Annotation of gforth/doc/gforth.ds, revision 1.81
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
170: * Concept Index:: A menu covering many topics
1.12 anton 171:
1.78 anton 172: @detailmenu --- The Detailed Node Listing ---
1.12 anton 173:
1.29 crook 174: Gforth Environment
175:
1.32 anton 176: * Invoking Gforth:: Getting in
177: * Leaving Gforth:: Getting out
178: * Command-line editing::
1.48 anton 179: * Environment variables:: that affect how Gforth starts up
1.32 anton 180: * Gforth Files:: What gets installed and where
1.48 anton 181: * Startup speed:: When 35ms is not fast enough ...
182:
183: Forth Tutorial
184:
185: * Starting Gforth Tutorial::
186: * Syntax Tutorial::
187: * Crash Course Tutorial::
188: * Stack Tutorial::
189: * Arithmetics Tutorial::
190: * Stack Manipulation Tutorial::
191: * Using files for Forth code Tutorial::
192: * Comments Tutorial::
193: * Colon Definitions Tutorial::
194: * Decompilation Tutorial::
195: * Stack-Effect Comments Tutorial::
196: * Types Tutorial::
197: * Factoring Tutorial::
198: * Designing the stack effect Tutorial::
199: * Local Variables Tutorial::
200: * Conditional execution Tutorial::
201: * Flags and Comparisons Tutorial::
202: * General Loops Tutorial::
203: * Counted loops Tutorial::
204: * Recursion Tutorial::
205: * Leaving definitions or loops Tutorial::
206: * Return Stack Tutorial::
207: * Memory Tutorial::
208: * Characters and Strings Tutorial::
209: * Alignment Tutorial::
210: * Interpretation and Compilation Semantics and Immediacy Tutorial::
211: * Execution Tokens Tutorial::
212: * Exceptions Tutorial::
213: * Defining Words Tutorial::
214: * Arrays and Records Tutorial::
215: * POSTPONE Tutorial::
216: * Literal Tutorial::
217: * Advanced macros Tutorial::
218: * Compilation Tokens Tutorial::
219: * Wordlists and Search Order Tutorial::
1.29 crook 220:
1.24 anton 221: An Introduction to ANS Forth
222:
1.67 anton 223: * Introducing the Text Interpreter::
224: * Stacks and Postfix notation::
225: * Your first definition::
226: * How does that work?::
227: * Forth is written in Forth::
228: * Review - elements of a Forth system::
229: * Where to go next::
230: * Exercises::
1.24 anton 231:
1.12 anton 232: Forth Words
233:
234: * Notation::
1.65 anton 235: * Case insensitivity::
236: * Comments::
237: * Boolean Flags::
1.12 anton 238: * Arithmetic::
239: * Stack Manipulation::
240: * Memory::
241: * Control Structures::
242: * Defining Words::
1.65 anton 243: * Interpretation and Compilation Semantics::
1.47 crook 244: * Tokens for Words::
1.81 ! anton 245: * Compiling 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::
1.78 anton 252: * Locals::
253: * Structures::
254: * Object-oriented Forth::
1.12 anton 255: * Programming Tools::
256: * Assembler and Code Words::
257: * Threading Words::
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: Locals
360:
361: * Gforth locals::
362: * ANS Forth locals::
363:
364: Gforth locals
365:
366: * Where are locals visible by name?::
367: * How long do locals live?::
1.78 anton 368: * Locals programming style::
369: * Locals implementation::
1.26 crook 370:
1.12 anton 371: Structures
372:
373: * Why explicit structure support?::
374: * Structure Usage::
375: * Structure Naming Convention::
376: * Structure Implementation::
377: * Structure Glossary::
378:
379: Object-oriented Forth
380:
1.48 anton 381: * Why object-oriented programming?::
382: * Object-Oriented Terminology::
383: * Objects::
384: * OOF::
385: * Mini-OOF::
1.23 crook 386: * Comparison with other object models::
1.12 anton 387:
1.24 anton 388: The @file{objects.fs} model
1.12 anton 389:
390: * Properties of the Objects model::
391: * Basic Objects Usage::
1.41 anton 392: * The Objects base class::
1.12 anton 393: * Creating objects::
394: * Object-Oriented Programming Style::
395: * Class Binding::
396: * Method conveniences::
397: * Classes and Scoping::
1.41 anton 398: * Dividing classes::
1.12 anton 399: * Object Interfaces::
400: * Objects Implementation::
401: * Objects Glossary::
402:
1.24 anton 403: The @file{oof.fs} model
1.12 anton 404:
1.67 anton 405: * Properties of the OOF model::
406: * Basic OOF Usage::
407: * The OOF base class::
408: * Class Declaration::
409: * Class Implementation::
1.12 anton 410:
1.24 anton 411: The @file{mini-oof.fs} model
1.23 crook 412:
1.48 anton 413: * Basic Mini-OOF Usage::
414: * Mini-OOF Example::
415: * Mini-OOF Implementation::
1.23 crook 416:
1.78 anton 417: Programming Tools
418:
419: * Examining::
420: * Forgetting words::
421: * Debugging:: Simple and quick.
422: * Assertions:: Making your programs self-checking.
423: * Singlestep Debugger:: Executing your program word by word.
424:
425: Assembler and Code Words
426:
427: * Code and ;code::
428: * Common Assembler:: Assembler Syntax
429: * Common Disassembler::
430: * 386 Assembler:: Deviations and special cases
431: * Alpha Assembler:: Deviations and special cases
432: * MIPS assembler:: Deviations and special cases
433: * Other assemblers:: How to write them
434:
1.12 anton 435: Tools
436:
437: * ANS Report:: Report the words used, sorted by wordset.
438:
439: ANS conformance
440:
441: * The Core Words::
442: * The optional Block word set::
443: * The optional Double Number word set::
444: * The optional Exception word set::
445: * The optional Facility word set::
446: * The optional File-Access word set::
447: * The optional Floating-Point word set::
448: * The optional Locals word set::
449: * The optional Memory-Allocation word set::
450: * The optional Programming-Tools word set::
451: * The optional Search-Order word set::
452:
453: The Core Words
454:
455: * core-idef:: Implementation Defined Options
456: * core-ambcond:: Ambiguous Conditions
457: * core-other:: Other System Documentation
458:
459: The optional Block word set
460:
461: * block-idef:: Implementation Defined Options
462: * block-ambcond:: Ambiguous Conditions
463: * block-other:: Other System Documentation
464:
465: The optional Double Number word set
466:
467: * double-ambcond:: Ambiguous Conditions
468:
469: The optional Exception word set
470:
471: * exception-idef:: Implementation Defined Options
472:
473: The optional Facility word set
474:
475: * facility-idef:: Implementation Defined Options
476: * facility-ambcond:: Ambiguous Conditions
477:
478: The optional File-Access word set
479:
480: * file-idef:: Implementation Defined Options
481: * file-ambcond:: Ambiguous Conditions
482:
483: The optional Floating-Point word set
484:
485: * floating-idef:: Implementation Defined Options
486: * floating-ambcond:: Ambiguous Conditions
487:
488: The optional Locals word set
489:
490: * locals-idef:: Implementation Defined Options
491: * locals-ambcond:: Ambiguous Conditions
492:
493: The optional Memory-Allocation word set
494:
495: * memory-idef:: Implementation Defined Options
496:
497: The optional Programming-Tools word set
498:
499: * programming-idef:: Implementation Defined Options
500: * programming-ambcond:: Ambiguous Conditions
501:
502: The optional Search-Order word set
503:
504: * search-idef:: Implementation Defined Options
505: * search-ambcond:: Ambiguous Conditions
506:
507: Image Files
508:
1.24 anton 509: * Image Licensing Issues:: Distribution terms for images.
510: * Image File Background:: Why have image files?
1.67 anton 511: * Non-Relocatable Image Files:: don't always work.
1.24 anton 512: * Data-Relocatable Image Files:: are better.
1.67 anton 513: * Fully Relocatable Image Files:: better yet.
1.24 anton 514: * Stack and Dictionary Sizes:: Setting the default sizes for an image.
1.32 anton 515: * Running Image Files:: @code{gforth -i @i{file}} or @i{file}.
1.24 anton 516: * Modifying the Startup Sequence:: and turnkey applications.
1.12 anton 517:
518: Fully Relocatable Image Files
519:
1.27 crook 520: * gforthmi:: The normal way
1.12 anton 521: * cross.fs:: The hard way
522:
523: Engine
524:
525: * Portability::
526: * Threading::
527: * Primitives::
528: * Performance::
529:
530: Threading
531:
532: * Scheduling::
533: * Direct or Indirect Threaded?::
534: * DOES>::
535:
536: Primitives
537:
538: * Automatic Generation::
539: * TOS Optimization::
540: * Produced code::
1.13 pazsan 541:
542: Cross Compiler
543:
1.67 anton 544: * Using the Cross Compiler::
545: * How the Cross Compiler Works::
1.13 pazsan 546:
1.24 anton 547: @end detailmenu
1.1 anton 548: @end menu
549:
1.26 crook 550: @node License, Goals, Top, Top
1.1 anton 551: @unnumbered GNU GENERAL PUBLIC LICENSE
552: @center Version 2, June 1991
553:
554: @display
555: Copyright @copyright{} 1989, 1991 Free Software Foundation, Inc.
556: 675 Mass Ave, Cambridge, MA 02139, USA
557:
558: Everyone is permitted to copy and distribute verbatim copies
559: of this license document, but changing it is not allowed.
560: @end display
561:
562: @unnumberedsec Preamble
563:
564: The licenses for most software are designed to take away your
565: freedom to share and change it. By contrast, the GNU General Public
566: License is intended to guarantee your freedom to share and change free
567: software---to make sure the software is free for all its users. This
568: General Public License applies to most of the Free Software
569: Foundation's software and to any other program whose authors commit to
570: using it. (Some other Free Software Foundation software is covered by
571: the GNU Library General Public License instead.) You can apply it to
572: your programs, too.
573:
574: When we speak of free software, we are referring to freedom, not
575: price. Our General Public Licenses are designed to make sure that you
576: have the freedom to distribute copies of free software (and charge for
577: this service if you wish), that you receive source code or can get it
578: if you want it, that you can change the software or use pieces of it
579: in new free programs; and that you know you can do these things.
580:
581: To protect your rights, we need to make restrictions that forbid
582: anyone to deny you these rights or to ask you to surrender the rights.
583: These restrictions translate to certain responsibilities for you if you
584: distribute copies of the software, or if you modify it.
585:
586: For example, if you distribute copies of such a program, whether
587: gratis or for a fee, you must give the recipients all the rights that
588: you have. You must make sure that they, too, receive or can get the
589: source code. And you must show them these terms so they know their
590: rights.
591:
592: We protect your rights with two steps: (1) copyright the software, and
593: (2) offer you this license which gives you legal permission to copy,
594: distribute and/or modify the software.
595:
596: Also, for each author's protection and ours, we want to make certain
597: that everyone understands that there is no warranty for this free
598: software. If the software is modified by someone else and passed on, we
599: want its recipients to know that what they have is not the original, so
600: that any problems introduced by others will not reflect on the original
601: authors' reputations.
602:
603: Finally, any free program is threatened constantly by software
604: patents. We wish to avoid the danger that redistributors of a free
605: program will individually obtain patent licenses, in effect making the
606: program proprietary. To prevent this, we have made it clear that any
607: patent must be licensed for everyone's free use or not licensed at all.
608:
609: The precise terms and conditions for copying, distribution and
610: modification follow.
611:
612: @iftex
613: @unnumberedsec TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
614: @end iftex
1.49 anton 615: @ifnottex
1.1 anton 616: @center TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
1.49 anton 617: @end ifnottex
1.1 anton 618:
619: @enumerate 0
620: @item
621: This License applies to any program or other work which contains
622: a notice placed by the copyright holder saying it may be distributed
623: under the terms of this General Public License. The ``Program'', below,
624: refers to any such program or work, and a ``work based on the Program''
625: means either the Program or any derivative work under copyright law:
626: that is to say, a work containing the Program or a portion of it,
627: either verbatim or with modifications and/or translated into another
628: language. (Hereinafter, translation is included without limitation in
629: the term ``modification''.) Each licensee is addressed as ``you''.
630:
631: Activities other than copying, distribution and modification are not
632: covered by this License; they are outside its scope. The act of
633: running the Program is not restricted, and the output from the Program
634: is covered only if its contents constitute a work based on the
635: Program (independent of having been made by running the Program).
636: Whether that is true depends on what the Program does.
637:
638: @item
639: You may copy and distribute verbatim copies of the Program's
640: source code as you receive it, in any medium, provided that you
641: conspicuously and appropriately publish on each copy an appropriate
642: copyright notice and disclaimer of warranty; keep intact all the
643: notices that refer to this License and to the absence of any warranty;
644: and give any other recipients of the Program a copy of this License
645: along with the Program.
646:
647: You may charge a fee for the physical act of transferring a copy, and
648: you may at your option offer warranty protection in exchange for a fee.
649:
650: @item
651: You may modify your copy or copies of the Program or any portion
652: of it, thus forming a work based on the Program, and copy and
653: distribute such modifications or work under the terms of Section 1
654: above, provided that you also meet all of these conditions:
655:
656: @enumerate a
657: @item
658: You must cause the modified files to carry prominent notices
659: stating that you changed the files and the date of any change.
660:
661: @item
662: You must cause any work that you distribute or publish, that in
663: whole or in part contains or is derived from the Program or any
664: part thereof, to be licensed as a whole at no charge to all third
665: parties under the terms of this License.
666:
667: @item
668: If the modified program normally reads commands interactively
669: when run, you must cause it, when started running for such
670: interactive use in the most ordinary way, to print or display an
671: announcement including an appropriate copyright notice and a
672: notice that there is no warranty (or else, saying that you provide
673: a warranty) and that users may redistribute the program under
674: these conditions, and telling the user how to view a copy of this
675: License. (Exception: if the Program itself is interactive but
676: does not normally print such an announcement, your work based on
677: the Program is not required to print an announcement.)
678: @end enumerate
679:
680: These requirements apply to the modified work as a whole. If
681: identifiable sections of that work are not derived from the Program,
682: and can be reasonably considered independent and separate works in
683: themselves, then this License, and its terms, do not apply to those
684: sections when you distribute them as separate works. But when you
685: distribute the same sections as part of a whole which is a work based
686: on the Program, the distribution of the whole must be on the terms of
687: this License, whose permissions for other licensees extend to the
688: entire whole, and thus to each and every part regardless of who wrote it.
689:
690: Thus, it is not the intent of this section to claim rights or contest
691: your rights to work written entirely by you; rather, the intent is to
692: exercise the right to control the distribution of derivative or
693: collective works based on the Program.
694:
695: In addition, mere aggregation of another work not based on the Program
696: with the Program (or with a work based on the Program) on a volume of
697: a storage or distribution medium does not bring the other work under
698: the scope of this License.
699:
700: @item
701: You may copy and distribute the Program (or a work based on it,
702: under Section 2) in object code or executable form under the terms of
703: Sections 1 and 2 above provided that you also do one of the following:
704:
705: @enumerate a
706: @item
707: Accompany it with the complete corresponding machine-readable
708: source code, which must be distributed under the terms of Sections
709: 1 and 2 above on a medium customarily used for software interchange; or,
710:
711: @item
712: Accompany it with a written offer, valid for at least three
713: years, to give any third party, for a charge no more than your
714: cost of physically performing source distribution, a complete
715: machine-readable copy of the corresponding source code, to be
716: distributed under the terms of Sections 1 and 2 above on a medium
717: customarily used for software interchange; or,
718:
719: @item
720: Accompany it with the information you received as to the offer
721: to distribute corresponding source code. (This alternative is
722: allowed only for noncommercial distribution and only if you
723: received the program in object code or executable form with such
724: an offer, in accord with Subsection b above.)
725: @end enumerate
726:
727: The source code for a work means the preferred form of the work for
728: making modifications to it. For an executable work, complete source
729: code means all the source code for all modules it contains, plus any
730: associated interface definition files, plus the scripts used to
731: control compilation and installation of the executable. However, as a
732: special exception, the source code distributed need not include
733: anything that is normally distributed (in either source or binary
734: form) with the major components (compiler, kernel, and so on) of the
735: operating system on which the executable runs, unless that component
736: itself accompanies the executable.
737:
738: If distribution of executable or object code is made by offering
739: access to copy from a designated place, then offering equivalent
740: access to copy the source code from the same place counts as
741: distribution of the source code, even though third parties are not
742: compelled to copy the source along with the object code.
743:
744: @item
745: You may not copy, modify, sublicense, or distribute the Program
746: except as expressly provided under this License. Any attempt
747: otherwise to copy, modify, sublicense or distribute the Program is
748: void, and will automatically terminate your rights under this License.
749: However, parties who have received copies, or rights, from you under
750: this License will not have their licenses terminated so long as such
751: parties remain in full compliance.
752:
753: @item
754: You are not required to accept this License, since you have not
755: signed it. However, nothing else grants you permission to modify or
756: distribute the Program or its derivative works. These actions are
757: prohibited by law if you do not accept this License. Therefore, by
758: modifying or distributing the Program (or any work based on the
759: Program), you indicate your acceptance of this License to do so, and
760: all its terms and conditions for copying, distributing or modifying
761: the Program or works based on it.
762:
763: @item
764: Each time you redistribute the Program (or any work based on the
765: Program), the recipient automatically receives a license from the
766: original licensor to copy, distribute or modify the Program subject to
767: these terms and conditions. You may not impose any further
768: restrictions on the recipients' exercise of the rights granted herein.
769: You are not responsible for enforcing compliance by third parties to
770: this License.
771:
772: @item
773: If, as a consequence of a court judgment or allegation of patent
774: infringement or for any other reason (not limited to patent issues),
775: conditions are imposed on you (whether by court order, agreement or
776: otherwise) that contradict the conditions of this License, they do not
777: excuse you from the conditions of this License. If you cannot
778: distribute so as to satisfy simultaneously your obligations under this
779: License and any other pertinent obligations, then as a consequence you
780: may not distribute the Program at all. For example, if a patent
781: license would not permit royalty-free redistribution of the Program by
782: all those who receive copies directly or indirectly through you, then
783: the only way you could satisfy both it and this License would be to
784: refrain entirely from distribution of the Program.
785:
786: If any portion of this section is held invalid or unenforceable under
787: any particular circumstance, the balance of the section is intended to
788: apply and the section as a whole is intended to apply in other
789: circumstances.
790:
791: It is not the purpose of this section to induce you to infringe any
792: patents or other property right claims or to contest validity of any
793: such claims; this section has the sole purpose of protecting the
794: integrity of the free software distribution system, which is
795: implemented by public license practices. Many people have made
796: generous contributions to the wide range of software distributed
797: through that system in reliance on consistent application of that
798: system; it is up to the author/donor to decide if he or she is willing
799: to distribute software through any other system and a licensee cannot
800: impose that choice.
801:
802: This section is intended to make thoroughly clear what is believed to
803: be a consequence of the rest of this License.
804:
805: @item
806: If the distribution and/or use of the Program is restricted in
807: certain countries either by patents or by copyrighted interfaces, the
808: original copyright holder who places the Program under this License
809: may add an explicit geographical distribution limitation excluding
810: those countries, so that distribution is permitted only in or among
811: countries not thus excluded. In such case, this License incorporates
812: the limitation as if written in the body of this License.
813:
814: @item
815: The Free Software Foundation may publish revised and/or new versions
816: of the General Public License from time to time. Such new versions will
817: be similar in spirit to the present version, but may differ in detail to
818: address new problems or concerns.
819:
820: Each version is given a distinguishing version number. If the Program
821: specifies a version number of this License which applies to it and ``any
822: later version'', you have the option of following the terms and conditions
823: either of that version or of any later version published by the Free
824: Software Foundation. If the Program does not specify a version number of
825: this License, you may choose any version ever published by the Free Software
826: Foundation.
827:
828: @item
829: If you wish to incorporate parts of the Program into other free
830: programs whose distribution conditions are different, write to the author
831: to ask for permission. For software which is copyrighted by the Free
832: Software Foundation, write to the Free Software Foundation; we sometimes
833: make exceptions for this. Our decision will be guided by the two goals
834: of preserving the free status of all derivatives of our free software and
835: of promoting the sharing and reuse of software generally.
836:
837: @iftex
838: @heading NO WARRANTY
839: @end iftex
1.49 anton 840: @ifnottex
1.1 anton 841: @center NO WARRANTY
1.49 anton 842: @end ifnottex
1.1 anton 843:
844: @item
845: BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY
846: FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW. EXCEPT WHEN
847: OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES
848: PROVIDE THE PROGRAM ``AS IS'' WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED
849: OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
850: MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS
851: TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU. SHOULD THE
852: PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING,
853: REPAIR OR CORRECTION.
854:
855: @item
856: IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
857: WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR
858: REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES,
859: INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING
860: OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED
861: TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY
862: YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER
863: PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE
864: POSSIBILITY OF SUCH DAMAGES.
865: @end enumerate
866:
867: @iftex
868: @heading END OF TERMS AND CONDITIONS
869: @end iftex
1.49 anton 870: @ifnottex
1.1 anton 871: @center END OF TERMS AND CONDITIONS
1.49 anton 872: @end ifnottex
1.1 anton 873:
874: @page
875: @unnumberedsec How to Apply These Terms to Your New Programs
876:
877: If you develop a new program, and you want it to be of the greatest
878: possible use to the public, the best way to achieve this is to make it
879: free software which everyone can redistribute and change under these terms.
880:
881: To do so, attach the following notices to the program. It is safest
882: to attach them to the start of each source file to most effectively
883: convey the exclusion of warranty; and each file should have at least
884: the ``copyright'' line and a pointer to where the full notice is found.
885:
886: @smallexample
887: @var{one line to give the program's name and a brief idea of what it does.}
888: Copyright (C) 19@var{yy} @var{name of author}
889:
890: This program is free software; you can redistribute it and/or modify
891: it under the terms of the GNU General Public License as published by
892: the Free Software Foundation; either version 2 of the License, or
893: (at your option) any later version.
894:
895: This program is distributed in the hope that it will be useful,
896: but WITHOUT ANY WARRANTY; without even the implied warranty of
897: MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
898: GNU General Public License for more details.
899:
900: You should have received a copy of the GNU General Public License
901: along with this program; if not, write to the Free Software
902: Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
903: @end smallexample
904:
905: Also add information on how to contact you by electronic and paper mail.
906:
907: If the program is interactive, make it output a short notice like this
908: when it starts in an interactive mode:
909:
910: @smallexample
911: Gnomovision version 69, Copyright (C) 19@var{yy} @var{name of author}
912: Gnomovision comes with ABSOLUTELY NO WARRANTY; for details
913: type `show w'.
914: This is free software, and you are welcome to redistribute it
915: under certain conditions; type `show c' for details.
916: @end smallexample
917:
918: The hypothetical commands @samp{show w} and @samp{show c} should show
919: the appropriate parts of the General Public License. Of course, the
920: commands you use may be called something other than @samp{show w} and
921: @samp{show c}; they could even be mouse-clicks or menu items---whatever
922: suits your program.
923:
924: You should also get your employer (if you work as a programmer) or your
925: school, if any, to sign a ``copyright disclaimer'' for the program, if
926: necessary. Here is a sample; alter the names:
927:
928: @smallexample
929: Yoyodyne, Inc., hereby disclaims all copyright interest in the program
930: `Gnomovision' (which makes passes at compilers) written by James Hacker.
931:
932: @var{signature of Ty Coon}, 1 April 1989
933: Ty Coon, President of Vice
934: @end smallexample
935:
936: This General Public License does not permit incorporating your program into
937: proprietary programs. If your program is a subroutine library, you may
938: consider it more useful to permit linking proprietary applications with the
939: library. If this is what you want to do, use the GNU Library General
940: Public License instead of this License.
941:
942: @iftex
943: @unnumbered Preface
944: @cindex Preface
1.21 crook 945: This manual documents Gforth. Some introductory material is provided for
946: readers who are unfamiliar with Forth or who are migrating to Gforth
947: from other Forth compilers. However, this manual is primarily a
948: reference manual.
1.1 anton 949: @end iftex
950:
1.28 crook 951: @comment TODO much more blurb here.
1.26 crook 952:
953: @c ******************************************************************
1.29 crook 954: @node Goals, Gforth Environment, License, Top
1.26 crook 955: @comment node-name, next, previous, up
956: @chapter Goals of Gforth
957: @cindex goals of the Gforth project
958: The goal of the Gforth Project is to develop a standard model for
959: ANS Forth. This can be split into several subgoals:
960:
961: @itemize @bullet
962: @item
963: Gforth should conform to the ANS Forth Standard.
964: @item
965: It should be a model, i.e. it should define all the
966: implementation-dependent things.
967: @item
968: It should become standard, i.e. widely accepted and used. This goal
969: is the most difficult one.
970: @end itemize
971:
972: To achieve these goals Gforth should be
973: @itemize @bullet
974: @item
975: Similar to previous models (fig-Forth, F83)
976: @item
977: Powerful. It should provide for all the things that are considered
978: necessary today and even some that are not yet considered necessary.
979: @item
980: Efficient. It should not get the reputation of being exceptionally
981: slow.
982: @item
983: Free.
984: @item
985: Available on many machines/easy to port.
986: @end itemize
987:
988: Have we achieved these goals? Gforth conforms to the ANS Forth
989: standard. It may be considered a model, but we have not yet documented
990: which parts of the model are stable and which parts we are likely to
991: change. It certainly has not yet become a de facto standard, but it
992: appears to be quite popular. It has some similarities to and some
993: differences from previous models. It has some powerful features, but not
994: yet everything that we envisioned. We certainly have achieved our
1.65 anton 995: execution speed goals (@pxref{Performance})@footnote{However, in 1998
996: the bar was raised when the major commercial Forth vendors switched to
997: native code compilers.}. It is free and available on many machines.
1.29 crook 998:
1.26 crook 999: @c ******************************************************************
1.48 anton 1000: @node Gforth Environment, Tutorial, Goals, Top
1.29 crook 1001: @chapter Gforth Environment
1002: @cindex Gforth environment
1.21 crook 1003:
1.45 crook 1004: Note: ultimately, the Gforth man page will be auto-generated from the
1.29 crook 1005: material in this chapter.
1.21 crook 1006:
1007: @menu
1.29 crook 1008: * Invoking Gforth:: Getting in
1009: * Leaving Gforth:: Getting out
1010: * Command-line editing::
1.48 anton 1011: * Environment variables:: that affect how Gforth starts up
1.29 crook 1012: * Gforth Files:: What gets installed and where
1.48 anton 1013: * Startup speed:: When 35ms is not fast enough ...
1.21 crook 1014: @end menu
1015:
1.49 anton 1016: For related information about the creation of images see @ref{Image Files}.
1.29 crook 1017:
1.21 crook 1018: @comment ----------------------------------------------
1.48 anton 1019: @node Invoking Gforth, Leaving Gforth, Gforth Environment, Gforth Environment
1.29 crook 1020: @section Invoking Gforth
1021: @cindex invoking Gforth
1022: @cindex running Gforth
1023: @cindex command-line options
1024: @cindex options on the command line
1025: @cindex flags on the command line
1.21 crook 1026:
1.30 anton 1027: Gforth is made up of two parts; an executable ``engine'' (named
1028: @file{gforth} or @file{gforth-fast}) and an image file. To start it, you
1029: will usually just say @code{gforth} -- this automatically loads the
1030: default image file @file{gforth.fi}. In many other cases the default
1031: Gforth image will be invoked like this:
1.21 crook 1032: @example
1.30 anton 1033: gforth [file | -e forth-code] ...
1.21 crook 1034: @end example
1.29 crook 1035: @noindent
1036: This interprets the contents of the files and the Forth code in the order they
1037: are given.
1.21 crook 1038:
1.30 anton 1039: In addition to the @file{gforth} engine, there is also an engine called
1040: @file{gforth-fast}, which is faster, but gives less informative error
1041: messages (@pxref{Error messages}).
1042:
1.29 crook 1043: In general, the command line looks like this:
1.21 crook 1044:
1045: @example
1.30 anton 1046: gforth[-fast] [engine options] [image options]
1.21 crook 1047: @end example
1048:
1.30 anton 1049: The engine options must come before the rest of the command
1.29 crook 1050: line. They are:
1.26 crook 1051:
1.29 crook 1052: @table @code
1053: @cindex -i, command-line option
1054: @cindex --image-file, command-line option
1055: @item --image-file @i{file}
1056: @itemx -i @i{file}
1057: Loads the Forth image @i{file} instead of the default
1058: @file{gforth.fi} (@pxref{Image Files}).
1.21 crook 1059:
1.39 anton 1060: @cindex --appl-image, command-line option
1061: @item --appl-image @i{file}
1062: Loads the image @i{file} and leaves all further command-line arguments
1.65 anton 1063: to the image (instead of processing them as engine options). This is
1064: useful for building executable application images on Unix, built with
1.39 anton 1065: @code{gforthmi --application ...}.
1066:
1.29 crook 1067: @cindex --path, command-line option
1068: @cindex -p, command-line option
1069: @item --path @i{path}
1070: @itemx -p @i{path}
1071: Uses @i{path} for searching the image file and Forth source code files
1072: instead of the default in the environment variable @code{GFORTHPATH} or
1073: the path specified at installation time (e.g.,
1074: @file{/usr/local/share/gforth/0.2.0:.}). A path is given as a list of
1075: directories, separated by @samp{:} (on Unix) or @samp{;} (on other OSs).
1.21 crook 1076:
1.29 crook 1077: @cindex --dictionary-size, command-line option
1078: @cindex -m, command-line option
1079: @cindex @i{size} parameters for command-line options
1080: @cindex size of the dictionary and the stacks
1081: @item --dictionary-size @i{size}
1082: @itemx -m @i{size}
1083: Allocate @i{size} space for the Forth dictionary space instead of
1084: using the default specified in the image (typically 256K). The
1085: @i{size} specification for this and subsequent options consists of
1086: an integer and a unit (e.g.,
1087: @code{4M}). The unit can be one of @code{b} (bytes), @code{e} (element
1088: size, in this case Cells), @code{k} (kilobytes), @code{M} (Megabytes),
1089: @code{G} (Gigabytes), and @code{T} (Terabytes). If no unit is specified,
1090: @code{e} is used.
1.21 crook 1091:
1.29 crook 1092: @cindex --data-stack-size, command-line option
1093: @cindex -d, command-line option
1094: @item --data-stack-size @i{size}
1095: @itemx -d @i{size}
1096: Allocate @i{size} space for the data stack instead of using the
1097: default specified in the image (typically 16K).
1.21 crook 1098:
1.29 crook 1099: @cindex --return-stack-size, command-line option
1100: @cindex -r, command-line option
1101: @item --return-stack-size @i{size}
1102: @itemx -r @i{size}
1103: Allocate @i{size} space for the return stack instead of using the
1104: default specified in the image (typically 15K).
1.21 crook 1105:
1.29 crook 1106: @cindex --fp-stack-size, command-line option
1107: @cindex -f, command-line option
1108: @item --fp-stack-size @i{size}
1109: @itemx -f @i{size}
1110: Allocate @i{size} space for the floating point stack instead of
1111: using the default specified in the image (typically 15.5K). In this case
1112: the unit specifier @code{e} refers to floating point numbers.
1.21 crook 1113:
1.48 anton 1114: @cindex --locals-stack-size, command-line option
1115: @cindex -l, command-line option
1116: @item --locals-stack-size @i{size}
1117: @itemx -l @i{size}
1118: Allocate @i{size} space for the locals stack instead of using the
1119: default specified in the image (typically 14.5K).
1120:
1121: @cindex -h, command-line option
1122: @cindex --help, command-line option
1123: @item --help
1124: @itemx -h
1125: Print a message about the command-line options
1126:
1127: @cindex -v, command-line option
1128: @cindex --version, command-line option
1129: @item --version
1130: @itemx -v
1131: Print version and exit
1132:
1133: @cindex --debug, command-line option
1134: @item --debug
1135: Print some information useful for debugging on startup.
1136:
1137: @cindex --offset-image, command-line option
1138: @item --offset-image
1139: Start the dictionary at a slightly different position than would be used
1140: otherwise (useful for creating data-relocatable images,
1141: @pxref{Data-Relocatable Image Files}).
1142:
1143: @cindex --no-offset-im, command-line option
1144: @item --no-offset-im
1145: Start the dictionary at the normal position.
1146:
1147: @cindex --clear-dictionary, command-line option
1148: @item --clear-dictionary
1149: Initialize all bytes in the dictionary to 0 before loading the image
1150: (@pxref{Data-Relocatable Image Files}).
1151:
1152: @cindex --die-on-signal, command-line-option
1153: @item --die-on-signal
1154: Normally Gforth handles most signals (e.g., the user interrupt SIGINT,
1155: or the segmentation violation SIGSEGV) by translating it into a Forth
1156: @code{THROW}. With this option, Gforth exits if it receives such a
1157: signal. This option is useful when the engine and/or the image might be
1158: severely broken (such that it causes another signal before recovering
1159: from the first); this option avoids endless loops in such cases.
1160: @end table
1161:
1162: @cindex loading files at startup
1163: @cindex executing code on startup
1164: @cindex batch processing with Gforth
1165: As explained above, the image-specific command-line arguments for the
1166: default image @file{gforth.fi} consist of a sequence of filenames and
1167: @code{-e @var{forth-code}} options that are interpreted in the sequence
1168: in which they are given. The @code{-e @var{forth-code}} or
1169: @code{--evaluate @var{forth-code}} option evaluates the Forth
1170: code. This option takes only one argument; if you want to evaluate more
1171: Forth words, you have to quote them or use @code{-e} several times. To exit
1172: after processing the command line (instead of entering interactive mode)
1173: append @code{-e bye} to the command line.
1174:
1175: @cindex versions, invoking other versions of Gforth
1176: If you have several versions of Gforth installed, @code{gforth} will
1177: invoke the version that was installed last. @code{gforth-@i{version}}
1178: invokes a specific version. If your environment contains the variable
1179: @code{GFORTHPATH}, you may want to override it by using the
1180: @code{--path} option.
1181:
1182: Not yet implemented:
1183: On startup the system first executes the system initialization file
1184: (unless the option @code{--no-init-file} is given; note that the system
1185: resulting from using this option may not be ANS Forth conformant). Then
1186: the user initialization file @file{.gforth.fs} is executed, unless the
1.62 crook 1187: option @code{--no-rc} is given; this file is searched for in @file{.},
1.48 anton 1188: then in @file{~}, then in the normal path (see above).
1189:
1190:
1191:
1192: @comment ----------------------------------------------
1193: @node Leaving Gforth, Command-line editing, Invoking Gforth, Gforth Environment
1194: @section Leaving Gforth
1195: @cindex Gforth - leaving
1196: @cindex leaving Gforth
1197:
1198: You can leave Gforth by typing @code{bye} or @kbd{Ctrl-d} (at the start
1199: of a line) or (if you invoked Gforth with the @code{--die-on-signal}
1200: option) @kbd{Ctrl-c}. When you leave Gforth, all of your definitions and
1.49 anton 1201: data are discarded. For ways of saving the state of the system before
1202: leaving Gforth see @ref{Image Files}.
1.48 anton 1203:
1204: doc-bye
1205:
1206:
1207: @comment ----------------------------------------------
1.65 anton 1208: @node Command-line editing, Environment variables, Leaving Gforth, Gforth Environment
1.48 anton 1209: @section Command-line editing
1210: @cindex command-line editing
1211:
1212: Gforth maintains a history file that records every line that you type to
1213: the text interpreter. This file is preserved between sessions, and is
1214: used to provide a command-line recall facility; if you type @kbd{Ctrl-P}
1215: repeatedly you can recall successively older commands from this (or
1216: previous) session(s). The full list of command-line editing facilities is:
1217:
1218: @itemize @bullet
1219: @item
1220: @kbd{Ctrl-p} (``previous'') (or up-arrow) to recall successively older
1221: commands from the history buffer.
1222: @item
1223: @kbd{Ctrl-n} (``next'') (or down-arrow) to recall successively newer commands
1224: from the history buffer.
1225: @item
1226: @kbd{Ctrl-f} (or right-arrow) to move the cursor right, non-destructively.
1227: @item
1228: @kbd{Ctrl-b} (or left-arrow) to move the cursor left, non-destructively.
1229: @item
1230: @kbd{Ctrl-h} (backspace) to delete the character to the left of the cursor,
1231: closing up the line.
1232: @item
1233: @kbd{Ctrl-k} to delete (``kill'') from the cursor to the end of the line.
1234: @item
1235: @kbd{Ctrl-a} to move the cursor to the start of the line.
1236: @item
1237: @kbd{Ctrl-e} to move the cursor to the end of the line.
1238: @item
1239: @key{RET} (@kbd{Ctrl-m}) or @key{LFD} (@kbd{Ctrl-j}) to submit the current
1240: line.
1241: @item
1242: @key{TAB} to step through all possible full-word completions of the word
1243: currently being typed.
1244: @item
1.65 anton 1245: @kbd{Ctrl-d} on an empty line line to terminate Gforth (gracefully,
1246: using @code{bye}).
1247: @item
1248: @kbd{Ctrl-x} (or @code{Ctrl-d} on a non-empty line) to delete the
1249: character under the cursor.
1.48 anton 1250: @end itemize
1251:
1252: When editing, displayable characters are inserted to the left of the
1253: cursor position; the line is always in ``insert'' (as opposed to
1254: ``overstrike'') mode.
1255:
1256: @cindex history file
1257: @cindex @file{.gforth-history}
1258: On Unix systems, the history file is @file{~/.gforth-history} by
1259: default@footnote{i.e. it is stored in the user's home directory.}. You
1260: can find out the name and location of your history file using:
1261:
1262: @example
1263: history-file type \ Unix-class systems
1264:
1265: history-file type \ Other systems
1266: history-dir type
1267: @end example
1268:
1269: If you enter long definitions by hand, you can use a text editor to
1270: paste them out of the history file into a Forth source file for reuse at
1271: a later time.
1272:
1273: Gforth never trims the size of the history file, so you should do this
1274: periodically, if necessary.
1275:
1276: @comment this is all defined in history.fs
1277: @comment NAC TODO the ctrl-D behaviour can either do a bye or a beep.. how is that option
1278: @comment chosen?
1279:
1280:
1281: @comment ----------------------------------------------
1.65 anton 1282: @node Environment variables, Gforth Files, Command-line editing, Gforth Environment
1.48 anton 1283: @section Environment variables
1284: @cindex environment variables
1285:
1286: Gforth uses these environment variables:
1287:
1288: @itemize @bullet
1289: @item
1290: @cindex @code{GFORTHHIST} -- environment variable
1291: @code{GFORTHHIST} -- (Unix systems only) specifies the directory in which to
1292: open/create the history file, @file{.gforth-history}. Default:
1293: @code{$HOME}.
1294:
1295: @item
1296: @cindex @code{GFORTHPATH} -- environment variable
1297: @code{GFORTHPATH} -- specifies the path used when searching for the gforth image file and
1298: for Forth source-code files.
1299:
1300: @item
1301: @cindex @code{GFORTH} -- environment variable
1.49 anton 1302: @code{GFORTH} -- used by @file{gforthmi}, @xref{gforthmi}.
1.48 anton 1303:
1304: @item
1305: @cindex @code{GFORTHD} -- environment variable
1.62 crook 1306: @code{GFORTHD} -- used by @file{gforthmi}, @xref{gforthmi}.
1.48 anton 1307:
1308: @item
1309: @cindex @code{TMP}, @code{TEMP} - environment variable
1310: @code{TMP}, @code{TEMP} - (non-Unix systems only) used as a potential
1311: location for the history file.
1312: @end itemize
1313:
1314: @comment also POSIXELY_CORRECT LINES COLUMNS HOME but no interest in
1315: @comment mentioning these.
1316:
1317: All the Gforth environment variables default to sensible values if they
1318: are not set.
1319:
1320:
1321: @comment ----------------------------------------------
1322: @node Gforth Files, Startup speed, Environment variables, Gforth Environment
1323: @section Gforth files
1324: @cindex Gforth files
1325:
1326: When you install Gforth on a Unix system, it installs files in these
1327: locations by default:
1328:
1329: @itemize @bullet
1330: @item
1331: @file{/usr/local/bin/gforth}
1332: @item
1333: @file{/usr/local/bin/gforthmi}
1334: @item
1335: @file{/usr/local/man/man1/gforth.1} - man page.
1336: @item
1337: @file{/usr/local/info} - the Info version of this manual.
1338: @item
1339: @file{/usr/local/lib/gforth/<version>/...} - Gforth @file{.fi} files.
1340: @item
1341: @file{/usr/local/share/gforth/<version>/TAGS} - Emacs TAGS file.
1342: @item
1343: @file{/usr/local/share/gforth/<version>/...} - Gforth source files.
1344: @item
1345: @file{.../emacs/site-lisp/gforth.el} - Emacs gforth mode.
1346: @end itemize
1347:
1348: You can select different places for installation by using
1349: @code{configure} options (listed with @code{configure --help}).
1350:
1351: @comment ----------------------------------------------
1352: @node Startup speed, , Gforth Files, Gforth Environment
1353: @section Startup speed
1354: @cindex Startup speed
1355: @cindex speed, startup
1356:
1357: If Gforth is used for CGI scripts or in shell scripts, its startup
1358: speed may become a problem. On a 300MHz 21064a under Linux-2.2.13 with
1359: glibc-2.0.7, @code{gforth -e bye} takes about 24.6ms user and 11.3ms
1360: system time.
1361:
1362: If startup speed is a problem, you may consider the following ways to
1363: improve it; or you may consider ways to reduce the number of startups
1.62 crook 1364: (for example, by using Fast-CGI).
1.48 anton 1365:
1366: The first step to improve startup speed is to statically link Gforth, by
1367: building it with @code{XLDFLAGS=-static}. This requires more memory for
1368: the code and will therefore slow down the first invocation, but
1369: subsequent invocations avoid the dynamic linking overhead. Another
1370: disadvantage is that Gforth won't profit from library upgrades. As a
1371: result, @code{gforth-static -e bye} takes about 17.1ms user and
1372: 8.2ms system time.
1373:
1374: The next step to improve startup speed is to use a non-relocatable image
1.65 anton 1375: (@pxref{Non-Relocatable Image Files}). You can create this image with
1.48 anton 1376: @code{gforth -e "savesystem gforthnr.fi bye"} and later use it with
1377: @code{gforth -i gforthnr.fi ...}. This avoids the relocation overhead
1378: and a part of the copy-on-write overhead. The disadvantage is that the
1.62 crook 1379: non-relocatable image does not work if the OS gives Gforth a different
1.48 anton 1380: address for the dictionary, for whatever reason; so you better provide a
1381: fallback on a relocatable image. @code{gforth-static -i gforthnr.fi -e
1382: bye} takes about 15.3ms user and 7.5ms system time.
1383:
1384: The final step is to disable dictionary hashing in Gforth. Gforth
1385: builds the hash table on startup, which takes much of the startup
1386: overhead. You can do this by commenting out the @code{include hash.fs}
1387: in @file{startup.fs} and everything that requires @file{hash.fs} (at the
1388: moment @file{table.fs} and @file{ekey.fs}) and then doing @code{make}.
1389: The disadvantages are that functionality like @code{table} and
1390: @code{ekey} is missing and that text interpretation (e.g., compiling)
1391: now takes much longer. So, you should only use this method if there is
1392: no significant text interpretation to perform (the script should be
1.62 crook 1393: compiled into the image, amongst other things). @code{gforth-static -i
1.48 anton 1394: gforthnrnh.fi -e bye} takes about 2.1ms user and 6.1ms system time.
1395:
1396: @c ******************************************************************
1397: @node Tutorial, Introduction, Gforth Environment, Top
1398: @chapter Forth Tutorial
1399: @cindex Tutorial
1400: @cindex Forth Tutorial
1401:
1.67 anton 1402: @c Topics from nac's Introduction that could be mentioned:
1403: @c press <ret> after each line
1404: @c Prompt
1405: @c numbers vs. words in dictionary on text interpretation
1406: @c what happens on redefinition
1407: @c parsing words (in particular, defining words)
1408:
1.62 crook 1409: This tutorial can be used with any ANS-compliant Forth; any
1410: Gforth-specific features are marked as such and you can skip them if you
1411: work with another Forth. This tutorial does not explain all features of
1412: Forth, just enough to get you started and give you some ideas about the
1413: facilities available in Forth. Read the rest of the manual and the
1414: standard when you are through this.
1.48 anton 1415:
1416: The intended way to use this tutorial is that you work through it while
1417: sitting in front of the console, take a look at the examples and predict
1418: what they will do, then try them out; if the outcome is not as expected,
1419: find out why (e.g., by trying out variations of the example), so you
1420: understand what's going on. There are also some assignments that you
1421: should solve.
1422:
1423: This tutorial assumes that you have programmed before and know what,
1424: e.g., a loop is.
1425:
1426: @c !! explain compat library
1427:
1428: @menu
1429: * Starting Gforth Tutorial::
1430: * Syntax Tutorial::
1431: * Crash Course Tutorial::
1432: * Stack Tutorial::
1433: * Arithmetics Tutorial::
1434: * Stack Manipulation Tutorial::
1435: * Using files for Forth code Tutorial::
1436: * Comments Tutorial::
1437: * Colon Definitions Tutorial::
1438: * Decompilation Tutorial::
1439: * Stack-Effect Comments Tutorial::
1440: * Types Tutorial::
1441: * Factoring Tutorial::
1442: * Designing the stack effect Tutorial::
1443: * Local Variables Tutorial::
1444: * Conditional execution Tutorial::
1445: * Flags and Comparisons Tutorial::
1446: * General Loops Tutorial::
1447: * Counted loops Tutorial::
1448: * Recursion Tutorial::
1449: * Leaving definitions or loops Tutorial::
1450: * Return Stack Tutorial::
1451: * Memory Tutorial::
1452: * Characters and Strings Tutorial::
1453: * Alignment Tutorial::
1454: * Interpretation and Compilation Semantics and Immediacy Tutorial::
1455: * Execution Tokens Tutorial::
1456: * Exceptions Tutorial::
1457: * Defining Words Tutorial::
1458: * Arrays and Records Tutorial::
1459: * POSTPONE Tutorial::
1460: * Literal Tutorial::
1461: * Advanced macros Tutorial::
1462: * Compilation Tokens Tutorial::
1463: * Wordlists and Search Order Tutorial::
1464: @end menu
1465:
1466: @node Starting Gforth Tutorial, Syntax Tutorial, Tutorial, Tutorial
1467: @section Starting Gforth
1.66 anton 1468: @cindex starting Gforth tutorial
1.48 anton 1469: You can start Gforth by typing its name:
1470:
1471: @example
1472: gforth
1473: @end example
1474:
1475: That puts you into interactive mode; you can leave Gforth by typing
1476: @code{bye}. While in Gforth, you can edit the command line and access
1477: the command line history with cursor keys, similar to bash.
1478:
1479:
1480: @node Syntax Tutorial, Crash Course Tutorial, Starting Gforth Tutorial, Tutorial
1481: @section Syntax
1.66 anton 1482: @cindex syntax tutorial
1.48 anton 1483:
1484: A @dfn{word} is a sequence of arbitrary characters (expcept white
1485: space). Words are separated by white space. E.g., each of the
1486: following lines contains exactly one word:
1487:
1488: @example
1489: word
1490: !@@#$%^&*()
1491: 1234567890
1492: 5!a
1493: @end example
1494:
1495: A frequent beginner's error is to leave away necessary white space,
1496: resulting in an error like @samp{Undefined word}; so if you see such an
1497: error, check if you have put spaces wherever necessary.
1498:
1499: @example
1500: ." hello, world" \ correct
1501: ."hello, world" \ gives an "Undefined word" error
1502: @end example
1503:
1.65 anton 1504: Gforth and most other Forth systems ignore differences in case (they are
1.48 anton 1505: case-insensitive), i.e., @samp{word} is the same as @samp{Word}. If
1506: your system is case-sensitive, you may have to type all the examples
1507: given here in upper case.
1508:
1509:
1510: @node Crash Course Tutorial, Stack Tutorial, Syntax Tutorial, Tutorial
1511: @section Crash Course
1512:
1513: Type
1514:
1515: @example
1516: 0 0 !
1517: here execute
1518: ' catch >body 20 erase abort
1519: ' (quit) >body 20 erase
1520: @end example
1521:
1522: The last two examples are guaranteed to destroy parts of Gforth (and
1523: most other systems), so you better leave Gforth afterwards (if it has
1524: not finished by itself). On some systems you may have to kill gforth
1525: from outside (e.g., in Unix with @code{kill}).
1526:
1527: Now that you know how to produce crashes (and that there's not much to
1528: them), let's learn how to produce meaningful programs.
1529:
1530:
1531: @node Stack Tutorial, Arithmetics Tutorial, Crash Course Tutorial, Tutorial
1532: @section Stack
1.66 anton 1533: @cindex stack tutorial
1.48 anton 1534:
1535: The most obvious feature of Forth is the stack. When you type in a
1536: number, it is pushed on the stack. You can display the content of the
1537: stack with @code{.s}.
1538:
1539: @example
1540: 1 2 .s
1541: 3 .s
1542: @end example
1543:
1544: @code{.s} displays the top-of-stack to the right, i.e., the numbers
1545: appear in @code{.s} output as they appeared in the input.
1546:
1547: You can print the top of stack element with @code{.}.
1548:
1549: @example
1550: 1 2 3 . . .
1551: @end example
1552:
1553: In general, words consume their stack arguments (@code{.s} is an
1554: exception).
1555:
1556: @assignment
1557: What does the stack contain after @code{5 6 7 .}?
1558: @endassignment
1559:
1560:
1561: @node Arithmetics Tutorial, Stack Manipulation Tutorial, Stack Tutorial, Tutorial
1562: @section Arithmetics
1.66 anton 1563: @cindex arithmetics tutorial
1.48 anton 1564:
1565: The words @code{+}, @code{-}, @code{*}, @code{/}, and @code{mod} always
1566: operate on the top two stack items:
1567:
1568: @example
1.67 anton 1569: 2 2 .s
1570: + .s
1571: .
1.48 anton 1572: 2 1 - .
1573: 7 3 mod .
1574: @end example
1575:
1576: The operands of @code{-}, @code{/}, and @code{mod} are in the same order
1577: as in the corresponding infix expression (this is generally the case in
1578: Forth).
1579:
1580: Parentheses are superfluous (and not available), because the order of
1581: the words unambiguously determines the order of evaluation and the
1582: operands:
1583:
1584: @example
1585: 3 4 + 5 * .
1586: 3 4 5 * + .
1587: @end example
1588:
1589: @assignment
1590: What are the infix expressions corresponding to the Forth code above?
1591: Write @code{6-7*8+9} in Forth notation@footnote{This notation is also
1592: known as Postfix or RPN (Reverse Polish Notation).}.
1593: @endassignment
1594:
1595: To change the sign, use @code{negate}:
1596:
1597: @example
1598: 2 negate .
1599: @end example
1600:
1601: @assignment
1602: Convert -(-3)*4-5 to Forth.
1603: @endassignment
1604:
1605: @code{/mod} performs both @code{/} and @code{mod}.
1606:
1607: @example
1608: 7 3 /mod . .
1609: @end example
1610:
1.66 anton 1611: Reference: @ref{Arithmetic}.
1612:
1613:
1.48 anton 1614: @node Stack Manipulation Tutorial, Using files for Forth code Tutorial, Arithmetics Tutorial, Tutorial
1615: @section Stack Manipulation
1.66 anton 1616: @cindex stack manipulation tutorial
1.48 anton 1617:
1618: Stack manipulation words rearrange the data on the stack.
1619:
1620: @example
1621: 1 .s drop .s
1622: 1 .s dup .s drop drop .s
1623: 1 2 .s over .s drop drop drop
1624: 1 2 .s swap .s drop drop
1625: 1 2 3 .s rot .s drop drop drop
1626: @end example
1627:
1628: These are the most important stack manipulation words. There are also
1629: variants that manipulate twice as many stack items:
1630:
1631: @example
1632: 1 2 3 4 .s 2swap .s 2drop 2drop
1633: @end example
1634:
1635: Two more stack manipulation words are:
1636:
1637: @example
1638: 1 2 .s nip .s drop
1639: 1 2 .s tuck .s 2drop drop
1640: @end example
1641:
1642: @assignment
1643: Replace @code{nip} and @code{tuck} with combinations of other stack
1644: manipulation words.
1645:
1646: @example
1647: Given: How do you get:
1648: 1 2 3 3 2 1
1649: 1 2 3 1 2 3 2
1650: 1 2 3 1 2 3 3
1651: 1 2 3 1 3 3
1652: 1 2 3 2 1 3
1653: 1 2 3 4 4 3 2 1
1654: 1 2 3 1 2 3 1 2 3
1655: 1 2 3 4 1 2 3 4 1 2
1656: 1 2 3
1657: 1 2 3 1 2 3 4
1658: 1 2 3 1 3
1659: @end example
1660: @endassignment
1661:
1662: @example
1663: 5 dup * .
1664: @end example
1665:
1666: @assignment
1667: Write 17^3 and 17^4 in Forth, without writing @code{17} more than once.
1668: Write a piece of Forth code that expects two numbers on the stack
1669: (@var{a} and @var{b}, with @var{b} on top) and computes
1670: @code{(a-b)(a+1)}.
1671: @endassignment
1672:
1.66 anton 1673: Reference: @ref{Stack Manipulation}.
1674:
1675:
1.48 anton 1676: @node Using files for Forth code Tutorial, Comments Tutorial, Stack Manipulation Tutorial, Tutorial
1677: @section Using files for Forth code
1.66 anton 1678: @cindex loading Forth code, tutorial
1679: @cindex files containing Forth code, tutorial
1.48 anton 1680:
1681: While working at the Forth command line is convenient for one-line
1682: examples and short one-off code, you probably want to store your source
1683: code in files for convenient editing and persistence. You can use your
1684: favourite editor (Gforth includes Emacs support, @pxref{Emacs and
1685: Gforth}) to create @var{file} and use
1686:
1687: @example
1688: s" @var{file}" included
1689: @end example
1690:
1691: to load it into your Forth system. The file name extension I use for
1692: Forth files is @samp{.fs}.
1693:
1694: You can easily start Gforth with some files loaded like this:
1695:
1696: @example
1697: gforth @var{file1} @var{file2}
1698: @end example
1699:
1700: If an error occurs during loading these files, Gforth terminates,
1701: whereas an error during @code{INCLUDED} within Gforth usually gives you
1702: a Gforth command line. Starting the Forth system every time gives you a
1703: clean start every time, without interference from the results of earlier
1704: tries.
1705:
1706: I often put all the tests in a file, then load the code and run the
1707: tests with
1708:
1709: @example
1710: gforth @var{code} @var{tests} -e bye
1711: @end example
1712:
1713: (often by performing this command with @kbd{C-x C-e} in Emacs). The
1714: @code{-e bye} ensures that Gforth terminates afterwards so that I can
1715: restart this command without ado.
1716:
1717: The advantage of this approach is that the tests can be repeated easily
1718: every time the program ist changed, making it easy to catch bugs
1719: introduced by the change.
1720:
1.66 anton 1721: Reference: @ref{Forth source files}.
1722:
1.48 anton 1723:
1724: @node Comments Tutorial, Colon Definitions Tutorial, Using files for Forth code Tutorial, Tutorial
1725: @section Comments
1.66 anton 1726: @cindex comments tutorial
1.48 anton 1727:
1728: @example
1729: \ That's a comment; it ends at the end of the line
1730: ( Another comment; it ends here: ) .s
1731: @end example
1732:
1733: @code{\} and @code{(} are ordinary Forth words and therefore have to be
1734: separated with white space from the following text.
1735:
1736: @example
1737: \This gives an "Undefined word" error
1738: @end example
1739:
1740: The first @code{)} ends a comment started with @code{(}, so you cannot
1741: nest @code{(}-comments; and you cannot comment out text containing a
1742: @code{)} with @code{( ... )}@footnote{therefore it's a good idea to
1743: avoid @code{)} in word names.}.
1744:
1745: I use @code{\}-comments for descriptive text and for commenting out code
1746: of one or more line; I use @code{(}-comments for describing the stack
1747: effect, the stack contents, or for commenting out sub-line pieces of
1748: code.
1749:
1750: The Emacs mode @file{gforth.el} (@pxref{Emacs and Gforth}) supports
1751: these uses by commenting out a region with @kbd{C-x \}, uncommenting a
1752: region with @kbd{C-u C-x \}, and filling a @code{\}-commented region
1753: with @kbd{M-q}.
1754:
1.66 anton 1755: Reference: @ref{Comments}.
1756:
1.48 anton 1757:
1758: @node Colon Definitions Tutorial, Decompilation Tutorial, Comments Tutorial, Tutorial
1759: @section Colon Definitions
1.66 anton 1760: @cindex colon definitions, tutorial
1761: @cindex definitions, tutorial
1762: @cindex procedures, tutorial
1763: @cindex functions, tutorial
1.48 anton 1764:
1765: are similar to procedures and functions in other programming languages.
1766:
1767: @example
1768: : squared ( n -- n^2 )
1769: dup * ;
1770: 5 squared .
1771: 7 squared .
1772: @end example
1773:
1774: @code{:} starts the colon definition; its name is @code{squared}. The
1775: following comment describes its stack effect. The words @code{dup *}
1776: are not executed, but compiled into the definition. @code{;} ends the
1777: colon definition.
1778:
1779: The newly-defined word can be used like any other word, including using
1780: it in other definitions:
1781:
1782: @example
1783: : cubed ( n -- n^3 )
1784: dup squared * ;
1785: -5 cubed .
1786: : fourth-power ( n -- n^4 )
1787: squared squared ;
1788: 3 fourth-power .
1789: @end example
1790:
1791: @assignment
1792: Write colon definitions for @code{nip}, @code{tuck}, @code{negate}, and
1793: @code{/mod} in terms of other Forth words, and check if they work (hint:
1794: test your tests on the originals first). Don't let the
1795: @samp{redefined}-Messages spook you, they are just warnings.
1796: @endassignment
1797:
1.66 anton 1798: Reference: @ref{Colon Definitions}.
1799:
1.48 anton 1800:
1801: @node Decompilation Tutorial, Stack-Effect Comments Tutorial, Colon Definitions Tutorial, Tutorial
1802: @section Decompilation
1.66 anton 1803: @cindex decompilation tutorial
1804: @cindex see tutorial
1.48 anton 1805:
1806: You can decompile colon definitions with @code{see}:
1807:
1808: @example
1809: see squared
1810: see cubed
1811: @end example
1812:
1813: In Gforth @code{see} shows you a reconstruction of the source code from
1814: the executable code. Informations that were present in the source, but
1815: not in the executable code, are lost (e.g., comments).
1816:
1.65 anton 1817: You can also decompile the predefined words:
1818:
1819: @example
1820: see .
1821: see +
1822: @end example
1823:
1824:
1.48 anton 1825: @node Stack-Effect Comments Tutorial, Types Tutorial, Decompilation Tutorial, Tutorial
1826: @section Stack-Effect Comments
1.66 anton 1827: @cindex stack-effect comments, tutorial
1828: @cindex --, tutorial
1.48 anton 1829: By convention the comment after the name of a definition describes the
1830: stack effect: The part in from of the @samp{--} describes the state of
1831: the stack before the execution of the definition, i.e., the parameters
1832: that are passed into the colon definition; the part behind the @samp{--}
1833: is the state of the stack after the execution of the definition, i.e.,
1834: the results of the definition. The stack comment only shows the top
1835: stack items that the definition accesses and/or changes.
