordlist searched first, then
3376: it displays the wordlist that will contain newly defined words.
1.21 crook 3377:
1.48 anton 3378: You can create a new, empty wordlist with @code{wordlist ( -- wid )}:
1.21 crook 3379:
1.48 anton 3380: @example
3381: wordlist constant mywords
3382: @end example
1.21 crook 3383:
1.48 anton 3384: @code{Set-current ( wid -- )} sets the wordlist that will contain newly
3385: defined words (the @emph{current} wordlist):
1.21 crook 3386:
1.48 anton 3387: @example
3388: mywords set-current
3389: order
3390: @end example
1.26 crook 3391:
1.48 anton 3392: Gforth does not display a name for the wordlist in @code{mywords}
3393: because this wordlist was created anonymously with @code{wordlist}.
1.21 crook 3394:
1.48 anton 3395: You can get the current wordlist with @code{get-current ( -- wid)}. If
3396: you want to put something into a specific wordlist without overall
3397: effect on the current wordlist, this typically looks like this:
1.21 crook 3398:
1.48 anton 3399: @example
3400: get-current mywords set-current ( wid )
3401: create someword
3402: ( wid ) set-current
3403: @end example
1.21 crook 3404:
1.48 anton 3405: You can write the search order with @code{set-order ( wid1 .. widn n --
3406: )} and read it with @code{get-order ( -- wid1 .. widn n )}. The first
3407: searched wordlist is topmost.
1.21 crook 3408:
1.48 anton 3409: @example
3410: get-order mywords swap 1+ set-order
3411: order
3412: @end example
1.21 crook 3413:
1.48 anton 3414: Yes, the order of wordlists in the output of @code{order} is reversed
3415: from stack comments and the output of @code{.s} and thus unintuitive.
1.21 crook 3416:
1.48 anton 3417: @assignment
3418: Define @code{>order ( wid -- )} with adds @code{wid} as first searched
3419: wordlist to the search order. Define @code{previous ( -- )}, which
3420: removes the first searched wordlist from the search order. Experiment
3421: with boundary conditions (you will see some crashes or situations that
3422: are hard or impossible to leave).
3423: @endassignment
1.21 crook 3424:
1.48 anton 3425: The search order is a powerful foundation for providing features similar
3426: to Modula-2 modules and C++ namespaces. However, trying to modularize
3427: programs in this way has disadvantages for debugging and reuse/factoring
3428: that overcome the advantages in my experience (I don't do huge projects,
1.55 anton 3429: though). These disadvantages are not so clear in other
1.48 anton 3430: languages/programming environments, because these langauges are not so
3431: strong in debugging and reuse.
1.21 crook 3432:
1.66 anton 3433: @c !! example
3434:
3435: Reference: @ref{Word Lists}.
1.21 crook 3436:
1.29 crook 3437: @c ******************************************************************
1.48 anton 3438: @node Introduction, Words, Tutorial, Top
1.29 crook 3439: @comment node-name, next, previous, up
3440: @chapter An Introduction to ANS Forth
3441: @cindex Forth - an introduction
1.21 crook 3442:
1.29 crook 3443: The primary purpose of this manual is to document Gforth. However, since
3444: Forth is not a widely-known language and there is a lack of up-to-date
3445: teaching material, it seems worthwhile to provide some introductory
1.49 anton 3446: material. For other sources of Forth-related
3447: information, see @ref{Forth-related information}.
1.21 crook 3448:
1.29 crook 3449: The examples in this section should work on any ANS Forth; the
3450: output shown was produced using Gforth. Each example attempts to
3451: reproduce the exact output that Gforth produces. If you try out the
3452: examples (and you should), what you should type is shown @kbd{like this}
3453: and Gforth's response is shown @code{like this}. The single exception is
1.30 anton 3454: that, where the example shows @key{RET} it means that you should
1.29 crook 3455: press the ``carriage return'' key. Unfortunately, some output formats for
3456: this manual cannot show the difference between @kbd{this} and
3457: @code{this} which will make trying out the examples harder (but not
3458: impossible).
1.21 crook 3459:
1.29 crook 3460: Forth is an unusual language. It provides an interactive development
3461: environment which includes both an interpreter and compiler. Forth
3462: programming style encourages you to break a problem down into many
3463: @cindex factoring
3464: small fragments (@dfn{factoring}), and then to develop and test each
3465: fragment interactively. Forth advocates assert that breaking the
3466: edit-compile-test cycle used by conventional programming languages can
3467: lead to great productivity improvements.
1.21 crook 3468:
1.29 crook 3469: @menu
1.67 anton 3470: * Introducing the Text Interpreter::
3471: * Stacks and Postfix notation::
3472: * Your first definition::
3473: * How does that work?::
3474: * Forth is written in Forth::
3475: * Review - elements of a Forth system::
3476: * Where to go next::
3477: * Exercises::
1.29 crook 3478: @end menu
1.21 crook 3479:
1.29 crook 3480: @comment ----------------------------------------------
3481: @node Introducing the Text Interpreter, Stacks and Postfix notation, Introduction, Introduction
3482: @section Introducing the Text Interpreter
3483: @cindex text interpreter
3484: @cindex outer interpreter
1.21 crook 3485:
1.30 anton 3486: @c IMO this is too detailed and the pace is too slow for
3487: @c an introduction. If you know German, take a look at
3488: @c http://www.complang.tuwien.ac.at/anton/lvas/skriptum-stack.html
3489: @c to see how I do it - anton
3490:
1.44 crook 3491: @c nac-> Where I have accepted your comments 100% and modified the text
3492: @c accordingly, I have deleted your comments. Elsewhere I have added a
3493: @c response like this to attempt to rationalise what I have done. Of
3494: @c course, this is a very clumsy mechanism for something that would be
3495: @c done far more efficiently over a beer. Please delete any dialogue
3496: @c you consider closed.
3497:
1.29 crook 3498: When you invoke the Forth image, you will see a startup banner printed
3499: and nothing else (if you have Gforth installed on your system, try
1.30 anton 3500: invoking it now, by typing @kbd{gforth@key{RET}}). Forth is now running
1.29 crook 3501: its command line interpreter, which is called the @dfn{Text Interpreter}
3502: (also known as the @dfn{Outer Interpreter}). (You will learn a lot
1.49 anton 3503: about the text interpreter as you read through this chapter, for more
3504: detail @pxref{The Text Interpreter}).
1.21 crook 3505:
1.29 crook 3506: Although it's not obvious, Forth is actually waiting for your
1.30 anton 3507: input. Type a number and press the @key{RET} key:
1.21 crook 3508:
1.26 crook 3509: @example
1.30 anton 3510: @kbd{45@key{RET}} ok
1.26 crook 3511: @end example
1.21 crook 3512:
1.29 crook 3513: Rather than give you a prompt to invite you to input something, the text
3514: interpreter prints a status message @i{after} it has processed a line
3515: of input. The status message in this case (``@code{ ok}'' followed by
3516: carriage-return) indicates that the text interpreter was able to process
3517: all of your input successfully. Now type something illegal:
3518:
3519: @example
1.30 anton 3520: @kbd{qwer341@key{RET}}
1.29 crook 3521: :1: Undefined word
3522: qwer341
3523: ^^^^^^^
3524: $400D2BA8 Bounce
3525: $400DBDA8 no.extensions
3526: @end example
1.23 crook 3527:
1.29 crook 3528: The exact text, other than the ``Undefined word'' may differ slightly on
3529: your system, but the effect is the same; when the text interpreter
3530: detects an error, it discards any remaining text on a line, resets
1.49 anton 3531: certain internal state and prints an error message. For a detailed description of error messages see @ref{Error
3532: messages}.
1.23 crook 3533:
1.29 crook 3534: The text interpreter waits for you to press carriage-return, and then
3535: processes your input line. Starting at the beginning of the line, it
3536: breaks the line into groups of characters separated by spaces. For each
3537: group of characters in turn, it makes two attempts to do something:
1.23 crook 3538:
1.29 crook 3539: @itemize @bullet
3540: @item
1.44 crook 3541: @cindex name dictionary
1.29 crook 3542: It tries to treat it as a command. It does this by searching a @dfn{name
3543: dictionary}. If the group of characters matches an entry in the name
3544: dictionary, the name dictionary provides the text interpreter with
3545: information that allows the text interpreter perform some actions. In
3546: Forth jargon, we say that the group
3547: @cindex word
3548: @cindex definition
3549: @cindex execution token
3550: @cindex xt
3551: of characters names a @dfn{word}, that the dictionary search returns an
3552: @dfn{execution token (xt)} corresponding to the @dfn{definition} of the
3553: word, and that the text interpreter executes the xt. Often, the terms
3554: @dfn{word} and @dfn{definition} are used interchangeably.
3555: @item
3556: If the text interpreter fails to find a match in the name dictionary, it
3557: tries to treat the group of characters as a number in the current number
3558: base (when you start up Forth, the current number base is base 10). If
3559: the group of characters legitimately represents a number, the text
3560: interpreter pushes the number onto a stack (we'll learn more about that
3561: in the next section).
3562: @end itemize
1.23 crook 3563:
1.29 crook 3564: If the text interpreter is unable to do either of these things with any
3565: group of characters, it discards the group of characters and the rest of
3566: the line, then prints an error message. If the text interpreter reaches
3567: the end of the line without error, it prints the status message ``@code{ ok}''
3568: followed by carriage-return.
1.21 crook 3569:
1.29 crook 3570: This is the simplest command we can give to the text interpreter:
1.23 crook 3571:
3572: @example
1.30 anton 3573: @key{RET} ok
1.23 crook 3574: @end example
1.21 crook 3575:
1.29 crook 3576: The text interpreter did everything we asked it to do (nothing) without
3577: an error, so it said that everything is ``@code{ ok}''. Try a slightly longer
3578: command:
1.21 crook 3579:
1.23 crook 3580: @example
1.30 anton 3581: @kbd{12 dup fred dup@key{RET}}
1.29 crook 3582: :1: Undefined word
3583: 12 dup fred dup
3584: ^^^^
3585: $400D2BA8 Bounce
3586: $400DBDA8 no.extensions
1.23 crook 3587: @end example
1.21 crook 3588:
1.29 crook 3589: When you press the carriage-return key, the text interpreter starts to
3590: work its way along the line:
1.21 crook 3591:
1.29 crook 3592: @itemize @bullet
3593: @item
3594: When it gets to the space after the @code{2}, it takes the group of
3595: characters @code{12} and looks them up in the name
3596: dictionary@footnote{We can't tell if it found them or not, but assume
3597: for now that it did not}. There is no match for this group of characters
3598: in the name dictionary, so it tries to treat them as a number. It is
3599: able to do this successfully, so it puts the number, 12, ``on the stack''
3600: (whatever that means).
3601: @item
3602: The text interpreter resumes scanning the line and gets the next group
3603: of characters, @code{dup}. It looks it up in the name dictionary and
3604: (you'll have to take my word for this) finds it, and executes the word
3605: @code{dup} (whatever that means).
3606: @item
3607: Once again, the text interpreter resumes scanning the line and gets the
3608: group of characters @code{fred}. It looks them up in the name
3609: dictionary, but can't find them. It tries to treat them as a number, but
3610: they don't represent any legal number.
3611: @end itemize
1.21 crook 3612:
1.29 crook 3613: At this point, the text interpreter gives up and prints an error
3614: message. The error message shows exactly how far the text interpreter
3615: got in processing the line. In particular, it shows that the text
3616: interpreter made no attempt to do anything with the final character
3617: group, @code{dup}, even though we have good reason to believe that the
3618: text interpreter would have no problem looking that word up and
3619: executing it a second time.
1.21 crook 3620:
3621:
1.29 crook 3622: @comment ----------------------------------------------
3623: @node Stacks and Postfix notation, Your first definition, Introducing the Text Interpreter, Introduction
3624: @section Stacks, postfix notation and parameter passing
3625: @cindex text interpreter
3626: @cindex outer interpreter
1.21 crook 3627:
1.29 crook 3628: In procedural programming languages (like C and Pascal), the
3629: building-block of programs is the @dfn{function} or @dfn{procedure}. These
3630: functions or procedures are called with @dfn{explicit parameters}. For
3631: example, in C we might write:
1.21 crook 3632:
1.23 crook 3633: @example
1.29 crook 3634: total = total + new_volume(length,height,depth);
1.23 crook 3635: @end example
1.21 crook 3636:
1.23 crook 3637: @noindent
1.29 crook 3638: where new_volume is a function-call to another piece of code, and total,
3639: length, height and depth are all variables. length, height and depth are
3640: parameters to the function-call.
1.21 crook 3641:
1.29 crook 3642: In Forth, the equivalent of the function or procedure is the
3643: @dfn{definition} and parameters are implicitly passed between
3644: definitions using a shared stack that is visible to the
3645: programmer. Although Forth does support variables, the existence of the
3646: stack means that they are used far less often than in most other
3647: programming languages. When the text interpreter encounters a number, it
3648: will place (@dfn{push}) it on the stack. There are several stacks (the
1.30 anton 3649: actual number is implementation-dependent ...) and the particular stack
1.29 crook 3650: used for any operation is implied unambiguously by the operation being
3651: performed. The stack used for all integer operations is called the @dfn{data
3652: stack} and, since this is the stack used most commonly, references to
3653: ``the data stack'' are often abbreviated to ``the stack''.
1.21 crook 3654:
1.29 crook 3655: The stacks have a last-in, first-out (LIFO) organisation. If you type:
1.21 crook 3656:
1.23 crook 3657: @example
1.30 anton 3658: @kbd{1 2 3@key{RET}} ok
1.23 crook 3659: @end example
1.21 crook 3660:
1.29 crook 3661: Then this instructs the text interpreter to placed three numbers on the
3662: (data) stack. An analogy for the behaviour of the stack is to take a
3663: pack of playing cards and deal out the ace (1), 2 and 3 into a pile on
3664: the table. The 3 was the last card onto the pile (``last-in'') and if
3665: you take a card off the pile then, unless you're prepared to fiddle a
3666: bit, the card that you take off will be the 3 (``first-out''). The
3667: number that will be first-out of the stack is called the @dfn{top of
3668: stack}, which
3669: @cindex TOS definition
3670: is often abbreviated to @dfn{TOS}.
1.21 crook 3671:
1.29 crook 3672: To understand how parameters are passed in Forth, consider the
3673: behaviour of the definition @code{+} (pronounced ``plus''). You will not
3674: be surprised to learn that this definition performs addition. More
3675: precisely, it adds two number together and produces a result. Where does
3676: it get the two numbers from? It takes the top two numbers off the
3677: stack. Where does it place the result? On the stack. You can act-out the
3678: behaviour of @code{+} with your playing cards like this:
1.21 crook 3679:
3680: @itemize @bullet
3681: @item
1.29 crook 3682: Pick up two cards from the stack on the table
1.21 crook 3683: @item
1.29 crook 3684: Stare at them intently and ask yourself ``what @i{is} the sum of these two
3685: numbers''
1.21 crook 3686: @item
1.29 crook 3687: Decide that the answer is 5
1.21 crook 3688: @item
1.29 crook 3689: Shuffle the two cards back into the pack and find a 5
1.21 crook 3690: @item
1.29 crook 3691: Put a 5 on the remaining ace that's on the table.
1.21 crook 3692: @end itemize
3693:
1.29 crook 3694: If you don't have a pack of cards handy but you do have Forth running,
3695: you can use the definition @code{.s} to show the current state of the stack,
3696: without affecting the stack. Type:
1.21 crook 3697:
3698: @example
1.30 anton 3699: @kbd{clearstack 1 2 3@key{RET}} ok
3700: @kbd{.s@key{RET}} <3> 1 2 3 ok
1.23 crook 3701: @end example
3702:
1.29 crook 3703: The text interpreter looks up the word @code{clearstack} and executes
3704: it; it tidies up the stack and removes any entries that may have been
3705: left on it by earlier examples. The text interpreter pushes each of the
3706: three numbers in turn onto the stack. Finally, the text interpreter
3707: looks up the word @code{.s} and executes it. The effect of executing
3708: @code{.s} is to print the ``<3>'' (the total number of items on the stack)
3709: followed by a list of all the items on the stack; the item on the far
3710: right-hand side is the TOS.
1.21 crook 3711:
1.29 crook 3712: You can now type:
1.21 crook 3713:
1.29 crook 3714: @example
1.30 anton 3715: @kbd{+ .s@key{RET}} <2> 1 5 ok
1.29 crook 3716: @end example
1.21 crook 3717:
1.29 crook 3718: @noindent
3719: which is correct; there are now 2 items on the stack and the result of
3720: the addition is 5.
1.23 crook 3721:
1.29 crook 3722: If you're playing with cards, try doing a second addition: pick up the
3723: two cards, work out that their sum is 6, shuffle them into the pack,
3724: look for a 6 and place that on the table. You now have just one item on
3725: the stack. What happens if you try to do a third addition? Pick up the
3726: first card, pick up the second card -- ah! There is no second card. This
3727: is called a @dfn{stack underflow} and consitutes an error. If you try to
3728: do the same thing with Forth it will report an error (probably a Stack
3729: Underflow or an Invalid Memory Address error).
1.23 crook 3730:
1.29 crook 3731: The opposite situation to a stack underflow is a @dfn{stack overflow},
3732: which simply accepts that there is a finite amount of storage space
3733: reserved for the stack. To stretch the playing card analogy, if you had
3734: enough packs of cards and you piled the cards up on the table, you would
3735: eventually be unable to add another card; you'd hit the ceiling. Gforth
3736: allows you to set the maximum size of the stacks. In general, the only
3737: time that you will get a stack overflow is because a definition has a
3738: bug in it and is generating data on the stack uncontrollably.
1.23 crook 3739:
1.29 crook 3740: There's one final use for the playing card analogy. If you model your
3741: stack using a pack of playing cards, the maximum number of items on
3742: your stack will be 52 (I assume you didn't use the Joker). The maximum
3743: @i{value} of any item on the stack is 13 (the King). In fact, the only
3744: possible numbers are positive integer numbers 1 through 13; you can't
3745: have (for example) 0 or 27 or 3.52 or -2. If you change the way you
3746: think about some of the cards, you can accommodate different
3747: numbers. For example, you could think of the Jack as representing 0,
3748: the Queen as representing -1 and the King as representing -2. Your
1.45 crook 3749: @i{range} remains unchanged (you can still only represent a total of 13
1.29 crook 3750: numbers) but the numbers that you can represent are -2 through 10.
1.28 crook 3751:
1.29 crook 3752: In that analogy, the limit was the amount of information that a single
3753: stack entry could hold, and Forth has a similar limit. In Forth, the
3754: size of a stack entry is called a @dfn{cell}. The actual size of a cell is
3755: implementation dependent and affects the maximum value that a stack
3756: entry can hold. A Standard Forth provides a cell size of at least
3757: 16-bits, and most desktop systems use a cell size of 32-bits.
1.21 crook 3758:
1.29 crook 3759: Forth does not do any type checking for you, so you are free to
3760: manipulate and combine stack items in any way you wish. A convenient way
3761: of treating stack items is as 2's complement signed integers, and that
3762: is what Standard words like @code{+} do. Therefore you can type:
1.21 crook 3763:
1.29 crook 3764: @example
1.30 anton 3765: @kbd{-5 12 + .s@key{RET}} <1> 7 ok
1.29 crook 3766: @end example
1.21 crook 3767:
1.29 crook 3768: If you use numbers and definitions like @code{+} in order to turn Forth
3769: into a great big pocket calculator, you will realise that it's rather
3770: different from a normal calculator. Rather than typing 2 + 3 = you had
3771: to type 2 3 + (ignore the fact that you had to use @code{.s} to see the
3772: result). The terminology used to describe this difference is to say that
3773: your calculator uses @dfn{Infix Notation} (parameters and operators are
3774: mixed) whilst Forth uses @dfn{Postfix Notation} (parameters and
3775: operators are separate), also called @dfn{Reverse Polish Notation}.
1.21 crook 3776:
1.29 crook 3777: Whilst postfix notation might look confusing to begin with, it has
3778: several important advantages:
1.21 crook 3779:
1.23 crook 3780: @itemize @bullet
3781: @item
1.29 crook 3782: it is unambiguous
1.23 crook 3783: @item
1.29 crook 3784: it is more concise
1.23 crook 3785: @item
1.29 crook 3786: it fits naturally with a stack-based system
1.23 crook 3787: @end itemize
1.21 crook 3788:
1.29 crook 3789: To examine these claims in more detail, consider these sums:
1.21 crook 3790:
1.29 crook 3791: @example
3792: 6 + 5 * 4 =
3793: 4 * 5 + 6 =
3794: @end example
1.21 crook 3795:
1.29 crook 3796: If you're just learning maths or your maths is very rusty, you will
3797: probably come up with the answer 44 for the first and 26 for the
3798: second. If you are a bit of a whizz at maths you will remember the
3799: @i{convention} that multiplication takes precendence over addition, and
3800: you'd come up with the answer 26 both times. To explain the answer 26
3801: to someone who got the answer 44, you'd probably rewrite the first sum
3802: like this:
1.21 crook 3803:
1.29 crook 3804: @example
3805: 6 + (5 * 4) =
3806: @end example
1.21 crook 3807:
1.29 crook 3808: If what you really wanted was to perform the addition before the
3809: multiplication, you would have to use parentheses to force it.
1.21 crook 3810:
1.29 crook 3811: If you did the first two sums on a pocket calculator you would probably
3812: get the right answers, unless you were very cautious and entered them using
3813: these keystroke sequences:
1.21 crook 3814:
1.29 crook 3815: 6 + 5 = * 4 =
3816: 4 * 5 = + 6 =
1.21 crook 3817:
1.29 crook 3818: Postfix notation is unambiguous because the order that the operators
3819: are applied is always explicit; that also means that parentheses are
3820: never required. The operators are @i{active} (the act of quoting the
3821: operator makes the operation occur) which removes the need for ``=''.
1.28 crook 3822:
1.29 crook 3823: The sum 6 + 5 * 4 can be written (in postfix notation) in two
3824: equivalent ways:
1.26 crook 3825:
3826: @example
1.29 crook 3827: 6 5 4 * + or:
3828: 5 4 * 6 +
1.26 crook 3829: @end example
1.23 crook 3830:
1.29 crook 3831: An important thing that you should notice about this notation is that
3832: the @i{order} of the numbers does not change; if you want to subtract
3833: 2 from 10 you type @code{10 2 -}.
1.1 anton 3834:
1.29 crook 3835: The reason that Forth uses postfix notation is very simple to explain: it
3836: makes the implementation extremely simple, and it follows naturally from
3837: using the stack as a mechanism for passing parameters. Another way of
3838: thinking about this is to realise that all Forth definitions are
3839: @i{active}; they execute as they are encountered by the text
3840: interpreter. The result of this is that the syntax of Forth is trivially
3841: simple.
1.1 anton 3842:
3843:
3844:
1.29 crook 3845: @comment ----------------------------------------------
3846: @node Your first definition, How does that work?, Stacks and Postfix notation, Introduction
3847: @section Your first Forth definition
3848: @cindex first definition
1.1 anton 3849:
1.29 crook 3850: Until now, the examples we've seen have been trivial; we've just been
3851: using Forth as a bigger-than-pocket calculator. Also, each calculation
3852: we've shown has been a ``one-off'' -- to repeat it we'd need to type it in
3853: again@footnote{That's not quite true. If you press the up-arrow key on
3854: your keyboard you should be able to scroll back to any earlier command,
3855: edit it and re-enter it.} In this section we'll see how to add new
3856: words to Forth's vocabulary.
1.1 anton 3857:
1.29 crook 3858: The easiest way to create a new word is to use a @dfn{colon
3859: definition}. We'll define a few and try them out before worrying too
3860: much about how they work. Try typing in these examples; be careful to
3861: copy the spaces accurately:
1.1 anton 3862:
1.29 crook 3863: @example
3864: : add-two 2 + . ;
3865: : greet ." Hello and welcome" ;
3866: : demo 5 add-two ;
3867: @end example
1.1 anton 3868:
1.29 crook 3869: @noindent
3870: Now try them out:
1.1 anton 3871:
1.29 crook 3872: @example
1.30 anton 3873: @kbd{greet@key{RET}} Hello and welcome ok
3874: @kbd{greet greet@key{RET}} Hello and welcomeHello and welcome ok
3875: @kbd{4 add-two@key{RET}} 6 ok
3876: @kbd{demo@key{RET}} 7 ok
3877: @kbd{9 greet demo add-two@key{RET}} Hello and welcome7 11 ok
1.29 crook 3878: @end example
1.1 anton 3879:
1.29 crook 3880: The first new thing that we've introduced here is the pair of words
3881: @code{:} and @code{;}. These are used to start and terminate a new
3882: definition, respectively. The first word after the @code{:} is the name
3883: for the new definition.
1.1 anton 3884:
1.29 crook 3885: As you can see from the examples, a definition is built up of words that
3886: have already been defined; Forth makes no distinction between
3887: definitions that existed when you started the system up, and those that
3888: you define yourself.
1.1 anton 3889:
1.29 crook 3890: The examples also introduce the words @code{.} (dot), @code{."}
3891: (dot-quote) and @code{dup} (dewp). Dot takes the value from the top of
3892: the stack and displays it. It's like @code{.s} except that it only
3893: displays the top item of the stack and it is destructive; after it has
3894: executed, the number is no longer on the stack. There is always one
3895: space printed after the number, and no spaces before it. Dot-quote
3896: defines a string (a sequence of characters) that will be printed when
3897: the word is executed. The string can contain any printable characters
3898: except @code{"}. A @code{"} has a special function; it is not a Forth
3899: word but it acts as a delimiter (the way that delimiters work is
3900: described in the next section). Finally, @code{dup} duplicates the value
3901: at the top of the stack. Try typing @code{5 dup .s} to see what it does.
1.1 anton 3902:
1.29 crook 3903: We already know that the text interpreter searches through the
3904: dictionary to locate names. If you've followed the examples earlier, you
3905: will already have a definition called @code{add-two}. Lets try modifying
3906: it by typing in a new definition:
1.1 anton 3907:
1.29 crook 3908: @example
1.30 anton 3909: @kbd{: add-two dup . ." + 2 =" 2 + . ;@key{RET}} redefined add-two ok
1.29 crook 3910: @end example
1.5 anton 3911:
1.29 crook 3912: Forth recognised that we were defining a word that already exists, and
3913: printed a message to warn us of that fact. Let's try out the new
3914: definition:
1.5 anton 3915:
1.29 crook 3916: @example
1.30 anton 3917: @kbd{9 add-two@key{RET}} 9 + 2 =11 ok
1.29 crook 3918: @end example
1.1 anton 3919:
1.29 crook 3920: @noindent
3921: All that we've actually done here, though, is to create a new
3922: definition, with a particular name. The fact that there was already a
3923: definition with the same name did not make any difference to the way
3924: that the new definition was created (except that Forth printed a warning
3925: message). The old definition of add-two still exists (try @code{demo}
3926: again to see that this is true). Any new definition will use the new
3927: definition of @code{add-two}, but old definitions continue to use the
3928: version that already existed at the time that they were @code{compiled}.
1.1 anton 3929:
1.29 crook 3930: Before you go on to the next section, try defining and redefining some
3931: words of your own.
1.1 anton 3932:
1.29 crook 3933: @comment ----------------------------------------------
3934: @node How does that work?, Forth is written in Forth, Your first definition, Introduction
3935: @section How does that work?
3936: @cindex parsing words
1.1 anton 3937:
1.30 anton 3938: @c That's pretty deep (IMO way too deep) for an introduction. - anton
3939:
3940: @c Is it a good idea to talk about the interpretation semantics of a
3941: @c number? We don't have an xt to go along with it. - anton
3942:
3943: @c Now that I have eliminated execution semantics, I wonder if it would not
3944: @c be better to keep them (or add run-time semantics), to make it easier to
3945: @c explain what compilation semantics usually does. - anton
3946:
1.44 crook 3947: @c nac-> I removed the term ``default compilation sematics'' from the
3948: @c introductory chapter. Removing ``execution semantics'' was making
3949: @c everything simpler to explain, then I think the use of this term made
3950: @c everything more complex again. I replaced it with ``default
3951: @c semantics'' (which is used elsewhere in the manual) by which I mean
3952: @c ``a definition that has neither the immediate nor the compile-only
3953: @c flag set''. I reworded big chunks of the ``how does that work''
3954: @c section (and, unusually for me, I think I even made it shorter!). See
3955: @c what you think -- I know I have not addressed your primary concern
3956: @c that it is too heavy-going for an introduction. From what I understood
3957: @c of your course notes it looks as though they might be a good framework.
3958: @c Things that I've tried to capture here are some things that came as a
3959: @c great revelation here when I first understood them. Also, I like the
3960: @c fact that a very simple code example shows up almost all of the issues
3961: @c that you need to understand to see how Forth works. That's unique and
3962: @c worthwhile to emphasise.
3963:
1.29 crook 3964: Now we're going to take another look at the definition of @code{add-two}
3965: from the previous section. From our knowledge of the way that the text
3966: interpreter works, we would have expected this result when we tried to
3967: define @code{add-two}:
1.21 crook 3968:
1.29 crook 3969: @example
1.44 crook 3970: @kbd{: add-two 2 + . ;@key{RET}}
1.29 crook 3971: ^^^^^^^
3972: Error: Undefined word
3973: @end example
1.28 crook 3974:
1.29 crook 3975: The reason that this didn't happen is bound up in the way that @code{:}
3976: works. The word @code{:} does two special things. The first special
3977: thing that it does prevents the text interpreter from ever seeing the
3978: characters @code{add-two}. The text interpreter uses a variable called
3979: @cindex modifying >IN
1.44 crook 3980: @code{>IN} (pronounced ``to-in'') to keep track of where it is in the
1.29 crook 3981: input line. When it encounters the word @code{:} it behaves in exactly
3982: the same way as it does for any other word; it looks it up in the name
3983: dictionary, finds its xt and executes it. When @code{:} executes, it
3984: looks at the input buffer, finds the word @code{add-two} and advances the
3985: value of @code{>IN} to point past it. It then does some other stuff
3986: associated with creating the new definition (including creating an entry
3987: for @code{add-two} in the name dictionary). When the execution of @code{:}
3988: completes, control returns to the text interpreter, which is oblivious
3989: to the fact that it has been tricked into ignoring part of the input
3990: line.
1.21 crook 3991:
1.29 crook 3992: @cindex parsing words
3993: Words like @code{:} -- words that advance the value of @code{>IN} and so
3994: prevent the text interpreter from acting on the whole of the input line
3995: -- are called @dfn{parsing words}.
1.21 crook 3996:
1.29 crook 3997: @cindex @code{state} - effect on the text interpreter
3998: @cindex text interpreter - effect of state
3999: The second special thing that @code{:} does is change the value of a
4000: variable called @code{state}, which affects the way that the text
4001: interpreter behaves. When Gforth starts up, @code{state} has the value
4002: 0, and the text interpreter is said to be @dfn{interpreting}. During a
4003: colon definition (started with @code{:}), @code{state} is set to -1 and
1.44 crook 4004: the text interpreter is said to be @dfn{compiling}.
4005:
4006: In this example, the text interpreter is compiling when it processes the
4007: string ``@code{2 + . ;}''. It still breaks the string down into
4008: character sequences in the same way. However, instead of pushing the
4009: number @code{2} onto the stack, it lays down (@dfn{compiles}) some magic
4010: into the definition of @code{add-two} that will make the number @code{2} get
4011: pushed onto the stack when @code{add-two} is @dfn{executed}. Similarly,
4012: the behaviours of @code{+} and @code{.} are also compiled into the
4013: definition.
4014:
4015: One category of words don't get compiled. These so-called @dfn{immediate
4016: words} get executed (performed @i{now}) regardless of whether the text
4017: interpreter is interpreting or compiling. The word @code{;} is an
4018: immediate word. Rather than being compiled into the definition, it
4019: executes. Its effect is to terminate the current definition, which
4020: includes changing the value of @code{state} back to 0.
4021:
4022: When you execute @code{add-two}, it has a @dfn{run-time effect} that is
4023: exactly the same as if you had typed @code{2 + . @key{RET}} outside of a
4024: definition.
1.28 crook 4025:
1.30 anton 4026: In Forth, every word or number can be described in terms of two
1.29 crook 4027: properties:
1.28 crook 4028:
4029: @itemize @bullet
4030: @item
1.29 crook 4031: @cindex interpretation semantics
1.44 crook 4032: Its @dfn{interpretation semantics} describe how it will behave when the
4033: text interpreter encounters it in @dfn{interpret} state. The
4034: interpretation semantics of a word are represented by an @dfn{execution
4035: token}.
1.28 crook 4036: @item
1.29 crook 4037: @cindex compilation semantics
1.44 crook 4038: Its @dfn{compilation semantics} describe how it will behave when the
4039: text interpreter encounters it in @dfn{compile} state. The compilation
4040: semantics of a word are represented in an implementation-dependent way;
4041: Gforth uses a @dfn{compilation token}.
1.29 crook 4042: @end itemize
4043:
4044: @noindent
4045: Numbers are always treated in a fixed way:
4046:
4047: @itemize @bullet
1.28 crook 4048: @item
1.44 crook 4049: When the number is @dfn{interpreted}, its behaviour is to push the
4050: number onto the stack.
1.28 crook 4051: @item
1.30 anton 4052: When the number is @dfn{compiled}, a piece of code is appended to the
4053: current definition that pushes the number when it runs. (In other words,
4054: the compilation semantics of a number are to postpone its interpretation
4055: semantics until the run-time of the definition that it is being compiled
4056: into.)
1.29 crook 4057: @end itemize
4058:
1.44 crook 4059: Words don't behave in such a regular way, but most have @i{default
4060: semantics} which means that they behave like this:
1.29 crook 4061:
4062: @itemize @bullet
1.28 crook 4063: @item
1.30 anton 4064: The @dfn{interpretation semantics} of the word are to do something useful.
4065: @item
1.29 crook 4066: The @dfn{compilation semantics} of the word are to append its
1.30 anton 4067: @dfn{interpretation semantics} to the current definition (so that its
4068: run-time behaviour is to do something useful).
1.28 crook 4069: @end itemize
4070:
1.30 anton 4071: @cindex immediate words
1.44 crook 4072: The actual behaviour of any particular word can be controlled by using
4073: the words @code{immediate} and @code{compile-only} when the word is
4074: defined. These words set flags in the name dictionary entry of the most
4075: recently defined word, and these flags are retrieved by the text
4076: interpreter when it finds the word in the name dictionary.
4077:
4078: A word that is marked as @dfn{immediate} has compilation semantics that
4079: are identical to its interpretation semantics. In other words, it
4080: behaves like this:
1.29 crook 4081:
4082: @itemize @bullet
4083: @item
1.30 anton 4084: The @dfn{interpretation semantics} of the word are to do something useful.
1.29 crook 4085: @item
1.30 anton 4086: The @dfn{compilation semantics} of the word are to do something useful
4087: (and actually the same thing); i.e., it is executed during compilation.
1.29 crook 4088: @end itemize
1.28 crook 4089:
1.44 crook 4090: Marking a word as @dfn{compile-only} prohibits the text interpreter from
4091: performing the interpretation semantics of the word directly; an attempt
4092: to do so will generate an error. It is never necessary to use
4093: @code{compile-only} (and it is not even part of ANS Forth, though it is
4094: provided by many implementations) but it is good etiquette to apply it
4095: to a word that will not behave correctly (and might have unexpected
4096: side-effects) in interpret state. For example, it is only legal to use
4097: the conditional word @code{IF} within a definition. If you forget this
4098: and try to use it elsewhere, the fact that (in Gforth) it is marked as
4099: @code{compile-only} allows the text interpreter to generate a helpful
4100: error message rather than subjecting you to the consequences of your
4101: folly.
4102:
1.29 crook 4103: This example shows the difference between an immediate and a
4104: non-immediate word:
1.28 crook 4105:
1.29 crook 4106: @example
4107: : show-state state @@ . ;
4108: : show-state-now show-state ; immediate
4109: : word1 show-state ;
4110: : word2 show-state-now ;
1.28 crook 4111: @end example
1.23 crook 4112:
1.29 crook 4113: The word @code{immediate} after the definition of @code{show-state-now}
4114: makes that word an immediate word. These definitions introduce a new
4115: word: @code{@@} (pronounced ``fetch''). This word fetches the value of a
4116: variable, and leaves it on the stack. Therefore, the behaviour of
4117: @code{show-state} is to print a number that represents the current value
4118: of @code{state}.
1.28 crook 4119:
1.29 crook 4120: When you execute @code{word1}, it prints the number 0, indicating that
4121: the system is interpreting. When the text interpreter compiled the
4122: definition of @code{word1}, it encountered @code{show-state} whose
1.30 anton 4123: compilation semantics are to append its interpretation semantics to the
1.29 crook 4124: current definition. When you execute @code{word1}, it performs the
1.30 anton 4125: interpretation semantics of @code{show-state}. At the time that @code{word1}
1.29 crook 4126: (and therefore @code{show-state}) are executed, the system is
4127: interpreting.
1.28 crook 4128:
1.30 anton 4129: When you pressed @key{RET} after entering the definition of @code{word2},
1.29 crook 4130: you should have seen the number -1 printed, followed by ``@code{
4131: ok}''. When the text interpreter compiled the definition of
4132: @code{word2}, it encountered @code{show-state-now}, an immediate word,
1.30 anton 4133: whose compilation semantics are therefore to perform its interpretation
1.29 crook 4134: semantics. It is executed straight away (even before the text
4135: interpreter has moved on to process another group of characters; the
4136: @code{;} in this example). The effect of executing it are to display the
4137: value of @code{state} @i{at the time that the definition of}
4138: @code{word2} @i{is being defined}. Printing -1 demonstrates that the
4139: system is compiling at this time. If you execute @code{word2} it does
4140: nothing at all.
1.28 crook 4141:
1.29 crook 4142: @cindex @code{."}, how it works
4143: Before leaving the subject of immediate words, consider the behaviour of
4144: @code{."} in the definition of @code{greet}, in the previous
4145: section. This word is both a parsing word and an immediate word. Notice
4146: that there is a space between @code{."} and the start of the text
4147: @code{Hello and welcome}, but that there is no space between the last
4148: letter of @code{welcome} and the @code{"} character. The reason for this
4149: is that @code{."} is a Forth word; it must have a space after it so that
4150: the text interpreter can identify it. The @code{"} is not a Forth word;
4151: it is a @dfn{delimiter}. The examples earlier show that, when the string
4152: is displayed, there is neither a space before the @code{H} nor after the
4153: @code{e}. Since @code{."} is an immediate word, it executes at the time
4154: that @code{greet} is defined. When it executes, its behaviour is to
4155: search forward in the input line looking for the delimiter. When it
4156: finds the delimiter, it updates @code{>IN} to point past the
4157: delimiter. It also compiles some magic code into the definition of
4158: @code{greet}; the xt of a run-time routine that prints a text string. It
4159: compiles the string @code{Hello and welcome} into memory so that it is
4160: available to be printed later. When the text interpreter gains control,
4161: the next word it finds in the input stream is @code{;} and so it
4162: terminates the definition of @code{greet}.
1.28 crook 4163:
4164:
4165: @comment ----------------------------------------------
1.29 crook 4166: @node Forth is written in Forth, Review - elements of a Forth system, How does that work?, Introduction
4167: @section Forth is written in Forth
4168: @cindex structure of Forth programs
4169:
4170: When you start up a Forth compiler, a large number of definitions
4171: already exist. In Forth, you develop a new application using bottom-up
4172: programming techniques to create new definitions that are defined in
4173: terms of existing definitions. As you create each definition you can
4174: test and debug it interactively.
4175:
4176: If you have tried out the examples in this section, you will probably
4177: have typed them in by hand; when you leave Gforth, your definitions will
4178: be lost. You can avoid this by using a text editor to enter Forth source
4179: code into a file, and then loading code from the file using
1.49 anton 4180: @code{include} (@pxref{Forth source files}). A Forth source file is
1.29 crook 4181: processed by the text interpreter, just as though you had typed it in by
4182: hand@footnote{Actually, there are some subtle differences -- see
4183: @ref{The Text Interpreter}.}.
4184:
4185: Gforth also supports the traditional Forth alternative to using text
1.49 anton 4186: files for program entry (@pxref{Blocks}).
1.28 crook 4187:
1.29 crook 4188: In common with many, if not most, Forth compilers, most of Gforth is
4189: actually written in Forth. All of the @file{.fs} files in the
4190: installation directory@footnote{For example,
1.30 anton 4191: @file{/usr/local/share/gforth...}} are Forth source files, which you can
1.29 crook 4192: study to see examples of Forth programming.
1.28 crook 4193:
1.29 crook 4194: Gforth maintains a history file that records every line that you type to
4195: the text interpreter. This file is preserved between sessions, and is
4196: used to provide a command-line recall facility. If you enter long
4197: definitions by hand, you can use a text editor to paste them out of the
4198: history file into a Forth source file for reuse at a later time
1.49 anton 4199: (for more information @pxref{Command-line editing}).
1.28 crook 4200:
4201:
4202: @comment ----------------------------------------------
1.29 crook 4203: @node Review - elements of a Forth system, Where to go next, Forth is written in Forth, Introduction
4204: @section Review - elements of a Forth system
4205: @cindex elements of a Forth system
1.28 crook 4206:
1.29 crook 4207: To summarise this chapter:
1.28 crook 4208:
4209: @itemize @bullet
4210: @item
1.29 crook 4211: Forth programs use @dfn{factoring} to break a problem down into small
4212: fragments called @dfn{words} or @dfn{definitions}.
4213: @item
4214: Forth program development is an interactive process.
4215: @item
4216: The main command loop that accepts input, and controls both
4217: interpretation and compilation, is called the @dfn{text interpreter}
4218: (also known as the @dfn{outer interpreter}).
4219: @item
4220: Forth has a very simple syntax, consisting of words and numbers
4221: separated by spaces or carriage-return characters. Any additional syntax
4222: is imposed by @dfn{parsing words}.
4223: @item
4224: Forth uses a stack to pass parameters between words. As a result, it
4225: uses postfix notation.
4226: @item
4227: To use a word that has previously been defined, the text interpreter
4228: searches for the word in the @dfn{name dictionary}.
4229: @item
1.30 anton 4230: Words have @dfn{interpretation semantics} and @dfn{compilation semantics}.
1.28 crook 4231: @item
1.29 crook 4232: The text interpreter uses the value of @code{state} to select between
4233: the use of the @dfn{interpretation semantics} and the @dfn{compilation
4234: semantics} of a word that it encounters.
1.28 crook 4235: @item
1.30 anton 4236: The relationship between the @dfn{interpretation semantics} and
4237: @dfn{compilation semantics} for a word
1.29 crook 4238: depend upon the way in which the word was defined (for example, whether
4239: it is an @dfn{immediate} word).
1.28 crook 4240: @item
1.29 crook 4241: Forth definitions can be implemented in Forth (called @dfn{high-level
4242: definitions}) or in some other way (usually a lower-level language and
4243: as a result often called @dfn{low-level definitions}, @dfn{code
4244: definitions} or @dfn{primitives}).
1.28 crook 4245: @item
1.29 crook 4246: Many Forth systems are implemented mainly in Forth.
1.28 crook 4247: @end itemize
4248:
4249:
1.29 crook 4250: @comment ----------------------------------------------
1.48 anton 4251: @node Where to go next, Exercises, Review - elements of a Forth system, Introduction
1.29 crook 4252: @section Where To Go Next
4253: @cindex where to go next
1.28 crook 4254:
1.29 crook 4255: Amazing as it may seem, if you have read (and understood) this far, you
4256: know almost all the fundamentals about the inner workings of a Forth
4257: system. You certainly know enough to be able to read and understand the
4258: rest of this manual and the ANS Forth document, to learn more about the
4259: facilities that Forth in general and Gforth in particular provide. Even
4260: scarier, you know almost enough to implement your own Forth system.
1.30 anton 4261: However, that's not a good idea just yet... better to try writing some
1.29 crook 4262: programs in Gforth.
1.28 crook 4263:
1.29 crook 4264: Forth has such a rich vocabulary that it can be hard to know where to
4265: start in learning it. This section suggests a few sets of words that are
4266: enough to write small but useful programs. Use the word index in this
4267: document to learn more about each word, then try it out and try to write
4268: small definitions using it. Start by experimenting with these words:
1.28 crook 4269:
4270: @itemize @bullet
4271: @item
1.29 crook 4272: Arithmetic: @code{+ - * / /MOD */ ABS INVERT}
4273: @item
4274: Comparison: @code{MIN MAX =}
4275: @item
4276: Logic: @code{AND OR XOR NOT}
4277: @item
4278: Stack manipulation: @code{DUP DROP SWAP OVER}
1.28 crook 4279: @item
1.29 crook 4280: Loops and decisions: @code{IF ELSE ENDIF ?DO I LOOP}
1.28 crook 4281: @item
1.29 crook 4282: Input/Output: @code{. ." EMIT CR KEY}
1.28 crook 4283: @item
1.29 crook 4284: Defining words: @code{: ; CREATE}
1.28 crook 4285: @item
1.29 crook 4286: Memory allocation words: @code{ALLOT ,}
1.28 crook 4287: @item
1.29 crook 4288: Tools: @code{SEE WORDS .S MARKER}
4289: @end itemize
4290:
4291: When you have mastered those, go on to:
4292:
4293: @itemize @bullet
1.28 crook 4294: @item
1.29 crook 4295: More defining words: @code{VARIABLE CONSTANT VALUE TO CREATE DOES>}
1.28 crook 4296: @item
1.29 crook 4297: Memory access: @code{@@ !}
1.28 crook 4298: @end itemize
1.23 crook 4299:
1.29 crook 4300: When you have mastered these, there's nothing for it but to read through
4301: the whole of this manual and find out what you've missed.
4302:
4303: @comment ----------------------------------------------
1.48 anton 4304: @node Exercises, , Where to go next, Introduction
1.29 crook 4305: @section Exercises
4306: @cindex exercises
4307:
4308: TODO: provide a set of programming excercises linked into the stuff done
4309: already and into other sections of the manual. Provide solutions to all
4310: the exercises in a .fs file in the distribution.
4311:
4312: @c Get some inspiration from Starting Forth and Kelly&Spies.
4313:
4314: @c excercises:
4315: @c 1. take inches and convert to feet and inches.
4316: @c 2. take temperature and convert from fahrenheight to celcius;
4317: @c may need to care about symmetric vs floored??
4318: @c 3. take input line and do character substitution
4319: @c to encipher or decipher
4320: @c 4. as above but work on a file for in and out
4321: @c 5. take input line and convert to pig-latin
4322: @c
4323: @c thing of sets of things to exercise then come up with
4324: @c problems that need those things.
4325:
4326:
1.26 crook 4327: @c ******************************************************************
1.29 crook 4328: @node Words, Error messages, Introduction, Top
1.1 anton 4329: @chapter Forth Words
1.26 crook 4330: @cindex words
1.1 anton 4331:
4332: @menu
4333: * Notation::
1.65 anton 4334: * Case insensitivity::
4335: * Comments::
4336: * Boolean Flags::
1.1 anton 4337: * Arithmetic::
4338: * Stack Manipulation::
1.5 anton 4339: * Memory::
1.1 anton 4340: * Control Structures::
4341: * Defining Words::
1.65 anton 4342: * Interpretation and Compilation Semantics::
1.47 crook 4343: * Tokens for Words::
1.65 anton 4344: * The Text Interpreter::
4345: * Word Lists::
4346: * Environmental Queries::
1.12 anton 4347: * Files::
4348: * Blocks::
4349: * Other I/O::
4350: * Programming Tools::
4351: * Assembler and Code Words::
4352: * Threading Words::
1.26 crook 4353: * Locals::
4354: * Structures::
4355: * Object-oriented Forth::
1.65 anton 4356: * Passing Commands to the OS::
4357: * Keeping track of Time::
4358: * Miscellaneous Words::
1.1 anton 4359: @end menu
4360:
1.65 anton 4361: @node Notation, Case insensitivity, Words, Words
1.1 anton 4362: @section Notation
4363: @cindex notation of glossary entries
4364: @cindex format of glossary entries
4365: @cindex glossary notation format
4366: @cindex word glossary entry format
4367:
4368: The Forth words are described in this section in the glossary notation
1.67 anton 4369: that has become a de-facto standard for Forth texts:
1.1 anton 4370:
4371: @format
1.29 crook 4372: @i{word} @i{Stack effect} @i{wordset} @i{pronunciation}
1.1 anton 4373: @end format
1.29 crook 4374: @i{Description}
1.1 anton 4375:
4376: @table @var
4377: @item word
1.28 crook 4378: The name of the word.
1.1 anton 4379:
4380: @item Stack effect
4381: @cindex stack effect
1.29 crook 4382: The stack effect is written in the notation @code{@i{before} --
4383: @i{after}}, where @i{before} and @i{after} describe the top of
1.1 anton 4384: stack entries before and after the execution of the word. The rest of
4385: the stack is not touched by the word. The top of stack is rightmost,
4386: i.e., a stack sequence is written as it is typed in. Note that Gforth
4387: uses a separate floating point stack, but a unified stack
1.29 crook 4388: notation. Also, return stack effects are not shown in @i{stack
4389: effect}, but in @i{Description}. The name of a stack item describes
1.1 anton 4390: the type and/or the function of the item. See below for a discussion of
4391: the types.
4392:
4393: All words have two stack effects: A compile-time stack effect and a
4394: run-time stack effect. The compile-time stack-effect of most words is
1.29 crook 4395: @i{ -- }. If the compile-time stack-effect of a word deviates from
1.1 anton 4396: this standard behaviour, or the word does other unusual things at
4397: compile time, both stack effects are shown; otherwise only the run-time
4398: stack effect is shown.
4399:
4400: @cindex pronounciation of words
4401: @item pronunciation
4402: How the word is pronounced.
4403:
4404: @cindex wordset
1.67 anton 4405: @cindex environment wordset
1.1 anton 4406: @item wordset
1.21 crook 4407: The ANS Forth standard is divided into several word sets. A standard
4408: system need not support all of them. Therefore, in theory, the fewer
4409: word sets your program uses the more portable it will be. However, we
4410: suspect that most ANS Forth systems on personal machines will feature
1.26 crook 4411: all word sets. Words that are not defined in ANS Forth have
1.21 crook 4412: @code{gforth} or @code{gforth-internal} as word set. @code{gforth}
1.1 anton 4413: describes words that will work in future releases of Gforth;
4414: @code{gforth-internal} words are more volatile. Environmental query
4415: strings are also displayed like words; you can recognize them by the
1.21 crook 4416: @code{environment} in the word set field.
1.1 anton 4417:
4418: @item Description
4419: A description of the behaviour of the word.
4420: @end table
4421:
4422: @cindex types of stack items
4423: @cindex stack item types
4424: The type of a stack item is specified by the character(s) the name
4425: starts with:
4426:
4427: @table @code
4428: @item f
4429: @cindex @code{f}, stack item type
4430: Boolean flags, i.e. @code{false} or @code{true}.
4431: @item c
4432: @cindex @code{c}, stack item type
4433: Char
4434: @item w
4435: @cindex @code{w}, stack item type
4436: Cell, can contain an integer or an address
4437: @item n
4438: @cindex @code{n}, stack item type
4439: signed integer
4440: @item u
4441: @cindex @code{u}, stack item type
4442: unsigned integer
4443: @item d
4444: @cindex @code{d}, stack item type
4445: double sized signed integer
4446: @item ud
4447: @cindex @code{ud}, stack item type
4448: double sized unsigned integer
4449: @item r
4450: @cindex @code{r}, stack item type
4451: Float (on the FP stack)
1.21 crook 4452: @item a-
1.1 anton 4453: @cindex @code{a_}, stack item type
4454: Cell-aligned address
1.21 crook 4455: @item c-
1.1 anton 4456: @cindex @code{c_}, stack item type
4457: Char-aligned address (note that a Char may have two bytes in Windows NT)
1.21 crook 4458: @item f-
1.1 anton 4459: @cindex @code{f_}, stack item type
4460: Float-aligned address
1.21 crook 4461: @item df-
1.1 anton 4462: @cindex @code{df_}, stack item type
4463: Address aligned for IEEE double precision float
1.21 crook 4464: @item sf-
1.1 anton 4465: @cindex @code{sf_}, stack item type
4466: Address aligned for IEEE single precision float
4467: @item xt
4468: @cindex @code{xt}, stack item type
4469: Execution token, same size as Cell
4470: @item wid
4471: @cindex @code{wid}, stack item type
1.21 crook 4472: Word list ID, same size as Cell
1.68 anton 4473: @item ior, wior
4474: @cindex ior type description
4475: @cindex wior type description
4476: I/O result code, cell-sized. In Gforth, you can @code{throw} iors.
1.1 anton 4477: @item f83name
4478: @cindex @code{f83name}, stack item type
4479: Pointer to a name structure
4480: @item "
4481: @cindex @code{"}, stack item type
1.12 anton 4482: string in the input stream (not on the stack). The terminating character
4483: is a blank by default. If it is not a blank, it is shown in @code{<>}
1.1 anton 4484: quotes.
4485: @end table
4486:
1.65 anton 4487: @comment ----------------------------------------------
4488: @node Case insensitivity, Comments, Notation, Words
4489: @section Case insensitivity
4490: @cindex case sensitivity
4491: @cindex upper and lower case
4492:
4493: Gforth is case-insensitive; you can enter definitions and invoke
4494: Standard words using upper, lower or mixed case (however,
4495: @pxref{core-idef, Implementation-defined options, Implementation-defined
4496: options}).
4497:
4498: ANS Forth only @i{requires} implementations to recognise Standard words
4499: when they are typed entirely in upper case. Therefore, a Standard
4500: program must use upper case for all Standard words. You can use whatever
4501: case you like for words that you define, but in a Standard program you
4502: have to use the words in the same case that you defined them.
4503:
4504: Gforth supports case sensitivity through @code{table}s (case-sensitive
4505: wordlists, @pxref{Word Lists}).
4506:
4507: Two people have asked how to convert Gforth to be case-sensitive; while
4508: we think this is a bad idea, you can change all wordlists into tables
4509: like this:
4510:
4511: @example
4512: ' table-find forth-wordlist wordlist-map @ !
4513: @end example
4514:
4515: Note that you now have to type the predefined words in the same case
4516: that we defined them, which are varying. You may want to convert them
4517: to your favourite case before doing this operation (I won't explain how,
4518: because if you are even contemplating doing this, you'd better have
4519: enough knowledge of Forth systems to know this already).
4520:
4521: @node Comments, Boolean Flags, Case insensitivity, Words
1.21 crook 4522: @section Comments
1.26 crook 4523: @cindex comments
1.21 crook 4524:
1.29 crook 4525: Forth supports two styles of comment; the traditional @i{in-line} comment,
4526: @code{(} and its modern cousin, the @i{comment to end of line}; @code{\}.
1.21 crook 4527:
1.44 crook 4528:
1.23 crook 4529: doc-(
1.21 crook 4530: doc-\
1.23 crook 4531: doc-\G
1.21 crook 4532:
1.44 crook 4533:
1.21 crook 4534: @node Boolean Flags, Arithmetic, Comments, Words
4535: @section Boolean Flags
1.26 crook 4536: @cindex Boolean flags
1.21 crook 4537:
4538: A Boolean flag is cell-sized. A cell with all bits clear represents the
4539: flag @code{false} and a flag with all bits set represents the flag
1.26 crook 4540: @code{true}. Words that check a flag (for example, @code{IF}) will treat
1.29 crook 4541: a cell that has @i{any} bit set as @code{true}.
1.67 anton 4542: @c on and off to Memory?
4543: @c true and false to "Bitwise operations" or "Numeric comparison"?
1.44 crook 4544:
1.21 crook 4545: doc-true
4546: doc-false
1.29 crook 4547: doc-on
4548: doc-off
1.21 crook 4549:
1.44 crook 4550:
1.21 crook 4551: @node Arithmetic, Stack Manipulation, Boolean Flags, Words
1.1 anton 4552: @section Arithmetic
4553: @cindex arithmetic words
4554:
4555: @cindex division with potentially negative operands
4556: Forth arithmetic is not checked, i.e., you will not hear about integer
4557: overflow on addition or multiplication, you may hear about division by
4558: zero if you are lucky. The operator is written after the operands, but
4559: the operands are still in the original order. I.e., the infix @code{2-1}
4560: corresponds to @code{2 1 -}. Forth offers a variety of division
4561: operators. If you perform division with potentially negative operands,
4562: you do not want to use @code{/} or @code{/mod} with its undefined
4563: behaviour, but rather @code{fm/mod} or @code{sm/mod} (probably the
4564: former, @pxref{Mixed precision}).
1.26 crook 4565: @comment TODO discuss the different division forms and the std approach
1.1 anton 4566:
4567: @menu
4568: * Single precision::
1.67 anton 4569: * Double precision:: Double-cell integer arithmetic
1.1 anton 4570: * Bitwise operations::
1.67 anton 4571: * Numeric comparison::
1.29 crook 4572: * Mixed precision:: Operations with single and double-cell integers
1.1 anton 4573: * Floating Point::
4574: @end menu
4575:
1.67 anton 4576: @node Single precision, Double precision, Arithmetic, Arithmetic
1.1 anton 4577: @subsection Single precision
4578: @cindex single precision arithmetic words
4579:
1.67 anton 4580: @c !! cell undefined
4581:
4582: By default, numbers in Forth are single-precision integers that are one
1.26 crook 4583: cell in size. They can be signed or unsigned, depending upon how you
1.49 anton 4584: treat them. For the rules used by the text interpreter for recognising
4585: single-precision integers see @ref{Number Conversion}.
1.21 crook 4586:
1.67 anton 4587: These words are all defined for signed operands, but some of them also
4588: work for unsigned numbers: @code{+}, @code{1+}, @code{-}, @code{1-},
4589: @code{*}.
1.44 crook 4590:
1.1 anton 4591: doc-+
1.21 crook 4592: doc-1+
1.1 anton 4593: doc--
1.21 crook 4594: doc-1-
1.1 anton 4595: doc-*
4596: doc-/
4597: doc-mod
4598: doc-/mod
4599: doc-negate
4600: doc-abs
4601: doc-min
4602: doc-max
1.27 crook 4603: doc-floored
1.1 anton 4604:
1.44 crook 4605:
1.67 anton 4606: @node Double precision, Bitwise operations, Single precision, Arithmetic
1.21 crook 4607: @subsection Double precision
4608: @cindex double precision arithmetic words
4609:
1.49 anton 4610: For the rules used by the text interpreter for
4611: recognising double-precision integers, see @ref{Number Conversion}.
1.21 crook 4612:
4613: A double precision number is represented by a cell pair, with the most
1.67 anton 4614: significant cell at the TOS. It is trivial to convert an unsigned single
4615: to a double: simply push a @code{0} onto the TOS. Since numbers are
4616: represented by Gforth using 2's complement arithmetic, converting a
4617: signed single to a (signed) double requires sign-extension across the
4618: most significant cell. This can be achieved using @code{s>d}. The moral
4619: of the story is that you cannot convert a number without knowing whether
4620: it represents an unsigned or a signed number.
1.21 crook 4621:
1.67 anton 4622: These words are all defined for signed operands, but some of them also
4623: work for unsigned numbers: @code{d+}, @code{d-}.
1.44 crook 4624:
1.21 crook 4625: doc-s>d
1.67 anton 4626: doc-d>s
1.21 crook 4627: doc-d+
4628: doc-d-
4629: doc-dnegate
4630: doc-dabs
4631: doc-dmin
4632: doc-dmax
4633:
1.44 crook 4634:
1.67 anton 4635: @node Bitwise operations, Numeric comparison, Double precision, Arithmetic
4636: @subsection Bitwise operations
4637: @cindex bitwise operation words
4638:
4639:
4640: doc-and
4641: doc-or
4642: doc-xor
4643: doc-invert
4644: doc-lshift
4645: doc-rshift
4646: doc-2*
4647: doc-d2*
4648: doc-2/
4649: doc-d2/
4650:
4651:
4652: @node Numeric comparison, Mixed precision, Bitwise operations, Arithmetic
1.21 crook 4653: @subsection Numeric comparison
4654: @cindex numeric comparison words
4655:
1.67 anton 4656: Note that the words that compare for equality (@code{= <> 0= 0<> d= d<>
4657: d0= d0<>}) work for for both signed and unsigned numbers.
1.44 crook 4658:
1.28 crook 4659: doc-<
4660: doc-<=
4661: doc-<>
4662: doc-=
4663: doc->
4664: doc->=
4665:
1.21 crook 4666: doc-0<
1.23 crook 4667: doc-0<=
1.21 crook 4668: doc-0<>
4669: doc-0=
1.23 crook 4670: doc-0>
4671: doc-0>=
1.28 crook 4672:
4673: doc-u<
4674: doc-u<=
1.44 crook 4675: @c u<> and u= exist but are the same as <> and =
1.31 anton 4676: @c doc-u<>
4677: @c doc-u=
1.28 crook 4678: doc-u>
4679: doc-u>=
4680:
4681: doc-within
4682:
4683: doc-d<
4684: doc-d<=
4685: doc-d<>
4686: doc-d=
4687: doc-d>
4688: doc-d>=
1.23 crook 4689:
1.21 crook 4690: doc-d0<
1.23 crook 4691: doc-d0<=
4692: doc-d0<>
1.21 crook 4693: doc-d0=
1.23 crook 4694: doc-d0>
4695: doc-d0>=
4696:
1.21 crook 4697: doc-du<
1.28 crook 4698: doc-du<=
1.44 crook 4699: @c du<> and du= exist but are the same as d<> and d=
1.31 anton 4700: @c doc-du<>
4701: @c doc-du=
1.28 crook 4702: doc-du>
4703: doc-du>=
1.1 anton 4704:
1.44 crook 4705:
1.21 crook 4706: @node Mixed precision, Floating Point, Numeric comparison, Arithmetic
1.1 anton 4707: @subsection Mixed precision
4708: @cindex mixed precision arithmetic words
4709:
1.44 crook 4710:
1.1 anton 4711: doc-m+
4712: doc-*/
4713: doc-*/mod
4714: doc-m*
4715: doc-um*
4716: doc-m*/
4717: doc-um/mod
4718: doc-fm/mod
4719: doc-sm/rem
4720:
1.44 crook 4721:
1.21 crook 4722: @node Floating Point, , Mixed precision, Arithmetic
1.1 anton 4723: @subsection Floating Point
4724: @cindex floating point arithmetic words
4725:
1.49 anton 4726: For the rules used by the text interpreter for
4727: recognising floating-point numbers see @ref{Number Conversion}.