1836:
1837: You should put a correct stack effect on every definition, even if it is
1838: just @code{( -- )}. You should also add some descriptive comment to
1839: more complicated words (I usually do this in the lines following
1840: @code{:}). If you don't do this, your code becomes unreadable (because
1841: you have to work through every definition before you can undertsand
1842: any).
1843:
1844: @assignment
1845: The stack effect of @code{swap} can be written like this: @code{x1 x2 --
1846: x2 x1}. Describe the stack effect of @code{-}, @code{drop}, @code{dup},
1847: @code{over}, @code{rot}, @code{nip}, and @code{tuck}. Hint: When you
1.65 anton 1848: are done, you can compare your stack effects to those in this manual
1.48 anton 1849: (@pxref{Word Index}).
1850: @endassignment
1851:
1852: Sometimes programmers put comments at various places in colon
1853: definitions that describe the contents of the stack at that place (stack
1854: comments); i.e., they are like the first part of a stack-effect
1855: comment. E.g.,
1856:
1857: @example
1858: : cubed ( n -- n^3 )
1859: dup squared ( n n^2 ) * ;
1860: @end example
1861:
1862: In this case the stack comment is pretty superfluous, because the word
1863: is simple enough. If you think it would be a good idea to add such a
1864: comment to increase readability, you should also consider factoring the
1865: word into several simpler words (@pxref{Factoring Tutorial,,
1.60 anton 1866: Factoring}), which typically eliminates the need for the stack comment;
1.48 anton 1867: however, if you decide not to refactor it, then having such a comment is
1868: better than not having it.
1869:
1870: The names of the stack items in stack-effect and stack comments in the
1871: standard, in this manual, and in many programs specify the type through
1872: a type prefix, similar to Fortran and Hungarian notation. The most
1873: frequent prefixes are:
1874:
1875: @table @code
1876: @item n
1877: signed integer
1878: @item u
1879: unsigned integer
1880: @item c
1881: character
1882: @item f
1883: Boolean flags, i.e. @code{false} or @code{true}.
1884: @item a-addr,a-
1885: Cell-aligned address
1886: @item c-addr,c-
1887: Char-aligned address (note that a Char may have two bytes in Windows NT)
1888: @item xt
1889: Execution token, same size as Cell
1890: @item w,x
1891: Cell, can contain an integer or an address. It usually takes 32, 64 or
1892: 16 bits (depending on your platform and Forth system). A cell is more
1893: commonly known as machine word, but the term @emph{word} already means
1894: something different in Forth.
1895: @item d
1896: signed double-cell integer
1897: @item ud
1898: unsigned double-cell integer
1899: @item r
1900: Float (on the FP stack)
1901: @end table
1902:
1903: You can find a more complete list in @ref{Notation}.
1904:
1905: @assignment
1906: Write stack-effect comments for all definitions you have written up to
1907: now.
1908: @endassignment
1909:
1910:
1911: @node Types Tutorial, Factoring Tutorial, Stack-Effect Comments Tutorial, Tutorial
1912: @section Types
1.66 anton 1913: @cindex types tutorial
1.48 anton 1914:
1915: In Forth the names of the operations are not overloaded; so similar
1916: operations on different types need different names; e.g., @code{+} adds
1917: integers, and you have to use @code{f+} to add floating-point numbers.
1918: The following prefixes are often used for related operations on
1919: different types:
1920:
1921: @table @code
1922: @item (none)
1923: signed integer
1924: @item u
1925: unsigned integer
1926: @item c
1927: character
1928: @item d
1929: signed double-cell integer
1930: @item ud, du
1931: unsigned double-cell integer
1932: @item 2
1933: two cells (not-necessarily double-cell numbers)
1934: @item m, um
1935: mixed single-cell and double-cell operations
1936: @item f
1937: floating-point (note that in stack comments @samp{f} represents flags,
1.66 anton 1938: and @samp{r} represents FP numbers).
1.48 anton 1939: @end table
1940:
1941: If there are no differences between the signed and the unsigned variant
1942: (e.g., for @code{+}), there is only the prefix-less variant.
1943:
1944: Forth does not perform type checking, neither at compile time, nor at
1945: run time. If you use the wrong oeration, the data are interpreted
1946: incorrectly:
1947:
1948: @example
1949: -1 u.
1950: @end example
1951:
1952: If you have only experience with type-checked languages until now, and
1953: have heard how important type-checking is, don't panic! In my
1954: experience (and that of other Forthers), type errors in Forth code are
1955: usually easy to find (once you get used to it), the increased vigilance
1956: of the programmer tends to catch some harder errors in addition to most
1957: type errors, and you never have to work around the type system, so in
1958: most situations the lack of type-checking seems to be a win (projects to
1959: add type checking to Forth have not caught on).
1960:
1961:
1962: @node Factoring Tutorial, Designing the stack effect Tutorial, Types Tutorial, Tutorial
1963: @section Factoring
1.66 anton 1964: @cindex factoring tutorial
1.48 anton 1965:
1966: If you try to write longer definitions, you will soon find it hard to
1967: keep track of the stack contents. Therefore, good Forth programmers
1968: tend to write only short definitions (e.g., three lines). The art of
1969: finding meaningful short definitions is known as factoring (as in
1970: factoring polynomials).
1971:
1972: Well-factored programs offer additional advantages: smaller, more
1973: general words, are easier to test and debug and can be reused more and
1974: better than larger, specialized words.
1975:
1976: So, if you run into difficulties with stack management, when writing
1977: code, try to define meaningful factors for the word, and define the word
1978: in terms of those. Even if a factor contains only two words, it is
1979: often helpful.
1980:
1.65 anton 1981: Good factoring is not easy, and it takes some practice to get the knack
1982: for it; but even experienced Forth programmers often don't find the
1983: right solution right away, but only when rewriting the program. So, if
1984: you don't come up with a good solution immediately, keep trying, don't
1985: despair.
1.48 anton 1986:
1987: @c example !!
1988:
1989:
1990: @node Designing the stack effect Tutorial, Local Variables Tutorial, Factoring Tutorial, Tutorial
1991: @section Designing the stack effect
1.66 anton 1992: @cindex Stack effect design, tutorial
1993: @cindex design of stack effects, tutorial
1.48 anton 1994:
1995: In other languages you can use an arbitrary order of parameters for a
1.65 anton 1996: function; and since there is only one result, you don't have to deal with
1.48 anton 1997: the order of results, either.
1998:
1999: In Forth (and other stack-based languages, e.g., Postscript) the
2000: parameter and result order of a definition is important and should be
2001: designed well. The general guideline is to design the stack effect such
2002: that the word is simple to use in most cases, even if that complicates
2003: the implementation of the word. Some concrete rules are:
2004:
2005: @itemize @bullet
2006:
2007: @item
2008: Words consume all of their parameters (e.g., @code{.}).
2009:
2010: @item
2011: If there is a convention on the order of parameters (e.g., from
2012: mathematics or another programming language), stick with it (e.g.,
2013: @code{-}).
2014:
2015: @item
2016: If one parameter usually requires only a short computation (e.g., it is
2017: a constant), pass it on the top of the stack. Conversely, parameters
2018: that usually require a long sequence of code to compute should be passed
2019: as the bottom (i.e., first) parameter. This makes the code easier to
2020: read, because reader does not need to keep track of the bottom item
2021: through a long sequence of code (or, alternatively, through stack
1.49 anton 2022: manipulations). E.g., @code{!} (store, @pxref{Memory}) expects the
1.48 anton 2023: address on top of the stack because it is usually simpler to compute
2024: than the stored value (often the address is just a variable).
2025:
2026: @item
2027: Similarly, results that are usually consumed quickly should be returned
2028: on the top of stack, whereas a result that is often used in long
2029: computations should be passed as bottom result. E.g., the file words
2030: like @code{open-file} return the error code on the top of stack, because
2031: it is usually consumed quickly by @code{throw}; moreover, the error code
2032: has to be checked before doing anything with the other results.
2033:
2034: @end itemize
2035:
2036: These rules are just general guidelines, don't lose sight of the overall
2037: goal to make the words easy to use. E.g., if the convention rule
2038: conflicts with the computation-length rule, you might decide in favour
2039: of the convention if the word will be used rarely, and in favour of the
2040: computation-length rule if the word will be used frequently (because
2041: with frequent use the cost of breaking the computation-length rule would
2042: be quite high, and frequent use makes it easier to remember an
2043: unconventional order).
2044:
2045: @c example !! structure package
2046:
1.65 anton 2047:
1.48 anton 2048: @node Local Variables Tutorial, Conditional execution Tutorial, Designing the stack effect Tutorial, Tutorial
2049: @section Local Variables
1.66 anton 2050: @cindex local variables, tutorial
1.48 anton 2051:
2052: You can define local variables (@emph{locals}) in a colon definition:
2053:
2054: @example
2055: : swap @{ a b -- b a @}
2056: b a ;
2057: 1 2 swap .s 2drop
2058: @end example
2059:
2060: (If your Forth system does not support this syntax, include
2061: @file{compat/anslocals.fs} first).
2062:
2063: In this example @code{@{ a b -- b a @}} is the locals definition; it
2064: takes two cells from the stack, puts the top of stack in @code{b} and
2065: the next stack element in @code{a}. @code{--} starts a comment ending
2066: with @code{@}}. After the locals definition, using the name of the
2067: local will push its value on the stack. You can leave the comment
2068: part (@code{-- b a}) away:
2069:
2070: @example
2071: : swap ( x1 x2 -- x2 x1 )
2072: @{ a b @} b a ;
2073: @end example
2074:
2075: In Gforth you can have several locals definitions, anywhere in a colon
2076: definition; in contrast, in a standard program you can have only one
2077: locals definition per colon definition, and that locals definition must
2078: be outside any controll structure.
2079:
2080: With locals you can write slightly longer definitions without running
2081: into stack trouble. However, I recommend trying to write colon
2082: definitions without locals for exercise purposes to help you gain the
2083: essential factoring skills.
2084:
2085: @assignment
2086: Rewrite your definitions until now with locals
2087: @endassignment
2088:
1.66 anton 2089: Reference: @ref{Locals}.
2090:
1.48 anton 2091:
2092: @node Conditional execution Tutorial, Flags and Comparisons Tutorial, Local Variables Tutorial, Tutorial
2093: @section Conditional execution
1.66 anton 2094: @cindex conditionals, tutorial
2095: @cindex if, tutorial
1.48 anton 2096:
2097: In Forth you can use control structures only inside colon definitions.
2098: An @code{if}-structure looks like this:
2099:
2100: @example
2101: : abs ( n1 -- +n2 )
2102: dup 0 < if
2103: negate
2104: endif ;
2105: 5 abs .
2106: -5 abs .
2107: @end example
2108:
2109: @code{if} takes a flag from the stack. If the flag is non-zero (true),
2110: the following code is performed, otherwise execution continues after the
1.51 pazsan 2111: @code{endif} (or @code{else}). @code{<} compares the top two stack
1.48 anton 2112: elements and prioduces a flag:
2113:
2114: @example
2115: 1 2 < .
2116: 2 1 < .
2117: 1 1 < .
2118: @end example
2119:
2120: Actually the standard name for @code{endif} is @code{then}. This
2121: tutorial presents the examples using @code{endif}, because this is often
2122: less confusing for people familiar with other programming languages
2123: where @code{then} has a different meaning. If your system does not have
2124: @code{endif}, define it with
2125:
2126: @example
2127: : endif postpone then ; immediate
2128: @end example
2129:
2130: You can optionally use an @code{else}-part:
2131:
2132: @example
2133: : min ( n1 n2 -- n )
2134: 2dup < if
2135: drop
2136: else
2137: nip
2138: endif ;
2139: 2 3 min .
2140: 3 2 min .
2141: @end example
2142:
2143: @assignment
2144: Write @code{min} without @code{else}-part (hint: what's the definition
2145: of @code{nip}?).
2146: @endassignment
2147:
1.66 anton 2148: Reference: @ref{Selection}.
2149:
1.48 anton 2150:
2151: @node Flags and Comparisons Tutorial, General Loops Tutorial, Conditional execution Tutorial, Tutorial
2152: @section Flags and Comparisons
1.66 anton 2153: @cindex flags tutorial
2154: @cindex comparison tutorial
1.48 anton 2155:
2156: In a false-flag all bits are clear (0 when interpreted as integer). In
2157: a canonical true-flag all bits are set (-1 as a twos-complement signed
2158: integer); in many contexts (e.g., @code{if}) any non-zero value is
2159: treated as true flag.
2160:
2161: @example
2162: false .
2163: true .
2164: true hex u. decimal
2165: @end example
2166:
2167: Comparison words produce canonical flags:
2168:
2169: @example
2170: 1 1 = .
2171: 1 0= .
2172: 0 1 < .
2173: 0 0 < .
2174: -1 1 u< . \ type error, u< interprets -1 as large unsigned number
2175: -1 1 < .
2176: @end example
2177:
1.66 anton 2178: Gforth supports all combinations of the prefixes @code{0 u d d0 du f f0}
2179: (or none) and the comparisons @code{= <> < > <= >=}. Only a part of
2180: these combinations are standard (for details see the standard,
2181: @ref{Numeric comparison}, @ref{Floating Point} or @ref{Word Index}).
1.48 anton 2182:
2183: You can use @code{and or xor invert} can be used as operations on
2184: canonical flags. Actually they are bitwise operations:
2185:
2186: @example
2187: 1 2 and .
2188: 1 2 or .
2189: 1 3 xor .
2190: 1 invert .
2191: @end example
2192:
2193: You can convert a zero/non-zero flag into a canonical flag with
2194: @code{0<>} (and complement it on the way with @code{0=}).
2195:
2196: @example
2197: 1 0= .
2198: 1 0<> .
2199: @end example
2200:
1.65 anton 2201: You can use the all-bits-set feature of canonical flags and the bitwise
1.48 anton 2202: operation of the Boolean operations to avoid @code{if}s:
2203:
2204: @example
2205: : foo ( n1 -- n2 )
2206: 0= if
2207: 14
2208: else
2209: 0
2210: endif ;
2211: 0 foo .
2212: 1 foo .
2213:
2214: : foo ( n1 -- n2 )
2215: 0= 14 and ;
2216: 0 foo .
2217: 1 foo .
2218: @end example
2219:
2220: @assignment
2221: Write @code{min} without @code{if}.
2222: @endassignment
2223:
1.66 anton 2224: For reference, see @ref{Boolean Flags}, @ref{Numeric comparison}, and
2225: @ref{Bitwise operations}.
2226:
1.48 anton 2227:
2228: @node General Loops Tutorial, Counted loops Tutorial, Flags and Comparisons Tutorial, Tutorial
2229: @section General Loops
1.66 anton 2230: @cindex loops, indefinite, tutorial
1.48 anton 2231:
2232: The endless loop is the most simple one:
2233:
2234: @example
2235: : endless ( -- )
2236: 0 begin
2237: dup . 1+
2238: again ;
2239: endless
2240: @end example
2241:
2242: Terminate this loop by pressing @kbd{Ctrl-C} (in Gforth). @code{begin}
2243: does nothing at run-time, @code{again} jumps back to @code{begin}.
2244:
2245: A loop with one exit at any place looks like this:
2246:
2247: @example
2248: : log2 ( +n1 -- n2 )
2249: \ logarithmus dualis of n1>0, rounded down to the next integer
2250: assert( dup 0> )
2251: 2/ 0 begin
2252: over 0> while
2253: 1+ swap 2/ swap
2254: repeat
2255: nip ;
2256: 7 log2 .
2257: 8 log2 .
2258: @end example
2259:
2260: At run-time @code{while} consumes a flag; if it is 0, execution
1.51 pazsan 2261: continues behind the @code{repeat}; if the flag is non-zero, execution
1.48 anton 2262: continues behind the @code{while}. @code{Repeat} jumps back to
2263: @code{begin}, just like @code{again}.
2264:
2265: In Forth there are many combinations/abbreviations, like @code{1+}.
2266: However, @code{2/} is not one of them; it shifts it's argument right by
2267: one bit (arithmetic shift right):
2268:
2269: @example
2270: -5 2 / .
2271: -5 2/ .
2272: @end example
2273:
2274: @code{assert(} is no standard word, but you can get it on systems other
2275: then Gforth by including @file{compat/assert.fs}. You can see what it
2276: does by trying
2277:
2278: @example
2279: 0 log2 .
2280: @end example
2281:
2282: Here's a loop with an exit at the end:
2283:
2284: @example
2285: : log2 ( +n1 -- n2 )
2286: \ logarithmus dualis of n1>0, rounded down to the next integer
2287: assert( dup 0 > )
2288: -1 begin
2289: 1+ swap 2/ swap
2290: over 0 <=
2291: until
2292: nip ;
2293: @end example
2294:
2295: @code{Until} consumes a flag; if it is non-zero, execution continues at
2296: the @code{begin}, otherwise after the @code{until}.
2297:
2298: @assignment
2299: Write a definition for computing the greatest common divisor.
2300: @endassignment
2301:
1.66 anton 2302: Reference: @ref{Simple Loops}.
2303:
1.48 anton 2304:
2305: @node Counted loops Tutorial, Recursion Tutorial, General Loops Tutorial, Tutorial
2306: @section Counted loops
1.66 anton 2307: @cindex loops, counted, tutorial
1.48 anton 2308:
2309: @example
2310: : ^ ( n1 u -- n )
2311: \ n = the uth power of u1
2312: 1 swap 0 u+do
2313: over *
2314: loop
2315: nip ;
2316: 3 2 ^ .
2317: 4 3 ^ .
2318: @end example
2319:
2320: @code{U+do} (from @file{compat/loops.fs}, if your Forth system doesn't
2321: have it) takes two numbers of the stack @code{( u3 u4 -- )}, and then
2322: performs the code between @code{u+do} and @code{loop} for @code{u3-u4}
2323: times (or not at all, if @code{u3-u4<0}).
2324:
2325: You can see the stack effect design rules at work in the stack effect of
2326: the loop start words: Since the start value of the loop is more
2327: frequently constant than the end value, the start value is passed on
2328: the top-of-stack.
2329:
2330: You can access the counter of a counted loop with @code{i}:
2331:
2332: @example
2333: : fac ( u -- u! )
2334: 1 swap 1+ 1 u+do
2335: i *
2336: loop ;
2337: 5 fac .
2338: 7 fac .
2339: @end example
2340:
2341: There is also @code{+do}, which expects signed numbers (important for
2342: deciding whether to enter the loop).
2343:
2344: @assignment
2345: Write a definition for computing the nth Fibonacci number.
2346: @endassignment
2347:
1.65 anton 2348: You can also use increments other than 1:
2349:
2350: @example
2351: : up2 ( n1 n2 -- )
2352: +do
2353: i .
2354: 2 +loop ;
2355: 10 0 up2
2356:
2357: : down2 ( n1 n2 -- )
2358: -do
2359: i .
2360: 2 -loop ;
2361: 0 10 down2
2362: @end example
1.48 anton 2363:
1.66 anton 2364: Reference: @ref{Counted Loops}.
2365:
1.48 anton 2366:
2367: @node Recursion Tutorial, Leaving definitions or loops Tutorial, Counted loops Tutorial, Tutorial
2368: @section Recursion
1.66 anton 2369: @cindex recursion tutorial
1.48 anton 2370:
2371: Usually the name of a definition is not visible in the definition; but
2372: earlier definitions are usually visible:
2373:
2374: @example
2375: 1 0 / . \ "Floating-point unidentified fault" in Gforth on most platforms
2376: : / ( n1 n2 -- n )
2377: dup 0= if
2378: -10 throw \ report division by zero
2379: endif
2380: / \ old version
2381: ;
2382: 1 0 /
2383: @end example
2384:
2385: For recursive definitions you can use @code{recursive} (non-standard) or
2386: @code{recurse}:
2387:
2388: @example
2389: : fac1 ( n -- n! ) recursive
2390: dup 0> if
2391: dup 1- fac1 *
2392: else
2393: drop 1
2394: endif ;
2395: 7 fac1 .
2396:
2397: : fac2 ( n -- n! )
2398: dup 0> if
2399: dup 1- recurse *
2400: else
2401: drop 1
2402: endif ;
2403: 8 fac2 .
2404: @end example
2405:
2406: @assignment
2407: Write a recursive definition for computing the nth Fibonacci number.
2408: @endassignment
2409:
1.66 anton 2410: Reference (including indirect recursion): @xref{Calls and returns}.
2411:
1.48 anton 2412:
2413: @node Leaving definitions or loops Tutorial, Return Stack Tutorial, Recursion Tutorial, Tutorial
2414: @section Leaving definitions or loops
1.66 anton 2415: @cindex leaving definitions, tutorial
2416: @cindex leaving loops, tutorial
1.48 anton 2417:
2418: @code{EXIT} exits the current definition right away. For every counted
2419: loop that is left in this way, an @code{UNLOOP} has to be performed
2420: before the @code{EXIT}:
2421:
2422: @c !! real examples
2423: @example
2424: : ...
2425: ... u+do
2426: ... if
2427: ... unloop exit
2428: endif
2429: ...
2430: loop
2431: ... ;
2432: @end example
2433:
2434: @code{LEAVE} leaves the innermost counted loop right away:
2435:
2436: @example
2437: : ...
2438: ... u+do
2439: ... if
2440: ... leave
2441: endif
2442: ...
2443: loop
2444: ... ;
2445: @end example
2446:
1.65 anton 2447: @c !! example
1.48 anton 2448:
1.66 anton 2449: Reference: @ref{Calls and returns}, @ref{Counted Loops}.
2450:
2451:
1.48 anton 2452: @node Return Stack Tutorial, Memory Tutorial, Leaving definitions or loops Tutorial, Tutorial
2453: @section Return Stack
1.66 anton 2454: @cindex return stack tutorial
1.48 anton 2455:
2456: In addition to the data stack Forth also has a second stack, the return
2457: stack; most Forth systems store the return addresses of procedure calls
2458: there (thus its name). Programmers can also use this stack:
2459:
2460: @example
2461: : foo ( n1 n2 -- )
2462: .s
2463: >r .s
1.50 anton 2464: r@@ .
1.48 anton 2465: >r .s
1.50 anton 2466: r@@ .
1.48 anton 2467: r> .
1.50 anton 2468: r@@ .
1.48 anton 2469: r> . ;
2470: 1 2 foo
2471: @end example
2472:
2473: @code{>r} takes an element from the data stack and pushes it onto the
2474: return stack; conversely, @code{r>} moves an elementm from the return to
2475: the data stack; @code{r@@} pushes a copy of the top of the return stack
2476: on the return stack.
2477:
2478: Forth programmers usually use the return stack for storing data
2479: temporarily, if using the data stack alone would be too complex, and
2480: factoring and locals are not an option:
2481:
2482: @example
2483: : 2swap ( x1 x2 x3 x4 -- x3 x4 x1 x2 )
2484: rot >r rot r> ;
2485: @end example
2486:
2487: The return address of the definition and the loop control parameters of
2488: counted loops usually reside on the return stack, so you have to take
2489: all items, that you have pushed on the return stack in a colon
2490: definition or counted loop, from the return stack before the definition
2491: or loop ends. You cannot access items that you pushed on the return
2492: stack outside some definition or loop within the definition of loop.
2493:
2494: If you miscount the return stack items, this usually ends in a crash:
2495:
2496: @example
2497: : crash ( n -- )
2498: >r ;
2499: 5 crash
2500: @end example
2501:
2502: You cannot mix using locals and using the return stack (according to the
2503: standard; Gforth has no problem). However, they solve the same
2504: problems, so this shouldn't be an issue.
2505:
2506: @assignment
2507: Can you rewrite any of the definitions you wrote until now in a better
2508: way using the return stack?
2509: @endassignment
2510:
1.66 anton 2511: Reference: @ref{Return stack}.
2512:
1.48 anton 2513:
2514: @node Memory Tutorial, Characters and Strings Tutorial, Return Stack Tutorial, Tutorial
2515: @section Memory
1.66 anton 2516: @cindex memory access/allocation tutorial
1.48 anton 2517:
2518: You can create a global variable @code{v} with
2519:
2520: @example
2521: variable v ( -- addr )
2522: @end example
2523:
2524: @code{v} pushes the address of a cell in memory on the stack. This cell
2525: was reserved by @code{variable}. You can use @code{!} (store) to store
2526: values into this cell and @code{@@} (fetch) to load the value from the
2527: stack into memory:
2528:
2529: @example
2530: v .
2531: 5 v ! .s
1.50 anton 2532: v @@ .
1.48 anton 2533: @end example
2534:
1.65 anton 2535: You can see a raw dump of memory with @code{dump}:
2536:
2537: @example
2538: v 1 cells .s dump
2539: @end example
2540:
2541: @code{Cells ( n1 -- n2 )} gives you the number of bytes (or, more
2542: generally, address units (aus)) that @code{n1 cells} occupy. You can
2543: also reserve more memory:
1.48 anton 2544:
2545: @example
2546: create v2 20 cells allot
1.65 anton 2547: v2 20 cells dump
1.48 anton 2548: @end example
2549:
1.65 anton 2550: creates a word @code{v2} and reserves 20 uninitialized cells; the
2551: address pushed by @code{v2} points to the start of these 20 cells. You
2552: can use address arithmetic to access these cells:
1.48 anton 2553:
2554: @example
2555: 3 v2 5 cells + !
1.65 anton 2556: v2 20 cells dump
1.48 anton 2557: @end example
2558:
2559: You can reserve and initialize memory with @code{,}:
2560:
2561: @example
2562: create v3
2563: 5 , 4 , 3 , 2 , 1 ,
1.50 anton 2564: v3 @@ .
2565: v3 cell+ @@ .
2566: v3 2 cells + @@ .
1.65 anton 2567: v3 5 cells dump
1.48 anton 2568: @end example
2569:
2570: @assignment
2571: Write a definition @code{vsum ( addr u -- n )} that computes the sum of
2572: @code{u} cells, with the first of these cells at @code{addr}, the next
2573: one at @code{addr cell+} etc.
2574: @endassignment
2575:
2576: You can also reserve memory without creating a new word:
2577:
2578: @example
1.60 anton 2579: here 10 cells allot .
2580: here .
1.48 anton 2581: @end example
2582:
2583: @code{Here} pushes the start address of the memory area. You should
2584: store it somewhere, or you will have a hard time finding the memory area
2585: again.
2586:
2587: @code{Allot} manages dictionary memory. The dictionary memory contains
2588: the system's data structures for words etc. on Gforth and most other
2589: Forth systems. It is managed like a stack: You can free the memory that
2590: you have just @code{allot}ed with
2591:
2592: @example
2593: -10 cells allot
1.60 anton 2594: here .
1.48 anton 2595: @end example
2596:
2597: Note that you cannot do this if you have created a new word in the
2598: meantime (because then your @code{allot}ed memory is no longer on the
2599: top of the dictionary ``stack'').
2600:
2601: Alternatively, you can use @code{allocate} and @code{free} which allow
2602: freeing memory in any order:
2603:
2604: @example
2605: 10 cells allocate throw .s
2606: 20 cells allocate throw .s
2607: swap
2608: free throw
2609: free throw
2610: @end example
2611:
2612: The @code{throw}s deal with errors (e.g., out of memory).
2613:
1.65 anton 2614: And there is also a
2615: @uref{http://www.complang.tuwien.ac.at/forth/garbage-collection.zip,
2616: garbage collector}, which eliminates the need to @code{free} memory
2617: explicitly.
1.48 anton 2618:
1.66 anton 2619: Reference: @ref{Memory}.
2620:
1.48 anton 2621:
2622: @node Characters and Strings Tutorial, Alignment Tutorial, Memory Tutorial, Tutorial
2623: @section Characters and Strings
1.66 anton 2624: @cindex strings tutorial
2625: @cindex characters tutorial
1.48 anton 2626:
2627: On the stack characters take up a cell, like numbers. In memory they
2628: have their own size (one 8-bit byte on most systems), and therefore
2629: require their own words for memory access:
2630:
2631: @example
2632: create v4
2633: 104 c, 97 c, 108 c, 108 c, 111 c,
1.50 anton 2634: v4 4 chars + c@@ .
1.65 anton 2635: v4 5 chars dump
1.48 anton 2636: @end example
2637:
2638: The preferred representation of strings on the stack is @code{addr
2639: u-count}, where @code{addr} is the address of the first character and
2640: @code{u-count} is the number of characters in the string.
2641:
2642: @example
2643: v4 5 type
2644: @end example
2645:
2646: You get a string constant with
2647:
2648: @example
2649: s" hello, world" .s
2650: type
2651: @end example
2652:
2653: Make sure you have a space between @code{s"} and the string; @code{s"}
2654: is a normal Forth word and must be delimited with white space (try what
2655: happens when you remove the space).
2656:
2657: However, this interpretive use of @code{s"} is quite restricted: the
2658: string exists only until the next call of @code{s"} (some Forth systems
2659: keep more than one of these strings, but usually they still have a
1.62 crook 2660: limited lifetime).
1.48 anton 2661:
2662: @example
2663: s" hello," s" world" .s
2664: type
2665: type
2666: @end example
2667:
1.62 crook 2668: You can also use @code{s"} in a definition, and the resulting
2669: strings then live forever (well, for as long as the definition):
1.48 anton 2670:
2671: @example
2672: : foo s" hello," s" world" ;
2673: foo .s
2674: type
2675: type
2676: @end example
2677:
2678: @assignment
2679: @code{Emit ( c -- )} types @code{c} as character (not a number).
2680: Implement @code{type ( addr u -- )}.
2681: @endassignment
2682:
1.66 anton 2683: Reference: @ref{Memory Blocks}.
2684:
2685:
1.48 anton 2686: @node Alignment Tutorial, Interpretation and Compilation Semantics and Immediacy Tutorial, Characters and Strings Tutorial, Tutorial
2687: @section Alignment
1.66 anton 2688: @cindex alignment tutorial
2689: @cindex memory alignment tutorial
1.48 anton 2690:
2691: On many processors cells have to be aligned in memory, if you want to
2692: access them with @code{@@} and @code{!} (and even if the processor does
1.62 crook 2693: not require alignment, access to aligned cells is faster).
1.48 anton 2694:
2695: @code{Create} aligns @code{here} (i.e., the place where the next
2696: allocation will occur, and that the @code{create}d word points to).
2697: Likewise, the memory produced by @code{allocate} starts at an aligned
2698: address. Adding a number of @code{cells} to an aligned address produces
2699: another aligned address.
2700:
2701: However, address arithmetic involving @code{char+} and @code{chars} can
2702: create an address that is not cell-aligned. @code{Aligned ( addr --
2703: a-addr )} produces the next aligned address:
2704:
2705: @example
1.50 anton 2706: v3 char+ aligned .s @@ .
2707: v3 char+ .s @@ .
1.48 anton 2708: @end example
2709:
2710: Similarly, @code{align} advances @code{here} to the next aligned
2711: address:
2712:
2713: @example
2714: create v5 97 c,
2715: here .
2716: align here .
2717: 1000 ,
2718: @end example
2719:
2720: Note that you should use aligned addresses even if your processor does
2721: not require them, if you want your program to be portable.
2722:
1.66 anton 2723: Reference: @ref{Address arithmetic}.
2724:
1.48 anton 2725:
2726: @node Interpretation and Compilation Semantics and Immediacy Tutorial, Execution Tokens Tutorial, Alignment Tutorial, Tutorial
2727: @section Interpretation and Compilation Semantics and Immediacy
1.66 anton 2728: @cindex semantics tutorial
2729: @cindex interpretation semantics tutorial
2730: @cindex compilation semantics tutorial
2731: @cindex immediate, tutorial
1.48 anton 2732:
2733: When a word is compiled, it behaves differently from being interpreted.
2734: E.g., consider @code{+}:
2735:
2736: @example
2737: 1 2 + .
2738: : foo + ;
2739: @end example
2740:
2741: These two behaviours are known as compilation and interpretation
2742: semantics. For normal words (e.g., @code{+}), the compilation semantics
2743: is to append the interpretation semantics to the currently defined word
2744: (@code{foo} in the example above). I.e., when @code{foo} is executed
2745: later, the interpretation semantics of @code{+} (i.e., adding two
2746: numbers) will be performed.
2747:
2748: However, there are words with non-default compilation semantics, e.g.,
2749: the control-flow words like @code{if}. You can use @code{immediate} to
2750: change the compilation semantics of the last defined word to be equal to
2751: the interpretation semantics:
2752:
2753: @example
2754: : [FOO] ( -- )
2755: 5 . ; immediate
2756:
2757: [FOO]
2758: : bar ( -- )
2759: [FOO] ;
2760: bar
2761: see bar
2762: @end example
2763:
2764: Two conventions to mark words with non-default compilation semnatics are
2765: names with brackets (more frequently used) and to write them all in
2766: upper case (less frequently used).
2767:
2768: In Gforth (and many other systems) you can also remove the
2769: interpretation semantics with @code{compile-only} (the compilation
2770: semantics is derived from the original interpretation semantics):
2771:
2772: @example
2773: : flip ( -- )
2774: 6 . ; compile-only \ but not immediate
2775: flip
2776:
2777: : flop ( -- )
2778: flip ;
2779: flop
2780: @end example
2781:
2782: In this example the interpretation semantics of @code{flop} is equal to
2783: the original interpretation semantics of @code{flip}.
2784:
2785: The text interpreter has two states: in interpret state, it performs the
2786: interpretation semantics of words it encounters; in compile state, it
2787: performs the compilation semantics of these words.
2788:
2789: Among other things, @code{:} switches into compile state, and @code{;}
2790: switches back to interpret state. They contain the factors @code{]}
2791: (switch to compile state) and @code{[} (switch to interpret state), that
2792: do nothing but switch the state.
2793:
2794: @example
2795: : xxx ( -- )
2796: [ 5 . ]
2797: ;
2798:
2799: xxx
2800: see xxx
2801: @end example
2802:
2803: These brackets are also the source of the naming convention mentioned
2804: above.
2805:
1.66 anton 2806: Reference: @ref{Interpretation and Compilation Semantics}.
2807:
1.48 anton 2808:
2809: @node Execution Tokens Tutorial, Exceptions Tutorial, Interpretation and Compilation Semantics and Immediacy Tutorial, Tutorial
2810: @section Execution Tokens
1.66 anton 2811: @cindex execution tokens tutorial
2812: @cindex XT tutorial
1.48 anton 2813:
2814: @code{' word} gives you the execution token (XT) of a word. The XT is a
2815: cell representing the interpretation semantics of a word. You can
2816: execute this semantics with @code{execute}:
2817:
2818: @example
2819: ' + .s
2820: 1 2 rot execute .
2821: @end example
2822:
2823: The XT is similar to a function pointer in C. However, parameter
2824: passing through the stack makes it a little more flexible:
2825:
2826: @example
2827: : map-array ( ... addr u xt -- ... )
1.50 anton 2828: \ executes xt ( ... x -- ... ) for every element of the array starting
2829: \ at addr and containing u elements
1.48 anton 2830: @{ xt @}
2831: cells over + swap ?do
1.50 anton 2832: i @@ xt execute
1.48 anton 2833: 1 cells +loop ;
2834:
2835: create a 3 , 4 , 2 , -1 , 4 ,
2836: a 5 ' . map-array .s
2837: 0 a 5 ' + map-array .
2838: s" max-n" environment? drop .s
2839: a 5 ' min map-array .
2840: @end example
2841:
2842: You can use map-array with the XTs of words that consume one element
2843: more than they produce. In theory you can also use it with other XTs,
2844: but the stack effect then depends on the size of the array, which is
2845: hard to understand.
2846:
1.51 pazsan 2847: Since XTs are cell-sized, you can store them in memory and manipulate
2848: them on the stack like other cells. You can also compile the XT into a
1.48 anton 2849: word with @code{compile,}:
2850:
2851: @example
2852: : foo1 ( n1 n2 -- n )
2853: [ ' + compile, ] ;
2854: see foo
2855: @end example
2856:
2857: This is non-standard, because @code{compile,} has no compilation
2858: semantics in the standard, but it works in good Forth systems. For the
2859: broken ones, use
2860:
2861: @example
2862: : [compile,] compile, ; immediate
2863:
2864: : foo1 ( n1 n2 -- n )
2865: [ ' + ] [compile,] ;
2866: see foo
2867: @end example
2868:
2869: @code{'} is a word with default compilation semantics; it parses the
2870: next word when its interpretation semantics are executed, not during
2871: compilation:
2872:
2873: @example
2874: : foo ( -- xt )
2875: ' ;
2876: see foo
2877: : bar ( ... "word" -- ... )
2878: ' execute ;
2879: see bar
1.60 anton 2880: 1 2 bar + .
1.48 anton 2881: @end example
2882:
2883: You often want to parse a word during compilation and compile its XT so
2884: it will be pushed on the stack at run-time. @code{[']} does this:
2885:
2886: @example
2887: : xt-+ ( -- xt )
2888: ['] + ;
2889: see xt-+
2890: 1 2 xt-+ execute .
2891: @end example
2892:
2893: Many programmers tend to see @code{'} and the word it parses as one
2894: unit, and expect it to behave like @code{[']} when compiled, and are
2895: confused by the actual behaviour. If you are, just remember that the
2896: Forth system just takes @code{'} as one unit and has no idea that it is
2897: a parsing word (attempts to convenience programmers in this issue have
2898: usually resulted in even worse pitfalls, see
1.66 anton 2899: @uref{http://www.complang.tuwien.ac.at/papers/ertl98.ps.gz,
2900: @code{State}-smartness---Why it is evil and How to Exorcise it}).
1.48 anton 2901:
2902: Note that the state of the interpreter does not come into play when
1.51 pazsan 2903: creating and executing XTs. I.e., even when you execute @code{'} in
1.48 anton 2904: compile state, it still gives you the interpretation semantics. And
2905: whatever that state is, @code{execute} performs the semantics
1.66 anton 2906: represented by the XT (i.e., for XTs produced with @code{'} the
2907: interpretation semantics).
2908:
2909: Reference: @ref{Tokens for Words}.
1.48 anton 2910:
2911:
2912: @node Exceptions Tutorial, Defining Words Tutorial, Execution Tokens Tutorial, Tutorial
2913: @section Exceptions
1.66 anton 2914: @cindex exceptions tutorial
1.48 anton 2915:
2916: @code{throw ( n -- )} causes an exception unless n is zero.
2917:
2918: @example
2919: 100 throw .s
2920: 0 throw .s
2921: @end example
2922:
2923: @code{catch ( ... xt -- ... n )} behaves similar to @code{execute}, but
2924: it catches exceptions and pushes the number of the exception on the
2925: stack (or 0, if the xt executed without exception). If there was an
2926: exception, the stacks have the same depth as when entering @code{catch}:
2927:
2928: @example
2929: .s
2930: 3 0 ' / catch .s
2931: 3 2 ' / catch .s
2932: @end example
2933:
2934: @assignment
2935: Try the same with @code{execute} instead of @code{catch}.
2936: @endassignment
2937:
2938: @code{Throw} always jumps to the dynamically next enclosing
2939: @code{catch}, even if it has to leave several call levels to achieve
2940: this:
2941:
2942: @example
2943: : foo 100 throw ;
2944: : foo1 foo ." after foo" ;
1.51 pazsan 2945: : bar ['] foo1 catch ;
1.60 anton 2946: bar .
1.48 anton 2947: @end example
2948:
2949: It is often important to restore a value upon leaving a definition, even
2950: if the definition is left through an exception. You can ensure this
2951: like this:
2952:
2953: @example
2954: : ...
2955: save-x
1.51 pazsan 2956: ['] word-changing-x catch ( ... n )
1.48 anton 2957: restore-x
2958: ( ... n ) throw ;
2959: @end example
2960:
1.55 anton 2961: Gforth provides an alternative syntax in addition to @code{catch}:
1.48 anton 2962: @code{try ... recover ... endtry}. If the code between @code{try} and
2963: @code{recover} has an exception, the stack depths are restored, the
2964: exception number is pushed on the stack, and the code between
2965: @code{recover} and @code{endtry} is performed. E.g., the definition for
2966: @code{catch} is
2967:
2968: @example
2969: : catch ( x1 .. xn xt -- y1 .. ym 0 / z1 .. zn error ) \ exception
2970: try
2971: execute 0
2972: recover
2973: nip
2974: endtry ;
2975: @end example
2976:
2977: The equivalent to the restoration code above is
2978:
2979: @example
2980: : ...
2981: save-x
2982: try
2983: word-changing-x
2984: end-try
2985: restore-x
2986: throw ;
2987: @end example
2988:
2989: As you can see, the @code{recover} part is optional.
2990:
1.66 anton 2991: Reference: @ref{Exception Handling}.
2992:
1.48 anton 2993:
2994: @node Defining Words Tutorial, Arrays and Records Tutorial, Exceptions Tutorial, Tutorial
2995: @section Defining Words
1.66 anton 2996: @cindex defining words tutorial
2997: @cindex does> tutorial
2998: @cindex create...does> tutorial
2999:
3000: @c before semantics?
1.48 anton 3001:
3002: @code{:}, @code{create}, and @code{variable} are definition words: They
3003: define other words. @code{Constant} is another definition word:
3004:
3005: @example
3006: 5 constant foo
3007: foo .
3008: @end example
3009:
3010: You can also use the prefixes @code{2} (double-cell) and @code{f}
3011: (floating point) with @code{variable} and @code{constant}.
3012:
3013: You can also define your own defining words. E.g.:
3014:
3015: @example
3016: : variable ( "name" -- )
3017: create 0 , ;
3018: @end example
3019:
3020: You can also define defining words that create words that do something
3021: other than just producing their address:
3022:
3023: @example
3024: : constant ( n "name" -- )
3025: create ,
3026: does> ( -- n )
1.50 anton 3027: ( addr ) @@ ;
1.48 anton 3028:
3029: 5 constant foo
3030: foo .
3031: @end example
3032:
3033: The definition of @code{constant} above ends at the @code{does>}; i.e.,
3034: @code{does>} replaces @code{;}, but it also does something else: It
3035: changes the last defined word such that it pushes the address of the
3036: body of the word and then performs the code after the @code{does>}
3037: whenever it is called.
3038:
3039: In the example above, @code{constant} uses @code{,} to store 5 into the
3040: body of @code{foo}. When @code{foo} executes, it pushes the address of
3041: the body onto the stack, then (in the code after the @code{does>})
3042: fetches the 5 from there.
3043:
3044: The stack comment near the @code{does>} reflects the stack effect of the
3045: defined word, not the stack effect of the code after the @code{does>}
3046: (the difference is that the code expects the address of the body that
3047: the stack comment does not show).
3048:
3049: You can use these definition words to do factoring in cases that involve
3050: (other) definition words. E.g., a field offset is always added to an
3051: address. Instead of defining
3052:
3053: @example
3054: 2 cells constant offset-field1
3055: @end example
3056:
3057: and using this like
3058:
3059: @example
3060: ( addr ) offset-field1 +
3061: @end example
3062:
3063: you can define a definition word
3064:
3065: @example
3066: : simple-field ( n "name" -- )
3067: create ,
3068: does> ( n1 -- n1+n )
1.50 anton 3069: ( addr ) @@ + ;
1.48 anton 3070: @end example
1.21 crook 3071:
1.48 anton 3072: Definition and use of field offsets now look like this:
1.21 crook 3073:
1.48 anton 3074: @example
3075: 2 cells simple-field field1
1.60 anton 3076: create mystruct 4 cells allot
3077: mystruct .s field1 .s drop
1.48 anton 3078: @end example
1.21 crook 3079:
1.48 anton 3080: If you want to do something with the word without performing the code
3081: after the @code{does>}, you can access the body of a @code{create}d word
3082: with @code{>body ( xt -- addr )}:
1.21 crook 3083:
1.48 anton 3084: @example
3085: : value ( n "name" -- )
3086: create ,
3087: does> ( -- n1 )
1.50 anton 3088: @@ ;
1.48 anton 3089: : to ( n "name" -- )
3090: ' >body ! ;
1.21 crook 3091:
1.48 anton 3092: 5 value foo
3093: foo .
3094: 7 to foo
3095: foo .
3096: @end example
1.21 crook 3097:
1.48 anton 3098: @assignment
3099: Define @code{defer ( "name" -- )}, which creates a word that stores an
3100: XT (at the start the XT of @code{abort}), and upon execution
3101: @code{execute}s the XT. Define @code{is ( xt "name" -- )} that stores
3102: @code{xt} into @code{name}, a word defined with @code{defer}. Indirect
3103: recursion is one application of @code{defer}.
3104: @endassignment
1.29 crook 3105:
1.66 anton 3106: Reference: @ref{User-defined Defining Words}.
3107:
3108:
1.48 anton 3109: @node Arrays and Records Tutorial, POSTPONE Tutorial, Defining Words Tutorial, Tutorial
3110: @section Arrays and Records
1.66 anton 3111: @cindex arrays tutorial
3112: @cindex records tutorial
3113: @cindex structs tutorial
1.29 crook 3114:
1.48 anton 3115: Forth has no standard words for defining data structures such as arrays
3116: and records (structs in C terminology), but you can build them yourself
3117: based on address arithmetic. You can also define words for defining
3118: arrays and records (@pxref{Defining Words Tutorial,, Defining Words}).
1.29 crook 3119:
1.48 anton 3120: One of the first projects a Forth newcomer sets out upon when learning
3121: about defining words is an array defining word (possibly for
3122: n-dimensional arrays). Go ahead and do it, I did it, too; you will
3123: learn something from it. However, don't be disappointed when you later
3124: learn that you have little use for these words (inappropriate use would
3125: be even worse). I have not yet found a set of useful array words yet;
3126: the needs are just too diverse, and named, global arrays (the result of
3127: naive use of defining words) are often not flexible enough (e.g.,
1.66 anton 3128: consider how to pass them as parameters). Another such project is a set
3129: of words to help dealing with strings.
1.29 crook 3130:
1.48 anton 3131: On the other hand, there is a useful set of record words, and it has
3132: been defined in @file{compat/struct.fs}; these words are predefined in
3133: Gforth. They are explained in depth elsewhere in this manual (see
3134: @pxref{Structures}). The @code{simple-field} example above is
3135: simplified variant of fields in this package.
1.21 crook 3136:
3137:
1.48 anton 3138: @node POSTPONE Tutorial, Literal Tutorial, Arrays and Records Tutorial, Tutorial
3139: @section @code{POSTPONE}
1.66 anton 3140: @cindex postpone tutorial
1.21 crook 3141:
1.48 anton 3142: You can compile the compilation semantics (instead of compiling the
3143: interpretation semantics) of a word with @code{POSTPONE}:
1.21 crook 3144:
1.48 anton 3145: @example
3146: : MY-+ ( Compilation: -- ; Run-time of compiled code: n1 n2 -- n )
1.51 pazsan 3147: POSTPONE + ; immediate
1.48 anton 3148: : foo ( n1 n2 -- n )
3149: MY-+ ;
3150: 1 2 foo .
3151: see foo
3152: @end example
1.21 crook 3153:
1.48 anton 3154: During the definition of @code{foo} the text interpreter performs the
3155: compilation semantics of @code{MY-+}, which performs the compilation
3156: semantics of @code{+}, i.e., it compiles @code{+} into @code{foo}.
3157:
3158: This example also displays separate stack comments for the compilation
3159: semantics and for the stack effect of the compiled code. For words with
3160: default compilation semantics these stack effects are usually not
3161: displayed; the stack effect of the compilation semantics is always
3162: @code{( -- )} for these words, the stack effect for the compiled code is
3163: the stack effect of the interpretation semantics.
3164:
3165: Note that the state of the interpreter does not come into play when
3166: performing the compilation semantics in this way. You can also perform
3167: it interpretively, e.g.:
3168:
3169: @example
3170: : foo2 ( n1 n2 -- n )
3171: [ MY-+ ] ;
3172: 1 2 foo .
3173: see foo
3174: @end example
1.21 crook 3175:
1.48 anton 3176: However, there are some broken Forth systems where this does not always
1.62 crook 3177: work, and therefore this practice was been declared non-standard in
1.48 anton 3178: 1999.
3179: @c !! repair.fs
3180:
3181: Here is another example for using @code{POSTPONE}:
1.44 crook 3182:
1.48 anton 3183: @example
3184: : MY-- ( Compilation: -- ; Run-time of compiled code: n1 n2 -- n )
3185: POSTPONE negate POSTPONE + ; immediate compile-only
3186: : bar ( n1 n2 -- n )
3187: MY-- ;
3188: 2 1 bar .
3189: see bar
3190: @end example
1.21 crook 3191:
1.48 anton 3192: You can define @code{ENDIF} in this way:
1.21 crook 3193:
1.48 anton 3194: @example
3195: : ENDIF ( Compilation: orig -- )
3196: POSTPONE then ; immediate
3197: @end example
1.21 crook 3198:
1.48 anton 3199: @assignment
3200: Write @code{MY-2DUP} that has compilation semantics equivalent to
3201: @code{2dup}, but compiles @code{over over}.
3202: @endassignment
1.29 crook 3203:
1.66 anton 3204: @c !! @xref{Macros} for reference
3205:
3206:
1.48 anton 3207: @node Literal Tutorial, Advanced macros Tutorial, POSTPONE Tutorial, Tutorial
3208: @section @code{Literal}
1.66 anton 3209: @cindex literal tutorial
1.29 crook 3210:
1.48 anton 3211: You cannot @code{POSTPONE} numbers:
1.21 crook 3212:
1.48 anton 3213: @example
3214: : [FOO] POSTPONE 500 ; immediate
1.21 crook 3215: @end example
3216:
1.48 anton 3217: Instead, you can use @code{LITERAL (compilation: n --; run-time: -- n )}:
1.29 crook 3218:
1.48 anton 3219: @example
3220: : [FOO] ( compilation: --; run-time: -- n )
3221: 500 POSTPONE literal ; immediate
1.29 crook 3222:
1.60 anton 3223: : flip [FOO] ;
1.48 anton 3224: flip .
3225: see flip
3226: @end example
1.29 crook 3227:
1.48 anton 3228: @code{LITERAL} consumes a number at compile-time (when it's compilation
3229: semantics are executed) and pushes it at run-time (when the code it
3230: compiled is executed). A frequent use of @code{LITERAL} is to compile a
3231: number computed at compile time into the current word:
1.29 crook 3232:
1.48 anton 3233: @example
3234: : bar ( -- n )
3235: [ 2 2 + ] literal ;
3236: see bar
3237: @end example
1.29 crook 3238:
1.48 anton 3239: @assignment
3240: Write @code{]L} which allows writing the example above as @code{: bar (
3241: -- n ) [ 2 2 + ]L ;}
3242: @endassignment
3243:
1.66 anton 3244: @c !! @xref{Macros} for reference
3245:
1.48 anton 3246:
3247: @node Advanced macros Tutorial, Compilation Tokens Tutorial, Literal Tutorial, Tutorial
3248: @section Advanced macros
1.66 anton 3249: @cindex macros, advanced tutorial
3250: @cindex run-time code generation, tutorial
1.48 anton 3251:
1.66 anton 3252: Reconsider @code{map-array} from @ref{Execution Tokens Tutorial,,
3253: Execution Tokens}. It frequently performs @code{execute}, a relatively
3254: expensive operation in some Forth implementations. You can use
1.48 anton 3255: @code{compile,} and @code{POSTPONE} to eliminate these @code{execute}s
3256: and produce a word that contains the word to be performed directly:
3257:
3258: @c use ]] ... [[
3259: @example
3260: : compile-map-array ( compilation: xt -- ; run-time: ... addr u -- ... )
3261: \ at run-time, execute xt ( ... x -- ... ) for each element of the
3262: \ array beginning at addr and containing u elements
3263: @{ xt @}
3264: POSTPONE cells POSTPONE over POSTPONE + POSTPONE swap POSTPONE ?do
1.50 anton 3265: POSTPONE i POSTPONE @@ xt compile,
1.48 anton 3266: 1 cells POSTPONE literal POSTPONE +loop ;
3267:
3268: : sum-array ( addr u -- n )
3269: 0 rot rot [ ' + compile-map-array ] ;
3270: see sum-array
3271: a 5 sum-array .
3272: @end example
3273:
3274: You can use the full power of Forth for generating the code; here's an
3275: example where the code is generated in a loop:
3276:
3277: @example
3278: : compile-vmul-step ( compilation: n --; run-time: n1 addr1 -- n2 addr2 )
3279: \ n2=n1+(addr1)*n, addr2=addr1+cell
1.50 anton 3280: POSTPONE tuck POSTPONE @@
1.48 anton 3281: POSTPONE literal POSTPONE * POSTPONE +
3282: POSTPONE swap POSTPONE cell+ ;
3283:
3284: : compile-vmul ( compilation: addr1 u -- ; run-time: addr2 -- n )
1.51 pazsan 3285: \ n=v1*v2 (inner product), where the v_i are represented as addr_i u
1.48 anton 3286: 0 postpone literal postpone swap
3287: [ ' compile-vmul-step compile-map-array ]
3288: postpone drop ;
3289: see compile-vmul
3290:
3291: : a-vmul ( addr -- n )
1.51 pazsan 3292: \ n=a*v, where v is a vector that's as long as a and starts at addr
1.48 anton 3293: [ a 5 compile-vmul ] ;
3294: see a-vmul
3295: a a-vmul .
3296: @end example
3297:
3298: This example uses @code{compile-map-array} to show off, but you could
1.66 anton 3299: also use @code{map-array} instead (try it now!).
1.48 anton 3300:
3301: You can use this technique for efficient multiplication of large
3302: matrices. In matrix multiplication, you multiply every line of one
3303: matrix with every column of the other matrix. You can generate the code
3304: for one line once, and use it for every column. The only downside of
3305: this technique is that it is cumbersome to recover the memory consumed
3306: by the generated code when you are done (and in more complicated cases
3307: it is not possible portably).
3308:
1.66 anton 3309: @c !! @xref{Macros} for reference
3310:
3311:
1.48 anton 3312: @node Compilation Tokens Tutorial, Wordlists and Search Order Tutorial, Advanced macros Tutorial, Tutorial
3313: @section Compilation Tokens
1.66 anton 3314: @cindex compilation tokens, tutorial
3315: @cindex CT, tutorial
1.48 anton 3316:
3317: This section is Gforth-specific. You can skip it.
3318:
3319: @code{' word compile,} compiles the interpretation semantics. For words
3320: with default compilation semantics this is the same as performing the
3321: compilation semantics. To represent the compilation semantics of other
3322: words (e.g., words like @code{if} that have no interpretation
3323: semantics), Gforth has the concept of a compilation token (CT,
3324: consisting of two cells), and words @code{comp'} and @code{[comp']}.
3325: You can perform the compilation semantics represented by a CT with
3326: @code{execute}:
1.29 crook 3327:
1.48 anton 3328: @example
3329: : foo2 ( n1 n2 -- n )
3330: [ comp' + execute ] ;
3331: see foo
3332: @end example
1.29 crook 3333:
1.48 anton 3334: You can compile the compilation semantics represented by a CT with
3335: @code{postpone,}:
1.30 anton 3336:
1.48 anton 3337: @example
3338: : foo3 ( -- )
3339: [ comp' + postpone, ] ;
3340: see foo3
3341: @end example
1.30 anton 3342:
1.51 pazsan 3343: @code{[ comp' word postpone, ]} is equivalent to @code{POSTPONE word}.
1.48 anton 3344: @code{comp'} is particularly useful for words that have no
3345: interpretation semantics:
1.29 crook 3346:
1.30 anton 3347: @example
1.48 anton 3348: ' if
1.60 anton 3349: comp' if .s 2drop
1.30 anton 3350: @end example
3351:
1.66 anton 3352: Reference: @ref{Tokens for Words}.
3353:
1.29 crook 3354:
1.48 anton 3355: @node Wordlists and Search Order Tutorial, , Compilation Tokens Tutorial, Tutorial
3356: @section Wordlists and Search Order
1.66 anton 3357: @cindex wordlists tutorial
3358: @cindex search order, tutorial
1.48 anton 3359:
3360: The dictionary is not just a memory area that allows you to allocate
3361: memory with @code{allot}, it also contains the Forth words, arranged in
3362: several wordlists. When searching for a word in a wordlist,
3363: conceptually you start searching at the youngest and proceed towards
3364: older words (in reality most systems nowadays use hash-tables); i.e., if
3365: you define a word with the same name as an older word, the new word
3366: shadows the older word.
3367:
3368: Which wordlists are searched in which order is determined by the search
3369: order. You can display the search order with @code{order}. It displays
3370: first the search order, starting with the wordlist searched first, then
3371: it displays the wordlist that will contain newly defined words.
1.21 crook 3372:
1.48 anton 3373: You can create a new, empty wordlist with @code{wordlist ( -- wid )}:
1.21 crook 3374:
1.48 anton 3375: @example
3376: wordlist constant mywords
3377: @end example
1.21 crook 3378:
1.48 anton 3379: @code{Set-current ( wid -- )} sets the wordlist that will contain newly
3380: defined words (the @emph{current} wordlist):
1.21 crook 3381:
1.48 anton 3382: @example
3383: mywords set-current
3384: order
3385: @end example
1.26 crook 3386:
1.48 anton 3387: Gforth does not display a name for the wordlist in @code{mywords}
3388: because this wordlist was created anonymously with @code{wordlist}.
1.21 crook 3389:
1.48 anton 3390: You can get the current wordlist with @code{get-current ( -- wid)}. If
3391: you want to put something into a specific wordlist without overall
3392: effect on the current wordlist, this typically looks like this:
1.21 crook 3393:
1.48 anton 3394: @example
3395: get-current mywords set-current ( wid )
3396: create someword
3397: ( wid ) set-current
3398: @end example
1.21 crook 3399:
1.48 anton 3400: You can write the search order with @code{set-order ( wid1 .. widn n --
3401: )} and read it with @code{get-order ( -- wid1 .. widn n )}. The first
3402: searched wordlist is topmost.
1.21 crook 3403:
1.48 anton 3404: @example
3405: get-order mywords swap 1+ set-order
3406: order
3407: @end example
1.21 crook 3408:
1.48 anton 3409: Yes, the order of wordlists in the output of @code{order} is reversed
3410: from stack comments and the output of @code{.s} and thus unintuitive.
1.21 crook 3411:
1.48 anton 3412: @assignment
3413: Define @code{>order ( wid -- )} with adds @code{wid} as first searched
3414: wordlist to the search order. Define @code{previous ( -- )}, which
3415: removes the first searched wordlist from the search order. Experiment
3416: with boundary conditions (you will see some crashes or situations that
3417: are hard or impossible to leave).
3418: @endassignment
1.21 crook 3419:
1.48 anton 3420: The search order is a powerful foundation for providing features similar
3421: to Modula-2 modules and C++ namespaces. However, trying to modularize
3422: programs in this way has disadvantages for debugging and reuse/factoring
3423: that overcome the advantages in my experience (I don't do huge projects,
1.55 anton 3424: though). These disadvantages are not so clear in other
1.48 anton 3425: languages/programming environments, because these langauges are not so
3426: strong in debugging and reuse.
1.21 crook 3427:
1.66 anton 3428: @c !! example
3429:
3430: Reference: @ref{Word Lists}.
1.21 crook 3431:
1.29 crook 3432: @c ******************************************************************
1.48 anton 3433: @node Introduction, Words, Tutorial, Top
1.29 crook 3434: @comment node-name, next, previous, up
3435: @chapter An Introduction to ANS Forth
3436: @cindex Forth - an introduction
1.21 crook 3437:
1.29 crook 3438: The primary purpose of this manual is to document Gforth. However, since
3439: Forth is not a widely-known language and there is a lack of up-to-date
3440: teaching material, it seems worthwhile to provide some introductory
1.49 anton 3441: material. For other sources of Forth-related
3442: information, see @ref{Forth-related information}.