1.1 anton 4728:
1.67 anton 4729: Gforth has a separate floating point stack, but the documentation uses
4730: the unified notation.@footnote{It's easy to generate the separate
4731: notation from that by just separating the floating-point numbers out:
4732: e.g. @code{( n r1 u r2 -- r3 )} becomes @code{( n u -- ) ( F: r1 r2 --
4733: r3 )}.}
1.1 anton 4734:
4735: @cindex floating-point arithmetic, pitfalls
4736: Floating point numbers have a number of unpleasant surprises for the
4737: unwary (e.g., floating point addition is not associative) and even a few
4738: for the wary. You should not use them unless you know what you are doing
4739: or you don't care that the results you get are totally bogus. If you
4740: want to learn about the problems of floating point numbers (and how to
1.66 anton 4741: avoid them), you might start with @cite{David Goldberg,
4742: @uref{http://www.validgh.com/goldberg/paper.ps,What Every Computer
4743: Scientist Should Know About Floating-Point Arithmetic}, ACM Computing
4744: Surveys 23(1):5@minus{}48, March 1991}.
1.1 anton 4745:
1.44 crook 4746:
1.21 crook 4747: doc-d>f
4748: doc-f>d
1.1 anton 4749: doc-f+
4750: doc-f-
4751: doc-f*
4752: doc-f/
4753: doc-fnegate
4754: doc-fabs
4755: doc-fmax
4756: doc-fmin
4757: doc-floor
4758: doc-fround
4759: doc-f**
4760: doc-fsqrt
4761: doc-fexp
4762: doc-fexpm1
4763: doc-fln
4764: doc-flnp1
4765: doc-flog
4766: doc-falog
1.32 anton 4767: doc-f2*
4768: doc-f2/
4769: doc-1/f
4770: doc-precision
4771: doc-set-precision
4772:
4773: @cindex angles in trigonometric operations
4774: @cindex trigonometric operations
4775: Angles in floating point operations are given in radians (a full circle
4776: has 2 pi radians).
4777:
1.1 anton 4778: doc-fsin
4779: doc-fcos
4780: doc-fsincos
4781: doc-ftan
4782: doc-fasin
4783: doc-facos
4784: doc-fatan
4785: doc-fatan2
4786: doc-fsinh
4787: doc-fcosh
4788: doc-ftanh
4789: doc-fasinh
4790: doc-facosh
4791: doc-fatanh
1.21 crook 4792: doc-pi
1.28 crook 4793:
1.32 anton 4794: @cindex equality of floats
4795: @cindex floating-point comparisons
1.31 anton 4796: One particular problem with floating-point arithmetic is that comparison
4797: for equality often fails when you would expect it to succeed. For this
4798: reason approximate equality is often preferred (but you still have to
1.67 anton 4799: know what you are doing). Also note that IEEE NaNs may compare
1.68 anton 4800: differently from what you might expect. The comparison words are:
1.31 anton 4801:
4802: doc-f~rel
4803: doc-f~abs
1.68 anton 4804: doc-f~
1.31 anton 4805: doc-f=
4806: doc-f<>
4807:
4808: doc-f<
4809: doc-f<=
4810: doc-f>
4811: doc-f>=
4812:
1.21 crook 4813: doc-f0<
1.28 crook 4814: doc-f0<=
4815: doc-f0<>
1.21 crook 4816: doc-f0=
1.28 crook 4817: doc-f0>
4818: doc-f0>=
4819:
1.1 anton 4820:
4821: @node Stack Manipulation, Memory, Arithmetic, Words
4822: @section Stack Manipulation
4823: @cindex stack manipulation words
4824:
4825: @cindex floating-point stack in the standard
1.21 crook 4826: Gforth maintains a number of separate stacks:
4827:
1.29 crook 4828: @cindex data stack
4829: @cindex parameter stack
1.21 crook 4830: @itemize @bullet
4831: @item
1.29 crook 4832: A data stack (also known as the @dfn{parameter stack}) -- for
4833: characters, cells, addresses, and double cells.
1.21 crook 4834:
1.29 crook 4835: @cindex floating-point stack
1.21 crook 4836: @item
1.44 crook 4837: A floating point stack -- for holding floating point (FP) numbers.
1.21 crook 4838:
1.29 crook 4839: @cindex return stack
1.21 crook 4840: @item
1.44 crook 4841: A return stack -- for holding the return addresses of colon
1.32 anton 4842: definitions and other (non-FP) data.
1.21 crook 4843:
1.29 crook 4844: @cindex locals stack
1.21 crook 4845: @item
1.44 crook 4846: A locals stack -- for holding local variables.
1.21 crook 4847: @end itemize
4848:
1.1 anton 4849: @menu
4850: * Data stack::
4851: * Floating point stack::
4852: * Return stack::
4853: * Locals stack::
4854: * Stack pointer manipulation::
4855: @end menu
4856:
4857: @node Data stack, Floating point stack, Stack Manipulation, Stack Manipulation
4858: @subsection Data stack
4859: @cindex data stack manipulation words
4860: @cindex stack manipulations words, data stack
4861:
1.44 crook 4862:
1.1 anton 4863: doc-drop
4864: doc-nip
4865: doc-dup
4866: doc-over
4867: doc-tuck
4868: doc-swap
1.21 crook 4869: doc-pick
1.1 anton 4870: doc-rot
4871: doc--rot
4872: doc-?dup
4873: doc-roll
4874: doc-2drop
4875: doc-2nip
4876: doc-2dup
4877: doc-2over
4878: doc-2tuck
4879: doc-2swap
4880: doc-2rot
4881:
1.44 crook 4882:
1.1 anton 4883: @node Floating point stack, Return stack, Data stack, Stack Manipulation
4884: @subsection Floating point stack
4885: @cindex floating-point stack manipulation words
4886: @cindex stack manipulation words, floating-point stack
4887:
1.32 anton 4888: Whilst every sane Forth has a separate floating-point stack, it is not
4889: strictly required; an ANS Forth system could theoretically keep
4890: floating-point numbers on the data stack. As an additional difficulty,
4891: you don't know how many cells a floating-point number takes. It is
4892: reportedly possible to write words in a way that they work also for a
4893: unified stack model, but we do not recommend trying it. Instead, just
4894: say that your program has an environmental dependency on a separate
4895: floating-point stack.
4896:
4897: doc-floating-stack
4898:
1.1 anton 4899: doc-fdrop
4900: doc-fnip
4901: doc-fdup
4902: doc-fover
4903: doc-ftuck
4904: doc-fswap
1.21 crook 4905: doc-fpick
1.1 anton 4906: doc-frot
4907:
1.44 crook 4908:
1.1 anton 4909: @node Return stack, Locals stack, Floating point stack, Stack Manipulation
4910: @subsection Return stack
4911: @cindex return stack manipulation words
4912: @cindex stack manipulation words, return stack
4913:
1.32 anton 4914: @cindex return stack and locals
4915: @cindex locals and return stack
4916: A Forth system is allowed to keep local variables on the
4917: return stack. This is reasonable, as local variables usually eliminate
4918: the need to use the return stack explicitly. So, if you want to produce
4919: a standard compliant program and you are using local variables in a
4920: word, forget about return stack manipulations in that word (refer to the
4921: standard document for the exact rules).
4922:
1.1 anton 4923: doc->r
4924: doc-r>
4925: doc-r@
4926: doc-rdrop
4927: doc-2>r
4928: doc-2r>
4929: doc-2r@
4930: doc-2rdrop
4931:
1.44 crook 4932:
1.1 anton 4933: @node Locals stack, Stack pointer manipulation, Return stack, Stack Manipulation
4934: @subsection Locals stack
4935:
1.47 crook 4936: Gforth uses an extra locals stack. It is described, along with the
4937: reasons for its existence, in @ref{Implementation,Implementation of locals}.
1.21 crook 4938:
1.1 anton 4939: @node Stack pointer manipulation, , Locals stack, Stack Manipulation
4940: @subsection Stack pointer manipulation
4941: @cindex stack pointer manipulation words
4942:
1.44 crook 4943: @c removed s0 r0 l0 -- they are obsolete aliases for sp0 rp0 lp0
1.21 crook 4944: doc-sp0
1.1 anton 4945: doc-sp@
4946: doc-sp!
1.21 crook 4947: doc-fp0
1.1 anton 4948: doc-fp@
4949: doc-fp!
1.21 crook 4950: doc-rp0
1.1 anton 4951: doc-rp@
4952: doc-rp!
1.21 crook 4953: doc-lp0
1.1 anton 4954: doc-lp@
4955: doc-lp!
4956:
1.44 crook 4957:
1.1 anton 4958: @node Memory, Control Structures, Stack Manipulation, Words
4959: @section Memory
1.26 crook 4960: @cindex memory words
1.1 anton 4961:
1.32 anton 4962: @menu
4963: * Memory model::
4964: * Dictionary allocation::
4965: * Heap Allocation::
4966: * Memory Access::
4967: * Address arithmetic::
4968: * Memory Blocks::
4969: @end menu
4970:
1.67 anton 4971: In addition to the standard Forth memory allocation words, there is also
4972: a @uref{http://www.complang.tuwien.ac.at/forth/garbage-collection.zip,
4973: garbage collector}.
4974:
1.32 anton 4975: @node Memory model, Dictionary allocation, Memory, Memory
4976: @subsection ANS Forth and Gforth memory models
4977:
4978: @c The ANS Forth description is a mess (e.g., is the heap part of
4979: @c the dictionary?), so let's not stick to closely with it.
4980:
1.67 anton 4981: ANS Forth considers a Forth system as consisting of several address
4982: spaces, of which only @dfn{data space} is managed and accessible with
4983: the memory words. Memory not necessarily in data space includes the
4984: stacks, the code (called code space) and the headers (called name
4985: space). In Gforth everything is in data space, but the code for the
4986: primitives is usually read-only.
1.32 anton 4987:
4988: Data space is divided into a number of areas: The (data space portion of
4989: the) dictionary@footnote{Sometimes, the term @dfn{dictionary} is used to
4990: refer to the search data structure embodied in word lists and headers,
4991: because it is used for looking up names, just as you would in a
4992: conventional dictionary.}, the heap, and a number of system-allocated
4993: buffers.
4994:
1.68 anton 4995: @cindex address arithmetic restrictions, ANS vs. Gforth
4996: @cindex contiguous regions, ANS vs. Gforth
1.32 anton 4997: In ANS Forth data space is also divided into contiguous regions. You
4998: can only use address arithmetic within a contiguous region, not between
4999: them. Usually each allocation gives you one contiguous region, but the
1.33 anton 5000: dictionary allocation words have additional rules (@pxref{Dictionary
1.32 anton 5001: allocation}).
5002:
5003: Gforth provides one big address space, and address arithmetic can be
5004: performed between any addresses. However, in the dictionary headers or
5005: code are interleaved with data, so almost the only contiguous data space
5006: regions there are those described by ANS Forth as contiguous; but you
5007: can be sure that the dictionary is allocated towards increasing
5008: addresses even between contiguous regions. The memory order of
5009: allocations in the heap is platform-dependent (and possibly different
5010: from one run to the next).
5011:
1.27 crook 5012:
1.32 anton 5013: @node Dictionary allocation, Heap Allocation, Memory model, Memory
5014: @subsection Dictionary allocation
1.27 crook 5015: @cindex reserving data space
5016: @cindex data space - reserving some
5017:
1.32 anton 5018: Dictionary allocation is a stack-oriented allocation scheme, i.e., if
5019: you want to deallocate X, you also deallocate everything
5020: allocated after X.
5021:
1.68 anton 5022: @cindex contiguous regions in dictionary allocation
1.32 anton 5023: The allocations using the words below are contiguous and grow the region
5024: towards increasing addresses. Other words that allocate dictionary
5025: memory of any kind (i.e., defining words including @code{:noname}) end
5026: the contiguous region and start a new one.
5027:
5028: In ANS Forth only @code{create}d words are guaranteed to produce an
5029: address that is the start of the following contiguous region. In
5030: particular, the cell allocated by @code{variable} is not guaranteed to
5031: be contiguous with following @code{allot}ed memory.
5032:
5033: You can deallocate memory by using @code{allot} with a negative argument
5034: (with some restrictions, see @code{allot}). For larger deallocations use
5035: @code{marker}.
1.27 crook 5036:
1.29 crook 5037:
1.27 crook 5038: doc-here
5039: doc-unused
5040: doc-allot
5041: doc-c,
1.29 crook 5042: doc-f,
1.27 crook 5043: doc-,
5044: doc-2,
5045:
1.32 anton 5046: Memory accesses have to be aligned (@pxref{Address arithmetic}). So of
5047: course you should allocate memory in an aligned way, too. I.e., before
5048: allocating allocating a cell, @code{here} must be cell-aligned, etc.
5049: The words below align @code{here} if it is not already. Basically it is
5050: only already aligned for a type, if the last allocation was a multiple
5051: of the size of this type and if @code{here} was aligned for this type
5052: before.
5053:
5054: After freshly @code{create}ing a word, @code{here} is @code{align}ed in
5055: ANS Forth (@code{maxalign}ed in Gforth).
5056:
5057: doc-align
5058: doc-falign
5059: doc-sfalign
5060: doc-dfalign
5061: doc-maxalign
5062: doc-cfalign
5063:
5064:
5065: @node Heap Allocation, Memory Access, Dictionary allocation, Memory
5066: @subsection Heap allocation
5067: @cindex heap allocation
5068: @cindex dynamic allocation of memory
5069: @cindex memory-allocation word set
5070:
1.68 anton 5071: @cindex contiguous regions and heap allocation
1.32 anton 5072: Heap allocation supports deallocation of allocated memory in any
5073: order. Dictionary allocation is not affected by it (i.e., it does not
5074: end a contiguous region). In Gforth, these words are implemented using
5075: the standard C library calls malloc(), free() and resize().
5076:
1.68 anton 5077: The memory region produced by one invocation of @code{allocate} or
5078: @code{resize} is internally contiguous. There is no contiguity between
5079: such a region and any other region (including others allocated from the
5080: heap).
5081:
1.32 anton 5082: doc-allocate
5083: doc-free
5084: doc-resize
5085:
1.27 crook 5086:
1.32 anton 5087: @node Memory Access, Address arithmetic, Heap Allocation, Memory
1.1 anton 5088: @subsection Memory Access
5089: @cindex memory access words
5090:
5091: doc-@
5092: doc-!
5093: doc-+!
5094: doc-c@
5095: doc-c!
5096: doc-2@
5097: doc-2!
5098: doc-f@
5099: doc-f!
5100: doc-sf@
5101: doc-sf!
5102: doc-df@
5103: doc-df!
5104:
1.68 anton 5105:
1.32 anton 5106: @node Address arithmetic, Memory Blocks, Memory Access, Memory
5107: @subsection Address arithmetic
1.1 anton 5108: @cindex address arithmetic words
5109:
1.67 anton 5110: Address arithmetic is the foundation on which you can build data
5111: structures like arrays, records (@pxref{Structures}) and objects
5112: (@pxref{Object-oriented Forth}).
1.32 anton 5113:
1.68 anton 5114: @cindex address unit
5115: @cindex au (address unit)
1.1 anton 5116: ANS Forth does not specify the sizes of the data types. Instead, it
5117: offers a number of words for computing sizes and doing address
1.29 crook 5118: arithmetic. Address arithmetic is performed in terms of address units
5119: (aus); on most systems the address unit is one byte. Note that a
5120: character may have more than one au, so @code{chars} is no noop (on
1.68 anton 5121: platforms where it is a noop, it compiles to nothing).
1.1 anton 5122:
1.67 anton 5123: The basic address arithmetic words are @code{+} and @code{-}. E.g., if
5124: you have the address of a cell, perform @code{1 cells +}, and you will
5125: have the address of the next cell.
5126:
1.68 anton 5127: @cindex contiguous regions and address arithmetic
1.67 anton 5128: In ANS Forth you can perform address arithmetic only within a contiguous
5129: region, i.e., if you have an address into one region, you can only add
5130: and subtract such that the result is still within the region; you can
5131: only subtract or compare addresses from within the same contiguous
5132: region. Reasons: several contiguous regions can be arranged in memory
5133: in any way; on segmented systems addresses may have unusual
5134: representations, such that address arithmetic only works within a
5135: region. Gforth provides a few more guarantees (linear address space,
5136: dictionary grows upwards), but in general I have found it easy to stay
5137: within contiguous regions (exception: computing and comparing to the
5138: address just beyond the end of an array).
5139:
1.1 anton 5140: @cindex alignment of addresses for types
5141: ANS Forth also defines words for aligning addresses for specific
5142: types. Many computers require that accesses to specific data types
5143: must only occur at specific addresses; e.g., that cells may only be
5144: accessed at addresses divisible by 4. Even if a machine allows unaligned
5145: accesses, it can usually perform aligned accesses faster.
5146:
5147: For the performance-conscious: alignment operations are usually only
5148: necessary during the definition of a data structure, not during the
5149: (more frequent) accesses to it.
5150:
5151: ANS Forth defines no words for character-aligning addresses. This is not
5152: an oversight, but reflects the fact that addresses that are not
5153: char-aligned have no use in the standard and therefore will not be
5154: created.
5155:
5156: @cindex @code{CREATE} and alignment
1.29 crook 5157: ANS Forth guarantees that addresses returned by @code{CREATE}d words
1.1 anton 5158: are cell-aligned; in addition, Gforth guarantees that these addresses
5159: are aligned for all purposes.
5160:
1.26 crook 5161: Note that the ANS Forth word @code{char} has nothing to do with address
5162: arithmetic.
1.1 anton 5163:
1.44 crook 5164:
1.1 anton 5165: doc-chars
5166: doc-char+
5167: doc-cells
5168: doc-cell+
5169: doc-cell
5170: doc-aligned
5171: doc-floats
5172: doc-float+
5173: doc-float
5174: doc-faligned
5175: doc-sfloats
5176: doc-sfloat+
5177: doc-sfaligned
5178: doc-dfloats
5179: doc-dfloat+
5180: doc-dfaligned
5181: doc-maxaligned
5182: doc-cfaligned
5183: doc-address-unit-bits
5184:
1.44 crook 5185:
1.32 anton 5186: @node Memory Blocks, , Address arithmetic, Memory
1.1 anton 5187: @subsection Memory Blocks
5188: @cindex memory block words
1.27 crook 5189: @cindex character strings - moving and copying
5190:
1.49 anton 5191: Memory blocks often represent character strings; For ways of storing
5192: character strings in memory see @ref{String Formats}. For other
5193: string-processing words see @ref{Displaying characters and strings}.
1.1 anton 5194:
1.67 anton 5195: A few of these words work on address unit blocks. In that case, you
5196: usually have to insert @code{CHARS} before the word when working on
5197: character strings. Most words work on character blocks, and expect a
5198: char-aligned address.
5199:
5200: When copying characters between overlapping memory regions, use
5201: @code{chars move} or choose carefully between @code{cmove} and
5202: @code{cmove>}.
1.44 crook 5203:
1.1 anton 5204: doc-move
5205: doc-erase
5206: doc-cmove
5207: doc-cmove>
5208: doc-fill
5209: doc-blank
1.21 crook 5210: doc-compare
5211: doc-search
1.27 crook 5212: doc--trailing
5213: doc-/string
5214:
1.44 crook 5215:
1.27 crook 5216: @comment TODO examples
5217:
1.1 anton 5218:
1.26 crook 5219: @node Control Structures, Defining Words, Memory, Words
1.1 anton 5220: @section Control Structures
5221: @cindex control structures
5222:
1.33 anton 5223: Control structures in Forth cannot be used interpretively, only in a
5224: colon definition@footnote{To be precise, they have no interpretation
5225: semantics (@pxref{Interpretation and Compilation Semantics}).}. We do
5226: not like this limitation, but have not seen a satisfying way around it
5227: yet, although many schemes have been proposed.
1.1 anton 5228:
5229: @menu
1.33 anton 5230: * Selection:: IF ... ELSE ... ENDIF
5231: * Simple Loops:: BEGIN ...
1.29 crook 5232: * Counted Loops:: DO
1.67 anton 5233: * Arbitrary control structures::
5234: * Calls and returns::
1.1 anton 5235: * Exception Handling::
5236: @end menu
5237:
5238: @node Selection, Simple Loops, Control Structures, Control Structures
5239: @subsection Selection
5240: @cindex selection control structures
5241: @cindex control structures for selection
5242:
5243: @cindex @code{IF} control structure
5244: @example
1.29 crook 5245: @i{flag}
1.1 anton 5246: IF
1.29 crook 5247: @i{code}
1.1 anton 5248: ENDIF
5249: @end example
1.21 crook 5250: @noindent
1.33 anton 5251:
1.44 crook 5252: If @i{flag} is non-zero (as far as @code{IF} etc. are concerned, a cell
5253: with any bit set represents truth) @i{code} is executed.
1.33 anton 5254:
1.1 anton 5255: @example
1.29 crook 5256: @i{flag}
1.1 anton 5257: IF
1.29 crook 5258: @i{code1}
1.1 anton 5259: ELSE
1.29 crook 5260: @i{code2}
1.1 anton 5261: ENDIF
5262: @end example
5263:
1.44 crook 5264: If @var{flag} is true, @i{code1} is executed, otherwise @i{code2} is
5265: executed.
1.33 anton 5266:
1.1 anton 5267: You can use @code{THEN} instead of @code{ENDIF}. Indeed, @code{THEN} is
5268: standard, and @code{ENDIF} is not, although it is quite popular. We
5269: recommend using @code{ENDIF}, because it is less confusing for people
5270: who also know other languages (and is not prone to reinforcing negative
5271: prejudices against Forth in these people). Adding @code{ENDIF} to a
5272: system that only supplies @code{THEN} is simple:
5273: @example
1.21 crook 5274: : ENDIF POSTPONE THEN ; immediate
1.1 anton 5275: @end example
5276:
5277: [According to @cite{Webster's New Encyclopedic Dictionary}, @dfn{then
5278: (adv.)} has the following meanings:
5279: @quotation
5280: ... 2b: following next after in order ... 3d: as a necessary consequence
5281: (if you were there, then you saw them).
5282: @end quotation
5283: Forth's @code{THEN} has the meaning 2b, whereas @code{THEN} in Pascal
5284: and many other programming languages has the meaning 3d.]
5285:
1.21 crook 5286: Gforth also provides the words @code{?DUP-IF} and @code{?DUP-0=-IF}, so
1.1 anton 5287: you can avoid using @code{?dup}. Using these alternatives is also more
1.26 crook 5288: efficient than using @code{?dup}. Definitions in ANS Forth
1.1 anton 5289: for @code{ENDIF}, @code{?DUP-IF} and @code{?DUP-0=-IF} are provided in
5290: @file{compat/control.fs}.
5291:
5292: @cindex @code{CASE} control structure
5293: @example
1.29 crook 5294: @i{n}
1.1 anton 5295: CASE
1.29 crook 5296: @i{n1} OF @i{code1} ENDOF
5297: @i{n2} OF @i{code2} ENDOF
1.1 anton 5298: @dots{}
1.68 anton 5299: ( n ) @i{default-code} ( n )
1.1 anton 5300: ENDCASE
5301: @end example
5302:
1.68 anton 5303: Executes the first @i{codei}, where the @i{ni} is equal to @i{n}. If no
5304: @i{ni} matches, the optional @i{default-code} is executed. The optional
5305: default case can be added by simply writing the code after the last
5306: @code{ENDOF}. It may use @i{n}, which is on top of the stack, but must
5307: not consume it.
1.1 anton 5308:
1.69 anton 5309: @progstyle
5310: To keep the code understandable, you should ensure that on all paths
5311: through a selection construct the stack is changed in the same way
5312: (wrt. number and types of stack items consumed and pushed).
5313:
1.1 anton 5314: @node Simple Loops, Counted Loops, Selection, Control Structures
5315: @subsection Simple Loops
5316: @cindex simple loops
5317: @cindex loops without count
5318:
5319: @cindex @code{WHILE} loop
5320: @example
5321: BEGIN
1.29 crook 5322: @i{code1}
5323: @i{flag}
1.1 anton 5324: WHILE
1.29 crook 5325: @i{code2}
1.1 anton 5326: REPEAT
5327: @end example
5328:
1.29 crook 5329: @i{code1} is executed and @i{flag} is computed. If it is true,
5330: @i{code2} is executed and the loop is restarted; If @i{flag} is
1.1 anton 5331: false, execution continues after the @code{REPEAT}.
5332:
5333: @cindex @code{UNTIL} loop
5334: @example
5335: BEGIN
1.29 crook 5336: @i{code}
5337: @i{flag}
1.1 anton 5338: UNTIL
5339: @end example
5340:
1.29 crook 5341: @i{code} is executed. The loop is restarted if @code{flag} is false.
1.1 anton 5342:
1.69 anton 5343: @progstyle
5344: To keep the code understandable, a complete iteration of the loop should
5345: not change the number and types of the items on the stacks.
5346:
1.1 anton 5347: @cindex endless loop
5348: @cindex loops, endless
5349: @example
5350: BEGIN
1.29 crook 5351: @i{code}
1.1 anton 5352: AGAIN
5353: @end example
5354:
5355: This is an endless loop.
5356:
5357: @node Counted Loops, Arbitrary control structures, Simple Loops, Control Structures
5358: @subsection Counted Loops
5359: @cindex counted loops
5360: @cindex loops, counted
5361: @cindex @code{DO} loops
5362:
5363: The basic counted loop is:
5364: @example
1.29 crook 5365: @i{limit} @i{start}
1.1 anton 5366: ?DO
1.29 crook 5367: @i{body}
1.1 anton 5368: LOOP
5369: @end example
5370:
1.29 crook 5371: This performs one iteration for every integer, starting from @i{start}
5372: and up to, but excluding @i{limit}. The counter, or @i{index}, can be
1.21 crook 5373: accessed with @code{i}. For example, the loop:
1.1 anton 5374: @example
5375: 10 0 ?DO
5376: i .
5377: LOOP
5378: @end example
1.21 crook 5379: @noindent
5380: prints @code{0 1 2 3 4 5 6 7 8 9}
5381:
1.1 anton 5382: The index of the innermost loop can be accessed with @code{i}, the index
5383: of the next loop with @code{j}, and the index of the third loop with
5384: @code{k}.
5385:
1.44 crook 5386:
1.1 anton 5387: doc-i
5388: doc-j
5389: doc-k
5390:
1.44 crook 5391:
1.1 anton 5392: The loop control data are kept on the return stack, so there are some
1.21 crook 5393: restrictions on mixing return stack accesses and counted loop words. In
5394: particuler, if you put values on the return stack outside the loop, you
5395: cannot read them inside the loop@footnote{well, not in a way that is
5396: portable.}. If you put values on the return stack within a loop, you
5397: have to remove them before the end of the loop and before accessing the
5398: index of the loop.
1.1 anton 5399:
5400: There are several variations on the counted loop:
5401:
1.21 crook 5402: @itemize @bullet
5403: @item
5404: @code{LEAVE} leaves the innermost counted loop immediately; execution
5405: continues after the associated @code{LOOP} or @code{NEXT}. For example:
5406:
5407: @example
5408: 10 0 ?DO i DUP . 3 = IF LEAVE THEN LOOP
5409: @end example
5410: prints @code{0 1 2 3}
5411:
1.1 anton 5412:
1.21 crook 5413: @item
5414: @code{UNLOOP} prepares for an abnormal loop exit, e.g., via
5415: @code{EXIT}. @code{UNLOOP} removes the loop control parameters from the
5416: return stack so @code{EXIT} can get to its return address. For example:
5417:
5418: @example
5419: : demo 10 0 ?DO i DUP . 3 = IF UNLOOP EXIT THEN LOOP ." Done" ;
5420: @end example
5421: prints @code{0 1 2 3}
5422:
5423:
5424: @item
1.29 crook 5425: If @i{start} is greater than @i{limit}, a @code{?DO} loop is entered
1.1 anton 5426: (and @code{LOOP} iterates until they become equal by wrap-around
5427: arithmetic). This behaviour is usually not what you want. Therefore,
5428: Gforth offers @code{+DO} and @code{U+DO} (as replacements for
1.29 crook 5429: @code{?DO}), which do not enter the loop if @i{start} is greater than
5430: @i{limit}; @code{+DO} is for signed loop parameters, @code{U+DO} for
1.1 anton 5431: unsigned loop parameters.
5432:
1.21 crook 5433: @item
5434: @code{?DO} can be replaced by @code{DO}. @code{DO} always enters
5435: the loop, independent of the loop parameters. Do not use @code{DO}, even
5436: if you know that the loop is entered in any case. Such knowledge tends
5437: to become invalid during maintenance of a program, and then the
5438: @code{DO} will make trouble.
5439:
5440: @item
1.29 crook 5441: @code{LOOP} can be replaced with @code{@i{n} +LOOP}; this updates the
5442: index by @i{n} instead of by 1. The loop is terminated when the border
5443: between @i{limit-1} and @i{limit} is crossed. E.g.:
1.1 anton 5444:
1.21 crook 5445: @example
5446: 4 0 +DO i . 2 +LOOP
5447: @end example
5448: @noindent
5449: prints @code{0 2}
5450:
5451: @example
5452: 4 1 +DO i . 2 +LOOP
5453: @end example
5454: @noindent
5455: prints @code{1 3}
1.1 anton 5456:
1.68 anton 5457: @item
1.1 anton 5458: @cindex negative increment for counted loops
5459: @cindex counted loops with negative increment
1.29 crook 5460: The behaviour of @code{@i{n} +LOOP} is peculiar when @i{n} is negative:
1.1 anton 5461:
1.21 crook 5462: @example
5463: -1 0 ?DO i . -1 +LOOP
5464: @end example
5465: @noindent
5466: prints @code{0 -1}
1.1 anton 5467:
1.21 crook 5468: @example
5469: 0 0 ?DO i . -1 +LOOP
5470: @end example
5471: prints nothing.
1.1 anton 5472:
1.29 crook 5473: Therefore we recommend avoiding @code{@i{n} +LOOP} with negative
5474: @i{n}. One alternative is @code{@i{u} -LOOP}, which reduces the
5475: index by @i{u} each iteration. The loop is terminated when the border
5476: between @i{limit+1} and @i{limit} is crossed. Gforth also provides
1.1 anton 5477: @code{-DO} and @code{U-DO} for down-counting loops. E.g.:
5478:
1.21 crook 5479: @example
5480: -2 0 -DO i . 1 -LOOP
5481: @end example
5482: @noindent
5483: prints @code{0 -1}
1.1 anton 5484:
1.21 crook 5485: @example
5486: -1 0 -DO i . 1 -LOOP
5487: @end example
5488: @noindent
5489: prints @code{0}
5490:
5491: @example
5492: 0 0 -DO i . 1 -LOOP
5493: @end example
5494: @noindent
5495: prints nothing.
1.1 anton 5496:
1.21 crook 5497: @end itemize
1.1 anton 5498:
5499: Unfortunately, @code{+DO}, @code{U+DO}, @code{-DO}, @code{U-DO} and
1.26 crook 5500: @code{-LOOP} are not defined in ANS Forth. However, an implementation
5501: for these words that uses only standard words is provided in
5502: @file{compat/loops.fs}.
1.1 anton 5503:
5504:
5505: @cindex @code{FOR} loops
1.26 crook 5506: Another counted loop is:
1.1 anton 5507: @example
1.29 crook 5508: @i{n}
1.1 anton 5509: FOR
1.29 crook 5510: @i{body}
1.1 anton 5511: NEXT
5512: @end example
5513: This is the preferred loop of native code compiler writers who are too
1.26 crook 5514: lazy to optimize @code{?DO} loops properly. This loop structure is not
1.29 crook 5515: defined in ANS Forth. In Gforth, this loop iterates @i{n+1} times;
5516: @code{i} produces values starting with @i{n} and ending with 0. Other
1.26 crook 5517: Forth systems may behave differently, even if they support @code{FOR}
5518: loops. To avoid problems, don't use @code{FOR} loops.
1.1 anton 5519:
5520: @node Arbitrary control structures, Calls and returns, Counted Loops, Control Structures
5521: @subsection Arbitrary control structures
5522: @cindex control structures, user-defined
5523:
5524: @cindex control-flow stack
5525: ANS Forth permits and supports using control structures in a non-nested
5526: way. Information about incomplete control structures is stored on the
5527: control-flow stack. This stack may be implemented on the Forth data
5528: stack, and this is what we have done in Gforth.
5529:
5530: @cindex @code{orig}, control-flow stack item
5531: @cindex @code{dest}, control-flow stack item
5532: An @i{orig} entry represents an unresolved forward branch, a @i{dest}
5533: entry represents a backward branch target. A few words are the basis for
5534: building any control structure possible (except control structures that
5535: need storage, like calls, coroutines, and backtracking).
5536:
1.44 crook 5537:
1.1 anton 5538: doc-if
5539: doc-ahead
5540: doc-then
5541: doc-begin
5542: doc-until
5543: doc-again
5544: doc-cs-pick
5545: doc-cs-roll
5546:
1.44 crook 5547:
1.21 crook 5548: The Standard words @code{CS-PICK} and @code{CS-ROLL} allow you to
5549: manipulate the control-flow stack in a portable way. Without them, you
5550: would need to know how many stack items are occupied by a control-flow
5551: entry (many systems use one cell. In Gforth they currently take three,
5552: but this may change in the future).
5553:
1.1 anton 5554: Some standard control structure words are built from these words:
5555:
1.44 crook 5556:
1.1 anton 5557: doc-else
5558: doc-while
5559: doc-repeat
5560:
1.44 crook 5561:
5562: @noindent
1.1 anton 5563: Gforth adds some more control-structure words:
5564:
1.44 crook 5565:
1.1 anton 5566: doc-endif
5567: doc-?dup-if
5568: doc-?dup-0=-if
5569:
1.44 crook 5570:
5571: @noindent
1.1 anton 5572: Counted loop words constitute a separate group of words:
5573:
1.44 crook 5574:
1.1 anton 5575: doc-?do
5576: doc-+do
5577: doc-u+do
5578: doc--do
5579: doc-u-do
5580: doc-do
5581: doc-for
5582: doc-loop
5583: doc-+loop
5584: doc--loop
5585: doc-next
5586: doc-leave
5587: doc-?leave
5588: doc-unloop
5589: doc-done
5590:
1.44 crook 5591:
1.21 crook 5592: The standard does not allow using @code{CS-PICK} and @code{CS-ROLL} on
5593: @i{do-sys}. Gforth allows it, but it's your job to ensure that for
1.1 anton 5594: every @code{?DO} etc. there is exactly one @code{UNLOOP} on any path
5595: through the definition (@code{LOOP} etc. compile an @code{UNLOOP} on the
5596: fall-through path). Also, you have to ensure that all @code{LEAVE}s are
5597: resolved (by using one of the loop-ending words or @code{DONE}).
5598:
1.44 crook 5599: @noindent
1.26 crook 5600: Another group of control structure words are:
1.1 anton 5601:
1.44 crook 5602:
1.1 anton 5603: doc-case
5604: doc-endcase
5605: doc-of
5606: doc-endof
5607:
1.44 crook 5608:
1.21 crook 5609: @i{case-sys} and @i{of-sys} cannot be processed using @code{CS-PICK} and
5610: @code{CS-ROLL}.
1.1 anton 5611:
5612: @subsubsection Programming Style
1.47 crook 5613: @cindex control structures programming style
5614: @cindex programming style, arbitrary control structures
1.1 anton 5615:
5616: In order to ensure readability we recommend that you do not create
5617: arbitrary control structures directly, but define new control structure
5618: words for the control structure you want and use these words in your
1.26 crook 5619: program. For example, instead of writing:
1.1 anton 5620:
5621: @example
1.26 crook 5622: BEGIN
1.1 anton 5623: ...
1.26 crook 5624: IF [ 1 CS-ROLL ]
1.1 anton 5625: ...
1.26 crook 5626: AGAIN THEN
1.1 anton 5627: @end example
5628:
1.21 crook 5629: @noindent
1.1 anton 5630: we recommend defining control structure words, e.g.,
5631:
5632: @example
1.26 crook 5633: : WHILE ( DEST -- ORIG DEST )
5634: POSTPONE IF
5635: 1 CS-ROLL ; immediate
5636:
5637: : REPEAT ( orig dest -- )
5638: POSTPONE AGAIN
5639: POSTPONE THEN ; immediate
1.1 anton 5640: @end example
5641:
1.21 crook 5642: @noindent
1.1 anton 5643: and then using these to create the control structure:
5644:
5645: @example
1.26 crook 5646: BEGIN
1.1 anton 5647: ...
1.26 crook 5648: WHILE
1.1 anton 5649: ...
1.26 crook 5650: REPEAT
1.1 anton 5651: @end example
5652:
5653: That's much easier to read, isn't it? Of course, @code{REPEAT} and
5654: @code{WHILE} are predefined, so in this example it would not be
5655: necessary to define them.
5656:
5657: @node Calls and returns, Exception Handling, Arbitrary control structures, Control Structures
5658: @subsection Calls and returns
5659: @cindex calling a definition
5660: @cindex returning from a definition
5661:
1.3 anton 5662: @cindex recursive definitions
5663: A definition can be called simply be writing the name of the definition
1.26 crook 5664: to be called. Normally a definition is invisible during its own
1.3 anton 5665: definition. If you want to write a directly recursive definition, you
1.26 crook 5666: can use @code{recursive} to make the current definition visible, or
5667: @code{recurse} to call the current definition directly.
1.3 anton 5668:
1.44 crook 5669:
1.3 anton 5670: doc-recursive
5671: doc-recurse
5672:
1.44 crook 5673:
1.21 crook 5674: @comment TODO add example of the two recursion methods
1.12 anton 5675: @quotation
5676: @progstyle
5677: I prefer using @code{recursive} to @code{recurse}, because calling the
5678: definition by name is more descriptive (if the name is well-chosen) than
5679: the somewhat cryptic @code{recurse}. E.g., in a quicksort
5680: implementation, it is much better to read (and think) ``now sort the
5681: partitions'' than to read ``now do a recursive call''.
5682: @end quotation
1.3 anton 5683:
1.29 crook 5684: For mutual recursion, use @code{Defer}red words, like this:
1.3 anton 5685:
5686: @example
1.28 crook 5687: Defer foo
1.3 anton 5688:
5689: : bar ( ... -- ... )
5690: ... foo ... ;
5691:
5692: :noname ( ... -- ... )
5693: ... bar ... ;
5694: IS foo
5695: @end example
5696:
1.44 crook 5697: Deferred words are discussed in more detail in @ref{Deferred words}.
1.33 anton 5698:
1.26 crook 5699: The current definition returns control to the calling definition when
1.33 anton 5700: the end of the definition is reached or @code{EXIT} is encountered.
1.1 anton 5701:
5702: doc-exit
5703: doc-;s
5704:
1.44 crook 5705:
1.1 anton 5706: @node Exception Handling, , Calls and returns, Control Structures
5707: @subsection Exception Handling
1.26 crook 5708: @cindex exceptions
1.1 anton 5709:
1.68 anton 5710: @c quit is a very bad idea for error handling,
5711: @c because it does not translate into a THROW
5712: @c it also does not belong into this chapter
5713:
5714: If a word detects an error condition that it cannot handle, it can
5715: @code{throw} an exception. In the simplest case, this will terminate
5716: your program, and report an appropriate error.
1.21 crook 5717:
1.68 anton 5718: doc-throw
1.1 anton 5719:
1.69 anton 5720: @code{Throw} consumes a cell-sized error number on the stack. There are
5721: some predefined error numbers in ANS Forth (see @file{errors.fs}). In
5722: Gforth (and most other systems) you can use the iors produced by various
5723: words as error numbers (e.g., a typical use of @code{allocate} is
5724: @code{allocate throw}). Gforth also provides the word @code{exception}
5725: to define your own error numbers (with decent error reporting); an ANS
5726: Forth version of this word (but without the error messages) is available
5727: in @code{compat/except.fs}. And finally, you can use your own error
1.68 anton 5728: numbers (anything outside the range -4095..0), but won't get nice error
5729: messages, only numbers. For example, try:
5730:
5731: @example
1.69 anton 5732: -10 throw \ ANS defined
5733: -267 throw \ system defined
5734: s" my error" exception throw \ user defined
5735: 7 throw \ arbitrary number
1.68 anton 5736: @end example
5737:
5738: doc---exception-exception
1.1 anton 5739:
1.69 anton 5740: A common idiom to @code{THROW} a specific error if a flag is true is
5741: this:
5742:
5743: @example
5744: @code{( flag ) 0<> @i{errno} and throw}
5745: @end example
5746:
5747: Your program can provide exception handlers to catch exceptions. An
5748: exception handler can be used to correct the problem, or to clean up
5749: some data structures and just throw the exception to the next exception
5750: handler. Note that @code{throw} jumps to the dynamically innermost
5751: exception handler. The system's exception handler is outermost, and just
5752: prints an error and restarts command-line interpretation (or, in batch
5753: mode (i.e., while processing the shell command line), leaves Gforth).
1.1 anton 5754:
1.68 anton 5755: The ANS Forth way to catch exceptions is @code{catch}:
1.1 anton 5756:
1.68 anton 5757: doc-catch
5758:
5759: The most common use of exception handlers is to clean up the state when
5760: an error happens. E.g.,
1.1 anton 5761:
1.26 crook 5762: @example
1.68 anton 5763: base @ >r hex \ actually the hex should be inside foo, or we h
5764: ['] foo catch ( nerror|0 )
5765: r> base !
1.69 anton 5766: ( nerror|0 ) throw \ pass it on
1.26 crook 5767: @end example
1.1 anton 5768:
1.69 anton 5769: A use of @code{catch} for handling the error @code{myerror} might look
5770: like this:
1.44 crook 5771:
1.68 anton 5772: @example
1.69 anton 5773: ['] foo catch
5774: CASE
5775: myerror OF ... ( do something about it ) ENDOF
5776: dup throw \ default: pass other errors on, do nothing on non-errors
5777: ENDCASE
1.68 anton 5778: @end example
1.44 crook 5779:
1.68 anton 5780: Having to wrap the code into a separate word is often cumbersome,
5781: therefore Gforth provides an alternative syntax:
1.1 anton 5782:
5783: @example
1.69 anton 5784: TRY
1.68 anton 5785: @i{code1}
1.69 anton 5786: RECOVER \ optional
1.68 anton 5787: @i{code2} \ optional
1.69 anton 5788: ENDTRY
1.1 anton 5789: @end example
5790:
1.68 anton 5791: This performs @i{Code1}. If @i{code1} completes normally, execution
5792: continues after the @code{endtry}. If @i{Code1} throws, the stacks are
5793: reset to the state during @code{try}, the throw value is pushed on the
5794: data stack, and execution constinues at @i{code2}, and finally falls
5795: through the @code{endtry} into the following code. If there is no
5796: @code{recover} clause, this works like an empty recover clause.
1.26 crook 5797:
1.68 anton 5798: doc-try
5799: doc-recover
5800: doc-endtry
1.26 crook 5801:
1.69 anton 5802: The cleanup example from above in this syntax:
1.26 crook 5803:
1.68 anton 5804: @example
1.69 anton 5805: base @ >r TRY
1.68 anton 5806: hex foo \ now the hex is placed correctly
1.69 anton 5807: 0 \ value for throw
5808: ENDTRY
1.68 anton 5809: r> base ! throw
1.1 anton 5810: @end example
5811:
1.69 anton 5812: And here's the error handling example:
1.1 anton 5813:
1.68 anton 5814: @example
1.69 anton 5815: TRY
1.68 anton 5816: foo
1.69 anton 5817: RECOVER
5818: CASE
5819: myerror OF ... ( do something about it ) ENDOF
5820: throw \ pass other errors on
5821: ENDCASE
5822: ENDTRY
1.68 anton 5823: @end example
1.1 anton 5824:
1.69 anton 5825: @progstyle
5826: As usual, you should ensure that the stack depth is statically known at
5827: the end: either after the @code{throw} for passing on errors, or after
5828: the @code{ENDTRY} (or, if you use @code{catch}, after the end of the
5829: selection construct for handling the error).
5830:
1.68 anton 5831: There are two alternatives to @code{throw}: @code{Abort"} is conditional
5832: and you can provide an error message. @code{Abort} just produces an
5833: ``Aborted'' error.
1.1 anton 5834:
1.68 anton 5835: The problem with these words is that exception handlers cannot
5836: differentiate between different @code{abort"}s; they just look like
5837: @code{-2 throw} to them (the error message cannot be accessed by
5838: standard programs). Similar @code{abort} looks like @code{-1 throw} to
5839: exception handlers.
1.44 crook 5840:
1.68 anton 5841: doc-abort"
1.26 crook 5842: doc-abort
1.29 crook 5843:
5844:
1.44 crook 5845:
1.29 crook 5846: @c -------------------------------------------------------------
1.47 crook 5847: @node Defining Words, Interpretation and Compilation Semantics, Control Structures, Words
1.29 crook 5848: @section Defining Words
5849: @cindex defining words
5850:
1.47 crook 5851: Defining words are used to extend Forth by creating new entries in the dictionary.
5852:
1.29 crook 5853: @menu
1.67 anton 5854: * CREATE::
1.44 crook 5855: * Variables:: Variables and user variables
1.67 anton 5856: * Constants::
1.44 crook 5857: * Values:: Initialised variables
1.67 anton 5858: * Colon Definitions::
1.44 crook 5859: * Anonymous Definitions:: Definitions without names
1.69 anton 5860: * Supplying names:: Passing definition names as strings
1.67 anton 5861: * User-defined Defining Words::
1.44 crook 5862: * Deferred words:: Allow forward references
1.67 anton 5863: * Aliases::
1.29 crook 5864: @end menu
5865:
1.44 crook 5866: @node CREATE, Variables, Defining Words, Defining Words
5867: @subsection @code{CREATE}
1.29 crook 5868: @cindex simple defining words
5869: @cindex defining words, simple
5870:
5871: Defining words are used to create new entries in the dictionary. The
5872: simplest defining word is @code{CREATE}. @code{CREATE} is used like
5873: this:
5874:
5875: @example
5876: CREATE new-word1
5877: @end example
5878:
1.69 anton 5879: @code{CREATE} is a parsing word, i.e., it takes an argument from the
5880: input stream (@code{new-word1} in our example). It generates a
5881: dictionary entry for @code{new-word1}. When @code{new-word1} is
5882: executed, all that it does is leave an address on the stack. The address
5883: represents the value of the data space pointer (@code{HERE}) at the time
5884: that @code{new-word1} was defined. Therefore, @code{CREATE} is a way of
5885: associating a name with the address of a region of memory.
1.29 crook 5886:
1.34 anton 5887: doc-create
5888:
1.69 anton 5889: Note that in ANS Forth guarantees only for @code{create} that its body
5890: is in dictionary data space (i.e., where @code{here}, @code{allot}
5891: etc. work, @pxref{Dictionary allocation}). Also, in ANS Forth only
5892: @code{create}d words can be modified with @code{does>}
5893: (@pxref{User-defined Defining Words}). And in ANS Forth @code{>body}
5894: can only be applied to @code{create}d words.
5895:
1.29 crook 5896: By extending this example to reserve some memory in data space, we end
1.69 anton 5897: up with something like a @i{variable}. Here are two different ways to do
5898: it:
1.29 crook 5899:
5900: @example
5901: CREATE new-word2 1 cells allot \ reserve 1 cell - initial value undefined
5902: CREATE new-word3 4 , \ reserve 1 cell and initialise it (to 4)
5903: @end example
5904:
5905: The variable can be examined and modified using @code{@@} (``fetch'') and
5906: @code{!} (``store'') like this:
5907:
5908: @example
5909: new-word2 @@ . \ get address, fetch from it and display
5910: 1234 new-word2 ! \ new value, get address, store to it
5911: @end example
5912:
1.44 crook 5913: @cindex arrays
5914: A similar mechanism can be used to create arrays. For example, an
5915: 80-character text input buffer:
1.29 crook 5916:
5917: @example
1.44 crook 5918: CREATE text-buf 80 chars allot
5919:
5920: text-buf 0 chars c@@ \ the 1st character (offset 0)
5921: text-buf 3 chars c@@ \ the 4th character (offset 3)
5922: @end example
1.29 crook 5923:
1.44 crook 5924: You can build arbitrarily complex data structures by allocating
1.49 anton 5925: appropriate areas of memory. For further discussions of this, and to
1.66 anton 5926: learn about some Gforth tools that make it easier,
1.49 anton 5927: @xref{Structures}.
1.44 crook 5928:
5929:
5930: @node Variables, Constants, CREATE, Defining Words
5931: @subsection Variables
5932: @cindex variables
5933:
5934: The previous section showed how a sequence of commands could be used to
5935: generate a variable. As a final refinement, the whole code sequence can
5936: be wrapped up in a defining word (pre-empting the subject of the next
5937: section), making it easier to create new variables:
5938:
5939: @example
5940: : myvariableX ( "name" -- a-addr ) CREATE 1 cells allot ;
5941: : myvariable0 ( "name" -- a-addr ) CREATE 0 , ;
5942:
5943: myvariableX foo \ variable foo starts off with an unknown value
5944: myvariable0 joe \ whilst joe is initialised to 0
1.29 crook 5945:
5946: 45 3 * foo ! \ set foo to 135
5947: 1234 joe ! \ set joe to 1234
5948: 3 joe +! \ increment joe by 3.. to 1237
5949: @end example
5950:
5951: Not surprisingly, there is no need to define @code{myvariable}, since
1.44 crook 5952: Forth already has a definition @code{Variable}. ANS Forth does not
1.69 anton 5953: guarantee that a @code{Variable} is initialised when it is created
5954: (i.e., it may behave like @code{myvariableX}). In contrast, Gforth's
5955: @code{Variable} initialises the variable to 0 (i.e., it behaves exactly
5956: like @code{myvariable0}). Forth also provides @code{2Variable} and
1.47 crook 5957: @code{fvariable} for double and floating-point variables, respectively
1.69 anton 5958: -- they are initialised to 0. and 0e in Gforth. If you use a @code{Variable} to
1.47 crook 5959: store a boolean, you can use @code{on} and @code{off} to toggle its
5960: state.
1.29 crook 5961:
1.34 anton 5962: doc-variable
5963: doc-2variable
5964: doc-fvariable
5965:
1.29 crook 5966: @cindex user variables
5967: @cindex user space
5968: The defining word @code{User} behaves in the same way as @code{Variable}.
5969: The difference is that it reserves space in @i{user (data) space} rather
5970: than normal data space. In a Forth system that has a multi-tasker, each
5971: task has its own set of user variables.
5972:
1.34 anton 5973: doc-user
1.67 anton 5974: @c doc-udp
5975: @c doc-uallot
1.34 anton 5976:
1.29 crook 5977: @comment TODO is that stuff about user variables strictly correct? Is it
5978: @comment just terminal tasks that have user variables?
5979: @comment should document tasker.fs (with some examples) elsewhere
5980: @comment in this manual, then expand on user space and user variables.
5981:
1.44 crook 5982: @node Constants, Values, Variables, Defining Words
5983: @subsection Constants
5984: @cindex constants
5985:
5986: @code{Constant} allows you to declare a fixed value and refer to it by
5987: name. For example:
1.29 crook 5988:
5989: @example
5990: 12 Constant INCHES-PER-FOOT
5991: 3E+08 fconstant SPEED-O-LIGHT
5992: @end example
5993:
5994: A @code{Variable} can be both read and written, so its run-time
5995: behaviour is to supply an address through which its current value can be
5996: manipulated. In contrast, the value of a @code{Constant} cannot be
5997: changed once it has been declared@footnote{Well, often it can be -- but
5998: not in a Standard, portable way. It's safer to use a @code{Value} (read
5999: on).} so it's not necessary to supply the address -- it is more
6000: efficient to return the value of the constant directly. That's exactly
6001: what happens; the run-time effect of a constant is to put its value on
1.49 anton 6002: the top of the stack (You can find one
6003: way of implementing @code{Constant} in @ref{User-defined Defining Words}).
1.29 crook 6004:
1.69 anton 6005: Forth also provides @code{2Constant} and @code{fconstant} for defining
1.29 crook 6006: double and floating-point constants, respectively.
6007:
1.34 anton 6008: doc-constant
6009: doc-2constant
6010: doc-fconstant
6011:
6012: @c that's too deep, and it's not necessarily true for all ANS Forths. - anton
1.44 crook 6013: @c nac-> How could that not be true in an ANS Forth? You can't define a
6014: @c constant, use it and then delete the definition of the constant..
1.69 anton 6015:
6016: @c anton->An ANS Forth system can compile a constant to a literal; On
6017: @c decompilation you would see only the number, just as if it had been used
6018: @c in the first place. The word will stay, of course, but it will only be
6019: @c used by the text interpreter (no run-time duties, except when it is
6020: @c POSTPONEd or somesuch).
6021:
6022: @c nac:
1.44 crook 6023: @c I agree that it's rather deep, but IMO it is an important difference
6024: @c relative to other programming languages.. often it's annoying: it
6025: @c certainly changes my programming style relative to C.
6026:
1.69 anton 6027: @c anton: In what way?
6028:
1.29 crook 6029: Constants in Forth behave differently from their equivalents in other
6030: programming languages. In other languages, a constant (such as an EQU in
6031: assembler or a #define in C) only exists at compile-time; in the
6032: executable program the constant has been translated into an absolute
6033: number and, unless you are using a symbolic debugger, it's impossible to
6034: know what abstract thing that number represents. In Forth a constant has
1.44 crook 6035: an entry in the header space and remains there after the code that uses
6036: it has been defined. In fact, it must remain in the dictionary since it
6037: has run-time duties to perform. For example:
1.29 crook 6038:
6039: @example
6040: 12 Constant INCHES-PER-FOOT
6041: : FEET-TO-INCHES ( n1 -- n2 ) INCHES-PER-FOOT * ;
6042: @end example
6043:
6044: @cindex in-lining of constants
6045: When @code{FEET-TO-INCHES} is executed, it will in turn execute the xt
6046: associated with the constant @code{INCHES-PER-FOOT}. If you use
6047: @code{see} to decompile the definition of @code{FEET-TO-INCHES}, you can
6048: see that it makes a call to @code{INCHES-PER-FOOT}. Some Forth compilers
6049: attempt to optimise constants by in-lining them where they are used. You
6050: can force Gforth to in-line a constant like this:
6051:
6052: @example
6053: : FEET-TO-INCHES ( n1 -- n2 ) [ INCHES-PER-FOOT ] LITERAL * ;
6054: @end example
6055:
6056: If you use @code{see} to decompile @i{this} version of
6057: @code{FEET-TO-INCHES}, you can see that @code{INCHES-PER-FOOT} is no
1.49 anton 6058: longer present. To understand how this works, read
6059: @ref{Interpret/Compile states}, and @ref{Literals}.
1.29 crook 6060:
6061: In-lining constants in this way might improve execution time
6062: fractionally, and can ensure that a constant is now only referenced at
6063: compile-time. However, the definition of the constant still remains in
6064: the dictionary. Some Forth compilers provide a mechanism for controlling
6065: a second dictionary for holding transient words such that this second
6066: dictionary can be deleted later in order to recover memory
6067: space. However, there is no standard way of doing this.
6068:
6069:
1.44 crook 6070: @node Values, Colon Definitions, Constants, Defining Words
6071: @subsection Values
6072: @cindex values
1.34 anton 6073:
1.69 anton 6074: A @code{Value} behaves like a @code{Constant}, but it can be changed.
6075: @code{TO} is a parsing word that changes a @code{Values}. In Gforth
6076: (not in ANS Forth) you can access (and change) a @code{value} also with
6077: @code{>body}.
6078:
6079: Here are some
6080: examples:
1.29 crook 6081:
6082: @example
1.69 anton 6083: 12 Value APPLES \ Define APPLES with an initial value of 12
6084: 34 TO APPLES \ Change the value of APPLES. TO is a parsing word
6085: 1 ' APPLES >body +! \ Increment APPLES. Non-standard usage.
6086: APPLES \ puts 35 on the top of the stack.
1.29 crook 6087: @end example
6088:
1.44 crook 6089: doc-value
6090: doc-to
1.29 crook 6091:
1.35 anton 6092:
1.69 anton 6093:
1.44 crook 6094: @node Colon Definitions, Anonymous Definitions, Values, Defining Words
6095: @subsection Colon Definitions
6096: @cindex colon definitions
1.35 anton 6097:
6098: @example
1.44 crook 6099: : name ( ... -- ... )
6100: word1 word2 word3 ;
1.29 crook 6101: @end example
6102:
1.44 crook 6103: @noindent
6104: Creates a word called @code{name} that, upon execution, executes
6105: @code{word1 word2 word3}. @code{name} is a @dfn{(colon) definition}.
1.29 crook 6106:
1.49 anton 6107: The explanation above is somewhat superficial. For simple examples of
6108: colon definitions see @ref{Your first definition}. For an in-depth
1.66 anton 6109: discussion of some of the issues involved, @xref{Interpretation and
1.49 anton 6110: Compilation Semantics}.
1.29 crook 6111:
1.44 crook 6112: doc-:
6113: doc-;
1.1 anton 6114:
1.34 anton 6115:
1.69 anton 6116: @node Anonymous Definitions, Supplying names, Colon Definitions, Defining Words
1.44 crook 6117: @subsection Anonymous Definitions
6118: @cindex colon definitions
6119: @cindex defining words without name
1.34 anton 6120:
1.44 crook 6121: Sometimes you want to define an @dfn{anonymous word}; a word without a
6122: name. You can do this with:
1.1 anton 6123:
1.44 crook 6124: doc-:noname
1.1 anton 6125:
1.44 crook 6126: This leaves the execution token for the word on the stack after the
6127: closing @code{;}. Here's an example in which a deferred word is
6128: initialised with an @code{xt} from an anonymous colon definition:
1.1 anton 6129:
1.29 crook 6130: @example
1.44 crook 6131: Defer deferred
6132: :noname ( ... -- ... )
6133: ... ;
6134: IS deferred
1.29 crook 6135: @end example
1.26 crook 6136:
1.44 crook 6137: @noindent
6138: Gforth provides an alternative way of doing this, using two separate
6139: words:
1.27 crook 6140:
1.44 crook 6141: doc-noname
6142: @cindex execution token of last defined word
6143: doc-lastxt
1.1 anton 6144:
1.44 crook 6145: @noindent
6146: The previous example can be rewritten using @code{noname} and
6147: @code{lastxt}:
1.1 anton 6148:
1.26 crook 6149: @example
1.44 crook 6150: Defer deferred
6151: noname : ( ... -- ... )
6152: ... ;
6153: lastxt IS deferred
1.26 crook 6154: @end example
1.1 anton 6155:
1.29 crook 6156: @noindent
1.44 crook 6157: @code{noname} works with any defining word, not just @code{:}.
6158:
6159: @code{lastxt} also works when the last word was not defined as
1.71 anton 6160: @code{noname}. It does not work for combined words, though. It also has
6161: the useful property that is is valid as soon as the header for a
6162: definition has been built. Thus:
1.44 crook 6163:
6164: @example
6165: lastxt . : foo [ lastxt . ] ; ' foo .
6166: @end example
1.1 anton 6167:
1.44 crook 6168: @noindent
6169: prints 3 numbers; the last two are the same.
1.26 crook 6170:
1.69 anton 6171: @node Supplying names, User-defined Defining Words, Anonymous Definitions, Defining Words
6172: @subsection Supplying the name of a defined word
6173: @cindex names for defined words
6174: @cindex defining words, name given in a string
6175:
6176: By default, a defining word takes the name for the defined word from the
6177: input stream. Sometimes you want to supply the name from a string. You
6178: can do this with:
6179:
6180: doc-nextname
6181:
6182: For example:
6183:
6184: @example
6185: s" foo" nextname create
6186: @end example
6187:
6188: @noindent
6189: is equivalent to:
6190:
6191: @example
6192: create foo
6193: @end example
6194:
6195: @noindent
6196: @code{nextname} works with any defining word.
6197:
1.1 anton 6198:
1.69 anton 6199: @node User-defined Defining Words, Deferred words, Supplying names, Defining Words
1.26 crook 6200: @subsection User-defined Defining Words
6201: @cindex user-defined defining words
6202: @cindex defining words, user-defined
1.1 anton 6203:
1.29 crook 6204: You can create a new defining word by wrapping defining-time code around
6205: an existing defining word and putting the sequence in a colon
1.69 anton 6206: definition.
6207:
6208: @c anton: This example is very complex and leads in a quite different
6209: @c direction from the CREATE-DOES> stuff that follows. It should probably
6210: @c be done elsewhere, or as a subsubsection of this subsection (or as a
6211: @c subsection of Defining Words)
6212:
6213: For example, suppose that you have a word @code{stats} that
1.29 crook 6214: gathers statistics about colon definitions given the @i{xt} of the
6215: definition, and you want every colon definition in your application to
6216: make a call to @code{stats}. You can define and use a new version of
6217: @code{:} like this:
6218:
6219: @example
6220: : stats ( xt -- ) DUP ." (Gathering statistics for " . ." )"
6221: ... ; \ other code
6222:
6223: : my: : lastxt postpone literal ['] stats compile, ;
6224:
6225: my: foo + - ;
6226: @end example
6227:
6228: When @code{foo} is defined using @code{my:} these steps occur:
6229:
6230: @itemize @bullet
6231: @item
6232: @code{my:} is executed.
6233: @item
6234: The @code{:} within the definition (the one between @code{my:} and
6235: @code{lastxt}) is executed, and does just what it always does; it parses
6236: the input stream for a name, builds a dictionary header for the name
6237: @code{foo} and switches @code{state} from interpret to compile.
6238: @item
6239: The word @code{lastxt} is executed. It puts the @i{xt} for the word that is
6240: being defined -- @code{foo} -- onto the stack.
6241: @item
6242: The code that was produced by @code{postpone literal} is executed; this
6243: causes the value on the stack to be compiled as a literal in the code
6244: area of @code{foo}.
6245: @item
6246: The code @code{['] stats} compiles a literal into the definition of
6247: @code{my:}. When @code{compile,} is executed, that literal -- the
6248: execution token for @code{stats} -- is layed down in the code area of
6249: @code{foo} , following the literal@footnote{Strictly speaking, the
6250: mechanism that @code{compile,} uses to convert an @i{xt} into something
6251: in the code area is implementation-dependent. A threaded implementation
6252: might spit out the execution token directly whilst another
6253: implementation might spit out a native code sequence.}.