1.21 crook 3443:
1.29 crook 3444: The examples in this section should work on any ANS Forth; the
3445: output shown was produced using Gforth. Each example attempts to
3446: reproduce the exact output that Gforth produces. If you try out the
3447: examples (and you should), what you should type is shown @kbd{like this}
3448: and Gforth's response is shown @code{like this}. The single exception is
1.30 anton 3449: that, where the example shows @key{RET} it means that you should
1.29 crook 3450: press the ``carriage return'' key. Unfortunately, some output formats for
3451: this manual cannot show the difference between @kbd{this} and
3452: @code{this} which will make trying out the examples harder (but not
3453: impossible).
1.21 crook 3454:
1.29 crook 3455: Forth is an unusual language. It provides an interactive development
3456: environment which includes both an interpreter and compiler. Forth
3457: programming style encourages you to break a problem down into many
3458: @cindex factoring
3459: small fragments (@dfn{factoring}), and then to develop and test each
3460: fragment interactively. Forth advocates assert that breaking the
3461: edit-compile-test cycle used by conventional programming languages can
3462: lead to great productivity improvements.
1.21 crook 3463:
1.29 crook 3464: @menu
1.67 anton 3465: * Introducing the Text Interpreter::
3466: * Stacks and Postfix notation::
3467: * Your first definition::
3468: * How does that work?::
3469: * Forth is written in Forth::
3470: * Review - elements of a Forth system::
3471: * Where to go next::
3472: * Exercises::
1.29 crook 3473: @end menu
1.21 crook 3474:
1.29 crook 3475: @comment ----------------------------------------------
3476: @node Introducing the Text Interpreter, Stacks and Postfix notation, Introduction, Introduction
3477: @section Introducing the Text Interpreter
3478: @cindex text interpreter
3479: @cindex outer interpreter
1.21 crook 3480:
1.30 anton 3481: @c IMO this is too detailed and the pace is too slow for
3482: @c an introduction. If you know German, take a look at
3483: @c http://www.complang.tuwien.ac.at/anton/lvas/skriptum-stack.html
3484: @c to see how I do it - anton
3485:
1.44 crook 3486: @c nac-> Where I have accepted your comments 100% and modified the text
3487: @c accordingly, I have deleted your comments. Elsewhere I have added a
3488: @c response like this to attempt to rationalise what I have done. Of
3489: @c course, this is a very clumsy mechanism for something that would be
3490: @c done far more efficiently over a beer. Please delete any dialogue
3491: @c you consider closed.
3492:
1.29 crook 3493: When you invoke the Forth image, you will see a startup banner printed
3494: and nothing else (if you have Gforth installed on your system, try
1.30 anton 3495: invoking it now, by typing @kbd{gforth@key{RET}}). Forth is now running
1.29 crook 3496: its command line interpreter, which is called the @dfn{Text Interpreter}
3497: (also known as the @dfn{Outer Interpreter}). (You will learn a lot
1.49 anton 3498: about the text interpreter as you read through this chapter, for more
3499: detail @pxref{The Text Interpreter}).
1.21 crook 3500:
1.29 crook 3501: Although it's not obvious, Forth is actually waiting for your
1.30 anton 3502: input. Type a number and press the @key{RET} key:
1.21 crook 3503:
1.26 crook 3504: @example
1.30 anton 3505: @kbd{45@key{RET}} ok
1.26 crook 3506: @end example
1.21 crook 3507:
1.29 crook 3508: Rather than give you a prompt to invite you to input something, the text
3509: interpreter prints a status message @i{after} it has processed a line
3510: of input. The status message in this case (``@code{ ok}'' followed by
3511: carriage-return) indicates that the text interpreter was able to process
3512: all of your input successfully. Now type something illegal:
3513:
3514: @example
1.30 anton 3515: @kbd{qwer341@key{RET}}
1.29 crook 3516: :1: Undefined word
3517: qwer341
3518: ^^^^^^^
3519: $400D2BA8 Bounce
3520: $400DBDA8 no.extensions
3521: @end example
1.23 crook 3522:
1.29 crook 3523: The exact text, other than the ``Undefined word'' may differ slightly on
3524: your system, but the effect is the same; when the text interpreter
3525: detects an error, it discards any remaining text on a line, resets
1.49 anton 3526: certain internal state and prints an error message. For a detailed description of error messages see @ref{Error
3527: messages}.
1.23 crook 3528:
1.29 crook 3529: The text interpreter waits for you to press carriage-return, and then
3530: processes your input line. Starting at the beginning of the line, it
3531: breaks the line into groups of characters separated by spaces. For each
3532: group of characters in turn, it makes two attempts to do something:
1.23 crook 3533:
1.29 crook 3534: @itemize @bullet
3535: @item
1.44 crook 3536: @cindex name dictionary
1.29 crook 3537: It tries to treat it as a command. It does this by searching a @dfn{name
3538: dictionary}. If the group of characters matches an entry in the name
3539: dictionary, the name dictionary provides the text interpreter with
3540: information that allows the text interpreter perform some actions. In
3541: Forth jargon, we say that the group
3542: @cindex word
3543: @cindex definition
3544: @cindex execution token
3545: @cindex xt
3546: of characters names a @dfn{word}, that the dictionary search returns an
3547: @dfn{execution token (xt)} corresponding to the @dfn{definition} of the
3548: word, and that the text interpreter executes the xt. Often, the terms
3549: @dfn{word} and @dfn{definition} are used interchangeably.
3550: @item
3551: If the text interpreter fails to find a match in the name dictionary, it
3552: tries to treat the group of characters as a number in the current number
3553: base (when you start up Forth, the current number base is base 10). If
3554: the group of characters legitimately represents a number, the text
3555: interpreter pushes the number onto a stack (we'll learn more about that
3556: in the next section).
3557: @end itemize
1.23 crook 3558:
1.29 crook 3559: If the text interpreter is unable to do either of these things with any
3560: group of characters, it discards the group of characters and the rest of
3561: the line, then prints an error message. If the text interpreter reaches
3562: the end of the line without error, it prints the status message ``@code{ ok}''
3563: followed by carriage-return.
1.21 crook 3564:
1.29 crook 3565: This is the simplest command we can give to the text interpreter:
1.23 crook 3566:
3567: @example
1.30 anton 3568: @key{RET} ok
1.23 crook 3569: @end example
1.21 crook 3570:
1.29 crook 3571: The text interpreter did everything we asked it to do (nothing) without
3572: an error, so it said that everything is ``@code{ ok}''. Try a slightly longer
3573: command:
1.21 crook 3574:
1.23 crook 3575: @example
1.30 anton 3576: @kbd{12 dup fred dup@key{RET}}
1.29 crook 3577: :1: Undefined word
3578: 12 dup fred dup
3579: ^^^^
3580: $400D2BA8 Bounce
3581: $400DBDA8 no.extensions
1.23 crook 3582: @end example
1.21 crook 3583:
1.29 crook 3584: When you press the carriage-return key, the text interpreter starts to
3585: work its way along the line:
1.21 crook 3586:
1.29 crook 3587: @itemize @bullet
3588: @item
3589: When it gets to the space after the @code{2}, it takes the group of
3590: characters @code{12} and looks them up in the name
3591: dictionary@footnote{We can't tell if it found them or not, but assume
3592: for now that it did not}. There is no match for this group of characters
3593: in the name dictionary, so it tries to treat them as a number. It is
3594: able to do this successfully, so it puts the number, 12, ``on the stack''
3595: (whatever that means).
3596: @item
3597: The text interpreter resumes scanning the line and gets the next group
3598: of characters, @code{dup}. It looks it up in the name dictionary and
3599: (you'll have to take my word for this) finds it, and executes the word
3600: @code{dup} (whatever that means).
3601: @item
3602: Once again, the text interpreter resumes scanning the line and gets the
3603: group of characters @code{fred}. It looks them up in the name
3604: dictionary, but can't find them. It tries to treat them as a number, but
3605: they don't represent any legal number.
3606: @end itemize
1.21 crook 3607:
1.29 crook 3608: At this point, the text interpreter gives up and prints an error
3609: message. The error message shows exactly how far the text interpreter
3610: got in processing the line. In particular, it shows that the text
3611: interpreter made no attempt to do anything with the final character
3612: group, @code{dup}, even though we have good reason to believe that the
3613: text interpreter would have no problem looking that word up and
3614: executing it a second time.
1.21 crook 3615:
3616:
1.29 crook 3617: @comment ----------------------------------------------
3618: @node Stacks and Postfix notation, Your first definition, Introducing the Text Interpreter, Introduction
3619: @section Stacks, postfix notation and parameter passing
3620: @cindex text interpreter
3621: @cindex outer interpreter
1.21 crook 3622:
1.29 crook 3623: In procedural programming languages (like C and Pascal), the
3624: building-block of programs is the @dfn{function} or @dfn{procedure}. These
3625: functions or procedures are called with @dfn{explicit parameters}. For
3626: example, in C we might write:
1.21 crook 3627:
1.23 crook 3628: @example
1.29 crook 3629: total = total + new_volume(length,height,depth);
1.23 crook 3630: @end example
1.21 crook 3631:
1.23 crook 3632: @noindent
1.29 crook 3633: where new_volume is a function-call to another piece of code, and total,
3634: length, height and depth are all variables. length, height and depth are
3635: parameters to the function-call.
1.21 crook 3636:
1.29 crook 3637: In Forth, the equivalent of the function or procedure is the
3638: @dfn{definition} and parameters are implicitly passed between
3639: definitions using a shared stack that is visible to the
3640: programmer. Although Forth does support variables, the existence of the
3641: stack means that they are used far less often than in most other
3642: programming languages. When the text interpreter encounters a number, it
3643: will place (@dfn{push}) it on the stack. There are several stacks (the
1.30 anton 3644: actual number is implementation-dependent ...) and the particular stack
1.29 crook 3645: used for any operation is implied unambiguously by the operation being
3646: performed. The stack used for all integer operations is called the @dfn{data
3647: stack} and, since this is the stack used most commonly, references to
3648: ``the data stack'' are often abbreviated to ``the stack''.
1.21 crook 3649:
1.29 crook 3650: The stacks have a last-in, first-out (LIFO) organisation. If you type:
1.21 crook 3651:
1.23 crook 3652: @example
1.30 anton 3653: @kbd{1 2 3@key{RET}} ok
1.23 crook 3654: @end example
1.21 crook 3655:
1.29 crook 3656: Then this instructs the text interpreter to placed three numbers on the
3657: (data) stack. An analogy for the behaviour of the stack is to take a
3658: pack of playing cards and deal out the ace (1), 2 and 3 into a pile on
3659: the table. The 3 was the last card onto the pile (``last-in'') and if
3660: you take a card off the pile then, unless you're prepared to fiddle a
3661: bit, the card that you take off will be the 3 (``first-out''). The
3662: number that will be first-out of the stack is called the @dfn{top of
3663: stack}, which
3664: @cindex TOS definition
3665: is often abbreviated to @dfn{TOS}.
1.21 crook 3666:
1.29 crook 3667: To understand how parameters are passed in Forth, consider the
3668: behaviour of the definition @code{+} (pronounced ``plus''). You will not
3669: be surprised to learn that this definition performs addition. More
3670: precisely, it adds two number together and produces a result. Where does
3671: it get the two numbers from? It takes the top two numbers off the
3672: stack. Where does it place the result? On the stack. You can act-out the
3673: behaviour of @code{+} with your playing cards like this:
1.21 crook 3674:
3675: @itemize @bullet
3676: @item
1.29 crook 3677: Pick up two cards from the stack on the table
1.21 crook 3678: @item
1.29 crook 3679: Stare at them intently and ask yourself ``what @i{is} the sum of these two
3680: numbers''
1.21 crook 3681: @item
1.29 crook 3682: Decide that the answer is 5
1.21 crook 3683: @item
1.29 crook 3684: Shuffle the two cards back into the pack and find a 5
1.21 crook 3685: @item
1.29 crook 3686: Put a 5 on the remaining ace that's on the table.
1.21 crook 3687: @end itemize
3688:
1.29 crook 3689: If you don't have a pack of cards handy but you do have Forth running,
3690: you can use the definition @code{.s} to show the current state of the stack,
3691: without affecting the stack. Type:
1.21 crook 3692:
3693: @example
1.30 anton 3694: @kbd{clearstack 1 2 3@key{RET}} ok
3695: @kbd{.s@key{RET}} <3> 1 2 3 ok
1.23 crook 3696: @end example
3697:
1.29 crook 3698: The text interpreter looks up the word @code{clearstack} and executes
3699: it; it tidies up the stack and removes any entries that may have been
3700: left on it by earlier examples. The text interpreter pushes each of the
3701: three numbers in turn onto the stack. Finally, the text interpreter
3702: looks up the word @code{.s} and executes it. The effect of executing
3703: @code{.s} is to print the ``<3>'' (the total number of items on the stack)
3704: followed by a list of all the items on the stack; the item on the far
3705: right-hand side is the TOS.
1.21 crook 3706:
1.29 crook 3707: You can now type:
1.21 crook 3708:
1.29 crook 3709: @example
1.30 anton 3710: @kbd{+ .s@key{RET}} <2> 1 5 ok
1.29 crook 3711: @end example
1.21 crook 3712:
1.29 crook 3713: @noindent
3714: which is correct; there are now 2 items on the stack and the result of
3715: the addition is 5.
1.23 crook 3716:
1.29 crook 3717: If you're playing with cards, try doing a second addition: pick up the
3718: two cards, work out that their sum is 6, shuffle them into the pack,
3719: look for a 6 and place that on the table. You now have just one item on
3720: the stack. What happens if you try to do a third addition? Pick up the
3721: first card, pick up the second card -- ah! There is no second card. This
3722: is called a @dfn{stack underflow} and consitutes an error. If you try to
3723: do the same thing with Forth it will report an error (probably a Stack
3724: Underflow or an Invalid Memory Address error).
1.23 crook 3725:
1.29 crook 3726: The opposite situation to a stack underflow is a @dfn{stack overflow},
3727: which simply accepts that there is a finite amount of storage space
3728: reserved for the stack. To stretch the playing card analogy, if you had
3729: enough packs of cards and you piled the cards up on the table, you would
3730: eventually be unable to add another card; you'd hit the ceiling. Gforth
3731: allows you to set the maximum size of the stacks. In general, the only
3732: time that you will get a stack overflow is because a definition has a
3733: bug in it and is generating data on the stack uncontrollably.
1.23 crook 3734:
1.29 crook 3735: There's one final use for the playing card analogy. If you model your
3736: stack using a pack of playing cards, the maximum number of items on
3737: your stack will be 52 (I assume you didn't use the Joker). The maximum
3738: @i{value} of any item on the stack is 13 (the King). In fact, the only
3739: possible numbers are positive integer numbers 1 through 13; you can't
3740: have (for example) 0 or 27 or 3.52 or -2. If you change the way you
3741: think about some of the cards, you can accommodate different
3742: numbers. For example, you could think of the Jack as representing 0,
3743: the Queen as representing -1 and the King as representing -2. Your
1.45 crook 3744: @i{range} remains unchanged (you can still only represent a total of 13
1.29 crook 3745: numbers) but the numbers that you can represent are -2 through 10.
1.28 crook 3746:
1.29 crook 3747: In that analogy, the limit was the amount of information that a single
3748: stack entry could hold, and Forth has a similar limit. In Forth, the
3749: size of a stack entry is called a @dfn{cell}. The actual size of a cell is
3750: implementation dependent and affects the maximum value that a stack
3751: entry can hold. A Standard Forth provides a cell size of at least
3752: 16-bits, and most desktop systems use a cell size of 32-bits.
1.21 crook 3753:
1.29 crook 3754: Forth does not do any type checking for you, so you are free to
3755: manipulate and combine stack items in any way you wish. A convenient way
3756: of treating stack items is as 2's complement signed integers, and that
3757: is what Standard words like @code{+} do. Therefore you can type:
1.21 crook 3758:
1.29 crook 3759: @example
1.30 anton 3760: @kbd{-5 12 + .s@key{RET}} <1> 7 ok
1.29 crook 3761: @end example
1.21 crook 3762:
1.29 crook 3763: If you use numbers and definitions like @code{+} in order to turn Forth
3764: into a great big pocket calculator, you will realise that it's rather
3765: different from a normal calculator. Rather than typing 2 + 3 = you had
3766: to type 2 3 + (ignore the fact that you had to use @code{.s} to see the
3767: result). The terminology used to describe this difference is to say that
3768: your calculator uses @dfn{Infix Notation} (parameters and operators are
3769: mixed) whilst Forth uses @dfn{Postfix Notation} (parameters and
3770: operators are separate), also called @dfn{Reverse Polish Notation}.
1.21 crook 3771:
1.29 crook 3772: Whilst postfix notation might look confusing to begin with, it has
3773: several important advantages:
1.21 crook 3774:
1.23 crook 3775: @itemize @bullet
3776: @item
1.29 crook 3777: it is unambiguous
1.23 crook 3778: @item
1.29 crook 3779: it is more concise
1.23 crook 3780: @item
1.29 crook 3781: it fits naturally with a stack-based system
1.23 crook 3782: @end itemize
1.21 crook 3783:
1.29 crook 3784: To examine these claims in more detail, consider these sums:
1.21 crook 3785:
1.29 crook 3786: @example
3787: 6 + 5 * 4 =
3788: 4 * 5 + 6 =
3789: @end example
1.21 crook 3790:
1.29 crook 3791: If you're just learning maths or your maths is very rusty, you will
3792: probably come up with the answer 44 for the first and 26 for the
3793: second. If you are a bit of a whizz at maths you will remember the
3794: @i{convention} that multiplication takes precendence over addition, and
3795: you'd come up with the answer 26 both times. To explain the answer 26
3796: to someone who got the answer 44, you'd probably rewrite the first sum
3797: like this:
1.21 crook 3798:
1.29 crook 3799: @example
3800: 6 + (5 * 4) =
3801: @end example
1.21 crook 3802:
1.29 crook 3803: If what you really wanted was to perform the addition before the
3804: multiplication, you would have to use parentheses to force it.
1.21 crook 3805:
1.29 crook 3806: If you did the first two sums on a pocket calculator you would probably
3807: get the right answers, unless you were very cautious and entered them using
3808: these keystroke sequences:
1.21 crook 3809:
1.29 crook 3810: 6 + 5 = * 4 =
3811: 4 * 5 = + 6 =
1.21 crook 3812:
1.29 crook 3813: Postfix notation is unambiguous because the order that the operators
3814: are applied is always explicit; that also means that parentheses are
3815: never required. The operators are @i{active} (the act of quoting the
3816: operator makes the operation occur) which removes the need for ``=''.
1.28 crook 3817:
1.29 crook 3818: The sum 6 + 5 * 4 can be written (in postfix notation) in two
3819: equivalent ways:
1.26 crook 3820:
3821: @example
1.29 crook 3822: 6 5 4 * + or:
3823: 5 4 * 6 +
1.26 crook 3824: @end example
1.23 crook 3825:
1.29 crook 3826: An important thing that you should notice about this notation is that
3827: the @i{order} of the numbers does not change; if you want to subtract
3828: 2 from 10 you type @code{10 2 -}.
1.1 anton 3829:
1.29 crook 3830: The reason that Forth uses postfix notation is very simple to explain: it
3831: makes the implementation extremely simple, and it follows naturally from
3832: using the stack as a mechanism for passing parameters. Another way of
3833: thinking about this is to realise that all Forth definitions are
3834: @i{active}; they execute as they are encountered by the text
3835: interpreter. The result of this is that the syntax of Forth is trivially
3836: simple.
1.1 anton 3837:
3838:
3839:
1.29 crook 3840: @comment ----------------------------------------------
3841: @node Your first definition, How does that work?, Stacks and Postfix notation, Introduction
3842: @section Your first Forth definition
3843: @cindex first definition
1.1 anton 3844:
1.29 crook 3845: Until now, the examples we've seen have been trivial; we've just been
3846: using Forth as a bigger-than-pocket calculator. Also, each calculation
3847: we've shown has been a ``one-off'' -- to repeat it we'd need to type it in
3848: again@footnote{That's not quite true. If you press the up-arrow key on
3849: your keyboard you should be able to scroll back to any earlier command,
3850: edit it and re-enter it.} In this section we'll see how to add new
3851: words to Forth's vocabulary.
1.1 anton 3852:
1.29 crook 3853: The easiest way to create a new word is to use a @dfn{colon
3854: definition}. We'll define a few and try them out before worrying too
3855: much about how they work. Try typing in these examples; be careful to
3856: copy the spaces accurately:
1.1 anton 3857:
1.29 crook 3858: @example
3859: : add-two 2 + . ;
3860: : greet ." Hello and welcome" ;
3861: : demo 5 add-two ;
3862: @end example
1.1 anton 3863:
1.29 crook 3864: @noindent
3865: Now try them out:
1.1 anton 3866:
1.29 crook 3867: @example
1.30 anton 3868: @kbd{greet@key{RET}} Hello and welcome ok
3869: @kbd{greet greet@key{RET}} Hello and welcomeHello and welcome ok
3870: @kbd{4 add-two@key{RET}} 6 ok
3871: @kbd{demo@key{RET}} 7 ok
3872: @kbd{9 greet demo add-two@key{RET}} Hello and welcome7 11 ok
1.29 crook 3873: @end example
1.1 anton 3874:
1.29 crook 3875: The first new thing that we've introduced here is the pair of words
3876: @code{:} and @code{;}. These are used to start and terminate a new
3877: definition, respectively. The first word after the @code{:} is the name
3878: for the new definition.
1.1 anton 3879:
1.29 crook 3880: As you can see from the examples, a definition is built up of words that
3881: have already been defined; Forth makes no distinction between
3882: definitions that existed when you started the system up, and those that
3883: you define yourself.
1.1 anton 3884:
1.29 crook 3885: The examples also introduce the words @code{.} (dot), @code{."}
3886: (dot-quote) and @code{dup} (dewp). Dot takes the value from the top of
3887: the stack and displays it. It's like @code{.s} except that it only
3888: displays the top item of the stack and it is destructive; after it has
3889: executed, the number is no longer on the stack. There is always one
3890: space printed after the number, and no spaces before it. Dot-quote
3891: defines a string (a sequence of characters) that will be printed when
3892: the word is executed. The string can contain any printable characters
3893: except @code{"}. A @code{"} has a special function; it is not a Forth
3894: word but it acts as a delimiter (the way that delimiters work is
3895: described in the next section). Finally, @code{dup} duplicates the value
3896: at the top of the stack. Try typing @code{5 dup .s} to see what it does.
1.1 anton 3897:
1.29 crook 3898: We already know that the text interpreter searches through the
3899: dictionary to locate names. If you've followed the examples earlier, you
3900: will already have a definition called @code{add-two}. Lets try modifying
3901: it by typing in a new definition:
1.1 anton 3902:
1.29 crook 3903: @example
1.30 anton 3904: @kbd{: add-two dup . ." + 2 =" 2 + . ;@key{RET}} redefined add-two ok
1.29 crook 3905: @end example
1.5 anton 3906:
1.29 crook 3907: Forth recognised that we were defining a word that already exists, and
3908: printed a message to warn us of that fact. Let's try out the new
3909: definition:
1.5 anton 3910:
1.29 crook 3911: @example
1.30 anton 3912: @kbd{9 add-two@key{RET}} 9 + 2 =11 ok
1.29 crook 3913: @end example
1.1 anton 3914:
1.29 crook 3915: @noindent
3916: All that we've actually done here, though, is to create a new
3917: definition, with a particular name. The fact that there was already a
3918: definition with the same name did not make any difference to the way
3919: that the new definition was created (except that Forth printed a warning
3920: message). The old definition of add-two still exists (try @code{demo}
3921: again to see that this is true). Any new definition will use the new
3922: definition of @code{add-two}, but old definitions continue to use the
3923: version that already existed at the time that they were @code{compiled}.
1.1 anton 3924:
1.29 crook 3925: Before you go on to the next section, try defining and redefining some
3926: words of your own.
1.1 anton 3927:
1.29 crook 3928: @comment ----------------------------------------------
3929: @node How does that work?, Forth is written in Forth, Your first definition, Introduction
3930: @section How does that work?
3931: @cindex parsing words
1.1 anton 3932:
1.30 anton 3933: @c That's pretty deep (IMO way too deep) for an introduction. - anton
3934:
3935: @c Is it a good idea to talk about the interpretation semantics of a
3936: @c number? We don't have an xt to go along with it. - anton
3937:
3938: @c Now that I have eliminated execution semantics, I wonder if it would not
3939: @c be better to keep them (or add run-time semantics), to make it easier to
3940: @c explain what compilation semantics usually does. - anton
3941:
1.44 crook 3942: @c nac-> I removed the term ``default compilation sematics'' from the
3943: @c introductory chapter. Removing ``execution semantics'' was making
3944: @c everything simpler to explain, then I think the use of this term made
3945: @c everything more complex again. I replaced it with ``default
3946: @c semantics'' (which is used elsewhere in the manual) by which I mean
3947: @c ``a definition that has neither the immediate nor the compile-only
3948: @c flag set''. I reworded big chunks of the ``how does that work''
3949: @c section (and, unusually for me, I think I even made it shorter!). See
3950: @c what you think -- I know I have not addressed your primary concern
3951: @c that it is too heavy-going for an introduction. From what I understood
3952: @c of your course notes it looks as though they might be a good framework.
3953: @c Things that I've tried to capture here are some things that came as a
3954: @c great revelation here when I first understood them. Also, I like the
3955: @c fact that a very simple code example shows up almost all of the issues
3956: @c that you need to understand to see how Forth works. That's unique and
3957: @c worthwhile to emphasise.
3958:
1.29 crook 3959: Now we're going to take another look at the definition of @code{add-two}
3960: from the previous section. From our knowledge of the way that the text
3961: interpreter works, we would have expected this result when we tried to
3962: define @code{add-two}:
1.21 crook 3963:
1.29 crook 3964: @example
1.44 crook 3965: @kbd{: add-two 2 + . ;@key{RET}}
1.29 crook 3966: ^^^^^^^
3967: Error: Undefined word
3968: @end example
1.28 crook 3969:
1.29 crook 3970: The reason that this didn't happen is bound up in the way that @code{:}
3971: works. The word @code{:} does two special things. The first special
3972: thing that it does prevents the text interpreter from ever seeing the
3973: characters @code{add-two}. The text interpreter uses a variable called
3974: @cindex modifying >IN
1.44 crook 3975: @code{>IN} (pronounced ``to-in'') to keep track of where it is in the
1.29 crook 3976: input line. When it encounters the word @code{:} it behaves in exactly
3977: the same way as it does for any other word; it looks it up in the name
3978: dictionary, finds its xt and executes it. When @code{:} executes, it
3979: looks at the input buffer, finds the word @code{add-two} and advances the
3980: value of @code{>IN} to point past it. It then does some other stuff
3981: associated with creating the new definition (including creating an entry
3982: for @code{add-two} in the name dictionary). When the execution of @code{:}
3983: completes, control returns to the text interpreter, which is oblivious
3984: to the fact that it has been tricked into ignoring part of the input
3985: line.
1.21 crook 3986:
1.29 crook 3987: @cindex parsing words
3988: Words like @code{:} -- words that advance the value of @code{>IN} and so
3989: prevent the text interpreter from acting on the whole of the input line
3990: -- are called @dfn{parsing words}.
1.21 crook 3991:
1.29 crook 3992: @cindex @code{state} - effect on the text interpreter
3993: @cindex text interpreter - effect of state
3994: The second special thing that @code{:} does is change the value of a
3995: variable called @code{state}, which affects the way that the text
3996: interpreter behaves. When Gforth starts up, @code{state} has the value
3997: 0, and the text interpreter is said to be @dfn{interpreting}. During a
3998: colon definition (started with @code{:}), @code{state} is set to -1 and
1.44 crook 3999: the text interpreter is said to be @dfn{compiling}.
4000:
4001: In this example, the text interpreter is compiling when it processes the
4002: string ``@code{2 + . ;}''. It still breaks the string down into
4003: character sequences in the same way. However, instead of pushing the
4004: number @code{2} onto the stack, it lays down (@dfn{compiles}) some magic
4005: into the definition of @code{add-two} that will make the number @code{2} get
4006: pushed onto the stack when @code{add-two} is @dfn{executed}. Similarly,
4007: the behaviours of @code{+} and @code{.} are also compiled into the
4008: definition.
4009:
4010: One category of words don't get compiled. These so-called @dfn{immediate
4011: words} get executed (performed @i{now}) regardless of whether the text
4012: interpreter is interpreting or compiling. The word @code{;} is an
4013: immediate word. Rather than being compiled into the definition, it
4014: executes. Its effect is to terminate the current definition, which
4015: includes changing the value of @code{state} back to 0.
4016:
4017: When you execute @code{add-two}, it has a @dfn{run-time effect} that is
4018: exactly the same as if you had typed @code{2 + . @key{RET}} outside of a
4019: definition.
1.28 crook 4020:
1.30 anton 4021: In Forth, every word or number can be described in terms of two
1.29 crook 4022: properties:
1.28 crook 4023:
4024: @itemize @bullet
4025: @item
1.29 crook 4026: @cindex interpretation semantics
1.44 crook 4027: Its @dfn{interpretation semantics} describe how it will behave when the
4028: text interpreter encounters it in @dfn{interpret} state. The
4029: interpretation semantics of a word are represented by an @dfn{execution
4030: token}.
1.28 crook 4031: @item
1.29 crook 4032: @cindex compilation semantics
1.44 crook 4033: Its @dfn{compilation semantics} describe how it will behave when the
4034: text interpreter encounters it in @dfn{compile} state. The compilation
4035: semantics of a word are represented in an implementation-dependent way;
4036: Gforth uses a @dfn{compilation token}.
1.29 crook 4037: @end itemize
4038:
4039: @noindent
4040: Numbers are always treated in a fixed way:
4041:
4042: @itemize @bullet
1.28 crook 4043: @item
1.44 crook 4044: When the number is @dfn{interpreted}, its behaviour is to push the
4045: number onto the stack.
1.28 crook 4046: @item
1.30 anton 4047: When the number is @dfn{compiled}, a piece of code is appended to the
4048: current definition that pushes the number when it runs. (In other words,
4049: the compilation semantics of a number are to postpone its interpretation
4050: semantics until the run-time of the definition that it is being compiled
4051: into.)
1.29 crook 4052: @end itemize
4053:
1.44 crook 4054: Words don't behave in such a regular way, but most have @i{default
4055: semantics} which means that they behave like this:
1.29 crook 4056:
4057: @itemize @bullet
1.28 crook 4058: @item
1.30 anton 4059: The @dfn{interpretation semantics} of the word are to do something useful.
4060: @item
1.29 crook 4061: The @dfn{compilation semantics} of the word are to append its
1.30 anton 4062: @dfn{interpretation semantics} to the current definition (so that its
4063: run-time behaviour is to do something useful).
1.28 crook 4064: @end itemize
4065:
1.30 anton 4066: @cindex immediate words
1.44 crook 4067: The actual behaviour of any particular word can be controlled by using
4068: the words @code{immediate} and @code{compile-only} when the word is
4069: defined. These words set flags in the name dictionary entry of the most
4070: recently defined word, and these flags are retrieved by the text
4071: interpreter when it finds the word in the name dictionary.
4072:
4073: A word that is marked as @dfn{immediate} has compilation semantics that
4074: are identical to its interpretation semantics. In other words, it
4075: behaves like this:
1.29 crook 4076:
4077: @itemize @bullet
4078: @item
1.30 anton 4079: The @dfn{interpretation semantics} of the word are to do something useful.
1.29 crook 4080: @item
1.30 anton 4081: The @dfn{compilation semantics} of the word are to do something useful
4082: (and actually the same thing); i.e., it is executed during compilation.
1.29 crook 4083: @end itemize
1.28 crook 4084:
1.44 crook 4085: Marking a word as @dfn{compile-only} prohibits the text interpreter from
4086: performing the interpretation semantics of the word directly; an attempt
4087: to do so will generate an error. It is never necessary to use
4088: @code{compile-only} (and it is not even part of ANS Forth, though it is
4089: provided by many implementations) but it is good etiquette to apply it
4090: to a word that will not behave correctly (and might have unexpected
4091: side-effects) in interpret state. For example, it is only legal to use
4092: the conditional word @code{IF} within a definition. If you forget this
4093: and try to use it elsewhere, the fact that (in Gforth) it is marked as
4094: @code{compile-only} allows the text interpreter to generate a helpful
4095: error message rather than subjecting you to the consequences of your
4096: folly.
4097:
1.29 crook 4098: This example shows the difference between an immediate and a
4099: non-immediate word:
1.28 crook 4100:
1.29 crook 4101: @example
4102: : show-state state @@ . ;
4103: : show-state-now show-state ; immediate
4104: : word1 show-state ;
4105: : word2 show-state-now ;
1.28 crook 4106: @end example
1.23 crook 4107:
1.29 crook 4108: The word @code{immediate} after the definition of @code{show-state-now}
4109: makes that word an immediate word. These definitions introduce a new
4110: word: @code{@@} (pronounced ``fetch''). This word fetches the value of a
4111: variable, and leaves it on the stack. Therefore, the behaviour of
4112: @code{show-state} is to print a number that represents the current value
4113: of @code{state}.
1.28 crook 4114:
1.29 crook 4115: When you execute @code{word1}, it prints the number 0, indicating that
4116: the system is interpreting. When the text interpreter compiled the
4117: definition of @code{word1}, it encountered @code{show-state} whose
1.30 anton 4118: compilation semantics are to append its interpretation semantics to the
1.29 crook 4119: current definition. When you execute @code{word1}, it performs the
1.30 anton 4120: interpretation semantics of @code{show-state}. At the time that @code{word1}
1.29 crook 4121: (and therefore @code{show-state}) are executed, the system is
4122: interpreting.
1.28 crook 4123:
1.30 anton 4124: When you pressed @key{RET} after entering the definition of @code{word2},
1.29 crook 4125: you should have seen the number -1 printed, followed by ``@code{
4126: ok}''. When the text interpreter compiled the definition of
4127: @code{word2}, it encountered @code{show-state-now}, an immediate word,
1.30 anton 4128: whose compilation semantics are therefore to perform its interpretation
1.29 crook 4129: semantics. It is executed straight away (even before the text
4130: interpreter has moved on to process another group of characters; the
4131: @code{;} in this example). The effect of executing it are to display the
4132: value of @code{state} @i{at the time that the definition of}
4133: @code{word2} @i{is being defined}. Printing -1 demonstrates that the
4134: system is compiling at this time. If you execute @code{word2} it does
4135: nothing at all.
1.28 crook 4136:
1.29 crook 4137: @cindex @code{."}, how it works
4138: Before leaving the subject of immediate words, consider the behaviour of
4139: @code{."} in the definition of @code{greet}, in the previous
4140: section. This word is both a parsing word and an immediate word. Notice
4141: that there is a space between @code{."} and the start of the text
4142: @code{Hello and welcome}, but that there is no space between the last
4143: letter of @code{welcome} and the @code{"} character. The reason for this
4144: is that @code{."} is a Forth word; it must have a space after it so that
4145: the text interpreter can identify it. The @code{"} is not a Forth word;
4146: it is a @dfn{delimiter}. The examples earlier show that, when the string
4147: is displayed, there is neither a space before the @code{H} nor after the
4148: @code{e}. Since @code{."} is an immediate word, it executes at the time
4149: that @code{greet} is defined. When it executes, its behaviour is to
4150: search forward in the input line looking for the delimiter. When it
4151: finds the delimiter, it updates @code{>IN} to point past the
4152: delimiter. It also compiles some magic code into the definition of
4153: @code{greet}; the xt of a run-time routine that prints a text string. It
4154: compiles the string @code{Hello and welcome} into memory so that it is
4155: available to be printed later. When the text interpreter gains control,
4156: the next word it finds in the input stream is @code{;} and so it
4157: terminates the definition of @code{greet}.
1.28 crook 4158:
4159:
4160: @comment ----------------------------------------------
1.29 crook 4161: @node Forth is written in Forth, Review - elements of a Forth system, How does that work?, Introduction
4162: @section Forth is written in Forth
4163: @cindex structure of Forth programs
4164:
4165: When you start up a Forth compiler, a large number of definitions
4166: already exist. In Forth, you develop a new application using bottom-up
4167: programming techniques to create new definitions that are defined in
4168: terms of existing definitions. As you create each definition you can
4169: test and debug it interactively.
4170:
4171: If you have tried out the examples in this section, you will probably
4172: have typed them in by hand; when you leave Gforth, your definitions will
4173: be lost. You can avoid this by using a text editor to enter Forth source
4174: code into a file, and then loading code from the file using
1.49 anton 4175: @code{include} (@pxref{Forth source files}). A Forth source file is
1.29 crook 4176: processed by the text interpreter, just as though you had typed it in by
4177: hand@footnote{Actually, there are some subtle differences -- see
4178: @ref{The Text Interpreter}.}.
4179:
4180: Gforth also supports the traditional Forth alternative to using text
1.49 anton 4181: files for program entry (@pxref{Blocks}).
1.28 crook 4182:
1.29 crook 4183: In common with many, if not most, Forth compilers, most of Gforth is
4184: actually written in Forth. All of the @file{.fs} files in the
4185: installation directory@footnote{For example,
1.30 anton 4186: @file{/usr/local/share/gforth...}} are Forth source files, which you can
1.29 crook 4187: study to see examples of Forth programming.
1.28 crook 4188:
1.29 crook 4189: Gforth maintains a history file that records every line that you type to
4190: the text interpreter. This file is preserved between sessions, and is
4191: used to provide a command-line recall facility. If you enter long
4192: definitions by hand, you can use a text editor to paste them out of the
4193: history file into a Forth source file for reuse at a later time
1.49 anton 4194: (for more information @pxref{Command-line editing}).
1.28 crook 4195:
4196:
4197: @comment ----------------------------------------------
1.29 crook 4198: @node Review - elements of a Forth system, Where to go next, Forth is written in Forth, Introduction
4199: @section Review - elements of a Forth system
4200: @cindex elements of a Forth system
1.28 crook 4201:
1.29 crook 4202: To summarise this chapter:
1.28 crook 4203:
4204: @itemize @bullet
4205: @item
1.29 crook 4206: Forth programs use @dfn{factoring} to break a problem down into small
4207: fragments called @dfn{words} or @dfn{definitions}.
4208: @item
4209: Forth program development is an interactive process.
4210: @item
4211: The main command loop that accepts input, and controls both
4212: interpretation and compilation, is called the @dfn{text interpreter}
4213: (also known as the @dfn{outer interpreter}).
4214: @item
4215: Forth has a very simple syntax, consisting of words and numbers
4216: separated by spaces or carriage-return characters. Any additional syntax
4217: is imposed by @dfn{parsing words}.
4218: @item
4219: Forth uses a stack to pass parameters between words. As a result, it
4220: uses postfix notation.
4221: @item
4222: To use a word that has previously been defined, the text interpreter
4223: searches for the word in the @dfn{name dictionary}.
4224: @item
1.30 anton 4225: Words have @dfn{interpretation semantics} and @dfn{compilation semantics}.
1.28 crook 4226: @item
1.29 crook 4227: The text interpreter uses the value of @code{state} to select between
4228: the use of the @dfn{interpretation semantics} and the @dfn{compilation
4229: semantics} of a word that it encounters.
1.28 crook 4230: @item
1.30 anton 4231: The relationship between the @dfn{interpretation semantics} and
4232: @dfn{compilation semantics} for a word
1.29 crook 4233: depend upon the way in which the word was defined (for example, whether
4234: it is an @dfn{immediate} word).
1.28 crook 4235: @item
1.29 crook 4236: Forth definitions can be implemented in Forth (called @dfn{high-level
4237: definitions}) or in some other way (usually a lower-level language and
4238: as a result often called @dfn{low-level definitions}, @dfn{code
4239: definitions} or @dfn{primitives}).
1.28 crook 4240: @item
1.29 crook 4241: Many Forth systems are implemented mainly in Forth.
1.28 crook 4242: @end itemize
4243:
4244:
1.29 crook 4245: @comment ----------------------------------------------
1.48 anton 4246: @node Where to go next, Exercises, Review - elements of a Forth system, Introduction
1.29 crook 4247: @section Where To Go Next
4248: @cindex where to go next
1.28 crook 4249:
1.29 crook 4250: Amazing as it may seem, if you have read (and understood) this far, you
4251: know almost all the fundamentals about the inner workings of a Forth
4252: system. You certainly know enough to be able to read and understand the
4253: rest of this manual and the ANS Forth document, to learn more about the
4254: facilities that Forth in general and Gforth in particular provide. Even
4255: scarier, you know almost enough to implement your own Forth system.
1.30 anton 4256: However, that's not a good idea just yet... better to try writing some
1.29 crook 4257: programs in Gforth.
1.28 crook 4258:
1.29 crook 4259: Forth has such a rich vocabulary that it can be hard to know where to
4260: start in learning it. This section suggests a few sets of words that are
4261: enough to write small but useful programs. Use the word index in this
4262: document to learn more about each word, then try it out and try to write
4263: small definitions using it. Start by experimenting with these words:
1.28 crook 4264:
4265: @itemize @bullet
4266: @item
1.29 crook 4267: Arithmetic: @code{+ - * / /MOD */ ABS INVERT}
4268: @item
4269: Comparison: @code{MIN MAX =}
4270: @item
4271: Logic: @code{AND OR XOR NOT}
4272: @item
4273: Stack manipulation: @code{DUP DROP SWAP OVER}
1.28 crook 4274: @item
1.29 crook 4275: Loops and decisions: @code{IF ELSE ENDIF ?DO I LOOP}
1.28 crook 4276: @item
1.29 crook 4277: Input/Output: @code{. ." EMIT CR KEY}
1.28 crook 4278: @item
1.29 crook 4279: Defining words: @code{: ; CREATE}
1.28 crook 4280: @item
1.29 crook 4281: Memory allocation words: @code{ALLOT ,}
1.28 crook 4282: @item
1.29 crook 4283: Tools: @code{SEE WORDS .S MARKER}
4284: @end itemize
4285:
4286: When you have mastered those, go on to:
4287:
4288: @itemize @bullet
1.28 crook 4289: @item
1.29 crook 4290: More defining words: @code{VARIABLE CONSTANT VALUE TO CREATE DOES>}
1.28 crook 4291: @item
1.29 crook 4292: Memory access: @code{@@ !}
1.28 crook 4293: @end itemize
1.23 crook 4294:
1.29 crook 4295: When you have mastered these, there's nothing for it but to read through
4296: the whole of this manual and find out what you've missed.
4297:
4298: @comment ----------------------------------------------
1.48 anton 4299: @node Exercises, , Where to go next, Introduction
1.29 crook 4300: @section Exercises
4301: @cindex exercises
4302:
4303: TODO: provide a set of programming excercises linked into the stuff done
4304: already and into other sections of the manual. Provide solutions to all
4305: the exercises in a .fs file in the distribution.
4306:
4307: @c Get some inspiration from Starting Forth and Kelly&Spies.
4308:
4309: @c excercises:
4310: @c 1. take inches and convert to feet and inches.
4311: @c 2. take temperature and convert from fahrenheight to celcius;
4312: @c may need to care about symmetric vs floored??
4313: @c 3. take input line and do character substitution
4314: @c to encipher or decipher
4315: @c 4. as above but work on a file for in and out
4316: @c 5. take input line and convert to pig-latin
4317: @c
4318: @c thing of sets of things to exercise then come up with
4319: @c problems that need those things.
4320:
4321:
1.26 crook 4322: @c ******************************************************************
1.29 crook 4323: @node Words, Error messages, Introduction, Top
1.1 anton 4324: @chapter Forth Words
1.26 crook 4325: @cindex words
1.1 anton 4326:
4327: @menu
4328: * Notation::
1.65 anton 4329: * Case insensitivity::
4330: * Comments::
4331: * Boolean Flags::
1.1 anton 4332: * Arithmetic::
4333: * Stack Manipulation::
1.5 anton 4334: * Memory::
1.1 anton 4335: * Control Structures::
4336: * Defining Words::
1.65 anton 4337: * Interpretation and Compilation Semantics::
1.47 crook 4338: * Tokens for Words::
1.81 ! anton 4339: * Compiling words::
1.65 anton 4340: * The Text Interpreter::
4341: * Word Lists::
4342: * Environmental Queries::
1.12 anton 4343: * Files::
4344: * Blocks::
4345: * Other I/O::
1.78 anton 4346: * Locals::
4347: * Structures::
4348: * Object-oriented Forth::
1.12 anton 4349: * Programming Tools::
4350: * Assembler and Code Words::
4351: * Threading Words::
1.65 anton 4352: * Passing Commands to the OS::
4353: * Keeping track of Time::
4354: * Miscellaneous Words::
1.1 anton 4355: @end menu
4356:
1.65 anton 4357: @node Notation, Case insensitivity, Words, Words
1.1 anton 4358: @section Notation
4359: @cindex notation of glossary entries
4360: @cindex format of glossary entries
4361: @cindex glossary notation format
4362: @cindex word glossary entry format
4363:
4364: The Forth words are described in this section in the glossary notation
1.67 anton 4365: that has become a de-facto standard for Forth texts:
1.1 anton 4366:
4367: @format
1.29 crook 4368: @i{word} @i{Stack effect} @i{wordset} @i{pronunciation}
1.1 anton 4369: @end format
1.29 crook 4370: @i{Description}
1.1 anton 4371:
4372: @table @var
4373: @item word
1.28 crook 4374: The name of the word.
1.1 anton 4375:
4376: @item Stack effect
4377: @cindex stack effect
1.29 crook 4378: The stack effect is written in the notation @code{@i{before} --
4379: @i{after}}, where @i{before} and @i{after} describe the top of
1.1 anton 4380: stack entries before and after the execution of the word. The rest of
4381: the stack is not touched by the word. The top of stack is rightmost,
4382: i.e., a stack sequence is written as it is typed in. Note that Gforth
4383: uses a separate floating point stack, but a unified stack
1.29 crook 4384: notation. Also, return stack effects are not shown in @i{stack
4385: effect}, but in @i{Description}. The name of a stack item describes
1.1 anton 4386: the type and/or the function of the item. See below for a discussion of
4387: the types.
4388:
4389: All words have two stack effects: A compile-time stack effect and a
4390: run-time stack effect. The compile-time stack-effect of most words is
1.29 crook 4391: @i{ -- }. If the compile-time stack-effect of a word deviates from
1.1 anton 4392: this standard behaviour, or the word does other unusual things at
4393: compile time, both stack effects are shown; otherwise only the run-time
4394: stack effect is shown.
4395:
4396: @cindex pronounciation of words
4397: @item pronunciation
4398: How the word is pronounced.
4399:
4400: @cindex wordset
1.67 anton 4401: @cindex environment wordset
1.1 anton 4402: @item wordset
1.21 crook 4403: The ANS Forth standard is divided into several word sets. A standard
4404: system need not support all of them. Therefore, in theory, the fewer
4405: word sets your program uses the more portable it will be. However, we
4406: suspect that most ANS Forth systems on personal machines will feature
1.26 crook 4407: all word sets. Words that are not defined in ANS Forth have
1.21 crook 4408: @code{gforth} or @code{gforth-internal} as word set. @code{gforth}
1.1 anton 4409: describes words that will work in future releases of Gforth;
4410: @code{gforth-internal} words are more volatile. Environmental query
4411: strings are also displayed like words; you can recognize them by the
1.21 crook 4412: @code{environment} in the word set field.
1.1 anton 4413:
4414: @item Description
4415: A description of the behaviour of the word.
4416: @end table
4417:
4418: @cindex types of stack items
4419: @cindex stack item types
4420: The type of a stack item is specified by the character(s) the name
4421: starts with:
4422:
4423: @table @code
4424: @item f
4425: @cindex @code{f}, stack item type
4426: Boolean flags, i.e. @code{false} or @code{true}.
4427: @item c
4428: @cindex @code{c}, stack item type
4429: Char
4430: @item w
4431: @cindex @code{w}, stack item type
4432: Cell, can contain an integer or an address
4433: @item n
4434: @cindex @code{n}, stack item type
4435: signed integer
4436: @item u
4437: @cindex @code{u}, stack item type
4438: unsigned integer
4439: @item d
4440: @cindex @code{d}, stack item type
4441: double sized signed integer
4442: @item ud
4443: @cindex @code{ud}, stack item type
4444: double sized unsigned integer
4445: @item r
4446: @cindex @code{r}, stack item type
4447: Float (on the FP stack)
1.21 crook 4448: @item a-
1.1 anton 4449: @cindex @code{a_}, stack item type
4450: Cell-aligned address
1.21 crook 4451: @item c-
1.1 anton 4452: @cindex @code{c_}, stack item type
4453: Char-aligned address (note that a Char may have two bytes in Windows NT)
1.21 crook 4454: @item f-
1.1 anton 4455: @cindex @code{f_}, stack item type
4456: Float-aligned address
1.21 crook 4457: @item df-
1.1 anton 4458: @cindex @code{df_}, stack item type
4459: Address aligned for IEEE double precision float
1.21 crook 4460: @item sf-
1.1 anton 4461: @cindex @code{sf_}, stack item type
4462: Address aligned for IEEE single precision float
4463: @item xt
4464: @cindex @code{xt}, stack item type
4465: Execution token, same size as Cell
4466: @item wid
4467: @cindex @code{wid}, stack item type
1.21 crook 4468: Word list ID, same size as Cell
1.68 anton 4469: @item ior, wior
4470: @cindex ior type description
4471: @cindex wior type description
4472: I/O result code, cell-sized. In Gforth, you can @code{throw} iors.
1.1 anton 4473: @item f83name
4474: @cindex @code{f83name}, stack item type
4475: Pointer to a name structure
4476: @item "
4477: @cindex @code{"}, stack item type
1.12 anton 4478: string in the input stream (not on the stack). The terminating character
4479: is a blank by default. If it is not a blank, it is shown in @code{<>}
1.1 anton 4480: quotes.
4481: @end table
4482:
1.65 anton 4483: @comment ----------------------------------------------
4484: @node Case insensitivity, Comments, Notation, Words
4485: @section Case insensitivity
4486: @cindex case sensitivity
4487: @cindex upper and lower case
4488:
4489: Gforth is case-insensitive; you can enter definitions and invoke
4490: Standard words using upper, lower or mixed case (however,
4491: @pxref{core-idef, Implementation-defined options, Implementation-defined
4492: options}).
4493:
4494: ANS Forth only @i{requires} implementations to recognise Standard words
4495: when they are typed entirely in upper case. Therefore, a Standard
4496: program must use upper case for all Standard words. You can use whatever
4497: case you like for words that you define, but in a Standard program you
4498: have to use the words in the same case that you defined them.
4499:
4500: Gforth supports case sensitivity through @code{table}s (case-sensitive
4501: wordlists, @pxref{Word Lists}).
4502:
4503: Two people have asked how to convert Gforth to be case-sensitive; while
4504: we think this is a bad idea, you can change all wordlists into tables
4505: like this:
4506:
4507: @example
4508: ' table-find forth-wordlist wordlist-map @ !
4509: @end example
4510:
4511: Note that you now have to type the predefined words in the same case
4512: that we defined them, which are varying. You may want to convert them
4513: to your favourite case before doing this operation (I won't explain how,
4514: because if you are even contemplating doing this, you'd better have
4515: enough knowledge of Forth systems to know this already).
4516:
4517: @node Comments, Boolean Flags, Case insensitivity, Words
1.21 crook 4518: @section Comments
1.26 crook 4519: @cindex comments
1.21 crook 4520:
1.29 crook 4521: Forth supports two styles of comment; the traditional @i{in-line} comment,
4522: @code{(} and its modern cousin, the @i{comment to end of line}; @code{\}.
1.21 crook 4523:
1.44 crook 4524:
1.23 crook 4525: doc-(
1.21 crook 4526: doc-\
1.23 crook 4527: doc-\G
1.21 crook 4528:
1.44 crook 4529:
1.21 crook 4530: @node Boolean Flags, Arithmetic, Comments, Words
4531: @section Boolean Flags
1.26 crook 4532: @cindex Boolean flags
1.21 crook 4533:
4534: A Boolean flag is cell-sized. A cell with all bits clear represents the
4535: flag @code{false} and a flag with all bits set represents the flag
1.26 crook 4536: @code{true}. Words that check a flag (for example, @code{IF}) will treat
1.29 crook 4537: a cell that has @i{any} bit set as @code{true}.
1.67 anton 4538: @c on and off to Memory?
4539: @c true and false to "Bitwise operations" or "Numeric comparison"?
1.44 crook 4540:
1.21 crook 4541: doc-true
4542: doc-false
1.29 crook 4543: doc-on
4544: doc-off
1.21 crook 4545:
1.44 crook 4546:
1.21 crook 4547: @node Arithmetic, Stack Manipulation, Boolean Flags, Words
1.1 anton 4548: @section Arithmetic
4549: @cindex arithmetic words
4550:
4551: @cindex division with potentially negative operands
4552: Forth arithmetic is not checked, i.e., you will not hear about integer
4553: overflow on addition or multiplication, you may hear about division by
4554: zero if you are lucky. The operator is written after the operands, but
4555: the operands are still in the original order. I.e., the infix @code{2-1}
4556: corresponds to @code{2 1 -}. Forth offers a variety of division
4557: operators. If you perform division with potentially negative operands,
4558: you do not want to use @code{/} or @code{/mod} with its undefined
4559: behaviour, but rather @code{fm/mod} or @code{sm/mod} (probably the
4560: former, @pxref{Mixed precision}).
1.26 crook 4561: @comment TODO discuss the different division forms and the std approach
1.1 anton 4562:
4563: @menu
4564: * Single precision::
1.67 anton 4565: * Double precision:: Double-cell integer arithmetic
1.1 anton 4566: * Bitwise operations::
1.67 anton 4567: * Numeric comparison::
1.29 crook 4568: * Mixed precision:: Operations with single and double-cell integers
1.1 anton 4569: * Floating Point::
4570: @end menu
4571:
1.67 anton 4572: @node Single precision, Double precision, Arithmetic, Arithmetic
1.1 anton 4573: @subsection Single precision
4574: @cindex single precision arithmetic words
4575:
1.67 anton 4576: @c !! cell undefined
4577:
4578: By default, numbers in Forth are single-precision integers that are one
1.26 crook 4579: cell in size. They can be signed or unsigned, depending upon how you
1.49 anton 4580: treat them. For the rules used by the text interpreter for recognising
4581: single-precision integers see @ref{Number Conversion}.
1.21 crook 4582:
1.67 anton 4583: These words are all defined for signed operands, but some of them also
4584: work for unsigned numbers: @code{+}, @code{1+}, @code{-}, @code{1-},
4585: @code{*}.
1.44 crook 4586:
1.1 anton 4587: doc-+
1.21 crook 4588: doc-1+
1.1 anton 4589: doc--
1.21 crook 4590: doc-1-
1.1 anton 4591: doc-*
4592: doc-/
4593: doc-mod
4594: doc-/mod
4595: doc-negate
4596: doc-abs
4597: doc-min
4598: doc-max
1.27 crook 4599: doc-floored
1.1 anton 4600:
1.44 crook 4601:
1.67 anton 4602: @node Double precision, Bitwise operations, Single precision, Arithmetic
1.21 crook 4603: @subsection Double precision
4604: @cindex double precision arithmetic words
4605:
1.49 anton 4606: For the rules used by the text interpreter for
4607: recognising double-precision integers, see @ref{Number Conversion}.
1.21 crook 4608:
4609: A double precision number is represented by a cell pair, with the most
1.67 anton 4610: significant cell at the TOS. It is trivial to convert an unsigned single
4611: to a double: simply push a @code{0} onto the TOS. Since numbers are
4612: represented by Gforth using 2's complement arithmetic, converting a
4613: signed single to a (signed) double requires sign-extension across the
4614: most significant cell. This can be achieved using @code{s>d}. The moral
4615: of the story is that you cannot convert a number without knowing whether
4616: it represents an unsigned or a signed number.
1.21 crook 4617:
1.67 anton 4618: These words are all defined for signed operands, but some of them also
4619: work for unsigned numbers: @code{d+}, @code{d-}.
1.44 crook 4620:
1.21 crook 4621: doc-s>d
1.67 anton 4622: doc-d>s
1.21 crook 4623: doc-d+
4624: doc-d-
4625: doc-dnegate
4626: doc-dabs
4627: doc-dmin
4628: doc-dmax
4629:
1.44 crook 4630:
1.67 anton 4631: @node Bitwise operations, Numeric comparison, Double precision, Arithmetic
4632: @subsection Bitwise operations
4633: @cindex bitwise operation words
4634:
4635:
4636: doc-and
4637: doc-or
4638: doc-xor
4639: doc-invert
4640: doc-lshift
4641: doc-rshift
4642: doc-2*
4643: doc-d2*
4644: doc-2/
4645: doc-d2/
4646:
4647:
4648: @node Numeric comparison, Mixed precision, Bitwise operations, Arithmetic
1.21 crook 4649: @subsection Numeric comparison
4650: @cindex numeric comparison words
4651:
1.67 anton 4652: Note that the words that compare for equality (@code{= <> 0= 0<> d= d<>
4653: d0= d0<>}) work for for both signed and unsigned numbers.
1.44 crook 4654:
1.28 crook 4655: doc-<
4656: doc-<=
4657: doc-<>
4658: doc-=
4659: doc->
4660: doc->=
4661:
1.21 crook 4662: doc-0<
1.23 crook 4663: doc-0<=
1.21 crook 4664: doc-0<>
4665: doc-0=
1.23 crook 4666: doc-0>
4667: doc-0>=
1.28 crook 4668:
4669: doc-u<
4670: doc-u<=
1.44 crook 4671: @c u<> and u= exist but are the same as <> and =
1.31 anton 4672: @c doc-u<>
4673: @c doc-u=
1.28 crook 4674: doc-u>
4675: doc-u>=
4676:
4677: doc-within
4678:
4679: doc-d<
4680: doc-d<=
4681: doc-d<>
4682: doc-d=
4683: doc-d>
4684: doc-d>=
1.23 crook 4685:
1.21 crook 4686: doc-d0<
1.23 crook 4687: doc-d0<=
4688: doc-d0<>
1.21 crook 4689: doc-d0=
1.23 crook 4690: doc-d0>
4691: doc-d0>=
4692:
1.21 crook 4693: doc-du<
1.28 crook 4694: doc-du<=
1.44 crook 4695: @c du<> and du= exist but are the same as d<> and d=
1.31 anton 4696: @c doc-du<>
4697: @c doc-du=
1.28 crook 4698: doc-du>
4699: doc-du>=
1.1 anton 4700:
1.44 crook 4701:
1.21 crook 4702: @node Mixed precision, Floating Point, Numeric comparison, Arithmetic
1.1 anton 4703: @subsection Mixed precision
4704: @cindex mixed precision arithmetic words
4705:
1.44 crook 4706:
1.1 anton 4707: doc-m+
4708: doc-*/
4709: doc-*/mod
4710: doc-m*
4711: doc-um*
4712: doc-m*/
4713: doc-um/mod
4714: doc-fm/mod
4715: doc-sm/rem
4716:
1.44 crook 4717:
1.21 crook 4718: @node Floating Point, , Mixed precision, Arithmetic
1.1 anton 4719: @subsection Floating Point
4720: @cindex floating point arithmetic words
4721:
1.49 anton 4722: For the rules used by the text interpreter for
4723: recognising floating-point numbers see @ref{Number Conversion}.
1.1 anton 4724:
1.67 anton 4725: Gforth has a separate floating point stack, but the documentation uses
4726: the unified notation.@footnote{It's easy to generate the separate
4727: notation from that by just separating the floating-point numbers out:
4728: e.g. @code{( n r1 u r2 -- r3 )} becomes @code{( n u -- ) ( F: r1 r2 --
4729: r3 )}.}
1.1 anton 4730:
4731: @cindex floating-point arithmetic, pitfalls
4732: Floating point numbers have a number of unpleasant surprises for the
4733: unwary (e.g., floating point addition is not associative) and even a few
4734: for the wary. You should not use them unless you know what you are doing
4735: or you don't care that the results you get are totally bogus. If you
4736: want to learn about the problems of floating point numbers (and how to
1.66 anton 4737: avoid them), you might start with @cite{David Goldberg,
4738: @uref{http://www.validgh.com/goldberg/paper.ps,What Every Computer
4739: Scientist Should Know About Floating-Point Arithmetic}, ACM Computing
4740: Surveys 23(1):5@minus{}48, March 1991}.
1.1 anton 4741:
1.44 crook 4742:
1.21 crook 4743: doc-d>f
4744: doc-f>d
1.1 anton 4745: doc-f+
4746: doc-f-
4747: doc-f*
4748: doc-f/
4749: doc-fnegate
4750: doc-fabs
4751: doc-fmax
4752: doc-fmin
4753: doc-floor
4754: doc-fround
4755: doc-f**
4756: doc-fsqrt
4757: doc-fexp
4758: doc-fexpm1
4759: doc-fln
4760: doc-flnp1
4761: doc-flog
4762: doc-falog
1.32 anton 4763: doc-f2*
4764: doc-f2/
4765: doc-1/f
4766: doc-precision
4767: doc-set-precision
4768:
4769: @cindex angles in trigonometric operations
4770: @cindex trigonometric operations
4771: Angles in floating point operations are given in radians (a full circle
4772: has 2 pi radians).
4773:
1.1 anton 4774: doc-fsin
4775: doc-fcos
4776: doc-fsincos
4777: doc-ftan
4778: doc-fasin
4779: doc-facos
4780: doc-fatan
4781: doc-fatan2
4782: doc-fsinh
4783: doc-fcosh
4784: doc-ftanh
4785: doc-fasinh
4786: doc-facosh
4787: doc-fatanh
1.21 crook 4788: doc-pi
1.28 crook 4789:
1.32 anton 4790: @cindex equality of floats
4791: @cindex floating-point comparisons
1.31 anton 4792: One particular problem with floating-point arithmetic is that comparison
4793: for equality often fails when you would expect it to succeed. For this
4794: reason approximate equality is often preferred (but you still have to
1.67 anton 4795: know what you are doing). Also note that IEEE NaNs may compare
1.68 anton 4796: differently from what you might expect. The comparison words are:
1.31 anton 4797:
4798: doc-f~rel
4799: doc-f~abs
1.68 anton 4800: doc-f~
1.31 anton 4801: doc-f=
4802: doc-f<>
4803:
4804: doc-f<
4805: doc-f<=
4806: doc-f>
4807: doc-f>=
4808:
1.21 crook 4809: doc-f0<
1.28 crook 4810: doc-f0<=
4811: doc-f0<>
1.21 crook 4812: doc-f0=
1.28 crook 4813: doc-f0>
4814: doc-f0>=
4815:
1.1 anton 4816:
4817: @node Stack Manipulation, Memory, Arithmetic, Words
4818: @section Stack Manipulation
4819: @cindex stack manipulation words
4820:
4821: @cindex floating-point stack in the standard
1.21 crook 4822: Gforth maintains a number of separate stacks:
4823:
1.29 crook 4824: @cindex data stack
4825: @cindex parameter stack
1.21 crook 4826: @itemize @bullet
4827: @item
1.29 crook 4828: A data stack (also known as the @dfn{parameter stack}) -- for
4829: characters, cells, addresses, and double cells.
1.21 crook 4830:
1.29 crook 4831: @cindex floating-point stack
1.21 crook 4832: @item
1.44 crook 4833: A floating point stack -- for holding floating point (FP) numbers.
1.21 crook 4834:
1.29 crook 4835: @cindex return stack
1.21 crook 4836: @item
1.44 crook 4837: A return stack -- for holding the return addresses of colon
1.32 anton 4838: definitions and other (non-FP) data.
1.21 crook 4839:
1.29 crook 4840: @cindex locals stack
1.21 crook 4841: @item
1.44 crook 4842: A locals stack -- for holding local variables.
1.21 crook 4843: @end itemize
4844:
1.1 anton 4845: @menu
4846: * Data stack::
4847: * Floating point stack::
4848: * Return stack::
4849: * Locals stack::
4850: * Stack pointer manipulation::
4851: @end menu
4852:
4853: @node Data stack, Floating point stack, Stack Manipulation, Stack Manipulation
4854: @subsection Data stack
4855: @cindex data stack manipulation words
4856: @cindex stack manipulations words, data stack
4857:
1.44 crook 4858:
1.1 anton 4859: doc-drop
4860: doc-nip
4861: doc-dup
4862: doc-over
4863: doc-tuck
4864: doc-swap
1.21 crook 4865: doc-pick
1.1 anton 4866: doc-rot
4867: doc--rot
4868: doc-?dup
4869: doc-roll
4870: doc-2drop
4871: doc-2nip
4872: doc-2dup
4873: doc-2over
4874: doc-2tuck
4875: doc-2swap
4876: doc-2rot
4877:
1.44 crook 4878:
1.1 anton 4879: @node Floating point stack, Return stack, Data stack, Stack Manipulation
4880: @subsection Floating point stack
4881: @cindex floating-point stack manipulation words
4882: @cindex stack manipulation words, floating-point stack
4883:
1.32 anton 4884: Whilst every sane Forth has a separate floating-point stack, it is not
4885: strictly required; an ANS Forth system could theoretically keep
4886: floating-point numbers on the data stack. As an additional difficulty,
4887: you don't know how many cells a floating-point number takes. It is
4888: reportedly possible to write words in a way that they work also for a
4889: unified stack model, but we do not recommend trying it. Instead, just
4890: say that your program has an environmental dependency on a separate
4891: floating-point stack.
4892:
4893: doc-floating-stack
4894:
1.1 anton 4895: doc-fdrop
4896: doc-fnip
4897: doc-fdup
4898: doc-fover
4899: doc-ftuck
4900: doc-fswap
1.21 crook 4901: doc-fpick
1.1 anton 4902: doc-frot
4903:
1.44 crook 4904:
1.1 anton 4905: @node Return stack, Locals stack, Floating point stack, Stack Manipulation
4906: @subsection Return stack
4907: @cindex return stack manipulation words
4908: @cindex stack manipulation words, return stack
4909:
1.32 anton 4910: @cindex return stack and locals
4911: @cindex locals and return stack
4912: A Forth system is allowed to keep local variables on the
4913: return stack. This is reasonable, as local variables usually eliminate
4914: the need to use the return stack explicitly. So, if you want to produce
4915: a standard compliant program and you are using local variables in a
4916: word, forget about return stack manipulations in that word (refer to the
4917: standard document for the exact rules).
4918:
1.1 anton 4919: doc->r
4920: doc-r>
4921: doc-r@
4922: doc-rdrop
4923: doc-2>r
4924: doc-2r>
4925: doc-2r@
4926: doc-2rdrop
4927:
1.44 crook 4928:
1.1 anton 4929: @node Locals stack, Stack pointer manipulation, Return stack, Stack Manipulation
4930: @subsection Locals stack
4931:
1.78 anton 4932: Gforth uses an extra locals stack. It is described, along with the
4933: reasons for its existence, in @ref{Locals implementation}.
1.21 crook 4934:
1.1 anton 4935: @node Stack pointer manipulation, , Locals stack, Stack Manipulation
4936: @subsection Stack pointer manipulation
4937: @cindex stack pointer manipulation words
4938:
1.44 crook 4939: @c removed s0 r0 l0 -- they are obsolete aliases for sp0 rp0 lp0
1.21 crook 4940: doc-sp0
1.1 anton 4941: doc-sp@
4942: doc-sp!
1.21 crook 4943: doc-fp0
1.1 anton 4944: doc-fp@
4945: doc-fp!
1.21 crook 4946: doc-rp0
1.1 anton 4947: doc-rp@
4948: doc-rp!
1.21 crook 4949: doc-lp0
1.1 anton 4950: doc-lp@
4951: doc-lp!
4952:
1.44 crook 4953:
1.1 anton 4954: @node Memory, Control Structures, Stack Manipulation, Words
4955: @section Memory
1.26 crook 4956: @cindex memory words
1.1 anton 4957:
1.32 anton 4958: @menu
4959: * Memory model::
4960: * Dictionary allocation::
4961: * Heap Allocation::
4962: * Memory Access::
4963: * Address arithmetic::
4964: * Memory Blocks::
4965: @end menu
4966:
1.67 anton 4967: In addition to the standard Forth memory allocation words, there is also
4968: a @uref{http://www.complang.tuwien.ac.at/forth/garbage-collection.zip,
4969: garbage collector}.
4970:
1.32 anton 4971: @node Memory model, Dictionary allocation, Memory, Memory
4972: @subsection ANS Forth and Gforth memory models
4973:
4974: @c The ANS Forth description is a mess (e.g., is the heap part of
4975: @c the dictionary?), so let's not stick to closely with it.
4976:
1.67 anton 4977: ANS Forth considers a Forth system as consisting of several address
4978: spaces, of which only @dfn{data space} is managed and accessible with
4979: the memory words. Memory not necessarily in data space includes the
4980: stacks, the code (called code space) and the headers (called name
4981: space). In Gforth everything is in data space, but the code for the
4982: primitives is usually read-only.
1.32 anton 4983:
4984: Data space is divided into a number of areas: The (data space portion of
4985: the) dictionary@footnote{Sometimes, the term @dfn{dictionary} is used to
4986: refer to the search data structure embodied in word lists and headers,
4987: because it is used for looking up names, just as you would in a
4988: conventional dictionary.}, the heap, and a number of system-allocated
4989: buffers.
4990:
1.68 anton 4991: @cindex address arithmetic restrictions, ANS vs. Gforth
4992: @cindex contiguous regions, ANS vs. Gforth
1.32 anton 4993: In ANS Forth data space is also divided into contiguous regions. You
4994: can only use address arithmetic within a contiguous region, not between
4995: them. Usually each allocation gives you one contiguous region, but the
1.33 anton 4996: dictionary allocation words have additional rules (@pxref{Dictionary
1.32 anton 4997: allocation}).
4998:
4999: Gforth provides one big address space, and address arithmetic can be
5000: performed between any addresses. However, in the dictionary headers or
5001: code are interleaved with data, so almost the only contiguous data space
5002: regions there are those described by ANS Forth as contiguous; but you
5003: can be sure that the dictionary is allocated towards increasing
5004: addresses even between contiguous regions. The memory order of
5005: allocations in the heap is platform-dependent (and possibly different
5006: from one run to the next).
5007:
1.27 crook 5008:
1.32 anton 5009: @node Dictionary allocation, Heap Allocation, Memory model, Memory
5010: @subsection Dictionary allocation
1.27 crook 5011: @cindex reserving data space
5012: @cindex data space - reserving some
5013:
1.32 anton 5014: Dictionary allocation is a stack-oriented allocation scheme, i.e., if
5015: you want to deallocate X, you also deallocate everything
5016: allocated after X.
5017:
1.68 anton 5018: @cindex contiguous regions in dictionary allocation
1.32 anton 5019: The allocations using the words below are contiguous and grow the region
5020: towards increasing addresses. Other words that allocate dictionary
5021: memory of any kind (i.e., defining words including @code{:noname}) end
5022: the contiguous region and start a new one.
5023:
5024: In ANS Forth only @code{create}d words are guaranteed to produce an
5025: address that is the start of the following contiguous region. In
5026: particular, the cell allocated by @code{variable} is not guaranteed to
5027: be contiguous with following @code{allot}ed memory.
5028:
5029: You can deallocate memory by using @code{allot} with a negative argument
5030: (with some restrictions, see @code{allot}). For larger deallocations use
5031: @code{marker}.
1.27 crook 5032:
1.29 crook 5033:
1.27 crook 5034: doc-here
5035: doc-unused
5036: doc-allot
5037: doc-c,
1.29 crook 5038: doc-f,
1.27 crook 5039: doc-,
5040: doc-2,
5041:
1.32 anton 5042: Memory accesses have to be aligned (@pxref{Address arithmetic}). So of
5043: course you should allocate memory in an aligned way, too. I.e., before
5044: allocating allocating a cell, @code{here} must be cell-aligned, etc.
5045: The words below align @code{here} if it is not already. Basically it is
5046: only already aligned for a type, if the last allocation was a multiple
5047: of the size of this type and if @code{here} was aligned for this type
5048: before.
5049:
5050: After freshly @code{create}ing a word, @code{here} is @code{align}ed in
5051: ANS Forth (@code{maxalign}ed in Gforth).
5052:
5053: doc-align
5054: doc-falign
5055: doc-sfalign
5056: doc-dfalign
5057: doc-maxalign
5058: doc-cfalign
5059:
5060:
5061: @node Heap Allocation, Memory Access, Dictionary allocation, Memory
5062: @subsection Heap allocation
5063: @cindex heap allocation
5064: @cindex dynamic allocation of memory
5065: @cindex memory-allocation word set
5066:
1.68 anton 5067: @cindex contiguous regions and heap allocation
1.32 anton 5068: Heap allocation supports deallocation of allocated memory in any
5069: order. Dictionary allocation is not affected by it (i.e., it does not
5070: end a contiguous region). In Gforth, these words are implemented using
5071: the standard C library calls malloc(), free() and resize().
5072:
1.68 anton 5073: The memory region produced by one invocation of @code{allocate} or
5074: @code{resize} is internally contiguous. There is no contiguity between
5075: such a region and any other region (including others allocated from the
5076: heap).
5077:
1.32 anton 5078: doc-allocate
5079: doc-free
5080: doc-resize
5081:
1.27 crook 5082:
1.32 anton 5083: @node Memory Access, Address arithmetic, Heap Allocation, Memory
1.1 anton 5084: @subsection Memory Access
5085: @cindex memory access words
5086:
5087: doc-@
5088: doc-!
5089: doc-+!
5090: doc-c@
5091: doc-c!
5092: doc-2@
5093: doc-2!
5094: doc-f@
5095: doc-f!
5096: doc-sf@
5097: doc-sf!
5098: doc-df@
5099: doc-df!
5100:
1.68 anton 5101:
1.32 anton 5102: @node Address arithmetic, Memory Blocks, Memory Access, Memory
5103: @subsection Address arithmetic
1.1 anton 5104: @cindex address arithmetic words
5105:
1.67 anton 5106: Address arithmetic is the foundation on which you can build data
5107: structures like arrays, records (@pxref{Structures}) and objects
5108: (@pxref{Object-oriented Forth}).
1.32 anton 5109:
1.68 anton 5110: @cindex address unit
5111: @cindex au (address unit)
1.1 anton 5112: ANS Forth does not specify the sizes of the data types. Instead, it
5113: offers a number of words for computing sizes and doing address
1.29 crook 5114: arithmetic. Address arithmetic is performed in terms of address units
5115: (aus); on most systems the address unit is one byte. Note that a
5116: character may have more than one au, so @code{chars} is no noop (on
1.68 anton 5117: platforms where it is a noop, it compiles to nothing).
1.1 anton 5118:
1.67 anton 5119: The basic address arithmetic words are @code{+} and @code{-}. E.g., if
5120: you have the address of a cell, perform @code{1 cells +}, and you will
5121: have the address of the next cell.
5122:
1.68 anton 5123: @cindex contiguous regions and address arithmetic
1.67 anton 5124: In ANS Forth you can perform address arithmetic only within a contiguous
5125: region, i.e., if you have an address into one region, you can only add
5126: and subtract such that the result is still within the region; you can
5127: only subtract or compare addresses from within the same contiguous
5128: region. Reasons: several contiguous regions can be arranged in memory
5129: in any way; on segmented systems addresses may have unusual
5130: representations, such that address arithmetic only works within a
5131: region. Gforth provides a few more guarantees (linear address space,
5132: dictionary grows upwards), but in general I have found it easy to stay
5133: within contiguous regions (exception: computing and comparing to the
5134: address just beyond the end of an array).
5135:
1.1 anton 5136: @cindex alignment of addresses for types
5137: ANS Forth also defines words for aligning addresses for specific
5138: types. Many computers require that accesses to specific data types
5139: must only occur at specific addresses; e.g., that cells may only be
5140: accessed at addresses divisible by 4. Even if a machine allows unaligned
5141: accesses, it can usually perform aligned accesses faster.
5142:
5143: For the performance-conscious: alignment operations are usually only
5144: necessary during the definition of a data structure, not during the
5145: (more frequent) accesses to it.
5146:
5147: ANS Forth defines no words for character-aligning addresses. This is not
5148: an oversight, but reflects the fact that addresses that are not
5149: char-aligned have no use in the standard and therefore will not be
5150: created.
5151:
5152: @cindex @code{CREATE} and alignment
1.29 crook 5153: ANS Forth guarantees that addresses returned by @code{CREATE}d words
1.1 anton 5154: are cell-aligned; in addition, Gforth guarantees that these addresses
5155: are aligned for all purposes.
5156:
1.26 crook 5157: Note that the ANS Forth word @code{char} has nothing to do with address
5158: arithmetic.
1.1 anton 5159:
1.44 crook 5160:
1.1 anton 5161: doc-chars
5162: doc-char+
5163: doc-cells
5164: doc-cell+
5165: doc-cell
5166: doc-aligned
5167: doc-floats
5168: doc-float+
5169: doc-float
5170: doc-faligned
5171: doc-sfloats
5172: doc-sfloat+
5173: doc-sfaligned
5174: doc-dfloats
5175: doc-dfloat+
5176: doc-dfaligned
5177: doc-maxaligned
5178: doc-cfaligned
5179: doc-address-unit-bits
5180:
1.44 crook 5181:
1.32 anton 5182: @node Memory Blocks, , Address arithmetic, Memory
1.1 anton 5183: @subsection Memory Blocks
5184: @cindex memory block words
1.27 crook 5185: @cindex character strings - moving and copying
5186:
1.49 anton 5187: Memory blocks often represent character strings; For ways of storing
5188: character strings in memory see @ref{String Formats}. For other
5189: string-processing words see @ref{Displaying characters and strings}.
1.1 anton 5190:
1.67 anton 5191: A few of these words work on address unit blocks. In that case, you
5192: usually have to insert @code{CHARS} before the word when working on
5193: character strings. Most words work on character blocks, and expect a
5194: char-aligned address.
5195:
5196: When copying characters between overlapping memory regions, use
5197: @code{chars move} or choose carefully between @code{cmove} and
5198: @code{cmove>}.
1.44 crook 5199:
1.1 anton 5200: doc-move
5201: doc-erase
5202: doc-cmove
5203: doc-cmove>
5204: doc-fill
5205: doc-blank
1.21 crook 5206: doc-compare
5207: doc-search
1.27 crook 5208: doc--trailing
5209: doc-/string
5210:
1.44 crook 5211:
1.27 crook 5212: @comment TODO examples
5213:
1.1 anton 5214:
1.26 crook 5215: @node Control Structures, Defining Words, Memory, Words
1.1 anton 5216: @section Control Structures
5217: @cindex control structures
5218:
1.33 anton 5219: Control structures in Forth cannot be used interpretively, only in a
5220: colon definition@footnote{To be precise, they have no interpretation
5221: semantics (@pxref{Interpretation and Compilation Semantics}).}. We do
5222: not like this limitation, but have not seen a satisfying way around it
5223: yet, although many schemes have been proposed.
1.1 anton 5224:
5225: @menu
1.33 anton 5226: * Selection:: IF ... ELSE ... ENDIF
5227: * Simple Loops:: BEGIN ...
1.29 crook 5228: * Counted Loops:: DO
1.67 anton 5229: * Arbitrary control structures::
5230: * Calls and returns::
1.1 anton 5231: * Exception Handling::
5232: @end menu
5233:
5234: @node Selection, Simple Loops, Control Structures, Control Structures
5235: @subsection Selection
5236: @cindex selection control structures
5237: @cindex control structures for selection
5238:
5239: @cindex @code{IF} control structure
5240: @example
1.29 crook 5241: @i{flag}
1.1 anton 5242: IF
1.29 crook 5243: @i{code}
1.1 anton 5244: ENDIF
5245: @end example
1.21 crook 5246: @noindent
1.33 anton 5247:
1.44 crook 5248: If @i{flag} is non-zero (as far as @code{IF} etc. are concerned, a cell
5249: with any bit set represents truth) @i{code} is executed.
1.33 anton 5250:
1.1 anton 5251: @example
1.29 crook 5252: @i{flag}
1.1 anton 5253: IF
1.29 crook 5254: @i{code1}
1.1 anton 5255: ELSE
1.29 crook 5256: @i{code2}
1.1 anton 5257: ENDIF
5258: @end example
5259:
1.44 crook 5260: If @var{flag} is true, @i{code1} is executed, otherwise @i{code2} is
5261: executed.
1.33 anton 5262:
1.1 anton 5263: You can use @code{THEN} instead of @code{ENDIF}. Indeed, @code{THEN} is
5264: standard, and @code{ENDIF} is not, although it is quite popular. We
5265: recommend using @code{ENDIF}, because it is less confusing for people
5266: who also know other languages (and is not prone to reinforcing negative
5267: prejudices against Forth in these people). Adding @code{ENDIF} to a
5268: system that only supplies @code{THEN} is simple:
5269: @example
1.21 crook 5270: : ENDIF POSTPONE THEN ; immediate
1.1 anton 5271: @end example
5272:
5273: [According to @cite{Webster's New Encyclopedic Dictionary}, @dfn{then
5274: (adv.)} has the following meanings:
5275: @quotation
5276: ... 2b: following next after in order ... 3d: as a necessary consequence
5277: (if you were there, then you saw them).
5278: @end quotation
5279: Forth's @code{THEN} has the meaning 2b, whereas @code{THEN} in Pascal
5280: and many other programming languages has the meaning 3d.]
5281:
1.21 crook 5282: Gforth also provides the words @code{?DUP-IF} and @code{?DUP-0=-IF}, so
1.1 anton 5283: you can avoid using @code{?dup}. Using these alternatives is also more
1.26 crook 5284: efficient than using @code{?dup}. Definitions in ANS Forth
1.1 anton 5285: for @code{ENDIF}, @code{?DUP-IF} and @code{?DUP-0=-IF} are provided in
5286: @file{compat/control.fs}.
5287:
5288: @cindex @code{CASE} control structure
5289: @example
1.29 crook 5290: @i{n}
1.1 anton 5291: CASE
1.29 crook 5292: @i{n1} OF @i{code1} ENDOF
5293: @i{n2} OF @i{code2} ENDOF
1.1 anton 5294: @dots{}
1.68 anton 5295: ( n ) @i{default-code} ( n )
1.1 anton 5296: ENDCASE
5297: @end example
5298:
1.68 anton 5299: Executes the first @i{codei}, where the @i{ni} is equal to @i{n}. If no
5300: @i{ni} matches, the optional @i{default-code} is executed. The optional
5301: default case can be added by simply writing the code after the last
5302: @code{ENDOF}. It may use @i{n}, which is on top of the stack, but must
5303: not consume it.
1.1 anton 5304:
1.69 anton 5305: @progstyle
5306: To keep the code understandable, you should ensure that on all paths
5307: through a selection construct the stack is changed in the same way
5308: (wrt. number and types of stack items consumed and pushed).
5309:
1.1 anton 5310: @node Simple Loops, Counted Loops, Selection, Control Structures
5311: @subsection Simple Loops
5312: @cindex simple loops
5313: @cindex loops without count
5314:
5315: @cindex @code{WHILE} loop
5316: @example
5317: BEGIN
1.29 crook 5318: @i{code1}
5319: @i{flag}
1.1 anton 5320: WHILE
1.29 crook 5321: @i{code2}
1.1 anton 5322: REPEAT
5323: @end example
5324:
1.29 crook 5325: @i{code1} is executed and @i{flag} is computed. If it is true,
5326: @i{code2} is executed and the loop is restarted; If @i{flag} is
1.1 anton 5327: false, execution continues after the @code{REPEAT}.
5328:
5329: @cindex @code{UNTIL} loop
5330: @example
5331: BEGIN
1.29 crook 5332: @i{code}
5333: @i{flag}
1.1 anton 5334: UNTIL
5335: @end example
5336:
1.29 crook 5337: @i{code} is executed. The loop is restarted if @code{flag} is false.
1.1 anton 5338:
1.69 anton 5339: @progstyle
5340: To keep the code understandable, a complete iteration of the loop should
5341: not change the number and types of the items on the stacks.
5342:
1.1 anton 5343: @cindex endless loop
5344: @cindex loops, endless
5345: @example
5346: BEGIN
1.29 crook 5347: @i{code}
1.1 anton 5348: AGAIN
5349: @end example
5350:
5351: This is an endless loop.
5352:
5353: @node Counted Loops, Arbitrary control structures, Simple Loops, Control Structures
5354: @subsection Counted Loops
5355: @cindex counted loops
5356: @cindex loops, counted
5357: @cindex @code{DO} loops
5358:
5359: The basic counted loop is:
5360: @example
1.29 crook 5361: @i{limit} @i{start}
1.1 anton 5362: ?DO
1.29 crook 5363: @i{body}
1.1 anton 5364: LOOP
5365: @end example
5366:
1.29 crook 5367: This performs one iteration for every integer, starting from @i{start}
5368: and up to, but excluding @i{limit}. The counter, or @i{index}, can be
1.21 crook 5369: accessed with @code{i}. For example, the loop:
1.1 anton 5370: @example
5371: 10 0 ?DO
5372: i .
5373: LOOP
5374: @end example
1.21 crook 5375: @noindent
5376: prints @code{0 1 2 3 4 5 6 7 8 9}
5377:
1.1 anton 5378: The index of the innermost loop can be accessed with @code{i}, the index
5379: of the next loop with @code{j}, and the index of the third loop with
5380: @code{k}.
5381:
1.44 crook 5382:
1.1 anton 5383: doc-i
5384: doc-j
5385: doc-k
5386:
1.44 crook 5387:
1.1 anton 5388: The loop control data are kept on the return stack, so there are some
1.21 crook 5389: restrictions on mixing return stack accesses and counted loop words. In
5390: particuler, if you put values on the return stack outside the loop, you
5391: cannot read them inside the loop@footnote{well, not in a way that is
5392: portable.}. If you put values on the return stack within a loop, you
5393: have to remove them before the end of the loop and before accessing the
5394: index of the loop.
1.1 anton 5395:
5396: There are several variations on the counted loop:
5397:
1.21 crook 5398: @itemize @bullet
5399: @item
5400: @code{LEAVE} leaves the innermost counted loop immediately; execution
5401: continues after the associated @code{LOOP} or @code{NEXT}. For example:
5402:
5403: @example
5404: 10 0 ?DO i DUP . 3 = IF LEAVE THEN LOOP
5405: @end example
5406: prints @code{0 1 2 3}
5407:
1.1 anton 5408:
1.21 crook 5409: @item
5410: @code{UNLOOP} prepares for an abnormal loop exit, e.g., via
5411: @code{EXIT}. @code{UNLOOP} removes the loop control parameters from the
5412: return stack so @code{EXIT} can get to its return address. For example:
5413:
5414: @example
5415: : demo 10 0 ?DO i DUP . 3 = IF UNLOOP EXIT THEN LOOP ." Done" ;
5416: @end example
5417: prints @code{0 1 2 3}
5418:
5419:
5420: @item
1.29 crook 5421: If @i{start} is greater than @i{limit}, a @code{?DO} loop is entered
1.1 anton 5422: (and @code{LOOP} iterates until they become equal by wrap-around
5423: arithmetic). This behaviour is usually not what you want. Therefore,
5424: Gforth offers @code{+DO} and @code{U+DO} (as replacements for
1.29 crook 5425: @code{?DO}), which do not enter the loop if @i{start} is greater than
5426: @i{limit}; @code{+DO} is for signed loop parameters, @code{U+DO} for
1.1 anton 5427: unsigned loop parameters.
5428:
1.21 crook 5429: @item
5430: @code{?DO} can be replaced by @code{DO}. @code{DO} always enters
5431: the loop, independent of the loop parameters. Do not use @code{DO}, even
5432: if you know that the loop is entered in any case. Such knowledge tends
5433: to become invalid during maintenance of a program, and then the
5434: @code{DO} will make trouble.
5435:
5436: @item
1.29 crook 5437: @code{LOOP} can be replaced with @code{@i{n} +LOOP}; this updates the
5438: index by @i{n} instead of by 1. The loop is terminated when the border
5439: between @i{limit-1} and @i{limit} is crossed. E.g.:
1.1 anton 5440:
1.21 crook 5441: @example
5442: 4 0 +DO i . 2 +LOOP
5443: @end example
5444: @noindent
5445: prints @code{0 2}
5446:
5447: @example
5448: 4 1 +DO i . 2 +LOOP
5449: @end example
5450: @noindent
5451: prints @code{1 3}
1.1 anton 5452:
1.68 anton 5453: @item
1.1 anton 5454: @cindex negative increment for counted loops
5455: @cindex counted loops with negative increment
1.29 crook 5456: The behaviour of @code{@i{n} +LOOP} is peculiar when @i{n} is negative:
1.1 anton 5457:
1.21 crook 5458: @example
5459: -1 0 ?DO i . -1 +LOOP
5460: @end example
5461: @noindent
5462: prints @code{0 -1}
1.1 anton 5463:
1.21 crook 5464: @example
5465: 0 0 ?DO i . -1 +LOOP
5466: @end example
5467: prints nothing.
1.1 anton 5468:
1.29 crook 5469: Therefore we recommend avoiding @code{@i{n} +LOOP} with negative
5470: @i{n}. One alternative is @code{@i{u} -LOOP}, which reduces the
5471: index by @i{u} each iteration. The loop is terminated when the border
5472: between @i{limit+1} and @i{limit} is crossed. Gforth also provides
1.1 anton 5473: @code{-DO} and @code{U-DO} for down-counting loops. E.g.:
5474:
1.21 crook 5475: @example
5476: -2 0 -DO i . 1 -LOOP
5477: @end example
5478: @noindent
5479: prints @code{0 -1}
1.1 anton 5480:
1.21 crook 5481: @example
5482: -1 0 -DO i . 1 -LOOP
5483: @end example
5484: @noindent
5485: prints @code{0}
5486:
5487: @example
5488: 0 0 -DO i . 1 -LOOP
5489: @end example
5490: @noindent
5491: prints nothing.
1.1 anton 5492:
1.21 crook 5493: @end itemize
1.1 anton 5494:
5495: Unfortunately, @code{+DO}, @code{U+DO}, @code{-DO}, @code{U-DO} and
1.26 crook 5496: @code{-LOOP} are not defined in ANS Forth. However, an implementation
5497: for these words that uses only standard words is provided in
5498: @file{compat/loops.fs}.
1.1 anton 5499:
5500:
5501: @cindex @code{FOR} loops
1.26 crook 5502: Another counted loop is:
1.1 anton 5503: @example
1.29 crook 5504: @i{n}
1.1 anton 5505: FOR
1.29 crook 5506: @i{body}
1.1 anton 5507: NEXT
5508: @end example
5509: This is the preferred loop of native code compiler writers who are too
1.26 crook 5510: lazy to optimize @code{?DO} loops properly. This loop structure is not
1.29 crook 5511: defined in ANS Forth. In Gforth, this loop iterates @i{n+1} times;
5512: @code{i} produces values starting with @i{n} and ending with 0. Other
1.26 crook 5513: Forth systems may behave differently, even if they support @code{FOR}
5514: loops. To avoid problems, don't use @code{FOR} loops.
1.1 anton 5515:
5516: @node Arbitrary control structures, Calls and returns, Counted Loops, Control Structures
5517: @subsection Arbitrary control structures
5518: @cindex control structures, user-defined
5519:
5520: @cindex control-flow stack
5521: ANS Forth permits and supports using control structures in a non-nested
5522: way. Information about incomplete control structures is stored on the
5523: control-flow stack. This stack may be implemented on the Forth data
5524: stack, and this is what we have done in Gforth.
5525:
5526: @cindex @code{orig}, control-flow stack item
5527: @cindex @code{dest}, control-flow stack item
5528: An @i{orig} entry represents an unresolved forward branch, a @i{dest}
5529: entry represents a backward branch target. A few words are the basis for
5530: building any control structure possible (except control structures that
5531: need storage, like calls, coroutines, and backtracking).
5532:
1.44 crook 5533:
1.1 anton 5534: doc-if
5535: doc-ahead
5536: doc-then
5537: doc-begin
5538: doc-until
5539: doc-again
5540: doc-cs-pick
5541: doc-cs-roll
5542:
1.44 crook 5543:
1.21 crook 5544: The Standard words @code{CS-PICK} and @code{CS-ROLL} allow you to
5545: manipulate the control-flow stack in a portable way. Without them, you
5546: would need to know how many stack items are occupied by a control-flow
5547: entry (many systems use one cell. In Gforth they currently take three,
5548: but this may change in the future).
5549:
1.1 anton 5550: Some standard control structure words are built from these words:
5551:
1.44 crook 5552:
1.1 anton 5553: doc-else
5554: doc-while
5555: doc-repeat
5556:
1.44 crook 5557:
5558: @noindent
1.1 anton 5559: Gforth adds some more control-structure words:
5560:
1.44 crook 5561:
1.1 anton 5562: doc-endif
5563: doc-?dup-if
5564: doc-?dup-0=-if
5565:
1.44 crook 5566:
5567: @noindent
1.1 anton 5568: Counted loop words constitute a separate group of words:
5569:
1.44 crook 5570:
1.1 anton 5571: doc-?do
5572: doc-+do
5573: doc-u+do
5574: doc--do
5575: doc-u-do
5576: doc-do
5577: doc-for
5578: doc-loop
5579: doc-+loop
5580: doc--loop
5581: doc-next
5582: doc-leave
5583: doc-?leave
5584: doc-unloop
5585: doc-done
5586:
1.44 crook 5587:
1.21 crook 5588: The standard does not allow using @code{CS-PICK} and @code{CS-ROLL} on
5589: @i{do-sys}. Gforth allows it, but it's your job to ensure that for
1.1 anton 5590: every @code{?DO} etc. there is exactly one @code{UNLOOP} on any path
5591: through the definition (@code{LOOP} etc. compile an @code{UNLOOP} on the
5592: fall-through path). Also, you have to ensure that all @code{LEAVE}s are
5593: resolved (by using one of the loop-ending words or @code{DONE}).
5594:
1.44 crook 5595: @noindent
1.26 crook 5596: Another group of control structure words are:
1.1 anton 5597:
1.44 crook 5598:
1.1 anton 5599: doc-case
5600: doc-endcase
5601: doc-of
5602: doc-endof
5603:
1.44 crook 5604:
1.21 crook 5605: @i{case-sys} and @i{of-sys} cannot be processed using @code{CS-PICK} and
5606: @code{CS-ROLL}.
1.1 anton 5607:
5608: @subsubsection Programming Style
1.47 crook 5609: @cindex control structures programming style
5610: @cindex programming style, arbitrary control structures
1.1 anton 5611:
5612: In order to ensure readability we recommend that you do not create
5613: arbitrary control structures directly, but define new control structure
5614: words for the control structure you want and use these words in your
1.26 crook 5615: program. For example, instead of writing:
1.1 anton 5616:
5617: @example
1.26 crook 5618: BEGIN
1.1 anton 5619: ...
1.26 crook 5620: IF [ 1 CS-ROLL ]
1.1 anton 5621: ...
1.26 crook 5622: AGAIN THEN
1.1 anton 5623: @end example
5624:
1.21 crook 5625: @noindent
1.1 anton 5626: we recommend defining control structure words, e.g.,
5627:
5628: @example
1.26 crook 5629: : WHILE ( DEST -- ORIG DEST )
5630: POSTPONE IF
5631: 1 CS-ROLL ; immediate
5632:
5633: : REPEAT ( orig dest -- )
5634: POSTPONE AGAIN
5635: POSTPONE THEN ; immediate
1.1 anton 5636: @end example
5637:
1.21 crook 5638: @noindent
1.1 anton 5639: and then using these to create the control structure:
5640:
5641: @example
1.26 crook 5642: BEGIN
1.1 anton 5643: ...
1.26 crook 5644: WHILE
1.1 anton 5645: ...
1.26 crook 5646: REPEAT
1.1 anton 5647: @end example
5648:
5649: That's much easier to read, isn't it? Of course, @code{REPEAT} and
5650: @code{WHILE} are predefined, so in this example it would not be
5651: necessary to define them.
5652:
5653: @node Calls and returns, Exception Handling, Arbitrary control structures, Control Structures
5654: @subsection Calls and returns
5655: @cindex calling a definition
5656: @cindex returning from a definition
5657:
1.3 anton 5658: @cindex recursive definitions
5659: A definition can be called simply be writing the name of the definition
1.26 crook 5660: to be called. Normally a definition is invisible during its own
1.3 anton 5661: definition. If you want to write a directly recursive definition, you
1.26 crook 5662: can use @code{recursive} to make the current definition visible, or
5663: @code{recurse} to call the current definition directly.
1.3 anton 5664:
1.44 crook 5665:
1.3 anton 5666: doc-recursive
5667: doc-recurse
5668:
1.44 crook 5669:
1.21 crook 5670: @comment TODO add example of the two recursion methods
1.12 anton 5671: @quotation
5672: @progstyle
5673: I prefer using @code{recursive} to @code{recurse}, because calling the
5674: definition by name is more descriptive (if the name is well-chosen) than
5675: the somewhat cryptic @code{recurse}. E.g., in a quicksort
5676: implementation, it is much better to read (and think) ``now sort the
5677: partitions'' than to read ``now do a recursive call''.
5678: @end quotation
1.3 anton 5679:
1.29 crook 5680: For mutual recursion, use @code{Defer}red words, like this:
1.3 anton 5681:
5682: @example
1.28 crook 5683: Defer foo
1.3 anton 5684:
5685: : bar ( ... -- ... )
5686: ... foo ... ;
5687:
5688: :noname ( ... -- ... )
5689: ... bar ... ;
5690: IS foo
5691: @end example
5692:
1.44 crook 5693: Deferred words are discussed in more detail in @ref{Deferred words}.
1.33 anton 5694:
1.26 crook 5695: The current definition returns control to the calling definition when
1.33 anton 5696: the end of the definition is reached or @code{EXIT} is encountered.
1.1 anton 5697:
5698: doc-exit
5699: doc-;s
5700:
1.44 crook 5701:
1.1 anton 5702: @node Exception Handling, , Calls and returns, Control Structures
5703: @subsection Exception Handling
1.26 crook 5704: @cindex exceptions
1.1 anton 5705:
1.68 anton 5706: @c quit is a very bad idea for error handling,
5707: @c because it does not translate into a THROW
5708: @c it also does not belong into this chapter
5709:
5710: If a word detects an error condition that it cannot handle, it can
5711: @code{throw} an exception. In the simplest case, this will terminate
5712: your program, and report an appropriate error.
1.21 crook 5713:
1.68 anton 5714: doc-throw
1.1 anton 5715:
1.69 anton 5716: @code{Throw} consumes a cell-sized error number on the stack. There are
5717: some predefined error numbers in ANS Forth (see @file{errors.fs}). In
5718: Gforth (and most other systems) you can use the iors produced by various
5719: words as error numbers (e.g., a typical use of @code{allocate} is
5720: @code{allocate throw}). Gforth also provides the word @code{exception}
5721: to define your own error numbers (with decent error reporting); an ANS
5722: Forth version of this word (but without the error messages) is available
5723: in @code{compat/except.fs}. And finally, you can use your own error
1.68 anton 5724: numbers (anything outside the range -4095..0), but won't get nice error
5725: messages, only numbers. For example, try:
5726:
5727: @example
1.69 anton 5728: -10 throw \ ANS defined
5729: -267 throw \ system defined
5730: s" my error" exception throw \ user defined
5731: 7 throw \ arbitrary number
1.68 anton 5732: @end example
5733:
5734: doc---exception-exception
1.1 anton 5735:
1.69 anton 5736: A common idiom to @code{THROW} a specific error if a flag is true is
5737: this:
5738:
5739: @example
5740: @code{( flag ) 0<> @i{errno} and throw}
5741: @end example
5742:
5743: Your program can provide exception handlers to catch exceptions. An
5744: exception handler can be used to correct the problem, or to clean up
5745: some data structures and just throw the exception to the next exception
5746: handler. Note that @code{throw} jumps to the dynamically innermost
5747: exception handler. The system's exception handler is outermost, and just
5748: prints an error and restarts command-line interpretation (or, in batch
5749: mode (i.e., while processing the shell command line), leaves Gforth).
1.1 anton 5750:
1.68 anton 5751: The ANS Forth way to catch exceptions is @code{catch}:
1.1 anton 5752:
1.68 anton 5753: doc-catch
5754:
5755: The most common use of exception handlers is to clean up the state when
5756: an error happens. E.g.,
1.1 anton 5757:
1.26 crook 5758: @example
1.68 anton 5759: base @ >r hex \ actually the hex should be inside foo, or we h
5760: ['] foo catch ( nerror|0 )
5761: r> base !
1.69 anton 5762: ( nerror|0 ) throw \ pass it on
1.26 crook 5763: @end example
1.1 anton 5764:
1.69 anton 5765: A use of @code{catch} for handling the error @code{myerror} might look
5766: like this:
1.44 crook 5767:
1.68 anton 5768: @example
1.69 anton 5769: ['] foo catch
5770: CASE
5771: myerror OF ... ( do something about it ) ENDOF
5772: dup throw \ default: pass other errors on, do nothing on non-errors
5773: ENDCASE
1.68 anton 5774: @end example
1.44 crook 5775:
1.68 anton 5776: Having to wrap the code into a separate word is often cumbersome,
5777: therefore Gforth provides an alternative syntax:
1.1 anton 5778:
5779: @example
1.69 anton 5780: TRY
1.68 anton 5781: @i{code1}
1.69 anton 5782: RECOVER \ optional
1.68 anton 5783: @i{code2} \ optional
1.69 anton 5784: ENDTRY
1.1 anton 5785: @end example
5786:
1.68 anton 5787: This performs @i{Code1}. If @i{code1} completes normally, execution
5788: continues after the @code{endtry}. If @i{Code1} throws, the stacks are
5789: reset to the state during @code{try}, the throw value is pushed on the
5790: data stack, and execution constinues at @i{code2}, and finally falls
5791: through the @code{endtry} into the following code. If there is no
5792: @code{recover} clause, this works like an empty recover clause.
1.26 crook 5793:
1.68 anton 5794: doc-try
5795: doc-recover
5796: doc-endtry
1.26 crook 5797:
1.69 anton 5798: The cleanup example from above in this syntax:
1.26 crook 5799:
1.68 anton 5800: @example
1.69 anton 5801: base @ >r TRY
1.68 anton 5802: hex foo \ now the hex is placed correctly
1.69 anton 5803: 0 \ value for throw
5804: ENDTRY
1.68 anton 5805: r> base ! throw
1.1 anton 5806: @end example
5807:
1.69 anton 5808: And here's the error handling example:
1.1 anton 5809:
1.68 anton 5810: @example
1.69 anton 5811: TRY
1.68 anton 5812: foo
1.69 anton 5813: RECOVER
5814: CASE
5815: myerror OF ... ( do something about it ) ENDOF
5816: throw \ pass other errors on
5817: ENDCASE
5818: ENDTRY
1.68 anton 5819: @end example
1.1 anton 5820:
1.69 anton 5821: @progstyle
5822: As usual, you should ensure that the stack depth is statically known at
5823: the end: either after the @code{throw} for passing on errors, or after
5824: the @code{ENDTRY} (or, if you use @code{catch}, after the end of the
5825: selection construct for handling the error).
5826:
1.68 anton 5827: There are two alternatives to @code{throw}: @code{Abort"} is conditional
5828: and you can provide an error message. @code{Abort} just produces an
5829: ``Aborted'' error.
1.1 anton 5830:
1.68 anton 5831: The problem with these words is that exception handlers cannot
5832: differentiate between different @code{abort"}s; they just look like
5833: @code{-2 throw} to them (the error message cannot be accessed by
5834: standard programs). Similar @code{abort} looks like @code{-1 throw} to
5835: exception handlers.
1.44 crook 5836:
1.68 anton 5837: doc-abort"
1.26 crook 5838: doc-abort
1.29 crook 5839:
5840:
1.44 crook 5841:
1.29 crook 5842: @c -------------------------------------------------------------
1.47 crook 5843: @node Defining Words, Interpretation and Compilation Semantics, Control Structures, Words
1.29 crook 5844: @section Defining Words
5845: @cindex defining words
5846:
1.47 crook 5847: Defining words are used to extend Forth by creating new entries in the dictionary.
5848:
1.29 crook 5849: @menu
1.67 anton 5850: * CREATE::
1.44 crook 5851: * Variables:: Variables and user variables
1.67 anton 5852: * Constants::
1.44 crook 5853: * Values:: Initialised variables
1.67 anton 5854: * Colon Definitions::
1.44 crook 5855: * Anonymous Definitions:: Definitions without names
1.69 anton 5856: * Supplying names:: Passing definition names as strings
1.67 anton 5857: * User-defined Defining Words::
1.44 crook 5858: * Deferred words:: Allow forward references
1.67 anton 5859: * Aliases::
1.29 crook 5860: @end menu
5861:
1.44 crook 5862: @node CREATE, Variables, Defining Words, Defining Words
5863: @subsection @code{CREATE}
1.29 crook 5864: @cindex simple defining words
5865: @cindex defining words, simple
5866:
5867: Defining words are used to create new entries in the dictionary. The
5868: simplest defining word is @code{CREATE}. @code{CREATE} is used like
5869: this:
5870:
5871: @example
5872: CREATE new-word1
5873: @end example
5874:
1.69 anton 5875: @code{CREATE} is a parsing word, i.e., it takes an argument from the
5876: input stream (@code{new-word1} in our example). It generates a
5877: dictionary entry for @code{new-word1}. When @code{new-word1} is
5878: executed, all that it does is leave an address on the stack. The address
5879: represents the value of the data space pointer (@code{HERE}) at the time
5880: that @code{new-word1} was defined. Therefore, @code{CREATE} is a way of
5881: associating a name with the address of a region of memory.
1.29 crook 5882:
1.34 anton 5883: doc-create
5884:
1.69 anton 5885: Note that in ANS Forth guarantees only for @code{create} that its body
5886: is in dictionary data space (i.e., where @code{here}, @code{allot}
5887: etc. work, @pxref{Dictionary allocation}). Also, in ANS Forth only
5888: @code{create}d words can be modified with @code{does>}
5889: (@pxref{User-defined Defining Words}). And in ANS Forth @code{>body}
5890: can only be applied to @code{create}d words.
5891:
1.29 crook 5892: By extending this example to reserve some memory in data space, we end
1.69 anton 5893: up with something like a @i{variable}. Here are two different ways to do
5894: it:
1.29 crook 5895:
5896: @example
5897: CREATE new-word2 1 cells allot \ reserve 1 cell - initial value undefined
5898: CREATE new-word3 4 , \ reserve 1 cell and initialise it (to 4)
5899: @end example
5900:
5901: The variable can be examined and modified using @code{@@} (``fetch'') and
5902: @code{!} (``store'') like this:
5903:
5904: @example
5905: new-word2 @@ . \ get address, fetch from it and display
5906: 1234 new-word2 ! \ new value, get address, store to it
5907: @end example
5908:
1.44 crook 5909: @cindex arrays
5910: A similar mechanism can be used to create arrays. For example, an
5911: 80-character text input buffer:
1.29 crook 5912:
5913: @example
1.44 crook 5914: CREATE text-buf 80 chars allot
5915:
5916: text-buf 0 chars c@@ \ the 1st character (offset 0)
5917: text-buf 3 chars c@@ \ the 4th character (offset 3)
5918: @end example
1.29 crook 5919:
1.44 crook 5920: You can build arbitrarily complex data structures by allocating
1.49 anton 5921: appropriate areas of memory. For further discussions of this, and to
1.66 anton 5922: learn about some Gforth tools that make it easier,
1.49 anton 5923: @xref{Structures}.
1.44 crook 5924:
5925:
5926: @node Variables, Constants, CREATE, Defining Words
5927: @subsection Variables
5928: @cindex variables
5929:
5930: The previous section showed how a sequence of commands could be used to
5931: generate a variable. As a final refinement, the whole code sequence can
5932: be wrapped up in a defining word (pre-empting the subject of the next
5933: section), making it easier to create new variables:
5934:
5935: @example
5936: : myvariableX ( "name" -- a-addr ) CREATE 1 cells allot ;
5937: : myvariable0 ( "name" -- a-addr ) CREATE 0 , ;
5938:
5939: myvariableX foo \ variable foo starts off with an unknown value
5940: myvariable0 joe \ whilst joe is initialised to 0
1.29 crook 5941:
5942: 45 3 * foo ! \ set foo to 135
5943: 1234 joe ! \ set joe to 1234
5944: 3 joe +! \ increment joe by 3.. to 1237
5945: @end example
5946:
5947: Not surprisingly, there is no need to define @code{myvariable}, since
1.44 crook 5948: Forth already has a definition @code{Variable}. ANS Forth does not
1.69 anton 5949: guarantee that a @code{Variable} is initialised when it is created
5950: (i.e., it may behave like @code{myvariableX}). In contrast, Gforth's
5951: @code{Variable} initialises the variable to 0 (i.e., it behaves exactly
5952: like @code{myvariable0}). Forth also provides @code{2Variable} and
1.47 crook 5953: @code{fvariable} for double and floating-point variables, respectively
1.69 anton 5954: -- they are initialised to 0. and 0e in Gforth. If you use a @code{Variable} to
1.47 crook 5955: store a boolean, you can use @code{on} and @code{off} to toggle its
5956: state.
1.29 crook 5957:
1.34 anton 5958: doc-variable
5959: doc-2variable
5960: doc-fvariable
5961:
1.29 crook 5962: @cindex user variables
5963: @cindex user space
5964: The defining word @code{User} behaves in the same way as @code{Variable}.
5965: The difference is that it reserves space in @i{user (data) space} rather
5966: than normal data space. In a Forth system that has a multi-tasker, each
5967: task has its own set of user variables.
5968:
1.34 anton 5969: doc-user
1.67 anton 5970: @c doc-udp
5971: @c doc-uallot
1.34 anton 5972:
1.29 crook 5973: @comment TODO is that stuff about user variables strictly correct? Is it
5974: @comment just terminal tasks that have user variables?
5975: @comment should document tasker.fs (with some examples) elsewhere
5976: @comment in this manual, then expand on user space and user variables.
5977:
1.44 crook 5978: @node Constants, Values, Variables, Defining Words
5979: @subsection Constants
5980: @cindex constants
5981:
5982: @code{Constant} allows you to declare a fixed value and refer to it by
5983: name. For example:
1.29 crook 5984:
5985: @example
5986: 12 Constant INCHES-PER-FOOT
5987: 3E+08 fconstant SPEED-O-LIGHT
5988: @end example
5989:
5990: A @code{Variable} can be both read and written, so its run-time
5991: behaviour is to supply an address through which its current value can be
5992: manipulated. In contrast, the value of a @code{Constant} cannot be
5993: changed once it has been declared@footnote{Well, often it can be -- but
5994: not in a Standard, portable way. It's safer to use a @code{Value} (read
5995: on).} so it's not necessary to supply the address -- it is more
5996: efficient to return the value of the constant directly. That's exactly
5997: what happens; the run-time effect of a constant is to put its value on
1.49 anton 5998: the top of the stack (You can find one
5999: way of implementing @code{Constant} in @ref{User-defined Defining Words}).
1.29 crook 6000:
1.69 anton 6001: Forth also provides @code{2Constant} and @code{fconstant} for defining
1.29 crook 6002: double and floating-point constants, respectively.
6003:
1.34 anton 6004: doc-constant
6005: doc-2constant
6006: doc-fconstant
6007:
6008: @c that's too deep, and it's not necessarily true for all ANS Forths. - anton
1.44 crook 6009: @c nac-> How could that not be true in an ANS Forth? You can't define a
6010: @c constant, use it and then delete the definition of the constant..
1.69 anton 6011:
6012: @c anton->An ANS Forth system can compile a constant to a literal; On
6013: @c decompilation you would see only the number, just as if it had been used
6014: @c in the first place. The word will stay, of course, but it will only be
6015: @c used by the text interpreter (no run-time duties, except when it is
6016: @c POSTPONEd or somesuch).
6017:
6018: @c nac:
1.44 crook 6019: @c I agree that it's rather deep, but IMO it is an important difference
6020: @c relative to other programming languages.. often it's annoying: it
6021: @c certainly changes my programming style relative to C.
6022:
1.69 anton 6023: @c anton: In what way?
6024:
1.29 crook 6025: Constants in Forth behave differently from their equivalents in other
6026: programming languages. In other languages, a constant (such as an EQU in
6027: assembler or a #define in C) only exists at compile-time; in the
6028: executable program the constant has been translated into an absolute
6029: number and, unless you are using a symbolic debugger, it's impossible to
6030: know what abstract thing that number represents. In Forth a constant has
1.44 crook 6031: an entry in the header space and remains there after the code that uses
6032: it has been defined. In fact, it must remain in the dictionary since it
6033: has run-time duties to perform. For example:
1.29 crook 6034:
6035: @example
6036: 12 Constant INCHES-PER-FOOT
6037: : FEET-TO-INCHES ( n1 -- n2 ) INCHES-PER-FOOT * ;
6038: @end example
6039:
6040: @cindex in-lining of constants
6041: When @code{FEET-TO-INCHES} is executed, it will in turn execute the xt
6042: associated with the constant @code{INCHES-PER-FOOT}. If you use
6043: @code{see} to decompile the definition of @code{FEET-TO-INCHES}, you can
6044: see that it makes a call to @code{INCHES-PER-FOOT}. Some Forth compilers
6045: attempt to optimise constants by in-lining them where they are used. You
6046: can force Gforth to in-line a constant like this:
6047:
6048: @example
6049: : FEET-TO-INCHES ( n1 -- n2 ) [ INCHES-PER-FOOT ] LITERAL * ;
6050: @end example
6051:
6052: If you use @code{see} to decompile @i{this} version of
6053: @code{FEET-TO-INCHES}, you can see that @code{INCHES-PER-FOOT} is no
1.49 anton 6054: longer present. To understand how this works, read
6055: @ref{Interpret/Compile states}, and @ref{Literals}.
1.29 crook 6056:
6057: In-lining constants in this way might improve execution time
6058: fractionally, and can ensure that a constant is now only referenced at
6059: compile-time. However, the definition of the constant still remains in
6060: the dictionary. Some Forth compilers provide a mechanism for controlling
6061: a second dictionary for holding transient words such that this second
6062: dictionary can be deleted later in order to recover memory
6063: space. However, there is no standard way of doing this.
6064:
6065:
1.44 crook 6066: @node Values, Colon Definitions, Constants, Defining Words
6067: @subsection Values
6068: @cindex values
1.34 anton 6069:
1.69 anton 6070: A @code{Value} behaves like a @code{Constant}, but it can be changed.
6071: @code{TO} is a parsing word that changes a @code{Values}. In Gforth
6072: (not in ANS Forth) you can access (and change) a @code{value} also with
6073: @code{>body}.
6074:
6075: Here are some
6076: examples:
1.29 crook 6077:
6078: @example
1.69 anton 6079: 12 Value APPLES \ Define APPLES with an initial value of 12
6080: 34 TO APPLES \ Change the value of APPLES. TO is a parsing word
6081: 1 ' APPLES >body +! \ Increment APPLES. Non-standard usage.
6082: APPLES \ puts 35 on the top of the stack.
1.29 crook 6083: @end example
6084:
1.44 crook 6085: doc-value
6086: doc-to
1.29 crook 6087:
1.35 anton 6088:
1.69 anton 6089:
1.44 crook 6090: @node Colon Definitions, Anonymous Definitions, Values, Defining Words
6091: @subsection Colon Definitions
6092: @cindex colon definitions
1.35 anton 6093:
6094: @example
1.44 crook 6095: : name ( ... -- ... )
6096: word1 word2 word3 ;
1.29 crook 6097: @end example
6098:
1.44 crook 6099: @noindent
6100: Creates a word called @code{name} that, upon execution, executes
6101: @code{word1 word2 word3}. @code{name} is a @dfn{(colon) definition}.
1.29 crook 6102:
1.49 anton 6103: The explanation above is somewhat superficial. For simple examples of
6104: colon definitions see @ref{Your first definition}. For an in-depth
1.66 anton 6105: discussion of some of the issues involved, @xref{Interpretation and
1.49 anton 6106: Compilation Semantics}.
1.29 crook 6107:
1.44 crook 6108: doc-:
6109: doc-;
1.1 anton 6110:
1.34 anton 6111:
1.69 anton 6112: @node Anonymous Definitions, Supplying names, Colon Definitions, Defining Words
1.44 crook 6113: @subsection Anonymous Definitions
6114: @cindex colon definitions
6115: @cindex defining words without name
1.34 anton 6116:
1.44 crook 6117: Sometimes you want to define an @dfn{anonymous word}; a word without a
6118: name. You can do this with:
1.1 anton 6119:
1.44 crook 6120: doc-:noname
1.1 anton 6121:
1.44 crook 6122: This leaves the execution token for the word on the stack after the
6123: closing @code{;}. Here's an example in which a deferred word is
6124: initialised with an @code{xt} from an anonymous colon definition:
1.1 anton 6125:
1.29 crook 6126: @example
1.44 crook 6127: Defer deferred
6128: :noname ( ... -- ... )
6129: ... ;
6130: IS deferred
1.29 crook 6131: @end example
1.26 crook 6132:
1.44 crook 6133: @noindent
6134: Gforth provides an alternative way of doing this, using two separate
6135: words:
1.27 crook 6136:
1.44 crook 6137: doc-noname
6138: @cindex execution token of last defined word
6139: doc-lastxt
1.1 anton 6140:
1.44 crook 6141: @noindent
6142: The previous example can be rewritten using @code{noname} and
6143: @code{lastxt}:
1.1 anton 6144:
1.26 crook 6145: @example
1.44 crook 6146: Defer deferred
6147: noname : ( ... -- ... )
6148: ... ;
6149: lastxt IS deferred
1.26 crook 6150: @end example
1.1 anton 6151:
1.29 crook 6152: @noindent
1.44 crook 6153: @code{noname} works with any defining word, not just @code{:}.
6154:
6155: @code{lastxt} also works when the last word was not defined as
1.71 anton 6156: @code{noname}. It does not work for combined words, though. It also has
6157: the useful property that is is valid as soon as the header for a
6158: definition has been built. Thus:
1.44 crook 6159:
6160: @example
6161: lastxt . : foo [ lastxt . ] ; ' foo .
6162: @end example
1.1 anton 6163:
1.44 crook 6164: @noindent
6165: prints 3 numbers; the last two are the same.
1.26 crook 6166:
1.69 anton 6167: @node Supplying names, User-defined Defining Words, Anonymous Definitions, Defining Words
6168: @subsection Supplying the name of a defined word
6169: @cindex names for defined words
6170: @cindex defining words, name given in a string
6171:
6172: By default, a defining word takes the name for the defined word from the
6173: input stream. Sometimes you want to supply the name from a string. You
6174: can do this with:
6175:
6176: doc-nextname
6177:
6178: For example:
6179:
6180: @example
6181: s" foo" nextname create
6182: @end example
6183:
6184: @noindent
6185: is equivalent to:
6186:
6187: @example
6188: create foo
6189: @end example
6190:
6191: @noindent
6192: @code{nextname} works with any defining word.