6254: @item
6255: At this point, the execution of @code{my:} is complete, and control
6256: returns to the text interpreter. The text interpreter is in compile
6257: state, so subsequent text @code{+ -} is compiled into the definition of
6258: @code{foo} and the @code{;} terminates the definition as always.
6259: @end itemize
6260:
6261: You can use @code{see} to decompile a word that was defined using
6262: @code{my:} and see how it is different from a normal @code{:}
6263: definition. For example:
6264:
6265: @example
6266: : bar + - ; \ like foo but using : rather than my:
6267: see bar
6268: : bar
6269: + - ;
6270: see foo
6271: : foo
6272: 107645672 stats + - ;
6273:
6274: \ use ' stats . to show that 107645672 is the xt for stats
6275: @end example
6276:
6277: You can use techniques like this to make new defining words in terms of
6278: @i{any} existing defining word.
1.1 anton 6279:
6280:
1.29 crook 6281: @cindex defining defining words
1.26 crook 6282: @cindex @code{CREATE} ... @code{DOES>}
6283: If you want the words defined with your defining words to behave
6284: differently from words defined with standard defining words, you can
6285: write your defining word like this:
1.1 anton 6286:
6287: @example
1.26 crook 6288: : def-word ( "name" -- )
1.29 crook 6289: CREATE @i{code1}
1.26 crook 6290: DOES> ( ... -- ... )
1.29 crook 6291: @i{code2} ;
1.26 crook 6292:
6293: def-word name
1.1 anton 6294: @end example
6295:
1.29 crook 6296: @cindex child words
6297: This fragment defines a @dfn{defining word} @code{def-word} and then
6298: executes it. When @code{def-word} executes, it @code{CREATE}s a new
6299: word, @code{name}, and executes the code @i{code1}. The code @i{code2}
6300: is not executed at this time. The word @code{name} is sometimes called a
6301: @dfn{child} of @code{def-word}.
6302:
6303: When you execute @code{name}, the address of the body of @code{name} is
6304: put on the data stack and @i{code2} is executed (the address of the body
6305: of @code{name} is the address @code{HERE} returns immediately after the
1.69 anton 6306: @code{CREATE}, i.e., the address a @code{create}d word returns by
6307: default).
6308:
6309: @c anton:
6310: @c www.dictionary.com says:
6311: @c at·a·vism: 1.The reappearance of a characteristic in an organism after
6312: @c several generations of absence, usually caused by the chance
6313: @c recombination of genes. 2.An individual or a part that exhibits
6314: @c atavism. Also called throwback. 3.The return of a trait or recurrence
6315: @c of previous behavior after a period of absence.
6316: @c
6317: @c Doesn't seem to fit.
1.29 crook 6318:
1.69 anton 6319: @c @cindex atavism in child words
1.33 anton 6320: You can use @code{def-word} to define a set of child words that behave
1.69 anton 6321: similarly; they all have a common run-time behaviour determined by
6322: @i{code2}. Typically, the @i{code1} sequence builds a data area in the
6323: body of the child word. The structure of the data is common to all
6324: children of @code{def-word}, but the data values are specific -- and
6325: private -- to each child word. When a child word is executed, the
6326: address of its private data area is passed as a parameter on TOS to be
6327: used and manipulated@footnote{It is legitimate both to read and write to
6328: this data area.} by @i{code2}.
1.29 crook 6329:
6330: The two fragments of code that make up the defining words act (are
6331: executed) at two completely separate times:
1.1 anton 6332:
1.29 crook 6333: @itemize @bullet
6334: @item
6335: At @i{define time}, the defining word executes @i{code1} to generate a
6336: child word
6337: @item
6338: At @i{child execution time}, when a child word is invoked, @i{code2}
6339: is executed, using parameters (data) that are private and specific to
6340: the child word.
6341: @end itemize
6342:
1.44 crook 6343: Another way of understanding the behaviour of @code{def-word} and
6344: @code{name} is to say that, if you make the following definitions:
1.33 anton 6345: @example
6346: : def-word1 ( "name" -- )
6347: CREATE @i{code1} ;
6348:
6349: : action1 ( ... -- ... )
6350: @i{code2} ;
6351:
6352: def-word1 name1
6353: @end example
6354:
1.44 crook 6355: @noindent
6356: Then using @code{name1 action1} is equivalent to using @code{name}.
1.1 anton 6357:
1.29 crook 6358: The classic example is that you can define @code{CONSTANT} in this way:
1.26 crook 6359:
1.1 anton 6360: @example
1.29 crook 6361: : CONSTANT ( w "name" -- )
6362: CREATE ,
1.26 crook 6363: DOES> ( -- w )
6364: @@ ;
1.1 anton 6365: @end example
6366:
1.29 crook 6367: @comment There is a beautiful description of how this works and what
6368: @comment it does in the Forthwrite 100th edition.. as well as an elegant
6369: @comment commentary on the Counting Fruits problem.
6370:
6371: When you create a constant with @code{5 CONSTANT five}, a set of
6372: define-time actions take place; first a new word @code{five} is created,
6373: then the value 5 is laid down in the body of @code{five} with
1.44 crook 6374: @code{,}. When @code{five} is executed, the address of the body is put on
1.29 crook 6375: the stack, and @code{@@} retrieves the value 5. The word @code{five} has
6376: no code of its own; it simply contains a data field and a pointer to the
6377: code that follows @code{DOES>} in its defining word. That makes words
6378: created in this way very compact.
6379:
6380: The final example in this section is intended to remind you that space
6381: reserved in @code{CREATE}d words is @i{data} space and therefore can be
6382: both read and written by a Standard program@footnote{Exercise: use this
6383: example as a starting point for your own implementation of @code{Value}
6384: and @code{TO} -- if you get stuck, investigate the behaviour of @code{'} and
6385: @code{[']}.}:
6386:
6387: @example
6388: : foo ( "name" -- )
6389: CREATE -1 ,
6390: DOES> ( -- )
1.33 anton 6391: @@ . ;
1.29 crook 6392:
6393: foo first-word
6394: foo second-word
6395:
6396: 123 ' first-word >BODY !
6397: @end example
6398:
6399: If @code{first-word} had been a @code{CREATE}d word, we could simply
6400: have executed it to get the address of its data field. However, since it
6401: was defined to have @code{DOES>} actions, its execution semantics are to
6402: perform those @code{DOES>} actions. To get the address of its data field
6403: it's necessary to use @code{'} to get its xt, then @code{>BODY} to
6404: translate the xt into the address of the data field. When you execute
6405: @code{first-word}, it will display @code{123}. When you execute
6406: @code{second-word} it will display @code{-1}.
1.26 crook 6407:
6408: @cindex stack effect of @code{DOES>}-parts
6409: @cindex @code{DOES>}-parts, stack effect
1.29 crook 6410: In the examples above the stack comment after the @code{DOES>} specifies
1.26 crook 6411: the stack effect of the defined words, not the stack effect of the
6412: following code (the following code expects the address of the body on
6413: the top of stack, which is not reflected in the stack comment). This is
6414: the convention that I use and recommend (it clashes a bit with using
6415: locals declarations for stack effect specification, though).
1.1 anton 6416:
1.53 anton 6417: @menu
6418: * CREATE..DOES> applications::
6419: * CREATE..DOES> details::
1.63 anton 6420: * Advanced does> usage example::
1.53 anton 6421: @end menu
6422:
6423: @node CREATE..DOES> applications, CREATE..DOES> details, User-defined Defining Words, User-defined Defining Words
1.26 crook 6424: @subsubsection Applications of @code{CREATE..DOES>}
6425: @cindex @code{CREATE} ... @code{DOES>}, applications
1.1 anton 6426:
1.26 crook 6427: You may wonder how to use this feature. Here are some usage patterns:
1.1 anton 6428:
1.26 crook 6429: @cindex factoring similar colon definitions
6430: When you see a sequence of code occurring several times, and you can
6431: identify a meaning, you will factor it out as a colon definition. When
6432: you see similar colon definitions, you can factor them using
6433: @code{CREATE..DOES>}. E.g., an assembler usually defines several words
6434: that look very similar:
1.1 anton 6435: @example
1.26 crook 6436: : ori, ( reg-target reg-source n -- )
6437: 0 asm-reg-reg-imm ;
6438: : andi, ( reg-target reg-source n -- )
6439: 1 asm-reg-reg-imm ;
1.1 anton 6440: @end example
6441:
1.26 crook 6442: @noindent
6443: This could be factored with:
6444: @example
6445: : reg-reg-imm ( op-code -- )
6446: CREATE ,
6447: DOES> ( reg-target reg-source n -- )
6448: @@ asm-reg-reg-imm ;
6449:
6450: 0 reg-reg-imm ori,
6451: 1 reg-reg-imm andi,
6452: @end example
1.1 anton 6453:
1.26 crook 6454: @cindex currying
6455: Another view of @code{CREATE..DOES>} is to consider it as a crude way to
6456: supply a part of the parameters for a word (known as @dfn{currying} in
6457: the functional language community). E.g., @code{+} needs two
6458: parameters. Creating versions of @code{+} with one parameter fixed can
6459: be done like this:
1.1 anton 6460: @example
1.26 crook 6461: : curry+ ( n1 -- )
6462: CREATE ,
6463: DOES> ( n2 -- n1+n2 )
6464: @@ + ;
6465:
6466: 3 curry+ 3+
6467: -2 curry+ 2-
1.1 anton 6468: @end example
6469:
1.63 anton 6470: @node CREATE..DOES> details, Advanced does> usage example, CREATE..DOES> applications, User-defined Defining Words
1.26 crook 6471: @subsubsection The gory details of @code{CREATE..DOES>}
6472: @cindex @code{CREATE} ... @code{DOES>}, details
1.1 anton 6473:
1.26 crook 6474: doc-does>
1.1 anton 6475:
1.26 crook 6476: @cindex @code{DOES>} in a separate definition
6477: This means that you need not use @code{CREATE} and @code{DOES>} in the
6478: same definition; you can put the @code{DOES>}-part in a separate
1.29 crook 6479: definition. This allows us to, e.g., select among different @code{DOES>}-parts:
1.26 crook 6480: @example
6481: : does1
6482: DOES> ( ... -- ... )
1.44 crook 6483: ... ;
6484:
6485: : does2
6486: DOES> ( ... -- ... )
6487: ... ;
6488:
6489: : def-word ( ... -- ... )
6490: create ...
6491: IF
6492: does1
6493: ELSE
6494: does2
6495: ENDIF ;
6496: @end example
6497:
6498: In this example, the selection of whether to use @code{does1} or
1.69 anton 6499: @code{does2} is made at definition-time; at the time that the child word is
1.44 crook 6500: @code{CREATE}d.
6501:
6502: @cindex @code{DOES>} in interpretation state
6503: In a standard program you can apply a @code{DOES>}-part only if the last
6504: word was defined with @code{CREATE}. In Gforth, the @code{DOES>}-part
6505: will override the behaviour of the last word defined in any case. In a
6506: standard program, you can use @code{DOES>} only in a colon
6507: definition. In Gforth, you can also use it in interpretation state, in a
6508: kind of one-shot mode; for example:
6509: @example
6510: CREATE name ( ... -- ... )
6511: @i{initialization}
6512: DOES>
6513: @i{code} ;
6514: @end example
6515:
6516: @noindent
6517: is equivalent to the standard:
6518: @example
6519: :noname
6520: DOES>
6521: @i{code} ;
6522: CREATE name EXECUTE ( ... -- ... )
6523: @i{initialization}
6524: @end example
6525:
1.53 anton 6526: doc->body
6527:
1.63 anton 6528: @node Advanced does> usage example, , CREATE..DOES> details, User-defined Defining Words
6529: @subsubsection Advanced does> usage example
6530:
6531: The MIPS disassembler (@file{arch/mips/disasm.fs}) contains many words
6532: for disassembling instructions, that follow a very repetetive scheme:
6533:
6534: @example
6535: :noname @var{disasm-operands} s" @var{inst-name}" type ;
6536: @var{entry-num} cells @var{table} + !
6537: @end example
6538:
6539: Of course, this inspires the idea to factor out the commonalities to
6540: allow a definition like
6541:
6542: @example
6543: @var{disasm-operands} @var{entry-num} @var{table} define-inst @var{inst-name}
6544: @end example
6545:
6546: The parameters @var{disasm-operands} and @var{table} are usually
1.69 anton 6547: correlated. Moreover, before I wrote the disassembler, there already
6548: existed code that defines instructions like this:
1.63 anton 6549:
6550: @example
6551: @var{entry-num} @var{inst-format} @var{inst-name}
6552: @end example
6553:
6554: This code comes from the assembler and resides in
6555: @file{arch/mips/insts.fs}.
6556:
6557: So I had to define the @var{inst-format} words that performed the scheme
6558: above when executed. At first I chose to use run-time code-generation:
6559:
6560: @example
6561: : @var{inst-format} ( entry-num "name" -- ; compiled code: addr w -- )
6562: :noname Postpone @var{disasm-operands}
6563: name Postpone sliteral Postpone type Postpone ;
6564: swap cells @var{table} + ! ;
6565: @end example
6566:
6567: Note that this supplies the other two parameters of the scheme above.
1.44 crook 6568:
1.63 anton 6569: An alternative would have been to write this using
6570: @code{create}/@code{does>}:
6571:
6572: @example
6573: : @var{inst-format} ( entry-num "name" -- )
6574: here name string, ( entry-num c-addr ) \ parse and save "name"
6575: noname create , ( entry-num )
6576: lastxt swap cells @var{table} + !
6577: does> ( addr w -- )
6578: \ disassemble instruction w at addr
6579: @@ >r
6580: @var{disasm-operands}
6581: r> count type ;
6582: @end example
6583:
6584: Somehow the first solution is simpler, mainly because it's simpler to
6585: shift a string from definition-time to use-time with @code{sliteral}
6586: than with @code{string,} and friends.
6587:
6588: I wrote a lot of words following this scheme and soon thought about
6589: factoring out the commonalities among them. Note that this uses a
6590: two-level defining word, i.e., a word that defines ordinary defining
6591: words.
6592:
6593: This time a solution involving @code{postpone} and friends seemed more
6594: difficult (try it as an exercise), so I decided to use a
6595: @code{create}/@code{does>} word; since I was already at it, I also used
6596: @code{create}/@code{does>} for the lower level (try using
6597: @code{postpone} etc. as an exercise), resulting in the following
6598: definition:
6599:
6600: @example
6601: : define-format ( disasm-xt table-xt -- )
6602: \ define an instruction format that uses disasm-xt for
6603: \ disassembling and enters the defined instructions into table
6604: \ table-xt
6605: create 2,
6606: does> ( u "inst" -- )
6607: \ defines an anonymous word for disassembling instruction inst,
6608: \ and enters it as u-th entry into table-xt
6609: 2@@ swap here name string, ( u table-xt disasm-xt c-addr ) \ remember string
6610: noname create 2, \ define anonymous word
6611: execute lastxt swap ! \ enter xt of defined word into table-xt
6612: does> ( addr w -- )
6613: \ disassemble instruction w at addr
6614: 2@@ >r ( addr w disasm-xt R: c-addr )
6615: execute ( R: c-addr ) \ disassemble operands
6616: r> count type ; \ print name
6617: @end example
6618:
6619: Note that the tables here (in contrast to above) do the @code{cells +}
6620: by themselves (that's why you have to pass an xt). This word is used in
6621: the following way:
6622:
6623: @example
6624: ' @var{disasm-operands} ' @var{table} define-format @var{inst-format}
6625: @end example
6626:
1.71 anton 6627: As shown above, the defined instruction format is then used like this:
6628:
6629: @example
6630: @var{entry-num} @var{inst-format} @var{inst-name}
6631: @end example
6632:
1.63 anton 6633: In terms of currying, this kind of two-level defining word provides the
6634: parameters in three stages: first @var{disasm-operands} and @var{table},
6635: then @var{entry-num} and @var{inst-name}, finally @code{addr w}, i.e.,
6636: the instruction to be disassembled.
6637:
6638: Of course this did not quite fit all the instruction format names used
6639: in @file{insts.fs}, so I had to define a few wrappers that conditioned
6640: the parameters into the right form.
6641:
6642: If you have trouble following this section, don't worry. First, this is
6643: involved and takes time (and probably some playing around) to
6644: understand; second, this is the first two-level
6645: @code{create}/@code{does>} word I have written in seventeen years of
6646: Forth; and if I did not have @file{insts.fs} to start with, I may well
6647: have elected to use just a one-level defining word (with some repeating
6648: of parameters when using the defining word). So it is not necessary to
6649: understand this, but it may improve your understanding of Forth.
1.44 crook 6650:
6651:
6652: @node Deferred words, Aliases, User-defined Defining Words, Defining Words
6653: @subsection Deferred words
6654: @cindex deferred words
6655:
6656: The defining word @code{Defer} allows you to define a word by name
6657: without defining its behaviour; the definition of its behaviour is
6658: deferred. Here are two situation where this can be useful:
6659:
6660: @itemize @bullet
6661: @item
6662: Where you want to allow the behaviour of a word to be altered later, and
6663: for all precompiled references to the word to change when its behaviour
6664: is changed.
6665: @item
6666: For mutual recursion; @xref{Calls and returns}.
6667: @end itemize
6668:
6669: In the following example, @code{foo} always invokes the version of
6670: @code{greet} that prints ``@code{Good morning}'' whilst @code{bar}
6671: always invokes the version that prints ``@code{Hello}''. There is no way
6672: of getting @code{foo} to use the later version without re-ordering the
6673: source code and recompiling it.
6674:
6675: @example
6676: : greet ." Good morning" ;
6677: : foo ... greet ... ;
6678: : greet ." Hello" ;
6679: : bar ... greet ... ;
6680: @end example
6681:
6682: This problem can be solved by defining @code{greet} as a @code{Defer}red
6683: word. The behaviour of a @code{Defer}red word can be defined and
6684: redefined at any time by using @code{IS} to associate the xt of a
6685: previously-defined word with it. The previous example becomes:
6686:
6687: @example
1.69 anton 6688: Defer greet ( -- )
1.44 crook 6689: : foo ... greet ... ;
6690: : bar ... greet ... ;
1.69 anton 6691: : greet1 ( -- ) ." Good morning" ;
6692: : greet2 ( -- ) ." Hello" ;
1.44 crook 6693: ' greet2 <IS> greet \ make greet behave like greet2
6694: @end example
6695:
1.69 anton 6696: @progstyle
6697: You should write a stack comment for every deferred word, and put only
6698: XTs into deferred words that conform to this stack effect. Otherwise
6699: it's too difficult to use the deferred word.
6700:
1.44 crook 6701: A deferred word can be used to improve the statistics-gathering example
6702: from @ref{User-defined Defining Words}; rather than edit the
6703: application's source code to change every @code{:} to a @code{my:}, do
6704: this:
6705:
6706: @example
6707: : real: : ; \ retain access to the original
6708: defer : \ redefine as a deferred word
1.69 anton 6709: ' my: <IS> : \ use special version of :
1.44 crook 6710: \
6711: \ load application here
6712: \
1.69 anton 6713: ' real: <IS> : \ go back to the original
1.44 crook 6714: @end example
6715:
6716:
6717: One thing to note is that @code{<IS>} consumes its name when it is
6718: executed. If you want to specify the name at compile time, use
6719: @code{[IS]}:
6720:
6721: @example
6722: : set-greet ( xt -- )
6723: [IS] greet ;
6724:
6725: ' greet1 set-greet
6726: @end example
6727:
1.69 anton 6728: A deferred word can only inherit execution semantics from the xt
6729: (because that is all that an xt can represent -- for more discussion of
6730: this @pxref{Tokens for Words}); by default it will have default
6731: interpretation and compilation semantics deriving from this execution
6732: semantics. However, you can change the interpretation and compilation
6733: semantics of the deferred word in the usual ways:
1.44 crook 6734:
6735: @example
6736: : bar .... ; compile-only
6737: Defer fred immediate
6738: Defer jim
6739:
6740: ' bar <IS> jim \ jim has default semantics
6741: ' bar <IS> fred \ fred is immediate
6742: @end example
6743:
6744: doc-defer
6745: doc-<is>
6746: doc-[is]
6747: doc-is
6748: @comment TODO document these: what's defers [is]
6749: doc-what's
6750: doc-defers
6751:
6752: @c Use @code{words-deferred} to see a list of deferred words.
6753:
6754: Definitions in ANS Forth for @code{defer}, @code{<is>} and @code{[is]}
6755: are provided in @file{compat/defer.fs}.
6756:
6757:
1.69 anton 6758: @node Aliases, , Deferred words, Defining Words
1.44 crook 6759: @subsection Aliases
6760: @cindex aliases
1.1 anton 6761:
1.44 crook 6762: The defining word @code{Alias} allows you to define a word by name that
6763: has the same behaviour as some other word. Here are two situation where
6764: this can be useful:
1.1 anton 6765:
1.44 crook 6766: @itemize @bullet
6767: @item
6768: When you want access to a word's definition from a different word list
6769: (for an example of this, see the definition of the @code{Root} word list
6770: in the Gforth source).
6771: @item
6772: When you want to create a synonym; a definition that can be known by
6773: either of two names (for example, @code{THEN} and @code{ENDIF} are
6774: aliases).
6775: @end itemize
1.1 anton 6776:
1.69 anton 6777: Like deferred words, an alias has default compilation and interpretation
6778: semantics at the beginning (not the modifications of the other word),
6779: but you can change them in the usual ways (@code{immediate},
6780: @code{compile-only}). For example:
1.1 anton 6781:
6782: @example
1.44 crook 6783: : foo ... ; immediate
6784:
6785: ' foo Alias bar \ bar is not an immediate word
6786: ' foo Alias fooby immediate \ fooby is an immediate word
1.1 anton 6787: @end example
6788:
1.44 crook 6789: Words that are aliases have the same xt, different headers in the
6790: dictionary, and consequently different name tokens (@pxref{Tokens for
6791: Words}) and possibly different immediate flags. An alias can only have
6792: default or immediate compilation semantics; you can define aliases for
6793: combined words with @code{interpret/compile:} -- see @ref{Combined words}.
1.1 anton 6794:
1.44 crook 6795: doc-alias
1.1 anton 6796:
6797:
1.47 crook 6798: @node Interpretation and Compilation Semantics, Tokens for Words, Defining Words, Words
6799: @section Interpretation and Compilation Semantics
1.26 crook 6800: @cindex semantics, interpretation and compilation
1.1 anton 6801:
1.71 anton 6802: @c !! state and ' are used without explanation
6803: @c example for immediate/compile-only? or is the tutorial enough
6804:
1.26 crook 6805: @cindex interpretation semantics
1.71 anton 6806: The @dfn{interpretation semantics} of a (named) word are what the text
1.26 crook 6807: interpreter does when it encounters the word in interpret state. It also
6808: appears in some other contexts, e.g., the execution token returned by
1.71 anton 6809: @code{' @i{word}} identifies the interpretation semantics of @i{word}
6810: (in other words, @code{' @i{word} execute} is equivalent to
1.29 crook 6811: interpret-state text interpretation of @code{@i{word}}).
1.1 anton 6812:
1.26 crook 6813: @cindex compilation semantics
1.71 anton 6814: The @dfn{compilation semantics} of a (named) word are what the text
6815: interpreter does when it encounters the word in compile state. It also
6816: appears in other contexts, e.g, @code{POSTPONE @i{word}}
6817: compiles@footnote{In standard terminology, ``appends to the current
6818: definition''.} the compilation semantics of @i{word}.
1.1 anton 6819:
1.26 crook 6820: @cindex execution semantics
6821: The standard also talks about @dfn{execution semantics}. They are used
6822: only for defining the interpretation and compilation semantics of many
6823: words. By default, the interpretation semantics of a word are to
6824: @code{execute} its execution semantics, and the compilation semantics of
6825: a word are to @code{compile,} its execution semantics.@footnote{In
6826: standard terminology: The default interpretation semantics are its
6827: execution semantics; the default compilation semantics are to append its
6828: execution semantics to the execution semantics of the current
6829: definition.}
6830:
1.71 anton 6831: Unnamed words (@pxref{Anonymous Definitions}) cannot be encountered by
6832: the text interpreter, ticked, or @code{postpone}d, so they have no
6833: interpretation or compilation semantics. Their behaviour is represented
6834: by their XT (@pxref{Tokens for Words}), and we call it execution
6835: semantics, too.
6836:
1.26 crook 6837: @comment TODO expand, make it co-operate with new sections on text interpreter.
6838:
6839: @cindex immediate words
6840: @cindex compile-only words
6841: You can change the semantics of the most-recently defined word:
6842:
1.44 crook 6843:
1.26 crook 6844: doc-immediate
6845: doc-compile-only
6846: doc-restrict
6847:
1.44 crook 6848:
1.26 crook 6849: Note that ticking (@code{'}) a compile-only word gives an error
6850: (``Interpreting a compile-only word'').
1.1 anton 6851:
1.47 crook 6852: @menu
1.67 anton 6853: * Combined words::
1.47 crook 6854: @end menu
1.44 crook 6855:
1.71 anton 6856:
1.48 anton 6857: @node Combined words, , Interpretation and Compilation Semantics, Interpretation and Compilation Semantics
1.44 crook 6858: @subsection Combined Words
6859: @cindex combined words
6860:
6861: Gforth allows you to define @dfn{combined words} -- words that have an
6862: arbitrary combination of interpretation and compilation semantics.
6863:
1.26 crook 6864: doc-interpret/compile:
1.1 anton 6865:
1.26 crook 6866: This feature was introduced for implementing @code{TO} and @code{S"}. I
6867: recommend that you do not define such words, as cute as they may be:
6868: they make it hard to get at both parts of the word in some contexts.
6869: E.g., assume you want to get an execution token for the compilation
6870: part. Instead, define two words, one that embodies the interpretation
6871: part, and one that embodies the compilation part. Once you have done
6872: that, you can define a combined word with @code{interpret/compile:} for
6873: the convenience of your users.
1.1 anton 6874:
1.26 crook 6875: You might try to use this feature to provide an optimizing
6876: implementation of the default compilation semantics of a word. For
6877: example, by defining:
1.1 anton 6878: @example
1.26 crook 6879: :noname
6880: foo bar ;
6881: :noname
6882: POSTPONE foo POSTPONE bar ;
1.29 crook 6883: interpret/compile: opti-foobar
1.1 anton 6884: @end example
1.26 crook 6885:
1.23 crook 6886: @noindent
1.26 crook 6887: as an optimizing version of:
6888:
1.1 anton 6889: @example
1.26 crook 6890: : foobar
6891: foo bar ;
1.1 anton 6892: @end example
6893:
1.26 crook 6894: Unfortunately, this does not work correctly with @code{[compile]},
6895: because @code{[compile]} assumes that the compilation semantics of all
6896: @code{interpret/compile:} words are non-default. I.e., @code{[compile]
1.29 crook 6897: opti-foobar} would compile compilation semantics, whereas
6898: @code{[compile] foobar} would compile interpretation semantics.
1.1 anton 6899:
1.26 crook 6900: @cindex state-smart words (are a bad idea)
1.29 crook 6901: Some people try to use @dfn{state-smart} words to emulate the feature provided
1.26 crook 6902: by @code{interpret/compile:} (words are state-smart if they check
6903: @code{STATE} during execution). E.g., they would try to code
6904: @code{foobar} like this:
1.1 anton 6905:
1.26 crook 6906: @example
6907: : foobar
6908: STATE @@
6909: IF ( compilation state )
6910: POSTPONE foo POSTPONE bar
6911: ELSE
6912: foo bar
6913: ENDIF ; immediate
6914: @end example
1.1 anton 6915:
1.26 crook 6916: Although this works if @code{foobar} is only processed by the text
6917: interpreter, it does not work in other contexts (like @code{'} or
6918: @code{POSTPONE}). E.g., @code{' foobar} will produce an execution token
6919: for a state-smart word, not for the interpretation semantics of the
6920: original @code{foobar}; when you execute this execution token (directly
6921: with @code{EXECUTE} or indirectly through @code{COMPILE,}) in compile
6922: state, the result will not be what you expected (i.e., it will not
6923: perform @code{foo bar}). State-smart words are a bad idea. Simply don't
6924: write them@footnote{For a more detailed discussion of this topic, see
1.66 anton 6925: M. Anton Ertl,
6926: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl98.ps.gz,@code{State}-smartness---Why
6927: it is Evil and How to Exorcise it}}, EuroForth '98.}!
1.1 anton 6928:
1.26 crook 6929: @cindex defining words with arbitrary semantics combinations
6930: It is also possible to write defining words that define words with
6931: arbitrary combinations of interpretation and compilation semantics. In
6932: general, they look like this:
1.1 anton 6933:
1.26 crook 6934: @example
6935: : def-word
6936: create-interpret/compile
1.29 crook 6937: @i{code1}
1.26 crook 6938: interpretation>
1.29 crook 6939: @i{code2}
1.26 crook 6940: <interpretation
6941: compilation>
1.29 crook 6942: @i{code3}
1.26 crook 6943: <compilation ;
6944: @end example
1.1 anton 6945:
1.29 crook 6946: For a @i{word} defined with @code{def-word}, the interpretation
6947: semantics are to push the address of the body of @i{word} and perform
6948: @i{code2}, and the compilation semantics are to push the address of
6949: the body of @i{word} and perform @i{code3}. E.g., @code{constant}
1.26 crook 6950: can also be defined like this (except that the defined constants don't
6951: behave correctly when @code{[compile]}d):
1.1 anton 6952:
1.26 crook 6953: @example
6954: : constant ( n "name" -- )
6955: create-interpret/compile
6956: ,
6957: interpretation> ( -- n )
6958: @@
6959: <interpretation
6960: compilation> ( compilation. -- ; run-time. -- n )
6961: @@ postpone literal
6962: <compilation ;
6963: @end example
1.1 anton 6964:
1.44 crook 6965:
1.26 crook 6966: doc-create-interpret/compile
6967: doc-interpretation>
6968: doc-<interpretation
6969: doc-compilation>
6970: doc-<compilation
1.1 anton 6971:
1.44 crook 6972:
1.29 crook 6973: Words defined with @code{interpret/compile:} and
1.26 crook 6974: @code{create-interpret/compile} have an extended header structure that
6975: differs from other words; however, unless you try to access them with
6976: plain address arithmetic, you should not notice this. Words for
6977: accessing the header structure usually know how to deal with this; e.g.,
1.29 crook 6978: @code{'} @i{word} @code{>body} also gives you the body of a word created
6979: with @code{create-interpret/compile}.
1.1 anton 6980:
1.44 crook 6981:
1.27 crook 6982: doc-postpone
1.44 crook 6983:
1.29 crook 6984: @comment TODO -- expand glossary text for POSTPONE
1.27 crook 6985:
1.47 crook 6986:
6987: @c -------------------------------------------------------------
6988: @node Tokens for Words, The Text Interpreter, Interpretation and Compilation Semantics, Words
6989: @section Tokens for Words
6990: @cindex tokens for words
6991:
6992: This section describes the creation and use of tokens that represent
6993: words.
6994:
1.71 anton 6995: @menu
6996: * Execution token:: represents execution/interpretation semantics
6997: * Compilation token:: represents compilation semantics
6998: * Name token:: represents named words
6999: @end menu
1.47 crook 7000:
1.71 anton 7001: @node Execution token, Compilation token, Tokens for Words, Tokens for Words
7002: @subsection Execution token
1.47 crook 7003:
7004: @cindex xt
7005: @cindex execution token
1.71 anton 7006: An @dfn{execution token} (@i{XT}) represents some behaviour of a word.
7007: You can use @code{execute} to invoke this behaviour.
1.47 crook 7008:
1.71 anton 7009: @cindex tick (')
7010: You can use @code{'} to get an execution token that represents the
7011: interpretation semantics of a named word:
1.47 crook 7012:
7013: @example
1.71 anton 7014: 5 ' .
7015: execute
7016: @end example
1.47 crook 7017:
1.71 anton 7018: doc-'
7019:
7020: @code{'} parses at run-time; there is also a word @code{[']} that parses
7021: when it is compiled, and compiles the resulting XT:
7022:
7023: @example
7024: : foo ['] . execute ;
7025: 5 foo
7026: : bar ' execute ; \ by contrast,
7027: 5 bar . \ ' parses "." when bar executes
7028: @end example
7029:
7030: doc-[']
7031:
7032: If you want the execution token of @i{word}, write @code{['] @i{word}}
7033: in compiled code and @code{' @i{word}} in interpreted code. Gforth's
7034: @code{'} and @code{[']} behave somewhat unusually by complaining about
7035: compile-only words (because these words have no interpretation
7036: semantics). You might get what you want by using @code{COMP' @i{word}
7037: DROP} or @code{[COMP'] @i{word} DROP} (for details @pxref{Compilation
7038: token}).
7039:
7040: Another way to get an XT is @code{:noname} or @code{lastxt}
7041: (@pxref{Anonymous Definitions}). For anonymous words this gives an xt
7042: for the only behaviour the word has (the execution semantics). For
7043: named words, @code{lastxt} produces an XT for the same behaviour it
7044: would produce if the word was defined anonymously.
7045:
7046: @example
7047: :noname ." hello" ;
7048: execute
1.47 crook 7049: @end example
7050:
1.71 anton 7051: An XT occupies one cell and can be manipulated like any other cell.
7052:
1.47 crook 7053: @cindex code field address
7054: @cindex CFA
1.71 anton 7055: In ANS Forth the XT is just an abstract data type (i.e., defined by the
7056: operations that produce or consume it). For old hands: In Gforth, the
7057: XT is implemented as a code field address (CFA).
7058:
7059: @c !! discuss "compile," some more (or in Macros).
7060:
7061: doc-execute
7062: doc-perform
7063: doc-compile,
7064:
7065: @node Compilation token, Name token, Execution token, Tokens for Words
7066: @subsection Compilation token
1.47 crook 7067:
7068: @cindex compilation token
1.71 anton 7069: @cindex CT (compilation token)
7070: Gforth represents the compilation semantics of a named word by a
1.47 crook 7071: @dfn{compilation token} consisting of two cells: @i{w xt}. The top cell
7072: @i{xt} is an execution token. The compilation semantics represented by
7073: the compilation token can be performed with @code{execute}, which
7074: consumes the whole compilation token, with an additional stack effect
7075: determined by the represented compilation semantics.
7076:
7077: At present, the @i{w} part of a compilation token is an execution token,
7078: and the @i{xt} part represents either @code{execute} or
7079: @code{compile,}@footnote{Depending upon the compilation semantics of the
7080: word. If the word has default compilation semantics, the @i{xt} will
7081: represent @code{compile,}. Otherwise (e.g., for immediate words), the
7082: @i{xt} will represent @code{execute}.}. However, don't rely on that
7083: knowledge, unless necessary; future versions of Gforth may introduce
7084: unusual compilation tokens (e.g., a compilation token that represents
7085: the compilation semantics of a literal).
7086:
1.71 anton 7087: You can perform the compilation semantics represented by the compilation
7088: token with @code{execute}. You can compile the compilation semantics
7089: with @code{postpone,}. I.e., @code{COMP' @i{word} postpone,} is
7090: equivalent to @code{postpone @i{word}}.
7091:
7092: doc-[comp']
7093: doc-comp'
7094: doc-postpone,
7095:
7096: @node Name token, , Compilation token, Tokens for Words
7097: @subsection Name token
1.47 crook 7098:
7099: @cindex name token
7100: @cindex name field address
7101: @cindex NFA
1.71 anton 7102: Gforth represents named words by the @dfn{name token}, (@i{nt}). In
1.47 crook 7103: Gforth, the abstract data type @emph{name token} is implemented as a
7104: name field address (NFA).
7105:
7106: doc-find-name
7107: doc-name>int
7108: doc-name?int
7109: doc-name>comp
7110: doc-name>string
7111:
7112:
1.26 crook 7113: @c ----------------------------------------------------------
1.47 crook 7114: @node The Text Interpreter, Word Lists, Tokens for Words, Words
1.26 crook 7115: @section The Text Interpreter
7116: @cindex interpreter - outer
7117: @cindex text interpreter
7118: @cindex outer interpreter
1.1 anton 7119:
1.34 anton 7120: @c Should we really describe all these ugly details? IMO the text
7121: @c interpreter should be much cleaner, but that may not be possible within
7122: @c ANS Forth. - anton
1.44 crook 7123: @c nac-> I wanted to explain how it works to show how you can exploit
7124: @c it in your own programs. When I was writing a cross-compiler, figuring out
7125: @c some of these gory details was very helpful to me. None of the textbooks
7126: @c I've seen cover it, and the most modern Forth textbook -- Forth Inc's,
7127: @c seems to positively avoid going into too much detail for some of
7128: @c the internals.
1.34 anton 7129:
1.71 anton 7130: @c anton: ok. I wonder, though, if this is the right place; for some stuff
7131: @c it is; for the ugly details, I would prefer another place. I wonder
7132: @c whether we should have a chapter before "Words" that describes some
7133: @c basic concepts referred to in words, and a chapter after "Words" that
7134: @c describes implementation details.
7135:
1.29 crook 7136: The text interpreter@footnote{This is an expanded version of the
7137: material in @ref{Introducing the Text Interpreter}.} is an endless loop
1.34 anton 7138: that processes input from the current input device. It is also called
7139: the outer interpreter, in contrast to the inner interpreter
7140: (@pxref{Engine}) which executes the compiled Forth code on interpretive
7141: implementations.
1.27 crook 7142:
1.29 crook 7143: @cindex interpret state
7144: @cindex compile state
7145: The text interpreter operates in one of two states: @dfn{interpret
7146: state} and @dfn{compile state}. The current state is defined by the
1.71 anton 7147: aptly-named variable @code{state}.
1.29 crook 7148:
7149: This section starts by describing how the text interpreter behaves when
7150: it is in interpret state, processing input from the user input device --
7151: the keyboard. This is the mode that a Forth system is in after it starts
7152: up.
7153:
7154: @cindex input buffer
7155: @cindex terminal input buffer
7156: The text interpreter works from an area of memory called the @dfn{input
7157: buffer}@footnote{When the text interpreter is processing input from the
7158: keyboard, this area of memory is called the @dfn{terminal input buffer}
7159: (TIB) and is addressed by the (obsolescent) words @code{TIB} and
7160: @code{#TIB}.}, which stores your keyboard input when you press the
1.30 anton 7161: @key{RET} key. Starting at the beginning of the input buffer, it skips
1.29 crook 7162: leading spaces (called @dfn{delimiters}) then parses a string (a
7163: sequence of non-space characters) until it reaches either a space
7164: character or the end of the buffer. Having parsed a string, it makes two
7165: attempts to process it:
1.27 crook 7166:
1.29 crook 7167: @cindex dictionary
1.27 crook 7168: @itemize @bullet
7169: @item
1.29 crook 7170: It looks for the string in a @dfn{dictionary} of definitions. If the
7171: string is found, the string names a @dfn{definition} (also known as a
7172: @dfn{word}) and the dictionary search returns information that allows
7173: the text interpreter to perform the word's @dfn{interpretation
7174: semantics}. In most cases, this simply means that the word will be
7175: executed.
1.27 crook 7176: @item
7177: If the string is not found in the dictionary, the text interpreter
1.29 crook 7178: attempts to treat it as a number, using the rules described in
7179: @ref{Number Conversion}. If the string represents a legal number in the
7180: current radix, the number is pushed onto a parameter stack (the data
7181: stack for integers, the floating-point stack for floating-point
7182: numbers).
7183: @end itemize
7184:
7185: If both attempts fail, or if the word is found in the dictionary but has
7186: no interpretation semantics@footnote{This happens if the word was
7187: defined as @code{COMPILE-ONLY}.} the text interpreter discards the
7188: remainder of the input buffer, issues an error message and waits for
7189: more input. If one of the attempts succeeds, the text interpreter
7190: repeats the parsing process until the whole of the input buffer has been
7191: processed, at which point it prints the status message ``@code{ ok}''
7192: and waits for more input.
7193:
1.71 anton 7194: @c anton: this should be in the input stream subsection (or below it)
7195:
1.29 crook 7196: @cindex parse area
7197: The text interpreter keeps track of its position in the input buffer by
7198: updating a variable called @code{>IN} (pronounced ``to-in''). The value
7199: of @code{>IN} starts out as 0, indicating an offset of 0 from the start
7200: of the input buffer. The region from offset @code{>IN @@} to the end of
7201: the input buffer is called the @dfn{parse area}@footnote{In other words,
7202: the text interpreter processes the contents of the input buffer by
7203: parsing strings from the parse area until the parse area is empty.}.
7204: This example shows how @code{>IN} changes as the text interpreter parses
7205: the input buffer:
7206:
7207: @example
7208: : remaining >IN @@ SOURCE 2 PICK - -ROT + SWAP
7209: CR ." ->" TYPE ." <-" ; IMMEDIATE
7210:
7211: 1 2 3 remaining + remaining .
7212:
7213: : foo 1 2 3 remaining SWAP remaining ;
7214: @end example
7215:
7216: @noindent
7217: The result is:
7218:
7219: @example
7220: ->+ remaining .<-
7221: ->.<-5 ok
7222:
7223: ->SWAP remaining ;-<
7224: ->;<- ok
7225: @end example
7226:
7227: @cindex parsing words
7228: The value of @code{>IN} can also be modified by a word in the input
7229: buffer that is executed by the text interpreter. This means that a word
7230: can ``trick'' the text interpreter into either skipping a section of the
7231: input buffer@footnote{This is how parsing words work.} or into parsing a
7232: section twice. For example:
1.27 crook 7233:
1.29 crook 7234: @example
1.71 anton 7235: : lat ." <<foo>>" ;
7236: : flat ." <<bar>>" >IN DUP @@ 3 - SWAP ! ;
1.29 crook 7237: @end example
7238:
7239: @noindent
7240: When @code{flat} is executed, this output is produced@footnote{Exercise
7241: for the reader: what would happen if the @code{3} were replaced with
7242: @code{4}?}:
7243:
7244: @example
1.71 anton 7245: <<bar>><<foo>>
1.29 crook 7246: @end example
7247:
1.71 anton 7248: This technique can be used to work around some of the interoperability
7249: problems of parsing words. Of course, it's better to avoid parsing
7250: words where possible.
7251:
1.29 crook 7252: @noindent
7253: Two important notes about the behaviour of the text interpreter:
1.27 crook 7254:
7255: @itemize @bullet
7256: @item
7257: It processes each input string to completion before parsing additional
1.29 crook 7258: characters from the input buffer.
7259: @item
7260: It treats the input buffer as a read-only region (and so must your code).
7261: @end itemize
7262:
7263: @noindent
7264: When the text interpreter is in compile state, its behaviour changes in
7265: these ways:
7266:
7267: @itemize @bullet
7268: @item
7269: If a parsed string is found in the dictionary, the text interpreter will
7270: perform the word's @dfn{compilation semantics}. In most cases, this
7271: simply means that the execution semantics of the word will be appended
7272: to the current definition.
1.27 crook 7273: @item
1.29 crook 7274: When a number is encountered, it is compiled into the current definition
7275: (as a literal) rather than being pushed onto a parameter stack.
7276: @item
7277: If an error occurs, @code{state} is modified to put the text interpreter
7278: back into interpret state.
7279: @item
7280: Each time a line is entered from the keyboard, Gforth prints
7281: ``@code{ compiled}'' rather than `` @code{ok}''.
7282: @end itemize
7283:
7284: @cindex text interpreter - input sources
7285: When the text interpreter is using an input device other than the
7286: keyboard, its behaviour changes in these ways:
7287:
7288: @itemize @bullet
7289: @item
7290: When the parse area is empty, the text interpreter attempts to refill
7291: the input buffer from the input source. When the input source is
1.71 anton 7292: exhausted, the input source is set back to the previous input source.
1.29 crook 7293: @item
7294: It doesn't print out ``@code{ ok}'' or ``@code{ compiled}'' messages each
7295: time the parse area is emptied.
7296: @item
7297: If an error occurs, the input source is set back to the user input
7298: device.
1.27 crook 7299: @end itemize
1.21 crook 7300:
1.49 anton 7301: You can read about this in more detail in @ref{Input Sources}.
1.44 crook 7302:
1.26 crook 7303: doc->in
1.27 crook 7304: doc-source
7305:
1.26 crook 7306: doc-tib
7307: doc-#tib
1.1 anton 7308:
1.44 crook 7309:
1.26 crook 7310: @menu
1.67 anton 7311: * Input Sources::
7312: * Number Conversion::
7313: * Interpret/Compile states::
7314: * Literals::
7315: * Interpreter Directives::
1.26 crook 7316: @end menu
1.1 anton 7317:
1.29 crook 7318: @node Input Sources, Number Conversion, The Text Interpreter, The Text Interpreter
7319: @subsection Input Sources
7320: @cindex input sources
7321: @cindex text interpreter - input sources
7322:
1.44 crook 7323: By default, the text interpreter processes input from the user input
1.29 crook 7324: device (the keyboard) when Forth starts up. The text interpreter can
7325: process input from any of these sources:
7326:
7327: @itemize @bullet
7328: @item
7329: The user input device -- the keyboard.
7330: @item
7331: A file, using the words described in @ref{Forth source files}.
7332: @item
7333: A block, using the words described in @ref{Blocks}.
7334: @item
7335: A text string, using @code{evaluate}.
7336: @end itemize
7337:
7338: A program can identify the current input device from the values of
7339: @code{source-id} and @code{blk}.
7340:
1.44 crook 7341:
1.29 crook 7342: doc-source-id
7343: doc-blk
7344:
7345: doc-save-input
7346: doc-restore-input
7347:
7348: doc-evaluate
1.1 anton 7349:
1.29 crook 7350:
1.44 crook 7351:
1.29 crook 7352: @node Number Conversion, Interpret/Compile states, Input Sources, The Text Interpreter
1.26 crook 7353: @subsection Number Conversion
7354: @cindex number conversion
7355: @cindex double-cell numbers, input format
7356: @cindex input format for double-cell numbers
7357: @cindex single-cell numbers, input format
7358: @cindex input format for single-cell numbers
7359: @cindex floating-point numbers, input format
7360: @cindex input format for floating-point numbers
1.1 anton 7361:
1.29 crook 7362: This section describes the rules that the text interpreter uses when it
7363: tries to convert a string into a number.
1.1 anton 7364:
1.26 crook 7365: Let <digit> represent any character that is a legal digit in the current
1.29 crook 7366: number base@footnote{For example, 0-9 when the number base is decimal or
7367: 0-9, A-F when the number base is hexadecimal.}.
1.1 anton 7368:
1.26 crook 7369: Let <decimal digit> represent any character in the range 0-9.
1.1 anton 7370:
1.29 crook 7371: Let @{@i{a b}@} represent the @i{optional} presence of any of the characters
7372: in the braces (@i{a} or @i{b} or neither).
1.1 anton 7373:
1.26 crook 7374: Let * represent any number of instances of the previous character
7375: (including none).
1.1 anton 7376:
1.26 crook 7377: Let any other character represent itself.
1.1 anton 7378:
1.29 crook 7379: @noindent
1.26 crook 7380: Now, the conversion rules are:
1.21 crook 7381:
1.26 crook 7382: @itemize @bullet
7383: @item
7384: A string of the form <digit><digit>* is treated as a single-precision
1.29 crook 7385: (cell-sized) positive integer. Examples are 0 123 6784532 32343212343456 42
1.26 crook 7386: @item
7387: A string of the form -<digit><digit>* is treated as a single-precision
1.29 crook 7388: (cell-sized) negative integer, and is represented using 2's-complement
1.26 crook 7389: arithmetic. Examples are -45 -5681 -0
7390: @item
7391: A string of the form <digit><digit>*.<digit>* is treated as a double-precision
1.29 crook 7392: (double-cell-sized) positive integer. Examples are 3465. 3.465 34.65
7393: (all three of these represent the same number).
1.26 crook 7394: @item
7395: A string of the form -<digit><digit>*.<digit>* is treated as a
1.29 crook 7396: double-precision (double-cell-sized) negative integer, and is
1.26 crook 7397: represented using 2's-complement arithmetic. Examples are -3465. -3.465
1.29 crook 7398: -34.65 (all three of these represent the same number).
1.26 crook 7399: @item
1.29 crook 7400: A string of the form @{+ -@}<decimal digit>@{.@}<decimal digit>*@{e
7401: E@}@{+ -@}<decimal digit><decimal digit>* is treated as a floating-point
1.35 anton 7402: number. Examples are 1e 1e0 1.e 1.e0 +1e+0 (which all represent the same
1.29 crook 7403: number) +12.E-4
1.26 crook 7404: @end itemize
1.1 anton 7405:
1.26 crook 7406: By default, the number base used for integer number conversion is given
1.35 anton 7407: by the contents of the variable @code{base}. Note that a lot of
7408: confusion can result from unexpected values of @code{base}. If you
7409: change @code{base} anywhere, make sure to save the old value and restore
7410: it afterwards. In general I recommend keeping @code{base} decimal, and
7411: using the prefixes described below for the popular non-decimal bases.
1.1 anton 7412:
1.29 crook 7413: doc-dpl
1.26 crook 7414: doc-base
7415: doc-hex
7416: doc-decimal
1.1 anton 7417:
1.44 crook 7418:
1.26 crook 7419: @cindex '-prefix for character strings
7420: @cindex &-prefix for decimal numbers
7421: @cindex %-prefix for binary numbers
7422: @cindex $-prefix for hexadecimal numbers
1.35 anton 7423: Gforth allows you to override the value of @code{base} by using a
1.29 crook 7424: prefix@footnote{Some Forth implementations provide a similar scheme by
7425: implementing @code{$} etc. as parsing words that process the subsequent
7426: number in the input stream and push it onto the stack. For example, see
7427: @cite{Number Conversion and Literals}, by Wil Baden; Forth Dimensions
7428: 20(3) pages 26--27. In such implementations, unlike in Gforth, a space
7429: is required between the prefix and the number.} before the first digit
7430: of an (integer) number. Four prefixes are supported:
1.1 anton 7431:
1.26 crook 7432: @itemize @bullet
7433: @item
1.35 anton 7434: @code{&} -- decimal
1.26 crook 7435: @item
1.35 anton 7436: @code{%} -- binary
1.26 crook 7437: @item
1.35 anton 7438: @code{$} -- hexadecimal
1.26 crook 7439: @item
1.35 anton 7440: @code{'} -- base @code{max-char+1}
1.26 crook 7441: @end itemize
1.1 anton 7442:
1.26 crook 7443: Here are some examples, with the equivalent decimal number shown after
7444: in braces:
1.1 anton 7445:
1.26 crook 7446: -$41 (-65), %1001101 (205), %1001.0001 (145 - a double-precision number),
7447: 'AB (16706; ascii A is 65, ascii B is 66, number is 65*256 + 66),
7448: 'ab (24930; ascii a is 97, ascii B is 98, number is 97*256 + 98),
7449: &905 (905), $abc (2478), $ABC (2478).
1.1 anton 7450:
1.26 crook 7451: @cindex number conversion - traps for the unwary
1.29 crook 7452: @noindent
1.26 crook 7453: Number conversion has a number of traps for the unwary:
1.1 anton 7454:
1.26 crook 7455: @itemize @bullet
7456: @item
7457: You cannot determine the current number base using the code sequence
1.35 anton 7458: @code{base @@ .} -- the number base is always 10 in the current number
7459: base. Instead, use something like @code{base @@ dec.}
1.26 crook 7460: @item
7461: If the number base is set to a value greater than 14 (for example,
7462: hexadecimal), the number 123E4 is ambiguous; the conversion rules allow
7463: it to be intepreted as either a single-precision integer or a
7464: floating-point number (Gforth treats it as an integer). The ambiguity
7465: can be resolved by explicitly stating the sign of the mantissa and/or
7466: exponent: 123E+4 or +123E4 -- if the number base is decimal, no
7467: ambiguity arises; either representation will be treated as a
7468: floating-point number.
7469: @item
1.29 crook 7470: There is a word @code{bin} but it does @i{not} set the number base!
1.26 crook 7471: It is used to specify file types.
7472: @item
1.72 anton 7473: ANS Forth requires the @code{.} of a double-precision number to be the
7474: final character in the string. Gforth allows the @code{.} to be
7475: anywhere after the first digit.
1.26 crook 7476: @item
7477: The number conversion process does not check for overflow.
7478: @item
1.72 anton 7479: In an ANS Forth program @code{base} is required to be decimal when
7480: converting floating-point numbers. In Gforth, number conversion to
7481: floating-point numbers always uses base &10, irrespective of the value
7482: of @code{base}.
1.26 crook 7483: @end itemize
1.1 anton 7484:
1.49 anton 7485: You can read numbers into your programs with the words described in
7486: @ref{Input}.
1.1 anton 7487:
1.26 crook 7488: @node Interpret/Compile states, Literals, Number Conversion, The Text Interpreter
7489: @subsection Interpret/Compile states
7490: @cindex Interpret/Compile states
1.1 anton 7491:
1.29 crook 7492: A standard program is not permitted to change @code{state}
7493: explicitly. However, it can change @code{state} implicitly, using the
7494: words @code{[} and @code{]}. When @code{[} is executed it switches
7495: @code{state} to interpret state, and therefore the text interpreter
7496: starts interpreting. When @code{]} is executed it switches @code{state}
7497: to compile state and therefore the text interpreter starts
1.44 crook 7498: compiling. The most common usage for these words is for switching into
7499: interpret state and back from within a colon definition; this technique
1.49 anton 7500: can be used to compile a literal (for an example, @pxref{Literals}) or
7501: for conditional compilation (for an example, @pxref{Interpreter
7502: Directives}).
1.44 crook 7503:
1.35 anton 7504:
7505: @c This is a bad example: It's non-standard, and it's not necessary.
7506: @c However, I can't think of a good example for switching into compile
7507: @c state when there is no current word (@code{state}-smart words are not a
7508: @c good reason). So maybe we should use an example for switching into
7509: @c interpret @code{state} in a colon def. - anton
1.44 crook 7510: @c nac-> I agree. I started out by putting in the example, then realised
7511: @c that it was non-ANS, so wrote more words around it. I hope this
7512: @c re-written version is acceptable to you. I do want to keep the example
7513: @c as it is helpful for showing what is and what is not portable, particularly
7514: @c where it outlaws a style in common use.
7515:
1.72 anton 7516: @c anton: it's more important to show what's portable. After we have done
7517: @c that, we can also show what's not. In any case, I intend to write a
7518: @c section Macros (or so) which will also deal with [ ].
1.35 anton 7519:
1.44 crook 7520: @code{[} and @code{]} also give you the ability to switch into compile
7521: state and back, but we cannot think of any useful Standard application
7522: for this ability. Pre-ANS Forth textbooks have examples like this:
1.29 crook 7523:
7524: @example
7525: : AA ." this is A" ;
7526: : BB ." this is B" ;
7527: : CC ." this is C" ;
7528:
1.44 crook 7529: create table ] aa bb cc [
7530:
1.29 crook 7531: : go ( n -- ) \ n is offset into table.. 0 for 1st entry
7532: cells table + @ execute ;
7533: @end example
7534:
1.44 crook 7535: This example builds a jump table; @code{0 go} will display ``@code{this
7536: is A}''. Using @code{[} and @code{]} in this example is equivalent to
7537: defining @code{table} like this:
1.29 crook 7538:
7539: @example
1.44 crook 7540: create table ' aa COMPILE, ' bb COMPILE, ' cc COMPILE,
1.29 crook 7541: @end example
7542:
1.44 crook 7543: The problem with this code is that the definition of @code{table} is not
7544: portable -- it @i{compile}s execution tokens into code space. Whilst it
7545: @i{may} work on systems where code space and data space co-incide, the
1.29 crook 7546: Standard only allows data space to be assigned for a @code{CREATE}d
7547: word. In addition, the Standard only allows @code{@@} to access data
7548: space, whilst this example is using it to access code space. The only
7549: portable, Standard way to build this table is to build it in data space,
7550: like this:
7551:
7552: @example
7553: create table ' aa , ' bb , ' cc ,
7554: @end example
7555:
1.26 crook 7556: doc-state
7557: doc-[
7558: doc-]
1.1 anton 7559:
1.44 crook 7560:
1.26 crook 7561: @node Literals, Interpreter Directives, Interpret/Compile states, The Text Interpreter
7562: @subsection Literals
7563: @cindex Literals
1.21 crook 7564:
1.29 crook 7565: Often, you want to use a number within a colon definition. When you do
7566: this, the text interpreter automatically compiles the number as a
7567: @i{literal}. A literal is a number whose run-time effect is to be pushed
7568: onto the stack. If you had to do some maths to generate the number, you
7569: might write it like this:
7570:
7571: @example
7572: : HOUR-TO-SEC ( n1 -- n2 )
7573: 60 * \ to minutes
7574: 60 * ; \ to seconds
7575: @end example
7576:
7577: It is very clear what this definition is doing, but it's inefficient
7578: since it is performing 2 multiples at run-time. An alternative would be
7579: to write:
7580:
7581: @example
7582: : HOUR-TO-SEC ( n1 -- n2 )
7583: 3600 * ; \ to seconds
7584: @end example
7585:
7586: Which does the same thing, and has the advantage of using a single
7587: multiply. Ideally, we'd like the efficiency of the second with the
7588: readability of the first.
7589:
7590: @code{Literal} allows us to achieve that. It takes a number from the
7591: stack and lays it down in the current definition just as though the
7592: number had been typed directly into the definition. Our first attempt
7593: might look like this:
7594:
7595: @example
7596: 60 \ mins per hour
7597: 60 * \ seconds per minute
7598: : HOUR-TO-SEC ( n1 -- n2 )
7599: Literal * ; \ to seconds
7600: @end example
7601:
7602: But this produces the error message @code{unstructured}. What happened?
7603: The stack notation for @code{:} is (@i{ -- colon-sys}) and the size of
7604: @i{colon-sys} is implementation-defined. In other words, once we start a
7605: colon definition we can't portably access anything that was on the stack
7606: before the definition began@footnote{@cite{Two Problems in ANS Forth},
7607: by Thomas Worthington; Forth Dimensions 20(2) pages 32--34 describes
7608: some situations where you might want to access stack items above
7609: colon-sys, and provides a solution to the problem.}. The correct way of
7610: solving this problem in this instance is to use @code{[ ]} like this:
7611:
7612: @example
7613: : HOUR-TO-SEC ( n1 -- n2 )
7614: [ 60 \ minutes per hour
7615: 60 * ] \ seconds per minute
7616: LITERAL * ; \ to seconds
7617: @end example
1.23 crook 7618:
1.44 crook 7619:
1.26 crook 7620: doc-literal
7621: doc-]L
7622: doc-2literal
7623: doc-fliteral
1.1 anton 7624:
1.44 crook 7625:
1.48 anton 7626: @node Interpreter Directives, , Literals, The Text Interpreter
1.26 crook 7627: @subsection Interpreter Directives
7628: @cindex interpreter directives
1.72 anton 7629: @cindex conditional compilation
1.1 anton 7630:
1.29 crook 7631: These words are usually used in interpret state; typically to control
7632: which parts of a source file are processed by the text
1.26 crook 7633: interpreter. There are only a few ANS Forth Standard words, but Gforth
7634: supplements these with a rich set of immediate control structure words
7635: to compensate for the fact that the non-immediate versions can only be
1.29 crook 7636: used in compile state (@pxref{Control Structures}). Typical usages:
7637:
7638: @example
1.72 anton 7639: FALSE Constant HAVE-ASSEMBLER
1.29 crook 7640: .
7641: .
1.72 anton 7642: HAVE-ASSEMBLER [IF]
1.29 crook 7643: : ASSEMBLER-FEATURE
7644: ...
7645: ;
7646: [ENDIF]
7647: .
7648: .
7649: : SEE
7650: ... \ general-purpose SEE code
1.72 anton 7651: [ HAVE-ASSEMBLER [IF] ]
1.29 crook 7652: ... \ assembler-specific SEE code
7653: [ [ENDIF] ]
7654: ;
7655: @end example
1.1 anton 7656:
1.44 crook 7657:
1.26 crook 7658: doc-[IF]
7659: doc-[ELSE]
7660: doc-[THEN]
7661: doc-[ENDIF]
1.1 anton 7662:
1.26 crook 7663: doc-[IFDEF]
7664: doc-[IFUNDEF]
1.1 anton 7665:
1.26 crook 7666: doc-[?DO]
7667: doc-[DO]
7668: doc-[FOR]
7669: doc-[LOOP]
7670: doc-[+LOOP]
7671: doc-[NEXT]
1.1 anton 7672:
1.26 crook 7673: doc-[BEGIN]
7674: doc-[UNTIL]
7675: doc-[AGAIN]
7676: doc-[WHILE]
7677: doc-[REPEAT]
1.1 anton 7678:
1.27 crook 7679:
1.26 crook 7680: @c -------------------------------------------------------------
1.47 crook 7681: @node Word Lists, Environmental Queries, The Text Interpreter, Words
1.26 crook 7682: @section Word Lists
7683: @cindex word lists
1.32 anton 7684: @cindex header space
1.1 anton 7685:
1.36 anton 7686: A wordlist is a list of named words; you can add new words and look up
7687: words by name (and you can remove words in a restricted way with
7688: markers). Every named (and @code{reveal}ed) word is in one wordlist.
7689:
7690: @cindex search order stack
7691: The text interpreter searches the wordlists present in the search order
7692: (a stack of wordlists), from the top to the bottom. Within each
7693: wordlist, the search starts conceptually at the newest word; i.e., if
7694: two words in a wordlist have the same name, the newer word is found.
1.1 anton 7695:
1.26 crook 7696: @cindex compilation word list
1.36 anton 7697: New words are added to the @dfn{compilation wordlist} (aka current
7698: wordlist).
1.1 anton 7699:
1.36 anton 7700: @cindex wid
7701: A word list is identified by a cell-sized word list identifier (@i{wid})
7702: in much the same way as a file is identified by a file handle. The
7703: numerical value of the wid has no (portable) meaning, and might change
7704: from session to session.
1.1 anton 7705:
1.29 crook 7706: The ANS Forth ``Search order'' word set is intended to provide a set of
7707: low-level tools that allow various different schemes to be
1.74 anton 7708: implemented. Gforth also provides @code{vocabulary}, a traditional Forth
1.26 crook 7709: word. @file{compat/vocabulary.fs} provides an implementation in ANS
1.45 crook 7710: Forth.
1.1 anton 7711:
1.27 crook 7712: @comment TODO: locals section refers to here, saying that every word list (aka
7713: @comment vocabulary) has its own methods for searching etc. Need to document that.