6193:
1.1 anton 6194:
1.69 anton 6195: @node User-defined Defining Words, Deferred words, Supplying names, Defining Words
1.26 crook 6196: @subsection User-defined Defining Words
6197: @cindex user-defined defining words
6198: @cindex defining words, user-defined
1.1 anton 6199:
1.29 crook 6200: You can create a new defining word by wrapping defining-time code around
6201: an existing defining word and putting the sequence in a colon
1.69 anton 6202: definition.
6203:
6204: @c anton: This example is very complex and leads in a quite different
6205: @c direction from the CREATE-DOES> stuff that follows. It should probably
6206: @c be done elsewhere, or as a subsubsection of this subsection (or as a
6207: @c subsection of Defining Words)
6208:
6209: For example, suppose that you have a word @code{stats} that
1.29 crook 6210: gathers statistics about colon definitions given the @i{xt} of the
6211: definition, and you want every colon definition in your application to
6212: make a call to @code{stats}. You can define and use a new version of
6213: @code{:} like this:
6214:
6215: @example
6216: : stats ( xt -- ) DUP ." (Gathering statistics for " . ." )"
6217: ... ; \ other code
6218:
6219: : my: : lastxt postpone literal ['] stats compile, ;
6220:
6221: my: foo + - ;
6222: @end example
6223:
6224: When @code{foo} is defined using @code{my:} these steps occur:
6225:
6226: @itemize @bullet
6227: @item
6228: @code{my:} is executed.
6229: @item
6230: The @code{:} within the definition (the one between @code{my:} and
6231: @code{lastxt}) is executed, and does just what it always does; it parses
6232: the input stream for a name, builds a dictionary header for the name
6233: @code{foo} and switches @code{state} from interpret to compile.
6234: @item
6235: The word @code{lastxt} is executed. It puts the @i{xt} for the word that is
6236: being defined -- @code{foo} -- onto the stack.
6237: @item
6238: The code that was produced by @code{postpone literal} is executed; this
6239: causes the value on the stack to be compiled as a literal in the code
6240: area of @code{foo}.
6241: @item
6242: The code @code{['] stats} compiles a literal into the definition of
6243: @code{my:}. When @code{compile,} is executed, that literal -- the
6244: execution token for @code{stats} -- is layed down in the code area of
6245: @code{foo} , following the literal@footnote{Strictly speaking, the
6246: mechanism that @code{compile,} uses to convert an @i{xt} into something
6247: in the code area is implementation-dependent. A threaded implementation
6248: might spit out the execution token directly whilst another
6249: implementation might spit out a native code sequence.}.
6250: @item
6251: At this point, the execution of @code{my:} is complete, and control
6252: returns to the text interpreter. The text interpreter is in compile
6253: state, so subsequent text @code{+ -} is compiled into the definition of
6254: @code{foo} and the @code{;} terminates the definition as always.
6255: @end itemize
6256:
6257: You can use @code{see} to decompile a word that was defined using
6258: @code{my:} and see how it is different from a normal @code{:}
6259: definition. For example:
6260:
6261: @example
6262: : bar + - ; \ like foo but using : rather than my:
6263: see bar
6264: : bar
6265: + - ;
6266: see foo
6267: : foo
6268: 107645672 stats + - ;
6269:
6270: \ use ' stats . to show that 107645672 is the xt for stats
6271: @end example
6272:
6273: You can use techniques like this to make new defining words in terms of
6274: @i{any} existing defining word.
1.1 anton 6275:
6276:
1.29 crook 6277: @cindex defining defining words
1.26 crook 6278: @cindex @code{CREATE} ... @code{DOES>}
6279: If you want the words defined with your defining words to behave
6280: differently from words defined with standard defining words, you can
6281: write your defining word like this:
1.1 anton 6282:
6283: @example
1.26 crook 6284: : def-word ( "name" -- )
1.29 crook 6285: CREATE @i{code1}
1.26 crook 6286: DOES> ( ... -- ... )
1.29 crook 6287: @i{code2} ;
1.26 crook 6288:
6289: def-word name
1.1 anton 6290: @end example
6291:
1.29 crook 6292: @cindex child words
6293: This fragment defines a @dfn{defining word} @code{def-word} and then
6294: executes it. When @code{def-word} executes, it @code{CREATE}s a new
6295: word, @code{name}, and executes the code @i{code1}. The code @i{code2}
6296: is not executed at this time. The word @code{name} is sometimes called a
6297: @dfn{child} of @code{def-word}.
6298:
6299: When you execute @code{name}, the address of the body of @code{name} is
6300: put on the data stack and @i{code2} is executed (the address of the body
6301: of @code{name} is the address @code{HERE} returns immediately after the
1.69 anton 6302: @code{CREATE}, i.e., the address a @code{create}d word returns by
6303: default).
6304:
6305: @c anton:
6306: @c www.dictionary.com says:
6307: @c at·a·vism: 1.The reappearance of a characteristic in an organism after
6308: @c several generations of absence, usually caused by the chance
6309: @c recombination of genes. 2.An individual or a part that exhibits
6310: @c atavism. Also called throwback. 3.The return of a trait or recurrence
6311: @c of previous behavior after a period of absence.
6312: @c
6313: @c Doesn't seem to fit.
1.29 crook 6314:
1.69 anton 6315: @c @cindex atavism in child words
1.33 anton 6316: You can use @code{def-word} to define a set of child words that behave
1.69 anton 6317: similarly; they all have a common run-time behaviour determined by
6318: @i{code2}. Typically, the @i{code1} sequence builds a data area in the
6319: body of the child word. The structure of the data is common to all
6320: children of @code{def-word}, but the data values are specific -- and
6321: private -- to each child word. When a child word is executed, the
6322: address of its private data area is passed as a parameter on TOS to be
6323: used and manipulated@footnote{It is legitimate both to read and write to
6324: this data area.} by @i{code2}.
1.29 crook 6325:
6326: The two fragments of code that make up the defining words act (are
6327: executed) at two completely separate times:
1.1 anton 6328:
1.29 crook 6329: @itemize @bullet
6330: @item
6331: At @i{define time}, the defining word executes @i{code1} to generate a
6332: child word
6333: @item
6334: At @i{child execution time}, when a child word is invoked, @i{code2}
6335: is executed, using parameters (data) that are private and specific to
6336: the child word.
6337: @end itemize
6338:
1.44 crook 6339: Another way of understanding the behaviour of @code{def-word} and
6340: @code{name} is to say that, if you make the following definitions:
1.33 anton 6341: @example
6342: : def-word1 ( "name" -- )
6343: CREATE @i{code1} ;
6344:
6345: : action1 ( ... -- ... )
6346: @i{code2} ;
6347:
6348: def-word1 name1
6349: @end example
6350:
1.44 crook 6351: @noindent
6352: Then using @code{name1 action1} is equivalent to using @code{name}.
1.1 anton 6353:
1.29 crook 6354: The classic example is that you can define @code{CONSTANT} in this way:
1.26 crook 6355:
1.1 anton 6356: @example
1.29 crook 6357: : CONSTANT ( w "name" -- )
6358: CREATE ,
1.26 crook 6359: DOES> ( -- w )
6360: @@ ;
1.1 anton 6361: @end example
6362:
1.29 crook 6363: @comment There is a beautiful description of how this works and what
6364: @comment it does in the Forthwrite 100th edition.. as well as an elegant
6365: @comment commentary on the Counting Fruits problem.
6366:
6367: When you create a constant with @code{5 CONSTANT five}, a set of
6368: define-time actions take place; first a new word @code{five} is created,
6369: then the value 5 is laid down in the body of @code{five} with
1.44 crook 6370: @code{,}. When @code{five} is executed, the address of the body is put on
1.29 crook 6371: the stack, and @code{@@} retrieves the value 5. The word @code{five} has
6372: no code of its own; it simply contains a data field and a pointer to the
6373: code that follows @code{DOES>} in its defining word. That makes words
6374: created in this way very compact.
6375:
6376: The final example in this section is intended to remind you that space
6377: reserved in @code{CREATE}d words is @i{data} space and therefore can be
6378: both read and written by a Standard program@footnote{Exercise: use this
6379: example as a starting point for your own implementation of @code{Value}
6380: and @code{TO} -- if you get stuck, investigate the behaviour of @code{'} and
6381: @code{[']}.}:
6382:
6383: @example
6384: : foo ( "name" -- )
6385: CREATE -1 ,
6386: DOES> ( -- )
1.33 anton 6387: @@ . ;
1.29 crook 6388:
6389: foo first-word
6390: foo second-word
6391:
6392: 123 ' first-word >BODY !
6393: @end example
6394:
6395: If @code{first-word} had been a @code{CREATE}d word, we could simply
6396: have executed it to get the address of its data field. However, since it
6397: was defined to have @code{DOES>} actions, its execution semantics are to
6398: perform those @code{DOES>} actions. To get the address of its data field
6399: it's necessary to use @code{'} to get its xt, then @code{>BODY} to
6400: translate the xt into the address of the data field. When you execute
6401: @code{first-word}, it will display @code{123}. When you execute
6402: @code{second-word} it will display @code{-1}.
1.26 crook 6403:
6404: @cindex stack effect of @code{DOES>}-parts
6405: @cindex @code{DOES>}-parts, stack effect
1.29 crook 6406: In the examples above the stack comment after the @code{DOES>} specifies
1.26 crook 6407: the stack effect of the defined words, not the stack effect of the
6408: following code (the following code expects the address of the body on
6409: the top of stack, which is not reflected in the stack comment). This is
6410: the convention that I use and recommend (it clashes a bit with using
6411: locals declarations for stack effect specification, though).
1.1 anton 6412:
1.53 anton 6413: @menu
6414: * CREATE..DOES> applications::
6415: * CREATE..DOES> details::
1.63 anton 6416: * Advanced does> usage example::
1.53 anton 6417: @end menu
6418:
6419: @node CREATE..DOES> applications, CREATE..DOES> details, User-defined Defining Words, User-defined Defining Words
1.26 crook 6420: @subsubsection Applications of @code{CREATE..DOES>}
6421: @cindex @code{CREATE} ... @code{DOES>}, applications
1.1 anton 6422:
1.26 crook 6423: You may wonder how to use this feature. Here are some usage patterns:
1.1 anton 6424:
1.26 crook 6425: @cindex factoring similar colon definitions
6426: When you see a sequence of code occurring several times, and you can
6427: identify a meaning, you will factor it out as a colon definition. When
6428: you see similar colon definitions, you can factor them using
6429: @code{CREATE..DOES>}. E.g., an assembler usually defines several words
6430: that look very similar:
1.1 anton 6431: @example
1.26 crook 6432: : ori, ( reg-target reg-source n -- )
6433: 0 asm-reg-reg-imm ;
6434: : andi, ( reg-target reg-source n -- )
6435: 1 asm-reg-reg-imm ;
1.1 anton 6436: @end example
6437:
1.26 crook 6438: @noindent
6439: This could be factored with:
6440: @example
6441: : reg-reg-imm ( op-code -- )
6442: CREATE ,
6443: DOES> ( reg-target reg-source n -- )
6444: @@ asm-reg-reg-imm ;
6445:
6446: 0 reg-reg-imm ori,
6447: 1 reg-reg-imm andi,
6448: @end example
1.1 anton 6449:
1.26 crook 6450: @cindex currying
6451: Another view of @code{CREATE..DOES>} is to consider it as a crude way to
6452: supply a part of the parameters for a word (known as @dfn{currying} in
6453: the functional language community). E.g., @code{+} needs two
6454: parameters. Creating versions of @code{+} with one parameter fixed can
6455: be done like this:
1.1 anton 6456: @example
1.26 crook 6457: : curry+ ( n1 -- )
6458: CREATE ,
6459: DOES> ( n2 -- n1+n2 )
6460: @@ + ;
6461:
6462: 3 curry+ 3+
6463: -2 curry+ 2-
1.1 anton 6464: @end example
6465:
1.63 anton 6466: @node CREATE..DOES> details, Advanced does> usage example, CREATE..DOES> applications, User-defined Defining Words
1.26 crook 6467: @subsubsection The gory details of @code{CREATE..DOES>}
6468: @cindex @code{CREATE} ... @code{DOES>}, details
1.1 anton 6469:
1.26 crook 6470: doc-does>
1.1 anton 6471:
1.26 crook 6472: @cindex @code{DOES>} in a separate definition
6473: This means that you need not use @code{CREATE} and @code{DOES>} in the
6474: same definition; you can put the @code{DOES>}-part in a separate
1.29 crook 6475: definition. This allows us to, e.g., select among different @code{DOES>}-parts:
1.26 crook 6476: @example
6477: : does1
6478: DOES> ( ... -- ... )
1.44 crook 6479: ... ;
6480:
6481: : does2
6482: DOES> ( ... -- ... )
6483: ... ;
6484:
6485: : def-word ( ... -- ... )
6486: create ...
6487: IF
6488: does1
6489: ELSE
6490: does2
6491: ENDIF ;
6492: @end example
6493:
6494: In this example, the selection of whether to use @code{does1} or
1.69 anton 6495: @code{does2} is made at definition-time; at the time that the child word is
1.44 crook 6496: @code{CREATE}d.
6497:
6498: @cindex @code{DOES>} in interpretation state
6499: In a standard program you can apply a @code{DOES>}-part only if the last
6500: word was defined with @code{CREATE}. In Gforth, the @code{DOES>}-part
6501: will override the behaviour of the last word defined in any case. In a
6502: standard program, you can use @code{DOES>} only in a colon
6503: definition. In Gforth, you can also use it in interpretation state, in a
6504: kind of one-shot mode; for example:
6505: @example
6506: CREATE name ( ... -- ... )
6507: @i{initialization}
6508: DOES>
6509: @i{code} ;
6510: @end example
6511:
6512: @noindent
6513: is equivalent to the standard:
6514: @example
6515: :noname
6516: DOES>
6517: @i{code} ;
6518: CREATE name EXECUTE ( ... -- ... )
6519: @i{initialization}
6520: @end example
6521:
1.53 anton 6522: doc->body
6523:
1.63 anton 6524: @node Advanced does> usage example, , CREATE..DOES> details, User-defined Defining Words
6525: @subsubsection Advanced does> usage example
6526:
6527: The MIPS disassembler (@file{arch/mips/disasm.fs}) contains many words
6528: for disassembling instructions, that follow a very repetetive scheme:
6529:
6530: @example
6531: :noname @var{disasm-operands} s" @var{inst-name}" type ;
6532: @var{entry-num} cells @var{table} + !
6533: @end example
6534:
6535: Of course, this inspires the idea to factor out the commonalities to
6536: allow a definition like
6537:
6538: @example
6539: @var{disasm-operands} @var{entry-num} @var{table} define-inst @var{inst-name}
6540: @end example
6541:
6542: The parameters @var{disasm-operands} and @var{table} are usually
1.69 anton 6543: correlated. Moreover, before I wrote the disassembler, there already
6544: existed code that defines instructions like this:
1.63 anton 6545:
6546: @example
6547: @var{entry-num} @var{inst-format} @var{inst-name}
6548: @end example
6549:
6550: This code comes from the assembler and resides in
6551: @file{arch/mips/insts.fs}.
6552:
6553: So I had to define the @var{inst-format} words that performed the scheme
6554: above when executed. At first I chose to use run-time code-generation:
6555:
6556: @example
6557: : @var{inst-format} ( entry-num "name" -- ; compiled code: addr w -- )
6558: :noname Postpone @var{disasm-operands}
6559: name Postpone sliteral Postpone type Postpone ;
6560: swap cells @var{table} + ! ;
6561: @end example
6562:
6563: Note that this supplies the other two parameters of the scheme above.
1.44 crook 6564:
1.63 anton 6565: An alternative would have been to write this using
6566: @code{create}/@code{does>}:
6567:
6568: @example
6569: : @var{inst-format} ( entry-num "name" -- )
6570: here name string, ( entry-num c-addr ) \ parse and save "name"
6571: noname create , ( entry-num )
6572: lastxt swap cells @var{table} + !
6573: does> ( addr w -- )
6574: \ disassemble instruction w at addr
6575: @@ >r
6576: @var{disasm-operands}
6577: r> count type ;
6578: @end example
6579:
6580: Somehow the first solution is simpler, mainly because it's simpler to
6581: shift a string from definition-time to use-time with @code{sliteral}
6582: than with @code{string,} and friends.
6583:
6584: I wrote a lot of words following this scheme and soon thought about
6585: factoring out the commonalities among them. Note that this uses a
6586: two-level defining word, i.e., a word that defines ordinary defining
6587: words.
6588:
6589: This time a solution involving @code{postpone} and friends seemed more
6590: difficult (try it as an exercise), so I decided to use a
6591: @code{create}/@code{does>} word; since I was already at it, I also used
6592: @code{create}/@code{does>} for the lower level (try using
6593: @code{postpone} etc. as an exercise), resulting in the following
6594: definition:
6595:
6596: @example
6597: : define-format ( disasm-xt table-xt -- )
6598: \ define an instruction format that uses disasm-xt for
6599: \ disassembling and enters the defined instructions into table
6600: \ table-xt
6601: create 2,
6602: does> ( u "inst" -- )
6603: \ defines an anonymous word for disassembling instruction inst,
6604: \ and enters it as u-th entry into table-xt
6605: 2@@ swap here name string, ( u table-xt disasm-xt c-addr ) \ remember string
6606: noname create 2, \ define anonymous word
6607: execute lastxt swap ! \ enter xt of defined word into table-xt
6608: does> ( addr w -- )
6609: \ disassemble instruction w at addr
6610: 2@@ >r ( addr w disasm-xt R: c-addr )
6611: execute ( R: c-addr ) \ disassemble operands
6612: r> count type ; \ print name
6613: @end example
6614:
6615: Note that the tables here (in contrast to above) do the @code{cells +}
6616: by themselves (that's why you have to pass an xt). This word is used in
6617: the following way:
6618:
6619: @example
6620: ' @var{disasm-operands} ' @var{table} define-format @var{inst-format}
6621: @end example
6622:
1.71 anton 6623: As shown above, the defined instruction format is then used like this:
6624:
6625: @example
6626: @var{entry-num} @var{inst-format} @var{inst-name}
6627: @end example
6628:
1.63 anton 6629: In terms of currying, this kind of two-level defining word provides the
6630: parameters in three stages: first @var{disasm-operands} and @var{table},
6631: then @var{entry-num} and @var{inst-name}, finally @code{addr w}, i.e.,
6632: the instruction to be disassembled.
6633:
6634: Of course this did not quite fit all the instruction format names used
6635: in @file{insts.fs}, so I had to define a few wrappers that conditioned
6636: the parameters into the right form.
6637:
6638: If you have trouble following this section, don't worry. First, this is
6639: involved and takes time (and probably some playing around) to
6640: understand; second, this is the first two-level
6641: @code{create}/@code{does>} word I have written in seventeen years of
6642: Forth; and if I did not have @file{insts.fs} to start with, I may well
6643: have elected to use just a one-level defining word (with some repeating
6644: of parameters when using the defining word). So it is not necessary to
6645: understand this, but it may improve your understanding of Forth.
1.44 crook 6646:
6647:
6648: @node Deferred words, Aliases, User-defined Defining Words, Defining Words
6649: @subsection Deferred words
6650: @cindex deferred words
6651:
6652: The defining word @code{Defer} allows you to define a word by name
6653: without defining its behaviour; the definition of its behaviour is
6654: deferred. Here are two situation where this can be useful:
6655:
6656: @itemize @bullet
6657: @item
6658: Where you want to allow the behaviour of a word to be altered later, and
6659: for all precompiled references to the word to change when its behaviour
6660: is changed.
6661: @item
6662: For mutual recursion; @xref{Calls and returns}.
6663: @end itemize
6664:
6665: In the following example, @code{foo} always invokes the version of
6666: @code{greet} that prints ``@code{Good morning}'' whilst @code{bar}
6667: always invokes the version that prints ``@code{Hello}''. There is no way
6668: of getting @code{foo} to use the later version without re-ordering the
6669: source code and recompiling it.
6670:
6671: @example
6672: : greet ." Good morning" ;
6673: : foo ... greet ... ;
6674: : greet ." Hello" ;
6675: : bar ... greet ... ;
6676: @end example
6677:
6678: This problem can be solved by defining @code{greet} as a @code{Defer}red
6679: word. The behaviour of a @code{Defer}red word can be defined and
6680: redefined at any time by using @code{IS} to associate the xt of a
6681: previously-defined word with it. The previous example becomes:
6682:
6683: @example
1.69 anton 6684: Defer greet ( -- )
1.44 crook 6685: : foo ... greet ... ;
6686: : bar ... greet ... ;
1.69 anton 6687: : greet1 ( -- ) ." Good morning" ;
6688: : greet2 ( -- ) ." Hello" ;
1.44 crook 6689: ' greet2 <IS> greet \ make greet behave like greet2
6690: @end example
6691:
1.69 anton 6692: @progstyle
6693: You should write a stack comment for every deferred word, and put only
6694: XTs into deferred words that conform to this stack effect. Otherwise
6695: it's too difficult to use the deferred word.
6696:
1.44 crook 6697: A deferred word can be used to improve the statistics-gathering example
6698: from @ref{User-defined Defining Words}; rather than edit the
6699: application's source code to change every @code{:} to a @code{my:}, do
6700: this:
6701:
6702: @example
6703: : real: : ; \ retain access to the original
6704: defer : \ redefine as a deferred word
1.69 anton 6705: ' my: <IS> : \ use special version of :
1.44 crook 6706: \
6707: \ load application here
6708: \
1.69 anton 6709: ' real: <IS> : \ go back to the original
1.44 crook 6710: @end example
6711:
6712:
6713: One thing to note is that @code{<IS>} consumes its name when it is
6714: executed. If you want to specify the name at compile time, use
6715: @code{[IS]}:
6716:
6717: @example
6718: : set-greet ( xt -- )
6719: [IS] greet ;
6720:
6721: ' greet1 set-greet
6722: @end example
6723:
1.69 anton 6724: A deferred word can only inherit execution semantics from the xt
6725: (because that is all that an xt can represent -- for more discussion of
6726: this @pxref{Tokens for Words}); by default it will have default
6727: interpretation and compilation semantics deriving from this execution
6728: semantics. However, you can change the interpretation and compilation
6729: semantics of the deferred word in the usual ways:
1.44 crook 6730:
6731: @example
6732: : bar .... ; compile-only
6733: Defer fred immediate
6734: Defer jim
6735:
6736: ' bar <IS> jim \ jim has default semantics
6737: ' bar <IS> fred \ fred is immediate
6738: @end example
6739:
6740: doc-defer
6741: doc-<is>
6742: doc-[is]
6743: doc-is
6744: @comment TODO document these: what's defers [is]
6745: doc-what's
6746: doc-defers
6747:
6748: @c Use @code{words-deferred} to see a list of deferred words.
6749:
6750: Definitions in ANS Forth for @code{defer}, @code{<is>} and @code{[is]}
6751: are provided in @file{compat/defer.fs}.
6752:
6753:
1.69 anton 6754: @node Aliases, , Deferred words, Defining Words
1.44 crook 6755: @subsection Aliases
6756: @cindex aliases
1.1 anton 6757:
1.44 crook 6758: The defining word @code{Alias} allows you to define a word by name that
6759: has the same behaviour as some other word. Here are two situation where
6760: this can be useful:
1.1 anton 6761:
1.44 crook 6762: @itemize @bullet
6763: @item
6764: When you want access to a word's definition from a different word list
6765: (for an example of this, see the definition of the @code{Root} word list
6766: in the Gforth source).
6767: @item
6768: When you want to create a synonym; a definition that can be known by
6769: either of two names (for example, @code{THEN} and @code{ENDIF} are
6770: aliases).
6771: @end itemize
1.1 anton 6772:
1.69 anton 6773: Like deferred words, an alias has default compilation and interpretation
6774: semantics at the beginning (not the modifications of the other word),
6775: but you can change them in the usual ways (@code{immediate},
6776: @code{compile-only}). For example:
1.1 anton 6777:
6778: @example
1.44 crook 6779: : foo ... ; immediate
6780:
6781: ' foo Alias bar \ bar is not an immediate word
6782: ' foo Alias fooby immediate \ fooby is an immediate word
1.1 anton 6783: @end example
6784:
1.44 crook 6785: Words that are aliases have the same xt, different headers in the
6786: dictionary, and consequently different name tokens (@pxref{Tokens for
6787: Words}) and possibly different immediate flags. An alias can only have
6788: default or immediate compilation semantics; you can define aliases for
6789: combined words with @code{interpret/compile:} -- see @ref{Combined words}.
1.1 anton 6790:
1.44 crook 6791: doc-alias
1.1 anton 6792:
6793:
1.47 crook 6794: @node Interpretation and Compilation Semantics, Tokens for Words, Defining Words, Words
6795: @section Interpretation and Compilation Semantics
1.26 crook 6796: @cindex semantics, interpretation and compilation
1.1 anton 6797:
1.71 anton 6798: @c !! state and ' are used without explanation
6799: @c example for immediate/compile-only? or is the tutorial enough
6800:
1.26 crook 6801: @cindex interpretation semantics
1.71 anton 6802: The @dfn{interpretation semantics} of a (named) word are what the text
1.26 crook 6803: interpreter does when it encounters the word in interpret state. It also
6804: appears in some other contexts, e.g., the execution token returned by
1.71 anton 6805: @code{' @i{word}} identifies the interpretation semantics of @i{word}
6806: (in other words, @code{' @i{word} execute} is equivalent to
1.29 crook 6807: interpret-state text interpretation of @code{@i{word}}).
1.1 anton 6808:
1.26 crook 6809: @cindex compilation semantics
1.71 anton 6810: The @dfn{compilation semantics} of a (named) word are what the text
6811: interpreter does when it encounters the word in compile state. It also
6812: appears in other contexts, e.g, @code{POSTPONE @i{word}}
6813: compiles@footnote{In standard terminology, ``appends to the current
6814: definition''.} the compilation semantics of @i{word}.
1.1 anton 6815:
1.26 crook 6816: @cindex execution semantics
6817: The standard also talks about @dfn{execution semantics}. They are used
6818: only for defining the interpretation and compilation semantics of many
6819: words. By default, the interpretation semantics of a word are to
6820: @code{execute} its execution semantics, and the compilation semantics of
6821: a word are to @code{compile,} its execution semantics.@footnote{In
6822: standard terminology: The default interpretation semantics are its
6823: execution semantics; the default compilation semantics are to append its
6824: execution semantics to the execution semantics of the current
6825: definition.}
6826:
1.71 anton 6827: Unnamed words (@pxref{Anonymous Definitions}) cannot be encountered by
6828: the text interpreter, ticked, or @code{postpone}d, so they have no
6829: interpretation or compilation semantics. Their behaviour is represented
6830: by their XT (@pxref{Tokens for Words}), and we call it execution
6831: semantics, too.
6832:
1.26 crook 6833: @comment TODO expand, make it co-operate with new sections on text interpreter.
6834:
6835: @cindex immediate words
6836: @cindex compile-only words
6837: You can change the semantics of the most-recently defined word:
6838:
1.44 crook 6839:
1.26 crook 6840: doc-immediate
6841: doc-compile-only
6842: doc-restrict
6843:
1.44 crook 6844:
1.26 crook 6845: Note that ticking (@code{'}) a compile-only word gives an error
6846: (``Interpreting a compile-only word'').
1.1 anton 6847:
1.47 crook 6848: @menu
1.67 anton 6849: * Combined words::
1.47 crook 6850: @end menu
1.44 crook 6851:
1.71 anton 6852:
1.48 anton 6853: @node Combined words, , Interpretation and Compilation Semantics, Interpretation and Compilation Semantics
1.44 crook 6854: @subsection Combined Words
6855: @cindex combined words
6856:
6857: Gforth allows you to define @dfn{combined words} -- words that have an
6858: arbitrary combination of interpretation and compilation semantics.
6859:
1.26 crook 6860: doc-interpret/compile:
1.1 anton 6861:
1.26 crook 6862: This feature was introduced for implementing @code{TO} and @code{S"}. I
6863: recommend that you do not define such words, as cute as they may be:
6864: they make it hard to get at both parts of the word in some contexts.
6865: E.g., assume you want to get an execution token for the compilation
6866: part. Instead, define two words, one that embodies the interpretation
6867: part, and one that embodies the compilation part. Once you have done
6868: that, you can define a combined word with @code{interpret/compile:} for
6869: the convenience of your users.
1.1 anton 6870:
1.26 crook 6871: You might try to use this feature to provide an optimizing
6872: implementation of the default compilation semantics of a word. For
6873: example, by defining:
1.1 anton 6874: @example
1.26 crook 6875: :noname
6876: foo bar ;
6877: :noname
6878: POSTPONE foo POSTPONE bar ;
1.29 crook 6879: interpret/compile: opti-foobar
1.1 anton 6880: @end example
1.26 crook 6881:
1.23 crook 6882: @noindent
1.26 crook 6883: as an optimizing version of:
6884:
1.1 anton 6885: @example
1.26 crook 6886: : foobar
6887: foo bar ;
1.1 anton 6888: @end example
6889:
1.26 crook 6890: Unfortunately, this does not work correctly with @code{[compile]},
6891: because @code{[compile]} assumes that the compilation semantics of all
6892: @code{interpret/compile:} words are non-default. I.e., @code{[compile]
1.29 crook 6893: opti-foobar} would compile compilation semantics, whereas
6894: @code{[compile] foobar} would compile interpretation semantics.
1.1 anton 6895:
1.26 crook 6896: @cindex state-smart words (are a bad idea)
1.29 crook 6897: Some people try to use @dfn{state-smart} words to emulate the feature provided
1.26 crook 6898: by @code{interpret/compile:} (words are state-smart if they check
6899: @code{STATE} during execution). E.g., they would try to code
6900: @code{foobar} like this:
1.1 anton 6901:
1.26 crook 6902: @example
6903: : foobar
6904: STATE @@
6905: IF ( compilation state )
6906: POSTPONE foo POSTPONE bar
6907: ELSE
6908: foo bar
6909: ENDIF ; immediate
6910: @end example
1.1 anton 6911:
1.26 crook 6912: Although this works if @code{foobar} is only processed by the text
6913: interpreter, it does not work in other contexts (like @code{'} or
6914: @code{POSTPONE}). E.g., @code{' foobar} will produce an execution token
6915: for a state-smart word, not for the interpretation semantics of the
6916: original @code{foobar}; when you execute this execution token (directly
6917: with @code{EXECUTE} or indirectly through @code{COMPILE,}) in compile
6918: state, the result will not be what you expected (i.e., it will not
6919: perform @code{foo bar}). State-smart words are a bad idea. Simply don't
6920: write them@footnote{For a more detailed discussion of this topic, see
1.66 anton 6921: M. Anton Ertl,
6922: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl98.ps.gz,@code{State}-smartness---Why
6923: it is Evil and How to Exorcise it}}, EuroForth '98.}!
1.1 anton 6924:
1.26 crook 6925: @cindex defining words with arbitrary semantics combinations
6926: It is also possible to write defining words that define words with
6927: arbitrary combinations of interpretation and compilation semantics. In
6928: general, they look like this:
1.1 anton 6929:
1.26 crook 6930: @example
6931: : def-word
6932: create-interpret/compile
1.29 crook 6933: @i{code1}
1.26 crook 6934: interpretation>
1.29 crook 6935: @i{code2}
1.26 crook 6936: <interpretation
6937: compilation>
1.29 crook 6938: @i{code3}
1.26 crook 6939: <compilation ;
6940: @end example
1.1 anton 6941:
1.29 crook 6942: For a @i{word} defined with @code{def-word}, the interpretation
6943: semantics are to push the address of the body of @i{word} and perform
6944: @i{code2}, and the compilation semantics are to push the address of
6945: the body of @i{word} and perform @i{code3}. E.g., @code{constant}
1.26 crook 6946: can also be defined like this (except that the defined constants don't
6947: behave correctly when @code{[compile]}d):
1.1 anton 6948:
1.26 crook 6949: @example
6950: : constant ( n "name" -- )
6951: create-interpret/compile
6952: ,
6953: interpretation> ( -- n )
6954: @@
6955: <interpretation
6956: compilation> ( compilation. -- ; run-time. -- n )
6957: @@ postpone literal
6958: <compilation ;
6959: @end example
1.1 anton 6960:
1.44 crook 6961:
1.26 crook 6962: doc-create-interpret/compile
6963: doc-interpretation>
6964: doc-<interpretation
6965: doc-compilation>
6966: doc-<compilation
1.1 anton 6967:
1.44 crook 6968:
1.29 crook 6969: Words defined with @code{interpret/compile:} and
1.26 crook 6970: @code{create-interpret/compile} have an extended header structure that
6971: differs from other words; however, unless you try to access them with
6972: plain address arithmetic, you should not notice this. Words for
6973: accessing the header structure usually know how to deal with this; e.g.,
1.29 crook 6974: @code{'} @i{word} @code{>body} also gives you the body of a word created
6975: with @code{create-interpret/compile}.
1.1 anton 6976:
1.44 crook 6977:
1.47 crook 6978: @c -------------------------------------------------------------
1.81 ! anton 6979: @node Tokens for Words, Compiling words, Interpretation and Compilation Semantics, Words
1.47 crook 6980: @section Tokens for Words
6981: @cindex tokens for words
6982:
6983: This section describes the creation and use of tokens that represent
6984: words.
6985:
1.71 anton 6986: @menu
6987: * Execution token:: represents execution/interpretation semantics
6988: * Compilation token:: represents compilation semantics
6989: * Name token:: represents named words
6990: @end menu
1.47 crook 6991:
1.71 anton 6992: @node Execution token, Compilation token, Tokens for Words, Tokens for Words
6993: @subsection Execution token
1.47 crook 6994:
6995: @cindex xt
6996: @cindex execution token
1.71 anton 6997: An @dfn{execution token} (@i{XT}) represents some behaviour of a word.
6998: You can use @code{execute} to invoke this behaviour.
1.47 crook 6999:
1.71 anton 7000: @cindex tick (')
7001: You can use @code{'} to get an execution token that represents the
7002: interpretation semantics of a named word:
1.47 crook 7003:
7004: @example
1.71 anton 7005: 5 ' .
7006: execute
7007: @end example
1.47 crook 7008:
1.71 anton 7009: doc-'
7010:
7011: @code{'} parses at run-time; there is also a word @code{[']} that parses
7012: when it is compiled, and compiles the resulting XT:
7013:
7014: @example
7015: : foo ['] . execute ;
7016: 5 foo
7017: : bar ' execute ; \ by contrast,
7018: 5 bar . \ ' parses "." when bar executes
7019: @end example
7020:
7021: doc-[']
7022:
7023: If you want the execution token of @i{word}, write @code{['] @i{word}}
7024: in compiled code and @code{' @i{word}} in interpreted code. Gforth's
7025: @code{'} and @code{[']} behave somewhat unusually by complaining about
7026: compile-only words (because these words have no interpretation
7027: semantics). You might get what you want by using @code{COMP' @i{word}
7028: DROP} or @code{[COMP'] @i{word} DROP} (for details @pxref{Compilation
7029: token}).
7030:
7031: Another way to get an XT is @code{:noname} or @code{lastxt}
7032: (@pxref{Anonymous Definitions}). For anonymous words this gives an xt
7033: for the only behaviour the word has (the execution semantics). For
7034: named words, @code{lastxt} produces an XT for the same behaviour it
7035: would produce if the word was defined anonymously.
7036:
7037: @example
7038: :noname ." hello" ;
7039: execute
1.47 crook 7040: @end example
7041:
1.71 anton 7042: An XT occupies one cell and can be manipulated like any other cell.
7043:
1.47 crook 7044: @cindex code field address
7045: @cindex CFA
1.71 anton 7046: In ANS Forth the XT is just an abstract data type (i.e., defined by the
7047: operations that produce or consume it). For old hands: In Gforth, the
7048: XT is implemented as a code field address (CFA).
7049:
7050: @c !! discuss "compile," some more (or in Macros).
7051:
7052: doc-execute
7053: doc-perform
7054: doc-compile,
7055:
7056: @node Compilation token, Name token, Execution token, Tokens for Words
7057: @subsection Compilation token
1.47 crook 7058:
7059: @cindex compilation token
1.71 anton 7060: @cindex CT (compilation token)
7061: Gforth represents the compilation semantics of a named word by a
1.47 crook 7062: @dfn{compilation token} consisting of two cells: @i{w xt}. The top cell
7063: @i{xt} is an execution token. The compilation semantics represented by
7064: the compilation token can be performed with @code{execute}, which
7065: consumes the whole compilation token, with an additional stack effect
7066: determined by the represented compilation semantics.
7067:
7068: At present, the @i{w} part of a compilation token is an execution token,
7069: and the @i{xt} part represents either @code{execute} or
7070: @code{compile,}@footnote{Depending upon the compilation semantics of the
7071: word. If the word has default compilation semantics, the @i{xt} will
7072: represent @code{compile,}. Otherwise (e.g., for immediate words), the
7073: @i{xt} will represent @code{execute}.}. However, don't rely on that
7074: knowledge, unless necessary; future versions of Gforth may introduce
7075: unusual compilation tokens (e.g., a compilation token that represents
7076: the compilation semantics of a literal).
7077:
1.71 anton 7078: You can perform the compilation semantics represented by the compilation
7079: token with @code{execute}. You can compile the compilation semantics
7080: with @code{postpone,}. I.e., @code{COMP' @i{word} postpone,} is
7081: equivalent to @code{postpone @i{word}}.
7082:
7083: doc-[comp']
7084: doc-comp'
7085: doc-postpone,
7086:
7087: @node Name token, , Compilation token, Tokens for Words
7088: @subsection Name token
1.47 crook 7089:
7090: @cindex name token
7091: @cindex name field address
7092: @cindex NFA
1.71 anton 7093: Gforth represents named words by the @dfn{name token}, (@i{nt}). In
1.47 crook 7094: Gforth, the abstract data type @emph{name token} is implemented as a
7095: name field address (NFA).
7096:
7097: doc-find-name
7098: doc-name>int
7099: doc-name?int
7100: doc-name>comp
7101: doc-name>string
7102:
1.81 ! anton 7103: @c ----------------------------------------------------------
! 7104: @node Compiling words, The Text Interpreter, Tokens for Words, Words
! 7105: @section Compiling words
! 7106: @cindex compiling words
! 7107: @cindex macros
! 7108:
! 7109: In contrast to most other languages, Forth has no strict boundary
! 7110: between compilation and run-time.
! 7111:
! 7112: doc-postpone
! 7113:
! 7114: @comment TODO -- expand glossary text for POSTPONE
! 7115:
! 7116: doc-literal
! 7117: doc-]L
! 7118: doc-2literal
! 7119: doc-fliteral
! 7120: doc-[
! 7121: doc-]
! 7122:
! 7123:
! 7124:
1.47 crook 7125:
1.26 crook 7126: @c ----------------------------------------------------------
1.81 ! anton 7127: @node The Text Interpreter, Word Lists, Compiling words, Words
1.26 crook 7128: @section The Text Interpreter
7129: @cindex interpreter - outer
7130: @cindex text interpreter
7131: @cindex outer interpreter
1.1 anton 7132:
1.34 anton 7133: @c Should we really describe all these ugly details? IMO the text
7134: @c interpreter should be much cleaner, but that may not be possible within
7135: @c ANS Forth. - anton
1.44 crook 7136: @c nac-> I wanted to explain how it works to show how you can exploit
7137: @c it in your own programs. When I was writing a cross-compiler, figuring out
7138: @c some of these gory details was very helpful to me. None of the textbooks
7139: @c I've seen cover it, and the most modern Forth textbook -- Forth Inc's,
7140: @c seems to positively avoid going into too much detail for some of
7141: @c the internals.
1.34 anton 7142:
1.71 anton 7143: @c anton: ok. I wonder, though, if this is the right place; for some stuff
7144: @c it is; for the ugly details, I would prefer another place. I wonder
7145: @c whether we should have a chapter before "Words" that describes some
7146: @c basic concepts referred to in words, and a chapter after "Words" that
7147: @c describes implementation details.
7148:
1.29 crook 7149: The text interpreter@footnote{This is an expanded version of the
7150: material in @ref{Introducing the Text Interpreter}.} is an endless loop
1.34 anton 7151: that processes input from the current input device. It is also called
7152: the outer interpreter, in contrast to the inner interpreter
7153: (@pxref{Engine}) which executes the compiled Forth code on interpretive
7154: implementations.
1.27 crook 7155:
1.29 crook 7156: @cindex interpret state
7157: @cindex compile state
7158: The text interpreter operates in one of two states: @dfn{interpret
7159: state} and @dfn{compile state}. The current state is defined by the
1.71 anton 7160: aptly-named variable @code{state}.
1.29 crook 7161:
7162: This section starts by describing how the text interpreter behaves when
7163: it is in interpret state, processing input from the user input device --
7164: the keyboard. This is the mode that a Forth system is in after it starts
7165: up.
7166:
7167: @cindex input buffer
7168: @cindex terminal input buffer
7169: The text interpreter works from an area of memory called the @dfn{input
7170: buffer}@footnote{When the text interpreter is processing input from the
7171: keyboard, this area of memory is called the @dfn{terminal input buffer}
7172: (TIB) and is addressed by the (obsolescent) words @code{TIB} and
7173: @code{#TIB}.}, which stores your keyboard input when you press the
1.30 anton 7174: @key{RET} key. Starting at the beginning of the input buffer, it skips
1.29 crook 7175: leading spaces (called @dfn{delimiters}) then parses a string (a
7176: sequence of non-space characters) until it reaches either a space
7177: character or the end of the buffer. Having parsed a string, it makes two
7178: attempts to process it:
1.27 crook 7179:
1.29 crook 7180: @cindex dictionary
1.27 crook 7181: @itemize @bullet
7182: @item
1.29 crook 7183: It looks for the string in a @dfn{dictionary} of definitions. If the
7184: string is found, the string names a @dfn{definition} (also known as a
7185: @dfn{word}) and the dictionary search returns information that allows
7186: the text interpreter to perform the word's @dfn{interpretation
7187: semantics}. In most cases, this simply means that the word will be
7188: executed.
1.27 crook 7189: @item
7190: If the string is not found in the dictionary, the text interpreter
1.29 crook 7191: attempts to treat it as a number, using the rules described in
7192: @ref{Number Conversion}. If the string represents a legal number in the
7193: current radix, the number is pushed onto a parameter stack (the data
7194: stack for integers, the floating-point stack for floating-point
7195: numbers).
7196: @end itemize
7197:
7198: If both attempts fail, or if the word is found in the dictionary but has
7199: no interpretation semantics@footnote{This happens if the word was
7200: defined as @code{COMPILE-ONLY}.} the text interpreter discards the
7201: remainder of the input buffer, issues an error message and waits for
7202: more input. If one of the attempts succeeds, the text interpreter
7203: repeats the parsing process until the whole of the input buffer has been
7204: processed, at which point it prints the status message ``@code{ ok}''
7205: and waits for more input.
7206:
1.71 anton 7207: @c anton: this should be in the input stream subsection (or below it)
7208:
1.29 crook 7209: @cindex parse area
7210: The text interpreter keeps track of its position in the input buffer by
7211: updating a variable called @code{>IN} (pronounced ``to-in''). The value
7212: of @code{>IN} starts out as 0, indicating an offset of 0 from the start
7213: of the input buffer. The region from offset @code{>IN @@} to the end of
7214: the input buffer is called the @dfn{parse area}@footnote{In other words,
7215: the text interpreter processes the contents of the input buffer by
7216: parsing strings from the parse area until the parse area is empty.}.
7217: This example shows how @code{>IN} changes as the text interpreter parses
7218: the input buffer:
7219:
7220: @example
7221: : remaining >IN @@ SOURCE 2 PICK - -ROT + SWAP
7222: CR ." ->" TYPE ." <-" ; IMMEDIATE
7223:
7224: 1 2 3 remaining + remaining .
7225:
7226: : foo 1 2 3 remaining SWAP remaining ;
7227: @end example
7228:
7229: @noindent
7230: The result is:
7231:
7232: @example
7233: ->+ remaining .<-
7234: ->.<-5 ok
7235:
7236: ->SWAP remaining ;-<
7237: ->;<- ok
7238: @end example
7239:
7240: @cindex parsing words
7241: The value of @code{>IN} can also be modified by a word in the input
7242: buffer that is executed by the text interpreter. This means that a word
7243: can ``trick'' the text interpreter into either skipping a section of the
7244: input buffer@footnote{This is how parsing words work.} or into parsing a
7245: section twice. For example:
1.27 crook 7246:
1.29 crook 7247: @example
1.71 anton 7248: : lat ." <<foo>>" ;
7249: : flat ." <<bar>>" >IN DUP @@ 3 - SWAP ! ;
1.29 crook 7250: @end example
7251:
7252: @noindent
7253: When @code{flat} is executed, this output is produced@footnote{Exercise
7254: for the reader: what would happen if the @code{3} were replaced with
7255: @code{4}?}:
7256:
7257: @example
1.71 anton 7258: <<bar>><<foo>>
1.29 crook 7259: @end example
7260:
1.71 anton 7261: This technique can be used to work around some of the interoperability
7262: problems of parsing words. Of course, it's better to avoid parsing
7263: words where possible.
7264:
1.29 crook 7265: @noindent
7266: Two important notes about the behaviour of the text interpreter:
1.27 crook 7267:
7268: @itemize @bullet
7269: @item
7270: It processes each input string to completion before parsing additional
1.29 crook 7271: characters from the input buffer.
7272: @item
7273: It treats the input buffer as a read-only region (and so must your code).
7274: @end itemize
7275:
7276: @noindent
7277: When the text interpreter is in compile state, its behaviour changes in
7278: these ways:
7279:
7280: @itemize @bullet
7281: @item
7282: If a parsed string is found in the dictionary, the text interpreter will
7283: perform the word's @dfn{compilation semantics}. In most cases, this
7284: simply means that the execution semantics of the word will be appended
7285: to the current definition.
1.27 crook 7286: @item
1.29 crook 7287: When a number is encountered, it is compiled into the current definition
7288: (as a literal) rather than being pushed onto a parameter stack.
7289: @item
7290: If an error occurs, @code{state} is modified to put the text interpreter
7291: back into interpret state.
7292: @item
7293: Each time a line is entered from the keyboard, Gforth prints
7294: ``@code{ compiled}'' rather than `` @code{ok}''.
7295: @end itemize
7296:
7297: @cindex text interpreter - input sources
7298: When the text interpreter is using an input device other than the
7299: keyboard, its behaviour changes in these ways:
7300:
7301: @itemize @bullet
7302: @item
7303: When the parse area is empty, the text interpreter attempts to refill
7304: the input buffer from the input source. When the input source is
1.71 anton 7305: exhausted, the input source is set back to the previous input source.
1.29 crook 7306: @item
7307: It doesn't print out ``@code{ ok}'' or ``@code{ compiled}'' messages each
7308: time the parse area is emptied.
7309: @item
7310: If an error occurs, the input source is set back to the user input
7311: device.
1.27 crook 7312: @end itemize
1.21 crook 7313:
1.49 anton 7314: You can read about this in more detail in @ref{Input Sources}.
1.44 crook 7315:
1.26 crook 7316: doc->in
1.27 crook 7317: doc-source
7318:
1.26 crook 7319: doc-tib
7320: doc-#tib
1.1 anton 7321:
1.44 crook 7322:
1.26 crook 7323: @menu
1.67 anton 7324: * Input Sources::
7325: * Number Conversion::
7326: * Interpret/Compile states::
7327: * Literals::
7328: * Interpreter Directives::
1.26 crook 7329: @end menu
1.1 anton 7330:
1.29 crook 7331: @node Input Sources, Number Conversion, The Text Interpreter, The Text Interpreter
7332: @subsection Input Sources
7333: @cindex input sources
7334: @cindex text interpreter - input sources
7335:
1.44 crook 7336: By default, the text interpreter processes input from the user input
1.29 crook 7337: device (the keyboard) when Forth starts up. The text interpreter can
7338: process input from any of these sources:
7339:
7340: @itemize @bullet
7341: @item
7342: The user input device -- the keyboard.
7343: @item
7344: A file, using the words described in @ref{Forth source files}.
7345: @item
7346: A block, using the words described in @ref{Blocks}.
7347: @item
7348: A text string, using @code{evaluate}.
7349: @end itemize
7350:
7351: A program can identify the current input device from the values of
7352: @code{source-id} and @code{blk}.
7353:
1.44 crook 7354:
1.29 crook 7355: doc-source-id
7356: doc-blk
7357:
7358: doc-save-input
7359: doc-restore-input
7360:
7361: doc-evaluate
1.1 anton 7362:
1.29 crook 7363:
1.44 crook 7364:
1.29 crook 7365: @node Number Conversion, Interpret/Compile states, Input Sources, The Text Interpreter
1.26 crook 7366: @subsection Number Conversion
7367: @cindex number conversion
7368: @cindex double-cell numbers, input format
7369: @cindex input format for double-cell numbers
7370: @cindex single-cell numbers, input format
7371: @cindex input format for single-cell numbers
7372: @cindex floating-point numbers, input format
7373: @cindex input format for floating-point numbers
1.1 anton 7374:
1.29 crook 7375: This section describes the rules that the text interpreter uses when it
7376: tries to convert a string into a number.
1.1 anton 7377:
1.26 crook 7378: Let <digit> represent any character that is a legal digit in the current
1.29 crook 7379: number base@footnote{For example, 0-9 when the number base is decimal or
7380: 0-9, A-F when the number base is hexadecimal.}.
1.1 anton 7381:
1.26 crook 7382: Let <decimal digit> represent any character in the range 0-9.
1.1 anton 7383:
1.29 crook 7384: Let @{@i{a b}@} represent the @i{optional} presence of any of the characters
7385: in the braces (@i{a} or @i{b} or neither).
1.1 anton 7386:
1.26 crook 7387: Let * represent any number of instances of the previous character
7388: (including none).
1.1 anton 7389:
1.26 crook 7390: Let any other character represent itself.
1.1 anton 7391:
1.29 crook 7392: @noindent
1.26 crook 7393: Now, the conversion rules are:
1.21 crook 7394:
1.26 crook 7395: @itemize @bullet
7396: @item
7397: A string of the form <digit><digit>* is treated as a single-precision
1.29 crook 7398: (cell-sized) positive integer. Examples are 0 123 6784532 32343212343456 42
1.26 crook 7399: @item
7400: A string of the form -<digit><digit>* is treated as a single-precision
1.29 crook 7401: (cell-sized) negative integer, and is represented using 2's-complement
1.26 crook 7402: arithmetic. Examples are -45 -5681 -0
7403: @item
7404: A string of the form <digit><digit>*.<digit>* is treated as a double-precision
1.29 crook 7405: (double-cell-sized) positive integer. Examples are 3465. 3.465 34.65
7406: (all three of these represent the same number).
1.26 crook 7407: @item
7408: A string of the form -<digit><digit>*.<digit>* is treated as a
1.29 crook 7409: double-precision (double-cell-sized) negative integer, and is
1.26 crook 7410: represented using 2's-complement arithmetic. Examples are -3465. -3.465
1.29 crook 7411: -34.65 (all three of these represent the same number).
1.26 crook 7412: @item
1.29 crook 7413: A string of the form @{+ -@}<decimal digit>@{.@}<decimal digit>*@{e
7414: E@}@{+ -@}<decimal digit><decimal digit>* is treated as a floating-point
1.35 anton 7415: number. Examples are 1e 1e0 1.e 1.e0 +1e+0 (which all represent the same
1.29 crook 7416: number) +12.E-4
1.26 crook 7417: @end itemize
1.1 anton 7418:
1.26 crook 7419: By default, the number base used for integer number conversion is given
1.35 anton 7420: by the contents of the variable @code{base}. Note that a lot of
7421: confusion can result from unexpected values of @code{base}. If you
7422: change @code{base} anywhere, make sure to save the old value and restore
7423: it afterwards. In general I recommend keeping @code{base} decimal, and
7424: using the prefixes described below for the popular non-decimal bases.
1.1 anton 7425:
1.29 crook 7426: doc-dpl
1.26 crook 7427: doc-base
7428: doc-hex
7429: doc-decimal
1.1 anton 7430:
1.44 crook 7431:
1.26 crook 7432: @cindex '-prefix for character strings
7433: @cindex &-prefix for decimal numbers
7434: @cindex %-prefix for binary numbers
7435: @cindex $-prefix for hexadecimal numbers
1.35 anton 7436: Gforth allows you to override the value of @code{base} by using a
1.29 crook 7437: prefix@footnote{Some Forth implementations provide a similar scheme by
7438: implementing @code{$} etc. as parsing words that process the subsequent
7439: number in the input stream and push it onto the stack. For example, see
7440: @cite{Number Conversion and Literals}, by Wil Baden; Forth Dimensions
7441: 20(3) pages 26--27. In such implementations, unlike in Gforth, a space
7442: is required between the prefix and the number.} before the first digit
7443: of an (integer) number. Four prefixes are supported:
1.1 anton 7444:
1.26 crook 7445: @itemize @bullet
7446: @item
1.35 anton 7447: @code{&} -- decimal
1.26 crook 7448: @item
1.35 anton 7449: @code{%} -- binary
1.26 crook 7450: @item
1.35 anton 7451: @code{$} -- hexadecimal
1.26 crook 7452: @item
1.35 anton 7453: @code{'} -- base @code{max-char+1}
1.26 crook 7454: @end itemize
1.1 anton 7455:
1.26 crook 7456: Here are some examples, with the equivalent decimal number shown after
7457: in braces:
1.1 anton 7458:
1.26 crook 7459: -$41 (-65), %1001101 (205), %1001.0001 (145 - a double-precision number),
7460: 'AB (16706; ascii A is 65, ascii B is 66, number is 65*256 + 66),
7461: 'ab (24930; ascii a is 97, ascii B is 98, number is 97*256 + 98),
7462: &905 (905), $abc (2478), $ABC (2478).
1.1 anton 7463:
1.26 crook 7464: @cindex number conversion - traps for the unwary
1.29 crook 7465: @noindent
1.26 crook 7466: Number conversion has a number of traps for the unwary:
1.1 anton 7467:
1.26 crook 7468: @itemize @bullet
7469: @item
7470: You cannot determine the current number base using the code sequence
1.35 anton 7471: @code{base @@ .} -- the number base is always 10 in the current number
7472: base. Instead, use something like @code{base @@ dec.}
1.26 crook 7473: @item
7474: If the number base is set to a value greater than 14 (for example,
7475: hexadecimal), the number 123E4 is ambiguous; the conversion rules allow
7476: it to be intepreted as either a single-precision integer or a
7477: floating-point number (Gforth treats it as an integer). The ambiguity
7478: can be resolved by explicitly stating the sign of the mantissa and/or
7479: exponent: 123E+4 or +123E4 -- if the number base is decimal, no
7480: ambiguity arises; either representation will be treated as a
7481: floating-point number.
7482: @item
1.29 crook 7483: There is a word @code{bin} but it does @i{not} set the number base!
1.26 crook 7484: It is used to specify file types.
7485: @item
1.72 anton 7486: ANS Forth requires the @code{.} of a double-precision number to be the
7487: final character in the string. Gforth allows the @code{.} to be
7488: anywhere after the first digit.
1.26 crook 7489: @item
7490: The number conversion process does not check for overflow.
7491: @item
1.72 anton 7492: In an ANS Forth program @code{base} is required to be decimal when
7493: converting floating-point numbers. In Gforth, number conversion to
7494: floating-point numbers always uses base &10, irrespective of the value
7495: of @code{base}.
1.26 crook 7496: @end itemize
1.1 anton 7497:
1.49 anton 7498: You can read numbers into your programs with the words described in
7499: @ref{Input}.
1.1 anton 7500:
1.26 crook 7501: @node Interpret/Compile states, Literals, Number Conversion, The Text Interpreter
7502: @subsection Interpret/Compile states
7503: @cindex Interpret/Compile states
1.1 anton 7504:
1.29 crook 7505: A standard program is not permitted to change @code{state}
7506: explicitly. However, it can change @code{state} implicitly, using the
7507: words @code{[} and @code{]}. When @code{[} is executed it switches
7508: @code{state} to interpret state, and therefore the text interpreter
7509: starts interpreting. When @code{]} is executed it switches @code{state}
7510: to compile state and therefore the text interpreter starts
1.44 crook 7511: compiling. The most common usage for these words is for switching into
7512: interpret state and back from within a colon definition; this technique
1.49 anton 7513: can be used to compile a literal (for an example, @pxref{Literals}) or
7514: for conditional compilation (for an example, @pxref{Interpreter
7515: Directives}).
1.44 crook 7516:
1.35 anton 7517:
7518: @c This is a bad example: It's non-standard, and it's not necessary.
7519: @c However, I can't think of a good example for switching into compile
7520: @c state when there is no current word (@code{state}-smart words are not a
7521: @c good reason). So maybe we should use an example for switching into
7522: @c interpret @code{state} in a colon def. - anton
1.44 crook 7523: @c nac-> I agree. I started out by putting in the example, then realised
7524: @c that it was non-ANS, so wrote more words around it. I hope this
7525: @c re-written version is acceptable to you. I do want to keep the example
7526: @c as it is helpful for showing what is and what is not portable, particularly
7527: @c where it outlaws a style in common use.
7528:
1.72 anton 7529: @c anton: it's more important to show what's portable. After we have done
7530: @c that, we can also show what's not. In any case, I intend to write a
7531: @c section Macros (or so) which will also deal with [ ].
1.35 anton 7532:
1.44 crook 7533: @code{[} and @code{]} also give you the ability to switch into compile
7534: state and back, but we cannot think of any useful Standard application
7535: for this ability. Pre-ANS Forth textbooks have examples like this:
1.29 crook 7536:
7537: @example
7538: : AA ." this is A" ;
7539: : BB ." this is B" ;
7540: : CC ." this is C" ;
7541:
1.44 crook 7542: create table ] aa bb cc [
7543:
1.29 crook 7544: : go ( n -- ) \ n is offset into table.. 0 for 1st entry
7545: cells table + @ execute ;
7546: @end example
7547:
1.44 crook 7548: This example builds a jump table; @code{0 go} will display ``@code{this
7549: is A}''. Using @code{[} and @code{]} in this example is equivalent to
7550: defining @code{table} like this:
1.29 crook 7551:
7552: @example
1.44 crook 7553: create table ' aa COMPILE, ' bb COMPILE, ' cc COMPILE,
1.29 crook 7554: @end example
7555:
1.44 crook 7556: The problem with this code is that the definition of @code{table} is not
7557: portable -- it @i{compile}s execution tokens into code space. Whilst it
7558: @i{may} work on systems where code space and data space co-incide, the
1.29 crook 7559: Standard only allows data space to be assigned for a @code{CREATE}d
7560: word. In addition, the Standard only allows @code{@@} to access data
7561: space, whilst this example is using it to access code space. The only
7562: portable, Standard way to build this table is to build it in data space,
7563: like this:
7564:
7565: @example
7566: create table ' aa , ' bb , ' cc ,
7567: @end example
7568:
1.26 crook 7569: doc-state
1.44 crook 7570:
1.26 crook 7571: @node Literals, Interpreter Directives, Interpret/Compile states, The Text Interpreter
7572: @subsection Literals
7573: @cindex Literals
1.21 crook 7574:
1.29 crook 7575: Often, you want to use a number within a colon definition. When you do
7576: this, the text interpreter automatically compiles the number as a
7577: @i{literal}. A literal is a number whose run-time effect is to be pushed
7578: onto the stack. If you had to do some maths to generate the number, you
7579: might write it like this:
7580:
7581: @example
7582: : HOUR-TO-SEC ( n1 -- n2 )
7583: 60 * \ to minutes
7584: 60 * ; \ to seconds
7585: @end example
7586:
7587: It is very clear what this definition is doing, but it's inefficient
7588: since it is performing 2 multiples at run-time. An alternative would be
7589: to write:
7590:
7591: @example
7592: : HOUR-TO-SEC ( n1 -- n2 )
7593: 3600 * ; \ to seconds
7594: @end example
7595:
7596: Which does the same thing, and has the advantage of using a single
7597: multiply. Ideally, we'd like the efficiency of the second with the
7598: readability of the first.
7599:
7600: @code{Literal} allows us to achieve that. It takes a number from the
7601: stack and lays it down in the current definition just as though the
7602: number had been typed directly into the definition. Our first attempt
7603: might look like this:
7604:
7605: @example
7606: 60 \ mins per hour
7607: 60 * \ seconds per minute
7608: : HOUR-TO-SEC ( n1 -- n2 )
7609: Literal * ; \ to seconds
7610: @end example
7611:
7612: But this produces the error message @code{unstructured}. What happened?
7613: The stack notation for @code{:} is (@i{ -- colon-sys}) and the size of
7614: @i{colon-sys} is implementation-defined. In other words, once we start a
7615: colon definition we can't portably access anything that was on the stack
7616: before the definition began@footnote{@cite{Two Problems in ANS Forth},
7617: by Thomas Worthington; Forth Dimensions 20(2) pages 32--34 describes
7618: some situations where you might want to access stack items above
7619: colon-sys, and provides a solution to the problem.}. The correct way of
7620: solving this problem in this instance is to use @code{[ ]} like this:
7621:
7622: @example
7623: : HOUR-TO-SEC ( n1 -- n2 )
7624: [ 60 \ minutes per hour
7625: 60 * ] \ seconds per minute
7626: LITERAL * ; \ to seconds
7627: @end example
1.23 crook 7628:
1.44 crook 7629:
7630:
1.48 anton 7631: @node Interpreter Directives, , Literals, The Text Interpreter
1.26 crook 7632: @subsection Interpreter Directives
7633: @cindex interpreter directives
1.72 anton 7634: @cindex conditional compilation
1.1 anton 7635:
1.29 crook 7636: These words are usually used in interpret state; typically to control
7637: which parts of a source file are processed by the text
1.26 crook 7638: interpreter. There are only a few ANS Forth Standard words, but Gforth
7639: supplements these with a rich set of immediate control structure words
7640: to compensate for the fact that the non-immediate versions can only be
1.29 crook 7641: used in compile state (@pxref{Control Structures}). Typical usages:
7642:
7643: @example
1.72 anton 7644: FALSE Constant HAVE-ASSEMBLER
1.29 crook 7645: .
7646: .
1.72 anton 7647: HAVE-ASSEMBLER [IF]
1.29 crook 7648: : ASSEMBLER-FEATURE
7649: ...
7650: ;
7651: [ENDIF]
7652: .
7653: .
7654: : SEE
7655: ... \ general-purpose SEE code
1.72 anton 7656: [ HAVE-ASSEMBLER [IF] ]
1.29 crook 7657: ... \ assembler-specific SEE code
7658: [ [ENDIF] ]
7659: ;
7660: @end example
1.1 anton 7661:
1.44 crook 7662:
1.26 crook 7663: doc-[IF]
7664: doc-[ELSE]
7665: doc-[THEN]
7666: doc-[ENDIF]
1.1 anton 7667:
1.26 crook 7668: doc-[IFDEF]
7669: doc-[IFUNDEF]
1.1 anton 7670:
1.26 crook 7671: doc-[?DO]
7672: doc-[DO]
7673: doc-[FOR]
7674: doc-[LOOP]
7675: doc-[+LOOP]
7676: doc-[NEXT]
1.1 anton 7677:
1.26 crook 7678: doc-[BEGIN]
7679: doc-[UNTIL]
7680: doc-[AGAIN]
7681: doc-[WHILE]
7682: doc-[REPEAT]
1.1 anton 7683:
1.27 crook 7684:
1.26 crook 7685: @c -------------------------------------------------------------
1.47 crook 7686: @node Word Lists, Environmental Queries, The Text Interpreter, Words
1.26 crook 7687: @section Word Lists
7688: @cindex word lists
1.32 anton 7689: @cindex header space
1.1 anton 7690:
1.36 anton 7691: A wordlist is a list of named words; you can add new words and look up
7692: words by name (and you can remove words in a restricted way with
7693: markers). Every named (and @code{reveal}ed) word is in one wordlist.
7694:
7695: @cindex search order stack
7696: The text interpreter searches the wordlists present in the search order
7697: (a stack of wordlists), from the top to the bottom. Within each
7698: wordlist, the search starts conceptually at the newest word; i.e., if
7699: two words in a wordlist have the same name, the newer word is found.
1.1 anton 7700:
1.26 crook 7701: @cindex compilation word list
1.36 anton 7702: New words are added to the @dfn{compilation wordlist} (aka current
7703: wordlist).
1.1 anton 7704:
1.36 anton 7705: @cindex wid
7706: A word list is identified by a cell-sized word list identifier (@i{wid})
7707: in much the same way as a file is identified by a file handle. The
7708: numerical value of the wid has no (portable) meaning, and might change
7709: from session to session.
1.1 anton 7710:
1.29 crook 7711: The ANS Forth ``Search order'' word set is intended to provide a set of
7712: low-level tools that allow various different schemes to be
1.74 anton 7713: implemented. Gforth also provides @code{vocabulary}, a traditional Forth
1.26 crook 7714: word. @file{compat/vocabulary.fs} provides an implementation in ANS
1.45 crook 7715: Forth.
1.1 anton 7716:
1.27 crook 7717: @comment TODO: locals section refers to here, saying that every word list (aka
7718: @comment vocabulary) has its own methods for searching etc. Need to document that.
1.78 anton 7719: @c anton: but better in a separate subsection on wordlist internals
1.1 anton 7720:
1.45 crook 7721: @comment TODO: document markers, reveal, tables, mappedwordlist
7722:
7723: @comment the gforthman- prefix is used to pick out the true definition of a
1.27 crook 7724: @comment word from the source files, rather than some alias.
1.44 crook 7725:
1.26 crook 7726: doc-forth-wordlist
7727: doc-definitions
7728: doc-get-current
7729: doc-set-current
7730: doc-get-order
1.45 crook 7731: doc---gforthman-set-order
1.26 crook 7732: doc-wordlist
1.30 anton 7733: doc-table
1.79 anton 7734: doc->order
1.36 anton 7735: doc-previous
1.26 crook 7736: doc-also
1.45 crook 7737: doc---gforthman-forth
1.26 crook 7738: doc-only
1.45 crook 7739: doc---gforthman-order
1.15 anton 7740:
1.26 crook 7741: doc-find
7742: doc-search-wordlist
1.15 anton 7743:
1.26 crook 7744: doc-words
7745: doc-vlist
1.44 crook 7746: @c doc-words-deferred
1.1 anton 7747:
1.74 anton 7748: @c doc-mappedwordlist @c map-structure undefined, implemantation-specific
1.26 crook 7749: doc-root
7750: doc-vocabulary
7751: doc-seal
7752: doc-vocs
7753: doc-current
7754: doc-context
1.1 anton 7755:
1.44 crook 7756:
1.26 crook 7757: @menu
1.75 anton 7758: * Vocabularies::
1.67 anton 7759: * Why use word lists?::
1.75 anton 7760: * Word list example::
1.26 crook 7761: @end menu
7762:
1.75 anton 7763: @node Vocabularies, Why use word lists?, Word Lists, Word Lists
7764: @subsection Vocabularies
7765: @cindex Vocabularies, detailed explanation
7766:
7767: Here is an example of creating and using a new wordlist using ANS
7768: Forth words:
7769:
7770: @example
7771: wordlist constant my-new-words-wordlist
7772: : my-new-words get-order nip my-new-words-wordlist swap set-order ;
7773:
7774: \ add it to the search order
7775: also my-new-words
7776:
7777: \ alternatively, add it to the search order and make it
7778: \ the compilation word list
7779: also my-new-words definitions
7780: \ type "order" to see the problem
7781: @end example
7782:
7783: The problem with this example is that @code{order} has no way to
7784: associate the name @code{my-new-words} with the wid of the word list (in
7785: Gforth, @code{order} and @code{vocs} will display @code{???} for a wid
7786: that has no associated name). There is no Standard way of associating a
7787: name with a wid.
7788:
7789: In Gforth, this example can be re-coded using @code{vocabulary}, which
7790: associates a name with a wid:
7791:
7792: @example
7793: vocabulary my-new-words
7794:
7795: \ add it to the search order
7796: also my-new-words
7797:
7798: \ alternatively, add it to the search order and make it
7799: \ the compilation word list
7800: my-new-words definitions
7801: \ type "order" to see that the problem is solved
7802: @end example
7803:
7804:
7805: @node Why use word lists?, Word list example, Vocabularies, Word Lists
1.26 crook 7806: @subsection Why use word lists?
7807: @cindex word lists - why use them?
7808:
1.74 anton 7809: Here are some reasons why people use wordlists:
1.26 crook 7810:
7811: @itemize @bullet
1.74 anton 7812:
7813: @c anton: Gforth's hashing implementation makes the search speed
7814: @c independent from the number of words. But it is linear with the number
7815: @c of wordlists that have to be searched, so in effect using more wordlists
7816: @c actually slows down compilation.
7817:
7818: @c @item
7819: @c To improve compilation speed by reducing the number of header space
7820: @c entries that must be searched. This is achieved by creating a new
7821: @c word list that contains all of the definitions that are used in the
7822: @c definition of a Forth system but which would not usually be used by
7823: @c programs running on that system. That word list would be on the search
7824: @c list when the Forth system was compiled but would be removed from the
7825: @c search list for normal operation. This can be a useful technique for
7826: @c low-performance systems (for example, 8-bit processors in embedded
7827: @c systems) but is unlikely to be necessary in high-performance desktop
7828: @c systems.
7829:
1.26 crook 7830: @item
7831: To prevent a set of words from being used outside the context in which
7832: they are valid. Two classic examples of this are an integrated editor
7833: (all of the edit commands are defined in a separate word list; the
7834: search order is set to the editor word list when the editor is invoked;
7835: the old search order is restored when the editor is terminated) and an
7836: integrated assembler (the op-codes for the machine are defined in a
7837: separate word list which is used when a @code{CODE} word is defined).
1.74 anton 7838:
7839: @item
7840: To organize the words of an application or library into a user-visible
7841: set (in @code{forth-wordlist} or some other common wordlist) and a set
7842: of helper words used just for the implementation (hidden in a separate
1.75 anton 7843: wordlist). This keeps @code{words}' output smaller, separates
7844: implementation and interface, and reduces the chance of name conflicts
7845: within the common wordlist.
1.74 anton 7846:
1.26 crook 7847: @item
7848: To prevent a name-space clash between multiple definitions with the same
7849: name. For example, when building a cross-compiler you might have a word
7850: @code{IF} that generates conditional code for your target system. By
7851: placing this definition in a different word list you can control whether
7852: the host system's @code{IF} or the target system's @code{IF} get used in
7853: any particular context by controlling the order of the word lists on the
7854: search order stack.
1.74 anton 7855:
1.26 crook 7856: @end itemize
1.1 anton 7857:
1.74 anton 7858: The downsides of using wordlists are:
7859:
7860: @itemize
7861:
7862: @item
7863: Debugging becomes more cumbersome.
7864:
7865: @item
7866: Name conflicts worked around with wordlists are still there, and you
7867: have to arrange the search order carefully to get the desired results;
7868: if you forget to do that, you get hard-to-find errors (as in any case
7869: where you read the code differently from the compiler; @code{see} can
1.75 anton 7870: help seeing which of several possible words the name resolves to in such
7871: cases). @code{See} displays just the name of the words, not what
7872: wordlist they belong to, so it might be misleading. Using unique names
7873: is a better approach to avoid name conflicts.
1.74 anton 7874:
7875: @item
7876: You have to explicitly undo any changes to the search order. In many
7877: cases it would be more convenient if this happened implicitly. Gforth
7878: currently does not provide such a feature, but it may do so in the
7879: future.
7880: @end itemize
7881:
7882:
1.75 anton 7883: @node Word list example, , Why use word lists?, Word Lists
7884: @subsection Word list example
7885: @cindex word lists - example
1.1 anton 7886:
1.74 anton 7887: The following example is from the
7888: @uref{http://www.complang.tuwien.ac.at/forth/garbage-collection.zip,
7889: garbage collector} and uses wordlists to separate public words from
7890: helper words:
7891:
7892: @example
7893: get-current ( wid )
7894: vocabulary garbage-collector also garbage-collector definitions
7895: ... \ define helper words
7896: ( wid ) set-current \ restore original (i.e., public) compilation wordlist
7897: ... \ define the public (i.e., API) words
7898: \ they can refer to the helper words
7899: previous \ restore original search order (helper words become invisible)
7900: @end example
7901:
1.26 crook 7902: @c -------------------------------------------------------------
7903: @node Environmental Queries, Files, Word Lists, Words
7904: @section Environmental Queries
7905: @cindex environmental queries
1.21 crook 7906:
1.26 crook 7907: ANS Forth introduced the idea of ``environmental queries'' as a way
7908: for a program running on a system to determine certain characteristics of the system.
7909: The Standard specifies a number of strings that might be recognised by a system.
1.21 crook 7910:
1.32 anton 7911: The Standard requires that the header space used for environmental queries
7912: be distinct from the header space used for definitions.
1.21 crook 7913:
1.26 crook 7914: Typically, environmental queries are supported by creating a set of
1.29 crook 7915: definitions in a word list that is @i{only} used during environmental
1.26 crook 7916: queries; that is what Gforth does. There is no Standard way of adding
7917: definitions to the set of recognised environmental queries, but any
7918: implementation that supports the loading of optional word sets must have
7919: some mechanism for doing this (after loading the word set, the
7920: associated environmental query string must return @code{true}). In
7921: Gforth, the word list used to honour environmental queries can be
7922: manipulated just like any other word list.
1.21 crook 7923:
1.44 crook 7924:
1.26 crook 7925: doc-environment?
7926: doc-environment-wordlist
1.21 crook 7927:
1.26 crook 7928: doc-gforth
7929: doc-os-class
1.21 crook 7930:
1.44 crook 7931:
1.26 crook 7932: Note that, whilst the documentation for (e.g.) @code{gforth} shows it
7933: returning two items on the stack, querying it using @code{environment?}
7934: will return an additional item; the @code{true} flag that shows that the
7935: string was recognised.
1.21 crook 7936:
1.26 crook 7937: @comment TODO Document the standard strings or note where they are documented herein
1.21 crook 7938:
1.26 crook 7939: Here are some examples of using environmental queries:
1.21 crook 7940:
1.26 crook 7941: @example
7942: s" address-unit-bits" environment? 0=
7943: [IF]
7944: cr .( environmental attribute address-units-bits unknown... ) cr
1.75 anton 7945: [ELSE]
7946: drop \ ensure balanced stack effect
1.26 crook 7947: [THEN]
1.21 crook 7948:
1.75 anton 7949: \ this might occur in the prelude of a standard program that uses THROW
7950: s" exception" environment? [IF]
7951: 0= [IF]
7952: : throw abort" exception thrown" ;
7953: [THEN]
7954: [ELSE] \ we don't know, so make sure
7955: : throw abort" exception thrown" ;
7956: [THEN]
1.21 crook 7957:
1.26 crook 7958: s" gforth" environment? [IF] .( Gforth version ) TYPE
7959: [ELSE] .( Not Gforth..) [THEN]
1.75 anton 7960:
7961: \ a program using v*
7962: s" gforth" environment? [IF]
7963: s" 0.5.0" compare 0< [IF] \ v* is a primitive since 0.5.0
7964: : v* ( f_addr1 nstride1 f_addr2 nstride2 ucount -- r )
7965: >r swap 2swap swap 0e r> 0 ?DO
7966: dup f@ over + 2swap dup f@ f* f+ over + 2swap
7967: LOOP
7968: 2drop 2drop ;
7969: [THEN]
7970: [ELSE] \
7971: : v* ( f_addr1 nstride1 f_addr2 nstride2 ucount -- r )
7972: ...
7973: [THEN]
1.26 crook 7974: @end example
1.21 crook 7975:
1.26 crook 7976: Here is an example of adding a definition to the environment word list:
1.21 crook 7977:
1.26 crook 7978: @example
7979: get-current environment-wordlist set-current
7980: true constant block
7981: true constant block-ext
7982: set-current
7983: @end example
1.21 crook 7984:
1.26 crook 7985: You can see what definitions are in the environment word list like this:
1.21 crook 7986:
1.26 crook 7987: @example
1.79 anton 7988: environment-wordlist >order words previous
1.26 crook 7989: @end example
1.21 crook 7990:
7991:
1.26 crook 7992: @c -------------------------------------------------------------
7993: @node Files, Blocks, Environmental Queries, Words
7994: @section Files
1.28 crook 7995: @cindex files
7996: @cindex I/O - file-handling
1.21 crook 7997:
1.26 crook 7998: Gforth provides facilities for accessing files that are stored in the
7999: host operating system's file-system. Files that are processed by Gforth
8000: can be divided into two categories:
1.21 crook 8001:
1.23 crook 8002: @itemize @bullet
8003: @item
1.29 crook 8004: Files that are processed by the Text Interpreter (@dfn{Forth source files}).
1.23 crook 8005: @item
1.29 crook 8006: Files that are processed by some other program (@dfn{general files}).
1.26 crook 8007: @end itemize
8008:
8009: @menu
1.48 anton 8010: * Forth source files::
8011: * General files::
8012: * Search Paths::
1.26 crook 8013: @end menu
8014:
8015: @c -------------------------------------------------------------
8016: @node Forth source files, General files, Files, Files
8017: @subsection Forth source files
8018: @cindex including files
8019: @cindex Forth source files
1.21 crook 8020:
1.26 crook 8021: The simplest way to interpret the contents of a file is to use one of
8022: these two formats:
1.21 crook 8023:
1.26 crook 8024: @example
8025: include mysource.fs
8026: s" mysource.fs" included
8027: @end example
1.21 crook 8028:
1.75 anton 8029: You usually want to include a file only if it is not included already
1.26 crook 8030: (by, say, another source file). In that case, you can use one of these
1.45 crook 8031: three formats:
1.21 crook 8032:
1.26 crook 8033: @example
8034: require mysource.fs
8035: needs mysource.fs
8036: s" mysource.fs" required
8037: @end example
1.21 crook 8038:
1.26 crook 8039: @cindex stack effect of included files
8040: @cindex including files, stack effect
1.45 crook 8041: It is good practice to write your source files such that interpreting them
8042: does not change the stack. Source files designed in this way can be used with
1.26 crook 8043: @code{required} and friends without complications. For example:
1.21 crook 8044:
1.26 crook 8045: @example
1.75 anton 8046: 1024 require foo.fs drop
1.26 crook 8047: @end example
1.21 crook 8048:
1.75 anton 8049: Here you want to pass the argument 1024 (e.g., a buffer size) to
8050: @file{foo.fs}. Interpreting @file{foo.fs} has the stack effect ( n -- n
8051: ), which allows its use with @code{require}. Of course with such
8052: parameters to required files, you have to ensure that the first
8053: @code{require} fits for all uses (i.e., @code{require} it early in the
8054: master load file).
1.44 crook 8055:
1.26 crook 8056: doc-include-file
8057: doc-included
1.28 crook 8058: doc-included?
1.26 crook 8059: doc-include
8060: doc-required
8061: doc-require
8062: doc-needs
1.75 anton 8063: @c doc-init-included-files @c internal
8064: @c doc-loadfilename @c internal word
8065: doc-sourcefilename
8066: doc-sourceline#
1.44 crook 8067:
1.26 crook 8068: A definition in ANS Forth for @code{required} is provided in
8069: @file{compat/required.fs}.
1.21 crook 8070:
1.26 crook 8071: @c -------------------------------------------------------------
8072: @node General files, Search Paths, Forth source files, Files
8073: @subsection General files
8074: @cindex general files
8075: @cindex file-handling
1.21 crook 8076:
1.75 anton 8077: Files are opened/created by name and type. The following file access
8078: methods (FAMs) are recognised:
1.44 crook 8079:
1.75 anton 8080: @cindex fam (file access method)
1.26 crook 8081: doc-r/o
8082: doc-r/w
8083: doc-w/o
8084: doc-bin
1.1 anton 8085:
1.44 crook 8086:
1.26 crook 8087: When a file is opened/created, it returns a file identifier,
1.29 crook 8088: @i{wfileid} that is used for all other file commands. All file
8089: commands also return a status value, @i{wior}, that is 0 for a
1.26 crook 8090: successful operation and an implementation-defined non-zero value in the
8091: case of an error.
1.21 crook 8092:
1.44 crook 8093:
1.26 crook 8094: doc-open-file
8095: doc-create-file
1.21 crook 8096:
1.26 crook 8097: doc-close-file
8098: doc-delete-file
8099: doc-rename-file
8100: doc-read-file
8101: doc-read-line
8102: doc-write-file
8103: doc-write-line
8104: doc-emit-file
8105: doc-flush-file
1.21 crook 8106:
1.26 crook 8107: doc-file-status
8108: doc-file-position
8109: doc-reposition-file
8110: doc-file-size
8111: doc-resize-file
1.21 crook 8112:
1.44 crook 8113:
1.26 crook 8114: @c ---------------------------------------------------------
1.48 anton 8115: @node Search Paths, , General files, Files
1.26 crook 8116: @subsection Search Paths
8117: @cindex path for @code{included}
8118: @cindex file search path
8119: @cindex @code{include} search path
8120: @cindex search path for files
1.21 crook 8121:
1.26 crook 8122: If you specify an absolute filename (i.e., a filename starting with
8123: @file{/} or @file{~}, or with @file{:} in the second position (as in
8124: @samp{C:...})) for @code{included} and friends, that file is included
8125: just as you would expect.
1.21 crook 8126:
1.75 anton 8127: If the filename starts with @file{./}, this refers to the directory that
8128: the present file was @code{included} from. This allows files to include
8129: other files relative to their own position (irrespective of the current
8130: working directory or the absolute position). This feature is essential
8131: for libraries consisting of several files, where a file may include
8132: other files from the library. It corresponds to @code{#include "..."}
8133: in C. If the current input source is not a file, @file{.} refers to the
8134: directory of the innermost file being included, or, if there is no file
8135: being included, to the current working directory.
8136:
8137: For relative filenames (not starting with @file{./}), Gforth uses a
8138: search path similar to Forth's search order (@pxref{Word Lists}). It
8139: tries to find the given filename in the directories present in the path,
8140: and includes the first one it finds. There are separate search paths for
8141: Forth source files and general files. If the search path contains the
8142: directory @file{.}, this refers to the directory of the current file, or
8143: the working directory, as if the file had been specified with @file{./}.
1.21 crook 8144:
1.26 crook 8145: Use @file{~+} to refer to the current working directory (as in the
8146: @code{bash}).
1.1 anton 8147:
1.75 anton 8148: @c anton: fold the following subsubsections into this subsection?
1.1 anton 8149:
1.48 anton 8150: @menu
1.75 anton 8151: * Source Search Paths::
1.48 anton 8152: * General Search Paths::
8153: @end menu
8154:
1.26 crook 8155: @c ---------------------------------------------------------
1.75 anton 8156: @node Source Search Paths, General Search Paths, Search Paths, Search Paths
8157: @subsubsection Source Search Paths
8158: @cindex search path control, source files
1.5 anton 8159:
1.26 crook 8160: The search path is initialized when you start Gforth (@pxref{Invoking
1.75 anton 8161: Gforth}). You can display it and change it using @code{fpath} in
8162: combination with the general path handling words.
1.5 anton 8163:
1.75 anton 8164: doc-fpath
8165: @c the functionality of the following words is easily available through
8166: @c fpath and the general path words. The may go away.
8167: @c doc-.fpath
8168: @c doc-fpath+
8169: @c doc-fpath=
8170: @c doc-open-fpath-file
1.44 crook 8171:
8172: @noindent
1.26 crook 8173: Here is an example of using @code{fpath} and @code{require}:
1.5 anton 8174:
1.26 crook 8175: @example
1.75 anton 8176: fpath path= /usr/lib/forth/|./
1.26 crook 8177: require timer.fs
8178: @end example
1.5 anton 8179:
1.75 anton 8180:
1.26 crook 8181: @c ---------------------------------------------------------
1.75 anton 8182: @node General Search Paths, , Source Search Paths, Search Paths
1.26 crook 8183: @subsubsection General Search Paths
1.75 anton 8184: @cindex search path control, source files
1.5 anton 8185:
1.26 crook 8186: Your application may need to search files in several directories, like
8187: @code{included} does. To facilitate this, Gforth allows you to define
8188: and use your own search paths, by providing generic equivalents of the
8189: Forth search path words:
1.5 anton 8190:
1.75 anton 8191: doc-open-path-file
8192: doc-path-allot
8193: doc-clear-path
8194: doc-also-path
1.26 crook 8195: doc-.path
8196: doc-path+
8197: doc-path=
1.5 anton 8198:
1.75 anton 8199: @c anton: better define a word for it, say "path-allot ( ucount -- path-addr )
1.44 crook 8200:
1.75 anton 8201: Here's an example of creating an empty search path:
8202: @c
1.26 crook 8203: @example
1.75 anton 8204: create mypath 500 path-allot \ maximum length 500 chars (is checked)
1.26 crook 8205: @end example
1.5 anton 8206:
1.26 crook 8207: @c -------------------------------------------------------------
8208: @node Blocks, Other I/O, Files, Words
8209: @section Blocks
1.28 crook 8210: @cindex I/O - blocks
8211: @cindex blocks
8212:
8213: When you run Gforth on a modern desk-top computer, it runs under the
8214: control of an operating system which provides certain services. One of
8215: these services is @var{file services}, which allows Forth source code
8216: and data to be stored in files and read into Gforth (@pxref{Files}).
8217:
8218: Traditionally, Forth has been an important programming language on
8219: systems where it has interfaced directly to the underlying hardware with
8220: no intervening operating system. Forth provides a mechanism, called
1.29 crook 8221: @dfn{blocks}, for accessing mass storage on such systems.
1.28 crook 8222:
8223: A block is a 1024-byte data area, which can be used to hold data or
8224: Forth source code. No structure is imposed on the contents of the
8225: block. A block is identified by its number; blocks are numbered
8226: contiguously from 1 to an implementation-defined maximum.
8227:
8228: A typical system that used blocks but no operating system might use a
8229: single floppy-disk drive for mass storage, with the disks formatted to
8230: provide 256-byte sectors. Blocks would be implemented by assigning the
8231: first four sectors of the disk to block 1, the second four sectors to
8232: block 2 and so on, up to the limit of the capacity of the disk. The disk
8233: would not contain any file system information, just the set of blocks.
8234:
1.29 crook 8235: @cindex blocks file
1.28 crook 8236: On systems that do provide file services, blocks are typically
1.29 crook 8237: implemented by storing a sequence of blocks within a single @dfn{blocks
1.28 crook 8238: file}. The size of the blocks file will be an exact multiple of 1024
8239: bytes, corresponding to the number of blocks it contains. This is the
8240: mechanism that Gforth uses.
8241:
1.29 crook 8242: @cindex @file{blocks.fb}
1.75 anton 8243: Only one blocks file can be open at a time. If you use block words without
1.28 crook 8244: having specified a blocks file, Gforth defaults to the blocks file
8245: @file{blocks.fb}. Gforth uses the Forth search path when attempting to
1.75 anton 8246: locate a blocks file (@pxref{Source Search Paths}).
1.28 crook 8247:
1.29 crook 8248: @cindex block buffers
1.28 crook 8249: When you read and write blocks under program control, Gforth uses a
1.29 crook 8250: number of @dfn{block buffers} as intermediate storage. These buffers are
1.28 crook 8251: not used when you use @code{load} to interpret the contents of a block.
8252:
1.75 anton 8253: The behaviour of the block buffers is analagous to that of a cache.
8254: Each block buffer has three states:
1.28 crook 8255:
8256: @itemize @bullet
8257: @item
8258: Unassigned
8259: @item
8260: Assigned-clean
8261: @item
8262: Assigned-dirty
8263: @end itemize
8264:
1.29 crook 8265: Initially, all block buffers are @i{unassigned}. In order to access a
1.28 crook 8266: block, the block (specified by its block number) must be assigned to a
8267: block buffer.
8268:
8269: The assignment of a block to a block buffer is performed by @code{block}
8270: or @code{buffer}. Use @code{block} when you wish to modify the existing
8271: contents of a block. Use @code{buffer} when you don't care about the
8272: existing contents of the block@footnote{The ANS Forth definition of
1.35 anton 8273: @code{buffer} is intended not to cause disk I/O; if the data associated
1.28 crook 8274: with the particular block is already stored in a block buffer due to an
8275: earlier @code{block} command, @code{buffer} will return that block
8276: buffer and the existing contents of the block will be
8277: available. Otherwise, @code{buffer} will simply assign a new, empty
1.29 crook 8278: block buffer for the block.}.
1.28 crook 8279:
1.47 crook 8280: Once a block has been assigned to a block buffer using @code{block} or
1.75 anton 8281: @code{buffer}, that block buffer becomes the @i{current block
8282: buffer}. Data may only be manipulated (read or written) within the
8283: current block buffer.
1.47 crook 8284:
8285: When the contents of the current block buffer has been modified it is
1.48 anton 8286: necessary, @emph{before calling @code{block} or @code{buffer} again}, to
1.75 anton 8287: either abandon the changes (by doing nothing) or mark the block as
8288: changed (assigned-dirty), using @code{update}. Using @code{update} does
8289: not change the blocks file; it simply changes a block buffer's state to
8290: @i{assigned-dirty}. The block will be written implicitly when it's
8291: buffer is needed for another block, or explicitly by @code{flush} or
8292: @code{save-buffers}.
8293:
8294: word @code{Flush} writes all @i{assigned-dirty} blocks back to the
8295: blocks file on disk. Leaving Gforth with @code{bye} also performs a
8296: @code{flush}.
1.28 crook 8297:
1.29 crook 8298: In Gforth, @code{block} and @code{buffer} use a @i{direct-mapped}
1.28 crook 8299: algorithm to assign a block buffer to a block. That means that any
8300: particular block can only be assigned to one specific block buffer,
1.29 crook 8301: called (for the particular operation) the @i{victim buffer}. If the
1.47 crook 8302: victim buffer is @i{unassigned} or @i{assigned-clean} it is allocated to
8303: the new block immediately. If it is @i{assigned-dirty} its current
8304: contents are written back to the blocks file on disk before it is
1.28 crook 8305: allocated to the new block.
8306:
8307: Although no structure is imposed on the contents of a block, it is
8308: traditional to display the contents as 16 lines each of 64 characters. A
8309: block provides a single, continuous stream of input (for example, it
8310: acts as a single parse area) -- there are no end-of-line characters
8311: within a block, and no end-of-file character at the end of a
8312: block. There are two consequences of this:
1.26 crook 8313:
1.28 crook 8314: @itemize @bullet
8315: @item
8316: The last character of one line wraps straight into the first character
8317: of the following line
8318: @item
8319: The word @code{\} -- comment to end of line -- requires special
8320: treatment; in the context of a block it causes all characters until the
8321: end of the current 64-character ``line'' to be ignored.
8322: @end itemize
8323:
8324: In Gforth, when you use @code{block} with a non-existent block number,
1.45 crook 8325: the current blocks file will be extended to the appropriate size and the
1.28 crook 8326: block buffer will be initialised with spaces.
8327:
1.47 crook 8328: Gforth includes a simple block editor (type @code{use blocked.fb 0 list}
8329: for details) but doesn't encourage the use of blocks; the mechanism is
8330: only provided for backward compatibility -- ANS Forth requires blocks to
8331: be available when files are.
1.28 crook 8332:
8333: Common techniques that are used when working with blocks include:
8334:
8335: @itemize @bullet
8336: @item
8337: A screen editor that allows you to edit blocks without leaving the Forth
8338: environment.
8339: @item
8340: Shadow screens; where every code block has an associated block
8341: containing comments (for example: code in odd block numbers, comments in
8342: even block numbers). Typically, the block editor provides a convenient
8343: mechanism to toggle between code and comments.
8344: @item
8345: Load blocks; a single block (typically block 1) contains a number of
8346: @code{thru} commands which @code{load} the whole of the application.
8347: @end itemize
1.26 crook 8348:
1.29 crook 8349: See Frank Sergeant's Pygmy Forth to see just how well blocks can be
8350: integrated into a Forth programming environment.
1.26 crook 8351:
8352: @comment TODO what about errors on open-blocks?
1.44 crook 8353:
1.26 crook 8354: doc-open-blocks
8355: doc-use
1.75 anton 8356: doc-block-offset
1.26 crook 8357: doc-get-block-fid
8358: doc-block-position
1.28 crook 8359:
1.75 anton 8360: doc-list
1.28 crook 8361: doc-scr
8362:
1.45 crook 8363: doc---gforthman-block
1.28 crook 8364: doc-buffer
8365:
1.75 anton 8366: doc-empty-buffers
8367: doc-empty-buffer
1.26 crook 8368: doc-update
1.28 crook 8369: doc-updated?
1.26 crook 8370: doc-save-buffers
1.75 anton 8371: doc-save-buffer
1.26 crook 8372: doc-flush
1.28 crook 8373:
1.26 crook 8374: doc-load
8375: doc-thru
8376: doc-+load
8377: doc-+thru
1.45 crook 8378: doc---gforthman--->
1.26 crook 8379: doc-block-included
8380:
1.44 crook 8381:
1.26 crook 8382: @c -------------------------------------------------------------
1.78 anton 8383: @node Other I/O, Locals, Blocks, Words
1.26 crook 8384: @section Other I/O
1.28 crook 8385: @cindex I/O - keyboard and display
1.26 crook 8386:
8387: @menu
8388: * Simple numeric output:: Predefined formats
8389: * Formatted numeric output:: Formatted (pictured) output
8390: * String Formats:: How Forth stores strings in memory
1.67 anton 8391: * Displaying characters and strings:: Other stuff
1.26 crook 8392: * Input:: Input
8393: @end menu
8394:
8395: @node Simple numeric output, Formatted numeric output, Other I/O, Other I/O
8396: @subsection Simple numeric output
1.28 crook 8397: @cindex numeric output - simple/free-format
1.5 anton 8398:
1.26 crook 8399: The simplest output functions are those that display numbers from the
8400: data or floating-point stacks. Floating-point output is always displayed
8401: using base 10. Numbers displayed from the data stack use the value stored
8402: in @code{base}.
1.5 anton 8403:
1.44 crook 8404:
1.26 crook 8405: doc-.
8406: doc-dec.
8407: doc-hex.
8408: doc-u.
8409: doc-.r
8410: doc-u.r
8411: doc-d.
8412: doc-ud.
8413: doc-d.r
8414: doc-ud.r
8415: doc-f.
8416: doc-fe.
8417: doc-fs.
1.5 anton 8418:
1.44 crook 8419:
1.26 crook 8420: Examples of printing the number 1234.5678E23 in the different floating-point output
8421: formats are shown below:
1.5 anton 8422:
8423: @example
1.26 crook 8424: f. 123456779999999000000000000.
8425: fe. 123.456779999999E24
8426: fs. 1.23456779999999E26
1.5 anton 8427: @end example
8428:
8429:
1.26 crook 8430: @node Formatted numeric output, String Formats, Simple numeric output, Other I/O
8431: @subsection Formatted numeric output
1.28 crook 8432: @cindex formatted numeric output
1.26 crook 8433: @cindex pictured numeric output
1.28 crook 8434: @cindex numeric output - formatted
1.26 crook 8435:
1.29 crook 8436: Forth traditionally uses a technique called @dfn{pictured numeric
1.26 crook 8437: output} for formatted printing of integers. In this technique, digits
8438: are extracted from the number (using the current output radix defined by
8439: @code{base}), converted to ASCII codes and appended to a string that is
8440: built in a scratch-pad area of memory (@pxref{core-idef,
8441: Implementation-defined options, Implementation-defined
8442: options}). Arbitrary characters can be appended to the string during the
8443: extraction process. The completed string is specified by an address
8444: and length and can be manipulated (@code{TYPE}ed, copied, modified)
8445: under program control.
1.5 anton 8446:
1.75 anton 8447: All of the integer output words described in the previous section
8448: (@pxref{Simple numeric output}) are implemented in Gforth using pictured
8449: numeric output.
1.5 anton 8450:
1.47 crook 8451: Three important things to remember about pictured numeric output:
1.5 anton 8452:
1.26 crook 8453: @itemize @bullet
8454: @item
1.28 crook 8455: It always operates on double-precision numbers; to display a
1.49 anton 8456: single-precision number, convert it first (for ways of doing this
8457: @pxref{Double precision}).
1.26 crook 8458: @item
1.28 crook 8459: It always treats the double-precision number as though it were
8460: unsigned. The examples below show ways of printing signed numbers.
1.26 crook 8461: @item
8462: The string is built up from right to left; least significant digit first.
8463: @end itemize
1.5 anton 8464:
1.44 crook 8465:
1.26 crook 8466: doc-<#
1.47 crook 8467: doc-<<#
1.26 crook 8468: doc-#
8469: doc-#s
8470: doc-hold
8471: doc-sign
8472: doc-#>
1.47 crook 8473: doc-#>>
1.5 anton 8474:
1.26 crook 8475: doc-represent
1.5 anton 8476:
1.44 crook 8477:
8478: @noindent
1.26 crook 8479: Here are some examples of using pictured numeric output:
1.5 anton 8480:
8481: @example
1.26 crook 8482: : my-u. ( u -- )
8483: \ Simplest use of pns.. behaves like Standard u.
8484: 0 \ convert to unsigned double
1.75 anton 8485: <<# \ start conversion
1.26 crook 8486: #s \ convert all digits
8487: #> \ complete conversion
1.75 anton 8488: TYPE SPACE \ display, with trailing space
8489: #>> ; \ release hold area
1.5 anton 8490:
1.26 crook 8491: : cents-only ( u -- )
8492: 0 \ convert to unsigned double
1.75 anton 8493: <<# \ start conversion
1.26 crook 8494: # # \ convert two least-significant digits
8495: #> \ complete conversion, discard other digits
1.75 anton 8496: TYPE SPACE \ display, with trailing space
8497: #>> ; \ release hold area
1.5 anton 8498:
1.26 crook 8499: : dollars-and-cents ( u -- )
8500: 0 \ convert to unsigned double
1.75 anton 8501: <<# \ start conversion
1.26 crook 8502: # # \ convert two least-significant digits
8503: [char] . hold \ insert decimal point
8504: #s \ convert remaining digits
8505: [char] $ hold \ append currency symbol
8506: #> \ complete conversion
1.75 anton 8507: TYPE SPACE \ display, with trailing space
8508: #>> ; \ release hold area
1.5 anton 8509:
1.26 crook 8510: : my-. ( n -- )
8511: \ handling negatives.. behaves like Standard .
8512: s>d \ convert to signed double
8513: swap over dabs \ leave sign byte followed by unsigned double
1.75 anton 8514: <<# \ start conversion
1.26 crook 8515: #s \ convert all digits
8516: rot sign \ get at sign byte, append "-" if needed
8517: #> \ complete conversion
1.75 anton 8518: TYPE SPACE \ display, with trailing space
8519: #>> ; \ release hold area
1.5 anton 8520:
1.26 crook 8521: : account. ( n -- )
1.75 anton 8522: \ accountants don't like minus signs, they use parentheses
1.26 crook 8523: \ for negative numbers
8524: s>d \ convert to signed double
8525: swap over dabs \ leave sign byte followed by unsigned double
1.75 anton 8526: <<# \ start conversion
1.26 crook 8527: 2 pick \ get copy of sign byte
8528: 0< IF [char] ) hold THEN \ right-most character of output
8529: #s \ convert all digits
8530: rot \ get at sign byte
8531: 0< IF [char] ( hold THEN
8532: #> \ complete conversion
1.75 anton 8533: TYPE SPACE \ display, with trailing space
8534: #>> ; \ release hold area
8535:
1.5 anton 8536: @end example
8537:
1.26 crook 8538: Here are some examples of using these words:
1.5 anton 8539:
8540: @example
1.26 crook 8541: 1 my-u. 1
8542: hex -1 my-u. decimal FFFFFFFF
8543: 1 cents-only 01
8544: 1234 cents-only 34
8545: 2 dollars-and-cents $0.02
8546: 1234 dollars-and-cents $12.34
8547: 123 my-. 123
8548: -123 my. -123
8549: 123 account. 123
8550: -456 account. (456)
1.5 anton 8551: @end example
8552:
8553:
1.26 crook 8554: @node String Formats, Displaying characters and strings, Formatted numeric output, Other I/O
8555: @subsection String Formats
1.27 crook 8556: @cindex strings - see character strings
8557: @cindex character strings - formats
1.28 crook 8558: @cindex I/O - see character strings
1.75 anton 8559: @cindex counted strings
8560:
8561: @c anton: this does not really belong here; maybe the memory section,
8562: @c or the principles chapter
1.26 crook 8563:
1.27 crook 8564: Forth commonly uses two different methods for representing character
8565: strings:
1.26 crook 8566:
8567: @itemize @bullet
8568: @item
8569: @cindex address of counted string
1.45 crook 8570: @cindex counted string
1.29 crook 8571: As a @dfn{counted string}, represented by a @i{c-addr}. The char
8572: addressed by @i{c-addr} contains a character-count, @i{n}, of the
8573: string and the string occupies the subsequent @i{n} char addresses in
1.26 crook 8574: memory.
8575: @item
1.29 crook 8576: As cell pair on the stack; @i{c-addr u}, where @i{u} is the length
8577: of the string in characters, and @i{c-addr} is the address of the
1.26 crook 8578: first byte of the string.
8579: @end itemize
8580:
8581: ANS Forth encourages the use of the second format when representing
1.75 anton 8582: strings.
1.26 crook 8583:
1.44 crook 8584:
1.26 crook 8585: doc-count
8586:
1.44 crook 8587:
1.49 anton 8588: For words that move, copy and search for strings see @ref{Memory
8589: Blocks}. For words that display characters and strings see
8590: @ref{Displaying characters and strings}.
1.26 crook 8591:
8592: @node Displaying characters and strings, Input, String Formats, Other I/O
8593: @subsection Displaying characters and strings
1.27 crook 8594: @cindex characters - compiling and displaying
8595: @cindex character strings - compiling and displaying
1.26 crook 8596:
8597: This section starts with a glossary of Forth words and ends with a set
8598: of examples.
8599:
1.44 crook 8600:
1.26 crook 8601: doc-bl
8602: doc-space
8603: doc-spaces
8604: doc-emit
8605: doc-toupper
8606: doc-."
8607: doc-.(
8608: doc-type
1.44 crook 8609: doc-typewhite
1.26 crook 8610: doc-cr
1.27 crook 8611: @cindex cursor control
1.26 crook 8612: doc-at-xy
8613: doc-page
8614: doc-s"
8615: doc-c"
8616: doc-char
8617: doc-[char]
8618: doc-sliteral
8619:
1.44 crook 8620:
8621: @noindent
1.26 crook 8622: As an example, consider the following text, stored in a file @file{test.fs}:
1.5 anton 8623:
8624: @example
1.26 crook 8625: .( text-1)
8626: : my-word
8627: ." text-2" cr
8628: .( text-3)
8629: ;
8630:
8631: ." text-4"
8632:
8633: : my-char
8634: [char] ALPHABET emit
8635: char emit
8636: ;
1.5 anton 8637: @end example
8638:
1.26 crook 8639: When you load this code into Gforth, the following output is generated:
1.5 anton 8640:
1.26 crook 8641: @example
1.30 anton 8642: @kbd{include test.fs @key{RET}} text-1text-3text-4 ok
1.26 crook 8643: @end example
1.5 anton 8644:
1.26 crook 8645: @itemize @bullet
8646: @item
8647: Messages @code{text-1} and @code{text-3} are displayed because @code{.(}
8648: is an immediate word; it behaves in the same way whether it is used inside
8649: or outside a colon definition.
8650: @item
8651: Message @code{text-4} is displayed because of Gforth's added interpretation
8652: semantics for @code{."}.
8653: @item
1.29 crook 8654: Message @code{text-2} is @i{not} displayed, because the text interpreter
1.26 crook 8655: performs the compilation semantics for @code{."} within the definition of
8656: @code{my-word}.
8657: @end itemize
1.5 anton 8658:
1.26 crook 8659: Here are some examples of executing @code{my-word} and @code{my-char}:
1.5 anton 8660:
1.26 crook 8661: @example
1.30 anton 8662: @kbd{my-word @key{RET}} text-2
1.26 crook 8663: ok
1.30 anton 8664: @kbd{my-char fred @key{RET}} Af ok
8665: @kbd{my-char jim @key{RET}} Aj ok
1.26 crook 8666: @end example
1.5 anton 8667:
8668: @itemize @bullet
8669: @item
1.26 crook 8670: Message @code{text-2} is displayed because of the run-time behaviour of
8671: @code{."}.
8672: @item
8673: @code{[char]} compiles the ``A'' from ``ALPHABET'' and puts its display code
8674: on the stack at run-time. @code{emit} always displays the character
8675: when @code{my-char} is executed.
8676: @item
8677: @code{char} parses a string at run-time and the second @code{emit} displays
8678: the first character of the string.
1.5 anton 8679: @item
1.26 crook 8680: If you type @code{see my-char} you can see that @code{[char]} discarded
8681: the text ``LPHABET'' and only compiled the display code for ``A'' into the
8682: definition of @code{my-char}.
1.5 anton 8683: @end itemize
8684:
8685:
8686:
1.48 anton 8687: @node Input, , Displaying characters and strings, Other I/O
1.26 crook 8688: @subsection Input
8689: @cindex input
1.28 crook 8690: @cindex I/O - see input
8691: @cindex parsing a string
1.5 anton 8692:
1.49 anton 8693: For ways of storing character strings in memory see @ref{String Formats}.
1.5 anton 8694:
1.27 crook 8695: @comment TODO examples for >number >float accept key key? pad parse word refill
1.29 crook 8696: @comment then index them
1.27 crook 8697:
1.44 crook 8698:
1.27 crook 8699: doc-key
8700: doc-key?
1.45 crook 8701: doc-ekey
8702: doc-ekey?
8703: doc-ekey>char
1.26 crook 8704: doc->number
8705: doc->float
8706: doc-accept
1.27 crook 8707: doc-pad
1.75 anton 8708: @c anton: these belong in the input stream section
1.27 crook 8709: doc-parse
8710: doc-word
8711: doc-sword
1.75 anton 8712: doc-name
1.27 crook 8713: doc-refill
8714: @comment obsolescent words..
8715: doc-convert
1.26 crook 8716: doc-query
8717: doc-expect
1.27 crook 8718: doc-span
1.5 anton 8719:
8720:
1.78 anton 8721: @c -------------------------------------------------------------
8722: @node Locals, Structures, Other I/O, Words
8723: @section Locals
8724: @cindex locals
8725:
8726: Local variables can make Forth programming more enjoyable and Forth
8727: programs easier to read. Unfortunately, the locals of ANS Forth are
8728: laden with restrictions. Therefore, we provide not only the ANS Forth
8729: locals wordset, but also our own, more powerful locals wordset (we
8730: implemented the ANS Forth locals wordset through our locals wordset).
1.44 crook 8731:
1.78 anton 8732: The ideas in this section have also been published in M. Anton Ertl,
8733: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl94l.ps.gz,
8734: Automatic Scoping of Local Variables}}, EuroForth '94.
1.12 anton 8735:
8736: @menu
1.78 anton 8737: * Gforth locals::
8738: * ANS Forth locals::
1.5 anton 8739: @end menu
8740:
1.78 anton 8741: @node Gforth locals, ANS Forth locals, Locals, Locals
8742: @subsection Gforth locals
8743: @cindex Gforth locals
8744: @cindex locals, Gforth style
1.5 anton 8745:
1.78 anton 8746: Locals can be defined with
1.44 crook 8747:
1.78 anton 8748: @example
8749: @{ local1 local2 ... -- comment @}
8750: @end example
8751: or
8752: @example
8753: @{ local1 local2 ... @}
8754: @end example
1.5 anton 8755:
1.78 anton 8756: E.g.,
8757: @example
8758: : max @{ n1 n2 -- n3 @}
8759: n1 n2 > if
8760: n1
8761: else
8762: n2
8763: endif ;
8764: @end example
1.44 crook 8765:
1.78 anton 8766: The similarity of locals definitions with stack comments is intended. A
8767: locals definition often replaces the stack comment of a word. The order
8768: of the locals corresponds to the order in a stack comment and everything
8769: after the @code{--} is really a comment.
1.77 anton 8770:
1.78 anton 8771: This similarity has one disadvantage: It is too easy to confuse locals
8772: declarations with stack comments, causing bugs and making them hard to
8773: find. However, this problem can be avoided by appropriate coding
8774: conventions: Do not use both notations in the same program. If you do,
8775: they should be distinguished using additional means, e.g. by position.
1.77 anton 8776:
1.78 anton 8777: @cindex types of locals
8778: @cindex locals types
8779: The name of the local may be preceded by a type specifier, e.g.,
8780: @code{F:} for a floating point value:
1.5 anton 8781:
1.78 anton 8782: @example
8783: : CX* @{ F: Ar F: Ai F: Br F: Bi -- Cr Ci @}
8784: \ complex multiplication
8785: Ar Br f* Ai Bi f* f-
8786: Ar Bi f* Ai Br f* f+ ;
8787: @end example
1.44 crook 8788:
1.78 anton 8789: @cindex flavours of locals
8790: @cindex locals flavours
8791: @cindex value-flavoured locals
8792: @cindex variable-flavoured locals
8793: Gforth currently supports cells (@code{W:}, @code{W^}), doubles
8794: (@code{D:}, @code{D^}), floats (@code{F:}, @code{F^}) and characters
8795: (@code{C:}, @code{C^}) in two flavours: a value-flavoured local (defined
8796: with @code{W:}, @code{D:} etc.) produces its value and can be changed
8797: with @code{TO}. A variable-flavoured local (defined with @code{W^} etc.)
8798: produces its address (which becomes invalid when the variable's scope is
8799: left). E.g., the standard word @code{emit} can be defined in terms of
8800: @code{type} like this:
1.5 anton 8801:
1.78 anton 8802: @example
8803: : emit @{ C^ char* -- @}
8804: char* 1 type ;
8805: @end example
1.5 anton 8806:
1.78 anton 8807: @cindex default type of locals
8808: @cindex locals, default type
8809: A local without type specifier is a @code{W:} local. Both flavours of
8810: locals are initialized with values from the data or FP stack.
1.44 crook 8811:
1.78 anton 8812: Currently there is no way to define locals with user-defined data
8813: structures, but we are working on it.
1.5 anton 8814:
1.78 anton 8815: Gforth allows defining locals everywhere in a colon definition. This
8816: poses the following questions:
1.5 anton 8817:
1.78 anton 8818: @menu
8819: * Where are locals visible by name?::
8820: * How long do locals live?::
8821: * Locals programming style::
8822: * Locals implementation::
8823: @end menu
1.44 crook 8824:
1.78 anton 8825: @node Where are locals visible by name?, How long do locals live?, Gforth locals, Gforth locals
8826: @subsubsection Where are locals visible by name?
8827: @cindex locals visibility
8828: @cindex visibility of locals
8829: @cindex scope of locals
1.5 anton 8830:
1.78 anton 8831: Basically, the answer is that locals are visible where you would expect
8832: it in block-structured languages, and sometimes a little longer. If you
8833: want to restrict the scope of a local, enclose its definition in
8834: @code{SCOPE}...@code{ENDSCOPE}.
1.5 anton 8835:
8836:
1.78 anton 8837: doc-scope
8838: doc-endscope
1.5 anton 8839:
8840:
1.78 anton 8841: These words behave like control structure words, so you can use them
8842: with @code{CS-PICK} and @code{CS-ROLL} to restrict the scope in
8843: arbitrary ways.
1.77 anton 8844:
1.78 anton 8845: If you want a more exact answer to the visibility question, here's the
8846: basic principle: A local is visible in all places that can only be
8847: reached through the definition of the local@footnote{In compiler
8848: construction terminology, all places dominated by the definition of the
8849: local.}. In other words, it is not visible in places that can be reached
8850: without going through the definition of the local. E.g., locals defined
8851: in @code{IF}...@code{ENDIF} are visible until the @code{ENDIF}, locals
8852: defined in @code{BEGIN}...@code{UNTIL} are visible after the
8853: @code{UNTIL} (until, e.g., a subsequent @code{ENDSCOPE}).
1.77 anton 8854:
1.78 anton 8855: The reasoning behind this solution is: We want to have the locals
8856: visible as long as it is meaningful. The user can always make the
8857: visibility shorter by using explicit scoping. In a place that can
8858: only be reached through the definition of a local, the meaning of a
8859: local name is clear. In other places it is not: How is the local
8860: initialized at the control flow path that does not contain the
8861: definition? Which local is meant, if the same name is defined twice in
8862: two independent control flow paths?
1.77 anton 8863:
1.78 anton 8864: This should be enough detail for nearly all users, so you can skip the
8865: rest of this section. If you really must know all the gory details and
8866: options, read on.
1.77 anton 8867:
1.78 anton 8868: In order to implement this rule, the compiler has to know which places
8869: are unreachable. It knows this automatically after @code{AHEAD},
8870: @code{AGAIN}, @code{EXIT} and @code{LEAVE}; in other cases (e.g., after
8871: most @code{THROW}s), you can use the word @code{UNREACHABLE} to tell the
8872: compiler that the control flow never reaches that place. If
8873: @code{UNREACHABLE} is not used where it could, the only consequence is
8874: that the visibility of some locals is more limited than the rule above
8875: says. If @code{UNREACHABLE} is used where it should not (i.e., if you
8876: lie to the compiler), buggy code will be produced.
1.77 anton 8877:
1.5 anton 8878:
1.78 anton 8879: doc-unreachable
1.5 anton 8880:
1.23 crook 8881:
1.78 anton 8882: Another problem with this rule is that at @code{BEGIN}, the compiler
8883: does not know which locals will be visible on the incoming
8884: back-edge. All problems discussed in the following are due to this
8885: ignorance of the compiler (we discuss the problems using @code{BEGIN}
8886: loops as examples; the discussion also applies to @code{?DO} and other
8887: loops). Perhaps the most insidious example is:
1.26 crook 8888: @example
1.78 anton 8889: AHEAD
8890: BEGIN
8891: x
8892: [ 1 CS-ROLL ] THEN
8893: @{ x @}
8894: ...
8895: UNTIL
1.26 crook 8896: @end example
1.23 crook 8897:
1.78 anton 8898: This should be legal according to the visibility rule. The use of
8899: @code{x} can only be reached through the definition; but that appears
8900: textually below the use.
8901:
8902: From this example it is clear that the visibility rules cannot be fully
8903: implemented without major headaches. Our implementation treats common
8904: cases as advertised and the exceptions are treated in a safe way: The
8905: compiler makes a reasonable guess about the locals visible after a
8906: @code{BEGIN}; if it is too pessimistic, the
8907: user will get a spurious error about the local not being defined; if the
8908: compiler is too optimistic, it will notice this later and issue a
8909: warning. In the case above the compiler would complain about @code{x}
8910: being undefined at its use. You can see from the obscure examples in
8911: this section that it takes quite unusual control structures to get the
8912: compiler into trouble, and even then it will often do fine.
1.23 crook 8913:
1.78 anton 8914: If the @code{BEGIN} is reachable from above, the most optimistic guess
8915: is that all locals visible before the @code{BEGIN} will also be
8916: visible after the @code{BEGIN}. This guess is valid for all loops that
8917: are entered only through the @code{BEGIN}, in particular, for normal
8918: @code{BEGIN}...@code{WHILE}...@code{REPEAT} and
8919: @code{BEGIN}...@code{UNTIL} loops and it is implemented in our
8920: compiler. When the branch to the @code{BEGIN} is finally generated by
8921: @code{AGAIN} or @code{UNTIL}, the compiler checks the guess and
8922: warns the user if it was too optimistic:
1.26 crook 8923: @example
1.78 anton 8924: IF
8925: @{ x @}
8926: BEGIN
8927: \ x ?
8928: [ 1 cs-roll ] THEN
8929: ...
8930: UNTIL
1.26 crook 8931: @end example
1.23 crook 8932:
1.78 anton 8933: Here, @code{x} lives only until the @code{BEGIN}, but the compiler
8934: optimistically assumes that it lives until the @code{THEN}. It notices
8935: this difference when it compiles the @code{UNTIL} and issues a
8936: warning. The user can avoid the warning, and make sure that @code{x}
8937: is not used in the wrong area by using explicit scoping:
8938: @example
8939: IF
8940: SCOPE
8941: @{ x @}
8942: ENDSCOPE
8943: BEGIN
8944: [ 1 cs-roll ] THEN
8945: ...
8946: UNTIL
8947: @end example
1.23 crook 8948:
1.78 anton 8949: Since the guess is optimistic, there will be no spurious error messages
8950: about undefined locals.
1.44 crook 8951:
1.78 anton 8952: If the @code{BEGIN} is not reachable from above (e.g., after
8953: @code{AHEAD} or @code{EXIT}), the compiler cannot even make an
8954: optimistic guess, as the locals visible after the @code{BEGIN} may be
8955: defined later. Therefore, the compiler assumes that no locals are
8956: visible after the @code{BEGIN}. However, the user can use
8957: @code{ASSUME-LIVE} to make the compiler assume that the same locals are
8958: visible at the BEGIN as at the point where the top control-flow stack
8959: item was created.
1.23 crook 8960:
1.44 crook 8961:
1.78 anton 8962: doc-assume-live
1.26 crook 8963:
1.23 crook 8964:
1.78 anton 8965: @noindent
8966: E.g.,
8967: @example
8968: @{ x @}
8969: AHEAD
8970: ASSUME-LIVE
8971: BEGIN
8972: x
8973: [ 1 CS-ROLL ] THEN
8974: ...
8975: UNTIL
8976: @end example
1.44 crook 8977:
1.78 anton 8978: Other cases where the locals are defined before the @code{BEGIN} can be
8979: handled by inserting an appropriate @code{CS-ROLL} before the
8980: @code{ASSUME-LIVE} (and changing the control-flow stack manipulation
8981: behind the @code{ASSUME-LIVE}).
1.23 crook 8982:
1.78 anton 8983: Cases where locals are defined after the @code{BEGIN} (but should be
8984: visible immediately after the @code{BEGIN}) can only be handled by
8985: rearranging the loop. E.g., the ``most insidious'' example above can be
8986: arranged into:
8987: @example
8988: BEGIN
8989: @{ x @}
8990: ... 0=
8991: WHILE
8992: x
8993: REPEAT
8994: @end example
1.44 crook 8995:
1.78 anton 8996: @node How long do locals live?, Locals programming style, Where are locals visible by name?, Gforth locals
8997: @subsubsection How long do locals live?