1.1 anton 7714:
1.45 crook 7715: @comment TODO: document markers, reveal, tables, mappedwordlist
7716:
7717: @comment the gforthman- prefix is used to pick out the true definition of a
1.27 crook 7718: @comment word from the source files, rather than some alias.
1.44 crook 7719:
1.26 crook 7720: doc-forth-wordlist
7721: doc-definitions
7722: doc-get-current
7723: doc-set-current
7724: doc-get-order
1.45 crook 7725: doc---gforthman-set-order
1.26 crook 7726: doc-wordlist
1.30 anton 7727: doc-table
1.36 anton 7728: doc-push-order
7729: doc-previous
1.26 crook 7730: doc-also
1.45 crook 7731: doc---gforthman-forth
1.26 crook 7732: doc-only
1.45 crook 7733: doc---gforthman-order
1.15 anton 7734:
1.26 crook 7735: doc-find
7736: doc-search-wordlist
1.15 anton 7737:
1.26 crook 7738: doc-words
7739: doc-vlist
1.44 crook 7740: @c doc-words-deferred
1.1 anton 7741:
1.74 anton 7742: @c doc-mappedwordlist @c map-structure undefined, implemantation-specific
1.26 crook 7743: doc-root
7744: doc-vocabulary
7745: doc-seal
7746: doc-vocs
7747: doc-current
7748: doc-context
1.1 anton 7749:
1.44 crook 7750:
1.26 crook 7751: @menu
1.75 ! anton 7752: * Vocabularies::
1.67 anton 7753: * Why use word lists?::
1.75 ! anton 7754: * Word list example::
1.26 crook 7755: @end menu
7756:
1.75 ! anton 7757: @node Vocabularies, Why use word lists?, Word Lists, Word Lists
! 7758: @subsection Vocabularies
! 7759: @cindex Vocabularies, detailed explanation
! 7760:
! 7761: Here is an example of creating and using a new wordlist using ANS
! 7762: Forth words:
! 7763:
! 7764: @example
! 7765: wordlist constant my-new-words-wordlist
! 7766: : my-new-words get-order nip my-new-words-wordlist swap set-order ;
! 7767:
! 7768: \ add it to the search order
! 7769: also my-new-words
! 7770:
! 7771: \ alternatively, add it to the search order and make it
! 7772: \ the compilation word list
! 7773: also my-new-words definitions
! 7774: \ type "order" to see the problem
! 7775: @end example
! 7776:
! 7777: The problem with this example is that @code{order} has no way to
! 7778: associate the name @code{my-new-words} with the wid of the word list (in
! 7779: Gforth, @code{order} and @code{vocs} will display @code{???} for a wid
! 7780: that has no associated name). There is no Standard way of associating a
! 7781: name with a wid.
! 7782:
! 7783: In Gforth, this example can be re-coded using @code{vocabulary}, which
! 7784: associates a name with a wid:
! 7785:
! 7786: @example
! 7787: vocabulary my-new-words
! 7788:
! 7789: \ add it to the search order
! 7790: also my-new-words
! 7791:
! 7792: \ alternatively, add it to the search order and make it
! 7793: \ the compilation word list
! 7794: my-new-words definitions
! 7795: \ type "order" to see that the problem is solved
! 7796: @end example
! 7797:
! 7798:
! 7799: @node Why use word lists?, Word list example, Vocabularies, Word Lists
1.26 crook 7800: @subsection Why use word lists?
7801: @cindex word lists - why use them?
7802:
1.74 anton 7803: Here are some reasons why people use wordlists:
1.26 crook 7804:
7805: @itemize @bullet
1.74 anton 7806:
7807: @c anton: Gforth's hashing implementation makes the search speed
7808: @c independent from the number of words. But it is linear with the number
7809: @c of wordlists that have to be searched, so in effect using more wordlists
7810: @c actually slows down compilation.
7811:
7812: @c @item
7813: @c To improve compilation speed by reducing the number of header space
7814: @c entries that must be searched. This is achieved by creating a new
7815: @c word list that contains all of the definitions that are used in the
7816: @c definition of a Forth system but which would not usually be used by
7817: @c programs running on that system. That word list would be on the search
7818: @c list when the Forth system was compiled but would be removed from the
7819: @c search list for normal operation. This can be a useful technique for
7820: @c low-performance systems (for example, 8-bit processors in embedded
7821: @c systems) but is unlikely to be necessary in high-performance desktop
7822: @c systems.
7823:
1.26 crook 7824: @item
7825: To prevent a set of words from being used outside the context in which
7826: they are valid. Two classic examples of this are an integrated editor
7827: (all of the edit commands are defined in a separate word list; the
7828: search order is set to the editor word list when the editor is invoked;
7829: the old search order is restored when the editor is terminated) and an
7830: integrated assembler (the op-codes for the machine are defined in a
7831: separate word list which is used when a @code{CODE} word is defined).
1.74 anton 7832:
7833: @item
7834: To organize the words of an application or library into a user-visible
7835: set (in @code{forth-wordlist} or some other common wordlist) and a set
7836: of helper words used just for the implementation (hidden in a separate
1.75 ! anton 7837: wordlist). This keeps @code{words}' output smaller, separates
! 7838: implementation and interface, and reduces the chance of name conflicts
! 7839: within the common wordlist.
1.74 anton 7840:
1.26 crook 7841: @item
7842: To prevent a name-space clash between multiple definitions with the same
7843: name. For example, when building a cross-compiler you might have a word
7844: @code{IF} that generates conditional code for your target system. By
7845: placing this definition in a different word list you can control whether
7846: the host system's @code{IF} or the target system's @code{IF} get used in
7847: any particular context by controlling the order of the word lists on the
7848: search order stack.
1.74 anton 7849:
1.26 crook 7850: @end itemize
1.1 anton 7851:
1.74 anton 7852: The downsides of using wordlists are:
7853:
7854: @itemize
7855:
7856: @item
7857: Debugging becomes more cumbersome.
7858:
7859: @item
7860: Name conflicts worked around with wordlists are still there, and you
7861: have to arrange the search order carefully to get the desired results;
7862: if you forget to do that, you get hard-to-find errors (as in any case
7863: where you read the code differently from the compiler; @code{see} can
1.75 ! anton 7864: help seeing which of several possible words the name resolves to in such
! 7865: cases). @code{See} displays just the name of the words, not what
! 7866: wordlist they belong to, so it might be misleading. Using unique names
! 7867: is a better approach to avoid name conflicts.
1.74 anton 7868:
7869: @item
7870: You have to explicitly undo any changes to the search order. In many
7871: cases it would be more convenient if this happened implicitly. Gforth
7872: currently does not provide such a feature, but it may do so in the
7873: future.
7874: @end itemize
7875:
7876:
1.75 ! anton 7877: @node Word list example, , Why use word lists?, Word Lists
! 7878: @subsection Word list example
! 7879: @cindex word lists - example
1.1 anton 7880:
1.74 anton 7881: The following example is from the
7882: @uref{http://www.complang.tuwien.ac.at/forth/garbage-collection.zip,
7883: garbage collector} and uses wordlists to separate public words from
7884: helper words:
7885:
7886: @example
7887: get-current ( wid )
7888: vocabulary garbage-collector also garbage-collector definitions
7889: ... \ define helper words
7890: ( wid ) set-current \ restore original (i.e., public) compilation wordlist
7891: ... \ define the public (i.e., API) words
7892: \ they can refer to the helper words
7893: previous \ restore original search order (helper words become invisible)
7894: @end example
7895:
1.26 crook 7896: @c -------------------------------------------------------------
7897: @node Environmental Queries, Files, Word Lists, Words
7898: @section Environmental Queries
7899: @cindex environmental queries
1.21 crook 7900:
1.26 crook 7901: ANS Forth introduced the idea of ``environmental queries'' as a way
7902: for a program running on a system to determine certain characteristics of the system.
7903: The Standard specifies a number of strings that might be recognised by a system.
1.21 crook 7904:
1.32 anton 7905: The Standard requires that the header space used for environmental queries
7906: be distinct from the header space used for definitions.
1.21 crook 7907:
1.26 crook 7908: Typically, environmental queries are supported by creating a set of
1.29 crook 7909: definitions in a word list that is @i{only} used during environmental
1.26 crook 7910: queries; that is what Gforth does. There is no Standard way of adding
7911: definitions to the set of recognised environmental queries, but any
7912: implementation that supports the loading of optional word sets must have
7913: some mechanism for doing this (after loading the word set, the
7914: associated environmental query string must return @code{true}). In
7915: Gforth, the word list used to honour environmental queries can be
7916: manipulated just like any other word list.
1.21 crook 7917:
1.44 crook 7918:
1.26 crook 7919: doc-environment?
7920: doc-environment-wordlist
1.21 crook 7921:
1.26 crook 7922: doc-gforth
7923: doc-os-class
1.21 crook 7924:
1.44 crook 7925:
1.26 crook 7926: Note that, whilst the documentation for (e.g.) @code{gforth} shows it
7927: returning two items on the stack, querying it using @code{environment?}
7928: will return an additional item; the @code{true} flag that shows that the
7929: string was recognised.
1.21 crook 7930:
1.26 crook 7931: @comment TODO Document the standard strings or note where they are documented herein
1.21 crook 7932:
1.26 crook 7933: Here are some examples of using environmental queries:
1.21 crook 7934:
1.26 crook 7935: @example
7936: s" address-unit-bits" environment? 0=
7937: [IF]
7938: cr .( environmental attribute address-units-bits unknown... ) cr
1.75 ! anton 7939: [ELSE]
! 7940: drop \ ensure balanced stack effect
1.26 crook 7941: [THEN]
1.21 crook 7942:
1.75 ! anton 7943: \ this might occur in the prelude of a standard program that uses THROW
! 7944: s" exception" environment? [IF]
! 7945: 0= [IF]
! 7946: : throw abort" exception thrown" ;
! 7947: [THEN]
! 7948: [ELSE] \ we don't know, so make sure
! 7949: : throw abort" exception thrown" ;
! 7950: [THEN]
1.21 crook 7951:
1.26 crook 7952: s" gforth" environment? [IF] .( Gforth version ) TYPE
7953: [ELSE] .( Not Gforth..) [THEN]
1.75 ! anton 7954:
! 7955: \ a program using v*
! 7956: s" gforth" environment? [IF]
! 7957: s" 0.5.0" compare 0< [IF] \ v* is a primitive since 0.5.0
! 7958: : v* ( f_addr1 nstride1 f_addr2 nstride2 ucount -- r )
! 7959: >r swap 2swap swap 0e r> 0 ?DO
! 7960: dup f@ over + 2swap dup f@ f* f+ over + 2swap
! 7961: LOOP
! 7962: 2drop 2drop ;
! 7963: [THEN]
! 7964: [ELSE] \
! 7965: : v* ( f_addr1 nstride1 f_addr2 nstride2 ucount -- r )
! 7966: ...
! 7967: [THEN]
1.26 crook 7968: @end example
1.21 crook 7969:
1.26 crook 7970: Here is an example of adding a definition to the environment word list:
1.21 crook 7971:
1.26 crook 7972: @example
7973: get-current environment-wordlist set-current
7974: true constant block
7975: true constant block-ext
7976: set-current
7977: @end example
1.21 crook 7978:
1.26 crook 7979: You can see what definitions are in the environment word list like this:
1.21 crook 7980:
1.26 crook 7981: @example
1.75 ! anton 7982: environment-wordlist push-order words previous
1.26 crook 7983: @end example
1.21 crook 7984:
7985:
1.26 crook 7986: @c -------------------------------------------------------------
7987: @node Files, Blocks, Environmental Queries, Words
7988: @section Files
1.28 crook 7989: @cindex files
7990: @cindex I/O - file-handling
1.21 crook 7991:
1.26 crook 7992: Gforth provides facilities for accessing files that are stored in the
7993: host operating system's file-system. Files that are processed by Gforth
7994: can be divided into two categories:
1.21 crook 7995:
1.23 crook 7996: @itemize @bullet
7997: @item
1.29 crook 7998: Files that are processed by the Text Interpreter (@dfn{Forth source files}).
1.23 crook 7999: @item
1.29 crook 8000: Files that are processed by some other program (@dfn{general files}).
1.26 crook 8001: @end itemize
8002:
8003: @menu
1.48 anton 8004: * Forth source files::
8005: * General files::
8006: * Search Paths::
1.26 crook 8007: @end menu
8008:
8009: @c -------------------------------------------------------------
8010: @node Forth source files, General files, Files, Files
8011: @subsection Forth source files
8012: @cindex including files
8013: @cindex Forth source files
1.21 crook 8014:
1.26 crook 8015: The simplest way to interpret the contents of a file is to use one of
8016: these two formats:
1.21 crook 8017:
1.26 crook 8018: @example
8019: include mysource.fs
8020: s" mysource.fs" included
8021: @end example
1.21 crook 8022:
1.75 ! anton 8023: You usually want to include a file only if it is not included already
1.26 crook 8024: (by, say, another source file). In that case, you can use one of these
1.45 crook 8025: three formats:
1.21 crook 8026:
1.26 crook 8027: @example
8028: require mysource.fs
8029: needs mysource.fs
8030: s" mysource.fs" required
8031: @end example
1.21 crook 8032:
1.26 crook 8033: @cindex stack effect of included files
8034: @cindex including files, stack effect
1.45 crook 8035: It is good practice to write your source files such that interpreting them
8036: does not change the stack. Source files designed in this way can be used with
1.26 crook 8037: @code{required} and friends without complications. For example:
1.21 crook 8038:
1.26 crook 8039: @example
1.75 ! anton 8040: 1024 require foo.fs drop
1.26 crook 8041: @end example
1.21 crook 8042:
1.75 ! anton 8043: Here you want to pass the argument 1024 (e.g., a buffer size) to
! 8044: @file{foo.fs}. Interpreting @file{foo.fs} has the stack effect ( n -- n
! 8045: ), which allows its use with @code{require}. Of course with such
! 8046: parameters to required files, you have to ensure that the first
! 8047: @code{require} fits for all uses (i.e., @code{require} it early in the
! 8048: master load file).
1.44 crook 8049:
1.26 crook 8050: doc-include-file
8051: doc-included
1.28 crook 8052: doc-included?
1.26 crook 8053: doc-include
8054: doc-required
8055: doc-require
8056: doc-needs
1.75 ! anton 8057: @c doc-init-included-files @c internal
! 8058: @c doc-loadfilename @c internal word
! 8059: doc-sourcefilename
! 8060: doc-sourceline#
1.44 crook 8061:
1.26 crook 8062: A definition in ANS Forth for @code{required} is provided in
8063: @file{compat/required.fs}.
1.21 crook 8064:
1.26 crook 8065: @c -------------------------------------------------------------
8066: @node General files, Search Paths, Forth source files, Files
8067: @subsection General files
8068: @cindex general files
8069: @cindex file-handling
1.21 crook 8070:
1.75 ! anton 8071: Files are opened/created by name and type. The following file access
! 8072: methods (FAMs) are recognised:
1.44 crook 8073:
1.75 ! anton 8074: @cindex fam (file access method)
1.26 crook 8075: doc-r/o
8076: doc-r/w
8077: doc-w/o
8078: doc-bin
1.1 anton 8079:
1.44 crook 8080:
1.26 crook 8081: When a file is opened/created, it returns a file identifier,
1.29 crook 8082: @i{wfileid} that is used for all other file commands. All file
8083: commands also return a status value, @i{wior}, that is 0 for a
1.26 crook 8084: successful operation and an implementation-defined non-zero value in the
8085: case of an error.
1.21 crook 8086:
1.44 crook 8087:
1.26 crook 8088: doc-open-file
8089: doc-create-file
1.21 crook 8090:
1.26 crook 8091: doc-close-file
8092: doc-delete-file
8093: doc-rename-file
8094: doc-read-file
8095: doc-read-line
8096: doc-write-file
8097: doc-write-line
8098: doc-emit-file
8099: doc-flush-file
1.21 crook 8100:
1.26 crook 8101: doc-file-status
8102: doc-file-position
8103: doc-reposition-file
8104: doc-file-size
8105: doc-resize-file
1.21 crook 8106:
1.44 crook 8107:
1.26 crook 8108: @c ---------------------------------------------------------
1.48 anton 8109: @node Search Paths, , General files, Files
1.26 crook 8110: @subsection Search Paths
8111: @cindex path for @code{included}
8112: @cindex file search path
8113: @cindex @code{include} search path
8114: @cindex search path for files
1.21 crook 8115:
1.26 crook 8116: If you specify an absolute filename (i.e., a filename starting with
8117: @file{/} or @file{~}, or with @file{:} in the second position (as in
8118: @samp{C:...})) for @code{included} and friends, that file is included
8119: just as you would expect.
1.21 crook 8120:
1.75 ! anton 8121: If the filename starts with @file{./}, this refers to the directory that
! 8122: the present file was @code{included} from. This allows files to include
! 8123: other files relative to their own position (irrespective of the current
! 8124: working directory or the absolute position). This feature is essential
! 8125: for libraries consisting of several files, where a file may include
! 8126: other files from the library. It corresponds to @code{#include "..."}
! 8127: in C. If the current input source is not a file, @file{.} refers to the
! 8128: directory of the innermost file being included, or, if there is no file
! 8129: being included, to the current working directory.
! 8130:
! 8131: For relative filenames (not starting with @file{./}), Gforth uses a
! 8132: search path similar to Forth's search order (@pxref{Word Lists}). It
! 8133: tries to find the given filename in the directories present in the path,
! 8134: and includes the first one it finds. There are separate search paths for
! 8135: Forth source files and general files. If the search path contains the
! 8136: directory @file{.}, this refers to the directory of the current file, or
! 8137: the working directory, as if the file had been specified with @file{./}.
1.21 crook 8138:
1.26 crook 8139: Use @file{~+} to refer to the current working directory (as in the
8140: @code{bash}).
1.1 anton 8141:
1.75 ! anton 8142: @c anton: fold the following subsubsections into this subsection?
1.1 anton 8143:
1.48 anton 8144: @menu
1.75 ! anton 8145: * Source Search Paths::
1.48 anton 8146: * General Search Paths::
8147: @end menu
8148:
1.26 crook 8149: @c ---------------------------------------------------------
1.75 ! anton 8150: @node Source Search Paths, General Search Paths, Search Paths, Search Paths
! 8151: @subsubsection Source Search Paths
! 8152: @cindex search path control, source files
1.5 anton 8153:
1.26 crook 8154: The search path is initialized when you start Gforth (@pxref{Invoking
1.75 ! anton 8155: Gforth}). You can display it and change it using @code{fpath} in
! 8156: combination with the general path handling words.
1.5 anton 8157:
1.75 ! anton 8158: doc-fpath
! 8159: @c the functionality of the following words is easily available through
! 8160: @c fpath and the general path words. The may go away.
! 8161: @c doc-.fpath
! 8162: @c doc-fpath+
! 8163: @c doc-fpath=
! 8164: @c doc-open-fpath-file
1.44 crook 8165:
8166: @noindent
1.26 crook 8167: Here is an example of using @code{fpath} and @code{require}:
1.5 anton 8168:
1.26 crook 8169: @example
1.75 ! anton 8170: fpath path= /usr/lib/forth/|./
1.26 crook 8171: require timer.fs
8172: @end example
1.5 anton 8173:
1.75 ! anton 8174:
1.26 crook 8175: @c ---------------------------------------------------------
1.75 ! anton 8176: @node General Search Paths, , Source Search Paths, Search Paths
1.26 crook 8177: @subsubsection General Search Paths
1.75 ! anton 8178: @cindex search path control, source files
1.5 anton 8179:
1.26 crook 8180: Your application may need to search files in several directories, like
8181: @code{included} does. To facilitate this, Gforth allows you to define
8182: and use your own search paths, by providing generic equivalents of the
8183: Forth search path words:
1.5 anton 8184:
1.75 ! anton 8185: doc-open-path-file
! 8186: doc-path-allot
! 8187: doc-clear-path
! 8188: doc-also-path
1.26 crook 8189: doc-.path
8190: doc-path+
8191: doc-path=
1.5 anton 8192:
1.75 ! anton 8193: @c anton: better define a word for it, say "path-allot ( ucount -- path-addr )
1.44 crook 8194:
1.75 ! anton 8195: Here's an example of creating an empty search path:
! 8196: @c
1.26 crook 8197: @example
1.75 ! anton 8198: create mypath 500 path-allot \ maximum length 500 chars (is checked)
1.26 crook 8199: @end example
1.5 anton 8200:
1.26 crook 8201: @c -------------------------------------------------------------
8202: @node Blocks, Other I/O, Files, Words
8203: @section Blocks
1.28 crook 8204: @cindex I/O - blocks
8205: @cindex blocks
8206:
8207: When you run Gforth on a modern desk-top computer, it runs under the
8208: control of an operating system which provides certain services. One of
8209: these services is @var{file services}, which allows Forth source code
8210: and data to be stored in files and read into Gforth (@pxref{Files}).
8211:
8212: Traditionally, Forth has been an important programming language on
8213: systems where it has interfaced directly to the underlying hardware with
8214: no intervening operating system. Forth provides a mechanism, called
1.29 crook 8215: @dfn{blocks}, for accessing mass storage on such systems.
1.28 crook 8216:
8217: A block is a 1024-byte data area, which can be used to hold data or
8218: Forth source code. No structure is imposed on the contents of the
8219: block. A block is identified by its number; blocks are numbered
8220: contiguously from 1 to an implementation-defined maximum.
8221:
8222: A typical system that used blocks but no operating system might use a
8223: single floppy-disk drive for mass storage, with the disks formatted to
8224: provide 256-byte sectors. Blocks would be implemented by assigning the
8225: first four sectors of the disk to block 1, the second four sectors to
8226: block 2 and so on, up to the limit of the capacity of the disk. The disk
8227: would not contain any file system information, just the set of blocks.
8228:
1.29 crook 8229: @cindex blocks file
1.28 crook 8230: On systems that do provide file services, blocks are typically
1.29 crook 8231: implemented by storing a sequence of blocks within a single @dfn{blocks
1.28 crook 8232: file}. The size of the blocks file will be an exact multiple of 1024
8233: bytes, corresponding to the number of blocks it contains. This is the
8234: mechanism that Gforth uses.
8235:
1.29 crook 8236: @cindex @file{blocks.fb}
1.75 ! anton 8237: Only one blocks file can be open at a time. If you use block words without
1.28 crook 8238: having specified a blocks file, Gforth defaults to the blocks file
8239: @file{blocks.fb}. Gforth uses the Forth search path when attempting to
1.75 ! anton 8240: locate a blocks file (@pxref{Source Search Paths}).
1.28 crook 8241:
1.29 crook 8242: @cindex block buffers
1.28 crook 8243: When you read and write blocks under program control, Gforth uses a
1.29 crook 8244: number of @dfn{block buffers} as intermediate storage. These buffers are
1.28 crook 8245: not used when you use @code{load} to interpret the contents of a block.
8246:
1.75 ! anton 8247: The behaviour of the block buffers is analagous to that of a cache.
! 8248: Each block buffer has three states:
1.28 crook 8249:
8250: @itemize @bullet
8251: @item
8252: Unassigned
8253: @item
8254: Assigned-clean
8255: @item
8256: Assigned-dirty
8257: @end itemize
8258:
1.29 crook 8259: Initially, all block buffers are @i{unassigned}. In order to access a
1.28 crook 8260: block, the block (specified by its block number) must be assigned to a
8261: block buffer.
8262:
8263: The assignment of a block to a block buffer is performed by @code{block}
8264: or @code{buffer}. Use @code{block} when you wish to modify the existing
8265: contents of a block. Use @code{buffer} when you don't care about the
8266: existing contents of the block@footnote{The ANS Forth definition of
1.35 anton 8267: @code{buffer} is intended not to cause disk I/O; if the data associated
1.28 crook 8268: with the particular block is already stored in a block buffer due to an
8269: earlier @code{block} command, @code{buffer} will return that block
8270: buffer and the existing contents of the block will be
8271: available. Otherwise, @code{buffer} will simply assign a new, empty
1.29 crook 8272: block buffer for the block.}.
1.28 crook 8273:
1.47 crook 8274: Once a block has been assigned to a block buffer using @code{block} or
1.75 ! anton 8275: @code{buffer}, that block buffer becomes the @i{current block
! 8276: buffer}. Data may only be manipulated (read or written) within the
! 8277: current block buffer.
1.47 crook 8278:
8279: When the contents of the current block buffer has been modified it is
1.48 anton 8280: necessary, @emph{before calling @code{block} or @code{buffer} again}, to
1.75 ! anton 8281: either abandon the changes (by doing nothing) or mark the block as
! 8282: changed (assigned-dirty), using @code{update}. Using @code{update} does
! 8283: not change the blocks file; it simply changes a block buffer's state to
! 8284: @i{assigned-dirty}. The block will be written implicitly when it's
! 8285: buffer is needed for another block, or explicitly by @code{flush} or
! 8286: @code{save-buffers}.
! 8287:
! 8288: word @code{Flush} writes all @i{assigned-dirty} blocks back to the
! 8289: blocks file on disk. Leaving Gforth with @code{bye} also performs a
! 8290: @code{flush}.
1.28 crook 8291:
1.29 crook 8292: In Gforth, @code{block} and @code{buffer} use a @i{direct-mapped}
1.28 crook 8293: algorithm to assign a block buffer to a block. That means that any
8294: particular block can only be assigned to one specific block buffer,
1.29 crook 8295: called (for the particular operation) the @i{victim buffer}. If the
1.47 crook 8296: victim buffer is @i{unassigned} or @i{assigned-clean} it is allocated to
8297: the new block immediately. If it is @i{assigned-dirty} its current
8298: contents are written back to the blocks file on disk before it is
1.28 crook 8299: allocated to the new block.
8300:
8301: Although no structure is imposed on the contents of a block, it is
8302: traditional to display the contents as 16 lines each of 64 characters. A
8303: block provides a single, continuous stream of input (for example, it
8304: acts as a single parse area) -- there are no end-of-line characters
8305: within a block, and no end-of-file character at the end of a
8306: block. There are two consequences of this:
1.26 crook 8307:
1.28 crook 8308: @itemize @bullet
8309: @item
8310: The last character of one line wraps straight into the first character
8311: of the following line
8312: @item
8313: The word @code{\} -- comment to end of line -- requires special
8314: treatment; in the context of a block it causes all characters until the
8315: end of the current 64-character ``line'' to be ignored.
8316: @end itemize
8317:
8318: In Gforth, when you use @code{block} with a non-existent block number,
1.45 crook 8319: the current blocks file will be extended to the appropriate size and the
1.28 crook 8320: block buffer will be initialised with spaces.
8321:
1.47 crook 8322: Gforth includes a simple block editor (type @code{use blocked.fb 0 list}
8323: for details) but doesn't encourage the use of blocks; the mechanism is
8324: only provided for backward compatibility -- ANS Forth requires blocks to
8325: be available when files are.
1.28 crook 8326:
8327: Common techniques that are used when working with blocks include:
8328:
8329: @itemize @bullet
8330: @item
8331: A screen editor that allows you to edit blocks without leaving the Forth
8332: environment.
8333: @item
8334: Shadow screens; where every code block has an associated block
8335: containing comments (for example: code in odd block numbers, comments in
8336: even block numbers). Typically, the block editor provides a convenient
8337: mechanism to toggle between code and comments.
8338: @item
8339: Load blocks; a single block (typically block 1) contains a number of
8340: @code{thru} commands which @code{load} the whole of the application.
8341: @end itemize
1.26 crook 8342:
1.29 crook 8343: See Frank Sergeant's Pygmy Forth to see just how well blocks can be
8344: integrated into a Forth programming environment.
1.26 crook 8345:
8346: @comment TODO what about errors on open-blocks?
1.44 crook 8347:
1.26 crook 8348: doc-open-blocks
8349: doc-use
1.75 ! anton 8350: doc-block-offset
1.26 crook 8351: doc-get-block-fid
8352: doc-block-position
1.28 crook 8353:
1.75 ! anton 8354: doc-list
1.28 crook 8355: doc-scr
8356:
1.45 crook 8357: doc---gforthman-block
1.28 crook 8358: doc-buffer
8359:
1.75 ! anton 8360: doc-empty-buffers
! 8361: doc-empty-buffer
1.26 crook 8362: doc-update
1.28 crook 8363: doc-updated?
1.26 crook 8364: doc-save-buffers
1.75 ! anton 8365: doc-save-buffer
1.26 crook 8366: doc-flush
1.28 crook 8367:
1.26 crook 8368: doc-load
8369: doc-thru
8370: doc-+load
8371: doc-+thru
1.45 crook 8372: doc---gforthman--->
1.26 crook 8373: doc-block-included
8374:
1.44 crook 8375:
1.26 crook 8376: @c -------------------------------------------------------------
8377: @node Other I/O, Programming Tools, Blocks, Words
8378: @section Other I/O
1.28 crook 8379: @cindex I/O - keyboard and display
1.26 crook 8380:
8381: @menu
8382: * Simple numeric output:: Predefined formats
8383: * Formatted numeric output:: Formatted (pictured) output
8384: * String Formats:: How Forth stores strings in memory
1.67 anton 8385: * Displaying characters and strings:: Other stuff
1.26 crook 8386: * Input:: Input
8387: @end menu
8388:
8389: @node Simple numeric output, Formatted numeric output, Other I/O, Other I/O
8390: @subsection Simple numeric output
1.28 crook 8391: @cindex numeric output - simple/free-format
1.5 anton 8392:
1.26 crook 8393: The simplest output functions are those that display numbers from the
8394: data or floating-point stacks. Floating-point output is always displayed
8395: using base 10. Numbers displayed from the data stack use the value stored
8396: in @code{base}.
1.5 anton 8397:
1.44 crook 8398:
1.26 crook 8399: doc-.
8400: doc-dec.
8401: doc-hex.
8402: doc-u.
8403: doc-.r
8404: doc-u.r
8405: doc-d.
8406: doc-ud.
8407: doc-d.r
8408: doc-ud.r
8409: doc-f.
8410: doc-fe.
8411: doc-fs.
1.5 anton 8412:
1.44 crook 8413:
1.26 crook 8414: Examples of printing the number 1234.5678E23 in the different floating-point output
8415: formats are shown below:
1.5 anton 8416:
8417: @example
1.26 crook 8418: f. 123456779999999000000000000.
8419: fe. 123.456779999999E24
8420: fs. 1.23456779999999E26
1.5 anton 8421: @end example
8422:
8423:
1.26 crook 8424: @node Formatted numeric output, String Formats, Simple numeric output, Other I/O
8425: @subsection Formatted numeric output
1.28 crook 8426: @cindex formatted numeric output
1.26 crook 8427: @cindex pictured numeric output
1.28 crook 8428: @cindex numeric output - formatted
1.26 crook 8429:
1.29 crook 8430: Forth traditionally uses a technique called @dfn{pictured numeric
1.26 crook 8431: output} for formatted printing of integers. In this technique, digits
8432: are extracted from the number (using the current output radix defined by
8433: @code{base}), converted to ASCII codes and appended to a string that is
8434: built in a scratch-pad area of memory (@pxref{core-idef,
8435: Implementation-defined options, Implementation-defined
8436: options}). Arbitrary characters can be appended to the string during the
8437: extraction process. The completed string is specified by an address
8438: and length and can be manipulated (@code{TYPE}ed, copied, modified)
8439: under program control.
1.5 anton 8440:
1.75 ! anton 8441: All of the integer output words described in the previous section
! 8442: (@pxref{Simple numeric output}) are implemented in Gforth using pictured
! 8443: numeric output.
1.5 anton 8444:
1.47 crook 8445: Three important things to remember about pictured numeric output:
1.5 anton 8446:
1.26 crook 8447: @itemize @bullet
8448: @item
1.28 crook 8449: It always operates on double-precision numbers; to display a
1.49 anton 8450: single-precision number, convert it first (for ways of doing this
8451: @pxref{Double precision}).
1.26 crook 8452: @item
1.28 crook 8453: It always treats the double-precision number as though it were
8454: unsigned. The examples below show ways of printing signed numbers.
1.26 crook 8455: @item
8456: The string is built up from right to left; least significant digit first.
8457: @end itemize
1.5 anton 8458:
1.44 crook 8459:
1.26 crook 8460: doc-<#
1.47 crook 8461: doc-<<#
1.26 crook 8462: doc-#
8463: doc-#s
8464: doc-hold
8465: doc-sign
8466: doc-#>
1.47 crook 8467: doc-#>>
1.5 anton 8468:
1.26 crook 8469: doc-represent
1.5 anton 8470:
1.44 crook 8471:
8472: @noindent
1.26 crook 8473: Here are some examples of using pictured numeric output:
1.5 anton 8474:
8475: @example
1.26 crook 8476: : my-u. ( u -- )
8477: \ Simplest use of pns.. behaves like Standard u.
8478: 0 \ convert to unsigned double
1.75 ! anton 8479: <<# \ start conversion
1.26 crook 8480: #s \ convert all digits
8481: #> \ complete conversion
1.75 ! anton 8482: TYPE SPACE \ display, with trailing space
! 8483: #>> ; \ release hold area
1.5 anton 8484:
1.26 crook 8485: : cents-only ( u -- )
8486: 0 \ convert to unsigned double
1.75 ! anton 8487: <<# \ start conversion
1.26 crook 8488: # # \ convert two least-significant digits
8489: #> \ complete conversion, discard other digits
1.75 ! anton 8490: TYPE SPACE \ display, with trailing space
! 8491: #>> ; \ release hold area
1.5 anton 8492:
1.26 crook 8493: : dollars-and-cents ( u -- )
8494: 0 \ convert to unsigned double
1.75 ! anton 8495: <<# \ start conversion
1.26 crook 8496: # # \ convert two least-significant digits
8497: [char] . hold \ insert decimal point
8498: #s \ convert remaining digits
8499: [char] $ hold \ append currency symbol
8500: #> \ complete conversion
1.75 ! anton 8501: TYPE SPACE \ display, with trailing space
! 8502: #>> ; \ release hold area
1.5 anton 8503:
1.26 crook 8504: : my-. ( n -- )
8505: \ handling negatives.. behaves like Standard .
8506: s>d \ convert to signed double
8507: swap over dabs \ leave sign byte followed by unsigned double
1.75 ! anton 8508: <<# \ start conversion
1.26 crook 8509: #s \ convert all digits
8510: rot sign \ get at sign byte, append "-" if needed
8511: #> \ complete conversion
1.75 ! anton 8512: TYPE SPACE \ display, with trailing space
! 8513: #>> ; \ release hold area
1.5 anton 8514:
1.26 crook 8515: : account. ( n -- )
1.75 ! anton 8516: \ accountants don't like minus signs, they use parentheses
1.26 crook 8517: \ for negative numbers
8518: s>d \ convert to signed double
8519: swap over dabs \ leave sign byte followed by unsigned double
1.75 ! anton 8520: <<# \ start conversion
1.26 crook 8521: 2 pick \ get copy of sign byte
8522: 0< IF [char] ) hold THEN \ right-most character of output
8523: #s \ convert all digits
8524: rot \ get at sign byte
8525: 0< IF [char] ( hold THEN
8526: #> \ complete conversion
1.75 ! anton 8527: TYPE SPACE \ display, with trailing space
! 8528: #>> ; \ release hold area
! 8529:
1.5 anton 8530: @end example
8531:
1.26 crook 8532: Here are some examples of using these words:
1.5 anton 8533:
8534: @example
1.26 crook 8535: 1 my-u. 1
8536: hex -1 my-u. decimal FFFFFFFF
8537: 1 cents-only 01
8538: 1234 cents-only 34
8539: 2 dollars-and-cents $0.02
8540: 1234 dollars-and-cents $12.34
8541: 123 my-. 123
8542: -123 my. -123
8543: 123 account. 123
8544: -456 account. (456)
1.5 anton 8545: @end example
8546:
8547:
1.26 crook 8548: @node String Formats, Displaying characters and strings, Formatted numeric output, Other I/O
8549: @subsection String Formats
1.27 crook 8550: @cindex strings - see character strings
8551: @cindex character strings - formats
1.28 crook 8552: @cindex I/O - see character strings
1.75 ! anton 8553: @cindex counted strings
! 8554:
! 8555: @c anton: this does not really belong here; maybe the memory section,
! 8556: @c or the principles chapter
1.26 crook 8557:
1.27 crook 8558: Forth commonly uses two different methods for representing character
8559: strings:
1.26 crook 8560:
8561: @itemize @bullet
8562: @item
8563: @cindex address of counted string
1.45 crook 8564: @cindex counted string
1.29 crook 8565: As a @dfn{counted string}, represented by a @i{c-addr}. The char
8566: addressed by @i{c-addr} contains a character-count, @i{n}, of the
8567: string and the string occupies the subsequent @i{n} char addresses in
1.26 crook 8568: memory.
8569: @item
1.29 crook 8570: As cell pair on the stack; @i{c-addr u}, where @i{u} is the length
8571: of the string in characters, and @i{c-addr} is the address of the
1.26 crook 8572: first byte of the string.
8573: @end itemize
8574:
8575: ANS Forth encourages the use of the second format when representing
1.75 ! anton 8576: strings.
1.26 crook 8577:
1.44 crook 8578:
1.26 crook 8579: doc-count
8580:
1.44 crook 8581:
1.49 anton 8582: For words that move, copy and search for strings see @ref{Memory
8583: Blocks}. For words that display characters and strings see
8584: @ref{Displaying characters and strings}.
1.26 crook 8585:
8586: @node Displaying characters and strings, Input, String Formats, Other I/O
8587: @subsection Displaying characters and strings
1.27 crook 8588: @cindex characters - compiling and displaying
8589: @cindex character strings - compiling and displaying
1.26 crook 8590:
8591: This section starts with a glossary of Forth words and ends with a set
8592: of examples.
8593:
1.44 crook 8594:
1.26 crook 8595: doc-bl
8596: doc-space
8597: doc-spaces
8598: doc-emit
8599: doc-toupper
8600: doc-."
8601: doc-.(
8602: doc-type
1.44 crook 8603: doc-typewhite
1.26 crook 8604: doc-cr
1.27 crook 8605: @cindex cursor control
1.26 crook 8606: doc-at-xy
8607: doc-page
8608: doc-s"
8609: doc-c"
8610: doc-char
8611: doc-[char]
8612: doc-sliteral
8613:
1.44 crook 8614:
8615: @noindent
1.26 crook 8616: As an example, consider the following text, stored in a file @file{test.fs}:
1.5 anton 8617:
8618: @example
1.26 crook 8619: .( text-1)
8620: : my-word
8621: ." text-2" cr
8622: .( text-3)
8623: ;
8624:
8625: ." text-4"
8626:
8627: : my-char
8628: [char] ALPHABET emit
8629: char emit
8630: ;
1.5 anton 8631: @end example
8632:
1.26 crook 8633: When you load this code into Gforth, the following output is generated:
1.5 anton 8634:
1.26 crook 8635: @example
1.30 anton 8636: @kbd{include test.fs @key{RET}} text-1text-3text-4 ok
1.26 crook 8637: @end example
1.5 anton 8638:
1.26 crook 8639: @itemize @bullet
8640: @item
8641: Messages @code{text-1} and @code{text-3} are displayed because @code{.(}
8642: is an immediate word; it behaves in the same way whether it is used inside
8643: or outside a colon definition.
8644: @item
8645: Message @code{text-4} is displayed because of Gforth's added interpretation
8646: semantics for @code{."}.
8647: @item
1.29 crook 8648: Message @code{text-2} is @i{not} displayed, because the text interpreter
1.26 crook 8649: performs the compilation semantics for @code{."} within the definition of
8650: @code{my-word}.
8651: @end itemize
1.5 anton 8652:
1.26 crook 8653: Here are some examples of executing @code{my-word} and @code{my-char}:
1.5 anton 8654:
1.26 crook 8655: @example
1.30 anton 8656: @kbd{my-word @key{RET}} text-2
1.26 crook 8657: ok
1.30 anton 8658: @kbd{my-char fred @key{RET}} Af ok
8659: @kbd{my-char jim @key{RET}} Aj ok
1.26 crook 8660: @end example
1.5 anton 8661:
8662: @itemize @bullet
8663: @item
1.26 crook 8664: Message @code{text-2} is displayed because of the run-time behaviour of
8665: @code{."}.
8666: @item
8667: @code{[char]} compiles the ``A'' from ``ALPHABET'' and puts its display code
8668: on the stack at run-time. @code{emit} always displays the character
8669: when @code{my-char} is executed.
8670: @item
8671: @code{char} parses a string at run-time and the second @code{emit} displays
8672: the first character of the string.
1.5 anton 8673: @item
1.26 crook 8674: If you type @code{see my-char} you can see that @code{[char]} discarded
8675: the text ``LPHABET'' and only compiled the display code for ``A'' into the
8676: definition of @code{my-char}.
1.5 anton 8677: @end itemize
8678:
8679:
8680:
1.48 anton 8681: @node Input, , Displaying characters and strings, Other I/O
1.26 crook 8682: @subsection Input
8683: @cindex input
1.28 crook 8684: @cindex I/O - see input
8685: @cindex parsing a string
1.5 anton 8686:
1.49 anton 8687: For ways of storing character strings in memory see @ref{String Formats}.
1.5 anton 8688:
1.27 crook 8689: @comment TODO examples for >number >float accept key key? pad parse word refill
1.29 crook 8690: @comment then index them
1.27 crook 8691:
1.44 crook 8692:
1.27 crook 8693: doc-key
8694: doc-key?
1.45 crook 8695: doc-ekey
8696: doc-ekey?
8697: doc-ekey>char
1.26 crook 8698: doc->number
8699: doc->float
8700: doc-accept
1.27 crook 8701: doc-pad
1.75 ! anton 8702: @c anton: these belong in the input stream section
1.27 crook 8703: doc-parse
8704: doc-word
8705: doc-sword
1.75 ! anton 8706: doc-name
1.27 crook 8707: doc-refill
8708: @comment obsolescent words..
8709: doc-convert
1.26 crook 8710: doc-query
8711: doc-expect
1.27 crook 8712: doc-span
1.5 anton 8713:
8714:
1.44 crook 8715:
1.5 anton 8716: @c -------------------------------------------------------------
1.26 crook 8717: @node Programming Tools, Assembler and Code Words, Other I/O, Words
8718: @section Programming Tools
8719: @cindex programming tools
1.12 anton 8720:
8721: @menu
1.26 crook 8722: * Debugging:: Simple and quick.
8723: * Assertions:: Making your programs self-checking.
1.46 pazsan 8724: * Singlestep Debugger:: Executing your program word by word.
1.5 anton 8725: @end menu
8726:
1.26 crook 8727: @node Debugging, Assertions, Programming Tools, Programming Tools
8728: @subsection Debugging
8729: @cindex debugging
1.5 anton 8730:
1.26 crook 8731: Languages with a slow edit/compile/link/test development loop tend to
8732: require sophisticated tracing/stepping debuggers to facilate
8733: productive debugging.
1.5 anton 8734:
1.26 crook 8735: A much better (faster) way in fast-compiling languages is to add
8736: printing code at well-selected places, let the program run, look at
8737: the output, see where things went wrong, add more printing code, etc.,
8738: until the bug is found.
1.5 anton 8739:
1.26 crook 8740: The simple debugging aids provided in @file{debugs.fs}
8741: are meant to support this style of debugging. In addition, there are
8742: words for non-destructively inspecting the stack and memory:
1.5 anton 8743:
1.44 crook 8744:
1.26 crook 8745: doc-.s
8746: doc-f.s
1.5 anton 8747:
1.44 crook 8748:
1.29 crook 8749: There is a word @code{.r} but it does @i{not} display the return
1.26 crook 8750: stack! It is used for formatted numeric output.
1.5 anton 8751:
1.44 crook 8752:
1.26 crook 8753: doc-depth
8754: doc-fdepth
8755: doc-clearstack
8756: doc-?
8757: doc-dump
1.5 anton 8758:
1.44 crook 8759:
1.26 crook 8760: The word @code{~~} prints debugging information (by default the source
8761: location and the stack contents). It is easy to insert. If you use Emacs
8762: it is also easy to remove (@kbd{C-x ~} in the Emacs Forth mode to
8763: query-replace them with nothing). The deferred words
8764: @code{printdebugdata} and @code{printdebugline} control the output of
8765: @code{~~}. The default source location output format works well with
8766: Emacs' compilation mode, so you can step through the program at the
8767: source level using @kbd{C-x `} (the advantage over a stepping debugger
8768: is that you can step in any direction and you know where the crash has
8769: happened or where the strange data has occurred).
1.5 anton 8770:
1.26 crook 8771: The default actions of @code{~~} clobber the contents of the pictured
8772: numeric output string, so you should not use @code{~~}, e.g., between
8773: @code{<#} and @code{#>}.
1.5 anton 8774:
1.44 crook 8775:
1.26 crook 8776: doc-~~
8777: doc-printdebugdata
8778: doc-printdebugline
1.5 anton 8779:
1.26 crook 8780: doc-see
8781: doc-marker
1.5 anton 8782:
1.44 crook 8783:
1.26 crook 8784: Here's an example of using @code{marker} at the start of a source file
8785: that you are debugging; it ensures that you only ever have one copy of
8786: the file's definitions compiled at any time:
1.5 anton 8787:
1.26 crook 8788: @example
8789: [IFDEF] my-code
8790: my-code
8791: [ENDIF]
1.5 anton 8792:
1.26 crook 8793: marker my-code
1.28 crook 8794: init-included-files
1.5 anton 8795:
1.26 crook 8796: \ .. definitions start here
8797: \ .
8798: \ .
8799: \ end
8800: @end example
1.5 anton 8801:
8802:
8803:
1.26 crook 8804: @node Assertions, Singlestep Debugger, Debugging, Programming Tools
8805: @subsection Assertions
8806: @cindex assertions
1.5 anton 8807:
1.26 crook 8808: It is a good idea to make your programs self-checking, especially if you
8809: make an assumption that may become invalid during maintenance (for
8810: example, that a certain field of a data structure is never zero). Gforth
1.29 crook 8811: supports @dfn{assertions} for this purpose. They are used like this:
1.23 crook 8812:
1.26 crook 8813: @example
1.29 crook 8814: assert( @i{flag} )
1.26 crook 8815: @end example
1.23 crook 8816:
1.26 crook 8817: The code between @code{assert(} and @code{)} should compute a flag, that
8818: should be true if everything is alright and false otherwise. It should
8819: not change anything else on the stack. The overall stack effect of the
8820: assertion is @code{( -- )}. E.g.
1.23 crook 8821:
1.26 crook 8822: @example
8823: assert( 1 1 + 2 = ) \ what we learn in school
8824: assert( dup 0<> ) \ assert that the top of stack is not zero
8825: assert( false ) \ this code should not be reached
8826: @end example
1.23 crook 8827:
1.26 crook 8828: The need for assertions is different at different times. During
8829: debugging, we want more checking, in production we sometimes care more
8830: for speed. Therefore, assertions can be turned off, i.e., the assertion
8831: becomes a comment. Depending on the importance of an assertion and the
8832: time it takes to check it, you may want to turn off some assertions and
8833: keep others turned on. Gforth provides several levels of assertions for
8834: this purpose:
1.23 crook 8835:
1.44 crook 8836:
1.26 crook 8837: doc-assert0(
8838: doc-assert1(
8839: doc-assert2(
8840: doc-assert3(
8841: doc-assert(
8842: doc-)
1.23 crook 8843:
1.44 crook 8844:
1.26 crook 8845: The variable @code{assert-level} specifies the highest assertions that
8846: are turned on. I.e., at the default @code{assert-level} of one,
8847: @code{assert0(} and @code{assert1(} assertions perform checking, while
8848: @code{assert2(} and @code{assert3(} assertions are treated as comments.
8849:
8850: The value of @code{assert-level} is evaluated at compile-time, not at
8851: run-time. Therefore you cannot turn assertions on or off at run-time;
8852: you have to set the @code{assert-level} appropriately before compiling a
8853: piece of code. You can compile different pieces of code at different
8854: @code{assert-level}s (e.g., a trusted library at level 1 and
8855: newly-written code at level 3).
1.23 crook 8856:
1.44 crook 8857:
1.26 crook 8858: doc-assert-level
1.23 crook 8859:
1.44 crook 8860:
1.26 crook 8861: If an assertion fails, a message compatible with Emacs' compilation mode
8862: is produced and the execution is aborted (currently with @code{ABORT"}.
8863: If there is interest, we will introduce a special throw code. But if you
8864: intend to @code{catch} a specific condition, using @code{throw} is
8865: probably more appropriate than an assertion).
1.23 crook 8866:
1.26 crook 8867: Definitions in ANS Forth for these assertion words are provided
8868: in @file{compat/assert.fs}.
1.23 crook 8869:
8870:
1.48 anton 8871: @node Singlestep Debugger, , Assertions, Programming Tools
1.26 crook 8872: @subsection Singlestep Debugger
8873: @cindex singlestep Debugger
8874: @cindex debugging Singlestep
1.23 crook 8875:
1.26 crook 8876: When you create a new word there's often the need to check whether it
8877: behaves correctly or not. You can do this by typing @code{dbg
8878: badword}. A debug session might look like this:
1.23 crook 8879:
1.26 crook 8880: @example
8881: : badword 0 DO i . LOOP ; ok
8882: 2 dbg badword
8883: : badword
8884: Scanning code...
1.23 crook 8885:
1.26 crook 8886: Nesting debugger ready!
1.23 crook 8887:
1.26 crook 8888: 400D4738 8049BC4 0 -> [ 2 ] 00002 00000
8889: 400D4740 8049F68 DO -> [ 0 ]
8890: 400D4744 804A0C8 i -> [ 1 ] 00000
8891: 400D4748 400C5E60 . -> 0 [ 0 ]
8892: 400D474C 8049D0C LOOP -> [ 0 ]
8893: 400D4744 804A0C8 i -> [ 1 ] 00001
8894: 400D4748 400C5E60 . -> 1 [ 0 ]
8895: 400D474C 8049D0C LOOP -> [ 0 ]
8896: 400D4758 804B384 ; -> ok
8897: @end example
1.23 crook 8898:
1.26 crook 8899: Each line displayed is one step. You always have to hit return to
8900: execute the next word that is displayed. If you don't want to execute
8901: the next word in a whole, you have to type @kbd{n} for @code{nest}. Here is
8902: an overview what keys are available:
1.23 crook 8903:
1.26 crook 8904: @table @i
1.23 crook 8905:
1.30 anton 8906: @item @key{RET}
1.26 crook 8907: Next; Execute the next word.
1.23 crook 8908:
1.26 crook 8909: @item n
8910: Nest; Single step through next word.
1.5 anton 8911:
1.26 crook 8912: @item u
8913: Unnest; Stop debugging and execute rest of word. If we got to this word
8914: with nest, continue debugging with the calling word.
1.5 anton 8915:
1.26 crook 8916: @item d
8917: Done; Stop debugging and execute rest.
1.5 anton 8918:
1.26 crook 8919: @item s
8920: Stop; Abort immediately.
1.5 anton 8921:
1.26 crook 8922: @end table
1.5 anton 8923:
1.26 crook 8924: Debugging large application with this mechanism is very difficult, because
8925: you have to nest very deeply into the program before the interesting part
8926: begins. This takes a lot of time.
1.5 anton 8927:
1.26 crook 8928: To do it more directly put a @code{BREAK:} command into your source code.
8929: When program execution reaches @code{BREAK:} the single step debugger is
8930: invoked and you have all the features described above.
1.23 crook 8931:
1.26 crook 8932: If you have more than one part to debug it is useful to know where the
8933: program has stopped at the moment. You can do this by the
8934: @code{BREAK" string"} command. This behaves like @code{BREAK:} except that
8935: string is typed out when the ``breakpoint'' is reached.
8936:
1.44 crook 8937:
1.26 crook 8938: doc-dbg
1.45 crook 8939: doc-break:
8940: doc-break"
1.26 crook 8941:
8942:
1.44 crook 8943:
1.26 crook 8944: @c -------------------------------------------------------------
8945: @node Assembler and Code Words, Threading Words, Programming Tools, Words
8946: @section Assembler and Code Words
8947: @cindex assembler
8948: @cindex code words
1.5 anton 8949:
1.52 anton 8950: @menu
1.53 anton 8951: * Code and ;code::
8952: * Common Assembler:: Assembler Syntax
1.52 anton 8953: * Common Disassembler::
8954: * 386 Assembler:: Deviations and special cases
8955: * Alpha Assembler:: Deviations and special cases
8956: * MIPS assembler:: Deviations and special cases
1.53 anton 8957: * Other assemblers:: How to write them
1.52 anton 8958: @end menu
8959:
1.53 anton 8960: @node Code and ;code, Common Assembler, Assembler and Code Words, Assembler and Code Words
8961: @subsection @code{Code} and @code{;code}
1.52 anton 8962:
1.26 crook 8963: Gforth provides some words for defining primitives (words written in
1.29 crook 8964: machine code), and for defining the machine-code equivalent of
1.26 crook 8965: @code{DOES>}-based defining words. However, the machine-independent
8966: nature of Gforth poses a few problems: First of all, Gforth runs on
8967: several architectures, so it can provide no standard assembler. What's
8968: worse is that the register allocation not only depends on the processor,
8969: but also on the @code{gcc} version and options used.
1.5 anton 8970:
1.29 crook 8971: The words that Gforth offers encapsulate some system dependences (e.g.,
8972: the header structure), so a system-independent assembler may be used in
1.26 crook 8973: Gforth. If you do not have an assembler, you can compile machine code
1.29 crook 8974: directly with @code{,} and @code{c,}@footnote{This isn't portable,
8975: because these words emit stuff in @i{data} space; it works because
8976: Gforth has unified code/data spaces. Assembler isn't likely to be
8977: portable anyway.}.
1.5 anton 8978:
1.44 crook 8979:
1.26 crook 8980: doc-assembler
1.45 crook 8981: doc-init-asm
1.26 crook 8982: doc-code
8983: doc-end-code
8984: doc-;code
8985: doc-flush-icache
1.5 anton 8986:
1.44 crook 8987:
1.26 crook 8988: If @code{flush-icache} does not work correctly, @code{code} words
8989: etc. will not work (reliably), either.
1.5 anton 8990:
1.29 crook 8991: The typical usage of these @code{code} words can be shown most easily by
8992: analogy to the equivalent high-level defining words:
8993:
8994: @example
1.53 anton 8995: : foo code foo
8996: <high-level Forth words> <assembler>
8997: ; end-code
8998:
8999: : bar : bar
9000: <high-level Forth words> <high-level Forth words>
9001: CREATE CREATE
9002: <high-level Forth words> <high-level Forth words>
9003: DOES> ;code
9004: <high-level Forth words> <assembler>
9005: ; end-code
1.29 crook 9006: @end example
9007:
1.26 crook 9008: @code{flush-icache} is always present. The other words are rarely used
9009: and reside in @code{code.fs}, which is usually not loaded. You can load
9010: it with @code{require code.fs}.
1.5 anton 9011:
1.26 crook 9012: @cindex registers of the inner interpreter
9013: In the assembly code you will want to refer to the inner interpreter's
9014: registers (e.g., the data stack pointer) and you may want to use other
9015: registers for temporary storage. Unfortunately, the register allocation
9016: is installation-dependent.
1.5 anton 9017:
1.26 crook 9018: The easiest solution is to use explicit register declarations
9019: (@pxref{Explicit Reg Vars, , Variables in Specified Registers, gcc.info,
9020: GNU C Manual}) for all of the inner interpreter's registers: You have to
9021: compile Gforth with @code{-DFORCE_REG} (configure option
9022: @code{--enable-force-reg}) and the appropriate declarations must be
9023: present in the @code{machine.h} file (see @code{mips.h} for an example;
9024: you can find a full list of all declarable register symbols with
9025: @code{grep register engine.c}). If you give explicit registers to all
9026: variables that are declared at the beginning of @code{engine()}, you
9027: should be able to use the other caller-saved registers for temporary
9028: storage. Alternatively, you can use the @code{gcc} option
9029: @code{-ffixed-REG} (@pxref{Code Gen Options, , Options for Code
9030: Generation Conventions, gcc.info, GNU C Manual}) to reserve a register
9031: (however, this restriction on register allocation may slow Gforth
9032: significantly).
1.5 anton 9033:
1.26 crook 9034: If this solution is not viable (e.g., because @code{gcc} does not allow
9035: you to explicitly declare all the registers you need), you have to find
9036: out by looking at the code where the inner interpreter's registers
9037: reside and which registers can be used for temporary storage. You can
9038: get an assembly listing of the engine's code with @code{make engine.s}.
1.5 anton 9039:
1.26 crook 9040: In any case, it is good practice to abstract your assembly code from the
9041: actual register allocation. E.g., if the data stack pointer resides in
9042: register @code{$17}, create an alias for this register called @code{sp},
9043: and use that in your assembly code.
1.5 anton 9044:
1.26 crook 9045: @cindex code words, portable
9046: Another option for implementing normal and defining words efficiently
9047: is to add the desired functionality to the source of Gforth. For normal
9048: words you just have to edit @file{primitives} (@pxref{Automatic
9049: Generation}). Defining words (equivalent to @code{;CODE} words, for fast
9050: defined words) may require changes in @file{engine.c}, @file{kernel.fs},
9051: @file{prims2x.fs}, and possibly @file{cross.fs}.
1.5 anton 9052:
1.53 anton 9053: @node Common Assembler, Common Disassembler, Code and ;code, Assembler and Code Words
9054: @subsection Common Assembler
9055:
9056: The assemblers in Gforth generally use a postfix syntax, i.e., the
9057: instruction name follows the operands.
9058:
9059: The operands are passed in the usual order (the same that is used in the
9060: manual of the architecture). Since they all are Forth words, they have
9061: to be separated by spaces; you can also use Forth words to compute the
9062: operands.
9063:
9064: The instruction names usually end with a @code{,}. This makes it easier
9065: to visually separate instructions if you put several of them on one
9066: line; it also avoids shadowing other Forth words (e.g., @code{and}).
9067:
1.55 anton 9068: Registers are usually specified by number; e.g., (decimal) @code{11}
9069: specifies registers R11 and F11 on the Alpha architecture (which one,
9070: depends on the instruction). The usual names are also available, e.g.,
9071: @code{s2} for R11 on Alpha.
9072:
1.53 anton 9073: Control flow is specified similar to normal Forth code (@pxref{Arbitrary
9074: control structures}), with @code{if,}, @code{ahead,}, @code{then,},
9075: @code{begin,}, @code{until,}, @code{again,}, @code{cs-roll},
9076: @code{cs-pick}, @code{else,}, @code{while,}, and @code{repeat,}. The
9077: conditions are specified in a way specific to each assembler.
9078:
1.57 anton 9079: Note that the register assignments of the Gforth engine can change
9080: between Gforth versions, or even between different compilations of the
9081: same Gforth version (e.g., if you use a different GCC version). So if
9082: you want to refer to Gforth's registers (e.g., the stack pointer or
9083: TOS), I recommend defining your own words for refering to these
9084: registers, and using them later on; then you can easily adapt to a
9085: changed register assignment. The stability of the register assignment
9086: is usually better if you build Gforth with @code{--enable-force-reg}.
9087:
9088: In particular, the resturn stack pointer and the instruction pointer are
9089: in memory in @code{gforth}, and usually in registers in
9090: @code{gforth-fast}. The most common use of these registers is to
9091: dispatch to the next word (the @code{next} routine). A portable way to
9092: do this is to jump to @code{' noop >code-address} (of course, this is
9093: less efficient than integrating the @code{next} code and scheduling it
9094: well).
9095:
1.52 anton 9096: @node Common Disassembler, 386 Assembler, Common Assembler, Assembler and Code Words
9097: @subsection Common Disassembler
9098:
9099: You can disassemble a @code{code} word with @code{see}
9100: (@pxref{Debugging}). You can disassemble a section of memory with
9101:
9102: doc-disasm
9103:
9104: The disassembler generally produces output that can be fed into the
9105: assembler (i.e., same syntax, etc.). It also includes additional
1.53 anton 9106: information in comments. In particular, the address of the instruction
9107: is given in a comment before the instruction.
9108:
9109: @code{See} may display more or less than the actual code of the word,
9110: because the recognition of the end of the code is unreliable. You can
9111: use @code{disasm} if it did not display enough. It may display more, if
9112: the code word is not immediately followed by a named word. If you have
9113: something else there, you can follow the word with @code{align last @ ,}
9114: to ensure that the end is recognized.
1.52 anton 9115:
9116: @node 386 Assembler, Alpha Assembler, Common Disassembler, Assembler and Code Words
9117: @subsection 386 Assembler
9118:
1.64 pazsan 9119: The 386 assembler included in Gforth was written by Bernd Paysan, it's
9120: available under GPL, and originally part of bigFORTH.
9121:
9122: The 386 disassembler included in Gforth was written by Andrew McKewan
9123: and is in the public domain.
1.57 anton 9124:
9125: The disassembler displays code in prefix Intel syntax.
9126:
1.64 pazsan 9127: The assembler uses a postfix syntax with reversed parameters.
9128:
9129: The assembler includes all instruction of the Athlon, i.e. 486 core
9130: instructions, Pentium and PPro extensions, floating point, MMX, 3Dnow!,
9131: but not ISSE. It's an integrated 16- and 32-bit assembler. Default is 32
9132: bit, you can switch to 16 bit with .86 and back to 32 bit with .386.
9133:
9134: There are several prefixes to switch between different operation sizes,
9135: @code{.b} for byte accesses, @code{.w} for word accesses, @code{.d} for
9136: double-word accesses. Addressing modes can be switched with @code{.wa}
9137: for 16 bit addresses, and @code{.da} for 32 bit addresses. You don't
9138: need a prefix for byte register names (@code{AL} et al).
9139:
9140: For floating point operations, the prefixes are @code{.fs} (IEEE
9141: single), @code{.fl} (IEEE double), @code{.fx} (extended), @code{.fw}
9142: (word), @code{.fd} (double-word), and @code{.fq} (quad-word).
9143:
9144: The MMX opcodes don't have size prefixes, they are spelled out like in
9145: the Intel assembler. Instead of move from and to memory, there are
9146: PLDQ/PLDD and PSTQ/PSTD.