8998: @cindex locals lifetime
8999: @cindex lifetime of locals
1.23 crook 9000:
1.78 anton 9001: The right answer for the lifetime question would be: A local lives at
9002: least as long as it can be accessed. For a value-flavoured local this
9003: means: until the end of its visibility. However, a variable-flavoured
9004: local could be accessed through its address far beyond its visibility
9005: scope. Ultimately, this would mean that such locals would have to be
9006: garbage collected. Since this entails un-Forth-like implementation
9007: complexities, I adopted the same cowardly solution as some other
9008: languages (e.g., C): The local lives only as long as it is visible;
9009: afterwards its address is invalid (and programs that access it
9010: afterwards are erroneous).
1.23 crook 9011:
1.78 anton 9012: @node Locals programming style, Locals implementation, How long do locals live?, Gforth locals
9013: @subsubsection Locals programming style
9014: @cindex locals programming style
9015: @cindex programming style, locals
1.23 crook 9016:
1.78 anton 9017: The freedom to define locals anywhere has the potential to change
9018: programming styles dramatically. In particular, the need to use the
9019: return stack for intermediate storage vanishes. Moreover, all stack
9020: manipulations (except @code{PICK}s and @code{ROLL}s with run-time
9021: determined arguments) can be eliminated: If the stack items are in the
9022: wrong order, just write a locals definition for all of them; then
9023: write the items in the order you want.
1.23 crook 9024:
1.78 anton 9025: This seems a little far-fetched and eliminating stack manipulations is
9026: unlikely to become a conscious programming objective. Still, the number
9027: of stack manipulations will be reduced dramatically if local variables
9028: are used liberally (e.g., compare @code{max} (@pxref{Gforth locals}) with
9029: a traditional implementation of @code{max}).
1.23 crook 9030:
1.78 anton 9031: This shows one potential benefit of locals: making Forth programs more
9032: readable. Of course, this benefit will only be realized if the
9033: programmers continue to honour the principle of factoring instead of
9034: using the added latitude to make the words longer.
1.23 crook 9035:
1.78 anton 9036: @cindex single-assignment style for locals
9037: Using @code{TO} can and should be avoided. Without @code{TO},
9038: every value-flavoured local has only a single assignment and many
9039: advantages of functional languages apply to Forth. I.e., programs are
9040: easier to analyse, to optimize and to read: It is clear from the
9041: definition what the local stands for, it does not turn into something
9042: different later.
1.23 crook 9043:
1.78 anton 9044: E.g., a definition using @code{TO} might look like this:
9045: @example
9046: : strcmp @{ addr1 u1 addr2 u2 -- n @}
9047: u1 u2 min 0
9048: ?do
9049: addr1 c@@ addr2 c@@ -
9050: ?dup-if
9051: unloop exit
9052: then
9053: addr1 char+ TO addr1
9054: addr2 char+ TO addr2
9055: loop
9056: u1 u2 - ;
1.26 crook 9057: @end example
1.78 anton 9058: Here, @code{TO} is used to update @code{addr1} and @code{addr2} at
9059: every loop iteration. @code{strcmp} is a typical example of the
9060: readability problems of using @code{TO}. When you start reading
9061: @code{strcmp}, you think that @code{addr1} refers to the start of the
9062: string. Only near the end of the loop you realize that it is something
9063: else.
1.23 crook 9064:
1.78 anton 9065: This can be avoided by defining two locals at the start of the loop that
9066: are initialized with the right value for the current iteration.
9067: @example
9068: : strcmp @{ addr1 u1 addr2 u2 -- n @}
9069: addr1 addr2
9070: u1 u2 min 0
9071: ?do @{ s1 s2 @}
9072: s1 c@@ s2 c@@ -
9073: ?dup-if
9074: unloop exit
9075: then
9076: s1 char+ s2 char+
9077: loop
9078: 2drop
9079: u1 u2 - ;
9080: @end example
9081: Here it is clear from the start that @code{s1} has a different value
9082: in every loop iteration.
1.23 crook 9083:
1.78 anton 9084: @node Locals implementation, , Locals programming style, Gforth locals
9085: @subsubsection Locals implementation
9086: @cindex locals implementation
9087: @cindex implementation of locals
1.23 crook 9088:
1.78 anton 9089: @cindex locals stack
9090: Gforth uses an extra locals stack. The most compelling reason for
9091: this is that the return stack is not float-aligned; using an extra stack
9092: also eliminates the problems and restrictions of using the return stack
9093: as locals stack. Like the other stacks, the locals stack grows toward
9094: lower addresses. A few primitives allow an efficient implementation:
9095:
9096:
9097: doc-@local#
9098: doc-f@local#
9099: doc-laddr#
9100: doc-lp+!#
9101: doc-lp!
9102: doc->l
9103: doc-f>l
9104:
9105:
9106: In addition to these primitives, some specializations of these
9107: primitives for commonly occurring inline arguments are provided for
9108: efficiency reasons, e.g., @code{@@local0} as specialization of
9109: @code{@@local#} for the inline argument 0. The following compiling words
9110: compile the right specialized version, or the general version, as
9111: appropriate:
1.23 crook 9112:
1.5 anton 9113:
1.78 anton 9114: doc-compile-@local
9115: doc-compile-f@local
9116: doc-compile-lp+!
1.5 anton 9117:
9118:
1.78 anton 9119: Combinations of conditional branches and @code{lp+!#} like
9120: @code{?branch-lp+!#} (the locals pointer is only changed if the branch
9121: is taken) are provided for efficiency and correctness in loops.
1.5 anton 9122:
1.78 anton 9123: A special area in the dictionary space is reserved for keeping the
9124: local variable names. @code{@{} switches the dictionary pointer to this
9125: area and @code{@}} switches it back and generates the locals
9126: initializing code. @code{W:} etc.@ are normal defining words. This
9127: special area is cleared at the start of every colon definition.
1.5 anton 9128:
1.78 anton 9129: @cindex word list for defining locals
9130: A special feature of Gforth's dictionary is used to implement the
9131: definition of locals without type specifiers: every word list (aka
9132: vocabulary) has its own methods for searching
9133: etc. (@pxref{Word Lists}). For the present purpose we defined a word list
9134: with a special search method: When it is searched for a word, it
9135: actually creates that word using @code{W:}. @code{@{} changes the search
9136: order to first search the word list containing @code{@}}, @code{W:} etc.,
9137: and then the word list for defining locals without type specifiers.
1.5 anton 9138:
1.78 anton 9139: The lifetime rules support a stack discipline within a colon
9140: definition: The lifetime of a local is either nested with other locals
9141: lifetimes or it does not overlap them.
1.23 crook 9142:
1.78 anton 9143: At @code{BEGIN}, @code{IF}, and @code{AHEAD} no code for locals stack
9144: pointer manipulation is generated. Between control structure words
9145: locals definitions can push locals onto the locals stack. @code{AGAIN}
9146: is the simplest of the other three control flow words. It has to
9147: restore the locals stack depth of the corresponding @code{BEGIN}
9148: before branching. The code looks like this:
9149: @format
9150: @code{lp+!#} current-locals-size @minus{} dest-locals-size
9151: @code{branch} <begin>
9152: @end format
1.26 crook 9153:
1.78 anton 9154: @code{UNTIL} is a little more complicated: If it branches back, it
9155: must adjust the stack just like @code{AGAIN}. But if it falls through,
9156: the locals stack must not be changed. The compiler generates the
9157: following code:
9158: @format
9159: @code{?branch-lp+!#} <begin> current-locals-size @minus{} dest-locals-size
9160: @end format
9161: The locals stack pointer is only adjusted if the branch is taken.
1.44 crook 9162:
1.78 anton 9163: @code{THEN} can produce somewhat inefficient code:
9164: @format
9165: @code{lp+!#} current-locals-size @minus{} orig-locals-size
9166: <orig target>:
9167: @code{lp+!#} orig-locals-size @minus{} new-locals-size
9168: @end format
9169: The second @code{lp+!#} adjusts the locals stack pointer from the
9170: level at the @i{orig} point to the level after the @code{THEN}. The
9171: first @code{lp+!#} adjusts the locals stack pointer from the current
9172: level to the level at the orig point, so the complete effect is an
9173: adjustment from the current level to the right level after the
9174: @code{THEN}.
1.26 crook 9175:
1.78 anton 9176: @cindex locals information on the control-flow stack
9177: @cindex control-flow stack items, locals information
9178: In a conventional Forth implementation a dest control-flow stack entry
9179: is just the target address and an orig entry is just the address to be
9180: patched. Our locals implementation adds a word list to every orig or dest
9181: item. It is the list of locals visible (or assumed visible) at the point
9182: described by the entry. Our implementation also adds a tag to identify
9183: the kind of entry, in particular to differentiate between live and dead
9184: (reachable and unreachable) orig entries.
1.26 crook 9185:
1.78 anton 9186: A few unusual operations have to be performed on locals word lists:
1.44 crook 9187:
1.5 anton 9188:
1.78 anton 9189: doc-common-list
9190: doc-sub-list?
9191: doc-list-size
1.52 anton 9192:
9193:
1.78 anton 9194: Several features of our locals word list implementation make these
9195: operations easy to implement: The locals word lists are organised as
9196: linked lists; the tails of these lists are shared, if the lists
9197: contain some of the same locals; and the address of a name is greater
9198: than the address of the names behind it in the list.
1.5 anton 9199:
1.78 anton 9200: Another important implementation detail is the variable
9201: @code{dead-code}. It is used by @code{BEGIN} and @code{THEN} to
9202: determine if they can be reached directly or only through the branch
9203: that they resolve. @code{dead-code} is set by @code{UNREACHABLE},
9204: @code{AHEAD}, @code{EXIT} etc., and cleared at the start of a colon
9205: definition, by @code{BEGIN} and usually by @code{THEN}.
1.5 anton 9206:
1.78 anton 9207: Counted loops are similar to other loops in most respects, but
9208: @code{LEAVE} requires special attention: It performs basically the same
9209: service as @code{AHEAD}, but it does not create a control-flow stack
9210: entry. Therefore the information has to be stored elsewhere;
9211: traditionally, the information was stored in the target fields of the
9212: branches created by the @code{LEAVE}s, by organizing these fields into a
9213: linked list. Unfortunately, this clever trick does not provide enough
9214: space for storing our extended control flow information. Therefore, we
9215: introduce another stack, the leave stack. It contains the control-flow
9216: stack entries for all unresolved @code{LEAVE}s.
1.44 crook 9217:
1.78 anton 9218: Local names are kept until the end of the colon definition, even if
9219: they are no longer visible in any control-flow path. In a few cases
9220: this may lead to increased space needs for the locals name area, but
9221: usually less than reclaiming this space would cost in code size.
1.5 anton 9222:
1.44 crook 9223:
1.78 anton 9224: @node ANS Forth locals, , Gforth locals, Locals
9225: @subsection ANS Forth locals
9226: @cindex locals, ANS Forth style
1.5 anton 9227:
1.78 anton 9228: The ANS Forth locals wordset does not define a syntax for locals, but
9229: words that make it possible to define various syntaxes. One of the
9230: possible syntaxes is a subset of the syntax we used in the Gforth locals
9231: wordset, i.e.:
1.29 crook 9232:
9233: @example
1.78 anton 9234: @{ local1 local2 ... -- comment @}
9235: @end example
9236: @noindent
9237: or
9238: @example
9239: @{ local1 local2 ... @}
1.29 crook 9240: @end example
9241:
1.78 anton 9242: The order of the locals corresponds to the order in a stack comment. The
9243: restrictions are:
1.5 anton 9244:
1.78 anton 9245: @itemize @bullet
9246: @item
9247: Locals can only be cell-sized values (no type specifiers are allowed).
9248: @item
9249: Locals can be defined only outside control structures.
9250: @item
9251: Locals can interfere with explicit usage of the return stack. For the
9252: exact (and long) rules, see the standard. If you don't use return stack
9253: accessing words in a definition using locals, you will be all right. The
9254: purpose of this rule is to make locals implementation on the return
9255: stack easier.
9256: @item
9257: The whole definition must be in one line.
9258: @end itemize
1.5 anton 9259:
1.78 anton 9260: Locals defined in ANS Forth behave like @code{VALUE}s
9261: (@pxref{Values}). I.e., they are initialized from the stack. Using their
9262: name produces their value. Their value can be changed using @code{TO}.
1.77 anton 9263:
1.78 anton 9264: Since the syntax above is supported by Gforth directly, you need not do
9265: anything to use it. If you want to port a program using this syntax to
9266: another ANS Forth system, use @file{compat/anslocal.fs} to implement the
9267: syntax on the other system.
1.5 anton 9268:
1.78 anton 9269: Note that a syntax shown in the standard, section A.13 looks
9270: similar, but is quite different in having the order of locals
9271: reversed. Beware!
1.5 anton 9272:
1.78 anton 9273: The ANS Forth locals wordset itself consists of one word:
1.5 anton 9274:
1.78 anton 9275: doc-(local)
1.5 anton 9276:
1.78 anton 9277: The ANS Forth locals extension wordset defines a syntax using
9278: @code{locals|}, but it is so awful that we strongly recommend not to use
9279: it. We have implemented this syntax to make porting to Gforth easy, but
9280: do not document it here. The problem with this syntax is that the locals
9281: are defined in an order reversed with respect to the standard stack
9282: comment notation, making programs harder to read, and easier to misread
9283: and miswrite. The only merit of this syntax is that it is easy to
9284: implement using the ANS Forth locals wordset.
1.53 anton 9285:
9286:
1.78 anton 9287: @c ----------------------------------------------------------
9288: @node Structures, Object-oriented Forth, Locals, Words
9289: @section Structures
9290: @cindex structures
9291: @cindex records
1.53 anton 9292:
1.78 anton 9293: This section presents the structure package that comes with Gforth. A
9294: version of the package implemented in ANS Forth is available in
9295: @file{compat/struct.fs}. This package was inspired by a posting on
9296: comp.lang.forth in 1989 (unfortunately I don't remember, by whom;
9297: possibly John Hayes). A version of this section has been published in
9298: M. Anton Ertl,
9299: @uref{http://www.complang.tuwien.ac.at/forth/objects/structs.html, Yet
9300: Another Forth Structures Package}, Forth Dimensions 19(3), pages
9301: 13--16. Marcel Hendrix provided helpful comments.
1.53 anton 9302:
1.78 anton 9303: @menu
9304: * Why explicit structure support?::
9305: * Structure Usage::
9306: * Structure Naming Convention::
9307: * Structure Implementation::
9308: * Structure Glossary::
9309: @end menu
1.55 anton 9310:
1.78 anton 9311: @node Why explicit structure support?, Structure Usage, Structures, Structures
9312: @subsection Why explicit structure support?
1.53 anton 9313:
1.78 anton 9314: @cindex address arithmetic for structures
9315: @cindex structures using address arithmetic
9316: If we want to use a structure containing several fields, we could simply
9317: reserve memory for it, and access the fields using address arithmetic
9318: (@pxref{Address arithmetic}). As an example, consider a structure with
9319: the following fields
1.57 anton 9320:
1.78 anton 9321: @table @code
9322: @item a
9323: is a float
9324: @item b
9325: is a cell
9326: @item c
9327: is a float
9328: @end table
1.57 anton 9329:
1.78 anton 9330: Given the (float-aligned) base address of the structure we get the
9331: address of the field
1.52 anton 9332:
1.78 anton 9333: @table @code
9334: @item a
9335: without doing anything further.
9336: @item b
9337: with @code{float+}
9338: @item c
9339: with @code{float+ cell+ faligned}
9340: @end table
1.52 anton 9341:
1.78 anton 9342: It is easy to see that this can become quite tiring.
1.52 anton 9343:
1.78 anton 9344: Moreover, it is not very readable, because seeing a
9345: @code{cell+} tells us neither which kind of structure is
9346: accessed nor what field is accessed; we have to somehow infer the kind
9347: of structure, and then look up in the documentation, which field of
9348: that structure corresponds to that offset.
1.53 anton 9349:
1.78 anton 9350: Finally, this kind of address arithmetic also causes maintenance
9351: troubles: If you add or delete a field somewhere in the middle of the
9352: structure, you have to find and change all computations for the fields
9353: afterwards.
1.52 anton 9354:
1.78 anton 9355: So, instead of using @code{cell+} and friends directly, how
9356: about storing the offsets in constants:
1.52 anton 9357:
1.78 anton 9358: @example
9359: 0 constant a-offset
9360: 0 float+ constant b-offset
9361: 0 float+ cell+ faligned c-offset
9362: @end example
1.64 pazsan 9363:
1.78 anton 9364: Now we can get the address of field @code{x} with @code{x-offset
9365: +}. This is much better in all respects. Of course, you still
9366: have to change all later offset definitions if you add a field. You can
9367: fix this by declaring the offsets in the following way:
1.57 anton 9368:
1.78 anton 9369: @example
9370: 0 constant a-offset
9371: a-offset float+ constant b-offset
9372: b-offset cell+ faligned constant c-offset
9373: @end example
1.57 anton 9374:
1.78 anton 9375: Since we always use the offsets with @code{+}, we could use a defining
9376: word @code{cfield} that includes the @code{+} in the action of the
9377: defined word:
1.64 pazsan 9378:
1.78 anton 9379: @example
9380: : cfield ( n "name" -- )
9381: create ,
9382: does> ( name execution: addr1 -- addr2 )
9383: @@ + ;
1.64 pazsan 9384:
1.78 anton 9385: 0 cfield a
9386: 0 a float+ cfield b
9387: 0 b cell+ faligned cfield c
9388: @end example
1.64 pazsan 9389:
1.78 anton 9390: Instead of @code{x-offset +}, we now simply write @code{x}.
1.64 pazsan 9391:
1.78 anton 9392: The structure field words now can be used quite nicely. However,
9393: their definition is still a bit cumbersome: We have to repeat the
9394: name, the information about size and alignment is distributed before
9395: and after the field definitions etc. The structure package presented
9396: here addresses these problems.
1.64 pazsan 9397:
1.78 anton 9398: @node Structure Usage, Structure Naming Convention, Why explicit structure support?, Structures
9399: @subsection Structure Usage
9400: @cindex structure usage
1.57 anton 9401:
1.78 anton 9402: @cindex @code{field} usage
9403: @cindex @code{struct} usage
9404: @cindex @code{end-struct} usage
9405: You can define a structure for a (data-less) linked list with:
1.57 anton 9406: @example
1.78 anton 9407: struct
9408: cell% field list-next
9409: end-struct list%
1.57 anton 9410: @end example
9411:
1.78 anton 9412: With the address of the list node on the stack, you can compute the
9413: address of the field that contains the address of the next node with
9414: @code{list-next}. E.g., you can determine the length of a list
9415: with:
1.57 anton 9416:
9417: @example
1.78 anton 9418: : list-length ( list -- n )
9419: \ "list" is a pointer to the first element of a linked list
9420: \ "n" is the length of the list
9421: 0 BEGIN ( list1 n1 )
9422: over
9423: WHILE ( list1 n1 )
9424: 1+ swap list-next @@ swap
9425: REPEAT
9426: nip ;
1.57 anton 9427: @end example
9428:
1.78 anton 9429: You can reserve memory for a list node in the dictionary with
9430: @code{list% %allot}, which leaves the address of the list node on the
9431: stack. For the equivalent allocation on the heap you can use @code{list%
9432: %alloc} (or, for an @code{allocate}-like stack effect (i.e., with ior),
9433: use @code{list% %allocate}). You can get the the size of a list
9434: node with @code{list% %size} and its alignment with @code{list%
9435: %alignment}.
9436:
9437: Note that in ANS Forth the body of a @code{create}d word is
9438: @code{aligned} but not necessarily @code{faligned};
9439: therefore, if you do a:
1.57 anton 9440:
9441: @example
1.78 anton 9442: create @emph{name} foo% %allot drop
1.57 anton 9443: @end example
9444:
1.78 anton 9445: @noindent
9446: then the memory alloted for @code{foo%} is guaranteed to start at the
9447: body of @code{@emph{name}} only if @code{foo%} contains only character,
9448: cell and double fields. Therefore, if your structure contains floats,
9449: better use
1.57 anton 9450:
9451: @example
1.78 anton 9452: foo% %allot constant @emph{name}
1.57 anton 9453: @end example
9454:
1.78 anton 9455: @cindex structures containing structures
9456: You can include a structure @code{foo%} as a field of
9457: another structure, like this:
1.65 anton 9458: @example
1.78 anton 9459: struct
9460: ...
9461: foo% field ...
9462: ...
9463: end-struct ...
1.65 anton 9464: @end example
1.52 anton 9465:
1.78 anton 9466: @cindex structure extension
9467: @cindex extended records
9468: Instead of starting with an empty structure, you can extend an
9469: existing structure. E.g., a plain linked list without data, as defined
9470: above, is hardly useful; You can extend it to a linked list of integers,
9471: like this:@footnote{This feature is also known as @emph{extended
9472: records}. It is the main innovation in the Oberon language; in other
9473: words, adding this feature to Modula-2 led Wirth to create a new
9474: language, write a new compiler etc. Adding this feature to Forth just
9475: required a few lines of code.}
1.52 anton 9476:
1.78 anton 9477: @example
9478: list%
9479: cell% field intlist-int
9480: end-struct intlist%
9481: @end example
1.55 anton 9482:
1.78 anton 9483: @code{intlist%} is a structure with two fields:
9484: @code{list-next} and @code{intlist-int}.
1.55 anton 9485:
1.78 anton 9486: @cindex structures containing arrays
9487: You can specify an array type containing @emph{n} elements of
9488: type @code{foo%} like this:
1.55 anton 9489:
9490: @example
1.78 anton 9491: foo% @emph{n} *
1.56 anton 9492: @end example
1.55 anton 9493:
1.78 anton 9494: You can use this array type in any place where you can use a normal
9495: type, e.g., when defining a @code{field}, or with
9496: @code{%allot}.
9497:
9498: @cindex first field optimization
9499: The first field is at the base address of a structure and the word for
9500: this field (e.g., @code{list-next}) actually does not change the address
9501: on the stack. You may be tempted to leave it away in the interest of
9502: run-time and space efficiency. This is not necessary, because the
9503: structure package optimizes this case: If you compile a first-field
9504: words, no code is generated. So, in the interest of readability and
9505: maintainability you should include the word for the field when accessing
9506: the field.
1.52 anton 9507:
9508:
1.78 anton 9509: @node Structure Naming Convention, Structure Implementation, Structure Usage, Structures
9510: @subsection Structure Naming Convention
9511: @cindex structure naming convention
1.52 anton 9512:
1.78 anton 9513: The field names that come to (my) mind are often quite generic, and,
9514: if used, would cause frequent name clashes. E.g., many structures
9515: probably contain a @code{counter} field. The structure names
9516: that come to (my) mind are often also the logical choice for the names
9517: of words that create such a structure.
1.52 anton 9518:
1.78 anton 9519: Therefore, I have adopted the following naming conventions:
1.52 anton 9520:
1.78 anton 9521: @itemize @bullet
9522: @cindex field naming convention
9523: @item
9524: The names of fields are of the form
9525: @code{@emph{struct}-@emph{field}}, where
9526: @code{@emph{struct}} is the basic name of the structure, and
9527: @code{@emph{field}} is the basic name of the field. You can
9528: think of field words as converting the (address of the)
9529: structure into the (address of the) field.
1.52 anton 9530:
1.78 anton 9531: @cindex structure naming convention
9532: @item
9533: The names of structures are of the form
9534: @code{@emph{struct}%}, where
9535: @code{@emph{struct}} is the basic name of the structure.
9536: @end itemize
1.52 anton 9537:
1.78 anton 9538: This naming convention does not work that well for fields of extended
9539: structures; e.g., the integer list structure has a field
9540: @code{intlist-int}, but has @code{list-next}, not
9541: @code{intlist-next}.
1.53 anton 9542:
1.78 anton 9543: @node Structure Implementation, Structure Glossary, Structure Naming Convention, Structures
9544: @subsection Structure Implementation
9545: @cindex structure implementation
9546: @cindex implementation of structures
1.52 anton 9547:
1.78 anton 9548: The central idea in the implementation is to pass the data about the
9549: structure being built on the stack, not in some global
9550: variable. Everything else falls into place naturally once this design
9551: decision is made.
1.53 anton 9552:
1.78 anton 9553: The type description on the stack is of the form @emph{align
9554: size}. Keeping the size on the top-of-stack makes dealing with arrays
9555: very simple.
1.53 anton 9556:
1.78 anton 9557: @code{field} is a defining word that uses @code{Create}
9558: and @code{DOES>}. The body of the field contains the offset
9559: of the field, and the normal @code{DOES>} action is simply:
1.53 anton 9560:
9561: @example
1.78 anton 9562: @@ +
1.53 anton 9563: @end example
9564:
1.78 anton 9565: @noindent
9566: i.e., add the offset to the address, giving the stack effect
9567: @i{addr1 -- addr2} for a field.
9568:
9569: @cindex first field optimization, implementation
9570: This simple structure is slightly complicated by the optimization
9571: for fields with offset 0, which requires a different
9572: @code{DOES>}-part (because we cannot rely on there being
9573: something on the stack if such a field is invoked during
9574: compilation). Therefore, we put the different @code{DOES>}-parts
9575: in separate words, and decide which one to invoke based on the
9576: offset. For a zero offset, the field is basically a noop; it is
9577: immediate, and therefore no code is generated when it is compiled.
1.53 anton 9578:
1.78 anton 9579: @node Structure Glossary, , Structure Implementation, Structures
9580: @subsection Structure Glossary
9581: @cindex structure glossary
1.53 anton 9582:
1.5 anton 9583:
1.78 anton 9584: doc-%align
9585: doc-%alignment
9586: doc-%alloc
9587: doc-%allocate
9588: doc-%allot
9589: doc-cell%
9590: doc-char%
9591: doc-dfloat%
9592: doc-double%
9593: doc-end-struct
9594: doc-field
9595: doc-float%
9596: doc-naligned
9597: doc-sfloat%
9598: doc-%size
9599: doc-struct
1.54 anton 9600:
9601:
1.26 crook 9602: @c -------------------------------------------------------------
1.78 anton 9603: @node Object-oriented Forth, Programming Tools, Structures, Words
9604: @section Object-oriented Forth
9605:
9606: Gforth comes with three packages for object-oriented programming:
9607: @file{objects.fs}, @file{oof.fs}, and @file{mini-oof.fs}; none of them
9608: is preloaded, so you have to @code{include} them before use. The most
9609: important differences between these packages (and others) are discussed
9610: in @ref{Comparison with other object models}. All packages are written
9611: in ANS Forth and can be used with any other ANS Forth.
1.5 anton 9612:
1.78 anton 9613: @menu
9614: * Why object-oriented programming?::
9615: * Object-Oriented Terminology::
9616: * Objects::
9617: * OOF::
9618: * Mini-OOF::
9619: * Comparison with other object models::
9620: @end menu
1.5 anton 9621:
1.78 anton 9622: @c ----------------------------------------------------------------
9623: @node Why object-oriented programming?, Object-Oriented Terminology, Object-oriented Forth, Object-oriented Forth
9624: @subsection Why object-oriented programming?
9625: @cindex object-oriented programming motivation
9626: @cindex motivation for object-oriented programming
1.44 crook 9627:
1.78 anton 9628: Often we have to deal with several data structures (@emph{objects}),
9629: that have to be treated similarly in some respects, but differently in
9630: others. Graphical objects are the textbook example: circles, triangles,
9631: dinosaurs, icons, and others, and we may want to add more during program
9632: development. We want to apply some operations to any graphical object,
9633: e.g., @code{draw} for displaying it on the screen. However, @code{draw}
9634: has to do something different for every kind of object.
9635: @comment TODO add some other operations eg perimeter, area
9636: @comment and tie in to concrete examples later..
1.5 anton 9637:
1.78 anton 9638: We could implement @code{draw} as a big @code{CASE}
9639: control structure that executes the appropriate code depending on the
9640: kind of object to be drawn. This would be not be very elegant, and,
9641: moreover, we would have to change @code{draw} every time we add
9642: a new kind of graphical object (say, a spaceship).
1.44 crook 9643:
1.78 anton 9644: What we would rather do is: When defining spaceships, we would tell
9645: the system: ``Here's how you @code{draw} a spaceship; you figure
9646: out the rest''.
1.5 anton 9647:
1.78 anton 9648: This is the problem that all systems solve that (rightfully) call
9649: themselves object-oriented; the object-oriented packages presented here
9650: solve this problem (and not much else).
9651: @comment TODO ?list properties of oo systems.. oo vs o-based?
1.44 crook 9652:
1.78 anton 9653: @c ------------------------------------------------------------------------
9654: @node Object-Oriented Terminology, Objects, Why object-oriented programming?, Object-oriented Forth
9655: @subsection Object-Oriented Terminology
9656: @cindex object-oriented terminology
9657: @cindex terminology for object-oriented programming
1.5 anton 9658:
1.78 anton 9659: This section is mainly for reference, so you don't have to understand
9660: all of it right away. The terminology is mainly Smalltalk-inspired. In
9661: short:
1.44 crook 9662:
1.78 anton 9663: @table @emph
9664: @cindex class
9665: @item class
9666: a data structure definition with some extras.
1.5 anton 9667:
1.78 anton 9668: @cindex object
9669: @item object
9670: an instance of the data structure described by the class definition.
1.5 anton 9671:
1.78 anton 9672: @cindex instance variables
9673: @item instance variables
9674: fields of the data structure.
1.5 anton 9675:
1.78 anton 9676: @cindex selector
9677: @cindex method selector
9678: @cindex virtual function
9679: @item selector
9680: (or @emph{method selector}) a word (e.g.,
9681: @code{draw}) that performs an operation on a variety of data
9682: structures (classes). A selector describes @emph{what} operation to
9683: perform. In C++ terminology: a (pure) virtual function.
1.5 anton 9684:
1.78 anton 9685: @cindex method
9686: @item method
9687: the concrete definition that performs the operation
9688: described by the selector for a specific class. A method specifies
9689: @emph{how} the operation is performed for a specific class.
1.5 anton 9690:
1.78 anton 9691: @cindex selector invocation
9692: @cindex message send
9693: @cindex invoking a selector
9694: @item selector invocation
9695: a call of a selector. One argument of the call (the TOS (top-of-stack))
9696: is used for determining which method is used. In Smalltalk terminology:
9697: a message (consisting of the selector and the other arguments) is sent
9698: to the object.
1.5 anton 9699:
1.78 anton 9700: @cindex receiving object
9701: @item receiving object
9702: the object used for determining the method executed by a selector
9703: invocation. In the @file{objects.fs} model, it is the object that is on
9704: the TOS when the selector is invoked. (@emph{Receiving} comes from
9705: the Smalltalk @emph{message} terminology.)
1.5 anton 9706:
1.78 anton 9707: @cindex child class
9708: @cindex parent class
9709: @cindex inheritance
9710: @item child class
9711: a class that has (@emph{inherits}) all properties (instance variables,
9712: selectors, methods) from a @emph{parent class}. In Smalltalk
9713: terminology: The subclass inherits from the superclass. In C++
9714: terminology: The derived class inherits from the base class.
1.5 anton 9715:
1.78 anton 9716: @end table
1.5 anton 9717:
1.78 anton 9718: @c If you wonder about the message sending terminology, it comes from
9719: @c a time when each object had it's own task and objects communicated via
9720: @c message passing; eventually the Smalltalk developers realized that
9721: @c they can do most things through simple (indirect) calls. They kept the
9722: @c terminology.
1.5 anton 9723:
1.78 anton 9724: @c --------------------------------------------------------------
9725: @node Objects, OOF, Object-Oriented Terminology, Object-oriented Forth
9726: @subsection The @file{objects.fs} model
9727: @cindex objects
9728: @cindex object-oriented programming
1.26 crook 9729:
1.78 anton 9730: @cindex @file{objects.fs}
9731: @cindex @file{oof.fs}
1.26 crook 9732:
1.78 anton 9733: This section describes the @file{objects.fs} package. This material also
9734: has been published in M. Anton Ertl,
9735: @cite{@uref{http://www.complang.tuwien.ac.at/forth/objects/objects.html,
9736: Yet Another Forth Objects Package}}, Forth Dimensions 19(2), pages
9737: 37--43.
9738: @c McKewan's and Zsoter's packages
1.26 crook 9739:
1.78 anton 9740: This section assumes that you have read @ref{Structures}.
1.5 anton 9741:
1.78 anton 9742: The techniques on which this model is based have been used to implement
9743: the parser generator, Gray, and have also been used in Gforth for
9744: implementing the various flavours of word lists (hashed or not,
9745: case-sensitive or not, special-purpose word lists for locals etc.).
1.5 anton 9746:
9747:
1.26 crook 9748: @menu
1.78 anton 9749: * Properties of the Objects model::
9750: * Basic Objects Usage::
9751: * The Objects base class::
9752: * Creating objects::
9753: * Object-Oriented Programming Style::
9754: * Class Binding::
9755: * Method conveniences::
9756: * Classes and Scoping::
9757: * Dividing classes::
9758: * Object Interfaces::
9759: * Objects Implementation::
9760: * Objects Glossary::
1.26 crook 9761: @end menu
1.5 anton 9762:
1.78 anton 9763: Marcel Hendrix provided helpful comments on this section.
1.5 anton 9764:
1.78 anton 9765: @node Properties of the Objects model, Basic Objects Usage, Objects, Objects
9766: @subsubsection Properties of the @file{objects.fs} model
9767: @cindex @file{objects.fs} properties
1.5 anton 9768:
1.78 anton 9769: @itemize @bullet
9770: @item
9771: It is straightforward to pass objects on the stack. Passing
9772: selectors on the stack is a little less convenient, but possible.
1.44 crook 9773:
1.78 anton 9774: @item
9775: Objects are just data structures in memory, and are referenced by their
9776: address. You can create words for objects with normal defining words
9777: like @code{constant}. Likewise, there is no difference between instance
9778: variables that contain objects and those that contain other data.
1.5 anton 9779:
1.78 anton 9780: @item
9781: Late binding is efficient and easy to use.
1.44 crook 9782:
1.78 anton 9783: @item
9784: It avoids parsing, and thus avoids problems with state-smartness
9785: and reduced extensibility; for convenience there are a few parsing
9786: words, but they have non-parsing counterparts. There are also a few
9787: defining words that parse. This is hard to avoid, because all standard
9788: defining words parse (except @code{:noname}); however, such
9789: words are not as bad as many other parsing words, because they are not
9790: state-smart.
1.5 anton 9791:
1.78 anton 9792: @item
9793: It does not try to incorporate everything. It does a few things and does
9794: them well (IMO). In particular, this model was not designed to support
9795: information hiding (although it has features that may help); you can use
9796: a separate package for achieving this.
1.5 anton 9797:
1.78 anton 9798: @item
9799: It is layered; you don't have to learn and use all features to use this
9800: model. Only a few features are necessary (@pxref{Basic Objects Usage},
9801: @pxref{The Objects base class}, @pxref{Creating objects}.), the others
9802: are optional and independent of each other.
1.5 anton 9803:
1.78 anton 9804: @item
9805: An implementation in ANS Forth is available.
1.5 anton 9806:
1.78 anton 9807: @end itemize
1.5 anton 9808:
1.44 crook 9809:
1.78 anton 9810: @node Basic Objects Usage, The Objects base class, Properties of the Objects model, Objects
9811: @subsubsection Basic @file{objects.fs} Usage
9812: @cindex basic objects usage
9813: @cindex objects, basic usage
1.5 anton 9814:
1.78 anton 9815: You can define a class for graphical objects like this:
1.44 crook 9816:
1.78 anton 9817: @cindex @code{class} usage
9818: @cindex @code{end-class} usage
9819: @cindex @code{selector} usage
1.5 anton 9820: @example
1.78 anton 9821: object class \ "object" is the parent class
9822: selector draw ( x y graphical -- )
9823: end-class graphical
9824: @end example
9825:
9826: This code defines a class @code{graphical} with an
9827: operation @code{draw}. We can perform the operation
9828: @code{draw} on any @code{graphical} object, e.g.:
9829:
9830: @example
9831: 100 100 t-rex draw
1.26 crook 9832: @end example
1.5 anton 9833:
1.78 anton 9834: @noindent
9835: where @code{t-rex} is a word (say, a constant) that produces a
9836: graphical object.
9837:
9838: @comment TODO add a 2nd operation eg perimeter.. and use for
9839: @comment a concrete example
1.5 anton 9840:
1.78 anton 9841: @cindex abstract class
9842: How do we create a graphical object? With the present definitions,
9843: we cannot create a useful graphical object. The class
9844: @code{graphical} describes graphical objects in general, but not
9845: any concrete graphical object type (C++ users would call it an
9846: @emph{abstract class}); e.g., there is no method for the selector
9847: @code{draw} in the class @code{graphical}.
1.5 anton 9848:
1.78 anton 9849: For concrete graphical objects, we define child classes of the
9850: class @code{graphical}, e.g.:
1.5 anton 9851:
1.78 anton 9852: @cindex @code{overrides} usage
9853: @cindex @code{field} usage in class definition
1.26 crook 9854: @example
1.78 anton 9855: graphical class \ "graphical" is the parent class
9856: cell% field circle-radius
1.5 anton 9857:
1.78 anton 9858: :noname ( x y circle -- )
9859: circle-radius @@ draw-circle ;
9860: overrides draw
1.5 anton 9861:
1.78 anton 9862: :noname ( n-radius circle -- )
9863: circle-radius ! ;
9864: overrides construct
1.5 anton 9865:
1.78 anton 9866: end-class circle
9867: @end example
1.44 crook 9868:
1.78 anton 9869: Here we define a class @code{circle} as a child of @code{graphical},
9870: with field @code{circle-radius} (which behaves just like a field
9871: (@pxref{Structures}); it defines (using @code{overrides}) new methods
9872: for the selectors @code{draw} and @code{construct} (@code{construct} is
9873: defined in @code{object}, the parent class of @code{graphical}).
1.5 anton 9874:
1.78 anton 9875: Now we can create a circle on the heap (i.e.,
9876: @code{allocate}d memory) with:
1.44 crook 9877:
1.78 anton 9878: @cindex @code{heap-new} usage
1.5 anton 9879: @example
1.78 anton 9880: 50 circle heap-new constant my-circle
1.5 anton 9881: @end example
9882:
1.78 anton 9883: @noindent
9884: @code{heap-new} invokes @code{construct}, thus
9885: initializing the field @code{circle-radius} with 50. We can draw
9886: this new circle at (100,100) with:
1.5 anton 9887:
9888: @example
1.78 anton 9889: 100 100 my-circle draw
1.5 anton 9890: @end example
9891:
1.78 anton 9892: @cindex selector invocation, restrictions
9893: @cindex class definition, restrictions
9894: Note: You can only invoke a selector if the object on the TOS
9895: (the receiving object) belongs to the class where the selector was
9896: defined or one of its descendents; e.g., you can invoke
9897: @code{draw} only for objects belonging to @code{graphical}
9898: or its descendents (e.g., @code{circle}). Immediately before
9899: @code{end-class}, the search order has to be the same as
9900: immediately after @code{class}.
9901:
9902: @node The Objects base class, Creating objects, Basic Objects Usage, Objects
9903: @subsubsection The @file{object.fs} base class
9904: @cindex @code{object} class
9905:
9906: When you define a class, you have to specify a parent class. So how do
9907: you start defining classes? There is one class available from the start:
9908: @code{object}. It is ancestor for all classes and so is the
9909: only class that has no parent. It has two selectors: @code{construct}
9910: and @code{print}.
9911:
9912: @node Creating objects, Object-Oriented Programming Style, The Objects base class, Objects
9913: @subsubsection Creating objects
9914: @cindex creating objects
9915: @cindex object creation
9916: @cindex object allocation options
9917:
9918: @cindex @code{heap-new} discussion
9919: @cindex @code{dict-new} discussion
9920: @cindex @code{construct} discussion
9921: You can create and initialize an object of a class on the heap with
9922: @code{heap-new} ( ... class -- object ) and in the dictionary
9923: (allocation with @code{allot}) with @code{dict-new} (
9924: ... class -- object ). Both words invoke @code{construct}, which
9925: consumes the stack items indicated by "..." above.
9926:
9927: @cindex @code{init-object} discussion
9928: @cindex @code{class-inst-size} discussion
9929: If you want to allocate memory for an object yourself, you can get its
9930: alignment and size with @code{class-inst-size 2@@} ( class --
9931: align size ). Once you have memory for an object, you can initialize
9932: it with @code{init-object} ( ... class object -- );
9933: @code{construct} does only a part of the necessary work.
9934:
9935: @node Object-Oriented Programming Style, Class Binding, Creating objects, Objects
9936: @subsubsection Object-Oriented Programming Style
9937: @cindex object-oriented programming style
9938: @cindex programming style, object-oriented
1.5 anton 9939:
1.78 anton 9940: This section is not exhaustive.
1.5 anton 9941:
1.78 anton 9942: @cindex stack effects of selectors
9943: @cindex selectors and stack effects
9944: In general, it is a good idea to ensure that all methods for the
9945: same selector have the same stack effect: when you invoke a selector,
9946: you often have no idea which method will be invoked, so, unless all
9947: methods have the same stack effect, you will not know the stack effect
9948: of the selector invocation.
1.5 anton 9949:
1.78 anton 9950: One exception to this rule is methods for the selector
9951: @code{construct}. We know which method is invoked, because we
9952: specify the class to be constructed at the same place. Actually, I
9953: defined @code{construct} as a selector only to give the users a
9954: convenient way to specify initialization. The way it is used, a
9955: mechanism different from selector invocation would be more natural
9956: (but probably would take more code and more space to explain).
1.5 anton 9957:
1.78 anton 9958: @node Class Binding, Method conveniences, Object-Oriented Programming Style, Objects
9959: @subsubsection Class Binding
9960: @cindex class binding
9961: @cindex early binding
1.5 anton 9962:
1.78 anton 9963: @cindex late binding
9964: Normal selector invocations determine the method at run-time depending
9965: on the class of the receiving object. This run-time selection is called
9966: @i{late binding}.
1.5 anton 9967:
1.78 anton 9968: Sometimes it's preferable to invoke a different method. For example,
9969: you might want to use the simple method for @code{print}ing
9970: @code{object}s instead of the possibly long-winded @code{print} method
9971: of the receiver class. You can achieve this by replacing the invocation
9972: of @code{print} with:
1.5 anton 9973:
1.78 anton 9974: @cindex @code{[bind]} usage
1.5 anton 9975: @example
1.78 anton 9976: [bind] object print
1.5 anton 9977: @end example
9978:
1.78 anton 9979: @noindent
9980: in compiled code or:
9981:
9982: @cindex @code{bind} usage
1.5 anton 9983: @example
1.78 anton 9984: bind object print
1.5 anton 9985: @end example
9986:
1.78 anton 9987: @cindex class binding, alternative to
9988: @noindent
9989: in interpreted code. Alternatively, you can define the method with a
9990: name (e.g., @code{print-object}), and then invoke it through the
9991: name. Class binding is just a (often more convenient) way to achieve
9992: the same effect; it avoids name clutter and allows you to invoke
9993: methods directly without naming them first.
1.5 anton 9994:
1.78 anton 9995: @cindex superclass binding
9996: @cindex parent class binding
9997: A frequent use of class binding is this: When we define a method
9998: for a selector, we often want the method to do what the selector does
9999: in the parent class, and a little more. There is a special word for
10000: this purpose: @code{[parent]}; @code{[parent]
10001: @emph{selector}} is equivalent to @code{[bind] @emph{parent
10002: selector}}, where @code{@emph{parent}} is the parent
10003: class of the current class. E.g., a method definition might look like:
1.44 crook 10004:
1.78 anton 10005: @cindex @code{[parent]} usage
10006: @example
10007: :noname
10008: dup [parent] foo \ do parent's foo on the receiving object
10009: ... \ do some more
10010: ; overrides foo
10011: @end example
1.6 pazsan 10012:
1.78 anton 10013: @cindex class binding as optimization
10014: In @cite{Object-oriented programming in ANS Forth} (Forth Dimensions,
10015: March 1997), Andrew McKewan presents class binding as an optimization
10016: technique. I recommend not using it for this purpose unless you are in
10017: an emergency. Late binding is pretty fast with this model anyway, so the
10018: benefit of using class binding is small; the cost of using class binding
10019: where it is not appropriate is reduced maintainability.
1.44 crook 10020:
1.78 anton 10021: While we are at programming style questions: You should bind
10022: selectors only to ancestor classes of the receiving object. E.g., say,
10023: you know that the receiving object is of class @code{foo} or its
10024: descendents; then you should bind only to @code{foo} and its
10025: ancestors.
1.12 anton 10026:
1.78 anton 10027: @node Method conveniences, Classes and Scoping, Class Binding, Objects
10028: @subsubsection Method conveniences
10029: @cindex method conveniences
1.44 crook 10030:
1.78 anton 10031: In a method you usually access the receiving object pretty often. If
10032: you define the method as a plain colon definition (e.g., with
10033: @code{:noname}), you may have to do a lot of stack
10034: gymnastics. To avoid this, you can define the method with @code{m:
10035: ... ;m}. E.g., you could define the method for
10036: @code{draw}ing a @code{circle} with
1.6 pazsan 10037:
1.78 anton 10038: @cindex @code{this} usage
10039: @cindex @code{m:} usage
10040: @cindex @code{;m} usage
10041: @example
10042: m: ( x y circle -- )
10043: ( x y ) this circle-radius @@ draw-circle ;m
10044: @end example
1.6 pazsan 10045:
1.78 anton 10046: @cindex @code{exit} in @code{m: ... ;m}
10047: @cindex @code{exitm} discussion
10048: @cindex @code{catch} in @code{m: ... ;m}
10049: When this method is executed, the receiver object is removed from the
10050: stack; you can access it with @code{this} (admittedly, in this
10051: example the use of @code{m: ... ;m} offers no advantage). Note
10052: that I specify the stack effect for the whole method (i.e. including
10053: the receiver object), not just for the code between @code{m:}
10054: and @code{;m}. You cannot use @code{exit} in
10055: @code{m:...;m}; instead, use
10056: @code{exitm}.@footnote{Moreover, for any word that calls
10057: @code{catch} and was defined before loading
10058: @code{objects.fs}, you have to redefine it like I redefined
10059: @code{catch}: @code{: catch this >r catch r> to-this ;}}
1.12 anton 10060:
1.78 anton 10061: @cindex @code{inst-var} usage
10062: You will frequently use sequences of the form @code{this
10063: @emph{field}} (in the example above: @code{this
10064: circle-radius}). If you use the field only in this way, you can
10065: define it with @code{inst-var} and eliminate the
10066: @code{this} before the field name. E.g., the @code{circle}
10067: class above could also be defined with:
1.6 pazsan 10068:
1.78 anton 10069: @example
10070: graphical class
10071: cell% inst-var radius
1.6 pazsan 10072:
1.78 anton 10073: m: ( x y circle -- )
10074: radius @@ draw-circle ;m
10075: overrides draw
1.6 pazsan 10076:
1.78 anton 10077: m: ( n-radius circle -- )
10078: radius ! ;m
10079: overrides construct
1.6 pazsan 10080:
1.78 anton 10081: end-class circle
10082: @end example
1.6 pazsan 10083:
1.78 anton 10084: @code{radius} can only be used in @code{circle} and its
10085: descendent classes and inside @code{m:...;m}.
1.6 pazsan 10086:
1.78 anton 10087: @cindex @code{inst-value} usage
10088: You can also define fields with @code{inst-value}, which is
10089: to @code{inst-var} what @code{value} is to
10090: @code{variable}. You can change the value of such a field with
10091: @code{[to-inst]}. E.g., we could also define the class
10092: @code{circle} like this:
1.44 crook 10093:
1.78 anton 10094: @example
10095: graphical class
10096: inst-value radius
1.6 pazsan 10097:
1.78 anton 10098: m: ( x y circle -- )
10099: radius draw-circle ;m
10100: overrides draw
1.44 crook 10101:
1.78 anton 10102: m: ( n-radius circle -- )
10103: [to-inst] radius ;m
10104: overrides construct
1.6 pazsan 10105:
1.78 anton 10106: end-class circle
10107: @end example
1.6 pazsan 10108:
1.78 anton 10109: @c !! :m is easy to confuse with m:. Another name would be better.
1.6 pazsan 10110:
1.78 anton 10111: @c Finally, you can define named methods with @code{:m}. One use of this
10112: @c feature is the definition of words that occur only in one class and are
10113: @c not intended to be overridden, but which still need method context
10114: @c (e.g., for accessing @code{inst-var}s). Another use is for methods that
10115: @c would be bound frequently, if defined anonymously.
1.6 pazsan 10116:
10117:
1.78 anton 10118: @node Classes and Scoping, Dividing classes, Method conveniences, Objects
10119: @subsubsection Classes and Scoping
10120: @cindex classes and scoping
10121: @cindex scoping and classes
1.6 pazsan 10122:
1.78 anton 10123: Inheritance is frequent, unlike structure extension. This exacerbates
10124: the problem with the field name convention (@pxref{Structure Naming
10125: Convention}): One always has to remember in which class the field was
10126: originally defined; changing a part of the class structure would require
10127: changes for renaming in otherwise unaffected code.
1.6 pazsan 10128:
1.78 anton 10129: @cindex @code{inst-var} visibility
10130: @cindex @code{inst-value} visibility
10131: To solve this problem, I added a scoping mechanism (which was not in my
10132: original charter): A field defined with @code{inst-var} (or
10133: @code{inst-value}) is visible only in the class where it is defined and in
10134: the descendent classes of this class. Using such fields only makes
10135: sense in @code{m:}-defined methods in these classes anyway.
1.6 pazsan 10136:
1.78 anton 10137: This scoping mechanism allows us to use the unadorned field name,
10138: because name clashes with unrelated words become much less likely.
1.6 pazsan 10139:
1.78 anton 10140: @cindex @code{protected} discussion
10141: @cindex @code{private} discussion
10142: Once we have this mechanism, we can also use it for controlling the
10143: visibility of other words: All words defined after
10144: @code{protected} are visible only in the current class and its
10145: descendents. @code{public} restores the compilation
10146: (i.e. @code{current}) word list that was in effect before. If you
10147: have several @code{protected}s without an intervening
10148: @code{public} or @code{set-current}, @code{public}
10149: will restore the compilation word list in effect before the first of
10150: these @code{protected}s.
1.6 pazsan 10151:
1.78 anton 10152: @node Dividing classes, Object Interfaces, Classes and Scoping, Objects
10153: @subsubsection Dividing classes
10154: @cindex Dividing classes
10155: @cindex @code{methods}...@code{end-methods}
1.6 pazsan 10156:
1.78 anton 10157: You may want to do the definition of methods separate from the
10158: definition of the class, its selectors, fields, and instance variables,
10159: i.e., separate the implementation from the definition. You can do this
10160: in the following way:
1.6 pazsan 10161:
1.78 anton 10162: @example
10163: graphical class
10164: inst-value radius
10165: end-class circle
1.6 pazsan 10166:
1.78 anton 10167: ... \ do some other stuff
1.6 pazsan 10168:
1.78 anton 10169: circle methods \ now we are ready
1.44 crook 10170:
1.78 anton 10171: m: ( x y circle -- )
10172: radius draw-circle ;m
10173: overrides draw
1.6 pazsan 10174:
1.78 anton 10175: m: ( n-radius circle -- )
10176: [to-inst] radius ;m
10177: overrides construct
1.44 crook 10178:
1.78 anton 10179: end-methods
10180: @end example
1.7 pazsan 10181:
1.78 anton 10182: You can use several @code{methods}...@code{end-methods} sections. The
10183: only things you can do to the class in these sections are: defining
10184: methods, and overriding the class's selectors. You must not define new
10185: selectors or fields.
1.7 pazsan 10186:
1.78 anton 10187: Note that you often have to override a selector before using it. In
10188: particular, you usually have to override @code{construct} with a new
10189: method before you can invoke @code{heap-new} and friends. E.g., you
10190: must not create a circle before the @code{overrides construct} sequence
10191: in the example above.
1.7 pazsan 10192:
1.78 anton 10193: @node Object Interfaces, Objects Implementation, Dividing classes, Objects
10194: @subsubsection Object Interfaces
10195: @cindex object interfaces
10196: @cindex interfaces for objects
1.7 pazsan 10197:
1.78 anton 10198: In this model you can only call selectors defined in the class of the
10199: receiving objects or in one of its ancestors. If you call a selector
10200: with a receiving object that is not in one of these classes, the
10201: result is undefined; if you are lucky, the program crashes
10202: immediately.
1.7 pazsan 10203:
1.78 anton 10204: @cindex selectors common to hardly-related classes
10205: Now consider the case when you want to have a selector (or several)
10206: available in two classes: You would have to add the selector to a
10207: common ancestor class, in the worst case to @code{object}. You
10208: may not want to do this, e.g., because someone else is responsible for
10209: this ancestor class.
1.7 pazsan 10210:
1.78 anton 10211: The solution for this problem is interfaces. An interface is a
10212: collection of selectors. If a class implements an interface, the
10213: selectors become available to the class and its descendents. A class
10214: can implement an unlimited number of interfaces. For the problem
10215: discussed above, we would define an interface for the selector(s), and
10216: both classes would implement the interface.
1.7 pazsan 10217:
1.78 anton 10218: As an example, consider an interface @code{storage} for
10219: writing objects to disk and getting them back, and a class
10220: @code{foo} that implements it. The code would look like this:
1.7 pazsan 10221:
1.78 anton 10222: @cindex @code{interface} usage
10223: @cindex @code{end-interface} usage
10224: @cindex @code{implementation} usage
10225: @example
10226: interface
10227: selector write ( file object -- )
10228: selector read1 ( file object -- )
10229: end-interface storage
1.13 pazsan 10230:
1.78 anton 10231: bar class
10232: storage implementation
1.13 pazsan 10233:
1.78 anton 10234: ... overrides write
10235: ... overrides read1
10236: ...
10237: end-class foo
10238: @end example
1.13 pazsan 10239:
1.78 anton 10240: @noindent
10241: (I would add a word @code{read} @i{( file -- object )} that uses
10242: @code{read1} internally, but that's beyond the point illustrated
10243: here.)
1.13 pazsan 10244:
1.78 anton 10245: Note that you cannot use @code{protected} in an interface; and
10246: of course you cannot define fields.
1.13 pazsan 10247:
1.78 anton 10248: In the Neon model, all selectors are available for all classes;
10249: therefore it does not need interfaces. The price you pay in this model
10250: is slower late binding, and therefore, added complexity to avoid late
10251: binding.
1.13 pazsan 10252:
1.78 anton 10253: @node Objects Implementation, Objects Glossary, Object Interfaces, Objects
10254: @subsubsection @file{objects.fs} Implementation
10255: @cindex @file{objects.fs} implementation
1.13 pazsan 10256:
1.78 anton 10257: @cindex @code{object-map} discussion
10258: An object is a piece of memory, like one of the data structures
10259: described with @code{struct...end-struct}. It has a field
10260: @code{object-map} that points to the method map for the object's
10261: class.
1.13 pazsan 10262:
1.78 anton 10263: @cindex method map
10264: @cindex virtual function table
10265: The @emph{method map}@footnote{This is Self terminology; in C++
10266: terminology: virtual function table.} is an array that contains the
10267: execution tokens (@i{xt}s) of the methods for the object's class. Each
10268: selector contains an offset into a method map.
1.13 pazsan 10269:
1.78 anton 10270: @cindex @code{selector} implementation, class
10271: @code{selector} is a defining word that uses
10272: @code{CREATE} and @code{DOES>}. The body of the
10273: selector contains the offset; the @code{DOES>} action for a
10274: class selector is, basically:
1.8 pazsan 10275:
10276: @example
1.78 anton 10277: ( object addr ) @@ over object-map @@ + @@ execute
1.13 pazsan 10278: @end example
10279:
1.78 anton 10280: Since @code{object-map} is the first field of the object, it
10281: does not generate any code. As you can see, calling a selector has a
10282: small, constant cost.
1.26 crook 10283:
1.78 anton 10284: @cindex @code{current-interface} discussion
10285: @cindex class implementation and representation
10286: A class is basically a @code{struct} combined with a method
10287: map. During the class definition the alignment and size of the class
10288: are passed on the stack, just as with @code{struct}s, so
10289: @code{field} can also be used for defining class
10290: fields. However, passing more items on the stack would be
10291: inconvenient, so @code{class} builds a data structure in memory,
10292: which is accessed through the variable
10293: @code{current-interface}. After its definition is complete, the
10294: class is represented on the stack by a pointer (e.g., as parameter for
10295: a child class definition).
1.26 crook 10296:
1.78 anton 10297: A new class starts off with the alignment and size of its parent,
10298: and a copy of the parent's method map. Defining new fields extends the
10299: size and alignment; likewise, defining new selectors extends the
10300: method map. @code{overrides} just stores a new @i{xt} in the method
10301: map at the offset given by the selector.
1.13 pazsan 10302:
1.78 anton 10303: @cindex class binding, implementation
10304: Class binding just gets the @i{xt} at the offset given by the selector
10305: from the class's method map and @code{compile,}s (in the case of
10306: @code{[bind]}) it.
1.13 pazsan 10307:
1.78 anton 10308: @cindex @code{this} implementation
10309: @cindex @code{catch} and @code{this}
10310: @cindex @code{this} and @code{catch}
10311: I implemented @code{this} as a @code{value}. At the
10312: start of an @code{m:...;m} method the old @code{this} is
10313: stored to the return stack and restored at the end; and the object on
10314: the TOS is stored @code{TO this}. This technique has one
10315: disadvantage: If the user does not leave the method via
10316: @code{;m}, but via @code{throw} or @code{exit},
10317: @code{this} is not restored (and @code{exit} may
10318: crash). To deal with the @code{throw} problem, I have redefined
10319: @code{catch} to save and restore @code{this}; the same
10320: should be done with any word that can catch an exception. As for
10321: @code{exit}, I simply forbid it (as a replacement, there is
10322: @code{exitm}).
1.13 pazsan 10323:
1.78 anton 10324: @cindex @code{inst-var} implementation
10325: @code{inst-var} is just the same as @code{field}, with
10326: a different @code{DOES>} action:
1.13 pazsan 10327: @example
1.78 anton 10328: @@ this +
1.8 pazsan 10329: @end example
1.78 anton 10330: Similar for @code{inst-value}.
1.8 pazsan 10331:
1.78 anton 10332: @cindex class scoping implementation
10333: Each class also has a word list that contains the words defined with
10334: @code{inst-var} and @code{inst-value}, and its protected
10335: words. It also has a pointer to its parent. @code{class} pushes
10336: the word lists of the class and all its ancestors onto the search order stack,
10337: and @code{end-class} drops them.
1.20 pazsan 10338:
1.78 anton 10339: @cindex interface implementation
10340: An interface is like a class without fields, parent and protected
10341: words; i.e., it just has a method map. If a class implements an
10342: interface, its method map contains a pointer to the method map of the
10343: interface. The positive offsets in the map are reserved for class
10344: methods, therefore interface map pointers have negative
10345: offsets. Interfaces have offsets that are unique throughout the
10346: system, unlike class selectors, whose offsets are only unique for the
10347: classes where the selector is available (invokable).
1.20 pazsan 10348:
1.78 anton 10349: This structure means that interface selectors have to perform one
10350: indirection more than class selectors to find their method. Their body
10351: contains the interface map pointer offset in the class method map, and
10352: the method offset in the interface method map. The
10353: @code{does>} action for an interface selector is, basically:
1.20 pazsan 10354:
10355: @example
1.78 anton 10356: ( object selector-body )
10357: 2dup selector-interface @@ ( object selector-body object interface-offset )
10358: swap object-map @@ + @@ ( object selector-body map )
10359: swap selector-offset @@ + @@ execute
1.20 pazsan 10360: @end example
10361:
1.78 anton 10362: where @code{object-map} and @code{selector-offset} are
10363: first fields and generate no code.