9147:
9148: The registers lack the 'e' prefix; even in 32 bit mode, eax is called
9149: ax. Immediate values are indicated by postfixing them with @code{#},
9150: e.g., @code{3 #}. Here are some examples of addressing modes:
1.57 anton 9151:
9152: @example
1.65 anton 9153: 3 # \ immediate
9154: ax \ register
9155: 100 di d) \ 100[edi]
9156: 4 bx cx di) \ 4[ebx][ecx]
9157: di ax *4 i) \ [edi][eax*4]
9158: 20 ax *4 i#) \ 20[eax*4]
1.57 anton 9159: @end example
9160:
9161: Some example of instructions are:
9162:
9163: @example
1.64 pazsan 9164: ax bx mov \ move ebx,eax
9165: 3 # ax mov \ mov eax,3
9166: 100 di ) ax mov \ mov eax,100[edi]
9167: 4 bx cx di) ax mov \ mov eax,4[ebx][ecx]
9168: .w ax bx mov \ mov bx,ax
1.57 anton 9169: @end example
9170:
1.64 pazsan 9171: The following forms are supported for binary instructions:
1.57 anton 9172:
9173: @example
9174: <reg> <reg> <inst>
9175: <n> # <reg> <inst>
9176: <mem> <reg> <inst>
9177: <reg> <mem> <inst>
9178: @end example
9179:
9180: Immediate to memory is not supported. The shift/rotate syntax is:
9181:
9182: @example
1.64 pazsan 9183: <reg/mem> 1 # shl \ shortens to shift without immediate
9184: <reg/mem> 4 # shl
9185: <reg/mem> cl shl
1.57 anton 9186: @end example
9187:
1.64 pazsan 9188: Precede string instructions (@code{movs} etc.) with @code{.b} to get
1.57 anton 9189: the byte version.
9190:
1.65 anton 9191: The control structure words @code{IF} @code{UNTIL} etc. must be preceded
9192: by one of these conditions: @code{vs vc u< u>= 0= 0<> u<= u> 0< 0>= ps
9193: pc < >= <= >}. (Note that most of these words shadow some Forth words
9194: when @code{assembler} is in front of @code{forth} in the search path,
9195: e.g., in @code{code} words). Currently the control structure words use
9196: one stack item, so you have to use @code{roll} instead of @code{cs-roll}
9197: to shuffle them (you can also use @code{swap} etc.).
9198:
9199: Here is an example of a @code{code} word (assumes that the stack pointer
9200: is in esi and the TOS is in ebx):
9201:
9202: @example
9203: code my+ ( n1 n2 -- n )
9204: 4 si D) bx add
9205: 4 # si add
9206: Next
9207: end-code
9208: @end example
1.52 anton 9209:
9210: @node Alpha Assembler, MIPS assembler, 386 Assembler, Assembler and Code Words
9211: @subsection Alpha Assembler
9212:
1.55 anton 9213: The Alpha assembler and disassembler were originally written by Bernd
9214: Thallner.
9215:
9216: The register names @code{a0}--@code{a5} are not available to avoid
9217: shadowing hex numbers.
9218:
9219: Immediate forms of arithmetic instructions are distinguished by a
9220: @code{#} just before the @code{,}, e.g., @code{and#,} (note: @code{lda,}
9221: does not count as arithmetic instruction).
9222:
9223: You have to specify all operands to an instruction, even those that
9224: other assemblers consider optional, e.g., the destination register for
9225: @code{br,}, or the destination register and hint for @code{jmp,}.
9226:
9227: You can specify conditions for @code{if,} by removing the first @code{b}
9228: and the trailing @code{,} from a branch with a corresponding name; e.g.,
9229:
9230: @example
9231: 11 fgt if, \ if F11>0e
9232: ...
9233: endif,
1.56 anton 9234: @end example
1.55 anton 9235:
9236: @code{fbgt,} gives @code{fgt}.
1.52 anton 9237:
1.53 anton 9238: @node MIPS assembler, Other assemblers, Alpha Assembler, Assembler and Code Words
1.52 anton 9239: @subsection MIPS assembler
9240:
9241: The MIPS assembler was originally written by Christian Pirker.
9242:
9243: Currently the assembler and disassembler only cover the MIPS-I
9244: architecture (R3000), and don't support FP instructions.
9245:
1.55 anton 9246: The register names @code{$a0}--@code{$a3} are not available to avoid
9247: shadowing hex numbers.
1.52 anton 9248:
9249: Because there is no way to distinguish registers from immediate values,
9250: you have to explicitly use the immediate forms of instructions, i.e.,
9251: @code{addiu,}, not just @code{addu,} (@command{as} does this
9252: implicitly).
9253:
9254: If the architecture manual specifies several formats for the instruction
9255: (e.g., for @code{jalr,}), you usually have to use the one with more
9256: arguments (i.e., two for @code{jalr,}). When in doubt, see
9257: @code{arch/mips/testasm.fs} for an example of correct use.
9258:
1.53 anton 9259: Branches and jumps in the MIPS architecture have a delay slot. You have
9260: to fill it yourself (the simplest way is to use @code{nop,}), the
9261: assembler does not do it for you (unlike @command{as}). Even
9262: @code{if,}, @code{ahead,}, @code{until,}, @code{again,}, @code{while,},
9263: @code{else,} and @code{repeat,} need a delay slot. Since @code{begin,}
9264: and @code{then,} just specify branch targets, they are not affected.
9265:
9266: Note that you must not put branches, jumps, or @code{li,} into the delay
9267: slot: @code{li,} may expand to several instructions, and control flow
9268: instructions may not be put into the branch delay slot in any case.
1.52 anton 9269:
9270: For branches the argument specifying the target is a relative address;
9271: You have to add the address of the delay slot to get the absolute
9272: address.
1.53 anton 9273:
9274: The MIPS architecture also has load delay slots and restrictions on
9275: using @code{mfhi,} and @code{mflo,}; you have to order the instructions
9276: yourself to satisfy these restrictions, the assembler does not do it for
9277: you.
9278:
9279: You can specify the conditions for @code{if,} etc. by taking a
9280: conditional branch and leaving away the @code{b} at the start and the
9281: @code{,} at the end. E.g.,
9282:
9283: @example
9284: 4 5 eq if,
9285: ... \ do something if $4 equals $5
9286: then,
9287: @end example
9288:
9289: @node Other assemblers, , MIPS assembler, Assembler and Code Words
9290: @subsection Other assemblers
9291:
9292: If you want to contribute another assembler/disassembler, please contact
9293: us (@email{bug-gforth@@gnu.org}) to check if we have such an assembler
9294: already. If you are writing them from scratch, please use a similar
9295: syntax style as the one we use (i.e., postfix, commas at the end of the
9296: instruction names, @pxref{Common Assembler}); make the output of the
9297: disassembler be valid input for the assembler, and keep the style
9298: similar to the style we used.
9299:
9300: Hints on implementation: The most important part is to have a good test
9301: suite that contains all instructions. Once you have that, the rest is
9302: easy. For actual coding you can take a look at
9303: @file{arch/mips/disasm.fs} to get some ideas on how to use data for both
9304: the assembler and disassembler, avoiding redundancy and some potential
1.63 anton 9305: bugs. You can also look at that file (and @pxref{Advanced does> usage
9306: example}) to get ideas how to factor a disassembler.
1.5 anton 9307:
1.54 anton 9308: Start with the disassembler, because it's easier to reuse data from the
9309: disassembler for the assembler than the other way round.
9310:
9311: For the assembler, take a look at @file{arch/alpha/asm.fs}, which shows
9312: how simple it can be.
9313:
1.26 crook 9314: @c -------------------------------------------------------------
9315: @node Threading Words, Locals, Assembler and Code Words, Words
9316: @section Threading Words
9317: @cindex threading words
1.5 anton 9318:
1.26 crook 9319: @cindex code address
9320: These words provide access to code addresses and other threading stuff
9321: in Gforth (and, possibly, other interpretive Forths). It more or less
9322: abstracts away the differences between direct and indirect threading
9323: (and, for direct threading, the machine dependences). However, at
9324: present this wordset is still incomplete. It is also pretty low-level;
9325: some day it will hopefully be made unnecessary by an internals wordset
9326: that abstracts implementation details away completely.
1.5 anton 9327:
1.44 crook 9328:
1.26 crook 9329: doc-threading-method
9330: doc->code-address
9331: doc->does-code
9332: doc-code-address!
9333: doc-does-code!
9334: doc-does-handler!
9335: doc-/does-handler
1.5 anton 9336:
1.44 crook 9337:
1.26 crook 9338: The code addresses produced by various defining words are produced by
9339: the following words:
1.5 anton 9340:
1.44 crook 9341:
1.26 crook 9342: doc-docol:
9343: doc-docon:
9344: doc-dovar:
9345: doc-douser:
9346: doc-dodefer:
9347: doc-dofield:
1.5 anton 9348:
1.44 crook 9349:
1.26 crook 9350: You can recognize words defined by a @code{CREATE}...@code{DOES>} word
9351: with @code{>does-code}. If the word was defined in that way, the value
9352: returned is non-zero and identifies the @code{DOES>} used by the
9353: defining word.
9354: @comment TODO should that be ``identifies the xt of the DOES> ??''
1.5 anton 9355:
1.26 crook 9356: @c -------------------------------------------------------------
9357: @node Locals, Structures, Threading Words, Words
9358: @section Locals
9359: @cindex locals
1.5 anton 9360:
1.26 crook 9361: Local variables can make Forth programming more enjoyable and Forth
9362: programs easier to read. Unfortunately, the locals of ANS Forth are
9363: laden with restrictions. Therefore, we provide not only the ANS Forth
9364: locals wordset, but also our own, more powerful locals wordset (we
9365: implemented the ANS Forth locals wordset through our locals wordset).
1.5 anton 9366:
1.66 anton 9367: The ideas in this section have also been published in M. Anton Ertl,
9368: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl94l.ps.gz,
9369: Automatic Scoping of Local Variables}}, EuroForth '94.
1.5 anton 9370:
1.26 crook 9371: @menu
9372: * Gforth locals::
9373: * ANS Forth locals::
9374: @end menu
1.5 anton 9375:
1.26 crook 9376: @node Gforth locals, ANS Forth locals, Locals, Locals
9377: @subsection Gforth locals
9378: @cindex Gforth locals
9379: @cindex locals, Gforth style
1.5 anton 9380:
1.26 crook 9381: Locals can be defined with
1.5 anton 9382:
9383: @example
1.26 crook 9384: @{ local1 local2 ... -- comment @}
9385: @end example
9386: or
9387: @example
9388: @{ local1 local2 ... @}
1.5 anton 9389: @end example
9390:
1.26 crook 9391: E.g.,
1.5 anton 9392: @example
1.26 crook 9393: : max @{ n1 n2 -- n3 @}
9394: n1 n2 > if
9395: n1
9396: else
9397: n2
9398: endif ;
1.5 anton 9399: @end example
9400:
1.26 crook 9401: The similarity of locals definitions with stack comments is intended. A
9402: locals definition often replaces the stack comment of a word. The order
9403: of the locals corresponds to the order in a stack comment and everything
9404: after the @code{--} is really a comment.
1.5 anton 9405:
1.26 crook 9406: This similarity has one disadvantage: It is too easy to confuse locals
9407: declarations with stack comments, causing bugs and making them hard to
9408: find. However, this problem can be avoided by appropriate coding
9409: conventions: Do not use both notations in the same program. If you do,
9410: they should be distinguished using additional means, e.g. by position.
9411:
9412: @cindex types of locals
9413: @cindex locals types
9414: The name of the local may be preceded by a type specifier, e.g.,
9415: @code{F:} for a floating point value:
9416:
9417: @example
9418: : CX* @{ F: Ar F: Ai F: Br F: Bi -- Cr Ci @}
9419: \ complex multiplication
9420: Ar Br f* Ai Bi f* f-
9421: Ar Bi f* Ai Br f* f+ ;
9422: @end example
9423:
9424: @cindex flavours of locals
9425: @cindex locals flavours
9426: @cindex value-flavoured locals
9427: @cindex variable-flavoured locals
9428: Gforth currently supports cells (@code{W:}, @code{W^}), doubles
9429: (@code{D:}, @code{D^}), floats (@code{F:}, @code{F^}) and characters
9430: (@code{C:}, @code{C^}) in two flavours: a value-flavoured local (defined
9431: with @code{W:}, @code{D:} etc.) produces its value and can be changed
9432: with @code{TO}. A variable-flavoured local (defined with @code{W^} etc.)
9433: produces its address (which becomes invalid when the variable's scope is
9434: left). E.g., the standard word @code{emit} can be defined in terms of
9435: @code{type} like this:
1.5 anton 9436:
9437: @example
1.26 crook 9438: : emit @{ C^ char* -- @}
9439: char* 1 type ;
1.5 anton 9440: @end example
9441:
1.26 crook 9442: @cindex default type of locals
9443: @cindex locals, default type
9444: A local without type specifier is a @code{W:} local. Both flavours of
9445: locals are initialized with values from the data or FP stack.
1.5 anton 9446:
1.26 crook 9447: Currently there is no way to define locals with user-defined data
9448: structures, but we are working on it.
1.5 anton 9449:
1.26 crook 9450: Gforth allows defining locals everywhere in a colon definition. This
9451: poses the following questions:
1.5 anton 9452:
1.26 crook 9453: @menu
9454: * Where are locals visible by name?::
9455: * How long do locals live?::
9456: * Programming Style::
9457: * Implementation::
9458: @end menu
1.5 anton 9459:
1.26 crook 9460: @node Where are locals visible by name?, How long do locals live?, Gforth locals, Gforth locals
9461: @subsubsection Where are locals visible by name?
9462: @cindex locals visibility
9463: @cindex visibility of locals
9464: @cindex scope of locals
1.5 anton 9465:
1.26 crook 9466: Basically, the answer is that locals are visible where you would expect
9467: it in block-structured languages, and sometimes a little longer. If you
9468: want to restrict the scope of a local, enclose its definition in
9469: @code{SCOPE}...@code{ENDSCOPE}.
1.5 anton 9470:
1.44 crook 9471:
1.26 crook 9472: doc-scope
9473: doc-endscope
1.5 anton 9474:
1.44 crook 9475:
1.26 crook 9476: These words behave like control structure words, so you can use them
9477: with @code{CS-PICK} and @code{CS-ROLL} to restrict the scope in
9478: arbitrary ways.
1.5 anton 9479:
1.26 crook 9480: If you want a more exact answer to the visibility question, here's the
9481: basic principle: A local is visible in all places that can only be
9482: reached through the definition of the local@footnote{In compiler
9483: construction terminology, all places dominated by the definition of the
9484: local.}. In other words, it is not visible in places that can be reached
9485: without going through the definition of the local. E.g., locals defined
9486: in @code{IF}...@code{ENDIF} are visible until the @code{ENDIF}, locals
9487: defined in @code{BEGIN}...@code{UNTIL} are visible after the
9488: @code{UNTIL} (until, e.g., a subsequent @code{ENDSCOPE}).
1.5 anton 9489:
1.26 crook 9490: The reasoning behind this solution is: We want to have the locals
9491: visible as long as it is meaningful. The user can always make the
9492: visibility shorter by using explicit scoping. In a place that can
9493: only be reached through the definition of a local, the meaning of a
9494: local name is clear. In other places it is not: How is the local
9495: initialized at the control flow path that does not contain the
9496: definition? Which local is meant, if the same name is defined twice in
9497: two independent control flow paths?
1.5 anton 9498:
1.26 crook 9499: This should be enough detail for nearly all users, so you can skip the
9500: rest of this section. If you really must know all the gory details and
9501: options, read on.
1.5 anton 9502:
1.26 crook 9503: In order to implement this rule, the compiler has to know which places
9504: are unreachable. It knows this automatically after @code{AHEAD},
9505: @code{AGAIN}, @code{EXIT} and @code{LEAVE}; in other cases (e.g., after
9506: most @code{THROW}s), you can use the word @code{UNREACHABLE} to tell the
9507: compiler that the control flow never reaches that place. If
9508: @code{UNREACHABLE} is not used where it could, the only consequence is
9509: that the visibility of some locals is more limited than the rule above
9510: says. If @code{UNREACHABLE} is used where it should not (i.e., if you
9511: lie to the compiler), buggy code will be produced.
1.5 anton 9512:
1.44 crook 9513:
1.26 crook 9514: doc-unreachable
1.5 anton 9515:
1.44 crook 9516:
1.26 crook 9517: Another problem with this rule is that at @code{BEGIN}, the compiler
9518: does not know which locals will be visible on the incoming
9519: back-edge. All problems discussed in the following are due to this
9520: ignorance of the compiler (we discuss the problems using @code{BEGIN}
9521: loops as examples; the discussion also applies to @code{?DO} and other
9522: loops). Perhaps the most insidious example is:
1.5 anton 9523: @example
1.26 crook 9524: AHEAD
9525: BEGIN
9526: x
9527: [ 1 CS-ROLL ] THEN
9528: @{ x @}
9529: ...
9530: UNTIL
9531: @end example
1.5 anton 9532:
1.26 crook 9533: This should be legal according to the visibility rule. The use of
9534: @code{x} can only be reached through the definition; but that appears
9535: textually below the use.
1.5 anton 9536:
1.26 crook 9537: From this example it is clear that the visibility rules cannot be fully
9538: implemented without major headaches. Our implementation treats common
9539: cases as advertised and the exceptions are treated in a safe way: The
9540: compiler makes a reasonable guess about the locals visible after a
9541: @code{BEGIN}; if it is too pessimistic, the
9542: user will get a spurious error about the local not being defined; if the
9543: compiler is too optimistic, it will notice this later and issue a
9544: warning. In the case above the compiler would complain about @code{x}
9545: being undefined at its use. You can see from the obscure examples in
9546: this section that it takes quite unusual control structures to get the
9547: compiler into trouble, and even then it will often do fine.
1.5 anton 9548:
1.26 crook 9549: If the @code{BEGIN} is reachable from above, the most optimistic guess
9550: is that all locals visible before the @code{BEGIN} will also be
9551: visible after the @code{BEGIN}. This guess is valid for all loops that
9552: are entered only through the @code{BEGIN}, in particular, for normal
9553: @code{BEGIN}...@code{WHILE}...@code{REPEAT} and
9554: @code{BEGIN}...@code{UNTIL} loops and it is implemented in our
9555: compiler. When the branch to the @code{BEGIN} is finally generated by
9556: @code{AGAIN} or @code{UNTIL}, the compiler checks the guess and
9557: warns the user if it was too optimistic:
9558: @example
9559: IF
9560: @{ x @}
9561: BEGIN
9562: \ x ?
9563: [ 1 cs-roll ] THEN
9564: ...
9565: UNTIL
1.5 anton 9566: @end example
9567:
1.26 crook 9568: Here, @code{x} lives only until the @code{BEGIN}, but the compiler
9569: optimistically assumes that it lives until the @code{THEN}. It notices
9570: this difference when it compiles the @code{UNTIL} and issues a
9571: warning. The user can avoid the warning, and make sure that @code{x}
9572: is not used in the wrong area by using explicit scoping:
9573: @example
9574: IF
9575: SCOPE
9576: @{ x @}
9577: ENDSCOPE
9578: BEGIN
9579: [ 1 cs-roll ] THEN
9580: ...
9581: UNTIL
9582: @end example
1.5 anton 9583:
1.26 crook 9584: Since the guess is optimistic, there will be no spurious error messages
9585: about undefined locals.
1.5 anton 9586:
1.26 crook 9587: If the @code{BEGIN} is not reachable from above (e.g., after
9588: @code{AHEAD} or @code{EXIT}), the compiler cannot even make an
9589: optimistic guess, as the locals visible after the @code{BEGIN} may be
9590: defined later. Therefore, the compiler assumes that no locals are
9591: visible after the @code{BEGIN}. However, the user can use
9592: @code{ASSUME-LIVE} to make the compiler assume that the same locals are
9593: visible at the BEGIN as at the point where the top control-flow stack
9594: item was created.
1.5 anton 9595:
1.44 crook 9596:
1.26 crook 9597: doc-assume-live
1.5 anton 9598:
1.44 crook 9599:
9600: @noindent
1.26 crook 9601: E.g.,
1.5 anton 9602: @example
1.26 crook 9603: @{ x @}
9604: AHEAD
9605: ASSUME-LIVE
9606: BEGIN
9607: x
9608: [ 1 CS-ROLL ] THEN
9609: ...
9610: UNTIL
1.5 anton 9611: @end example
9612:
1.26 crook 9613: Other cases where the locals are defined before the @code{BEGIN} can be
9614: handled by inserting an appropriate @code{CS-ROLL} before the
9615: @code{ASSUME-LIVE} (and changing the control-flow stack manipulation
9616: behind the @code{ASSUME-LIVE}).
1.5 anton 9617:
1.26 crook 9618: Cases where locals are defined after the @code{BEGIN} (but should be
9619: visible immediately after the @code{BEGIN}) can only be handled by
9620: rearranging the loop. E.g., the ``most insidious'' example above can be
9621: arranged into:
1.5 anton 9622: @example
1.26 crook 9623: BEGIN
9624: @{ x @}
9625: ... 0=
9626: WHILE
9627: x
9628: REPEAT
1.5 anton 9629: @end example
9630:
1.26 crook 9631: @node How long do locals live?, Programming Style, Where are locals visible by name?, Gforth locals
9632: @subsubsection How long do locals live?
9633: @cindex locals lifetime
9634: @cindex lifetime of locals
1.5 anton 9635:
1.26 crook 9636: The right answer for the lifetime question would be: A local lives at
9637: least as long as it can be accessed. For a value-flavoured local this
9638: means: until the end of its visibility. However, a variable-flavoured
9639: local could be accessed through its address far beyond its visibility
9640: scope. Ultimately, this would mean that such locals would have to be
9641: garbage collected. Since this entails un-Forth-like implementation
9642: complexities, I adopted the same cowardly solution as some other
9643: languages (e.g., C): The local lives only as long as it is visible;
9644: afterwards its address is invalid (and programs that access it
9645: afterwards are erroneous).
1.5 anton 9646:
1.26 crook 9647: @node Programming Style, Implementation, How long do locals live?, Gforth locals
9648: @subsubsection Programming Style
9649: @cindex locals programming style
9650: @cindex programming style, locals
1.5 anton 9651:
1.26 crook 9652: The freedom to define locals anywhere has the potential to change
9653: programming styles dramatically. In particular, the need to use the
9654: return stack for intermediate storage vanishes. Moreover, all stack
9655: manipulations (except @code{PICK}s and @code{ROLL}s with run-time
9656: determined arguments) can be eliminated: If the stack items are in the
9657: wrong order, just write a locals definition for all of them; then
9658: write the items in the order you want.
1.5 anton 9659:
1.26 crook 9660: This seems a little far-fetched and eliminating stack manipulations is
9661: unlikely to become a conscious programming objective. Still, the number
9662: of stack manipulations will be reduced dramatically if local variables
1.49 anton 9663: are used liberally (e.g., compare @code{max} (@pxref{Gforth locals}) with
1.26 crook 9664: a traditional implementation of @code{max}).
1.5 anton 9665:
1.26 crook 9666: This shows one potential benefit of locals: making Forth programs more
9667: readable. Of course, this benefit will only be realized if the
9668: programmers continue to honour the principle of factoring instead of
9669: using the added latitude to make the words longer.
1.5 anton 9670:
1.26 crook 9671: @cindex single-assignment style for locals
9672: Using @code{TO} can and should be avoided. Without @code{TO},
9673: every value-flavoured local has only a single assignment and many
9674: advantages of functional languages apply to Forth. I.e., programs are
9675: easier to analyse, to optimize and to read: It is clear from the
9676: definition what the local stands for, it does not turn into something
9677: different later.
1.5 anton 9678:
1.26 crook 9679: E.g., a definition using @code{TO} might look like this:
1.5 anton 9680: @example
1.26 crook 9681: : strcmp @{ addr1 u1 addr2 u2 -- n @}
9682: u1 u2 min 0
9683: ?do
9684: addr1 c@@ addr2 c@@ -
9685: ?dup-if
9686: unloop exit
9687: then
9688: addr1 char+ TO addr1
9689: addr2 char+ TO addr2
9690: loop
9691: u1 u2 - ;
1.5 anton 9692: @end example
1.26 crook 9693: Here, @code{TO} is used to update @code{addr1} and @code{addr2} at
9694: every loop iteration. @code{strcmp} is a typical example of the
9695: readability problems of using @code{TO}. When you start reading
9696: @code{strcmp}, you think that @code{addr1} refers to the start of the
9697: string. Only near the end of the loop you realize that it is something
9698: else.
1.5 anton 9699:
1.26 crook 9700: This can be avoided by defining two locals at the start of the loop that
9701: are initialized with the right value for the current iteration.
1.5 anton 9702: @example
1.26 crook 9703: : strcmp @{ addr1 u1 addr2 u2 -- n @}
9704: addr1 addr2
9705: u1 u2 min 0
9706: ?do @{ s1 s2 @}
9707: s1 c@@ s2 c@@ -
9708: ?dup-if
9709: unloop exit
9710: then
9711: s1 char+ s2 char+
9712: loop
9713: 2drop
9714: u1 u2 - ;
1.5 anton 9715: @end example
1.26 crook 9716: Here it is clear from the start that @code{s1} has a different value
9717: in every loop iteration.
1.5 anton 9718:
1.26 crook 9719: @node Implementation, , Programming Style, Gforth locals
9720: @subsubsection Implementation
9721: @cindex locals implementation
9722: @cindex implementation of locals
1.5 anton 9723:
1.26 crook 9724: @cindex locals stack
9725: Gforth uses an extra locals stack. The most compelling reason for
9726: this is that the return stack is not float-aligned; using an extra stack
9727: also eliminates the problems and restrictions of using the return stack
9728: as locals stack. Like the other stacks, the locals stack grows toward
9729: lower addresses. A few primitives allow an efficient implementation:
1.5 anton 9730:
1.44 crook 9731:
1.26 crook 9732: doc-@local#
9733: doc-f@local#
9734: doc-laddr#
9735: doc-lp+!#
9736: doc-lp!
9737: doc->l
9738: doc-f>l
1.5 anton 9739:
1.44 crook 9740:
1.26 crook 9741: In addition to these primitives, some specializations of these
9742: primitives for commonly occurring inline arguments are provided for
9743: efficiency reasons, e.g., @code{@@local0} as specialization of
9744: @code{@@local#} for the inline argument 0. The following compiling words
9745: compile the right specialized version, or the general version, as
9746: appropriate:
1.6 pazsan 9747:
1.44 crook 9748:
1.26 crook 9749: doc-compile-@local
9750: doc-compile-f@local
9751: doc-compile-lp+!
1.12 anton 9752:
1.44 crook 9753:
1.26 crook 9754: Combinations of conditional branches and @code{lp+!#} like
9755: @code{?branch-lp+!#} (the locals pointer is only changed if the branch
9756: is taken) are provided for efficiency and correctness in loops.
1.6 pazsan 9757:
1.26 crook 9758: A special area in the dictionary space is reserved for keeping the
9759: local variable names. @code{@{} switches the dictionary pointer to this
9760: area and @code{@}} switches it back and generates the locals
9761: initializing code. @code{W:} etc.@ are normal defining words. This
9762: special area is cleared at the start of every colon definition.
1.6 pazsan 9763:
1.26 crook 9764: @cindex word list for defining locals
9765: A special feature of Gforth's dictionary is used to implement the
9766: definition of locals without type specifiers: every word list (aka
9767: vocabulary) has its own methods for searching
9768: etc. (@pxref{Word Lists}). For the present purpose we defined a word list
9769: with a special search method: When it is searched for a word, it
9770: actually creates that word using @code{W:}. @code{@{} changes the search
9771: order to first search the word list containing @code{@}}, @code{W:} etc.,
9772: and then the word list for defining locals without type specifiers.
1.12 anton 9773:
1.26 crook 9774: The lifetime rules support a stack discipline within a colon
9775: definition: The lifetime of a local is either nested with other locals
9776: lifetimes or it does not overlap them.
1.6 pazsan 9777:
1.26 crook 9778: At @code{BEGIN}, @code{IF}, and @code{AHEAD} no code for locals stack
9779: pointer manipulation is generated. Between control structure words
9780: locals definitions can push locals onto the locals stack. @code{AGAIN}
9781: is the simplest of the other three control flow words. It has to
9782: restore the locals stack depth of the corresponding @code{BEGIN}
9783: before branching. The code looks like this:
9784: @format
9785: @code{lp+!#} current-locals-size @minus{} dest-locals-size
9786: @code{branch} <begin>
9787: @end format
1.6 pazsan 9788:
1.26 crook 9789: @code{UNTIL} is a little more complicated: If it branches back, it
9790: must adjust the stack just like @code{AGAIN}. But if it falls through,
9791: the locals stack must not be changed. The compiler generates the
9792: following code:
9793: @format
9794: @code{?branch-lp+!#} <begin> current-locals-size @minus{} dest-locals-size
9795: @end format
9796: The locals stack pointer is only adjusted if the branch is taken.
1.6 pazsan 9797:
1.26 crook 9798: @code{THEN} can produce somewhat inefficient code:
9799: @format
9800: @code{lp+!#} current-locals-size @minus{} orig-locals-size
9801: <orig target>:
9802: @code{lp+!#} orig-locals-size @minus{} new-locals-size
9803: @end format
9804: The second @code{lp+!#} adjusts the locals stack pointer from the
1.29 crook 9805: level at the @i{orig} point to the level after the @code{THEN}. The
1.26 crook 9806: first @code{lp+!#} adjusts the locals stack pointer from the current
9807: level to the level at the orig point, so the complete effect is an
9808: adjustment from the current level to the right level after the
9809: @code{THEN}.
1.6 pazsan 9810:
1.26 crook 9811: @cindex locals information on the control-flow stack
9812: @cindex control-flow stack items, locals information
9813: In a conventional Forth implementation a dest control-flow stack entry
9814: is just the target address and an orig entry is just the address to be
9815: patched. Our locals implementation adds a word list to every orig or dest
9816: item. It is the list of locals visible (or assumed visible) at the point
9817: described by the entry. Our implementation also adds a tag to identify
9818: the kind of entry, in particular to differentiate between live and dead
9819: (reachable and unreachable) orig entries.
1.6 pazsan 9820:
1.26 crook 9821: A few unusual operations have to be performed on locals word lists:
1.6 pazsan 9822:
1.44 crook 9823:
1.26 crook 9824: doc-common-list
9825: doc-sub-list?
9826: doc-list-size
1.6 pazsan 9827:
1.44 crook 9828:
1.26 crook 9829: Several features of our locals word list implementation make these
9830: operations easy to implement: The locals word lists are organised as
9831: linked lists; the tails of these lists are shared, if the lists
9832: contain some of the same locals; and the address of a name is greater
9833: than the address of the names behind it in the list.
1.6 pazsan 9834:
1.26 crook 9835: Another important implementation detail is the variable
9836: @code{dead-code}. It is used by @code{BEGIN} and @code{THEN} to
9837: determine if they can be reached directly or only through the branch
9838: that they resolve. @code{dead-code} is set by @code{UNREACHABLE},
9839: @code{AHEAD}, @code{EXIT} etc., and cleared at the start of a colon
9840: definition, by @code{BEGIN} and usually by @code{THEN}.
1.6 pazsan 9841:
1.26 crook 9842: Counted loops are similar to other loops in most respects, but
9843: @code{LEAVE} requires special attention: It performs basically the same
9844: service as @code{AHEAD}, but it does not create a control-flow stack
9845: entry. Therefore the information has to be stored elsewhere;
9846: traditionally, the information was stored in the target fields of the
9847: branches created by the @code{LEAVE}s, by organizing these fields into a
9848: linked list. Unfortunately, this clever trick does not provide enough
9849: space for storing our extended control flow information. Therefore, we
9850: introduce another stack, the leave stack. It contains the control-flow
9851: stack entries for all unresolved @code{LEAVE}s.
1.6 pazsan 9852:
1.26 crook 9853: Local names are kept until the end of the colon definition, even if
9854: they are no longer visible in any control-flow path. In a few cases
9855: this may lead to increased space needs for the locals name area, but
9856: usually less than reclaiming this space would cost in code size.
1.6 pazsan 9857:
9858:
1.26 crook 9859: @node ANS Forth locals, , Gforth locals, Locals
9860: @subsection ANS Forth locals
9861: @cindex locals, ANS Forth style
1.6 pazsan 9862:
1.26 crook 9863: The ANS Forth locals wordset does not define a syntax for locals, but
9864: words that make it possible to define various syntaxes. One of the
9865: possible syntaxes is a subset of the syntax we used in the Gforth locals
9866: wordset, i.e.:
1.6 pazsan 9867:
9868: @example
1.26 crook 9869: @{ local1 local2 ... -- comment @}
1.6 pazsan 9870: @end example
1.23 crook 9871: @noindent
1.26 crook 9872: or
1.6 pazsan 9873: @example
1.26 crook 9874: @{ local1 local2 ... @}
1.6 pazsan 9875: @end example
9876:
1.26 crook 9877: The order of the locals corresponds to the order in a stack comment. The
9878: restrictions are:
1.6 pazsan 9879:
9880: @itemize @bullet
9881: @item
1.26 crook 9882: Locals can only be cell-sized values (no type specifiers are allowed).
1.6 pazsan 9883: @item
1.26 crook 9884: Locals can be defined only outside control structures.
1.6 pazsan 9885: @item
1.26 crook 9886: Locals can interfere with explicit usage of the return stack. For the
9887: exact (and long) rules, see the standard. If you don't use return stack
9888: accessing words in a definition using locals, you will be all right. The
9889: purpose of this rule is to make locals implementation on the return
9890: stack easier.
1.6 pazsan 9891: @item
1.26 crook 9892: The whole definition must be in one line.
9893: @end itemize
1.6 pazsan 9894:
1.44 crook 9895: Locals defined in this way behave like @code{VALUE}s
1.49 anton 9896: (@pxref{Values}). I.e., they are initialized from the stack. Using their
1.26 crook 9897: name produces their value. Their value can be changed using @code{TO}.
1.6 pazsan 9898:
1.26 crook 9899: Since this syntax is supported by Gforth directly, you need not do
9900: anything to use it. If you want to port a program using this syntax to
9901: another ANS Forth system, use @file{compat/anslocal.fs} to implement the
9902: syntax on the other system.
1.6 pazsan 9903:
1.26 crook 9904: Note that a syntax shown in the standard, section A.13 looks
9905: similar, but is quite different in having the order of locals
9906: reversed. Beware!
1.6 pazsan 9907:
1.26 crook 9908: The ANS Forth locals wordset itself consists of a word:
1.6 pazsan 9909:
1.44 crook 9910:
1.26 crook 9911: doc-(local)
1.6 pazsan 9912:
1.44 crook 9913:
1.26 crook 9914: The ANS Forth locals extension wordset defines a syntax using @code{locals|}, but it is so
9915: awful that we strongly recommend not to use it. We have implemented this
9916: syntax to make porting to Gforth easy, but do not document it here. The
9917: problem with this syntax is that the locals are defined in an order
9918: reversed with respect to the standard stack comment notation, making
9919: programs harder to read, and easier to misread and miswrite. The only
9920: merit of this syntax is that it is easy to implement using the ANS Forth
9921: locals wordset.
1.7 pazsan 9922:
9923:
1.26 crook 9924: @c ----------------------------------------------------------
9925: @node Structures, Object-oriented Forth, Locals, Words
9926: @section Structures
9927: @cindex structures
9928: @cindex records
1.7 pazsan 9929:
1.26 crook 9930: This section presents the structure package that comes with Gforth. A
9931: version of the package implemented in ANS Forth is available in
9932: @file{compat/struct.fs}. This package was inspired by a posting on
9933: comp.lang.forth in 1989 (unfortunately I don't remember, by whom;
9934: possibly John Hayes). A version of this section has been published in
9935: ???. Marcel Hendrix provided helpful comments.
1.7 pazsan 9936:
1.26 crook 9937: @menu
9938: * Why explicit structure support?::
9939: * Structure Usage::
9940: * Structure Naming Convention::
9941: * Structure Implementation::
9942: * Structure Glossary::
9943: @end menu
1.7 pazsan 9944:
1.26 crook 9945: @node Why explicit structure support?, Structure Usage, Structures, Structures
9946: @subsection Why explicit structure support?
1.7 pazsan 9947:
1.26 crook 9948: @cindex address arithmetic for structures
9949: @cindex structures using address arithmetic
9950: If we want to use a structure containing several fields, we could simply
9951: reserve memory for it, and access the fields using address arithmetic
1.32 anton 9952: (@pxref{Address arithmetic}). As an example, consider a structure with
1.26 crook 9953: the following fields
1.7 pazsan 9954:
1.26 crook 9955: @table @code
9956: @item a
9957: is a float
9958: @item b
9959: is a cell
9960: @item c
9961: is a float
9962: @end table
1.7 pazsan 9963:
1.26 crook 9964: Given the (float-aligned) base address of the structure we get the
9965: address of the field
1.13 pazsan 9966:
1.26 crook 9967: @table @code
9968: @item a
9969: without doing anything further.
9970: @item b
9971: with @code{float+}
9972: @item c
9973: with @code{float+ cell+ faligned}
9974: @end table
1.13 pazsan 9975:
1.26 crook 9976: It is easy to see that this can become quite tiring.
1.13 pazsan 9977:
1.26 crook 9978: Moreover, it is not very readable, because seeing a
9979: @code{cell+} tells us neither which kind of structure is
9980: accessed nor what field is accessed; we have to somehow infer the kind
9981: of structure, and then look up in the documentation, which field of
9982: that structure corresponds to that offset.
1.13 pazsan 9983:
1.26 crook 9984: Finally, this kind of address arithmetic also causes maintenance
9985: troubles: If you add or delete a field somewhere in the middle of the
9986: structure, you have to find and change all computations for the fields
9987: afterwards.
1.13 pazsan 9988:
1.26 crook 9989: So, instead of using @code{cell+} and friends directly, how
9990: about storing the offsets in constants:
1.13 pazsan 9991:
9992: @example
1.26 crook 9993: 0 constant a-offset
9994: 0 float+ constant b-offset
9995: 0 float+ cell+ faligned c-offset
1.13 pazsan 9996: @end example
9997:
1.26 crook 9998: Now we can get the address of field @code{x} with @code{x-offset
9999: +}. This is much better in all respects. Of course, you still
10000: have to change all later offset definitions if you add a field. You can
10001: fix this by declaring the offsets in the following way:
1.13 pazsan 10002:
10003: @example
1.26 crook 10004: 0 constant a-offset
10005: a-offset float+ constant b-offset
10006: b-offset cell+ faligned constant c-offset
1.13 pazsan 10007: @end example
10008:
1.26 crook 10009: Since we always use the offsets with @code{+}, we could use a defining
10010: word @code{cfield} that includes the @code{+} in the action of the
10011: defined word:
1.8 pazsan 10012:
10013: @example
1.26 crook 10014: : cfield ( n "name" -- )
10015: create ,
10016: does> ( name execution: addr1 -- addr2 )
10017: @@ + ;
1.13 pazsan 10018:
1.26 crook 10019: 0 cfield a
10020: 0 a float+ cfield b
10021: 0 b cell+ faligned cfield c
1.13 pazsan 10022: @end example
10023:
1.26 crook 10024: Instead of @code{x-offset +}, we now simply write @code{x}.
10025:
10026: The structure field words now can be used quite nicely. However,
10027: their definition is still a bit cumbersome: We have to repeat the
10028: name, the information about size and alignment is distributed before
10029: and after the field definitions etc. The structure package presented
10030: here addresses these problems.
10031:
10032: @node Structure Usage, Structure Naming Convention, Why explicit structure support?, Structures
10033: @subsection Structure Usage
10034: @cindex structure usage
1.13 pazsan 10035:
1.26 crook 10036: @cindex @code{field} usage
10037: @cindex @code{struct} usage
10038: @cindex @code{end-struct} usage
10039: You can define a structure for a (data-less) linked list with:
1.13 pazsan 10040: @example
1.26 crook 10041: struct
10042: cell% field list-next
10043: end-struct list%
1.13 pazsan 10044: @end example
10045:
1.26 crook 10046: With the address of the list node on the stack, you can compute the
10047: address of the field that contains the address of the next node with
10048: @code{list-next}. E.g., you can determine the length of a list
10049: with:
1.13 pazsan 10050:
10051: @example
1.26 crook 10052: : list-length ( list -- n )
10053: \ "list" is a pointer to the first element of a linked list
10054: \ "n" is the length of the list
10055: 0 BEGIN ( list1 n1 )
10056: over
10057: WHILE ( list1 n1 )
10058: 1+ swap list-next @@ swap
10059: REPEAT
10060: nip ;
1.13 pazsan 10061: @end example
10062:
1.26 crook 10063: You can reserve memory for a list node in the dictionary with
10064: @code{list% %allot}, which leaves the address of the list node on the
10065: stack. For the equivalent allocation on the heap you can use @code{list%
10066: %alloc} (or, for an @code{allocate}-like stack effect (i.e., with ior),
10067: use @code{list% %allocate}). You can get the the size of a list
10068: node with @code{list% %size} and its alignment with @code{list%
10069: %alignment}.
1.13 pazsan 10070:
1.26 crook 10071: Note that in ANS Forth the body of a @code{create}d word is
10072: @code{aligned} but not necessarily @code{faligned};
10073: therefore, if you do a:
1.13 pazsan 10074: @example
1.26 crook 10075: create @emph{name} foo% %allot
1.8 pazsan 10076: @end example
10077:
1.26 crook 10078: @noindent
10079: then the memory alloted for @code{foo%} is
10080: guaranteed to start at the body of @code{@emph{name}} only if
10081: @code{foo%} contains only character, cell and double fields.
1.20 pazsan 10082:
1.45 crook 10083: @cindex structures containing structures
1.26 crook 10084: You can include a structure @code{foo%} as a field of
10085: another structure, like this:
1.20 pazsan 10086: @example
1.26 crook 10087: struct
10088: ...
10089: foo% field ...
10090: ...
10091: end-struct ...
1.20 pazsan 10092: @end example
10093:
1.26 crook 10094: @cindex structure extension
10095: @cindex extended records
10096: Instead of starting with an empty structure, you can extend an
10097: existing structure. E.g., a plain linked list without data, as defined
10098: above, is hardly useful; You can extend it to a linked list of integers,
10099: like this:@footnote{This feature is also known as @emph{extended
10100: records}. It is the main innovation in the Oberon language; in other
10101: words, adding this feature to Modula-2 led Wirth to create a new
10102: language, write a new compiler etc. Adding this feature to Forth just
10103: required a few lines of code.}
1.20 pazsan 10104:
10105: @example
1.26 crook 10106: list%
10107: cell% field intlist-int
10108: end-struct intlist%
1.20 pazsan 10109: @end example
10110:
1.26 crook 10111: @code{intlist%} is a structure with two fields:
10112: @code{list-next} and @code{intlist-int}.
1.20 pazsan 10113:
1.26 crook 10114: @cindex structures containing arrays
10115: You can specify an array type containing @emph{n} elements of
10116: type @code{foo%} like this:
1.20 pazsan 10117:
10118: @example
1.26 crook 10119: foo% @emph{n} *
1.20 pazsan 10120: @end example
10121:
1.26 crook 10122: You can use this array type in any place where you can use a normal
10123: type, e.g., when defining a @code{field}, or with
10124: @code{%allot}.
1.20 pazsan 10125:
1.26 crook 10126: @cindex first field optimization
10127: The first field is at the base address of a structure and the word
10128: for this field (e.g., @code{list-next}) actually does not change
10129: the address on the stack. You may be tempted to leave it away in the
10130: interest of run-time and space efficiency. This is not necessary,
10131: because the structure package optimizes this case and compiling such
10132: words does not generate any code. So, in the interest of readability
10133: and maintainability you should include the word for the field when
10134: accessing the field.
1.20 pazsan 10135:
1.26 crook 10136: @node Structure Naming Convention, Structure Implementation, Structure Usage, Structures
10137: @subsection Structure Naming Convention
10138: @cindex structure naming convention
1.20 pazsan 10139:
1.26 crook 10140: The field names that come to (my) mind are often quite generic, and,
10141: if used, would cause frequent name clashes. E.g., many structures
10142: probably contain a @code{counter} field. The structure names
10143: that come to (my) mind are often also the logical choice for the names
10144: of words that create such a structure.
1.20 pazsan 10145:
1.26 crook 10146: Therefore, I have adopted the following naming conventions:
1.20 pazsan 10147:
1.26 crook 10148: @itemize @bullet
10149: @cindex field naming convention
10150: @item
10151: The names of fields are of the form
10152: @code{@emph{struct}-@emph{field}}, where
10153: @code{@emph{struct}} is the basic name of the structure, and
10154: @code{@emph{field}} is the basic name of the field. You can
10155: think of field words as converting the (address of the)
10156: structure into the (address of the) field.
1.20 pazsan 10157:
1.26 crook 10158: @cindex structure naming convention
10159: @item
10160: The names of structures are of the form
10161: @code{@emph{struct}%}, where
10162: @code{@emph{struct}} is the basic name of the structure.
10163: @end itemize
1.20 pazsan 10164:
1.26 crook 10165: This naming convention does not work that well for fields of extended
10166: structures; e.g., the integer list structure has a field
10167: @code{intlist-int}, but has @code{list-next}, not
10168: @code{intlist-next}.
1.20 pazsan 10169:
1.26 crook 10170: @node Structure Implementation, Structure Glossary, Structure Naming Convention, Structures
10171: @subsection Structure Implementation
10172: @cindex structure implementation
10173: @cindex implementation of structures
1.20 pazsan 10174:
1.26 crook 10175: The central idea in the implementation is to pass the data about the
10176: structure being built on the stack, not in some global
10177: variable. Everything else falls into place naturally once this design
10178: decision is made.
1.20 pazsan 10179:
1.26 crook 10180: The type description on the stack is of the form @emph{align
10181: size}. Keeping the size on the top-of-stack makes dealing with arrays
10182: very simple.
1.20 pazsan 10183:
1.26 crook 10184: @code{field} is a defining word that uses @code{Create}
10185: and @code{DOES>}. The body of the field contains the offset
10186: of the field, and the normal @code{DOES>} action is simply:
1.20 pazsan 10187:
10188: @example
1.48 anton 10189: @@ +
1.20 pazsan 10190: @end example
10191:
1.23 crook 10192: @noindent
1.26 crook 10193: i.e., add the offset to the address, giving the stack effect
1.29 crook 10194: @i{addr1 -- addr2} for a field.
1.20 pazsan 10195:
1.26 crook 10196: @cindex first field optimization, implementation
10197: This simple structure is slightly complicated by the optimization
10198: for fields with offset 0, which requires a different
10199: @code{DOES>}-part (because we cannot rely on there being
10200: something on the stack if such a field is invoked during
10201: compilation). Therefore, we put the different @code{DOES>}-parts
10202: in separate words, and decide which one to invoke based on the
10203: offset. For a zero offset, the field is basically a noop; it is
10204: immediate, and therefore no code is generated when it is compiled.
1.20 pazsan 10205:
1.26 crook 10206: @node Structure Glossary, , Structure Implementation, Structures
10207: @subsection Structure Glossary
10208: @cindex structure glossary
1.20 pazsan 10209:
1.44 crook 10210:
1.26 crook 10211: doc-%align
10212: doc-%alignment
10213: doc-%alloc
10214: doc-%allocate
10215: doc-%allot
10216: doc-cell%
10217: doc-char%
10218: doc-dfloat%
10219: doc-double%
10220: doc-end-struct
10221: doc-field
10222: doc-float%
10223: doc-naligned
10224: doc-sfloat%
10225: doc-%size
10226: doc-struct
1.23 crook 10227:
1.44 crook 10228:
1.26 crook 10229: @c -------------------------------------------------------------
10230: @node Object-oriented Forth, Passing Commands to the OS, Structures, Words
10231: @section Object-oriented Forth
1.20 pazsan 10232:
1.26 crook 10233: Gforth comes with three packages for object-oriented programming:
10234: @file{objects.fs}, @file{oof.fs}, and @file{mini-oof.fs}; none of them
10235: is preloaded, so you have to @code{include} them before use. The most
10236: important differences between these packages (and others) are discussed
10237: in @ref{Comparison with other object models}. All packages are written
10238: in ANS Forth and can be used with any other ANS Forth.
1.20 pazsan 10239:
1.26 crook 10240: @menu
1.48 anton 10241: * Why object-oriented programming?::
10242: * Object-Oriented Terminology::
10243: * Objects::
10244: * OOF::
10245: * Mini-OOF::
1.26 crook 10246: * Comparison with other object models::
10247: @end menu
1.20 pazsan 10248:
1.48 anton 10249: @c ----------------------------------------------------------------
10250: @node Why object-oriented programming?, Object-Oriented Terminology, Object-oriented Forth, Object-oriented Forth
10251: @subsection Why object-oriented programming?
1.26 crook 10252: @cindex object-oriented programming motivation
10253: @cindex motivation for object-oriented programming
1.23 crook 10254:
1.26 crook 10255: Often we have to deal with several data structures (@emph{objects}),
10256: that have to be treated similarly in some respects, but differently in
10257: others. Graphical objects are the textbook example: circles, triangles,
10258: dinosaurs, icons, and others, and we may want to add more during program
10259: development. We want to apply some operations to any graphical object,
10260: e.g., @code{draw} for displaying it on the screen. However, @code{draw}
10261: has to do something different for every kind of object.
10262: @comment TODO add some other operations eg perimeter, area
10263: @comment and tie in to concrete examples later..
1.23 crook 10264:
1.26 crook 10265: We could implement @code{draw} as a big @code{CASE}
10266: control structure that executes the appropriate code depending on the
10267: kind of object to be drawn. This would be not be very elegant, and,
10268: moreover, we would have to change @code{draw} every time we add
10269: a new kind of graphical object (say, a spaceship).
1.23 crook 10270:
1.26 crook 10271: What we would rather do is: When defining spaceships, we would tell
10272: the system: ``Here's how you @code{draw} a spaceship; you figure
10273: out the rest''.
1.23 crook 10274:
1.26 crook 10275: This is the problem that all systems solve that (rightfully) call
10276: themselves object-oriented; the object-oriented packages presented here
10277: solve this problem (and not much else).
10278: @comment TODO ?list properties of oo systems.. oo vs o-based?
1.23 crook 10279:
1.48 anton 10280: @c ------------------------------------------------------------------------
1.26 crook 10281: @node Object-Oriented Terminology, Objects, Why object-oriented programming?, Object-oriented Forth
1.48 anton 10282: @subsection Object-Oriented Terminology
1.26 crook 10283: @cindex object-oriented terminology
10284: @cindex terminology for object-oriented programming
1.23 crook 10285:
1.26 crook 10286: This section is mainly for reference, so you don't have to understand
10287: all of it right away. The terminology is mainly Smalltalk-inspired. In
10288: short:
1.23 crook 10289:
1.26 crook 10290: @table @emph
10291: @cindex class
10292: @item class
10293: a data structure definition with some extras.
1.23 crook 10294:
1.26 crook 10295: @cindex object
10296: @item object
10297: an instance of the data structure described by the class definition.
1.23 crook 10298:
1.26 crook 10299: @cindex instance variables
10300: @item instance variables
10301: fields of the data structure.
1.23 crook 10302:
1.26 crook 10303: @cindex selector
10304: @cindex method selector
10305: @cindex virtual function
10306: @item selector
10307: (or @emph{method selector}) a word (e.g.,
10308: @code{draw}) that performs an operation on a variety of data
10309: structures (classes). A selector describes @emph{what} operation to
10310: perform. In C++ terminology: a (pure) virtual function.
1.23 crook 10311:
1.26 crook 10312: @cindex method
10313: @item method
10314: the concrete definition that performs the operation
10315: described by the selector for a specific class. A method specifies
10316: @emph{how} the operation is performed for a specific class.
1.23 crook 10317:
1.26 crook 10318: @cindex selector invocation
10319: @cindex message send
10320: @cindex invoking a selector
10321: @item selector invocation
10322: a call of a selector. One argument of the call (the TOS (top-of-stack))
10323: is used for determining which method is used. In Smalltalk terminology:
10324: a message (consisting of the selector and the other arguments) is sent
10325: to the object.
1.1 anton 10326:
1.26 crook 10327: @cindex receiving object
10328: @item receiving object
10329: the object used for determining the method executed by a selector
10330: invocation. In the @file{objects.fs} model, it is the object that is on
10331: the TOS when the selector is invoked. (@emph{Receiving} comes from
10332: the Smalltalk @emph{message} terminology.)
1.1 anton 10333:
1.26 crook 10334: @cindex child class
10335: @cindex parent class
10336: @cindex inheritance
10337: @item child class
10338: a class that has (@emph{inherits}) all properties (instance variables,
10339: selectors, methods) from a @emph{parent class}. In Smalltalk
10340: terminology: The subclass inherits from the superclass. In C++
10341: terminology: The derived class inherits from the base class.
1.1 anton 10342:
1.26 crook 10343: @end table
1.21 crook 10344:
1.26 crook 10345: @c If you wonder about the message sending terminology, it comes from
10346: @c a time when each object had it's own task and objects communicated via
10347: @c message passing; eventually the Smalltalk developers realized that
10348: @c they can do most things through simple (indirect) calls. They kept the
10349: @c terminology.
1.1 anton 10350:
1.48 anton 10351: @c --------------------------------------------------------------
1.26 crook 10352: @node Objects, OOF, Object-Oriented Terminology, Object-oriented Forth
10353: @subsection The @file{objects.fs} model
10354: @cindex objects
10355: @cindex object-oriented programming
1.1 anton 10356:
1.26 crook 10357: @cindex @file{objects.fs}
10358: @cindex @file{oof.fs}
1.1 anton 10359:
1.37 anton 10360: This section describes the @file{objects.fs} package. This material also
1.66 anton 10361: has been published in M. Anton Ertl,
10362: @cite{@uref{http://www.complang.tuwien.ac.at/forth/objects/objects.html,
10363: Yet Another Forth Objects Package}}, Forth Dimensions 19(2), pages
10364: 37--43.
1.26 crook 10365: @c McKewan's and Zsoter's packages
1.1 anton 10366:
1.26 crook 10367: This section assumes that you have read @ref{Structures}.
1.1 anton 10368:
1.26 crook 10369: The techniques on which this model is based have been used to implement
10370: the parser generator, Gray, and have also been used in Gforth for
10371: implementing the various flavours of word lists (hashed or not,
10372: case-sensitive or not, special-purpose word lists for locals etc.).
1.1 anton 10373:
10374:
1.26 crook 10375: @menu
10376: * Properties of the Objects model::
10377: * Basic Objects Usage::
1.37 anton 10378: * The Objects base class::
1.26 crook 10379: * Creating objects::
10380: * Object-Oriented Programming Style::
10381: * Class Binding::
10382: * Method conveniences::
10383: * Classes and Scoping::
1.37 anton 10384: * Dividing classes::
1.26 crook 10385: * Object Interfaces::
10386: * Objects Implementation::
10387: * Objects Glossary::
10388: @end menu
1.1 anton 10389:
1.26 crook 10390: Marcel Hendrix provided helpful comments on this section. Andras Zsoter
10391: and Bernd Paysan helped me with the related works section.
1.1 anton 10392:
1.26 crook 10393: @node Properties of the Objects model, Basic Objects Usage, Objects, Objects
10394: @subsubsection Properties of the @file{objects.fs} model
10395: @cindex @file{objects.fs} properties
1.1 anton 10396:
1.26 crook 10397: @itemize @bullet
10398: @item
10399: It is straightforward to pass objects on the stack. Passing
10400: selectors on the stack is a little less convenient, but possible.
1.1 anton 10401:
1.26 crook 10402: @item
10403: Objects are just data structures in memory, and are referenced by their
10404: address. You can create words for objects with normal defining words
10405: like @code{constant}. Likewise, there is no difference between instance
10406: variables that contain objects and those that contain other data.
1.1 anton 10407:
1.26 crook 10408: @item
10409: Late binding is efficient and easy to use.
1.21 crook 10410:
1.26 crook 10411: @item
10412: It avoids parsing, and thus avoids problems with state-smartness
10413: and reduced extensibility; for convenience there are a few parsing
10414: words, but they have non-parsing counterparts. There are also a few
10415: defining words that parse. This is hard to avoid, because all standard
10416: defining words parse (except @code{:noname}); however, such
10417: words are not as bad as many other parsing words, because they are not
10418: state-smart.
1.21 crook 10419:
1.26 crook 10420: @item
10421: It does not try to incorporate everything. It does a few things and does
10422: them well (IMO). In particular, this model was not designed to support
10423: information hiding (although it has features that may help); you can use
10424: a separate package for achieving this.
1.21 crook 10425:
1.26 crook 10426: @item
10427: It is layered; you don't have to learn and use all features to use this
1.49 anton 10428: model. Only a few features are necessary (@pxref{Basic Objects Usage},
10429: @pxref{The Objects base class}, @pxref{Creating objects}.), the others
1.26 crook 10430: are optional and independent of each other.
1.21 crook 10431:
1.26 crook 10432: @item
10433: An implementation in ANS Forth is available.
1.21 crook 10434:
1.26 crook 10435: @end itemize
1.21 crook 10436:
10437:
1.26 crook 10438: @node Basic Objects Usage, The Objects base class, Properties of the Objects model, Objects
10439: @subsubsection Basic @file{objects.fs} Usage
10440: @cindex basic objects usage
10441: @cindex objects, basic usage
1.21 crook 10442:
1.26 crook 10443: You can define a class for graphical objects like this:
1.21 crook 10444:
1.26 crook 10445: @cindex @code{class} usage
10446: @cindex @code{end-class} usage
10447: @cindex @code{selector} usage
10448: @example
10449: object class \ "object" is the parent class
10450: selector draw ( x y graphical -- )
10451: end-class graphical
10452: @end example
1.21 crook 10453:
1.26 crook 10454: This code defines a class @code{graphical} with an
10455: operation @code{draw}. We can perform the operation
10456: @code{draw} on any @code{graphical} object, e.g.:
1.21 crook 10457:
1.26 crook 10458: @example
10459: 100 100 t-rex draw
10460: @end example
1.21 crook 10461:
1.26 crook 10462: @noindent
10463: where @code{t-rex} is a word (say, a constant) that produces a
10464: graphical object.
1.21 crook 10465:
1.29 crook 10466: @comment TODO add a 2nd operation eg perimeter.. and use for
1.26 crook 10467: @comment a concrete example
1.21 crook 10468:
1.26 crook 10469: @cindex abstract class
10470: How do we create a graphical object? With the present definitions,
10471: we cannot create a useful graphical object. The class
10472: @code{graphical} describes graphical objects in general, but not
10473: any concrete graphical object type (C++ users would call it an
10474: @emph{abstract class}); e.g., there is no method for the selector
10475: @code{draw} in the class @code{graphical}.
1.21 crook 10476:
1.26 crook 10477: For concrete graphical objects, we define child classes of the
10478: class @code{graphical}, e.g.:
1.21 crook 10479:
1.26 crook 10480: @cindex @code{overrides} usage
10481: @cindex @code{field} usage in class definition
10482: @example
10483: graphical class \ "graphical" is the parent class
10484: cell% field circle-radius
1.21 crook 10485:
1.26 crook 10486: :noname ( x y circle -- )
10487: circle-radius @@ draw-circle ;
10488: overrides draw
1.21 crook 10489:
1.26 crook 10490: :noname ( n-radius circle -- )
10491: circle-radius ! ;
10492: overrides construct
1.21 crook 10493:
1.26 crook 10494: end-class circle
1.21 crook 10495: @end example
10496:
1.26 crook 10497: Here we define a class @code{circle} as a child of @code{graphical},
10498: with field @code{circle-radius} (which behaves just like a field
10499: (@pxref{Structures}); it defines (using @code{overrides}) new methods
10500: for the selectors @code{draw} and @code{construct} (@code{construct} is
10501: defined in @code{object}, the parent class of @code{graphical}).
1.21 crook 10502:
1.26 crook 10503: Now we can create a circle on the heap (i.e.,
10504: @code{allocate}d memory) with:
1.21 crook 10505:
1.26 crook 10506: @cindex @code{heap-new} usage
1.21 crook 10507: @example
1.26 crook 10508: 50 circle heap-new constant my-circle
10509: @end example
1.21 crook 10510:
1.26 crook 10511: @noindent
10512: @code{heap-new} invokes @code{construct}, thus
10513: initializing the field @code{circle-radius} with 50. We can draw
10514: this new circle at (100,100) with:
1.21 crook 10515:
1.26 crook 10516: @example
10517: 100 100 my-circle draw
1.21 crook 10518: @end example
10519:
1.26 crook 10520: @cindex selector invocation, restrictions
10521: @cindex class definition, restrictions
10522: Note: You can only invoke a selector if the object on the TOS
10523: (the receiving object) belongs to the class where the selector was
10524: defined or one of its descendents; e.g., you can invoke
10525: @code{draw} only for objects belonging to @code{graphical}
10526: or its descendents (e.g., @code{circle}). Immediately before
10527: @code{end-class}, the search order has to be the same as
10528: immediately after @code{class}.
1.21 crook 10529:
1.26 crook 10530: @node The Objects base class, Creating objects, Basic Objects Usage, Objects
10531: @subsubsection The @file{object.fs} base class
10532: @cindex @code{object} class
1.21 crook 10533:
1.26 crook 10534: When you define a class, you have to specify a parent class. So how do
10535: you start defining classes? There is one class available from the start:
10536: @code{object}. It is ancestor for all classes and so is the
10537: only class that has no parent. It has two selectors: @code{construct}
10538: and @code{print}.
1.21 crook 10539:
1.26 crook 10540: @node Creating objects, Object-Oriented Programming Style, The Objects base class, Objects
10541: @subsubsection Creating objects
10542: @cindex creating objects
10543: @cindex object creation
10544: @cindex object allocation options
1.21 crook 10545:
1.26 crook 10546: @cindex @code{heap-new} discussion
10547: @cindex @code{dict-new} discussion
10548: @cindex @code{construct} discussion
10549: You can create and initialize an object of a class on the heap with
10550: @code{heap-new} ( ... class -- object ) and in the dictionary
10551: (allocation with @code{allot}) with @code{dict-new} (
10552: ... class -- object ). Both words invoke @code{construct}, which
10553: consumes the stack items indicated by "..." above.
1.21 crook 10554:
1.26 crook 10555: @cindex @code{init-object} discussion
10556: @cindex @code{class-inst-size} discussion
10557: If you want to allocate memory for an object yourself, you can get its
10558: alignment and size with @code{class-inst-size 2@@} ( class --
10559: align size ). Once you have memory for an object, you can initialize
10560: it with @code{init-object} ( ... class object -- );
10561: @code{construct} does only a part of the necessary work.
1.21 crook 10562:
1.26 crook 10563: @node Object-Oriented Programming Style, Class Binding, Creating objects, Objects
10564: @subsubsection Object-Oriented Programming Style
10565: @cindex object-oriented programming style
1.47 crook 10566: @cindex programming style, object-oriented
1.21 crook 10567:
1.26 crook 10568: This section is not exhaustive.
1.1 anton 10569:
1.26 crook 10570: @cindex stack effects of selectors
10571: @cindex selectors and stack effects
10572: In general, it is a good idea to ensure that all methods for the
10573: same selector have the same stack effect: when you invoke a selector,
10574: you often have no idea which method will be invoked, so, unless all
10575: methods have the same stack effect, you will not know the stack effect
10576: of the selector invocation.
1.21 crook 10577:
1.26 crook 10578: One exception to this rule is methods for the selector
10579: @code{construct}. We know which method is invoked, because we
10580: specify the class to be constructed at the same place. Actually, I
10581: defined @code{construct} as a selector only to give the users a
10582: convenient way to specify initialization. The way it is used, a
10583: mechanism different from selector invocation would be more natural
10584: (but probably would take more code and more space to explain).
1.21 crook 10585:
1.26 crook 10586: @node Class Binding, Method conveniences, Object-Oriented Programming Style, Objects
10587: @subsubsection Class Binding
10588: @cindex class binding
10589: @cindex early binding
1.21 crook 10590:
1.26 crook 10591: @cindex late binding
10592: Normal selector invocations determine the method at run-time depending
10593: on the class of the receiving object. This run-time selection is called
1.29 crook 10594: @i{late binding}.
1.21 crook 10595:
1.26 crook 10596: Sometimes it's preferable to invoke a different method. For example,
10597: you might want to use the simple method for @code{print}ing
10598: @code{object}s instead of the possibly long-winded @code{print} method
10599: of the receiver class. You can achieve this by replacing the invocation
10600: of @code{print} with:
1.21 crook 10601:
1.26 crook 10602: @cindex @code{[bind]} usage
10603: @example
10604: [bind] object print
1.21 crook 10605: @end example
10606:
1.26 crook 10607: @noindent
10608: in compiled code or:
1.21 crook 10609:
1.26 crook 10610: @cindex @code{bind} usage
1.21 crook 10611: @example
1.26 crook 10612: bind object print
1.21 crook 10613: @end example
10614:
1.26 crook 10615: @cindex class binding, alternative to
10616: @noindent
10617: in interpreted code. Alternatively, you can define the method with a
10618: name (e.g., @code{print-object}), and then invoke it through the
10619: name. Class binding is just a (often more convenient) way to achieve
10620: the same effect; it avoids name clutter and allows you to invoke
10621: methods directly without naming them first.
10622:
10623: @cindex superclass binding
10624: @cindex parent class binding
10625: A frequent use of class binding is this: When we define a method
10626: for a selector, we often want the method to do what the selector does
10627: in the parent class, and a little more. There is a special word for
10628: this purpose: @code{[parent]}; @code{[parent]
10629: @emph{selector}} is equivalent to @code{[bind] @emph{parent
10630: selector}}, where @code{@emph{parent}} is the parent
10631: class of the current class. E.g., a method definition might look like:
1.21 crook 10632:
1.26 crook 10633: @cindex @code{[parent]} usage
1.21 crook 10634: @example
1.26 crook 10635: :noname
10636: dup [parent] foo \ do parent's foo on the receiving object
10637: ... \ do some more
10638: ; overrides foo
1.21 crook 10639: @end example
10640:
1.26 crook 10641: @cindex class binding as optimization
10642: In @cite{Object-oriented programming in ANS Forth} (Forth Dimensions,
10643: March 1997), Andrew McKewan presents class binding as an optimization
10644: technique. I recommend not using it for this purpose unless you are in
10645: an emergency. Late binding is pretty fast with this model anyway, so the
10646: benefit of using class binding is small; the cost of using class binding
10647: where it is not appropriate is reduced maintainability.
1.21 crook 10648:
1.26 crook 10649: While we are at programming style questions: You should bind
10650: selectors only to ancestor classes of the receiving object. E.g., say,
10651: you know that the receiving object is of class @code{foo} or its
10652: descendents; then you should bind only to @code{foo} and its
10653: ancestors.
1.21 crook 10654:
1.26 crook 10655: @node Method conveniences, Classes and Scoping, Class Binding, Objects
10656: @subsubsection Method conveniences
10657: @cindex method conveniences
1.1 anton 10658:
1.26 crook 10659: In a method you usually access the receiving object pretty often. If
10660: you define the method as a plain colon definition (e.g., with
10661: @code{:noname}), you may have to do a lot of stack
10662: gymnastics. To avoid this, you can define the method with @code{m:
10663: ... ;m}. E.g., you could define the method for
10664: @code{draw}ing a @code{circle} with
1.20 pazsan 10665:
1.26 crook 10666: @cindex @code{this} usage
10667: @cindex @code{m:} usage
10668: @cindex @code{;m} usage
10669: @example
10670: m: ( x y circle -- )
10671: ( x y ) this circle-radius @@ draw-circle ;m
10672: @end example
1.20 pazsan 10673:
1.26 crook 10674: @cindex @code{exit} in @code{m: ... ;m}
10675: @cindex @code{exitm} discussion
10676: @cindex @code{catch} in @code{m: ... ;m}
10677: When this method is executed, the receiver object is removed from the
10678: stack; you can access it with @code{this} (admittedly, in this
10679: example the use of @code{m: ... ;m} offers no advantage). Note
10680: that I specify the stack effect for the whole method (i.e. including
10681: the receiver object), not just for the code between @code{m:}
10682: and @code{;m}. You cannot use @code{exit} in
10683: @code{m:...;m}; instead, use
10684: @code{exitm}.@footnote{Moreover, for any word that calls
10685: @code{catch} and was defined before loading
10686: @code{objects.fs}, you have to redefine it like I redefined
10687: @code{catch}: @code{: catch this >r catch r> to-this ;}}
1.20 pazsan 10688:
1.26 crook 10689: @cindex @code{inst-var} usage
10690: You will frequently use sequences of the form @code{this
10691: @emph{field}} (in the example above: @code{this
10692: circle-radius}). If you use the field only in this way, you can
10693: define it with @code{inst-var} and eliminate the
10694: @code{this} before the field name. E.g., the @code{circle}
10695: class above could also be defined with:
1.20 pazsan 10696:
1.26 crook 10697: @example
10698: graphical class
10699: cell% inst-var radius
1.20 pazsan 10700:
1.26 crook 10701: m: ( x y circle -- )
10702: radius @@ draw-circle ;m
10703: overrides draw
1.20 pazsan 10704:
1.26 crook 10705: m: ( n-radius circle -- )
10706: radius ! ;m
10707: overrides construct
1.12 anton 10708:
1.26 crook 10709: end-class circle
10710: @end example
1.12 anton 10711:
1.26 crook 10712: @code{radius} can only be used in @code{circle} and its
10713: descendent classes and inside @code{m:...;m}.
1.12 anton 10714:
1.26 crook 10715: @cindex @code{inst-value} usage
10716: You can also define fields with @code{inst-value}, which is
10717: to @code{inst-var} what @code{value} is to
10718: @code{variable}. You can change the value of such a field with
10719: @code{[to-inst]}. E.g., we could also define the class
10720: @code{circle} like this:
1.12 anton 10721:
1.26 crook 10722: @example
10723: graphical class
10724: inst-value radius
1.12 anton 10725:
1.26 crook 10726: m: ( x y circle -- )
10727: radius draw-circle ;m
10728: overrides draw
1.12 anton 10729:
1.26 crook 10730: m: ( n-radius circle -- )
10731: [to-inst] radius ;m
10732: overrides construct
1.21 crook 10733:
1.26 crook 10734: end-class circle
1.12 anton 10735: @end example
10736:
1.38 anton 10737: Finally, you can define named methods with @code{:m}. One use of this
10738: feature is the definition of words that occur only in one class and are
10739: not intended to be overridden, but which still need method context
10740: (e.g., for accessing @code{inst-var}s). Another use is for methods that
10741: would be bound frequently, if defined anonymously.
10742:
1.12 anton 10743:
1.37 anton 10744: @node Classes and Scoping, Dividing classes, Method conveniences, Objects
1.26 crook 10745: @subsubsection Classes and Scoping
10746: @cindex classes and scoping
10747: @cindex scoping and classes
1.12 anton 10748:
1.26 crook 10749: Inheritance is frequent, unlike structure extension. This exacerbates
10750: the problem with the field name convention (@pxref{Structure Naming
10751: Convention}): One always has to remember in which class the field was
10752: originally defined; changing a part of the class structure would require
10753: changes for renaming in otherwise unaffected code.
1.12 anton 10754:
1.26 crook 10755: @cindex @code{inst-var} visibility
10756: @cindex @code{inst-value} visibility
10757: To solve this problem, I added a scoping mechanism (which was not in my
10758: original charter): A field defined with @code{inst-var} (or
10759: @code{inst-value}) is visible only in the class where it is defined and in
10760: the descendent classes of this class. Using such fields only makes
10761: sense in @code{m:}-defined methods in these classes anyway.
1.12 anton 10762:
1.26 crook 10763: This scoping mechanism allows us to use the unadorned field name,
10764: because name clashes with unrelated words become much less likely.
1.12 anton 10765:
1.26 crook 10766: @cindex @code{protected} discussion
10767: @cindex @code{private} discussion
10768: Once we have this mechanism, we can also use it for controlling the
10769: visibility of other words: All words defined after
10770: @code{protected} are visible only in the current class and its
10771: descendents. @code{public} restores the compilation
10772: (i.e. @code{current}) word list that was in effect before. If you
10773: have several @code{protected}s without an intervening
10774: @code{public} or @code{set-current}, @code{public}
10775: will restore the compilation word list in effect before the first of
10776: these @code{protected}s.
1.12 anton 10777:
1.37 anton 10778: @node Dividing classes, Object Interfaces, Classes and Scoping, Objects
10779: @subsubsection Dividing classes
10780: @cindex Dividing classes
10781: @cindex @code{methods}...@code{end-methods}
10782:
10783: You may want to do the definition of methods separate from the
10784: definition of the class, its selectors, fields, and instance variables,
10785: i.e., separate the implementation from the definition. You can do this
10786: in the following way:
10787:
10788: @example
10789: graphical class
10790: inst-value radius
10791: end-class circle
10792:
10793: ... \ do some other stuff
10794:
10795: circle methods \ now we are ready
10796:
10797: m: ( x y circle -- )
10798: radius draw-circle ;m
10799: overrides draw
10800:
10801: m: ( n-radius circle -- )
10802: [to-inst] radius ;m
10803: overrides construct
10804:
10805: end-methods
10806: @end example
10807:
10808: You can use several @code{methods}...@code{end-methods} sections. The
10809: only things you can do to the class in these sections are: defining
10810: methods, and overriding the class's selectors. You must not define new
10811: selectors or fields.
10812:
10813: Note that you often have to override a selector before using it. In
10814: particular, you usually have to override @code{construct} with a new
10815: method before you can invoke @code{heap-new} and friends. E.g., you
10816: must not create a circle before the @code{overrides construct} sequence
10817: in the example above.
10818:
10819: @node Object Interfaces, Objects Implementation, Dividing classes, Objects
1.26 crook 10820: @subsubsection Object Interfaces
10821: @cindex object interfaces
10822: @cindex interfaces for objects
1.12 anton 10823:
1.26 crook 10824: In this model you can only call selectors defined in the class of the
10825: receiving objects or in one of its ancestors. If you call a selector
10826: with a receiving object that is not in one of these classes, the
10827: result is undefined; if you are lucky, the program crashes
10828: immediately.
1.12 anton 10829:
1.26 crook 10830: @cindex selectors common to hardly-related classes
10831: Now consider the case when you want to have a selector (or several)
10832: available in two classes: You would have to add the selector to a
10833: common ancestor class, in the worst case to @code{object}. You
10834: may not want to do this, e.g., because someone else is responsible for
10835: this ancestor class.
1.12 anton 10836:
1.26 crook 10837: The solution for this problem is interfaces. An interface is a
10838: collection of selectors. If a class implements an interface, the
10839: selectors become available to the class and its descendents. A class
10840: can implement an unlimited number of interfaces. For the problem
10841: discussed above, we would define an interface for the selector(s), and
10842: both classes would implement the interface.
1.12 anton 10843:
1.26 crook 10844: As an example, consider an interface @code{storage} for
10845: writing objects to disk and getting them back, and a class
10846: @code{foo} that implements it. The code would look like this:
1.12 anton 10847:
1.26 crook 10848: @cindex @code{interface} usage
10849: @cindex @code{end-interface} usage
10850: @cindex @code{implementation} usage
10851: @example
10852: interface
10853: selector write ( file object -- )
10854: selector read1 ( file object -- )
10855: end-interface storage
1.12 anton 10856:
1.26 crook 10857: bar class
10858: storage implementation
1.12 anton 10859:
1.26 crook 10860: ... overrides write
1.37 anton 10861: ... overrides read1
1.26 crook 10862: ...
10863: end-class foo
1.12 anton 10864: @end example
10865:
1.26 crook 10866: @noindent
1.29 crook 10867: (I would add a word @code{read} @i{( file -- object )} that uses
1.26 crook 10868: @code{read1} internally, but that's beyond the point illustrated
10869: here.)
1.12 anton 10870:
1.26 crook 10871: Note that you cannot use @code{protected} in an interface; and
10872: of course you cannot define fields.
1.12 anton 10873:
1.26 crook 10874: In the Neon model, all selectors are available for all classes;
10875: therefore it does not need interfaces. The price you pay in this model
10876: is slower late binding, and therefore, added complexity to avoid late
10877: binding.
1.12 anton 10878:
1.26 crook 10879: @node Objects Implementation, Objects Glossary, Object Interfaces, Objects
10880: @subsubsection @file{objects.fs} Implementation
10881: @cindex @file{objects.fs} implementation
1.12 anton 10882:
1.26 crook 10883: @cindex @code{object-map} discussion
10884: An object is a piece of memory, like one of the data structures
10885: described with @code{struct...end-struct}. It has a field
10886: @code{object-map} that points to the method map for the object's
10887: class.
1.12 anton 10888:
1.26 crook 10889: @cindex method map
10890: @cindex virtual function table
10891: The @emph{method map}@footnote{This is Self terminology; in C++
10892: terminology: virtual function table.} is an array that contains the
1.29 crook 10893: execution tokens (@i{xt}s) of the methods for the object's class. Each
1.26 crook 10894: selector contains an offset into a method map.
1.12 anton 10895:
1.26 crook 10896: @cindex @code{selector} implementation, class
10897: @code{selector} is a defining word that uses
10898: @code{CREATE} and @code{DOES>}. The body of the
1.44 crook 10899: selector contains the offset; the @code{DOES>} action for a
1.26 crook 10900: class selector is, basically:
1.21 crook 10901:
1.26 crook 10902: @example
10903: ( object addr ) @@ over object-map @@ + @@ execute
10904: @end example
1.12 anton 10905:
1.26 crook 10906: Since @code{object-map} is the first field of the object, it
10907: does not generate any code. As you can see, calling a selector has a
10908: small, constant cost.
1.12 anton 10909:
1.26 crook 10910: @cindex @code{current-interface} discussion
10911: @cindex class implementation and representation
10912: A class is basically a @code{struct} combined with a method
10913: map. During the class definition the alignment and size of the class
10914: are passed on the stack, just as with @code{struct}s, so
10915: @code{field} can also be used for defining class
10916: fields. However, passing more items on the stack would be
10917: inconvenient, so @code{class} builds a data structure in memory,
10918: which is accessed through the variable
10919: @code{current-interface}. After its definition is complete, the
10920: class is represented on the stack by a pointer (e.g., as parameter for
10921: a child class definition).
1.1 anton 10922:
1.26 crook 10923: A new class starts off with the alignment and size of its parent,
10924: and a copy of the parent's method map. Defining new fields extends the
10925: size and alignment; likewise, defining new selectors extends the
1.29 crook 10926: method map. @code{overrides} just stores a new @i{xt} in the method
1.26 crook 10927: map at the offset given by the selector.
1.20 pazsan 10928:
1.26 crook 10929: @cindex class binding, implementation
1.29 crook 10930: Class binding just gets the @i{xt} at the offset given by the selector
1.26 crook 10931: from the class's method map and @code{compile,}s (in the case of
10932: @code{[bind]}) it.
1.21 crook 10933:
1.26 crook 10934: @cindex @code{this} implementation
10935: @cindex @code{catch} and @code{this}
10936: @cindex @code{this} and @code{catch}
10937: I implemented @code{this} as a @code{value}. At the
10938: start of an @code{m:...;m} method the old @code{this} is
10939: stored to the return stack and restored at the end; and the object on
10940: the TOS is stored @code{TO this}. This technique has one
10941: disadvantage: If the user does not leave the method via
10942: @code{;m}, but via @code{throw} or @code{exit},
10943: @code{this} is not restored (and @code{exit} may
10944: crash). To deal with the @code{throw} problem, I have redefined
10945: @code{catch} to save and restore @code{this}; the same
10946: should be done with any word that can catch an exception. As for
10947: @code{exit}, I simply forbid it (as a replacement, there is
10948: @code{exitm}).
1.21 crook 10949:
1.26 crook 10950: @cindex @code{inst-var} implementation
10951: @code{inst-var} is just the same as @code{field}, with
10952: a different @code{DOES>} action:
10953: @example
10954: @@ this +
10955: @end example
10956: Similar for @code{inst-value}.
1.21 crook 10957:
1.26 crook 10958: @cindex class scoping implementation
10959: Each class also has a word list that contains the words defined with
10960: @code{inst-var} and @code{inst-value}, and its protected
10961: words. It also has a pointer to its parent. @code{class} pushes
10962: the word lists of the class and all its ancestors onto the search order stack,
10963: and @code{end-class} drops them.
1.21 crook 10964:
1.26 crook 10965: @cindex interface implementation
10966: An interface is like a class without fields, parent and protected
10967: words; i.e., it just has a method map. If a class implements an
10968: interface, its method map contains a pointer to the method map of the
10969: interface. The positive offsets in the map are reserved for class
10970: methods, therefore interface map pointers have negative
10971: offsets. Interfaces have offsets that are unique throughout the
10972: system, unlike class selectors, whose offsets are only unique for the
10973: classes where the selector is available (invokable).
1.21 crook 10974:
1.26 crook 10975: This structure means that interface selectors have to perform one
10976: indirection more than class selectors to find their method. Their body
10977: contains the interface map pointer offset in the class method map, and
10978: the method offset in the interface method map. The
10979: @code{does>} action for an interface selector is, basically:
1.21 crook 10980:
10981: @example
1.26 crook 10982: ( object selector-body )
10983: 2dup selector-interface @@ ( object selector-body object interface-offset )
10984: swap object-map @@ + @@ ( object selector-body map )
10985: swap selector-offset @@ + @@ execute
1.21 crook 10986: @end example
10987:
1.26 crook 10988: where @code{object-map} and @code{selector-offset} are
10989: first fields and generate no code.
10990:
10991: As a concrete example, consider the following code:
1.21 crook 10992:
1.26 crook 10993: @example
10994: interface
10995: selector if1sel1
10996: selector if1sel2
10997: end-interface if1
1.21 crook 10998:
1.26 crook 10999: object class
11000: if1 implementation
11001: selector cl1sel1
11002: cell% inst-var cl1iv1
1.21 crook 11003:
1.26 crook 11004: ' m1 overrides construct
11005: ' m2 overrides if1sel1
11006: ' m3 overrides if1sel2
11007: ' m4 overrides cl1sel2
11008: end-class cl1
1.21 crook 11009:
1.26 crook 11010: create obj1 object dict-new drop
11011: create obj2 cl1 dict-new drop
11012: @end example
1.21 crook 11013:
1.26 crook 11014: The data structure created by this code (including the data structure
11015: for @code{object}) is shown in the <a
11016: href="objects-implementation.eps">figure</a>, assuming a cell size of 4.
1.29 crook 11017: @comment TODO add this diagram..
1.21 crook 11018:
1.26 crook 11019: @node Objects Glossary, , Objects Implementation, Objects
11020: @subsubsection @file{objects.fs} Glossary
11021: @cindex @file{objects.fs} Glossary
1.21 crook 11022:
1.44 crook 11023:
1.26 crook 11024: doc---objects-bind
11025: doc---objects-<bind>
11026: doc---objects-bind'
11027: doc---objects-[bind]
11028: doc---objects-class
11029: doc---objects-class->map
11030: doc---objects-class-inst-size
11031: doc---objects-class-override!
11032: doc---objects-construct
11033: doc---objects-current'
11034: doc---objects-[current]
11035: doc---objects-current-interface
11036: doc---objects-dict-new
11037: doc---objects-drop-order
11038: doc---objects-end-class
11039: doc---objects-end-class-noname
11040: doc---objects-end-interface
11041: doc---objects-end-interface-noname
1.37 anton 11042: doc---objects-end-methods
1.26 crook 11043: doc---objects-exitm
11044: doc---objects-heap-new
11045: doc---objects-implementation
11046: doc---objects-init-object
11047: doc---objects-inst-value
11048: doc---objects-inst-var
11049: doc---objects-interface
1.38 anton 11050: doc---objects-m:
11051: doc---objects-:m
1.26 crook 11052: doc---objects-;m
11053: doc---objects-method
1.37 anton 11054: doc---objects-methods
1.26 crook 11055: doc---objects-object
11056: doc---objects-overrides
11057: doc---objects-[parent]
11058: doc---objects-print
11059: doc---objects-protected
11060: doc---objects-public
11061: doc---objects-push-order
11062: doc---objects-selector
11063: doc---objects-this
11064: doc---objects-<to-inst>
11065: doc---objects-[to-inst]
11066: doc---objects-to-this
11067: doc---objects-xt-new
1.21 crook 11068:
1.44 crook 11069:
1.26 crook 11070: @c -------------------------------------------------------------
11071: @node OOF, Mini-OOF, Objects, Object-oriented Forth
11072: @subsection The @file{oof.fs} model
11073: @cindex oof
11074: @cindex object-oriented programming
1.21 crook 11075:
1.26 crook 11076: @cindex @file{objects.fs}
11077: @cindex @file{oof.fs}
1.21 crook 11078:
1.26 crook 11079: This section describes the @file{oof.fs} package.
1.21 crook 11080:
1.26 crook 11081: The package described in this section has been used in bigFORTH since 1991, and
11082: used for two large applications: a chromatographic system used to
11083: create new medicaments, and a graphic user interface library (MINOS).
1.21 crook 11084:
1.26 crook 11085: You can find a description (in German) of @file{oof.fs} in @cite{Object
11086: oriented bigFORTH} by Bernd Paysan, published in @cite{Vierte Dimension}
11087: 10(2), 1994.
1.21 crook 11088:
1.26 crook 11089: @menu
1.67 anton 11090: * Properties of the OOF model::
11091: * Basic OOF Usage::
11092: * The OOF base class::
11093: * Class Declaration::
11094: * Class Implementation::
1.26 crook 11095: @end menu
1.21 crook 11096:
1.26 crook 11097: @node Properties of the OOF model, Basic OOF Usage, OOF, OOF
11098: @subsubsection Properties of the @file{oof.fs} model
11099: @cindex @file{oof.fs} properties
1.21 crook 11100:
1.26 crook 11101: @itemize @bullet
11102: @item
11103: This model combines object oriented programming with information
11104: hiding. It helps you writing large application, where scoping is
11105: necessary, because it provides class-oriented scoping.
1.21 crook 11106:
1.26 crook 11107: @item
11108: Named objects, object pointers, and object arrays can be created,
11109: selector invocation uses the ``object selector'' syntax. Selector invocation
11110: to objects and/or selectors on the stack is a bit less convenient, but
11111: possible.
1.21 crook 11112:
1.26 crook 11113: @item
11114: Selector invocation and instance variable usage of the active object is
11115: straightforward, since both make use of the active object.
1.21 crook 11116:
1.26 crook 11117: @item
11118: Late binding is efficient and easy to use.
1.21 crook 11119:
1.26 crook 11120: @item
11121: State-smart objects parse selectors. However, extensibility is provided
11122: using a (parsing) selector @code{postpone} and a selector @code{'}.
1.21 crook 11123:
11124: @item
1.26 crook 11125: An implementation in ANS Forth is available.
11126:
1.21 crook 11127: @end itemize
11128:
11129:
1.26 crook 11130: @node Basic OOF Usage, The OOF base class, Properties of the OOF model, OOF
11131: @subsubsection Basic @file{oof.fs} Usage
11132: @cindex @file{oof.fs} usage
11133:
11134: This section uses the same example as for @code{objects} (@pxref{Basic Objects Usage}).
1.21 crook 11135:
1.26 crook 11136: You can define a class for graphical objects like this:
1.21 crook 11137:
1.26 crook 11138: @cindex @code{class} usage
11139: @cindex @code{class;} usage
11140: @cindex @code{method} usage
11141: @example
11142: object class graphical \ "object" is the parent class
11143: method draw ( x y graphical -- )
11144: class;
11145: @end example
1.21 crook 11146:
1.26 crook 11147: This code defines a class @code{graphical} with an
11148: operation @code{draw}. We can perform the operation
11149: @code{draw} on any @code{graphical} object, e.g.:
1.21 crook 11150:
1.26 crook 11151: @example
11152: 100 100 t-rex draw
11153: @end example
1.21 crook 11154:
1.26 crook 11155: @noindent
11156: where @code{t-rex} is an object or object pointer, created with e.g.
11157: @code{graphical : t-rex}.
1.21 crook 11158:
1.26 crook 11159: @cindex abstract class
11160: How do we create a graphical object? With the present definitions,
11161: we cannot create a useful graphical object. The class
11162: @code{graphical} describes graphical objects in general, but not
11163: any concrete graphical object type (C++ users would call it an
11164: @emph{abstract class}); e.g., there is no method for the selector
11165: @code{draw} in the class @code{graphical}.
1.21 crook 11166:
1.26 crook 11167: For concrete graphical objects, we define child classes of the
11168: class @code{graphical}, e.g.:
1.21 crook 11169:
11170: @example
1.26 crook 11171: graphical class circle \ "graphical" is the parent class
11172: cell var circle-radius
11173: how:
11174: : draw ( x y -- )
11175: circle-radius @@ draw-circle ;
11176:
11177: : init ( n-radius -- (
11178: circle-radius ! ;
11179: class;
11180: @end example
11181:
11182: Here we define a class @code{circle} as a child of @code{graphical},
11183: with a field @code{circle-radius}; it defines new methods for the
11184: selectors @code{draw} and @code{init} (@code{init} is defined in
11185: @code{object}, the parent class of @code{graphical}).
1.21 crook 11186:
1.26 crook 11187: Now we can create a circle in the dictionary with:
1.21 crook 11188:
1.26 crook 11189: @example
11190: 50 circle : my-circle
1.21 crook 11191: @end example
11192:
1.26 crook 11193: @noindent
11194: @code{:} invokes @code{init}, thus initializing the field
11195: @code{circle-radius} with 50. We can draw this new circle at (100,100)
11196: with:
1.21 crook 11197:
11198: @example
1.26 crook 11199: 100 100 my-circle draw
1.21 crook 11200: @end example
11201:
1.26 crook 11202: @cindex selector invocation, restrictions
11203: @cindex class definition, restrictions
11204: Note: You can only invoke a selector if the receiving object belongs to
11205: the class where the selector was defined or one of its descendents;
11206: e.g., you can invoke @code{draw} only for objects belonging to
11207: @code{graphical} or its descendents (e.g., @code{circle}). The scoping
11208: mechanism will check if you try to invoke a selector that is not
11209: defined in this class hierarchy, so you'll get an error at compilation
11210: time.
11211:
11212:
11213: @node The OOF base class, Class Declaration, Basic OOF Usage, OOF
11214: @subsubsection The @file{oof.fs} base class
11215: @cindex @file{oof.fs} base class
11216:
11217: When you define a class, you have to specify a parent class. So how do
11218: you start defining classes? There is one class available from the start:
11219: @code{object}. You have to use it as ancestor for all classes. It is the
11220: only class that has no parent. Classes are also objects, except that
11221: they don't have instance variables; class manipulation such as
11222: inheritance or changing definitions of a class is handled through
11223: selectors of the class @code{object}.
11224:
11225: @code{object} provides a number of selectors:
11226:
1.21 crook 11227: @itemize @bullet
11228: @item
1.26 crook 11229: @code{class} for subclassing, @code{definitions} to add definitions
11230: later on, and @code{class?} to get type informations (is the class a
11231: subclass of the class passed on the stack?).
1.44 crook 11232:
1.26 crook 11233: doc---object-class
11234: doc---object-definitions
11235: doc---object-class?
11236:
1.44 crook 11237:
1.21 crook 11238: @item
1.26 crook 11239: @code{init} and @code{dispose} as constructor and destructor of the
11240: object. @code{init} is invocated after the object's memory is allocated,
11241: while @code{dispose} also handles deallocation. Thus if you redefine
11242: @code{dispose}, you have to call the parent's dispose with @code{super
11243: dispose}, too.
1.44 crook 11244:
1.26 crook 11245: doc---object-init
11246: doc---object-dispose
11247:
1.44 crook 11248:
1.21 crook 11249: @item
1.26 crook 11250: @code{new}, @code{new[]}, @code{:}, @code{ptr}, @code{asptr}, and
11251: @code{[]} to create named and unnamed objects and object arrays or
11252: object pointers.
1.44 crook 11253:
1.26 crook 11254: doc---object-new
11255: doc---object-new[]
11256: doc---object-:
11257: doc---object-ptr
11258: doc---object-asptr
11259: doc---object-[]
1.21 crook 11260:
1.44 crook 11261:
1.26 crook 11262: @item
11263: @code{::} and @code{super} for explicit scoping. You should use explicit
11264: scoping only for super classes or classes with the same set of instance
11265: variables. Explicitly-scoped selectors use early binding.
1.44 crook 11266:
1.26 crook 11267: doc---object-::
11268: doc---object-super
1.21 crook 11269:
1.44 crook 11270:
1.26 crook 11271: @item
11272: @code{self} to get the address of the object
1.44 crook 11273:
1.26 crook 11274: doc---object-self
1.21 crook 11275:
1.44 crook 11276:
1.21 crook 11277: @item
1.26 crook 11278: @code{bind}, @code{bound}, @code{link}, and @code{is} to assign object
11279: pointers and instance defers.
1.44 crook 11280:
1.26 crook 11281: doc---object-bind
11282: doc---object-bound
11283: doc---object-link
11284: doc---object-is
11285:
1.44 crook 11286:
1.21 crook 11287: @item
1.26 crook 11288: @code{'} to obtain selector tokens, @code{send} to invocate selectors
11289: form the stack, and @code{postpone} to generate selector invocation code.
1.44 crook 11290:
1.26 crook 11291: doc---object-'
11292: doc---object-postpone
11293:
1.44 crook 11294:
1.21 crook 11295: @item
1.26 crook 11296: @code{with} and @code{endwith} to select the active object from the
11297: stack, and enable its scope. Using @code{with} and @code{endwith}
11298: also allows you to create code using selector @code{postpone} without being
11299: trapped by the state-smart objects.
1.44 crook 11300:
1.26 crook 11301: doc---object-with
11302: doc---object-endwith
11303:
1.44 crook 11304:
1.21 crook 11305: @end itemize
11306:
1.26 crook 11307: @node Class Declaration, Class Implementation, The OOF base class, OOF
11308: @subsubsection Class Declaration
11309: @cindex class declaration
11310:
11311: @itemize @bullet
11312: @item
11313: Instance variables
1.44 crook 11314:
1.26 crook 11315: doc---oof-var
1.21 crook 11316:
1.44 crook 11317:
1.26 crook 11318: @item
11319: Object pointers
1.44 crook 11320:
1.26 crook 11321: doc---oof-ptr
11322: doc---oof-asptr
1.21 crook 11323:
1.44 crook 11324:
1.26 crook 11325: @item
11326: Instance defers
1.44 crook 11327:
1.26 crook 11328: doc---oof-defer
1.21 crook 11329:
1.44 crook 11330:
1.26 crook 11331: @item
11332: Method selectors
1.44 crook 11333:
1.26 crook 11334: doc---oof-early
11335: doc---oof-method
1.21 crook 11336:
1.44 crook 11337:
1.26 crook 11338: @item
11339: Class-wide variables
1.44 crook 11340:
1.26 crook 11341: doc---oof-static
1.21 crook 11342:
1.44 crook 11343:
1.26 crook 11344: @item
11345: End declaration
1.44 crook 11346:
1.26 crook 11347: doc---oof-how:
11348: doc---oof-class;
1.21 crook 11349:
1.44 crook 11350:
1.26 crook 11351: @end itemize
1.21 crook 11352:
1.26 crook 11353: @c -------------------------------------------------------------
11354: @node Class Implementation, , Class Declaration, OOF
11355: @subsubsection Class Implementation
11356: @cindex class implementation
1.21 crook 11357:
1.26 crook 11358: @c -------------------------------------------------------------
11359: @node Mini-OOF, Comparison with other object models, OOF, Object-oriented Forth
11360: @subsection The @file{mini-oof.fs} model
11361: @cindex mini-oof
1.1 anton 11362:
1.26 crook 11363: Gforth's third object oriented Forth package is a 12-liner. It uses a
11364: mixture of the @file{object.fs} and the @file{oof.fs} syntax,
11365: and reduces to the bare minimum of features. This is based on a posting
1.70 pazsan 11366: of Bernd Paysan in comp.lang.forth.
1.1 anton 11367:
11368: @menu
1.48 anton 11369: * Basic Mini-OOF Usage::
11370: * Mini-OOF Example::
11371: * Mini-OOF Implementation::
1.1 anton 11372: @end menu
11373:
1.26 crook 11374: @c -------------------------------------------------------------
1.48 anton 11375: @node Basic Mini-OOF Usage, Mini-OOF Example, Mini-OOF, Mini-OOF
1.26 crook 11376: @subsubsection Basic @file{mini-oof.fs} Usage
11377: @cindex mini-oof usage
1.1 anton 11378:
1.28 crook 11379: There is a base class (@code{class}, which allocates one cell for the
11380: object pointer) plus seven other words: to define a method, a variable,
11381: a class; to end a class, to resolve binding, to allocate an object and
11382: to compile a class method.
1.26 crook 11383: @comment TODO better description of the last one
1.1 anton 11384:
1.44 crook 11385:
1.26 crook 11386: doc-object
11387: doc-method
11388: doc-var
11389: doc-class
11390: doc-end-class
11391: doc-defines
11392: doc-new
11393: doc-::
1.1 anton 11394:
1.21 crook 11395:
1.44 crook 11396:
1.26 crook 11397: @c -------------------------------------------------------------
11398: @node Mini-OOF Example, Mini-OOF Implementation, Basic Mini-OOF Usage, Mini-OOF
11399: @subsubsection Mini-OOF Example
11400: @cindex mini-oof example
1.21 crook 11401:
1.26 crook 11402: A short example shows how to use this package. This example, in slightly
11403: extended form, is supplied as @file{moof-exm.fs}
1.29 crook 11404: @comment TODO could flesh this out with some comments from the Forthwrite article
1.21 crook 11405:
1.26 crook 11406: @example
11407: object class
11408: method init
11409: method draw
11410: end-class graphical
11411: @end example
1.21 crook 11412:
1.26 crook 11413: This code defines a class @code{graphical} with an
11414: operation @code{draw}. We can perform the operation
11415: @code{draw} on any @code{graphical} object, e.g.:
1.1 anton 11416:
1.26 crook 11417: @example
11418: 100 100 t-rex draw
11419: @end example
1.1 anton 11420:
1.26 crook 11421: where @code{t-rex} is an object or object pointer, created with e.g.
11422: @code{graphical new Constant t-rex}.
1.1 anton 11423:
1.26 crook 11424: For concrete graphical objects, we define child classes of the
11425: class @code{graphical}, e.g.:
1.21 crook 11426:
11427: @example
1.26 crook 11428: graphical class
11429: cell var circle-radius
11430: end-class circle \ "graphical" is the parent class
1.21 crook 11431:
1.26 crook 11432: :noname ( x y -- )
11433: circle-radius @@ draw-circle ; circle defines draw
11434: :noname ( r -- )
11435: circle-radius ! ; circle defines init
1.21 crook 11436: @end example
11437:
1.26 crook 11438: There is no implicit init method, so we have to define one. The creation
11439: code of the object now has to call init explicitely.
1.21 crook 11440:
1.26 crook 11441: @example
11442: circle new Constant my-circle
11443: 50 my-circle init
11444: @end example
1.21 crook 11445:
1.26 crook 11446: It is also possible to add a function to create named objects with
11447: automatic call of @code{init}, given that all objects have @code{init}
11448: on the same place:
1.1 anton 11449:
11450: @example
1.26 crook 11451: : new: ( .. o "name" -- )
11452: new dup Constant init ;
11453: 80 circle new: large-circle
1.1 anton 11454: @end example
11455:
1.26 crook 11456: We can draw this new circle at (100,100) with:
1.1 anton 11457:
11458: @example
1.26 crook 11459: 100 100 my-circle draw
1.1 anton 11460: @end example
11461:
1.48 anton 11462: @node Mini-OOF Implementation, , Mini-OOF Example, Mini-OOF
1.26 crook 11463: @subsubsection @file{mini-oof.fs} Implementation
1.1 anton 11464:
1.26 crook 11465: Object-oriented systems with late binding typically use a
11466: ``vtable''-approach: the first variable in each object is a pointer to a
11467: table, which contains the methods as function pointers. The vtable
11468: may also contain other information.
1.1 anton 11469:
1.26 crook 11470: So first, let's declare methods:
1.1 anton 11471:
1.26 crook 11472: @example
11473: : method ( m v -- m' v ) Create over , swap cell+ swap
1.73 anton 11474: DOES> ( ... o -- ... ) @@ over @@ + @@ execute ;
1.26 crook 11475: @end example
1.1 anton 11476:
1.26 crook 11477: During method declaration, the number of methods and instance
11478: variables is on the stack (in address units). @code{method} creates
11479: one method and increments the method number. To execute a method, it
11480: takes the object, fetches the vtable pointer, adds the offset, and
1.29 crook 11481: executes the @i{xt} stored there. Each method takes the object it is
1.26 crook 11482: invoked from as top of stack parameter. The method itself should
11483: consume that object.
1.1 anton 11484:
1.26 crook 11485: Now, we also have to declare instance variables
1.21 crook 11486:
1.26 crook 11487: @example
11488: : var ( m v size -- m v' ) Create over , +
1.73 anton 11489: DOES> ( o -- addr ) @@ + ;
1.26 crook 11490: @end example
1.21 crook 11491:
1.26 crook 11492: As before, a word is created with the current offset. Instance
11493: variables can have different sizes (cells, floats, doubles, chars), so
11494: all we do is take the size and add it to the offset. If your machine
11495: has alignment restrictions, put the proper @code{aligned} or
11496: @code{faligned} before the variable, to adjust the variable
11497: offset. That's why it is on the top of stack.
1.2 jwilke 11498:
1.26 crook 11499: We need a starting point (the base object) and some syntactic sugar:
1.21 crook 11500:
1.26 crook 11501: @example
11502: Create object 1 cells , 2 cells ,
1.73 anton 11503: : class ( class -- class methods vars ) dup 2@@ ;
1.26 crook 11504: @end example
1.21 crook 11505:
1.26 crook 11506: For inheritance, the vtable of the parent object has to be
11507: copied when a new, derived class is declared. This gives all the
11508: methods of the parent class, which can be overridden, though.
1.21 crook 11509:
1.2 jwilke 11510: @example
1.26 crook 11511: : end-class ( class methods vars -- )
11512: Create here >r , dup , 2 cells ?DO ['] noop , 1 cells +LOOP
1.73 anton 11513: cell+ dup cell+ r> rot @@ 2 cells /string move ;
1.26 crook 11514: @end example
11515:
11516: The first line creates the vtable, initialized with
11517: @code{noop}s. The second line is the inheritance mechanism, it
11518: copies the xts from the parent vtable.
1.2 jwilke 11519:
1.26 crook 11520: We still have no way to define new methods, let's do that now:
1.2 jwilke 11521:
1.26 crook 11522: @example
1.73 anton 11523: : defines ( xt class -- ) ' >body @@ + ! ;
1.2 jwilke 11524: @end example
11525:
1.26 crook 11526: To allocate a new object, we need a word, too:
1.2 jwilke 11527:
1.26 crook 11528: @example
1.73 anton 11529: : new ( class -- o ) here over @@ allot swap over ! ;
1.26 crook 11530: @end example
1.2 jwilke 11531:
1.26 crook 11532: Sometimes derived classes want to access the method of the
11533: parent object. There are two ways to achieve this with Mini-OOF:
11534: first, you could use named words, and second, you could look up the
11535: vtable of the parent object.
1.2 jwilke 11536:
1.26 crook 11537: @example
1.73 anton 11538: : :: ( class "name" -- ) ' >body @@ + @@ compile, ;
1.26 crook 11539: @end example
1.2 jwilke 11540:
11541:
1.26 crook 11542: Nothing can be more confusing than a good example, so here is
11543: one. First let's declare a text object (called
11544: @code{button}), that stores text and position:
1.2 jwilke 11545:
1.26 crook 11546: @example
11547: object class
11548: cell var text
11549: cell var len
11550: cell var x
11551: cell var y
11552: method init
11553: method draw
11554: end-class button
11555: @end example
1.2 jwilke 11556:
1.26 crook 11557: @noindent
11558: Now, implement the two methods, @code{draw} and @code{init}:
1.2 jwilke 11559:
1.26 crook 11560: @example
11561: :noname ( o -- )
1.73 anton 11562: >r r@@ x @@ r@@ y @@ at-xy r@@ text @@ r> len @@ type ;
1.26 crook 11563: button defines draw
11564: :noname ( addr u o -- )
1.73 anton 11565: >r 0 r@@ x ! 0 r@@ y ! r@@ len ! r> text ! ;
1.26 crook 11566: button defines init
11567: @end example
1.2 jwilke 11568:
1.26 crook 11569: @noindent
11570: To demonstrate inheritance, we define a class @code{bold-button}, with no
11571: new data and no new methods:
1.2 jwilke 11572:
1.26 crook 11573: @example
11574: button class
11575: end-class bold-button
1.1 anton 11576:
1.26 crook 11577: : bold 27 emit ." [1m" ;
11578: : normal 27 emit ." [0m" ;
11579: @end example
1.1 anton 11580:
1.26 crook 11581: @noindent
11582: The class @code{bold-button} has a different draw method to
11583: @code{button}, but the new method is defined in terms of the draw method
11584: for @code{button}:
1.1 anton 11585:
1.26 crook 11586: @example
11587: :noname bold [ button :: draw ] normal ; bold-button defines draw
11588: @end example
1.1 anton 11589:
1.26 crook 11590: @noindent
11591: Finally, create two objects and apply methods:
1.1 anton 11592:
1.26 crook 11593: @example
11594: button new Constant foo
11595: s" thin foo" foo init
11596: page
11597: foo draw
11598: bold-button new Constant bar
11599: s" fat bar" bar init
11600: 1 bar y !
11601: bar draw
11602: @end example
1.1 anton 11603:
11604:
1.48 anton 11605: @node Comparison with other object models, , Mini-OOF, Object-oriented Forth
11606: @subsection Comparison with other object models
1.26 crook 11607: @cindex comparison of object models
11608: @cindex object models, comparison
1.1 anton 11609:
1.26 crook 11610: Many object-oriented Forth extensions have been proposed (@cite{A survey
11611: of object-oriented Forths} (SIGPLAN Notices, April 1996) by Bradford
11612: J. Rodriguez and W. F. S. Poehlman lists 17). This section discusses the
11613: relation of the object models described here to two well-known and two
11614: closely-related (by the use of method maps) models.
1.1 anton 11615:
1.26 crook 11616: @cindex Neon model
11617: The most popular model currently seems to be the Neon model (see
11618: @cite{Object-oriented programming in ANS Forth} (Forth Dimensions, March
11619: 1997) by Andrew McKewan) but this model has a number of limitations
11620: @footnote{A longer version of this critique can be
11621: found in @cite{On Standardizing Object-Oriented Forth Extensions} (Forth
11622: Dimensions, May 1997) by Anton Ertl.}:
1.1 anton 11623:
1.26 crook 11624: @itemize @bullet
11625: @item
1.48 anton 11626: It uses a @code{@emph{selector object}} syntax, which makes it unnatural
11627: to pass objects on the stack.
1.1 anton 11628:
1.26 crook 11629: @item
11630: It requires that the selector parses the input stream (at
11631: compile time); this leads to reduced extensibility and to bugs that are+
11632: hard to find.
1.1 anton 11633:
1.26 crook 11634: @item
11635: It allows using every selector to every object;
11636: this eliminates the need for classes, but makes it harder to create
11637: efficient implementations.
11638: @end itemize
1.1 anton 11639:
1.26 crook 11640: @cindex Pountain's object-oriented model
11641: Another well-known publication is @cite{Object-Oriented Forth} (Academic
11642: Press, London, 1987) by Dick Pountain. However, it is not really about
11643: object-oriented programming, because it hardly deals with late
11644: binding. Instead, it focuses on features like information hiding and
11645: overloading that are characteristic of modular languages like Ada (83).
1.1 anton 11646:
1.26 crook 11647: @cindex Zsoter's object-oriented model
1.48 anton 11648: In @cite{Does late binding have to be slow?} (Forth Dimensions 18(1)
11649: 1996, pages 31-35) Andras Zsoter describes a model that makes heavy use
11650: of an active object (like @code{this} in @file{objects.fs}): The active
11651: object is not only used for accessing all fields, but also specifies the
11652: receiving object of every selector invocation; you have to change the
11653: active object explicitly with @code{@{ ... @}}, whereas in
11654: @file{objects.fs} it changes more or less implicitly at @code{m:
11655: ... ;m}. Such a change at the method entry point is unnecessary with the
11656: Zsoter's model, because the receiving object is the active object
11657: already. On the other hand, the explicit change is absolutely necessary
11658: in that model, because otherwise no one could ever change the active
11659: object. An ANS Forth implementation of this model is available at
11660: @uref{http://www.forth.org/fig/oopf.html}.
1.1 anton 11661:
1.26 crook 11662: @cindex @file{oof.fs}, differences to other models
11663: The @file{oof.fs} model combines information hiding and overloading
11664: resolution (by keeping names in various word lists) with object-oriented
11665: programming. It sets the active object implicitly on method entry, but
11666: also allows explicit changing (with @code{>o...o>} or with
11667: @code{with...endwith}). It uses parsing and state-smart objects and
11668: classes for resolving overloading and for early binding: the object or
11669: class parses the selector and determines the method from this. If the
11670: selector is not parsed by an object or class, it performs a call to the
11671: selector for the active object (late binding), like Zsoter's model.
11672: Fields are always accessed through the active object. The big
11673: disadvantage of this model is the parsing and the state-smartness, which
11674: reduces extensibility and increases the opportunities for subtle bugs;
11675: essentially, you are only safe if you never tick or @code{postpone} an
11676: object or class (Bernd disagrees, but I (Anton) am not convinced).
1.1 anton 11677:
1.26 crook 11678: @cindex @file{mini-oof.fs}, differences to other models
1.48 anton 11679: The @file{mini-oof.fs} model is quite similar to a very stripped-down
11680: version of the @file{objects.fs} model, but syntactically it is a
11681: mixture of the @file{objects.fs} and @file{oof.fs} models.
1.1 anton 11682:
1.26 crook 11683: @c -------------------------------------------------------------
1.47 crook 11684: @node Passing Commands to the OS, Keeping track of Time, Object-oriented Forth, Words
1.21 crook 11685: @section Passing Commands to the Operating System
11686: @cindex operating system - passing commands
11687: @cindex shell commands
11688:
11689: Gforth allows you to pass an arbitrary string to the host operating
11690: system shell (if such a thing exists) for execution.
11691:
1.44 crook 11692:
1.21 crook 11693: doc-sh
11694: doc-system
11695: doc-$?
1.23 crook 11696: doc-getenv
1.21 crook 11697:
1.44 crook 11698:
1.26 crook 11699: @c -------------------------------------------------------------
1.47 crook 11700: @node Keeping track of Time, Miscellaneous Words, Passing Commands to the OS, Words
11701: @section Keeping track of Time
11702: @cindex time-related words
11703:
11704: Gforth implements time-related operations by making calls to the C
11705: library function, @code{gettimeofday}.
11706:
11707: doc-ms
11708: doc-time&date
11709:
11710:
11711:
11712: @c -------------------------------------------------------------
11713: @node Miscellaneous Words, , Keeping track of Time, Words
1.21 crook 11714: @section Miscellaneous Words
11715: @cindex miscellaneous words
11716:
1.29 crook 11717: @comment TODO find homes for these
11718:
1.26 crook 11719: These section lists the ANS Forth words that are not documented
1.21 crook 11720: elsewhere in this manual. Ultimately, they all need proper homes.
11721:
11722: doc-[compile]
1.68 anton 11723: doc-quit
1.44 crook 11724:
1.26 crook 11725: The following ANS Forth words are not currently supported by Gforth
1.27 crook 11726: (@pxref{ANS conformance}):
1.21 crook 11727:
11728: @code{EDITOR}
11729: @code{EMIT?}
11730: @code{FORGET}
11731:
1.24 anton 11732: @c ******************************************************************
11733: @node Error messages, Tools, Words, Top
11734: @chapter Error messages
11735: @cindex error messages
11736: @cindex backtrace
11737:
11738: A typical Gforth error message looks like this:
11739:
11740: @example
11741: in file included from :-1
11742: in file included from ./yyy.fs:1
11743: ./xxx.fs:4: Invalid memory address
11744: bar
11745: ^^^
1.25 anton 11746: $400E664C @@
11747: $400E6664 foo
1.24 anton 11748: @end example
11749:
11750: The message identifying the error is @code{Invalid memory address}. The
11751: error happened when text-interpreting line 4 of the file
11752: @file{./xxx.fs}. This line is given (it contains @code{bar}), and the
11753: word on the line where the error happened, is pointed out (with
11754: @code{^^^}).
11755:
11756: The file containing the error was included in line 1 of @file{./yyy.fs},
11757: and @file{yyy.fs} was included from a non-file (in this case, by giving
11758: @file{yyy.fs} as command-line parameter to Gforth).
11759:
11760: At the end of the error message you find a return stack dump that can be
11761: interpreted as a backtrace (possibly empty). On top you find the top of
11762: the return stack when the @code{throw} happened, and at the bottom you
11763: find the return stack entry just above the return stack of the topmost
11764: text interpreter.
11765:
11766: To the right of most return stack entries you see a guess for the word
11767: that pushed that return stack entry as its return address. This gives a
11768: backtrace. In our case we see that @code{bar} called @code{foo}, and
11769: @code{foo} called @code{@@} (and @code{@@} had an @emph{Invalid memory
11770: address} exception).
11771:
11772: Note that the backtrace is not perfect: We don't know which return stack
11773: entries are return addresses (so we may get false positives); and in
11774: some cases (e.g., for @code{abort"}) we cannot determine from the return
11775: address the word that pushed the return address, so for some return
11776: addresses you see no names in the return stack dump.
1.25 anton 11777:
11778: @cindex @code{catch} and backtraces
11779: The return stack dump represents the return stack at the time when a
11780: specific @code{throw} was executed. In programs that make use of
11781: @code{catch}, it is not necessarily clear which @code{throw} should be
11782: used for the return stack dump (e.g., consider one @code{throw} that
11783: indicates an error, which is caught, and during recovery another error
1.42 anton 11784: happens; which @code{throw} should be used for the stack dump?). Gforth
1.25 anton 11785: presents the return stack dump for the first @code{throw} after the last
11786: executed (not returned-to) @code{catch}; this works well in the usual
11787: case.
11788:
11789: @cindex @code{gforth-fast} and backtraces
11790: @cindex @code{gforth-fast}, difference from @code{gforth}
11791: @cindex backtraces with @code{gforth-fast}
11792: @cindex return stack dump with @code{gforth-fast}
11793: @code{gforth} is able to do a return stack dump for throws generated
11794: from primitives (e.g., invalid memory address, stack empty etc.);
11795: @code{gforth-fast} is only able to do a return stack dump from a
11796: directly called @code{throw} (including @code{abort} etc.). This is the
1.30 anton 11797: only difference (apart from a speed factor of between 1.15 (K6-2) and
11798: 1.6 (21164A)) between @code{gforth} and @code{gforth-fast}. Given an
11799: exception caused by a primitive in @code{gforth-fast}, you will
11800: typically see no return stack dump at all; however, if the exception is
11801: caught by @code{catch} (e.g., for restoring some state), and then
11802: @code{throw}n again, the return stack dump will be for the first such
11803: @code{throw}.
1.2 jwilke 11804:
1.5 anton 11805: @c ******************************************************************
1.24 anton 11806: @node Tools, ANS conformance, Error messages, Top
1.1 anton 11807: @chapter Tools
11808:
11809: @menu
11810: * ANS Report:: Report the words used, sorted by wordset.
11811: @end menu
11812:
11813: See also @ref{Emacs and Gforth}.
11814:
11815: @node ANS Report, , Tools, Tools
11816: @section @file{ans-report.fs}: Report the words used, sorted by wordset
11817: @cindex @file{ans-report.fs}
11818: @cindex report the words used in your program
11819: @cindex words used in your program
11820:
11821: If you want to label a Forth program as ANS Forth Program, you must
11822: document which wordsets the program uses; for extension wordsets, it is
11823: helpful to list the words the program requires from these wordsets
11824: (because Forth systems are allowed to provide only some words of them).
11825:
11826: The @file{ans-report.fs} tool makes it easy for you to determine which
11827: words from which wordset and which non-ANS words your application
11828: uses. You simply have to include @file{ans-report.fs} before loading the
11829: program you want to check. After loading your program, you can get the
11830: report with @code{print-ans-report}. A typical use is to run this as
11831: batch job like this:
11832: @example
11833: gforth ans-report.fs myprog.fs -e "print-ans-report bye"
11834: @end example
11835:
11836: The output looks like this (for @file{compat/control.fs}):
11837: @example
11838: The program uses the following words
11839: from CORE :
11840: : POSTPONE THEN ; immediate ?dup IF 0=
11841: from BLOCK-EXT :
11842: \
11843: from FILE :
11844: (
11845: @end example
11846:
11847: @subsection Caveats
11848:
11849: Note that @file{ans-report.fs} just checks which words are used, not whether
11850: they are used in an ANS Forth conforming way!
11851:
11852: Some words are defined in several wordsets in the
11853: standard. @file{ans-report.fs} reports them for only one of the
11854: wordsets, and not necessarily the one you expect. It depends on usage
11855: which wordset is the right one to specify. E.g., if you only use the
11856: compilation semantics of @code{S"}, it is a Core word; if you also use
11857: its interpretation semantics, it is a File word.
11858:
11859: @c ******************************************************************
1.65 anton 11860: @node ANS conformance, Standard vs Extensions, Tools, Top
1.1 anton 11861: @chapter ANS conformance
11862: @cindex ANS conformance of Gforth
11863:
11864: To the best of our knowledge, Gforth is an
11865:
11866: ANS Forth System
11867: @itemize @bullet
11868: @item providing the Core Extensions word set
11869: @item providing the Block word set
11870: @item providing the Block Extensions word set
11871: @item providing the Double-Number word set
11872: @item providing the Double-Number Extensions word set
11873: @item providing the Exception word set
11874: @item providing the Exception Extensions word set
11875: @item providing the Facility word set
1.40 anton 11876: @item providing @code{EKEY}, @code{EKEY>CHAR}, @code{EKEY?}, @code{MS} and @code{TIME&DATE} from the Facility Extensions word set
1.1 anton 11877: @item providing the File Access word set
11878: @item providing the File Access Extensions word set
11879: @item providing the Floating-Point word set
11880: @item providing the Floating-Point Extensions word set
11881: @item providing the Locals word set
11882: @item providing the Locals Extensions word set
11883: @item providing the Memory-Allocation word set
11884: @item providing the Memory-Allocation Extensions word set (that one's easy)
11885: @item providing the Programming-Tools word set
11886: @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
11887: @item providing the Search-Order word set
11888: @item providing the Search-Order Extensions word set
11889: @item providing the String word set
11890: @item providing the String Extensions word set (another easy one)
11891: @end itemize
11892:
11893: @cindex system documentation
11894: In addition, ANS Forth systems are required to document certain
11895: implementation choices. This chapter tries to meet these
11896: requirements. In many cases it gives a way to ask the system for the
11897: information instead of providing the information directly, in
11898: particular, if the information depends on the processor, the operating
11899: system or the installation options chosen, or if they are likely to
11900: change during the maintenance of Gforth.
11901:
11902: @comment The framework for the rest has been taken from pfe.
11903:
11904: @menu
11905: * The Core Words::
11906: * The optional Block word set::
11907: * The optional Double Number word set::
11908: * The optional Exception word set::
11909: * The optional Facility word set::
11910: * The optional File-Access word set::
11911: * The optional Floating-Point word set::
11912: * The optional Locals word set::
11913: * The optional Memory-Allocation word set::
11914: * The optional Programming-Tools word set::
11915: * The optional Search-Order word set::
11916: @end menu
11917:
11918:
11919: @c =====================================================================
11920: @node The Core Words, The optional Block word set, ANS conformance, ANS conformance
11921: @comment node-name, next, previous, up
11922: @section The Core Words
11923: @c =====================================================================
11924: @cindex core words, system documentation
11925: @cindex system documentation, core words
11926:
11927: @menu
11928: * core-idef:: Implementation Defined Options
11929: * core-ambcond:: Ambiguous Conditions
11930: * core-other:: Other System Documentation
11931: @end menu
11932:
11933: @c ---------------------------------------------------------------------
11934: @node core-idef, core-ambcond, The Core Words, The Core Words
11935: @subsection Implementation Defined Options
11936: @c ---------------------------------------------------------------------
11937: @cindex core words, implementation-defined options
11938: @cindex implementation-defined options, core words
11939:
11940:
11941: @table @i
11942: @item (Cell) aligned addresses:
11943: @cindex cell-aligned addresses
11944: @cindex aligned addresses
11945: processor-dependent. Gforth's alignment words perform natural alignment
11946: (e.g., an address aligned for a datum of size 8 is divisible by
11947: 8). Unaligned accesses usually result in a @code{-23 THROW}.
11948:
11949: @item @code{EMIT} and non-graphic characters:
11950: @cindex @code{EMIT} and non-graphic characters
11951: @cindex non-graphic characters and @code{EMIT}
11952: The character is output using the C library function (actually, macro)
11953: @code{putc}.
11954:
11955: @item character editing of @code{ACCEPT} and @code{EXPECT}:
11956: @cindex character editing of @code{ACCEPT} and @code{EXPECT}
11957: @cindex editing in @code{ACCEPT} and @code{EXPECT}
11958: @cindex @code{ACCEPT}, editing
11959: @cindex @code{EXPECT}, editing
11960: This is modeled on the GNU readline library (@pxref{Readline
11961: Interaction, , Command Line Editing, readline, The GNU Readline
11962: Library}) with Emacs-like key bindings. @kbd{Tab} deviates a little by
11963: producing a full word completion every time you type it (instead of
1.28 crook 11964: producing the common prefix of all completions). @xref{Command-line editing}.
1.1 anton 11965:
11966: @item character set:
11967: @cindex character set
11968: The character set of your computer and display device. Gforth is
11969: 8-bit-clean (but some other component in your system may make trouble).
11970:
11971: @item Character-aligned address requirements:
11972: @cindex character-aligned address requirements
11973: installation-dependent. Currently a character is represented by a C
11974: @code{unsigned char}; in the future we might switch to @code{wchar_t}
11975: (Comments on that requested).
11976:
11977: @item character-set extensions and matching of names:
11978: @cindex character-set extensions and matching of names
1.26 crook 11979: @cindex case-sensitivity for name lookup
11980: @cindex name lookup, case-sensitivity
11981: @cindex locale and case-sensitivity
1.21 crook 11982: Any character except the ASCII NUL character can be used in a
1.1 anton 11983: name. Matching is case-insensitive (except in @code{TABLE}s). The
1.47 crook 11984: matching is performed using the C library function @code{strncasecmp}, whose
1.1 anton 11985: function is probably influenced by the locale. E.g., the @code{C} locale
11986: does not know about accents and umlauts, so they are matched
11987: case-sensitively in that locale. For portability reasons it is best to
11988: write programs such that they work in the @code{C} locale. Then one can
11989: use libraries written by a Polish programmer (who might use words
11990: containing ISO Latin-2 encoded characters) and by a French programmer
11991: (ISO Latin-1) in the same program (of course, @code{WORDS} will produce
11992: funny results for some of the words (which ones, depends on the font you
11993: are using)). Also, the locale you prefer may not be available in other
11994: operating systems. Hopefully, Unicode will solve these problems one day.
11995:
11996: @item conditions under which control characters match a space delimiter:
11997: @cindex space delimiters
11998: @cindex control characters as delimiters
11999: If @code{WORD} is called with the space character as a delimiter, all
12000: white-space characters (as identified by the C macro @code{isspace()})
12001: are delimiters. @code{PARSE}, on the other hand, treats space like other
1.44 crook 12002: delimiters. @code{SWORD} treats space like @code{WORD}, but behaves
1.1 anton 12003: like @code{PARSE} otherwise. @code{(NAME)}, which is used by the outer
12004: interpreter (aka text interpreter) by default, treats all white-space
12005: characters as delimiters.
12006:
1.26 crook 12007: @item format of the control-flow stack:
12008: @cindex control-flow stack, format
12009: The data stack is used as control-flow stack. The size of a control-flow
1.1 anton 12010: stack item in cells is given by the constant @code{cs-item-size}. At the
12011: time of this writing, an item consists of a (pointer to a) locals list
12012: (third), an address in the code (second), and a tag for identifying the
12013: item (TOS). The following tags are used: @code{defstart},
12014: @code{live-orig}, @code{dead-orig}, @code{dest}, @code{do-dest},
12015: @code{scopestart}.
12016:
12017: @item conversion of digits > 35
12018: @cindex digits > 35
12019: The characters @code{[\]^_'} are the digits with the decimal value
12020: 36@minus{}41. There is no way to input many of the larger digits.
12021:
12022: @item display after input terminates in @code{ACCEPT} and @code{EXPECT}:
12023: @cindex @code{EXPECT}, display after end of input
12024: @cindex @code{ACCEPT}, display after end of input
12025: The cursor is moved to the end of the entered string. If the input is
12026: terminated using the @kbd{Return} key, a space is typed.
12027:
12028: @item exception abort sequence of @code{ABORT"}:
12029: @cindex exception abort sequence of @code{ABORT"}
12030: @cindex @code{ABORT"}, exception abort sequence
12031: The error string is stored into the variable @code{"error} and a
12032: @code{-2 throw} is performed.
12033:
12034: @item input line terminator:
12035: @cindex input line terminator
12036: @cindex line terminator on input
1.26 crook 12037: @cindex newline character on input
1.1 anton 12038: For interactive input, @kbd{C-m} (CR) and @kbd{C-j} (LF) terminate
12039: lines. One of these characters is typically produced when you type the
12040: @kbd{Enter} or @kbd{Return} key.
12041:
12042: @item maximum size of a counted string:
12043: @cindex maximum size of a counted string
12044: @cindex counted string, maximum size
12045: @code{s" /counted-string" environment? drop .}. Currently 255 characters
12046: on all ports, but this may change.
12047:
12048: @item maximum size of a parsed string:
12049: @cindex maximum size of a parsed string
12050: @cindex parsed string, maximum size
12051: Given by the constant @code{/line}. Currently 255 characters.
12052:
12053: @item maximum size of a definition name, in characters:
12054: @cindex maximum size of a definition name, in characters
12055: @cindex name, maximum length
12056: 31
12057:
12058: @item maximum string length for @code{ENVIRONMENT?}, in characters:
12059: @cindex maximum string length for @code{ENVIRONMENT?}, in characters
12060: @cindex @code{ENVIRONMENT?} string length, maximum
12061: 31
12062:
12063: @item method of selecting the user input device:
12064: @cindex user input device, method of selecting
12065: The user input device is the standard input. There is currently no way to
12066: change it from within Gforth. However, the input can typically be
12067: redirected in the command line that starts Gforth.
12068:
12069: @item method of selecting the user output device:
12070: @cindex user output device, method of selecting
12071: @code{EMIT} and @code{TYPE} output to the file-id stored in the value
1.10 anton 12072: @code{outfile-id} (@code{stdout} by default). Gforth uses unbuffered
12073: output when the user output device is a terminal, otherwise the output
12074: is buffered.
1.1 anton 12075:
12076: @item methods of dictionary compilation:
12077: What are we expected to document here?
12078:
12079: @item number of bits in one address unit:
12080: @cindex number of bits in one address unit
12081: @cindex address unit, size in bits
12082: @code{s" address-units-bits" environment? drop .}. 8 in all current
12083: ports.
12084:
12085: @item number representation and arithmetic:
12086: @cindex number representation and arithmetic
12087: Processor-dependent. Binary two's complement on all current ports.
12088:
12089: @item ranges for integer types:
12090: @cindex ranges for integer types
12091: @cindex integer types, ranges
12092: Installation-dependent. Make environmental queries for @code{MAX-N},
12093: @code{MAX-U}, @code{MAX-D} and @code{MAX-UD}. The lower bounds for
12094: unsigned (and positive) types is 0. The lower bound for signed types on
12095: two's complement and one's complement machines machines can be computed
12096: by adding 1 to the upper bound.
12097:
12098: @item read-only data space regions:
12099: @cindex read-only data space regions
12100: @cindex data-space, read-only regions
12101: The whole Forth data space is writable.
12102:
12103: @item size of buffer at @code{WORD}:
12104: @cindex size of buffer at @code{WORD}
12105: @cindex @code{WORD} buffer size
12106: @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
12107: shared with the pictured numeric output string. If overwriting
12108: @code{PAD} is acceptable, it is as large as the remaining dictionary
12109: space, although only as much can be sensibly used as fits in a counted
12110: string.
12111:
12112: @item size of one cell in address units:
12113: @cindex cell size
12114: @code{1 cells .}.
12115:
12116: @item size of one character in address units:
12117: @cindex char size
12118: @code{1 chars .}. 1 on all current ports.
12119:
12120: @item size of the keyboard terminal buffer:
12121: @cindex size of the keyboard terminal buffer
12122: @cindex terminal buffer, size
12123: Varies. You can determine the size at a specific time using @code{lp@@
12124: tib - .}. It is shared with the locals stack and TIBs of files that
12125: include the current file. You can change the amount of space for TIBs
12126: and locals stack at Gforth startup with the command line option
12127: @code{-l}.
12128:
12129: @item size of the pictured numeric output buffer:
12130: @cindex size of the pictured numeric output buffer
12131: @cindex pictured numeric output buffer, size
12132: @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
12133: shared with @code{WORD}.
12134:
12135: @item size of the scratch area returned by @code{PAD}:
12136: @cindex size of the scratch area returned by @code{PAD}
12137: @cindex @code{PAD} size
12138: The remainder of dictionary space. @code{unused pad here - - .}.
12139:
12140: @item system case-sensitivity characteristics:
12141: @cindex case-sensitivity characteristics
1.26 crook 12142: Dictionary searches are case-insensitive (except in
1.1 anton 12143: @code{TABLE}s). However, as explained above under @i{character-set
12144: extensions}, the matching for non-ASCII characters is determined by the
12145: locale you are using. In the default @code{C} locale all non-ASCII
12146: characters are matched case-sensitively.
12147:
12148: @item system prompt:
12149: @cindex system prompt
12150: @cindex prompt
12151: @code{ ok} in interpret state, @code{ compiled} in compile state.
12152:
12153: @item division rounding:
12154: @cindex division rounding
12155: installation dependent. @code{s" floored" environment? drop .}. We leave
12156: the choice to @code{gcc} (what to use for @code{/}) and to you (whether
12157: to use @code{fm/mod}, @code{sm/rem} or simply @code{/}).
12158:
12159: @item values of @code{STATE} when true:
12160: @cindex @code{STATE} values
12161: -1.
12162:
12163: @item values returned after arithmetic overflow:
12164: On two's complement machines, arithmetic is performed modulo
12165: 2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
12166: arithmetic (with appropriate mapping for signed types). Division by zero
12167: typically results in a @code{-55 throw} (Floating-point unidentified
12168: fault), although a @code{-10 throw} (divide by zero) would be more
12169: appropriate.
12170:
12171: @item whether the current definition can be found after @t{DOES>}:
12172: @cindex @t{DOES>}, visibility of current definition
12173: No.
12174:
12175: @end table
12176:
12177: @c ---------------------------------------------------------------------
12178: @node core-ambcond, core-other, core-idef, The Core Words
12179: @subsection Ambiguous conditions
12180: @c ---------------------------------------------------------------------
12181: @cindex core words, ambiguous conditions
12182: @cindex ambiguous conditions, core words
12183:
12184: @table @i
12185:
12186: @item a name is neither a word nor a number:
12187: @cindex name not found
1.26 crook 12188: @cindex undefined word
1.1 anton 12189: @code{-13 throw} (Undefined word). Actually, @code{-13 bounce}, which
12190: preserves the data and FP stack, so you don't lose more work than
12191: necessary.
12192:
12193: @item a definition name exceeds the maximum length allowed:
1.26 crook 12194: @cindex word name too long
1.1 anton 12195: @code{-19 throw} (Word name too long)
12196:
12197: @item addressing a region not inside the various data spaces of the forth system:
12198: @cindex Invalid memory address
1.32 anton 12199: The stacks, code space and header space are accessible. Machine code space is
1.1 anton 12200: typically readable. Accessing other addresses gives results dependent on
12201: the operating system. On decent systems: @code{-9 throw} (Invalid memory
12202: address).
12203:
12204: @item argument type incompatible with parameter:
1.26 crook 12205: @cindex argument type mismatch
1.1 anton 12206: This is usually not caught. Some words perform checks, e.g., the control
12207: flow words, and issue a @code{ABORT"} or @code{-12 THROW} (Argument type
12208: mismatch).
12209:
12210: @item attempting to obtain the execution token of a word with undefined execution semantics:
12211: @cindex Interpreting a compile-only word, for @code{'} etc.
12212: @cindex execution token of words with undefined execution semantics
12213: @code{-14 throw} (Interpreting a compile-only word). In some cases, you
12214: get an execution token for @code{compile-only-error} (which performs a
12215: @code{-14 throw} when executed).
12216:
12217: @item dividing by zero:
12218: @cindex dividing by zero
12219: @cindex floating point unidentified fault, integer division
1.24 anton 12220: On better platforms, this produces a @code{-10 throw} (Division by
12221: zero); on other systems, this typically results in a @code{-55 throw}
12222: (Floating-point unidentified fault).
1.1 anton 12223:
12224: @item insufficient data stack or return stack space:
12225: @cindex insufficient data stack or return stack space
12226: @cindex stack overflow
1.26 crook 12227: @cindex address alignment exception, stack overflow
1.1 anton 12228: @cindex Invalid memory address, stack overflow
12229: Depending on the operating system, the installation, and the invocation
12230: of Gforth, this is either checked by the memory management hardware, or
1.24 anton 12231: it is not checked. If it is checked, you typically get a @code{-3 throw}
12232: (Stack overflow), @code{-5 throw} (Return stack overflow), or @code{-9
12233: throw} (Invalid memory address) (depending on the platform and how you
12234: achieved the overflow) as soon as the overflow happens. If it is not
12235: checked, overflows typically result in mysterious illegal memory
12236: accesses, producing @code{-9 throw} (Invalid memory address) or
12237: @code{-23 throw} (Address alignment exception); they might also destroy
12238: the internal data structure of @code{ALLOCATE} and friends, resulting in
12239: various errors in these words.
1.1 anton 12240:
12241: @item insufficient space for loop control parameters:
12242: @cindex insufficient space for loop control parameters
12243: like other return stack overflows.
12244:
12245: @item insufficient space in the dictionary:
12246: @cindex insufficient space in the dictionary
12247: @cindex dictionary overflow
1.12 anton 12248: If you try to allot (either directly with @code{allot}, or indirectly
12249: with @code{,}, @code{create} etc.) more memory than available in the
12250: dictionary, you get a @code{-8 throw} (Dictionary overflow). If you try
12251: to access memory beyond the end of the dictionary, the results are
12252: similar to stack overflows.
1.1 anton 12253:
12254: @item interpreting a word with undefined interpretation semantics:
12255: @cindex interpreting a word with undefined interpretation semantics
12256: @cindex Interpreting a compile-only word
12257: For some words, we have defined interpretation semantics. For the
12258: others: @code{-14 throw} (Interpreting a compile-only word).
12259:
12260: @item modifying the contents of the input buffer or a string literal:
12261: @cindex modifying the contents of the input buffer or a string literal
12262: These are located in writable memory and can be modified.
12263:
12264: @item overflow of the pictured numeric output string:
12265: @cindex overflow of the pictured numeric output string
12266: @cindex pictured numeric output string, overflow
1.24 anton 12267: @code{-17 throw} (Pictured numeric ouput string overflow).
1.1 anton 12268:
12269: @item parsed string overflow:
12270: @cindex parsed string overflow
12271: @code{PARSE} cannot overflow. @code{WORD} does not check for overflow.
12272:
12273: @item producing a result out of range:
12274: @cindex result out of range
12275: On two's complement machines, arithmetic is performed modulo
12276: 2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
12277: arithmetic (with appropriate mapping for signed types). Division by zero
1.24 anton 12278: typically results in a @code{-10 throw} (divide by zero) or @code{-55
12279: throw} (floating point unidentified fault). @code{convert} and
12280: @code{>number} currently overflow silently.
1.1 anton 12281:
12282: @item reading from an empty data or return stack:
12283: @cindex stack empty
12284: @cindex stack underflow
1.24 anton 12285: @cindex return stack underflow
1.1 anton 12286: The data stack is checked by the outer (aka text) interpreter after
12287: every word executed. If it has underflowed, a @code{-4 throw} (Stack
12288: underflow) is performed. Apart from that, stacks may be checked or not,
1.24 anton 12289: depending on operating system, installation, and invocation. If they are
12290: caught by a check, they typically result in @code{-4 throw} (Stack
12291: underflow), @code{-6 throw} (Return stack underflow) or @code{-9 throw}
12292: (Invalid memory address), depending on the platform and which stack
12293: underflows and by how much. Note that even if the system uses checking
12294: (through the MMU), your program may have to underflow by a significant
12295: number of stack items to trigger the reaction (the reason for this is
12296: that the MMU, and therefore the checking, works with a page-size
12297: granularity). If there is no checking, the symptoms resulting from an
12298: underflow are similar to those from an overflow. Unbalanced return
12299: stack errors result in a variaty of symptoms, including @code{-9 throw}
12300: (Invalid memory address) and Illegal Instruction (typically @code{-260
12301: throw}).
1.1 anton 12302:
12303: @item unexpected end of the input buffer, resulting in an attempt to use a zero-length string as a name:
12304: @cindex unexpected end of the input buffer
12305: @cindex zero-length string as a name
12306: @cindex Attempt to use zero-length string as a name
12307: @code{Create} and its descendants perform a @code{-16 throw} (Attempt to
12308: use zero-length string as a name). Words like @code{'} probably will not
12309: find what they search. Note that it is possible to create zero-length
12310: names with @code{nextname} (should it not?).
12311:
12312: @item @code{>IN} greater than input buffer:
12313: @cindex @code{>IN} greater than input buffer
12314: The next invocation of a parsing word returns a string with length 0.
12315:
12316: @item @code{RECURSE} appears after @code{DOES>}:
12317: @cindex @code{RECURSE} appears after @code{DOES>}
12318: Compiles a recursive call to the defining word, not to the defined word.
12319:
12320: @item argument input source different than current input source for @code{RESTORE-INPUT}:
12321: @cindex argument input source different than current input source for @code{RESTORE-INPUT}
1.26 crook 12322: @cindex argument type mismatch, @code{RESTORE-INPUT}
1.1 anton 12323: @cindex @code{RESTORE-INPUT}, Argument type mismatch
12324: @code{-12 THROW}. Note that, once an input file is closed (e.g., because
12325: the end of the file was reached), its source-id may be
12326: reused. Therefore, restoring an input source specification referencing a
12327: closed file may lead to unpredictable results instead of a @code{-12
12328: THROW}.
12329:
12330: In the future, Gforth may be able to restore input source specifications
12331: from other than the current input source.
12332:
12333: @item data space containing definitions gets de-allocated:
12334: @cindex data space containing definitions gets de-allocated
12335: Deallocation with @code{allot} is not checked. This typically results in
12336: memory access faults or execution of illegal instructions.
12337:
12338: @item data space read/write with incorrect alignment:
12339: @cindex data space read/write with incorrect alignment
12340: @cindex alignment faults
1.26 crook 12341: @cindex address alignment exception
1.1 anton 12342: Processor-dependent. Typically results in a @code{-23 throw} (Address
1.12 anton 12343: alignment exception). Under Linux-Intel on a 486 or later processor with
1.1 anton 12344: alignment turned on, incorrect alignment results in a @code{-9 throw}
12345: (Invalid memory address). There are reportedly some processors with
1.12 anton 12346: alignment restrictions that do not report violations.
1.1 anton 12347:
12348: @item data space pointer not properly aligned, @code{,}, @code{C,}:
12349: @cindex data space pointer not properly aligned, @code{,}, @code{C,}
12350: Like other alignment errors.
12351:
12352: @item less than u+2 stack items (@code{PICK} and @code{ROLL}):
12353: Like other stack underflows.
12354:
12355: @item loop control parameters not available:
12356: @cindex loop control parameters not available
12357: Not checked. The counted loop words simply assume that the top of return
12358: stack items are loop control parameters and behave accordingly.
12359:
12360: @item most recent definition does not have a name (@code{IMMEDIATE}):
12361: @cindex most recent definition does not have a name (@code{IMMEDIATE})
12362: @cindex last word was headerless
12363: @code{abort" last word was headerless"}.
12364:
12365: @item name not defined by @code{VALUE} used by @code{TO}:
12366: @cindex name not defined by @code{VALUE} used by @code{TO}
12367: @cindex @code{TO} on non-@code{VALUE}s
12368: @cindex Invalid name argument, @code{TO}
12369: @code{-32 throw} (Invalid name argument) (unless name is a local or was
12370: defined by @code{CONSTANT}; in the latter case it just changes the constant).
12371:
12372: @item name not found (@code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]}):
12373: @cindex name not found (@code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]})
1.26 crook 12374: @cindex undefined word, @code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]}
1.1 anton 12375: @code{-13 throw} (Undefined word)
12376:
12377: @item parameters are not of the same type (@code{DO}, @code{?DO}, @code{WITHIN}):
12378: @cindex parameters are not of the same type (@code{DO}, @code{?DO}, @code{WITHIN})
12379: Gforth behaves as if they were of the same type. I.e., you can predict
12380: the behaviour by interpreting all parameters as, e.g., signed.
12381:
12382: @item @code{POSTPONE} or @code{[COMPILE]} applied to @code{TO}:
12383: @cindex @code{POSTPONE} or @code{[COMPILE]} applied to @code{TO}
12384: Assume @code{: X POSTPONE TO ; IMMEDIATE}. @code{X} performs the
12385: compilation semantics of @code{TO}.
12386:
12387: @item String longer than a counted string returned by @code{WORD}:
1.26 crook 12388: @cindex string longer than a counted string returned by @code{WORD}
1.1 anton 12389: @cindex @code{WORD}, string overflow
12390: Not checked. The string will be ok, but the count will, of course,
12391: contain only the least significant bits of the length.
12392:
12393: @item u greater than or equal to the number of bits in a cell (@code{LSHIFT}, @code{RSHIFT}):
12394: @cindex @code{LSHIFT}, large shift counts
12395: @cindex @code{RSHIFT}, large shift counts
12396: Processor-dependent. Typical behaviours are returning 0 and using only
12397: the low bits of the shift count.
12398:
12399: @item word not defined via @code{CREATE}:
12400: @cindex @code{>BODY} of non-@code{CREATE}d words
12401: @code{>BODY} produces the PFA of the word no matter how it was defined.
12402:
12403: @cindex @code{DOES>} of non-@code{CREATE}d words
12404: @code{DOES>} changes the execution semantics of the last defined word no
12405: matter how it was defined. E.g., @code{CONSTANT DOES>} is equivalent to
12406: @code{CREATE , DOES>}.
12407:
12408: @item words improperly used outside @code{<#} and @code{#>}:
12409: Not checked. As usual, you can expect memory faults.
12410:
12411: @end table
12412:
12413:
12414: @c ---------------------------------------------------------------------
12415: @node core-other, , core-ambcond, The Core Words
12416: @subsection Other system documentation
12417: @c ---------------------------------------------------------------------
12418: @cindex other system documentation, core words
12419: @cindex core words, other system documentation
12420:
12421: @table @i
12422: @item nonstandard words using @code{PAD}:
12423: @cindex @code{PAD} use by nonstandard words
12424: None.
12425:
12426: @item operator's terminal facilities available:
12427: @cindex operator's terminal facilities available
12428: After processing the command line, Gforth goes into interactive mode,
12429: and you can give commands to Gforth interactively. The actual facilities
12430: available depend on how you invoke Gforth.
12431:
12432: @item program data space available:
12433: @cindex program data space available
12434: @cindex data space available
12435: @code{UNUSED .} gives the remaining dictionary space. The total
12436: dictionary space can be specified with the @code{-m} switch
12437: (@pxref{Invoking Gforth}) when Gforth starts up.
12438:
12439: @item return stack space available:
12440: @cindex return stack space available
12441: You can compute the total return stack space in cells with
12442: @code{s" RETURN-STACK-CELLS" environment? drop .}. You can specify it at
12443: startup time with the @code{-r} switch (@pxref{Invoking Gforth}).
12444:
12445: @item stack space available:
12446: @cindex stack space available
12447: You can compute the total data stack space in cells with
12448: @code{s" STACK-CELLS" environment? drop .}. You can specify it at
12449: startup time with the @code{-d} switch (@pxref{Invoking Gforth}).
12450:
12451: @item system dictionary space required, in address units:
12452: @cindex system dictionary space required, in address units
12453: Type @code{here forthstart - .} after startup. At the time of this
12454: writing, this gives 80080 (bytes) on a 32-bit system.
12455: @end table
12456:
12457:
12458: @c =====================================================================
12459: @node The optional Block word set, The optional Double Number word set, The Core Words, ANS conformance
12460: @section The optional Block word set
12461: @c =====================================================================
12462: @cindex system documentation, block words
12463: @cindex block words, system documentation
12464:
12465: @menu
12466: * block-idef:: Implementation Defined Options
12467: * block-ambcond:: Ambiguous Conditions
12468: * block-other:: Other System Documentation
12469: @end menu
12470:
12471:
12472: @c ---------------------------------------------------------------------
12473: @node block-idef, block-ambcond, The optional Block word set, The optional Block word set
12474: @subsection Implementation Defined Options
12475: @c ---------------------------------------------------------------------
12476: @cindex implementation-defined options, block words
12477: @cindex block words, implementation-defined options
12478:
12479: @table @i
12480: @item the format for display by @code{LIST}:
12481: @cindex @code{LIST} display format
12482: First the screen number is displayed, then 16 lines of 64 characters,
12483: each line preceded by the line number.
12484:
12485: @item the length of a line affected by @code{\}:
12486: @cindex length of a line affected by @code{\}
12487: @cindex @code{\}, line length in blocks
12488: 64 characters.
12489: @end table
12490:
12491:
12492: @c ---------------------------------------------------------------------
12493: @node block-ambcond, block-other, block-idef, The optional Block word set
12494: @subsection Ambiguous conditions
12495: @c ---------------------------------------------------------------------
12496: @cindex block words, ambiguous conditions
12497: @cindex ambiguous conditions, block words
12498:
12499: @table @i
12500: @item correct block read was not possible:
12501: @cindex block read not possible
12502: Typically results in a @code{throw} of some OS-derived value (between
12503: -512 and -2048). If the blocks file was just not long enough, blanks are
12504: supplied for the missing portion.
12505:
12506: @item I/O exception in block transfer:
12507: @cindex I/O exception in block transfer
12508: @cindex block transfer, I/O exception
12509: Typically results in a @code{throw} of some OS-derived value (between
12510: -512 and -2048).
12511:
12512: @item invalid block number:
12513: @cindex invalid block number
12514: @cindex block number invalid
12515: @code{-35 throw} (Invalid block number)
12516:
12517: @item a program directly alters the contents of @code{BLK}:
12518: @cindex @code{BLK}, altering @code{BLK}
12519: The input stream is switched to that other block, at the same
12520: position. If the storing to @code{BLK} happens when interpreting
12521: non-block input, the system will get quite confused when the block ends.
12522:
12523: @item no current block buffer for @code{UPDATE}:
12524: @cindex @code{UPDATE}, no current block buffer
12525: @code{UPDATE} has no effect.
12526:
12527: @end table
12528:
12529: @c ---------------------------------------------------------------------
12530: @node block-other, , block-ambcond, The optional Block word set
12531: @subsection Other system documentation
12532: @c ---------------------------------------------------------------------
12533: @cindex other system documentation, block words
12534: @cindex block words, other system documentation
12535:
12536: @table @i
12537: @item any restrictions a multiprogramming system places on the use of buffer addresses:
12538: No restrictions (yet).
12539:
12540: @item the number of blocks available for source and data:
12541: depends on your disk space.
12542:
12543: @end table
12544:
12545:
12546: @c =====================================================================
12547: @node The optional Double Number word set, The optional Exception word set, The optional Block word set, ANS conformance
12548: @section The optional Double Number word set
12549: @c =====================================================================
12550: @cindex system documentation, double words
12551: @cindex double words, system documentation
12552:
12553: @menu
12554: * double-ambcond:: Ambiguous Conditions
12555: @end menu
12556:
12557:
12558: @c ---------------------------------------------------------------------
12559: @node double-ambcond, , The optional Double Number word set, The optional Double Number word set
12560: @subsection Ambiguous conditions
12561: @c ---------------------------------------------------------------------
12562: @cindex double words, ambiguous conditions
12563: @cindex ambiguous conditions, double words
12564:
12565: @table @i
1.29 crook 12566: @item @i{d} outside of range of @i{n} in @code{D>S}:
12567: @cindex @code{D>S}, @i{d} out of range of @i{n}
12568: The least significant cell of @i{d} is produced.
1.1 anton 12569:
12570: @end table
12571:
12572:
12573: @c =====================================================================
12574: @node The optional Exception word set, The optional Facility word set, The optional Double Number word set, ANS conformance
12575: @section The optional Exception word set
12576: @c =====================================================================
12577: @cindex system documentation, exception words
12578: @cindex exception words, system documentation
12579:
12580: @menu
12581: * exception-idef:: Implementation Defined Options
12582: @end menu
12583:
12584:
12585: @c ---------------------------------------------------------------------
12586: @node exception-idef, , The optional Exception word set, The optional Exception word set
12587: @subsection Implementation Defined Options
12588: @c ---------------------------------------------------------------------
12589: @cindex implementation-defined options, exception words
12590: @cindex exception words, implementation-defined options
12591:
12592: @table @i
12593: @item @code{THROW}-codes used in the system:
12594: @cindex @code{THROW}-codes used in the system
12595: The codes -256@minus{}-511 are used for reporting signals. The mapping
1.29 crook 12596: from OS signal numbers to throw codes is -256@minus{}@i{signal}. The
1.1 anton 12597: codes -512@minus{}-2047 are used for OS errors (for file and memory
12598: allocation operations). The mapping from OS error numbers to throw codes
12599: is -512@minus{}@code{errno}. One side effect of this mapping is that
12600: undefined OS errors produce a message with a strange number; e.g.,
12601: @code{-1000 THROW} results in @code{Unknown error 488} on my system.
12602: @end table
12603:
12604: @c =====================================================================
12605: @node The optional Facility word set, The optional File-Access word set, The optional Exception word set, ANS conformance
12606: @section The optional Facility word set
12607: @c =====================================================================
12608: @cindex system documentation, facility words
12609: @cindex facility words, system documentation
12610:
12611: @menu
12612: * facility-idef:: Implementation Defined Options
12613: * facility-ambcond:: Ambiguous Conditions
12614: @end menu
12615:
12616:
12617: @c ---------------------------------------------------------------------
12618: @node facility-idef, facility-ambcond, The optional Facility word set, The optional Facility word set
12619: @subsection Implementation Defined Options
12620: @c ---------------------------------------------------------------------
12621: @cindex implementation-defined options, facility words
12622: @cindex facility words, implementation-defined options
12623:
12624: @table @i
12625: @item encoding of keyboard events (@code{EKEY}):
12626: @cindex keyboard events, encoding in @code{EKEY}
12627: @cindex @code{EKEY}, encoding of keyboard events
1.40 anton 12628: Keys corresponding to ASCII characters are encoded as ASCII characters.
1.41 anton 12629: Other keys are encoded with the constants @code{k-left}, @code{k-right},
12630: @code{k-up}, @code{k-down}, @code{k-home}, @code{k-end}, @code{k1},
12631: @code{k2}, @code{k3}, @code{k4}, @code{k5}, @code{k6}, @code{k7},
12632: @code{k8}, @code{k9}, @code{k10}, @code{k11}, @code{k12}.
1.40 anton 12633:
1.1 anton 12634:
12635: @item duration of a system clock tick:
12636: @cindex duration of a system clock tick
12637: @cindex clock tick duration
12638: System dependent. With respect to @code{MS}, the time is specified in
12639: microseconds. How well the OS and the hardware implement this, is
12640: another question.
12641:
12642: @item repeatability to be expected from the execution of @code{MS}:
12643: @cindex repeatability to be expected from the execution of @code{MS}
12644: @cindex @code{MS}, repeatability to be expected
12645: System dependent. On Unix, a lot depends on load. If the system is
12646: lightly loaded, and the delay is short enough that Gforth does not get
12647: swapped out, the performance should be acceptable. Under MS-DOS and
12648: other single-tasking systems, it should be good.
12649:
12650: @end table
12651:
12652:
12653: @c ---------------------------------------------------------------------
12654: @node facility-ambcond, , facility-idef, The optional Facility word set
12655: @subsection Ambiguous conditions
12656: @c ---------------------------------------------------------------------
12657: @cindex facility words, ambiguous conditions
12658: @cindex ambiguous conditions, facility words
12659:
12660: @table @i
12661: @item @code{AT-XY} can't be performed on user output device:
12662: @cindex @code{AT-XY} can't be performed on user output device
12663: Largely terminal dependent. No range checks are done on the arguments.
12664: No errors are reported. You may see some garbage appearing, you may see
12665: simply nothing happen.
12666:
12667: @end table
12668:
12669:
12670: @c =====================================================================
12671: @node The optional File-Access word set, The optional Floating-Point word set, The optional Facility word set, ANS conformance
12672: @section The optional File-Access word set
12673: @c =====================================================================
12674: @cindex system documentation, file words
12675: @cindex file words, system documentation
12676:
12677: @menu
12678: * file-idef:: Implementation Defined Options
12679: * file-ambcond:: Ambiguous Conditions
12680: @end menu
12681:
12682: @c ---------------------------------------------------------------------
12683: @node file-idef, file-ambcond, The optional File-Access word set, The optional File-Access word set
12684: @subsection Implementation Defined Options
12685: @c ---------------------------------------------------------------------
12686: @cindex implementation-defined options, file words
12687: @cindex file words, implementation-defined options
12688:
12689: @table @i
12690: @item file access methods used:
12691: @cindex file access methods used
12692: @code{R/O}, @code{R/W} and @code{BIN} work as you would
12693: expect. @code{W/O} translates into the C file opening mode @code{w} (or
12694: @code{wb}): The file is cleared, if it exists, and created, if it does
12695: not (with both @code{open-file} and @code{create-file}). Under Unix
12696: @code{create-file} creates a file with 666 permissions modified by your
12697: umask.
12698:
12699: @item file exceptions:
12700: @cindex file exceptions
12701: The file words do not raise exceptions (except, perhaps, memory access
12702: faults when you pass illegal addresses or file-ids).
12703:
12704: @item file line terminator:
12705: @cindex file line terminator
12706: System-dependent. Gforth uses C's newline character as line
12707: terminator. What the actual character code(s) of this are is
12708: system-dependent.
12709:
12710: @item file name format:
12711: @cindex file name format
12712: System dependent. Gforth just uses the file name format of your OS.
12713:
12714: @item information returned by @code{FILE-STATUS}:
12715: @cindex @code{FILE-STATUS}, returned information
12716: @code{FILE-STATUS} returns the most powerful file access mode allowed
12717: for the file: Either @code{R/O}, @code{W/O} or @code{R/W}. If the file
12718: cannot be accessed, @code{R/O BIN} is returned. @code{BIN} is applicable
12719: along with the returned mode.
12720:
12721: @item input file state after an exception when including source:
12722: @cindex exception when including source
12723: All files that are left via the exception are closed.
12724:
1.29 crook 12725: @item @i{ior} values and meaning:
12726: @cindex @i{ior} values and meaning
1.68 anton 12727: @cindex @i{wior} values and meaning
1.29 crook 12728: The @i{ior}s returned by the file and memory allocation words are
1.1 anton 12729: intended as throw codes. They typically are in the range
12730: -512@minus{}-2047 of OS errors. The mapping from OS error numbers to
1.29 crook 12731: @i{ior}s is -512@minus{}@i{errno}.
1.1 anton 12732:
12733: @item maximum depth of file input nesting:
12734: @cindex maximum depth of file input nesting
12735: @cindex file input nesting, maximum depth
12736: limited by the amount of return stack, locals/TIB stack, and the number
12737: of open files available. This should not give you troubles.
12738:
12739: @item maximum size of input line:
12740: @cindex maximum size of input line
12741: @cindex input line size, maximum
12742: @code{/line}. Currently 255.
12743:
12744: @item methods of mapping block ranges to files:
12745: @cindex mapping block ranges to files
12746: @cindex files containing blocks
12747: @cindex blocks in files
12748: By default, blocks are accessed in the file @file{blocks.fb} in the
12749: current working directory. The file can be switched with @code{USE}.
12750:
12751: @item number of string buffers provided by @code{S"}:
12752: @cindex @code{S"}, number of string buffers
12753: 1
12754:
12755: @item size of string buffer used by @code{S"}:
12756: @cindex @code{S"}, size of string buffer
12757: @code{/line}. currently 255.
12758:
12759: @end table
12760:
12761: @c ---------------------------------------------------------------------
12762: @node file-ambcond, , file-idef, The optional File-Access word set
12763: @subsection Ambiguous conditions
12764: @c ---------------------------------------------------------------------
12765: @cindex file words, ambiguous conditions
12766: @cindex ambiguous conditions, file words
12767:
12768: @table @i
12769: @item attempting to position a file outside its boundaries:
12770: @cindex @code{REPOSITION-FILE}, outside the file's boundaries
12771: @code{REPOSITION-FILE} is performed as usual: Afterwards,
12772: @code{FILE-POSITION} returns the value given to @code{REPOSITION-FILE}.
12773:
12774: @item attempting to read from file positions not yet written:
12775: @cindex reading from file positions not yet written
12776: End-of-file, i.e., zero characters are read and no error is reported.
12777:
1.29 crook 12778: @item @i{file-id} is invalid (@code{INCLUDE-FILE}):
12779: @cindex @code{INCLUDE-FILE}, @i{file-id} is invalid
1.1 anton 12780: An appropriate exception may be thrown, but a memory fault or other
12781: problem is more probable.
12782:
1.29 crook 12783: @item I/O exception reading or closing @i{file-id} (@code{INCLUDE-FILE}, @code{INCLUDED}):
12784: @cindex @code{INCLUDE-FILE}, I/O exception reading or closing @i{file-id}
12785: @cindex @code{INCLUDED}, I/O exception reading or closing @i{file-id}
12786: The @i{ior} produced by the operation, that discovered the problem, is
1.1 anton 12787: thrown.
12788:
12789: @item named file cannot be opened (@code{INCLUDED}):
12790: @cindex @code{INCLUDED}, named file cannot be opened
1.29 crook 12791: The @i{ior} produced by @code{open-file} is thrown.
1.1 anton 12792:
12793: @item requesting an unmapped block number:
12794: @cindex unmapped block numbers
12795: There are no unmapped legal block numbers. On some operating systems,
12796: writing a block with a large number may overflow the file system and
12797: have an error message as consequence.
12798:
12799: @item using @code{source-id} when @code{blk} is non-zero:
12800: @cindex @code{SOURCE-ID}, behaviour when @code{BLK} is non-zero
12801: @code{source-id} performs its function. Typically it will give the id of
12802: the source which loaded the block. (Better ideas?)
12803:
12804: @end table
12805:
12806:
12807: @c =====================================================================
12808: @node The optional Floating-Point word set, The optional Locals word set, The optional File-Access word set, ANS conformance
12809: @section The optional Floating-Point word set
12810: @c =====================================================================
12811: @cindex system documentation, floating-point words
12812: @cindex floating-point words, system documentation
12813:
12814: @menu
12815: * floating-idef:: Implementation Defined Options
12816: * floating-ambcond:: Ambiguous Conditions
12817: @end menu
12818:
12819:
12820: @c ---------------------------------------------------------------------
12821: @node floating-idef, floating-ambcond, The optional Floating-Point word set, The optional Floating-Point word set
12822: @subsection Implementation Defined Options
12823: @c ---------------------------------------------------------------------
12824: @cindex implementation-defined options, floating-point words
12825: @cindex floating-point words, implementation-defined options
12826:
12827: @table @i
12828: @item format and range of floating point numbers:
12829: @cindex format and range of floating point numbers
12830: @cindex floating point numbers, format and range
12831: System-dependent; the @code{double} type of C.
12832:
1.29 crook 12833: @item results of @code{REPRESENT} when @i{float} is out of range:
12834: @cindex @code{REPRESENT}, results when @i{float} is out of range
1.1 anton 12835: System dependent; @code{REPRESENT} is implemented using the C library
12836: function @code{ecvt()} and inherits its behaviour in this respect.
12837:
12838: @item rounding or truncation of floating-point numbers:
12839: @cindex rounding of floating-point numbers
12840: @cindex truncation of floating-point numbers
12841: @cindex floating-point numbers, rounding or truncation
12842: System dependent; the rounding behaviour is inherited from the hosting C
12843: compiler. IEEE-FP-based (i.e., most) systems by default round to
12844: nearest, and break ties by rounding to even (i.e., such that the last
12845: bit of the mantissa is 0).
12846:
12847: @item size of floating-point stack:
12848: @cindex floating-point stack size
12849: @code{s" FLOATING-STACK" environment? drop .} gives the total size of
12850: the floating-point stack (in floats). You can specify this on startup
12851: with the command-line option @code{-f} (@pxref{Invoking Gforth}).
12852:
12853: @item width of floating-point stack:
12854: @cindex floating-point stack width
12855: @code{1 floats}.
12856:
12857: @end table
12858:
12859:
12860: @c ---------------------------------------------------------------------
12861: @node floating-ambcond, , floating-idef, The optional Floating-Point word set
12862: @subsection Ambiguous conditions
12863: @c ---------------------------------------------------------------------
12864: @cindex floating-point words, ambiguous conditions
12865: @cindex ambiguous conditions, floating-point words
12866:
12867: @table @i
12868: @item @code{df@@} or @code{df!} used with an address that is not double-float aligned:
12869: @cindex @code{df@@} or @code{df!} used with an address that is not double-float aligned
12870: System-dependent. Typically results in a @code{-23 THROW} like other
12871: alignment violations.
12872:
12873: @item @code{f@@} or @code{f!} used with an address that is not float aligned:
12874: @cindex @code{f@@} used with an address that is not float aligned
12875: @cindex @code{f!} used with an address that is not float aligned
12876: System-dependent. Typically results in a @code{-23 THROW} like other
12877: alignment violations.
12878:
12879: @item floating-point result out of range:
12880: @cindex floating-point result out of range
12881: System-dependent. Can result in a @code{-55 THROW} (Floating-point
12882: unidentified fault), or can produce a special value representing, e.g.,
12883: Infinity.
12884:
12885: @item @code{sf@@} or @code{sf!} used with an address that is not single-float aligned:
12886: @cindex @code{sf@@} or @code{sf!} used with an address that is not single-float aligned
12887: System-dependent. Typically results in an alignment fault like other
12888: alignment violations.
12889:
1.35 anton 12890: @item @code{base} is not decimal (@code{REPRESENT}, @code{F.}, @code{FE.}, @code{FS.}):
12891: @cindex @code{base} is not decimal (@code{REPRESENT}, @code{F.}, @code{FE.}, @code{FS.})
1.1 anton 12892: The floating-point number is converted into decimal nonetheless.
12893:
12894: @item Both arguments are equal to zero (@code{FATAN2}):
12895: @cindex @code{FATAN2}, both arguments are equal to zero
12896: System-dependent. @code{FATAN2} is implemented using the C library
12897: function @code{atan2()}.
12898:
1.29 crook 12899: @item Using @code{FTAN} on an argument @i{r1} where cos(@i{r1}) is zero:
12900: @cindex @code{FTAN} on an argument @i{r1} where cos(@i{r1}) is zero
12901: System-dependent. Anyway, typically the cos of @i{r1} will not be zero
1.1 anton 12902: because of small errors and the tan will be a very large (or very small)
12903: but finite number.
12904:
1.29 crook 12905: @item @i{d} cannot be presented precisely as a float in @code{D>F}:
12906: @cindex @code{D>F}, @i{d} cannot be presented precisely as a float
1.1 anton 12907: The result is rounded to the nearest float.
12908:
12909: @item dividing by zero:
12910: @cindex dividing by zero, floating-point
12911: @cindex floating-point dividing by zero
12912: @cindex floating-point unidentified fault, FP divide-by-zero
12913: @code{-55 throw} (Floating-point unidentified fault)
12914:
12915: @item exponent too big for conversion (@code{DF!}, @code{DF@@}, @code{SF!}, @code{SF@@}):
12916: @cindex exponent too big for conversion (@code{DF!}, @code{DF@@}, @code{SF!}, @code{SF@@})
12917: System dependent. On IEEE-FP based systems the number is converted into
12918: an infinity.
12919:
1.29 crook 12920: @item @i{float}<1 (@code{FACOSH}):
12921: @cindex @code{FACOSH}, @i{float}<1
1.1 anton 12922: @cindex floating-point unidentified fault, @code{FACOSH}
12923: @code{-55 throw} (Floating-point unidentified fault)
12924:
1.29 crook 12925: @item @i{float}=<-1 (@code{FLNP1}):
12926: @cindex @code{FLNP1}, @i{float}=<-1
1.1 anton 12927: @cindex floating-point unidentified fault, @code{FLNP1}
12928: @code{-55 throw} (Floating-point unidentified fault). On IEEE-FP systems
1.29 crook 12929: negative infinity is typically produced for @i{float}=-1.
1.1 anton 12930:
1.29 crook 12931: @item @i{float}=<0 (@code{FLN}, @code{FLOG}):
12932: @cindex @code{FLN}, @i{float}=<0
12933: @cindex @code{FLOG}, @i{float}=<0
1.1 anton 12934: @cindex floating-point unidentified fault, @code{FLN} or @code{FLOG}
12935: @code{-55 throw} (Floating-point unidentified fault). On IEEE-FP systems
1.29 crook 12936: negative infinity is typically produced for @i{float}=0.
1.1 anton 12937:
1.29 crook 12938: @item @i{float}<0 (@code{FASINH}, @code{FSQRT}):
12939: @cindex @code{FASINH}, @i{float}<0
12940: @cindex @code{FSQRT}, @i{float}<0
1.1 anton 12941: @cindex floating-point unidentified fault, @code{FASINH} or @code{FSQRT}
12942: @code{-55 throw} (Floating-point unidentified fault). @code{fasinh}
12943: produces values for these inputs on my Linux box (Bug in the C library?)
12944:
1.29 crook 12945: @item |@i{float}|>1 (@code{FACOS}, @code{FASIN}, @code{FATANH}):
12946: @cindex @code{FACOS}, |@i{float}|>1
12947: @cindex @code{FASIN}, |@i{float}|>1
12948: @cindex @code{FATANH}, |@i{float}|>1
1.1 anton 12949: @cindex floating-point unidentified fault, @code{FACOS}, @code{FASIN} or @code{FATANH}
12950: @code{-55 throw} (Floating-point unidentified fault).
12951:
1.29 crook 12952: @item integer part of float cannot be represented by @i{d} in @code{F>D}:
12953: @cindex @code{F>D}, integer part of float cannot be represented by @i{d}
1.1 anton 12954: @cindex floating-point unidentified fault, @code{F>D}
12955: @code{-55 throw} (Floating-point unidentified fault).
12956:
12957: @item string larger than pictured numeric output area (@code{f.}, @code{fe.}, @code{fs.}):
12958: @cindex string larger than pictured numeric output area (@code{f.}, @code{fe.}, @code{fs.})
12959: This does not happen.
12960: @end table
12961:
12962: @c =====================================================================
12963: @node The optional Locals word set, The optional Memory-Allocation word set, The optional Floating-Point word set, ANS conformance
12964: @section The optional Locals word set
12965: @c =====================================================================
12966: @cindex system documentation, locals words
12967: @cindex locals words, system documentation
12968:
12969: @menu
12970: * locals-idef:: Implementation Defined Options
12971: * locals-ambcond:: Ambiguous Conditions
12972: @end menu
12973:
12974:
12975: @c ---------------------------------------------------------------------
12976: @node locals-idef, locals-ambcond, The optional Locals word set, The optional Locals word set
12977: @subsection Implementation Defined Options
12978: @c ---------------------------------------------------------------------
12979: @cindex implementation-defined options, locals words
12980: @cindex locals words, implementation-defined options
12981:
12982: @table @i
12983: @item maximum number of locals in a definition:
12984: @cindex maximum number of locals in a definition
12985: @cindex locals, maximum number in a definition
12986: @code{s" #locals" environment? drop .}. Currently 15. This is a lower
12987: bound, e.g., on a 32-bit machine there can be 41 locals of up to 8
12988: characters. The number of locals in a definition is bounded by the size
12989: of locals-buffer, which contains the names of the locals.
12990:
12991: @end table
12992:
12993:
12994: @c ---------------------------------------------------------------------
12995: @node locals-ambcond, , locals-idef, The optional Locals word set
12996: @subsection Ambiguous conditions
12997: @c ---------------------------------------------------------------------
12998: @cindex locals words, ambiguous conditions
12999: @cindex ambiguous conditions, locals words
13000:
13001: @table @i
13002: @item executing a named local in interpretation state:
13003: @cindex local in interpretation state
13004: @cindex Interpreting a compile-only word, for a local
13005: Locals have no interpretation semantics. If you try to perform the
13006: interpretation semantics, you will get a @code{-14 throw} somewhere
13007: (Interpreting a compile-only word). If you perform the compilation
13008: semantics, the locals access will be compiled (irrespective of state).
13009:
1.29 crook 13010: @item @i{name} not defined by @code{VALUE} or @code{(LOCAL)} (@code{TO}):
1.1 anton 13011: @cindex name not defined by @code{VALUE} or @code{(LOCAL)} used by @code{TO}
13012: @cindex @code{TO} on non-@code{VALUE}s and non-locals
13013: @cindex Invalid name argument, @code{TO}
13014: @code{-32 throw} (Invalid name argument)
13015:
13016: @end table
13017:
13018:
13019: @c =====================================================================
13020: @node The optional Memory-Allocation word set, The optional Programming-Tools word set, The optional Locals word set, ANS conformance
13021: @section The optional Memory-Allocation word set
13022: @c =====================================================================
13023: @cindex system documentation, memory-allocation words
13024: @cindex memory-allocation words, system documentation
13025:
13026: @menu
13027: * memory-idef:: Implementation Defined Options
13028: @end menu
13029:
13030:
13031: @c ---------------------------------------------------------------------
13032: @node memory-idef, , The optional Memory-Allocation word set, The optional Memory-Allocation word set
13033: @subsection Implementation Defined Options
13034: @c ---------------------------------------------------------------------
13035: @cindex implementation-defined options, memory-allocation words
13036: @cindex memory-allocation words, implementation-defined options
13037:
13038: @table @i
1.29 crook 13039: @item values and meaning of @i{ior}:
13040: @cindex @i{ior} values and meaning
13041: The @i{ior}s returned by the file and memory allocation words are
1.1 anton 13042: intended as throw codes. They typically are in the range
13043: -512@minus{}-2047 of OS errors. The mapping from OS error numbers to
1.29 crook 13044: @i{ior}s is -512@minus{}@i{errno}.
1.1 anton 13045:
13046: @end table
13047:
13048: @c =====================================================================
13049: @node The optional Programming-Tools word set, The optional Search-Order word set, The optional Memory-Allocation word set, ANS conformance
13050: @section The optional Programming-Tools word set
13051: @c =====================================================================
13052: @cindex system documentation, programming-tools words
13053: @cindex programming-tools words, system documentation
13054:
13055: @menu
13056: * programming-idef:: Implementation Defined Options
13057: * programming-ambcond:: Ambiguous Conditions
13058: @end menu
13059:
13060:
13061: @c ---------------------------------------------------------------------
13062: @node programming-idef, programming-ambcond, The optional Programming-Tools word set, The optional Programming-Tools word set
13063: @subsection Implementation Defined Options
13064: @c ---------------------------------------------------------------------
13065: @cindex implementation-defined options, programming-tools words
13066: @cindex programming-tools words, implementation-defined options
13067:
13068: @table @i
13069: @item ending sequence for input following @code{;CODE} and @code{CODE}:
13070: @cindex @code{;CODE} ending sequence
13071: @cindex @code{CODE} ending sequence
13072: @code{END-CODE}
13073:
13074: @item manner of processing input following @code{;CODE} and @code{CODE}:
13075: @cindex @code{;CODE}, processing input
13076: @cindex @code{CODE}, processing input
13077: The @code{ASSEMBLER} vocabulary is pushed on the search order stack, and
13078: the input is processed by the text interpreter, (starting) in interpret
13079: state.
13080:
13081: @item search order capability for @code{EDITOR} and @code{ASSEMBLER}:
13082: @cindex @code{ASSEMBLER}, search order capability
13083: The ANS Forth search order word set.
13084:
13085: @item source and format of display by @code{SEE}:
13086: @cindex @code{SEE}, source and format of output
13087: The source for @code{see} is the intermediate code used by the inner
13088: interpreter. The current @code{see} tries to output Forth source code
13089: as well as possible.
13090:
13091: @end table
13092:
13093: @c ---------------------------------------------------------------------
13094: @node programming-ambcond, , programming-idef, The optional Programming-Tools word set
13095: @subsection Ambiguous conditions
13096: @c ---------------------------------------------------------------------
13097: @cindex programming-tools words, ambiguous conditions
13098: @cindex ambiguous conditions, programming-tools words
13099:
13100: @table @i
13101:
1.21 crook 13102: @item deleting the compilation word list (@code{FORGET}):
13103: @cindex @code{FORGET}, deleting the compilation word list
1.1 anton 13104: Not implemented (yet).
13105:
1.29 crook 13106: @item fewer than @i{u}+1 items on the control-flow stack (@code{CS-PICK}, @code{CS-ROLL}):
13107: @cindex @code{CS-PICK}, fewer than @i{u}+1 items on the control flow-stack
13108: @cindex @code{CS-ROLL}, fewer than @i{u}+1 items on the control flow-stack
1.1 anton 13109: @cindex control-flow stack underflow
13110: This typically results in an @code{abort"} with a descriptive error
13111: message (may change into a @code{-22 throw} (Control structure mismatch)
13112: in the future). You may also get a memory access error. If you are
13113: unlucky, this ambiguous condition is not caught.
13114:
1.29 crook 13115: @item @i{name} can't be found (@code{FORGET}):
13116: @cindex @code{FORGET}, @i{name} can't be found
1.1 anton 13117: Not implemented (yet).
13118:
1.29 crook 13119: @item @i{name} not defined via @code{CREATE}:
13120: @cindex @code{;CODE}, @i{name} not defined via @code{CREATE}
1.1 anton 13121: @code{;CODE} behaves like @code{DOES>} in this respect, i.e., it changes
13122: the execution semantics of the last defined word no matter how it was
13123: defined.
13124:
13125: @item @code{POSTPONE} applied to @code{[IF]}:
13126: @cindex @code{POSTPONE} applied to @code{[IF]}
13127: @cindex @code{[IF]} and @code{POSTPONE}
13128: After defining @code{: X POSTPONE [IF] ; IMMEDIATE}. @code{X} is
13129: equivalent to @code{[IF]}.
13130:
13131: @item reaching the end of the input source before matching @code{[ELSE]} or @code{[THEN]}:
13132: @cindex @code{[IF]}, end of the input source before matching @code{[ELSE]} or @code{[THEN]}
13133: Continue in the same state of conditional compilation in the next outer
13134: input source. Currently there is no warning to the user about this.
13135:
13136: @item removing a needed definition (@code{FORGET}):
13137: @cindex @code{FORGET}, removing a needed definition
13138: Not implemented (yet).
13139:
13140: @end table
13141:
13142:
13143: @c =====================================================================
13144: @node The optional Search-Order word set, , The optional Programming-Tools word set, ANS conformance
13145: @section The optional Search-Order word set
13146: @c =====================================================================
13147: @cindex system documentation, search-order words
13148: @cindex search-order words, system documentation
13149:
13150: @menu
13151: * search-idef:: Implementation Defined Options
13152: * search-ambcond:: Ambiguous Conditions
13153: @end menu
13154:
13155:
13156: @c ---------------------------------------------------------------------
13157: @node search-idef, search-ambcond, The optional Search-Order word set, The optional Search-Order word set
13158: @subsection Implementation Defined Options
13159: @c ---------------------------------------------------------------------
13160: @cindex implementation-defined options, search-order words
13161: @cindex search-order words, implementation-defined options
13162:
13163: @table @i
13164: @item maximum number of word lists in search order:
13165: @cindex maximum number of word lists in search order
13166: @cindex search order, maximum depth
13167: @code{s" wordlists" environment? drop .}. Currently 16.
13168:
13169: @item minimum search order:
13170: @cindex minimum search order
13171: @cindex search order, minimum
13172: @code{root root}.
13173:
13174: @end table
13175:
13176: @c ---------------------------------------------------------------------
13177: @node search-ambcond, , search-idef, The optional Search-Order word set
13178: @subsection Ambiguous conditions
13179: @c ---------------------------------------------------------------------
13180: @cindex search-order words, ambiguous conditions
13181: @cindex ambiguous conditions, search-order words
13182:
13183: @table @i
1.21 crook 13184: @item changing the compilation word list (during compilation):
13185: @cindex changing the compilation word list (during compilation)
13186: @cindex compilation word list, change before definition ends
13187: The word is entered into the word list that was the compilation word list
1.1 anton 13188: at the start of the definition. Any changes to the name field (e.g.,
13189: @code{immediate}) or the code field (e.g., when executing @code{DOES>})
13190: are applied to the latest defined word (as reported by @code{last} or
1.21 crook 13191: @code{lastxt}), if possible, irrespective of the compilation word list.
1.1 anton 13192:
13193: @item search order empty (@code{previous}):
13194: @cindex @code{previous}, search order empty
1.26 crook 13195: @cindex vocstack empty, @code{previous}
1.1 anton 13196: @code{abort" Vocstack empty"}.
13197:
13198: @item too many word lists in search order (@code{also}):
13199: @cindex @code{also}, too many word lists in search order
1.26 crook 13200: @cindex vocstack full, @code{also}
1.1 anton 13201: @code{abort" Vocstack full"}.
13202:
13203: @end table
13204:
13205: @c ***************************************************************
1.65 anton 13206: @node Standard vs Extensions, Model, ANS conformance, Top
13207: @chapter Should I use Gforth extensions?
13208: @cindex Gforth extensions
13209:
13210: As you read through the rest of this manual, you will see documentation
13211: for @i{Standard} words, and documentation for some appealing Gforth
13212: @i{extensions}. You might ask yourself the question: @i{``Should I
13213: restrict myself to the standard, or should I use the extensions?''}
13214:
13215: The answer depends on the goals you have for the program you are working
13216: on:
13217:
13218: @itemize @bullet
13219:
13220: @item Is it just for yourself or do you want to share it with others?
13221:
13222: @item
13223: If you want to share it, do the others all use Gforth?
13224:
13225: @item
13226: If it is just for yourself, do you want to restrict yourself to Gforth?
13227:
13228: @end itemize
13229:
13230: If restricting the program to Gforth is ok, then there is no reason not
13231: to use extensions. It is still a good idea to keep to the standard
13232: where it is easy, in case you want to reuse these parts in another
13233: program that you want to be portable.
13234:
13235: If you want to be able to port the program to other Forth systems, there
13236: are the following points to consider:
13237:
13238: @itemize @bullet
13239:
13240: @item
13241: Most Forth systems that are being maintained support the ANS Forth
13242: standard. So if your program complies with the standard, it will be
13243: portable among many systems.
13244:
13245: @item
13246: A number of the Gforth extensions can be implemented in ANS Forth using
13247: public-domain files provided in the @file{compat/} directory. These are
13248: mentioned in the text in passing. There is no reason not to use these
13249: extensions, your program will still be ANS Forth compliant; just include
13250: the appropriate compat files with your program.
13251:
13252: @item
13253: The tool @file{ans-report.fs} (@pxref{ANS Report}) makes it easy to
13254: analyse your program and determine what non-Standard words it relies
13255: upon. However, it does not check whether you use standard words in a
13256: non-standard way.
13257:
13258: @item
13259: Some techniques are not standardized by ANS Forth, and are hard or
13260: impossible to implement in a standard way, but can be implemented in
13261: most Forth systems easily, and usually in similar ways (e.g., accessing
13262: word headers). Forth has a rich historical precedent for programmers
13263: taking advantage of implementation-dependent features of their tools
13264: (for example, relying on a knowledge of the dictionary
13265: structure). Sometimes these techniques are necessary to extract every
13266: last bit of performance from the hardware, sometimes they are just a
13267: programming shorthand.
13268:
13269: @item
13270: Does using a Gforth extension save more work than the porting this part
13271: to other Forth systems (if any) will cost?
13272:
13273: @item
13274: Is the additional functionality worth the reduction in portability and
13275: the additional porting problems?
13276:
13277: @end itemize
13278:
13279: In order to perform these consideratios, you need to know what's
13280: standard and what's not. This manual generally states if something is
13281: non-standard, but the authoritative source is the standard document.
13282: Appendix A of the Standard (@var{Rationale}) provides a valuable insight
13283: into the thought processes of the technical committee.
13284:
13285: Note also that portability between Forth systems is not the only
13286: portability issue; there is also the issue of portability between
13287: different platforms (processor/OS combinations).
13288:
13289: @c ***************************************************************
13290: @node Model, Integrating Gforth, Standard vs Extensions, Top
1.1 anton 13291: @chapter Model
13292:
13293: This chapter has yet to be written. It will contain information, on
13294: which internal structures you can rely.
13295:
13296: @c ***************************************************************
13297: @node Integrating Gforth, Emacs and Gforth, Model, Top
13298: @chapter Integrating Gforth into C programs
13299:
13300: This is not yet implemented.
13301:
13302: Several people like to use Forth as scripting language for applications
13303: that are otherwise written in C, C++, or some other language.
13304:
13305: The Forth system ATLAST provides facilities for embedding it into
13306: applications; unfortunately it has several disadvantages: most
13307: importantly, it is not based on ANS Forth, and it is apparently dead
13308: (i.e., not developed further and not supported). The facilities
1.21 crook 13309: provided by Gforth in this area are inspired by ATLAST's facilities, so
1.1 anton 13310: making the switch should not be hard.
13311:
13312: We also tried to design the interface such that it can easily be
13313: implemented by other Forth systems, so that we may one day arrive at a
13314: standardized interface. Such a standard interface would allow you to
13315: replace the Forth system without having to rewrite C code.
13316:
13317: You embed the Gforth interpreter by linking with the library
13318: @code{libgforth.a} (give the compiler the option @code{-lgforth}). All
13319: global symbols in this library that belong to the interface, have the
13320: prefix @code{forth_}. (Global symbols that are used internally have the
13321: prefix @code{gforth_}).
13322:
13323: You can include the declarations of Forth types and the functions and
13324: variables of the interface with @code{#include <forth.h>}.
13325:
13326: Types.
13327:
13328: Variables.
13329:
13330: Data and FP Stack pointer. Area sizes.
13331:
13332: functions.
13333:
13334: forth_init(imagefile)
13335: forth_evaluate(string) exceptions?
13336: forth_goto(address) (or forth_execute(xt)?)
13337: forth_continue() (a corountining mechanism)
13338:
13339: Adding primitives.
13340:
13341: No checking.
13342:
13343: Signals?
13344:
13345: Accessing the Stacks
13346:
1.26 crook 13347: @c ******************************************************************
1.1 anton 13348: @node Emacs and Gforth, Image Files, Integrating Gforth, Top
13349: @chapter Emacs and Gforth
13350: @cindex Emacs and Gforth
13351:
13352: @cindex @file{gforth.el}
13353: @cindex @file{forth.el}
13354: @cindex Rydqvist, Goran
13355: @cindex comment editing commands
13356: @cindex @code{\}, editing with Emacs
13357: @cindex debug tracer editing commands
13358: @cindex @code{~~}, removal with Emacs
13359: @cindex Forth mode in Emacs
13360: Gforth comes with @file{gforth.el}, an improved version of
13361: @file{forth.el} by Goran Rydqvist (included in the TILE package). The
1.26 crook 13362: improvements are:
13363:
13364: @itemize @bullet
13365: @item
13366: A better (but still not perfect) handling of indentation.
13367: @item
13368: Comment paragraph filling (@kbd{M-q})
13369: @item
13370: Commenting (@kbd{C-x \}) and uncommenting (@kbd{C-u C-x \}) of regions
13371: @item
13372: Removal of debugging tracers (@kbd{C-x ~}, @pxref{Debugging}).
1.41 anton 13373: @item
13374: Support of the @code{info-lookup} feature for looking up the
13375: documentation of a word.
1.26 crook 13376: @end itemize
13377:
13378: I left the stuff I do not use alone, even though some of it only makes
13379: sense for TILE. To get a description of these features, enter Forth mode
13380: and type @kbd{C-h m}.
1.1 anton 13381:
13382: @cindex source location of error or debugging output in Emacs
13383: @cindex error output, finding the source location in Emacs
13384: @cindex debugging output, finding the source location in Emacs
13385: In addition, Gforth supports Emacs quite well: The source code locations
13386: given in error messages, debugging output (from @code{~~}) and failed
13387: assertion messages are in the right format for Emacs' compilation mode
13388: (@pxref{Compilation, , Running Compilations under Emacs, emacs, Emacs
13389: Manual}) so the source location corresponding to an error or other
13390: message is only a few keystrokes away (@kbd{C-x `} for the next error,
13391: @kbd{C-c C-c} for the error under the cursor).
13392:
13393: @cindex @file{TAGS} file
13394: @cindex @file{etags.fs}
13395: @cindex viewing the source of a word in Emacs
1.43 anton 13396: @cindex @code{require}, placement in files
13397: @cindex @code{include}, placement in files
13398: Also, if you @code{require} @file{etags.fs}, a new @file{TAGS} file will
1.26 crook 13399: be produced (@pxref{Tags, , Tags Tables, emacs, Emacs Manual}) that
1.1 anton 13400: contains the definitions of all words defined afterwards. You can then
13401: find the source for a word using @kbd{M-.}. Note that emacs can use
13402: several tags files at the same time (e.g., one for the Gforth sources
13403: and one for your program, @pxref{Select Tags Table,,Selecting a Tags
13404: Table,emacs, Emacs Manual}). The TAGS file for the preloaded words is
13405: @file{$(datadir)/gforth/$(VERSION)/TAGS} (e.g.,
1.43 anton 13406: @file{/usr/local/share/gforth/0.2.0/TAGS}). To get the best behaviour
13407: with @file{etags.fs}, you should avoid putting definitions both before
13408: and after @code{require} etc., otherwise you will see the same file
13409: visited several times by commands like @code{tags-search}.
1.1 anton 13410:
1.41 anton 13411: @cindex viewing the documentation of a word in Emacs
13412: @cindex context-sensitive help
13413: Moreover, for words documented in this manual, you can look up the
13414: glossary entry quickly by using @kbd{C-h TAB}
13415: (@code{info-lookup-symbol}, see @pxref{Documentation, ,Documentation
13416: Commands, emacs, Emacs Manual}). This feature requires Emacs 20.3 or
1.42 anton 13417: later and does not work for words containing @code{:}.
1.41 anton 13418:
13419:
1.1 anton 13420: @cindex @file{.emacs}
13421: To get all these benefits, add the following lines to your @file{.emacs}
13422: file:
13423:
13424: @example
13425: (autoload 'forth-mode "gforth.el")
13426: (setq auto-mode-alist (cons '("\\.fs\\'" . forth-mode) auto-mode-alist))
13427: @end example
13428:
1.26 crook 13429: @c ******************************************************************
1.1 anton 13430: @node Image Files, Engine, Emacs and Gforth, Top
13431: @chapter Image Files
1.26 crook 13432: @cindex image file
13433: @cindex @file{.fi} files
1.1 anton 13434: @cindex precompiled Forth code
13435: @cindex dictionary in persistent form
13436: @cindex persistent form of dictionary
13437:
13438: An image file is a file containing an image of the Forth dictionary,
13439: i.e., compiled Forth code and data residing in the dictionary. By
13440: convention, we use the extension @code{.fi} for image files.
13441:
13442: @menu
1.18 anton 13443: * Image Licensing Issues:: Distribution terms for images.
13444: * Image File Background:: Why have image files?
1.67 anton 13445: * Non-Relocatable Image Files:: don't always work.
1.18 anton 13446: * Data-Relocatable Image Files:: are better.
1.67 anton 13447: * Fully Relocatable Image Files:: better yet.
1.18 anton 13448: * Stack and Dictionary Sizes:: Setting the default sizes for an image.
1.29 crook 13449: * Running Image Files:: @code{gforth -i @i{file}} or @i{file}.
1.18 anton 13450: * Modifying the Startup Sequence:: and turnkey applications.
1.1 anton 13451: @end menu
13452:
1.18 anton 13453: @node Image Licensing Issues, Image File Background, Image Files, Image Files
13454: @section Image Licensing Issues
13455: @cindex license for images
13456: @cindex image license
13457:
13458: An image created with @code{gforthmi} (@pxref{gforthmi}) or
13459: @code{savesystem} (@pxref{Non-Relocatable Image Files}) includes the
13460: original image; i.e., according to copyright law it is a derived work of
13461: the original image.
13462:
13463: Since Gforth is distributed under the GNU GPL, the newly created image
13464: falls under the GNU GPL, too. In particular, this means that if you
13465: distribute the image, you have to make all of the sources for the image
13466: available, including those you wrote. For details see @ref{License, ,
13467: GNU General Public License (Section 3)}.
13468:
13469: If you create an image with @code{cross} (@pxref{cross.fs}), the image
13470: contains only code compiled from the sources you gave it; if none of
13471: these sources is under the GPL, the terms discussed above do not apply
13472: to the image. However, if your image needs an engine (a gforth binary)
13473: that is under the GPL, you should make sure that you distribute both in
13474: a way that is at most a @emph{mere aggregation}, if you don't want the
13475: terms of the GPL to apply to the image.
13476:
13477: @node Image File Background, Non-Relocatable Image Files, Image Licensing Issues, Image Files
1.1 anton 13478: @section Image File Background
13479: @cindex image file background
13480:
13481: Our Forth system consists not only of primitives, but also of
13482: definitions written in Forth. Since the Forth compiler itself belongs to
13483: those definitions, it is not possible to start the system with the
13484: primitives and the Forth source alone. Therefore we provide the Forth
1.26 crook 13485: code as an image file in nearly executable form. When Gforth starts up,
13486: a C routine loads the image file into memory, optionally relocates the
13487: addresses, then sets up the memory (stacks etc.) according to
13488: information in the image file, and (finally) starts executing Forth
13489: code.
1.1 anton 13490:
13491: The image file variants represent different compromises between the
13492: goals of making it easy to generate image files and making them
13493: portable.
13494:
13495: @cindex relocation at run-time
1.26 crook 13496: Win32Forth 3.4 and Mitch Bradley's @code{cforth} use relocation at
1.1 anton 13497: run-time. This avoids many of the complications discussed below (image
13498: files are data relocatable without further ado), but costs performance
13499: (one addition per memory access).
13500:
13501: @cindex relocation at load-time
1.26 crook 13502: By contrast, the Gforth loader performs relocation at image load time. The
13503: loader also has to replace tokens that represent primitive calls with the
1.1 anton 13504: appropriate code-field addresses (or code addresses in the case of
13505: direct threading).
13506:
13507: There are three kinds of image files, with different degrees of
13508: relocatability: non-relocatable, data-relocatable, and fully relocatable
13509: image files.
13510:
13511: @cindex image file loader
13512: @cindex relocating loader
13513: @cindex loader for image files
13514: These image file variants have several restrictions in common; they are
13515: caused by the design of the image file loader:
13516:
13517: @itemize @bullet
13518: @item
13519: There is only one segment; in particular, this means, that an image file
13520: cannot represent @code{ALLOCATE}d memory chunks (and pointers to
1.26 crook 13521: them). The contents of the stacks are not represented, either.
1.1 anton 13522:
13523: @item
13524: The only kinds of relocation supported are: adding the same offset to
13525: all cells that represent data addresses; and replacing special tokens
13526: with code addresses or with pieces of machine code.
13527:
13528: If any complex computations involving addresses are performed, the
13529: results cannot be represented in the image file. Several applications that
13530: use such computations come to mind:
13531: @itemize @minus
13532: @item
13533: Hashing addresses (or data structures which contain addresses) for table
13534: lookup. If you use Gforth's @code{table}s or @code{wordlist}s for this
13535: purpose, you will have no problem, because the hash tables are
13536: recomputed automatically when the system is started. If you use your own
13537: hash tables, you will have to do something similar.
13538:
13539: @item
13540: There's a cute implementation of doubly-linked lists that uses
13541: @code{XOR}ed addresses. You could represent such lists as singly-linked
13542: in the image file, and restore the doubly-linked representation on
13543: startup.@footnote{In my opinion, though, you should think thrice before
13544: using a doubly-linked list (whatever implementation).}
13545:
13546: @item
13547: The code addresses of run-time routines like @code{docol:} cannot be
13548: represented in the image file (because their tokens would be replaced by
13549: machine code in direct threaded implementations). As a workaround,
13550: compute these addresses at run-time with @code{>code-address} from the
13551: executions tokens of appropriate words (see the definitions of
13552: @code{docol:} and friends in @file{kernel.fs}).
13553:
13554: @item
13555: On many architectures addresses are represented in machine code in some
13556: shifted or mangled form. You cannot put @code{CODE} words that contain
13557: absolute addresses in this form in a relocatable image file. Workarounds
13558: are representing the address in some relative form (e.g., relative to
13559: the CFA, which is present in some register), or loading the address from
13560: a place where it is stored in a non-mangled form.
13561: @end itemize
13562: @end itemize
13563:
13564: @node Non-Relocatable Image Files, Data-Relocatable Image Files, Image File Background, Image Files
13565: @section Non-Relocatable Image Files
13566: @cindex non-relocatable image files
1.26 crook 13567: @cindex image file, non-relocatable
1.1 anton 13568:
13569: These files are simple memory dumps of the dictionary. They are specific
13570: to the executable (i.e., @file{gforth} file) they were created
13571: with. What's worse, they are specific to the place on which the
13572: dictionary resided when the image was created. Now, there is no
13573: guarantee that the dictionary will reside at the same place the next
13574: time you start Gforth, so there's no guarantee that a non-relocatable
13575: image will work the next time (Gforth will complain instead of crashing,
13576: though).
13577:
13578: You can create a non-relocatable image file with
13579:
1.44 crook 13580:
1.1 anton 13581: doc-savesystem
13582:
1.44 crook 13583:
1.1 anton 13584: @node Data-Relocatable Image Files, Fully Relocatable Image Files, Non-Relocatable Image Files, Image Files
13585: @section Data-Relocatable Image Files
13586: @cindex data-relocatable image files
1.26 crook 13587: @cindex image file, data-relocatable
1.1 anton 13588:
13589: These files contain relocatable data addresses, but fixed code addresses
13590: (instead of tokens). They are specific to the executable (i.e.,
13591: @file{gforth} file) they were created with. For direct threading on some
13592: architectures (e.g., the i386), data-relocatable images do not work. You
13593: get a data-relocatable image, if you use @file{gforthmi} with a
13594: Gforth binary that is not doubly indirect threaded (@pxref{Fully
13595: Relocatable Image Files}).
13596:
13597: @node Fully Relocatable Image Files, Stack and Dictionary Sizes, Data-Relocatable Image Files, Image Files
13598: @section Fully Relocatable Image Files
13599: @cindex fully relocatable image files
1.26 crook 13600: @cindex image file, fully relocatable
1.1 anton 13601:
13602: @cindex @file{kern*.fi}, relocatability
13603: @cindex @file{gforth.fi}, relocatability
13604: These image files have relocatable data addresses, and tokens for code
13605: addresses. They can be used with different binaries (e.g., with and
13606: without debugging) on the same machine, and even across machines with
13607: the same data formats (byte order, cell size, floating point
13608: format). However, they are usually specific to the version of Gforth
13609: they were created with. The files @file{gforth.fi} and @file{kernl*.fi}
13610: are fully relocatable.
13611:
13612: There are two ways to create a fully relocatable image file:
13613:
13614: @menu
1.29 crook 13615: * gforthmi:: The normal way
1.1 anton 13616: * cross.fs:: The hard way
13617: @end menu
13618:
13619: @node gforthmi, cross.fs, Fully Relocatable Image Files, Fully Relocatable Image Files
13620: @subsection @file{gforthmi}
13621: @cindex @file{comp-i.fs}
13622: @cindex @file{gforthmi}
13623:
13624: You will usually use @file{gforthmi}. If you want to create an
1.29 crook 13625: image @i{file} that contains everything you would load by invoking
13626: Gforth with @code{gforth @i{options}}, you simply say:
1.1 anton 13627: @example
1.29 crook 13628: gforthmi @i{file} @i{options}
1.1 anton 13629: @end example
13630:
13631: E.g., if you want to create an image @file{asm.fi} that has the file
13632: @file{asm.fs} loaded in addition to the usual stuff, you could do it
13633: like this:
13634:
13635: @example
13636: gforthmi asm.fi asm.fs
13637: @end example
13638:
1.27 crook 13639: @file{gforthmi} is implemented as a sh script and works like this: It
13640: produces two non-relocatable images for different addresses and then
13641: compares them. Its output reflects this: first you see the output (if
1.62 crook 13642: any) of the two Gforth invocations that produce the non-relocatable image
1.27 crook 13643: files, then you see the output of the comparing program: It displays the
13644: offset used for data addresses and the offset used for code addresses;
1.1 anton 13645: moreover, for each cell that cannot be represented correctly in the
1.44 crook 13646: image files, it displays a line like this:
1.1 anton 13647:
13648: @example
13649: 78DC BFFFFA50 BFFFFA40
13650: @end example
13651:
13652: This means that at offset $78dc from @code{forthstart}, one input image
13653: contains $bffffa50, and the other contains $bffffa40. Since these cells
13654: cannot be represented correctly in the output image, you should examine
13655: these places in the dictionary and verify that these cells are dead
13656: (i.e., not read before they are written).
1.39 anton 13657:
13658: @cindex --application, @code{gforthmi} option
13659: If you insert the option @code{--application} in front of the image file
13660: name, you will get an image that uses the @code{--appl-image} option
13661: instead of the @code{--image-file} option (@pxref{Invoking
13662: Gforth}). When you execute such an image on Unix (by typing the image
13663: name as command), the Gforth engine will pass all options to the image
13664: instead of trying to interpret them as engine options.
1.1 anton 13665:
1.27 crook 13666: If you type @file{gforthmi} with no arguments, it prints some usage
13667: instructions.
13668:
1.1 anton 13669: @cindex @code{savesystem} during @file{gforthmi}
13670: @cindex @code{bye} during @file{gforthmi}
13671: @cindex doubly indirect threaded code
1.44 crook 13672: @cindex environment variables
13673: @cindex @code{GFORTHD} -- environment variable
13674: @cindex @code{GFORTH} -- environment variable
1.1 anton 13675: @cindex @code{gforth-ditc}
1.29 crook 13676: There are a few wrinkles: After processing the passed @i{options}, the
1.1 anton 13677: words @code{savesystem} and @code{bye} must be visible. A special doubly
13678: indirect threaded version of the @file{gforth} executable is used for
1.62 crook 13679: creating the non-relocatable images; you can pass the exact filename of
1.1 anton 13680: this executable through the environment variable @code{GFORTHD}
13681: (default: @file{gforth-ditc}); if you pass a version that is not doubly
13682: indirect threaded, you will not get a fully relocatable image, but a
1.27 crook 13683: data-relocatable image (because there is no code address offset). The
13684: normal @file{gforth} executable is used for creating the relocatable
13685: image; you can pass the exact filename of this executable through the
13686: environment variable @code{GFORTH}.
1.1 anton 13687:
13688: @node cross.fs, , gforthmi, Fully Relocatable Image Files
13689: @subsection @file{cross.fs}
13690: @cindex @file{cross.fs}
13691: @cindex cross-compiler
13692: @cindex metacompiler
1.47 crook 13693: @cindex target compiler
1.1 anton 13694:
13695: You can also use @code{cross}, a batch compiler that accepts a Forth-like
1.47 crook 13696: programming language (@pxref{Cross Compiler}).
1.1 anton 13697:
1.47 crook 13698: @code{cross} allows you to create image files for machines with
1.1 anton 13699: different data sizes and data formats than the one used for generating
13700: the image file. You can also use it to create an application image that
13701: does not contain a Forth compiler. These features are bought with
13702: restrictions and inconveniences in programming. E.g., addresses have to
13703: be stored in memory with special words (@code{A!}, @code{A,}, etc.) in
13704: order to make the code relocatable.
13705:
13706:
13707: @node Stack and Dictionary Sizes, Running Image Files, Fully Relocatable Image Files, Image Files
13708: @section Stack and Dictionary Sizes
13709: @cindex image file, stack and dictionary sizes
13710: @cindex dictionary size default
13711: @cindex stack size default
13712:
13713: If you invoke Gforth with a command line flag for the size
13714: (@pxref{Invoking Gforth}), the size you specify is stored in the
13715: dictionary. If you save the dictionary with @code{savesystem} or create
13716: an image with @file{gforthmi}, this size will become the default
13717: for the resulting image file. E.g., the following will create a
1.21 crook 13718: fully relocatable version of @file{gforth.fi} with a 1MB dictionary:
1.1 anton 13719:
13720: @example
13721: gforthmi gforth.fi -m 1M
13722: @end example
13723:
13724: In other words, if you want to set the default size for the dictionary
13725: and the stacks of an image, just invoke @file{gforthmi} with the
13726: appropriate options when creating the image.
13727:
13728: @cindex stack size, cache-friendly
13729: Note: For cache-friendly behaviour (i.e., good performance), you should
13730: make the sizes of the stacks modulo, say, 2K, somewhat different. E.g.,
13731: the default stack sizes are: data: 16k (mod 2k=0); fp: 15.5k (mod
13732: 2k=1.5k); return: 15k(mod 2k=1k); locals: 14.5k (mod 2k=0.5k).
13733:
13734: @node Running Image Files, Modifying the Startup Sequence, Stack and Dictionary Sizes, Image Files
13735: @section Running Image Files
13736: @cindex running image files
13737: @cindex invoking image files
13738: @cindex image file invocation
13739:
13740: @cindex -i, invoke image file
13741: @cindex --image file, invoke image file
1.29 crook 13742: You can invoke Gforth with an image file @i{image} instead of the
1.1 anton 13743: default @file{gforth.fi} with the @code{-i} flag (@pxref{Invoking Gforth}):
13744: @example
1.29 crook 13745: gforth -i @i{image}
1.1 anton 13746: @end example
13747:
13748: @cindex executable image file
1.26 crook 13749: @cindex image file, executable
1.1 anton 13750: If your operating system supports starting scripts with a line of the
13751: form @code{#! ...}, you just have to type the image file name to start
13752: Gforth with this image file (note that the file extension @code{.fi} is
1.29 crook 13753: just a convention). I.e., to run Gforth with the image file @i{image},
13754: you can just type @i{image} instead of @code{gforth -i @i{image}}.
1.27 crook 13755: This works because every @code{.fi} file starts with a line of this
13756: format:
13757:
13758: @example
13759: #! /usr/local/bin/gforth-0.4.0 -i
13760: @end example
13761:
13762: The file and pathname for the Gforth engine specified on this line is
13763: the specific Gforth executable that it was built against; i.e. the value
13764: of the environment variable @code{GFORTH} at the time that
13765: @file{gforthmi} was executed.
1.1 anton 13766:
1.27 crook 13767: You can make use of the same shell capability to make a Forth source
13768: file into an executable. For example, if you place this text in a file:
1.26 crook 13769:
13770: @example
13771: #! /usr/local/bin/gforth
13772:
13773: ." Hello, world" CR
13774: bye
13775: @end example
13776:
13777: @noindent
1.27 crook 13778: and then make the file executable (chmod +x in Unix), you can run it
1.26 crook 13779: directly from the command line. The sequence @code{#!} is used in two
13780: ways; firstly, it is recognised as a ``magic sequence'' by the operating
1.29 crook 13781: system@footnote{The Unix kernel actually recognises two types of files:
13782: executable files and files of data, where the data is processed by an
13783: interpreter that is specified on the ``interpreter line'' -- the first
13784: line of the file, starting with the sequence #!. There may be a small
13785: limit (e.g., 32) on the number of characters that may be specified on
13786: the interpreter line.} secondly it is treated as a comment character by
13787: Gforth. Because of the second usage, a space is required between
13788: @code{#!} and the path to the executable.
1.27 crook 13789:
13790: The disadvantage of this latter technique, compared with using
13791: @file{gforthmi}, is that it is slower; the Forth source code is compiled
13792: on-the-fly, each time the program is invoked.
13793:
1.26 crook 13794:
1.1 anton 13795: doc-#!
13796:
1.44 crook 13797:
1.1 anton 13798: @node Modifying the Startup Sequence, , Running Image Files, Image Files
13799: @section Modifying the Startup Sequence
13800: @cindex startup sequence for image file
13801: @cindex image file initialization sequence
13802: @cindex initialization sequence of image file
13803:
13804: You can add your own initialization to the startup sequence through the
1.26 crook 13805: deferred word @code{'cold}. @code{'cold} is invoked just before the
13806: image-specific command line processing (by default, loading files and
13807: evaluating (@code{-e}) strings) starts.
1.1 anton 13808:
13809: A sequence for adding your initialization usually looks like this:
13810:
13811: @example
13812: :noname
13813: Defers 'cold \ do other initialization stuff (e.g., rehashing wordlists)
13814: ... \ your stuff
13815: ; IS 'cold
13816: @end example
13817:
13818: @cindex turnkey image files
1.26 crook 13819: @cindex image file, turnkey applications
1.1 anton 13820: You can make a turnkey image by letting @code{'cold} execute a word
13821: (your turnkey application) that never returns; instead, it exits Gforth
13822: via @code{bye} or @code{throw}.
13823:
13824: @cindex command-line arguments, access
13825: @cindex arguments on the command line, access
13826: You can access the (image-specific) command-line arguments through the
1.26 crook 13827: variables @code{argc} and @code{argv}. @code{arg} provides convenient
1.1 anton 13828: access to @code{argv}.
13829:
1.26 crook 13830: If @code{'cold} exits normally, Gforth processes the command-line
13831: arguments as files to be loaded and strings to be evaluated. Therefore,
13832: @code{'cold} should remove the arguments it has used in this case.
13833:
1.44 crook 13834:
13835:
1.26 crook 13836: doc-'cold
1.1 anton 13837: doc-argc
13838: doc-argv
13839: doc-arg
13840:
13841:
1.44 crook 13842:
1.1 anton 13843: @c ******************************************************************
1.13 pazsan 13844: @node Engine, Binding to System Library, Image Files, Top
1.1 anton 13845: @chapter Engine
13846: @cindex engine
13847: @cindex virtual machine
13848:
1.26 crook 13849: Reading this chapter is not necessary for programming with Gforth. It
1.1 anton 13850: may be helpful for finding your way in the Gforth sources.
13851:
1.66 anton 13852: The ideas in this section have also been published in Bernd Paysan,
13853: @cite{ANS fig/GNU/??? Forth} (in German), Forth-Tagung '93 and M. Anton
13854: Ertl, @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl93.ps.Z, A
13855: Portable Forth Engine}}, EuroForth '93.
1.1 anton 13856:
13857: @menu
13858: * Portability::
13859: * Threading::
13860: * Primitives::
13861: * Performance::
13862: @end menu
13863:
13864: @node Portability, Threading, Engine, Engine
13865: @section Portability
13866: @cindex engine portability
13867:
1.26 crook 13868: An important goal of the Gforth Project is availability across a wide
13869: range of personal machines. fig-Forth, and, to a lesser extent, F83,
13870: achieved this goal by manually coding the engine in assembly language
13871: for several then-popular processors. This approach is very
13872: labor-intensive and the results are short-lived due to progress in
13873: computer architecture.
1.1 anton 13874:
13875: @cindex C, using C for the engine
13876: Others have avoided this problem by coding in C, e.g., Mitch Bradley
13877: (cforth), Mikael Patel (TILE) and Dirk Zoller (pfe). This approach is
13878: particularly popular for UNIX-based Forths due to the large variety of
13879: architectures of UNIX machines. Unfortunately an implementation in C
13880: does not mix well with the goals of efficiency and with using
13881: traditional techniques: Indirect or direct threading cannot be expressed
13882: in C, and switch threading, the fastest technique available in C, is
13883: significantly slower. Another problem with C is that it is very
13884: cumbersome to express double integer arithmetic.
13885:
13886: @cindex GNU C for the engine
13887: @cindex long long
13888: Fortunately, there is a portable language that does not have these
13889: limitations: GNU C, the version of C processed by the GNU C compiler
13890: (@pxref{C Extensions, , Extensions to the C Language Family, gcc.info,
13891: GNU C Manual}). Its labels as values feature (@pxref{Labels as Values, ,
13892: Labels as Values, gcc.info, GNU C Manual}) makes direct and indirect
13893: threading possible, its @code{long long} type (@pxref{Long Long, ,
13894: Double-Word Integers, gcc.info, GNU C Manual}) corresponds to Forth's
13895: double numbers@footnote{Unfortunately, long longs are not implemented
13896: properly on all machines (e.g., on alpha-osf1, long longs are only 64
13897: bits, the same size as longs (and pointers), but they should be twice as
1.4 anton 13898: long according to @pxref{Long Long, , Double-Word Integers, gcc.info, GNU
1.1 anton 13899: C Manual}). So, we had to implement doubles in C after all. Still, on
13900: most machines we can use long longs and achieve better performance than
13901: with the emulation package.}. GNU C is available for free on all
13902: important (and many unimportant) UNIX machines, VMS, 80386s running
13903: MS-DOS, the Amiga, and the Atari ST, so a Forth written in GNU C can run
13904: on all these machines.
13905:
13906: Writing in a portable language has the reputation of producing code that
13907: is slower than assembly. For our Forth engine we repeatedly looked at
13908: the code produced by the compiler and eliminated most compiler-induced
13909: inefficiencies by appropriate changes in the source code.
13910:
13911: @cindex explicit register declarations
13912: @cindex --enable-force-reg, configuration flag
13913: @cindex -DFORCE_REG
13914: However, register allocation cannot be portably influenced by the
13915: programmer, leading to some inefficiencies on register-starved
13916: machines. We use explicit register declarations (@pxref{Explicit Reg
13917: Vars, , Variables in Specified Registers, gcc.info, GNU C Manual}) to
13918: improve the speed on some machines. They are turned on by using the
13919: configuration flag @code{--enable-force-reg} (@code{gcc} switch
13920: @code{-DFORCE_REG}). Unfortunately, this feature not only depends on the
13921: machine, but also on the compiler version: On some machines some
13922: compiler versions produce incorrect code when certain explicit register
13923: declarations are used. So by default @code{-DFORCE_REG} is not used.
13924:
13925: @node Threading, Primitives, Portability, Engine
13926: @section Threading
13927: @cindex inner interpreter implementation
13928: @cindex threaded code implementation
13929:
13930: @cindex labels as values
13931: GNU C's labels as values extension (available since @code{gcc-2.0},
13932: @pxref{Labels as Values, , Labels as Values, gcc.info, GNU C Manual})
1.29 crook 13933: makes it possible to take the address of @i{label} by writing
13934: @code{&&@i{label}}. This address can then be used in a statement like
13935: @code{goto *@i{address}}. I.e., @code{goto *&&x} is the same as
1.1 anton 13936: @code{goto x}.
13937:
1.26 crook 13938: @cindex @code{NEXT}, indirect threaded
1.1 anton 13939: @cindex indirect threaded inner interpreter
13940: @cindex inner interpreter, indirect threaded
1.26 crook 13941: With this feature an indirect threaded @code{NEXT} looks like:
1.1 anton 13942: @example
13943: cfa = *ip++;
13944: ca = *cfa;
13945: goto *ca;
13946: @end example
13947: @cindex instruction pointer
13948: For those unfamiliar with the names: @code{ip} is the Forth instruction
13949: pointer; the @code{cfa} (code-field address) corresponds to ANS Forths
13950: execution token and points to the code field of the next word to be
13951: executed; The @code{ca} (code address) fetched from there points to some
13952: executable code, e.g., a primitive or the colon definition handler
13953: @code{docol}.
13954:
1.26 crook 13955: @cindex @code{NEXT}, direct threaded
1.1 anton 13956: @cindex direct threaded inner interpreter
13957: @cindex inner interpreter, direct threaded
13958: Direct threading is even simpler:
13959: @example
13960: ca = *ip++;
13961: goto *ca;
13962: @end example
13963:
13964: Of course we have packaged the whole thing neatly in macros called
1.26 crook 13965: @code{NEXT} and @code{NEXT1} (the part of @code{NEXT} after fetching the cfa).
1.1 anton 13966:
13967: @menu
13968: * Scheduling::
13969: * Direct or Indirect Threaded?::
13970: * DOES>::
13971: @end menu
13972:
13973: @node Scheduling, Direct or Indirect Threaded?, Threading, Threading
13974: @subsection Scheduling
13975: @cindex inner interpreter optimization
13976:
13977: There is a little complication: Pipelined and superscalar processors,
13978: i.e., RISC and some modern CISC machines can process independent
13979: instructions while waiting for the results of an instruction. The
13980: compiler usually reorders (schedules) the instructions in a way that
13981: achieves good usage of these delay slots. However, on our first tries
13982: the compiler did not do well on scheduling primitives. E.g., for
13983: @code{+} implemented as
13984: @example
13985: n=sp[0]+sp[1];
13986: sp++;
13987: sp[0]=n;
13988: NEXT;
13989: @end example
1.26 crook 13990: the @code{NEXT} comes strictly after the other code, i.e., there is nearly no
1.1 anton 13991: scheduling. After a little thought the problem becomes clear: The
1.21 crook 13992: compiler cannot know that @code{sp} and @code{ip} point to different
13993: addresses (and the version of @code{gcc} we used would not know it even
13994: if it was possible), so it could not move the load of the cfa above the
13995: store to the TOS. Indeed the pointers could be the same, if code on or
13996: very near the top of stack were executed. In the interest of speed we
13997: chose to forbid this probably unused ``feature'' and helped the compiler
1.26 crook 13998: in scheduling: @code{NEXT} is divided into the loading part (@code{NEXT_P1})
1.21 crook 13999: and the goto part (@code{NEXT_P2}). @code{+} now looks like:
1.1 anton 14000: @example
14001: n=sp[0]+sp[1];
14002: sp++;
14003: NEXT_P1;
14004: sp[0]=n;
14005: NEXT_P2;
14006: @end example
14007: This can be scheduled optimally by the compiler.
14008:
14009: This division can be turned off with the switch @code{-DCISC_NEXT}. This
14010: switch is on by default on machines that do not profit from scheduling
14011: (e.g., the 80386), in order to preserve registers.
14012:
14013: @node Direct or Indirect Threaded?, DOES>, Scheduling, Threading
14014: @subsection Direct or Indirect Threaded?
14015: @cindex threading, direct or indirect?
14016:
14017: @cindex -DDIRECT_THREADED
14018: Both! After packaging the nasty details in macro definitions we
14019: realized that we could switch between direct and indirect threading by
14020: simply setting a compilation flag (@code{-DDIRECT_THREADED}) and
14021: defining a few machine-specific macros for the direct-threading case.
14022: On the Forth level we also offer access words that hide the
14023: differences between the threading methods (@pxref{Threading Words}).
14024:
14025: Indirect threading is implemented completely machine-independently.
14026: Direct threading needs routines for creating jumps to the executable
1.21 crook 14027: code (e.g. to @code{docol} or @code{dodoes}). These routines are inherently
14028: machine-dependent, but they do not amount to many source lines. Therefore,
14029: even porting direct threading to a new machine requires little effort.
1.1 anton 14030:
14031: @cindex --enable-indirect-threaded, configuration flag
14032: @cindex --enable-direct-threaded, configuration flag
14033: The default threading method is machine-dependent. You can enforce a
14034: specific threading method when building Gforth with the configuration
14035: flag @code{--enable-direct-threaded} or
14036: @code{--enable-indirect-threaded}. Note that direct threading is not
14037: supported on all machines.
14038:
14039: @node DOES>, , Direct or Indirect Threaded?, Threading
14040: @subsection DOES>
14041: @cindex @code{DOES>} implementation
14042:
1.26 crook 14043: @cindex @code{dodoes} routine
14044: @cindex @code{DOES>}-code
1.1 anton 14045: One of the most complex parts of a Forth engine is @code{dodoes}, i.e.,
14046: the chunk of code executed by every word defined by a
14047: @code{CREATE}...@code{DOES>} pair. The main problem here is: How to find
14048: the Forth code to be executed, i.e. the code after the
1.26 crook 14049: @code{DOES>} (the @code{DOES>}-code)? There are two solutions:
1.1 anton 14050:
1.21 crook 14051: In fig-Forth the code field points directly to the @code{dodoes} and the
1.45 crook 14052: @code{DOES>}-code address is stored in the cell after the code address (i.e. at
1.29 crook 14053: @code{@i{CFA} cell+}). It may seem that this solution is illegal in
1.1 anton 14054: the Forth-79 and all later standards, because in fig-Forth this address
14055: lies in the body (which is illegal in these standards). However, by
14056: making the code field larger for all words this solution becomes legal
14057: again. We use this approach for the indirect threaded version and for
14058: direct threading on some machines. Leaving a cell unused in most words
14059: is a bit wasteful, but on the machines we are targeting this is hardly a
14060: problem. The other reason for having a code field size of two cells is
14061: to avoid having different image files for direct and indirect threaded
14062: systems (direct threaded systems require two-cell code fields on many
14063: machines).
14064:
1.26 crook 14065: @cindex @code{DOES>}-handler
1.1 anton 14066: The other approach is that the code field points or jumps to the cell
1.26 crook 14067: after @code{DOES>}. In this variant there is a jump to @code{dodoes} at
14068: this address (the @code{DOES>}-handler). @code{dodoes} can then get the
14069: @code{DOES>}-code address by computing the code address, i.e., the address of
1.45 crook 14070: the jump to @code{dodoes}, and add the length of that jump field. A variant of
1.1 anton 14071: this is to have a call to @code{dodoes} after the @code{DOES>}; then the
14072: return address (which can be found in the return register on RISCs) is
1.26 crook 14073: the @code{DOES>}-code address. Since the two cells available in the code field
1.1 anton 14074: are used up by the jump to the code address in direct threading on many
14075: architectures, we use this approach for direct threading on these
14076: architectures. We did not want to add another cell to the code field.
14077:
14078: @node Primitives, Performance, Threading, Engine
14079: @section Primitives
14080: @cindex primitives, implementation
14081: @cindex virtual machine instructions, implementation
14082:
14083: @menu
14084: * Automatic Generation::
14085: * TOS Optimization::
14086: * Produced code::
14087: @end menu
14088:
14089: @node Automatic Generation, TOS Optimization, Primitives, Primitives
14090: @subsection Automatic Generation
14091: @cindex primitives, automatic generation
14092:
14093: @cindex @file{prims2x.fs}
14094: Since the primitives are implemented in a portable language, there is no
14095: longer any need to minimize the number of primitives. On the contrary,
14096: having many primitives has an advantage: speed. In order to reduce the
14097: number of errors in primitives and to make programming them easier, we
14098: provide a tool, the primitive generator (@file{prims2x.fs}), that
14099: automatically generates most (and sometimes all) of the C code for a
14100: primitive from the stack effect notation. The source for a primitive
14101: has the following form:
14102:
14103: @cindex primitive source format
14104: @format
1.58 anton 14105: @i{Forth-name} ( @i{stack-effect} ) @i{category} [@i{pronounc.}]
1.29 crook 14106: [@code{""}@i{glossary entry}@code{""}]
14107: @i{C code}
1.1 anton 14108: [@code{:}
1.29 crook 14109: @i{Forth code}]
1.1 anton 14110: @end format
14111:
14112: The items in brackets are optional. The category and glossary fields
14113: are there for generating the documentation, the Forth code is there
14114: for manual implementations on machines without GNU C. E.g., the source
14115: for the primitive @code{+} is:
14116: @example
1.58 anton 14117: + ( n1 n2 -- n ) core plus
1.1 anton 14118: n = n1+n2;
14119: @end example
14120:
14121: This looks like a specification, but in fact @code{n = n1+n2} is C
14122: code. Our primitive generation tool extracts a lot of information from
14123: the stack effect notations@footnote{We use a one-stack notation, even
14124: though we have separate data and floating-point stacks; The separate
14125: notation can be generated easily from the unified notation.}: The number
14126: of items popped from and pushed on the stack, their type, and by what
14127: name they are referred to in the C code. It then generates a C code
14128: prelude and postlude for each primitive. The final C code for @code{+}
14129: looks like this:
14130:
14131: @example
1.46 pazsan 14132: I_plus: /* + ( n1 n2 -- n ) */ /* label, stack effect */
1.1 anton 14133: /* */ /* documentation */
14134: @{
14135: DEF_CA /* definition of variable ca (indirect threading) */
14136: Cell n1; /* definitions of variables */
14137: Cell n2;
14138: Cell n;
14139: n1 = (Cell) sp[1]; /* input */
14140: n2 = (Cell) TOS;
14141: sp += 1; /* stack adjustment */
14142: NAME("+") /* debugging output (with -DDEBUG) */
14143: @{
14144: n = n1+n2; /* C code taken from the source */
14145: @}
14146: NEXT_P1; /* NEXT part 1 */
14147: TOS = (Cell)n; /* output */
14148: NEXT_P2; /* NEXT part 2 */
14149: @}
14150: @end example
14151:
14152: This looks long and inefficient, but the GNU C compiler optimizes quite
14153: well and produces optimal code for @code{+} on, e.g., the R3000 and the
14154: HP RISC machines: Defining the @code{n}s does not produce any code, and
14155: using them as intermediate storage also adds no cost.
14156:
1.26 crook 14157: There are also other optimizations that are not illustrated by this
14158: example: assignments between simple variables are usually for free (copy
1.1 anton 14159: propagation). If one of the stack items is not used by the primitive
14160: (e.g. in @code{drop}), the compiler eliminates the load from the stack
14161: (dead code elimination). On the other hand, there are some things that
14162: the compiler does not do, therefore they are performed by
14163: @file{prims2x.fs}: The compiler does not optimize code away that stores
14164: a stack item to the place where it just came from (e.g., @code{over}).
14165:
14166: While programming a primitive is usually easy, there are a few cases
14167: where the programmer has to take the actions of the generator into
14168: account, most notably @code{?dup}, but also words that do not (always)
1.26 crook 14169: fall through to @code{NEXT}.
1.1 anton 14170:
14171: @node TOS Optimization, Produced code, Automatic Generation, Primitives
14172: @subsection TOS Optimization
14173: @cindex TOS optimization for primitives
14174: @cindex primitives, keeping the TOS in a register
14175:
14176: An important optimization for stack machine emulators, e.g., Forth
14177: engines, is keeping one or more of the top stack items in
1.29 crook 14178: registers. If a word has the stack effect @i{in1}...@i{inx} @code{--}
14179: @i{out1}...@i{outy}, keeping the top @i{n} items in registers
1.1 anton 14180: @itemize @bullet
14181: @item
1.29 crook 14182: is better than keeping @i{n-1} items, if @i{x>=n} and @i{y>=n},
1.1 anton 14183: due to fewer loads from and stores to the stack.
1.29 crook 14184: @item is slower than keeping @i{n-1} items, if @i{x<>y} and @i{x<n} and
14185: @i{y<n}, due to additional moves between registers.
1.1 anton 14186: @end itemize
14187:
14188: @cindex -DUSE_TOS
14189: @cindex -DUSE_NO_TOS
14190: In particular, keeping one item in a register is never a disadvantage,
14191: if there are enough registers. Keeping two items in registers is a
14192: disadvantage for frequent words like @code{?branch}, constants,
14193: variables, literals and @code{i}. Therefore our generator only produces
14194: code that keeps zero or one items in registers. The generated C code
14195: covers both cases; the selection between these alternatives is made at
14196: C-compile time using the switch @code{-DUSE_TOS}. @code{TOS} in the C
14197: code for @code{+} is just a simple variable name in the one-item case,
14198: otherwise it is a macro that expands into @code{sp[0]}. Note that the
14199: GNU C compiler tries to keep simple variables like @code{TOS} in
14200: registers, and it usually succeeds, if there are enough registers.
14201:
14202: @cindex -DUSE_FTOS
14203: @cindex -DUSE_NO_FTOS
14204: The primitive generator performs the TOS optimization for the
14205: floating-point stack, too (@code{-DUSE_FTOS}). For floating-point
14206: operations the benefit of this optimization is even larger:
14207: floating-point operations take quite long on most processors, but can be
14208: performed in parallel with other operations as long as their results are
14209: not used. If the FP-TOS is kept in a register, this works. If
14210: it is kept on the stack, i.e., in memory, the store into memory has to
14211: wait for the result of the floating-point operation, lengthening the
14212: execution time of the primitive considerably.
14213:
14214: The TOS optimization makes the automatic generation of primitives a
14215: bit more complicated. Just replacing all occurrences of @code{sp[0]} by
14216: @code{TOS} is not sufficient. There are some special cases to
14217: consider:
14218: @itemize @bullet
14219: @item In the case of @code{dup ( w -- w w )} the generator must not
14220: eliminate the store to the original location of the item on the stack,
14221: if the TOS optimization is turned on.
14222: @item Primitives with stack effects of the form @code{--}
1.29 crook 14223: @i{out1}...@i{outy} must store the TOS to the stack at the start.
14224: Likewise, primitives with the stack effect @i{in1}...@i{inx} @code{--}
1.1 anton 14225: must load the TOS from the stack at the end. But for the null stack
14226: effect @code{--} no stores or loads should be generated.
14227: @end itemize
14228:
14229: @node Produced code, , TOS Optimization, Primitives
14230: @subsection Produced code
14231: @cindex primitives, assembly code listing
14232:
14233: @cindex @file{engine.s}
14234: To see what assembly code is produced for the primitives on your machine
14235: with your compiler and your flag settings, type @code{make engine.s} and
14236: look at the resulting file @file{engine.s}.
14237:
14238: @node Performance, , Primitives, Engine
14239: @section Performance
14240: @cindex performance of some Forth interpreters
14241: @cindex engine performance
14242: @cindex benchmarking Forth systems
14243: @cindex Gforth performance
14244:
14245: On RISCs the Gforth engine is very close to optimal; i.e., it is usually
14246: impossible to write a significantly faster engine.
14247:
14248: On register-starved machines like the 386 architecture processors
14249: improvements are possible, because @code{gcc} does not utilize the
14250: registers as well as a human, even with explicit register declarations;
14251: e.g., Bernd Beuster wrote a Forth system fragment in assembly language
14252: and hand-tuned it for the 486; this system is 1.19 times faster on the
14253: Sieve benchmark on a 486DX2/66 than Gforth compiled with
1.40 anton 14254: @code{gcc-2.6.3} with @code{-DFORCE_REG}. The situation has improved
14255: with gcc-2.95 and gforth-0.4.9; now the most important virtual machine
14256: registers fit in real registers (and we can even afford to use the TOS
14257: optimization), resulting in a speedup of 1.14 on the sieve over the
14258: earlier results.
1.1 anton 14259:
14260: @cindex Win32Forth performance
14261: @cindex NT Forth performance
14262: @cindex eforth performance
14263: @cindex ThisForth performance
14264: @cindex PFE performance
14265: @cindex TILE performance
1.40 anton 14266: The potential advantage of assembly language implementations
1.1 anton 14267: is not necessarily realized in complete Forth systems: We compared
1.40 anton 14268: Gforth-0.4.9 (direct threaded, compiled with @code{gcc-2.95.1} and
1.1 anton 14269: @code{-DFORCE_REG}) with Win32Forth 1.2093, LMI's NT Forth (Beta, May
14270: 1994) and Eforth (with and without peephole (aka pinhole) optimization
14271: of the threaded code); all these systems were written in assembly
14272: language. We also compared Gforth with three systems written in C:
14273: PFE-0.9.14 (compiled with @code{gcc-2.6.3} with the default
14274: configuration for Linux: @code{-O2 -fomit-frame-pointer -DUSE_REGS
1.21 crook 14275: -DUNROLL_NEXT}), ThisForth Beta (compiled with @code{gcc-2.6.3 -O3
14276: -fomit-frame-pointer}; ThisForth employs peephole optimization of the
1.1 anton 14277: threaded code) and TILE (compiled with @code{make opt}). We benchmarked
14278: Gforth, PFE, ThisForth and TILE on a 486DX2/66 under Linux. Kenneth
14279: O'Heskin kindly provided the results for Win32Forth and NT Forth on a
14280: 486DX2/66 with similar memory performance under Windows NT. Marcel
14281: Hendrix ported Eforth to Linux, then extended it to run the benchmarks,
14282: added the peephole optimizer, ran the benchmarks and reported the
14283: results.
1.40 anton 14284:
1.1 anton 14285: We used four small benchmarks: the ubiquitous Sieve; bubble-sorting and
14286: matrix multiplication come from the Stanford integer benchmarks and have
14287: been translated into Forth by Martin Fraeman; we used the versions
14288: included in the TILE Forth package, but with bigger data set sizes; and
14289: a recursive Fibonacci number computation for benchmarking calling
14290: performance. The following table shows the time taken for the benchmarks
14291: scaled by the time taken by Gforth (in other words, it shows the speedup
14292: factor that Gforth achieved over the other systems).
14293:
14294: @example
1.40 anton 14295: relative Win32- NT eforth This-
1.1 anton 14296: time Gforth Forth Forth eforth +opt PFE Forth TILE
1.40 anton 14297: sieve 1.00 1.58 1.30 1.58 0.97 1.80 3.63 9.79
14298: bubble 1.00 1.55 1.67 1.75 1.04 1.78 4.59
14299: matmul 1.00 1.67 1.53 1.66 0.84 1.79 4.63
14300: fib 1.00 1.75 1.53 1.40 0.99 1.99 3.43 4.93
1.1 anton 14301: @end example
14302:
1.26 crook 14303: You may be quite surprised by the good performance of Gforth when
14304: compared with systems written in assembly language. One important reason
14305: for the disappointing performance of these other systems is probably
14306: that they are not written optimally for the 486 (e.g., they use the
14307: @code{lods} instruction). In addition, Win32Forth uses a comfortable,
14308: but costly method for relocating the Forth image: like @code{cforth}, it
14309: computes the actual addresses at run time, resulting in two address
14310: computations per @code{NEXT} (@pxref{Image File Background}).
14311:
1.40 anton 14312: Only Eforth with the peephole optimizer performs comparable to
14313: Gforth. The speedups achieved with peephole optimization of threaded
14314: code are quite remarkable. Adding a peephole optimizer to Gforth should
14315: cause similar speedups.
1.1 anton 14316:
14317: The speedup of Gforth over PFE, ThisForth and TILE can be easily
14318: explained with the self-imposed restriction of the latter systems to
14319: standard C, which makes efficient threading impossible (however, the
1.4 anton 14320: measured implementation of PFE uses a GNU C extension: @pxref{Global Reg
1.1 anton 14321: Vars, , Defining Global Register Variables, gcc.info, GNU C Manual}).
14322: Moreover, current C compilers have a hard time optimizing other aspects
14323: of the ThisForth and the TILE source.
14324:
1.26 crook 14325: The performance of Gforth on 386 architecture processors varies widely
14326: with the version of @code{gcc} used. E.g., @code{gcc-2.5.8} failed to
14327: allocate any of the virtual machine registers into real machine
14328: registers by itself and would not work correctly with explicit register
1.40 anton 14329: declarations, giving a 1.5 times slower engine (on a 486DX2/66 running
1.26 crook 14330: the Sieve) than the one measured above.
1.1 anton 14331:
1.26 crook 14332: Note that there have been several releases of Win32Forth since the
14333: release presented here, so the results presented above may have little
1.40 anton 14334: predictive value for the performance of Win32Forth today (results for
14335: the current release on an i486DX2/66 are welcome).
1.1 anton 14336:
14337: @cindex @file{Benchres}
1.66 anton 14338: In
14339: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl&maierhofer95.ps.gz,
14340: Translating Forth to Efficient C}} by M. Anton Ertl and Martin
1.1 anton 14341: Maierhofer (presented at EuroForth '95), an indirect threaded version of
1.66 anton 14342: Gforth is compared with Win32Forth, NT Forth, PFE, ThisForth, and
14343: several native code systems; that version of Gforth is slower on a 486
14344: than the direct threaded version used here. You can find a newer version
14345: of these measurements at
1.47 crook 14346: @uref{http://www.complang.tuwien.ac.at/forth/performance.html}. You can
1.1 anton 14347: find numbers for Gforth on various machines in @file{Benchres}.
14348:
1.26 crook 14349: @c ******************************************************************
1.13 pazsan 14350: @node Binding to System Library, Cross Compiler, Engine, Top
1.14 pazsan 14351: @chapter Binding to System Library
1.13 pazsan 14352:
14353: @node Cross Compiler, Bugs, Binding to System Library, Top
1.14 pazsan 14354: @chapter Cross Compiler
1.47 crook 14355: @cindex @file{cross.fs}
14356: @cindex cross-compiler
14357: @cindex metacompiler
14358: @cindex target compiler
1.13 pazsan 14359:
1.46 pazsan 14360: The cross compiler is used to bootstrap a Forth kernel. Since Gforth is
14361: mostly written in Forth, including crucial parts like the outer
14362: interpreter and compiler, it needs compiled Forth code to get
14363: started. The cross compiler allows to create new images for other
14364: architectures, even running under another Forth system.
1.13 pazsan 14365:
14366: @menu
1.67 anton 14367: * Using the Cross Compiler::
14368: * How the Cross Compiler Works::
1.13 pazsan 14369: @end menu
14370:
1.21 crook 14371: @node Using the Cross Compiler, How the Cross Compiler Works, Cross Compiler, Cross Compiler
1.14 pazsan 14372: @section Using the Cross Compiler
1.46 pazsan 14373:
14374: The cross compiler uses a language that resembles Forth, but isn't. The
14375: main difference is that you can execute Forth code after definition,
14376: while you usually can't execute the code compiled by cross, because the
14377: code you are compiling is typically for a different computer than the
14378: one you are compiling on.
14379:
14380: The Makefile is already set up to allow you to create kernels for new
14381: architectures with a simple make command. The generic kernels using the
14382: GCC compiled virtual machine are created in the normal build process
14383: with @code{make}. To create a embedded Gforth executable for e.g. the
14384: 8086 processor (running on a DOS machine), type
14385:
14386: @example
14387: make kernl-8086.fi
14388: @end example
14389:
14390: This will use the machine description from the @file{arch/8086}
14391: directory to create a new kernel. A machine file may look like that:
14392:
14393: @example
14394: \ Parameter for target systems 06oct92py
14395:
14396: 4 Constant cell \ cell size in bytes
14397: 2 Constant cell<< \ cell shift to bytes
14398: 5 Constant cell>bit \ cell shift to bits
14399: 8 Constant bits/char \ bits per character
14400: 8 Constant bits/byte \ bits per byte [default: 8]
14401: 8 Constant float \ bytes per float
14402: 8 Constant /maxalign \ maximum alignment in bytes
14403: false Constant bigendian \ byte order
14404: ( true=big, false=little )
14405:
14406: include machpc.fs \ feature list
14407: @end example
14408:
14409: This part is obligatory for the cross compiler itself, the feature list
14410: is used by the kernel to conditionally compile some features in and out,
14411: depending on whether the target supports these features.
14412:
14413: There are some optional features, if you define your own primitives,
14414: have an assembler, or need special, nonstandard preparation to make the
14415: boot process work. @code{asm-include} include an assembler,
14416: @code{prims-include} includes primitives, and @code{>boot} prepares for
14417: booting.
14418:
14419: @example
14420: : asm-include ." Include assembler" cr
14421: s" arch/8086/asm.fs" included ;
14422:
14423: : prims-include ." Include primitives" cr
14424: s" arch/8086/prim.fs" included ;
14425:
14426: : >boot ." Prepare booting" cr
14427: s" ' boot >body into-forth 1+ !" evaluate ;
14428: @end example
14429:
14430: These words are used as sort of macro during the cross compilation in
14431: the file @file{kernel/main.fs}. Instead of using this macros, it would
14432: be possible --- but more complicated --- to write a new kernel project
14433: file, too.
14434:
14435: @file{kernel/main.fs} expects the machine description file name on the
14436: stack; the cross compiler itself (@file{cross.fs}) assumes that either
14437: @code{mach-file} leaves a counted string on the stack, or
14438: @code{machine-file} leaves an address, count pair of the filename on the
14439: stack.
14440:
14441: The feature list is typically controlled using @code{SetValue}, generic
14442: files that are used by several projects can use @code{DefaultValue}
14443: instead. Both functions work like @code{Value}, when the value isn't
14444: defined, but @code{SetValue} works like @code{to} if the value is
14445: defined, and @code{DefaultValue} doesn't set anything, if the value is
14446: defined.
14447:
14448: @example
14449: \ generic mach file for pc gforth 03sep97jaw
14450:
14451: true DefaultValue NIL \ relocating
14452:
14453: >ENVIRON
14454:
14455: true DefaultValue file \ controls the presence of the
14456: \ file access wordset
14457: true DefaultValue OS \ flag to indicate a operating system
14458:
14459: true DefaultValue prims \ true: primitives are c-code
14460:
14461: true DefaultValue floating \ floating point wordset is present
14462:
14463: true DefaultValue glocals \ gforth locals are present
14464: \ will be loaded
14465: true DefaultValue dcomps \ double number comparisons
14466:
14467: true DefaultValue hash \ hashing primitives are loaded/present
14468:
14469: true DefaultValue xconds \ used together with glocals,
14470: \ special conditionals supporting gforths'
14471: \ local variables
14472: true DefaultValue header \ save a header information
14473:
14474: true DefaultValue backtrace \ enables backtrace code
14475:
14476: false DefaultValue ec
14477: false DefaultValue crlf
14478:
14479: cell 2 = [IF] &32 [ELSE] &256 [THEN] KB DefaultValue kernel-size
14480:
14481: &16 KB DefaultValue stack-size
14482: &15 KB &512 + DefaultValue fstack-size
14483: &15 KB DefaultValue rstack-size
14484: &14 KB &512 + DefaultValue lstack-size
14485: @end example
1.13 pazsan 14486:
1.48 anton 14487: @node How the Cross Compiler Works, , Using the Cross Compiler, Cross Compiler
1.14 pazsan 14488: @section How the Cross Compiler Works
1.13 pazsan 14489:
14490: @node Bugs, Origin, Cross Compiler, Top
1.21 crook 14491: @appendix Bugs
1.1 anton 14492: @cindex bug reporting
14493:
1.21 crook 14494: Known bugs are described in the file @file{BUGS} in the Gforth distribution.
1.1 anton 14495:
14496: If you find a bug, please send a bug report to
1.33 anton 14497: @email{bug-gforth@@gnu.org}. A bug report should include this
1.21 crook 14498: information:
14499:
14500: @itemize @bullet
14501: @item
14502: The Gforth version used (it is announced at the start of an
14503: interactive Gforth session).
14504: @item
14505: The machine and operating system (on Unix
14506: systems @code{uname -a} will report this information).
14507: @item
14508: The installation options (send the file @file{config.status}).
14509: @item
14510: A complete list of changes (if any) you (or your installer) have made to the
14511: Gforth sources.
14512: @item
14513: A program (or a sequence of keyboard commands) that reproduces the bug.
14514: @item
14515: A description of what you think constitutes the buggy behaviour.
14516: @end itemize
1.1 anton 14517:
14518: For a thorough guide on reporting bugs read @ref{Bug Reporting, , How
14519: to Report Bugs, gcc.info, GNU C Manual}.
14520:
14521:
1.21 crook 14522: @node Origin, Forth-related information, Bugs, Top
14523: @appendix Authors and Ancestors of Gforth
1.1 anton 14524:
14525: @section Authors and Contributors
14526: @cindex authors of Gforth
14527: @cindex contributors to Gforth
14528:
14529: The Gforth project was started in mid-1992 by Bernd Paysan and Anton
14530: Ertl. The third major author was Jens Wilke. Lennart Benschop (who was
14531: one of Gforth's first users, in mid-1993) and Stuart Ramsden inspired us
14532: with their continuous feedback. Lennart Benshop contributed
14533: @file{glosgen.fs}, while Stuart Ramsden has been working on automatic
14534: support for calling C libraries. Helpful comments also came from Paul
14535: Kleinrubatscher, Christian Pirker, Dirk Zoller, Marcel Hendrix, John
1.58 anton 14536: Wavrik, Barrie Stott, Marc de Groot, Jorge Acerada, Bruce Hoyt, and
14537: Robert Epprecht. Since the release of Gforth-0.2.1 there were also
14538: helpful comments from many others; thank you all, sorry for not listing
14539: you here (but digging through my mailbox to extract your names is on my
14540: to-do list). Since the release of Gforth-0.4.0 Neal Crook worked on the
14541: manual.
1.1 anton 14542:
14543: Gforth also owes a lot to the authors of the tools we used (GCC, CVS,
14544: and autoconf, among others), and to the creators of the Internet: Gforth
1.21 crook 14545: was developed across the Internet, and its authors did not meet
1.20 pazsan 14546: physically for the first 4 years of development.
1.1 anton 14547:
14548: @section Pedigree
1.26 crook 14549: @cindex pedigree of Gforth
1.1 anton 14550:
1.20 pazsan 14551: Gforth descends from bigFORTH (1993) and fig-Forth. Gforth and PFE (by
1.1 anton 14552: Dirk Zoller) will cross-fertilize each other. Of course, a significant
14553: part of the design of Gforth was prescribed by ANS Forth.
14554:
1.20 pazsan 14555: Bernd Paysan wrote bigFORTH, a descendent from TurboForth, an unreleased
1.1 anton 14556: 32 bit native code version of VolksForth for the Atari ST, written
14557: mostly by Dietrich Weineck.
14558:
14559: VolksForth descends from F83. It was written by Klaus Schleisiek, Bernd
14560: Pennemann, Georg Rehfeld and Dietrich Weineck for the C64 (called
14561: UltraForth there) in the mid-80s and ported to the Atari ST in 1986.
14562:
14563: Henry Laxen and Mike Perry wrote F83 as a model implementation of the
14564: Forth-83 standard. !! Pedigree? When?
14565:
14566: A team led by Bill Ragsdale implemented fig-Forth on many processors in
14567: 1979. Robert Selzer and Bill Ragsdale developed the original
14568: implementation of fig-Forth for the 6502 based on microForth.
14569:
14570: The principal architect of microForth was Dean Sanderson. microForth was
14571: FORTH, Inc.'s first off-the-shelf product. It was developed in 1976 for
14572: the 1802, and subsequently implemented on the 8080, the 6800 and the
14573: Z80.
14574:
14575: All earlier Forth systems were custom-made, usually by Charles Moore,
14576: who discovered (as he puts it) Forth during the late 60s. The first full
14577: Forth existed in 1971.
14578:
14579: A part of the information in this section comes from @cite{The Evolution
14580: of Forth} by Elizabeth D. Rather, Donald R. Colburn and Charles
14581: H. Moore, presented at the HOPL-II conference and preprinted in SIGPLAN
14582: Notices 28(3), 1993. You can find more historical and genealogical
14583: information about Forth there.
14584:
1.21 crook 14585: @node Forth-related information, Word Index, Origin, Top
14586: @appendix Other Forth-related information
14587: @cindex Forth-related information
14588:
14589: @menu
1.67 anton 14590: * Internet resources::
14591: * Books::
14592: * The Forth Interest Group::
14593: * Conferences::
1.21 crook 14594: @end menu
14595:
14596:
14597: @node Internet resources, Books, Forth-related information, Forth-related information
14598: @section Internet resources
1.26 crook 14599: @cindex internet resources
1.21 crook 14600:
14601: @cindex comp.lang.forth
14602: @cindex frequently asked questions
1.45 crook 14603: There is an active news group (comp.lang.forth) discussing Forth and
1.21 crook 14604: Forth-related issues. A frequently-asked-questions (FAQ) list
1.45 crook 14605: is posted to the news group regularly, and archived at these sites:
1.21 crook 14606:
14607: @itemize @bullet
14608: @item
1.47 crook 14609: @uref{ftp://rtfm.mit.edu/pub/usenet-by-group/comp.lang.forth/}
1.21 crook 14610: @item
1.47 crook 14611: @uref{ftp://ftp.forth.org/pub/Forth/FAQ/}
1.21 crook 14612: @end itemize
14613:
14614: The FAQ list should be considered mandatory reading before posting to
1.45 crook 14615: the news group.
1.21 crook 14616:
14617: Here are some other web sites holding Forth-related material:
14618:
14619: @itemize @bullet
14620: @item
1.47 crook 14621: @uref{http://www.taygeta.com/forth.html} -- Skip Carter's Forth pages.
1.21 crook 14622: @item
1.47 crook 14623: @uref{http://www.jwdt.com/~paysan/gforth.html} -- the Gforth home page.
1.21 crook 14624: @item
1.47 crook 14625: @uref{http://www.minerva.com/uathena.htm} -- home of ANS Forth Standard.
1.21 crook 14626: @item
1.47 crook 14627: @uref{http://dec.bournemouth.ac.uk/forth/index.html} -- the Forth
1.21 crook 14628: Research page, including links to the Journal of Forth Application and
14629: Research (JFAR) and a searchable Forth bibliography.
14630: @end itemize
14631:
14632:
14633: @node Books, The Forth Interest Group, Internet resources, Forth-related information
14634: @section Books
1.26 crook 14635: @cindex books on Forth
1.21 crook 14636:
14637: As the Standard is relatively new, there are not many books out yet. It
14638: is not recommended to learn Forth by using Gforth and a book that is not
14639: written for ANS Forth, as you will not know your mistakes from the
14640: deviations of the book. However, books based on the Forth-83 standard
14641: should be ok, because ANS Forth is primarily an extension of Forth-83.
1.44 crook 14642: Refer to the Forth FAQ for details of Forth-related books.
1.21 crook 14643:
14644: @cindex standard document for ANS Forth
14645: @cindex ANS Forth document
14646: The definite reference if you want to write ANS Forth programs is, of
1.26 crook 14647: course, the ANS Forth document. It is available in printed form from the
1.21 crook 14648: National Standards Institute Sales Department (Tel.: USA (212) 642-4900;
14649: Fax.: USA (212) 302-1286) as document @cite{X3.215-1994} for about
14650: $200. You can also get it from Global Engineering Documents (Tel.: USA
14651: (800) 854-7179; Fax.: (303) 843-9880) for about $300.
14652:
14653: @cite{dpANS6}, the last draft of the standard, which was then submitted
14654: to ANSI for publication is available electronically and for free in some
14655: MS Word format, and it has been converted to HTML
1.47 crook 14656: (@uref{http://www.taygeta.com/forth/dpans.html}; this HTML version also
1.44 crook 14657: includes the answers to Requests for Interpretation (RFIs). Some
14658: pointers to these versions can be found through
1.47 crook 14659: @*@uref{http://www.complang.tuwien.ac.at/projects/forth.html}.
1.44 crook 14660:
1.21 crook 14661:
14662: @node The Forth Interest Group, Conferences, Books, Forth-related information
14663: @section The Forth Interest Group
14664: @cindex Forth interest group (FIG)
14665:
14666: The Forth Interest Group (FIG) is a world-wide, non-profit,
1.26 crook 14667: member-supported organisation. It publishes a regular magazine,
14668: @var{FORTH Dimensions}, and offers other benefits of membership. You can
14669: contact the FIG through their office email address:
14670: @email{office@@forth.org} or by visiting their web site at
1.47 crook 14671: @uref{http://www.forth.org/}. This web site also includes links to FIG
1.26 crook 14672: chapters in other countries and American cities
1.47 crook 14673: (@uref{http://www.forth.org/chapters.html}).
1.21 crook 14674:
1.48 anton 14675: @node Conferences, , The Forth Interest Group, Forth-related information
1.21 crook 14676: @section Conferences
14677: @cindex Conferences
14678:
14679: There are several regular conferences related to Forth. They are all
1.26 crook 14680: well-publicised in @var{FORTH Dimensions} and on the comp.lang.forth
1.45 crook 14681: news group:
1.21 crook 14682:
14683: @itemize @bullet
14684: @item
14685: FORML -- the Forth modification laboratory convenes every year near
14686: Monterey, California.
14687: @item
14688: The Rochester Forth Conference -- an annual conference traditionally
14689: held in Rochester, New York.
14690: @item
14691: EuroForth -- this European conference takes place annually.
14692: @end itemize
14693:
14694:
1.41 anton 14695: @node Word Index, Name Index, Forth-related information, Top
1.1 anton 14696: @unnumbered Word Index
14697:
1.26 crook 14698: This index is a list of Forth words that have ``glossary'' entries
14699: within this manual. Each word is listed with its stack effect and
14700: wordset.
1.1 anton 14701:
14702: @printindex fn
14703:
1.41 anton 14704: @node Name Index, Concept Index, Word Index, Top
14705: @unnumbered Name Index
14706:
14707: This index is a list of Forth words that have ``glossary'' entries
14708: within this manual.
14709:
14710: @printindex ky
14711:
14712: @node Concept Index, , Name Index, Top
1.1 anton 14713: @unnumbered Concept and Word Index
14714:
1.26 crook 14715: Not all entries listed in this index are present verbatim in the
14716: text. This index also duplicates, in abbreviated form, all of the words
14717: listed in the Word Index (only the names are listed for the words here).
1.1 anton 14718:
14719: @printindex cp
14720:
14721: @contents
14722: @bye
14723:
FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>