1.20 pazsan 10364:
1.78 anton 10365: As a concrete example, consider the following code:
1.20 pazsan 10366:
10367: @example
1.78 anton 10368: interface
10369: selector if1sel1
10370: selector if1sel2
10371: end-interface if1
1.20 pazsan 10372:
1.78 anton 10373: object class
10374: if1 implementation
10375: selector cl1sel1
10376: cell% inst-var cl1iv1
1.20 pazsan 10377:
1.78 anton 10378: ' m1 overrides construct
10379: ' m2 overrides if1sel1
10380: ' m3 overrides if1sel2
10381: ' m4 overrides cl1sel2
10382: end-class cl1
1.20 pazsan 10383:
1.78 anton 10384: create obj1 object dict-new drop
10385: create obj2 cl1 dict-new drop
10386: @end example
1.20 pazsan 10387:
1.78 anton 10388: The data structure created by this code (including the data structure
10389: for @code{object}) is shown in the
10390: @uref{objects-implementation.eps,figure}, assuming a cell size of 4.
10391: @comment TODO add this diagram..
1.20 pazsan 10392:
1.78 anton 10393: @node Objects Glossary, , Objects Implementation, Objects
10394: @subsubsection @file{objects.fs} Glossary
10395: @cindex @file{objects.fs} Glossary
1.20 pazsan 10396:
10397:
1.78 anton 10398: doc---objects-bind
10399: doc---objects-<bind>
10400: doc---objects-bind'
10401: doc---objects-[bind]
10402: doc---objects-class
10403: doc---objects-class->map
10404: doc---objects-class-inst-size
10405: doc---objects-class-override!
1.79 anton 10406: doc---objects-class-previous
10407: doc---objects-class>order
1.78 anton 10408: doc---objects-construct
10409: doc---objects-current'
10410: doc---objects-[current]
10411: doc---objects-current-interface
10412: doc---objects-dict-new
10413: doc---objects-end-class
10414: doc---objects-end-class-noname
10415: doc---objects-end-interface
10416: doc---objects-end-interface-noname
10417: doc---objects-end-methods
10418: doc---objects-exitm
10419: doc---objects-heap-new
10420: doc---objects-implementation
10421: doc---objects-init-object
10422: doc---objects-inst-value
10423: doc---objects-inst-var
10424: doc---objects-interface
10425: doc---objects-m:
10426: doc---objects-:m
10427: doc---objects-;m
10428: doc---objects-method
10429: doc---objects-methods
10430: doc---objects-object
10431: doc---objects-overrides
10432: doc---objects-[parent]
10433: doc---objects-print
10434: doc---objects-protected
10435: doc---objects-public
10436: doc---objects-selector
10437: doc---objects-this
10438: doc---objects-<to-inst>
10439: doc---objects-[to-inst]
10440: doc---objects-to-this
10441: doc---objects-xt-new
1.20 pazsan 10442:
10443:
1.78 anton 10444: @c -------------------------------------------------------------
10445: @node OOF, Mini-OOF, Objects, Object-oriented Forth
10446: @subsection The @file{oof.fs} model
10447: @cindex oof
10448: @cindex object-oriented programming
1.20 pazsan 10449:
1.78 anton 10450: @cindex @file{objects.fs}
10451: @cindex @file{oof.fs}
1.20 pazsan 10452:
1.78 anton 10453: This section describes the @file{oof.fs} package.
1.20 pazsan 10454:
1.78 anton 10455: The package described in this section has been used in bigFORTH since 1991, and
10456: used for two large applications: a chromatographic system used to
10457: create new medicaments, and a graphic user interface library (MINOS).
1.20 pazsan 10458:
1.78 anton 10459: You can find a description (in German) of @file{oof.fs} in @cite{Object
10460: oriented bigFORTH} by Bernd Paysan, published in @cite{Vierte Dimension}
10461: 10(2), 1994.
1.20 pazsan 10462:
1.78 anton 10463: @menu
10464: * Properties of the OOF model::
10465: * Basic OOF Usage::
10466: * The OOF base class::
10467: * Class Declaration::
10468: * Class Implementation::
10469: @end menu
1.20 pazsan 10470:
1.78 anton 10471: @node Properties of the OOF model, Basic OOF Usage, OOF, OOF
10472: @subsubsection Properties of the @file{oof.fs} model
10473: @cindex @file{oof.fs} properties
1.20 pazsan 10474:
1.78 anton 10475: @itemize @bullet
10476: @item
10477: This model combines object oriented programming with information
10478: hiding. It helps you writing large application, where scoping is
10479: necessary, because it provides class-oriented scoping.
1.20 pazsan 10480:
1.78 anton 10481: @item
10482: Named objects, object pointers, and object arrays can be created,
10483: selector invocation uses the ``object selector'' syntax. Selector invocation
10484: to objects and/or selectors on the stack is a bit less convenient, but
10485: possible.
1.44 crook 10486:
1.78 anton 10487: @item
10488: Selector invocation and instance variable usage of the active object is
10489: straightforward, since both make use of the active object.
1.44 crook 10490:
1.78 anton 10491: @item
10492: Late binding is efficient and easy to use.
1.20 pazsan 10493:
1.78 anton 10494: @item
10495: State-smart objects parse selectors. However, extensibility is provided
10496: using a (parsing) selector @code{postpone} and a selector @code{'}.
1.20 pazsan 10497:
1.78 anton 10498: @item
10499: An implementation in ANS Forth is available.
1.20 pazsan 10500:
1.78 anton 10501: @end itemize
1.23 crook 10502:
10503:
1.78 anton 10504: @node Basic OOF Usage, The OOF base class, Properties of the OOF model, OOF
10505: @subsubsection Basic @file{oof.fs} Usage
10506: @cindex @file{oof.fs} usage
1.23 crook 10507:
1.78 anton 10508: This section uses the same example as for @code{objects} (@pxref{Basic Objects Usage}).
1.23 crook 10509:
1.78 anton 10510: You can define a class for graphical objects like this:
1.23 crook 10511:
1.78 anton 10512: @cindex @code{class} usage
10513: @cindex @code{class;} usage
10514: @cindex @code{method} usage
10515: @example
10516: object class graphical \ "object" is the parent class
10517: method draw ( x y graphical -- )
10518: class;
10519: @end example
1.23 crook 10520:
1.78 anton 10521: This code defines a class @code{graphical} with an
10522: operation @code{draw}. We can perform the operation
10523: @code{draw} on any @code{graphical} object, e.g.:
1.23 crook 10524:
1.78 anton 10525: @example
10526: 100 100 t-rex draw
10527: @end example
1.23 crook 10528:
1.78 anton 10529: @noindent
10530: where @code{t-rex} is an object or object pointer, created with e.g.
10531: @code{graphical : t-rex}.
1.23 crook 10532:
1.78 anton 10533: @cindex abstract class
10534: How do we create a graphical object? With the present definitions,
10535: we cannot create a useful graphical object. The class
10536: @code{graphical} describes graphical objects in general, but not
10537: any concrete graphical object type (C++ users would call it an
10538: @emph{abstract class}); e.g., there is no method for the selector
10539: @code{draw} in the class @code{graphical}.
1.23 crook 10540:
1.78 anton 10541: For concrete graphical objects, we define child classes of the
10542: class @code{graphical}, e.g.:
1.23 crook 10543:
1.78 anton 10544: @example
10545: graphical class circle \ "graphical" is the parent class
10546: cell var circle-radius
10547: how:
10548: : draw ( x y -- )
10549: circle-radius @@ draw-circle ;
1.23 crook 10550:
1.78 anton 10551: : init ( n-radius -- (
10552: circle-radius ! ;
10553: class;
10554: @end example
1.1 anton 10555:
1.78 anton 10556: Here we define a class @code{circle} as a child of @code{graphical},
10557: with a field @code{circle-radius}; it defines new methods for the
10558: selectors @code{draw} and @code{init} (@code{init} is defined in
10559: @code{object}, the parent class of @code{graphical}).
1.1 anton 10560:
1.78 anton 10561: Now we can create a circle in the dictionary with:
1.1 anton 10562:
1.78 anton 10563: @example
10564: 50 circle : my-circle
10565: @end example
1.21 crook 10566:
1.78 anton 10567: @noindent
10568: @code{:} invokes @code{init}, thus initializing the field
10569: @code{circle-radius} with 50. We can draw this new circle at (100,100)
10570: with:
1.1 anton 10571:
1.78 anton 10572: @example
10573: 100 100 my-circle draw
10574: @end example
1.1 anton 10575:
1.78 anton 10576: @cindex selector invocation, restrictions
10577: @cindex class definition, restrictions
10578: Note: You can only invoke a selector if the receiving object belongs to
10579: the class where the selector was defined or one of its descendents;
10580: e.g., you can invoke @code{draw} only for objects belonging to
10581: @code{graphical} or its descendents (e.g., @code{circle}). The scoping
10582: mechanism will check if you try to invoke a selector that is not
10583: defined in this class hierarchy, so you'll get an error at compilation
10584: time.
1.1 anton 10585:
10586:
1.78 anton 10587: @node The OOF base class, Class Declaration, Basic OOF Usage, OOF
10588: @subsubsection The @file{oof.fs} base class
10589: @cindex @file{oof.fs} base class
1.1 anton 10590:
1.78 anton 10591: When you define a class, you have to specify a parent class. So how do
10592: you start defining classes? There is one class available from the start:
10593: @code{object}. You have to use it as ancestor for all classes. It is the
10594: only class that has no parent. Classes are also objects, except that
10595: they don't have instance variables; class manipulation such as
10596: inheritance or changing definitions of a class is handled through
10597: selectors of the class @code{object}.
1.1 anton 10598:
1.78 anton 10599: @code{object} provides a number of selectors:
1.1 anton 10600:
1.78 anton 10601: @itemize @bullet
10602: @item
10603: @code{class} for subclassing, @code{definitions} to add definitions
10604: later on, and @code{class?} to get type informations (is the class a
10605: subclass of the class passed on the stack?).
1.1 anton 10606:
1.78 anton 10607: doc---object-class
10608: doc---object-definitions
10609: doc---object-class?
1.1 anton 10610:
10611:
1.26 crook 10612: @item
1.78 anton 10613: @code{init} and @code{dispose} as constructor and destructor of the
10614: object. @code{init} is invocated after the object's memory is allocated,
10615: while @code{dispose} also handles deallocation. Thus if you redefine
10616: @code{dispose}, you have to call the parent's dispose with @code{super
10617: dispose}, too.
10618:
10619: doc---object-init
10620: doc---object-dispose
10621:
1.1 anton 10622:
1.26 crook 10623: @item
1.78 anton 10624: @code{new}, @code{new[]}, @code{:}, @code{ptr}, @code{asptr}, and
10625: @code{[]} to create named and unnamed objects and object arrays or
10626: object pointers.
10627:
10628: doc---object-new
10629: doc---object-new[]
10630: doc---object-:
10631: doc---object-ptr
10632: doc---object-asptr
10633: doc---object-[]
10634:
1.1 anton 10635:
1.26 crook 10636: @item
1.78 anton 10637: @code{::} and @code{super} for explicit scoping. You should use explicit
10638: scoping only for super classes or classes with the same set of instance
10639: variables. Explicitly-scoped selectors use early binding.
1.21 crook 10640:
1.78 anton 10641: doc---object-::
10642: doc---object-super
1.21 crook 10643:
10644:
1.26 crook 10645: @item
1.78 anton 10646: @code{self} to get the address of the object
1.21 crook 10647:
1.78 anton 10648: doc---object-self
1.21 crook 10649:
10650:
1.78 anton 10651: @item
10652: @code{bind}, @code{bound}, @code{link}, and @code{is} to assign object
10653: pointers and instance defers.
1.21 crook 10654:
1.78 anton 10655: doc---object-bind
10656: doc---object-bound
10657: doc---object-link
10658: doc---object-is
1.21 crook 10659:
10660:
1.78 anton 10661: @item
10662: @code{'} to obtain selector tokens, @code{send} to invocate selectors
10663: form the stack, and @code{postpone} to generate selector invocation code.
1.21 crook 10664:
1.78 anton 10665: doc---object-'
10666: doc---object-postpone
1.21 crook 10667:
10668:
1.78 anton 10669: @item
10670: @code{with} and @code{endwith} to select the active object from the
10671: stack, and enable its scope. Using @code{with} and @code{endwith}
10672: also allows you to create code using selector @code{postpone} without being
10673: trapped by the state-smart objects.
1.21 crook 10674:
1.78 anton 10675: doc---object-with
10676: doc---object-endwith
1.21 crook 10677:
10678:
1.78 anton 10679: @end itemize
1.21 crook 10680:
1.78 anton 10681: @node Class Declaration, Class Implementation, The OOF base class, OOF
10682: @subsubsection Class Declaration
10683: @cindex class declaration
1.21 crook 10684:
1.78 anton 10685: @itemize @bullet
10686: @item
10687: Instance variables
1.21 crook 10688:
1.78 anton 10689: doc---oof-var
1.21 crook 10690:
10691:
1.78 anton 10692: @item
10693: Object pointers
1.21 crook 10694:
1.78 anton 10695: doc---oof-ptr
10696: doc---oof-asptr
1.21 crook 10697:
10698:
1.78 anton 10699: @item
10700: Instance defers
1.21 crook 10701:
1.78 anton 10702: doc---oof-defer
1.21 crook 10703:
10704:
1.78 anton 10705: @item
10706: Method selectors
1.21 crook 10707:
1.78 anton 10708: doc---oof-early
10709: doc---oof-method
1.21 crook 10710:
10711:
1.78 anton 10712: @item
10713: Class-wide variables
1.21 crook 10714:
1.78 anton 10715: doc---oof-static
1.21 crook 10716:
10717:
1.78 anton 10718: @item
10719: End declaration
1.1 anton 10720:
1.78 anton 10721: doc---oof-how:
10722: doc---oof-class;
1.21 crook 10723:
10724:
1.78 anton 10725: @end itemize
1.21 crook 10726:
1.78 anton 10727: @c -------------------------------------------------------------
10728: @node Class Implementation, , Class Declaration, OOF
10729: @subsubsection Class Implementation
10730: @cindex class implementation
1.21 crook 10731:
1.78 anton 10732: @c -------------------------------------------------------------
10733: @node Mini-OOF, Comparison with other object models, OOF, Object-oriented Forth
10734: @subsection The @file{mini-oof.fs} model
10735: @cindex mini-oof
1.21 crook 10736:
1.78 anton 10737: Gforth's third object oriented Forth package is a 12-liner. It uses a
1.79 anton 10738: mixture of the @file{objects.fs} and the @file{oof.fs} syntax,
1.78 anton 10739: and reduces to the bare minimum of features. This is based on a posting
10740: of Bernd Paysan in comp.lang.forth.
1.21 crook 10741:
1.78 anton 10742: @menu
10743: * Basic Mini-OOF Usage::
10744: * Mini-OOF Example::
10745: * Mini-OOF Implementation::
10746: @end menu
1.21 crook 10747:
1.78 anton 10748: @c -------------------------------------------------------------
10749: @node Basic Mini-OOF Usage, Mini-OOF Example, Mini-OOF, Mini-OOF
10750: @subsubsection Basic @file{mini-oof.fs} Usage
10751: @cindex mini-oof usage
1.21 crook 10752:
1.78 anton 10753: There is a base class (@code{class}, which allocates one cell for the
10754: object pointer) plus seven other words: to define a method, a variable,
10755: a class; to end a class, to resolve binding, to allocate an object and
10756: to compile a class method.
10757: @comment TODO better description of the last one
1.26 crook 10758:
1.21 crook 10759:
1.78 anton 10760: doc-object
10761: doc-method
10762: doc-var
10763: doc-class
10764: doc-end-class
10765: doc-defines
10766: doc-new
10767: doc-::
1.21 crook 10768:
10769:
10770:
1.78 anton 10771: @c -------------------------------------------------------------
10772: @node Mini-OOF Example, Mini-OOF Implementation, Basic Mini-OOF Usage, Mini-OOF
10773: @subsubsection Mini-OOF Example
10774: @cindex mini-oof example
1.1 anton 10775:
1.78 anton 10776: A short example shows how to use this package. This example, in slightly
10777: extended form, is supplied as @file{moof-exm.fs}
10778: @comment TODO could flesh this out with some comments from the Forthwrite article
1.20 pazsan 10779:
1.26 crook 10780: @example
1.78 anton 10781: object class
10782: method init
10783: method draw
10784: end-class graphical
1.26 crook 10785: @end example
1.20 pazsan 10786:
1.78 anton 10787: This code defines a class @code{graphical} with an
10788: operation @code{draw}. We can perform the operation
10789: @code{draw} on any @code{graphical} object, e.g.:
1.20 pazsan 10790:
1.26 crook 10791: @example
1.78 anton 10792: 100 100 t-rex draw
1.26 crook 10793: @end example
1.12 anton 10794:
1.78 anton 10795: where @code{t-rex} is an object or object pointer, created with e.g.
10796: @code{graphical new Constant t-rex}.
1.12 anton 10797:
1.78 anton 10798: For concrete graphical objects, we define child classes of the
10799: class @code{graphical}, e.g.:
1.12 anton 10800:
1.26 crook 10801: @example
10802: graphical class
1.78 anton 10803: cell var circle-radius
10804: end-class circle \ "graphical" is the parent class
1.12 anton 10805:
1.78 anton 10806: :noname ( x y -- )
10807: circle-radius @@ draw-circle ; circle defines draw
10808: :noname ( r -- )
10809: circle-radius ! ; circle defines init
10810: @end example
1.12 anton 10811:
1.78 anton 10812: There is no implicit init method, so we have to define one. The creation
10813: code of the object now has to call init explicitely.
1.21 crook 10814:
1.78 anton 10815: @example
10816: circle new Constant my-circle
10817: 50 my-circle init
1.12 anton 10818: @end example
10819:
1.78 anton 10820: It is also possible to add a function to create named objects with
10821: automatic call of @code{init}, given that all objects have @code{init}
10822: on the same place:
1.38 anton 10823:
1.78 anton 10824: @example
10825: : new: ( .. o "name" -- )
10826: new dup Constant init ;
10827: 80 circle new: large-circle
10828: @end example
1.12 anton 10829:
1.78 anton 10830: We can draw this new circle at (100,100) with:
1.12 anton 10831:
1.78 anton 10832: @example
10833: 100 100 my-circle draw
10834: @end example
1.12 anton 10835:
1.78 anton 10836: @node Mini-OOF Implementation, , Mini-OOF Example, Mini-OOF
10837: @subsubsection @file{mini-oof.fs} Implementation
1.12 anton 10838:
1.78 anton 10839: Object-oriented systems with late binding typically use a
10840: ``vtable''-approach: the first variable in each object is a pointer to a
10841: table, which contains the methods as function pointers. The vtable
10842: may also contain other information.
1.12 anton 10843:
1.79 anton 10844: So first, let's declare selectors:
1.37 anton 10845:
10846: @example
1.79 anton 10847: : method ( m v "name" -- m' v ) Create over , swap cell+ swap
1.78 anton 10848: DOES> ( ... o -- ... ) @@ over @@ + @@ execute ;
10849: @end example
1.37 anton 10850:
1.79 anton 10851: During selector declaration, the number of selectors and instance
10852: variables is on the stack (in address units). @code{method} creates one
10853: selector and increments the selector number. To execute a selector, it
1.78 anton 10854: takes the object, fetches the vtable pointer, adds the offset, and
1.79 anton 10855: executes the method @i{xt} stored there. Each selector takes the object
10856: it is invoked with as top of stack parameter; it passes the parameters
10857: (including the object) unchanged to the appropriate method which should
1.78 anton 10858: consume that object.
1.37 anton 10859:
1.78 anton 10860: Now, we also have to declare instance variables
1.37 anton 10861:
1.78 anton 10862: @example
1.79 anton 10863: : var ( m v size "name" -- m v' ) Create over , +
1.78 anton 10864: DOES> ( o -- addr ) @@ + ;
1.37 anton 10865: @end example
10866:
1.78 anton 10867: As before, a word is created with the current offset. Instance
10868: variables can have different sizes (cells, floats, doubles, chars), so
10869: all we do is take the size and add it to the offset. If your machine
10870: has alignment restrictions, put the proper @code{aligned} or
10871: @code{faligned} before the variable, to adjust the variable
10872: offset. That's why it is on the top of stack.
1.37 anton 10873:
1.78 anton 10874: We need a starting point (the base object) and some syntactic sugar:
1.37 anton 10875:
1.78 anton 10876: @example
10877: Create object 1 cells , 2 cells ,
1.79 anton 10878: : class ( class -- class selectors vars ) dup 2@@ ;
1.78 anton 10879: @end example
1.12 anton 10880:
1.78 anton 10881: For inheritance, the vtable of the parent object has to be
10882: copied when a new, derived class is declared. This gives all the
10883: methods of the parent class, which can be overridden, though.
1.12 anton 10884:
1.78 anton 10885: @example
1.79 anton 10886: : end-class ( class selectors vars "name" -- )
1.78 anton 10887: Create here >r , dup , 2 cells ?DO ['] noop , 1 cells +LOOP
10888: cell+ dup cell+ r> rot @@ 2 cells /string move ;
10889: @end example
1.12 anton 10890:
1.78 anton 10891: The first line creates the vtable, initialized with
10892: @code{noop}s. The second line is the inheritance mechanism, it
10893: copies the xts from the parent vtable.
1.12 anton 10894:
1.78 anton 10895: We still have no way to define new methods, let's do that now:
1.12 anton 10896:
1.26 crook 10897: @example
1.79 anton 10898: : defines ( xt class "name" -- ) ' >body @@ + ! ;
1.78 anton 10899: @end example
1.12 anton 10900:
1.78 anton 10901: To allocate a new object, we need a word, too:
1.12 anton 10902:
1.78 anton 10903: @example
10904: : new ( class -- o ) here over @@ allot swap over ! ;
1.12 anton 10905: @end example
10906:
1.78 anton 10907: Sometimes derived classes want to access the method of the
10908: parent object. There are two ways to achieve this with Mini-OOF:
10909: first, you could use named words, and second, you could look up the
10910: vtable of the parent object.
1.12 anton 10911:
1.78 anton 10912: @example
10913: : :: ( class "name" -- ) ' >body @@ + @@ compile, ;
10914: @end example
1.12 anton 10915:
10916:
1.78 anton 10917: Nothing can be more confusing than a good example, so here is
10918: one. First let's declare a text object (called
10919: @code{button}), that stores text and position:
1.12 anton 10920:
1.78 anton 10921: @example
10922: object class
10923: cell var text
10924: cell var len
10925: cell var x
10926: cell var y
10927: method init
10928: method draw
10929: end-class button
10930: @end example
1.12 anton 10931:
1.78 anton 10932: @noindent
10933: Now, implement the two methods, @code{draw} and @code{init}:
1.21 crook 10934:
1.26 crook 10935: @example
1.78 anton 10936: :noname ( o -- )
10937: >r r@@ x @@ r@@ y @@ at-xy r@@ text @@ r> len @@ type ;
10938: button defines draw
10939: :noname ( addr u o -- )
10940: >r 0 r@@ x ! 0 r@@ y ! r@@ len ! r> text ! ;
10941: button defines init
1.26 crook 10942: @end example
1.12 anton 10943:
1.78 anton 10944: @noindent
10945: To demonstrate inheritance, we define a class @code{bold-button}, with no
1.79 anton 10946: new data and no new selectors:
1.78 anton 10947:
10948: @example
10949: button class
10950: end-class bold-button
1.12 anton 10951:
1.78 anton 10952: : bold 27 emit ." [1m" ;
10953: : normal 27 emit ." [0m" ;
10954: @end example
1.1 anton 10955:
1.78 anton 10956: @noindent
10957: The class @code{bold-button} has a different draw method to
10958: @code{button}, but the new method is defined in terms of the draw method
10959: for @code{button}:
1.20 pazsan 10960:
1.78 anton 10961: @example
10962: :noname bold [ button :: draw ] normal ; bold-button defines draw
10963: @end example
1.21 crook 10964:
1.78 anton 10965: @noindent
1.79 anton 10966: Finally, create two objects and apply selectors:
1.21 crook 10967:
1.26 crook 10968: @example
1.78 anton 10969: button new Constant foo
10970: s" thin foo" foo init
10971: page
10972: foo draw
10973: bold-button new Constant bar
10974: s" fat bar" bar init
10975: 1 bar y !
10976: bar draw
1.26 crook 10977: @end example
1.21 crook 10978:
10979:
1.78 anton 10980: @node Comparison with other object models, , Mini-OOF, Object-oriented Forth
10981: @subsection Comparison with other object models
10982: @cindex comparison of object models
10983: @cindex object models, comparison
10984:
10985: Many object-oriented Forth extensions have been proposed (@cite{A survey
10986: of object-oriented Forths} (SIGPLAN Notices, April 1996) by Bradford
10987: J. Rodriguez and W. F. S. Poehlman lists 17). This section discusses the
10988: relation of the object models described here to two well-known and two
10989: closely-related (by the use of method maps) models. Andras Zsoter
10990: helped us with this section.
10991:
10992: @cindex Neon model
10993: The most popular model currently seems to be the Neon model (see
10994: @cite{Object-oriented programming in ANS Forth} (Forth Dimensions, March
10995: 1997) by Andrew McKewan) but this model has a number of limitations
10996: @footnote{A longer version of this critique can be
10997: found in @cite{On Standardizing Object-Oriented Forth Extensions} (Forth
10998: Dimensions, May 1997) by Anton Ertl.}:
10999:
11000: @itemize @bullet
11001: @item
11002: It uses a @code{@emph{selector object}} syntax, which makes it unnatural
11003: to pass objects on the stack.
1.21 crook 11004:
1.78 anton 11005: @item
11006: It requires that the selector parses the input stream (at
1.79 anton 11007: compile time); this leads to reduced extensibility and to bugs that are
1.78 anton 11008: hard to find.
1.21 crook 11009:
1.78 anton 11010: @item
1.79 anton 11011: It allows using every selector on every object; this eliminates the
11012: need for interfaces, but makes it harder to create efficient
11013: implementations.
1.78 anton 11014: @end itemize
1.21 crook 11015:
1.78 anton 11016: @cindex Pountain's object-oriented model
11017: Another well-known publication is @cite{Object-Oriented Forth} (Academic
11018: Press, London, 1987) by Dick Pountain. However, it is not really about
11019: object-oriented programming, because it hardly deals with late
11020: binding. Instead, it focuses on features like information hiding and
11021: overloading that are characteristic of modular languages like Ada (83).
1.26 crook 11022:
1.78 anton 11023: @cindex Zsoter's object-oriented model
1.79 anton 11024: In @uref{http://www.forth.org/oopf.html, Does late binding have to be
11025: slow?} (Forth Dimensions 18(1) 1996, pages 31-35) Andras Zsoter
11026: describes a model that makes heavy use of an active object (like
11027: @code{this} in @file{objects.fs}): The active object is not only used
11028: for accessing all fields, but also specifies the receiving object of
11029: every selector invocation; you have to change the active object
11030: explicitly with @code{@{ ... @}}, whereas in @file{objects.fs} it
11031: changes more or less implicitly at @code{m: ... ;m}. Such a change at
11032: the method entry point is unnecessary with Zsoter's model, because the
11033: receiving object is the active object already. On the other hand, the
11034: explicit change is absolutely necessary in that model, because otherwise
11035: no one could ever change the active object. An ANS Forth implementation
11036: of this model is available through
11037: @uref{http://www.forth.org/oopf.html}.
1.21 crook 11038:
1.78 anton 11039: @cindex @file{oof.fs}, differences to other models
11040: The @file{oof.fs} model combines information hiding and overloading
11041: resolution (by keeping names in various word lists) with object-oriented
11042: programming. It sets the active object implicitly on method entry, but
11043: also allows explicit changing (with @code{>o...o>} or with
11044: @code{with...endwith}). It uses parsing and state-smart objects and
11045: classes for resolving overloading and for early binding: the object or
11046: class parses the selector and determines the method from this. If the
11047: selector is not parsed by an object or class, it performs a call to the
11048: selector for the active object (late binding), like Zsoter's model.
11049: Fields are always accessed through the active object. The big
11050: disadvantage of this model is the parsing and the state-smartness, which
11051: reduces extensibility and increases the opportunities for subtle bugs;
11052: essentially, you are only safe if you never tick or @code{postpone} an
11053: object or class (Bernd disagrees, but I (Anton) am not convinced).
1.21 crook 11054:
1.78 anton 11055: @cindex @file{mini-oof.fs}, differences to other models
11056: The @file{mini-oof.fs} model is quite similar to a very stripped-down
11057: version of the @file{objects.fs} model, but syntactically it is a
11058: mixture of the @file{objects.fs} and @file{oof.fs} models.
1.21 crook 11059:
11060:
1.78 anton 11061: @c -------------------------------------------------------------
11062: @node Programming Tools, Assembler and Code Words, Object-oriented Forth, Words
11063: @section Programming Tools
11064: @cindex programming tools
1.21 crook 11065:
1.78 anton 11066: @c !! move this and assembler down below OO stuff.
1.21 crook 11067:
1.78 anton 11068: @menu
11069: * Examining::
11070: * Forgetting words::
11071: * Debugging:: Simple and quick.
11072: * Assertions:: Making your programs self-checking.
11073: * Singlestep Debugger:: Executing your program word by word.
11074: @end menu
1.21 crook 11075:
1.78 anton 11076: @node Examining, Forgetting words, Programming Tools, Programming Tools
11077: @subsection Examining data and code
11078: @cindex examining data and code
11079: @cindex data examination
11080: @cindex code examination
1.44 crook 11081:
1.78 anton 11082: The following words inspect the stack non-destructively:
1.21 crook 11083:
1.78 anton 11084: doc-.s
11085: doc-f.s
1.44 crook 11086:
1.78 anton 11087: There is a word @code{.r} but it does @i{not} display the return stack!
11088: It is used for formatted numeric output (@pxref{Simple numeric output}).
1.21 crook 11089:
1.78 anton 11090: doc-depth
11091: doc-fdepth
11092: doc-clearstack
1.21 crook 11093:
1.78 anton 11094: The following words inspect memory.
1.21 crook 11095:
1.78 anton 11096: doc-?
11097: doc-dump
1.21 crook 11098:
1.78 anton 11099: And finally, @code{see} allows to inspect code:
1.21 crook 11100:
1.78 anton 11101: doc-see
11102: doc-xt-see
1.21 crook 11103:
1.78 anton 11104: @node Forgetting words, Debugging, Examining, Programming Tools
11105: @subsection Forgetting words
11106: @cindex words, forgetting
11107: @cindex forgeting words
1.21 crook 11108:
1.78 anton 11109: @c anton: other, maybe better places for this subsection: Defining Words;
11110: @c Dictionary allocation. At least a reference should be there.
1.21 crook 11111:
1.78 anton 11112: Forth allows you to forget words (and everything that was alloted in the
11113: dictonary after them) in a LIFO manner.
1.21 crook 11114:
1.78 anton 11115: doc-marker
1.21 crook 11116:
1.78 anton 11117: The most common use of this feature is during progam development: when
11118: you change a source file, forget all the words it defined and load it
11119: again (since you also forget everything defined after the source file
11120: was loaded, you have to reload that, too). Note that effects like
11121: storing to variables and destroyed system words are not undone when you
11122: forget words. With a system like Gforth, that is fast enough at
11123: starting up and compiling, I find it more convenient to exit and restart
11124: Gforth, as this gives me a clean slate.
1.21 crook 11125:
1.78 anton 11126: Here's an example of using @code{marker} at the start of a source file
11127: that you are debugging; it ensures that you only ever have one copy of
11128: the file's definitions compiled at any time:
1.21 crook 11129:
1.78 anton 11130: @example
11131: [IFDEF] my-code
11132: my-code
11133: [ENDIF]
1.26 crook 11134:
1.78 anton 11135: marker my-code
11136: init-included-files
1.21 crook 11137:
1.78 anton 11138: \ .. definitions start here
11139: \ .
11140: \ .
11141: \ end
11142: @end example
1.21 crook 11143:
1.26 crook 11144:
1.78 anton 11145: @node Debugging, Assertions, Forgetting words, Programming Tools
11146: @subsection Debugging
11147: @cindex debugging
1.21 crook 11148:
1.78 anton 11149: Languages with a slow edit/compile/link/test development loop tend to
11150: require sophisticated tracing/stepping debuggers to facilate debugging.
1.21 crook 11151:
1.78 anton 11152: A much better (faster) way in fast-compiling languages is to add
11153: printing code at well-selected places, let the program run, look at
11154: the output, see where things went wrong, add more printing code, etc.,
11155: until the bug is found.
1.21 crook 11156:
1.78 anton 11157: The simple debugging aids provided in @file{debugs.fs}
11158: are meant to support this style of debugging.
1.21 crook 11159:
1.78 anton 11160: The word @code{~~} prints debugging information (by default the source
11161: location and the stack contents). It is easy to insert. If you use Emacs
11162: it is also easy to remove (@kbd{C-x ~} in the Emacs Forth mode to
11163: query-replace them with nothing). The deferred words
11164: @code{printdebugdata} and @code{printdebugline} control the output of
11165: @code{~~}. The default source location output format works well with
11166: Emacs' compilation mode, so you can step through the program at the
11167: source level using @kbd{C-x `} (the advantage over a stepping debugger
11168: is that you can step in any direction and you know where the crash has
11169: happened or where the strange data has occurred).
1.21 crook 11170:
1.78 anton 11171: doc-~~
11172: doc-printdebugdata
11173: doc-printdebugline
1.21 crook 11174:
1.78 anton 11175: @node Assertions, Singlestep Debugger, Debugging, Programming Tools
11176: @subsection Assertions
11177: @cindex assertions
1.21 crook 11178:
1.78 anton 11179: It is a good idea to make your programs self-checking, especially if you
11180: make an assumption that may become invalid during maintenance (for
11181: example, that a certain field of a data structure is never zero). Gforth
11182: supports @dfn{assertions} for this purpose. They are used like this:
1.21 crook 11183:
11184: @example
1.78 anton 11185: assert( @i{flag} )
1.26 crook 11186: @end example
11187:
1.78 anton 11188: The code between @code{assert(} and @code{)} should compute a flag, that
11189: should be true if everything is alright and false otherwise. It should
11190: not change anything else on the stack. The overall stack effect of the
11191: assertion is @code{( -- )}. E.g.
1.21 crook 11192:
1.26 crook 11193: @example
1.78 anton 11194: assert( 1 1 + 2 = ) \ what we learn in school
11195: assert( dup 0<> ) \ assert that the top of stack is not zero
11196: assert( false ) \ this code should not be reached
1.21 crook 11197: @end example
11198:
1.78 anton 11199: The need for assertions is different at different times. During
11200: debugging, we want more checking, in production we sometimes care more
11201: for speed. Therefore, assertions can be turned off, i.e., the assertion
11202: becomes a comment. Depending on the importance of an assertion and the
11203: time it takes to check it, you may want to turn off some assertions and
11204: keep others turned on. Gforth provides several levels of assertions for
11205: this purpose:
11206:
11207:
11208: doc-assert0(
11209: doc-assert1(
11210: doc-assert2(
11211: doc-assert3(
11212: doc-assert(
11213: doc-)
1.21 crook 11214:
11215:
1.78 anton 11216: The variable @code{assert-level} specifies the highest assertions that
11217: are turned on. I.e., at the default @code{assert-level} of one,
11218: @code{assert0(} and @code{assert1(} assertions perform checking, while
11219: @code{assert2(} and @code{assert3(} assertions are treated as comments.
1.26 crook 11220:
1.78 anton 11221: The value of @code{assert-level} is evaluated at compile-time, not at
11222: run-time. Therefore you cannot turn assertions on or off at run-time;
11223: you have to set the @code{assert-level} appropriately before compiling a
11224: piece of code. You can compile different pieces of code at different
11225: @code{assert-level}s (e.g., a trusted library at level 1 and
11226: newly-written code at level 3).
1.26 crook 11227:
11228:
1.78 anton 11229: doc-assert-level
1.26 crook 11230:
11231:
1.78 anton 11232: If an assertion fails, a message compatible with Emacs' compilation mode
11233: is produced and the execution is aborted (currently with @code{ABORT"}.
11234: If there is interest, we will introduce a special throw code. But if you
11235: intend to @code{catch} a specific condition, using @code{throw} is
11236: probably more appropriate than an assertion).
1.44 crook 11237:
1.78 anton 11238: Definitions in ANS Forth for these assertion words are provided
11239: in @file{compat/assert.fs}.
1.26 crook 11240:
1.44 crook 11241:
1.78 anton 11242: @node Singlestep Debugger, , Assertions, Programming Tools
11243: @subsection Singlestep Debugger
11244: @cindex singlestep Debugger
11245: @cindex debugging Singlestep
1.44 crook 11246:
1.78 anton 11247: When you create a new word there's often the need to check whether it
11248: behaves correctly or not. You can do this by typing @code{dbg
11249: badword}. A debug session might look like this:
1.26 crook 11250:
1.78 anton 11251: @example
11252: : badword 0 DO i . LOOP ; ok
11253: 2 dbg badword
11254: : badword
11255: Scanning code...
1.44 crook 11256:
1.78 anton 11257: Nesting debugger ready!
1.44 crook 11258:
1.78 anton 11259: 400D4738 8049BC4 0 -> [ 2 ] 00002 00000
11260: 400D4740 8049F68 DO -> [ 0 ]
11261: 400D4744 804A0C8 i -> [ 1 ] 00000
11262: 400D4748 400C5E60 . -> 0 [ 0 ]
11263: 400D474C 8049D0C LOOP -> [ 0 ]
11264: 400D4744 804A0C8 i -> [ 1 ] 00001
11265: 400D4748 400C5E60 . -> 1 [ 0 ]
11266: 400D474C 8049D0C LOOP -> [ 0 ]
11267: 400D4758 804B384 ; -> ok
11268: @end example
1.21 crook 11269:
1.78 anton 11270: Each line displayed is one step. You always have to hit return to
11271: execute the next word that is displayed. If you don't want to execute
11272: the next word in a whole, you have to type @kbd{n} for @code{nest}. Here is
11273: an overview what keys are available:
1.44 crook 11274:
1.78 anton 11275: @table @i
1.44 crook 11276:
1.78 anton 11277: @item @key{RET}
11278: Next; Execute the next word.
1.21 crook 11279:
1.78 anton 11280: @item n
11281: Nest; Single step through next word.
1.44 crook 11282:
1.78 anton 11283: @item u
11284: Unnest; Stop debugging and execute rest of word. If we got to this word
11285: with nest, continue debugging with the calling word.
1.44 crook 11286:
1.78 anton 11287: @item d
11288: Done; Stop debugging and execute rest.
1.21 crook 11289:
1.78 anton 11290: @item s
11291: Stop; Abort immediately.
1.44 crook 11292:
1.78 anton 11293: @end table
1.44 crook 11294:
1.78 anton 11295: Debugging large application with this mechanism is very difficult, because
11296: you have to nest very deeply into the program before the interesting part
11297: begins. This takes a lot of time.
1.26 crook 11298:
1.78 anton 11299: To do it more directly put a @code{BREAK:} command into your source code.
11300: When program execution reaches @code{BREAK:} the single step debugger is
11301: invoked and you have all the features described above.
1.44 crook 11302:
1.78 anton 11303: If you have more than one part to debug it is useful to know where the
11304: program has stopped at the moment. You can do this by the
11305: @code{BREAK" string"} command. This behaves like @code{BREAK:} except that
11306: string is typed out when the ``breakpoint'' is reached.
1.44 crook 11307:
1.26 crook 11308:
1.78 anton 11309: doc-dbg
11310: doc-break:
11311: doc-break"
1.44 crook 11312:
11313:
1.26 crook 11314:
1.78 anton 11315: @c -------------------------------------------------------------
11316: @node Assembler and Code Words, Threading Words, Programming Tools, Words
11317: @section Assembler and Code Words
11318: @cindex assembler
11319: @cindex code words
1.44 crook 11320:
1.78 anton 11321: @menu
11322: * Code and ;code::
11323: * Common Assembler:: Assembler Syntax
11324: * Common Disassembler::
11325: * 386 Assembler:: Deviations and special cases
11326: * Alpha Assembler:: Deviations and special cases
11327: * MIPS assembler:: Deviations and special cases
11328: * Other assemblers:: How to write them
11329: @end menu
1.21 crook 11330:
1.78 anton 11331: @node Code and ;code, Common Assembler, Assembler and Code Words, Assembler and Code Words
11332: @subsection @code{Code} and @code{;code}
1.26 crook 11333:
1.78 anton 11334: Gforth provides some words for defining primitives (words written in
11335: machine code), and for defining the machine-code equivalent of
11336: @code{DOES>}-based defining words. However, the machine-independent
11337: nature of Gforth poses a few problems: First of all, Gforth runs on
11338: several architectures, so it can provide no standard assembler. What's
11339: worse is that the register allocation not only depends on the processor,
11340: but also on the @code{gcc} version and options used.
1.44 crook 11341:
1.78 anton 11342: The words that Gforth offers encapsulate some system dependences (e.g.,
11343: the header structure), so a system-independent assembler may be used in
11344: Gforth. If you do not have an assembler, you can compile machine code
11345: directly with @code{,} and @code{c,}@footnote{This isn't portable,
11346: because these words emit stuff in @i{data} space; it works because
11347: Gforth has unified code/data spaces. Assembler isn't likely to be
11348: portable anyway.}.
1.21 crook 11349:
1.44 crook 11350:
1.78 anton 11351: doc-assembler
11352: doc-init-asm
11353: doc-code
11354: doc-end-code
11355: doc-;code
11356: doc-flush-icache
1.44 crook 11357:
1.21 crook 11358:
1.78 anton 11359: If @code{flush-icache} does not work correctly, @code{code} words
11360: etc. will not work (reliably), either.
1.44 crook 11361:
1.78 anton 11362: The typical usage of these @code{code} words can be shown most easily by
11363: analogy to the equivalent high-level defining words:
1.44 crook 11364:
1.78 anton 11365: @example
11366: : foo code foo
11367: <high-level Forth words> <assembler>
11368: ; end-code
11369:
11370: : bar : bar
11371: <high-level Forth words> <high-level Forth words>
11372: CREATE CREATE
11373: <high-level Forth words> <high-level Forth words>
11374: DOES> ;code
11375: <high-level Forth words> <assembler>
11376: ; end-code
11377: @end example
1.21 crook 11378:
1.78 anton 11379: @c anton: the following stuff is also in "Common Assembler", in less detail.
1.44 crook 11380:
1.78 anton 11381: @cindex registers of the inner interpreter
11382: In the assembly code you will want to refer to the inner interpreter's
11383: registers (e.g., the data stack pointer) and you may want to use other
11384: registers for temporary storage. Unfortunately, the register allocation
11385: is installation-dependent.
1.44 crook 11386:
1.78 anton 11387: In particular, @code{ip} (Forth instruction pointer) and @code{rp}
11388: (return stack pointer) are in different places in @code{gforth} and
11389: @code{gforth-fast}. This means that you cannot write a @code{NEXT}
11390: routine that works on both versions; so for doing @code{NEXT}, I
11391: recomment jumping to @code{' noop >code-address}, which contains nothing
11392: but a @code{NEXT}.
1.21 crook 11393:
1.78 anton 11394: For general accesses to the inner interpreter's registers, the easiest
11395: solution is to use explicit register declarations (@pxref{Explicit Reg
11396: Vars, , Variables in Specified Registers, gcc.info, GNU C Manual}) for
11397: all of the inner interpreter's registers: You have to compile Gforth
11398: with @code{-DFORCE_REG} (configure option @code{--enable-force-reg}) and
11399: the appropriate declarations must be present in the @code{machine.h}
11400: file (see @code{mips.h} for an example; you can find a full list of all
11401: declarable register symbols with @code{grep register engine.c}). If you
11402: give explicit registers to all variables that are declared at the
11403: beginning of @code{engine()}, you should be able to use the other
11404: caller-saved registers for temporary storage. Alternatively, you can use
11405: the @code{gcc} option @code{-ffixed-REG} (@pxref{Code Gen Options, ,
11406: Options for Code Generation Conventions, gcc.info, GNU C Manual}) to
11407: reserve a register (however, this restriction on register allocation may
11408: slow Gforth significantly).
1.44 crook 11409:
1.78 anton 11410: If this solution is not viable (e.g., because @code{gcc} does not allow
11411: you to explicitly declare all the registers you need), you have to find
11412: out by looking at the code where the inner interpreter's registers
11413: reside and which registers can be used for temporary storage. You can
11414: get an assembly listing of the engine's code with @code{make engine.s}.
1.44 crook 11415:
1.78 anton 11416: In any case, it is good practice to abstract your assembly code from the
11417: actual register allocation. E.g., if the data stack pointer resides in
11418: register @code{$17}, create an alias for this register called @code{sp},
11419: and use that in your assembly code.
1.21 crook 11420:
1.78 anton 11421: @cindex code words, portable
11422: Another option for implementing normal and defining words efficiently
11423: is to add the desired functionality to the source of Gforth. For normal
11424: words you just have to edit @file{primitives} (@pxref{Automatic
11425: Generation}). Defining words (equivalent to @code{;CODE} words, for fast
11426: defined words) may require changes in @file{engine.c}, @file{kernel.fs},
11427: @file{prims2x.fs}, and possibly @file{cross.fs}.
1.44 crook 11428:
1.78 anton 11429: @node Common Assembler, Common Disassembler, Code and ;code, Assembler and Code Words
11430: @subsection Common Assembler
1.44 crook 11431:
1.78 anton 11432: The assemblers in Gforth generally use a postfix syntax, i.e., the
11433: instruction name follows the operands.
1.21 crook 11434:
1.78 anton 11435: The operands are passed in the usual order (the same that is used in the
11436: manual of the architecture). Since they all are Forth words, they have
11437: to be separated by spaces; you can also use Forth words to compute the
11438: operands.
1.44 crook 11439:
1.78 anton 11440: The instruction names usually end with a @code{,}. This makes it easier
11441: to visually separate instructions if you put several of them on one
11442: line; it also avoids shadowing other Forth words (e.g., @code{and}).
1.21 crook 11443:
1.78 anton 11444: Registers are usually specified by number; e.g., (decimal) @code{11}
11445: specifies registers R11 and F11 on the Alpha architecture (which one,
11446: depends on the instruction). The usual names are also available, e.g.,
11447: @code{s2} for R11 on Alpha.
1.21 crook 11448:
1.78 anton 11449: Control flow is specified similar to normal Forth code (@pxref{Arbitrary
11450: control structures}), with @code{if,}, @code{ahead,}, @code{then,},
11451: @code{begin,}, @code{until,}, @code{again,}, @code{cs-roll},
11452: @code{cs-pick}, @code{else,}, @code{while,}, and @code{repeat,}. The
11453: conditions are specified in a way specific to each assembler.
1.1 anton 11454:
1.78 anton 11455: Note that the register assignments of the Gforth engine can change
11456: between Gforth versions, or even between different compilations of the
11457: same Gforth version (e.g., if you use a different GCC version). So if
11458: you want to refer to Gforth's registers (e.g., the stack pointer or
11459: TOS), I recommend defining your own words for refering to these
11460: registers, and using them later on; then you can easily adapt to a
11461: changed register assignment. The stability of the register assignment
11462: is usually better if you build Gforth with @code{--enable-force-reg}.
1.1 anton 11463:
1.78 anton 11464: In particular, the return stack pointer and the instruction pointer are
11465: in memory in @code{gforth}, and usually in registers in
11466: @code{gforth-fast}. The most common use of these registers is to
11467: dispatch to the next word (the @code{next} routine). A portable way to
11468: do this is to jump to @code{' noop >code-address} (of course, this is
11469: less efficient than integrating the @code{next} code and scheduling it
11470: well).
1.1 anton 11471:
1.78 anton 11472: @node Common Disassembler, 386 Assembler, Common Assembler, Assembler and Code Words
11473: @subsection Common Disassembler
1.1 anton 11474:
1.78 anton 11475: You can disassemble a @code{code} word with @code{see}
11476: (@pxref{Debugging}). You can disassemble a section of memory with
1.1 anton 11477:
1.78 anton 11478: doc-disasm
1.44 crook 11479:
1.78 anton 11480: The disassembler generally produces output that can be fed into the
11481: assembler (i.e., same syntax, etc.). It also includes additional
11482: information in comments. In particular, the address of the instruction
11483: is given in a comment before the instruction.
1.1 anton 11484:
1.78 anton 11485: @code{See} may display more or less than the actual code of the word,
11486: because the recognition of the end of the code is unreliable. You can
11487: use @code{disasm} if it did not display enough. It may display more, if
11488: the code word is not immediately followed by a named word. If you have
11489: something else there, you can follow the word with @code{align last @ ,}
11490: to ensure that the end is recognized.
1.21 crook 11491:
1.78 anton 11492: @node 386 Assembler, Alpha Assembler, Common Disassembler, Assembler and Code Words
11493: @subsection 386 Assembler
1.44 crook 11494:
1.78 anton 11495: The 386 assembler included in Gforth was written by Bernd Paysan, it's
11496: available under GPL, and originally part of bigFORTH.
1.21 crook 11497:
1.78 anton 11498: The 386 disassembler included in Gforth was written by Andrew McKewan
11499: and is in the public domain.
1.21 crook 11500:
1.78 anton 11501: The disassembler displays code in prefix Intel syntax.
1.21 crook 11502:
1.78 anton 11503: The assembler uses a postfix syntax with reversed parameters.
1.1 anton 11504:
1.78 anton 11505: The assembler includes all instruction of the Athlon, i.e. 486 core
11506: instructions, Pentium and PPro extensions, floating point, MMX, 3Dnow!,
11507: but not ISSE. It's an integrated 16- and 32-bit assembler. Default is 32
11508: bit, you can switch to 16 bit with .86 and back to 32 bit with .386.
1.1 anton 11509:
1.78 anton 11510: There are several prefixes to switch between different operation sizes,
11511: @code{.b} for byte accesses, @code{.w} for word accesses, @code{.d} for
11512: double-word accesses. Addressing modes can be switched with @code{.wa}
11513: for 16 bit addresses, and @code{.da} for 32 bit addresses. You don't
11514: need a prefix for byte register names (@code{AL} et al).
1.1 anton 11515:
1.78 anton 11516: For floating point operations, the prefixes are @code{.fs} (IEEE
11517: single), @code{.fl} (IEEE double), @code{.fx} (extended), @code{.fw}
11518: (word), @code{.fd} (double-word), and @code{.fq} (quad-word).
1.21 crook 11519:
1.78 anton 11520: The MMX opcodes don't have size prefixes, they are spelled out like in
11521: the Intel assembler. Instead of move from and to memory, there are
11522: PLDQ/PLDD and PSTQ/PSTD.
1.21 crook 11523:
1.78 anton 11524: The registers lack the 'e' prefix; even in 32 bit mode, eax is called
11525: ax. Immediate values are indicated by postfixing them with @code{#},
11526: e.g., @code{3 #}. Here are some examples of addressing modes:
1.21 crook 11527:
1.26 crook 11528: @example
1.78 anton 11529: 3 # \ immediate
11530: ax \ register
11531: 100 di d) \ 100[edi]
11532: 4 bx cx di) \ 4[ebx][ecx]
11533: di ax *4 i) \ [edi][eax*4]
11534: 20 ax *4 i#) \ 20[eax*4]
1.26 crook 11535: @end example
1.21 crook 11536:
1.78 anton 11537: Some example of instructions are:
1.1 anton 11538:
11539: @example
1.78 anton 11540: ax bx mov \ move ebx,eax
11541: 3 # ax mov \ mov eax,3
11542: 100 di ) ax mov \ mov eax,100[edi]
11543: 4 bx cx di) ax mov \ mov eax,4[ebx][ecx]
11544: .w ax bx mov \ mov bx,ax
1.1 anton 11545: @end example
11546:
1.78 anton 11547: The following forms are supported for binary instructions:
1.1 anton 11548:
11549: @example
1.78 anton 11550: <reg> <reg> <inst>
11551: <n> # <reg> <inst>
11552: <mem> <reg> <inst>
11553: <reg> <mem> <inst>
1.1 anton 11554: @end example
11555:
1.78 anton 11556: Immediate to memory is not supported. The shift/rotate syntax is:
1.1 anton 11557:
1.26 crook 11558: @example
1.78 anton 11559: <reg/mem> 1 # shl \ shortens to shift without immediate
11560: <reg/mem> 4 # shl
11561: <reg/mem> cl shl
1.26 crook 11562: @end example
1.1 anton 11563:
1.78 anton 11564: Precede string instructions (@code{movs} etc.) with @code{.b} to get
11565: the byte version.
1.1 anton 11566:
1.78 anton 11567: The control structure words @code{IF} @code{UNTIL} etc. must be preceded
11568: by one of these conditions: @code{vs vc u< u>= 0= 0<> u<= u> 0< 0>= ps
11569: pc < >= <= >}. (Note that most of these words shadow some Forth words
11570: when @code{assembler} is in front of @code{forth} in the search path,
11571: e.g., in @code{code} words). Currently the control structure words use
11572: one stack item, so you have to use @code{roll} instead of @code{cs-roll}
11573: to shuffle them (you can also use @code{swap} etc.).
1.21 crook 11574:
1.78 anton 11575: Here is an example of a @code{code} word (assumes that the stack pointer
11576: is in esi and the TOS is in ebx):
1.21 crook 11577:
1.26 crook 11578: @example
1.78 anton 11579: code my+ ( n1 n2 -- n )
11580: 4 si D) bx add
11581: 4 # si add
11582: Next
11583: end-code
1.26 crook 11584: @end example
1.21 crook 11585:
1.78 anton 11586: @node Alpha Assembler, MIPS assembler, 386 Assembler, Assembler and Code Words
11587: @subsection Alpha Assembler
1.21 crook 11588:
1.78 anton 11589: The Alpha assembler and disassembler were originally written by Bernd
11590: Thallner.
1.26 crook 11591:
1.78 anton 11592: The register names @code{a0}--@code{a5} are not available to avoid
11593: shadowing hex numbers.
1.2 jwilke 11594:
1.78 anton 11595: Immediate forms of arithmetic instructions are distinguished by a
11596: @code{#} just before the @code{,}, e.g., @code{and#,} (note: @code{lda,}
11597: does not count as arithmetic instruction).
1.2 jwilke 11598:
1.78 anton 11599: You have to specify all operands to an instruction, even those that
11600: other assemblers consider optional, e.g., the destination register for
11601: @code{br,}, or the destination register and hint for @code{jmp,}.
1.2 jwilke 11602:
1.78 anton 11603: You can specify conditions for @code{if,} by removing the first @code{b}
11604: and the trailing @code{,} from a branch with a corresponding name; e.g.,
1.2 jwilke 11605:
1.26 crook 11606: @example
1.78 anton 11607: 11 fgt if, \ if F11>0e
11608: ...
11609: endif,
1.26 crook 11610: @end example
1.2 jwilke 11611:
1.78 anton 11612: @code{fbgt,} gives @code{fgt}.
11613:
11614: @node MIPS assembler, Other assemblers, Alpha Assembler, Assembler and Code Words
11615: @subsection MIPS assembler
1.2 jwilke 11616:
1.78 anton 11617: The MIPS assembler was originally written by Christian Pirker.
1.2 jwilke 11618:
1.78 anton 11619: Currently the assembler and disassembler only cover the MIPS-I
11620: architecture (R3000), and don't support FP instructions.
1.2 jwilke 11621:
1.78 anton 11622: The register names @code{$a0}--@code{$a3} are not available to avoid
11623: shadowing hex numbers.
1.2 jwilke 11624:
1.78 anton 11625: Because there is no way to distinguish registers from immediate values,
11626: you have to explicitly use the immediate forms of instructions, i.e.,
11627: @code{addiu,}, not just @code{addu,} (@command{as} does this
11628: implicitly).
1.2 jwilke 11629:
1.78 anton 11630: If the architecture manual specifies several formats for the instruction
11631: (e.g., for @code{jalr,}), you usually have to use the one with more
11632: arguments (i.e., two for @code{jalr,}). When in doubt, see
11633: @code{arch/mips/testasm.fs} for an example of correct use.
1.2 jwilke 11634:
1.78 anton 11635: Branches and jumps in the MIPS architecture have a delay slot. You have
11636: to fill it yourself (the simplest way is to use @code{nop,}), the
11637: assembler does not do it for you (unlike @command{as}). Even
11638: @code{if,}, @code{ahead,}, @code{until,}, @code{again,}, @code{while,},
11639: @code{else,} and @code{repeat,} need a delay slot. Since @code{begin,}
11640: and @code{then,} just specify branch targets, they are not affected.
1.2 jwilke 11641:
1.78 anton 11642: Note that you must not put branches, jumps, or @code{li,} into the delay
11643: slot: @code{li,} may expand to several instructions, and control flow
11644: instructions may not be put into the branch delay slot in any case.
1.2 jwilke 11645:
1.78 anton 11646: For branches the argument specifying the target is a relative address;
11647: You have to add the address of the delay slot to get the absolute
11648: address.
1.1 anton 11649:
1.78 anton 11650: The MIPS architecture also has load delay slots and restrictions on
11651: using @code{mfhi,} and @code{mflo,}; you have to order the instructions
11652: yourself to satisfy these restrictions, the assembler does not do it for
11653: you.
1.1 anton 11654:
1.78 anton 11655: You can specify the conditions for @code{if,} etc. by taking a
11656: conditional branch and leaving away the @code{b} at the start and the
11657: @code{,} at the end. E.g.,
1.1 anton 11658:
1.26 crook 11659: @example
1.78 anton 11660: 4 5 eq if,
11661: ... \ do something if $4 equals $5
11662: then,
1.26 crook 11663: @end example
1.1 anton 11664:
1.78 anton 11665: @node Other assemblers, , MIPS assembler, Assembler and Code Words
11666: @subsection Other assemblers
11667:
11668: If you want to contribute another assembler/disassembler, please contact
11669: us (@email{bug-gforth@@gnu.org}) to check if we have such an assembler
11670: already. If you are writing them from scratch, please use a similar
11671: syntax style as the one we use (i.e., postfix, commas at the end of the
11672: instruction names, @pxref{Common Assembler}); make the output of the
11673: disassembler be valid input for the assembler, and keep the style
11674: similar to the style we used.
11675:
11676: Hints on implementation: The most important part is to have a good test
11677: suite that contains all instructions. Once you have that, the rest is
11678: easy. For actual coding you can take a look at
11679: @file{arch/mips/disasm.fs} to get some ideas on how to use data for both
11680: the assembler and disassembler, avoiding redundancy and some potential
11681: bugs. You can also look at that file (and @pxref{Advanced does> usage
11682: example}) to get ideas how to factor a disassembler.
11683:
11684: Start with the disassembler, because it's easier to reuse data from the
11685: disassembler for the assembler than the other way round.
1.1 anton 11686:
1.78 anton 11687: For the assembler, take a look at @file{arch/alpha/asm.fs}, which shows
11688: how simple it can be.
1.1 anton 11689:
1.78 anton 11690: @c -------------------------------------------------------------
11691: @node Threading Words, Passing Commands to the OS, Assembler and Code Words, Words
11692: @section Threading Words
11693: @cindex threading words
1.1 anton 11694:
1.78 anton 11695: @cindex code address
11696: These words provide access to code addresses and other threading stuff
11697: in Gforth (and, possibly, other interpretive Forths). It more or less
11698: abstracts away the differences between direct and indirect threading
11699: (and, for direct threading, the machine dependences). However, at
11700: present this wordset is still incomplete. It is also pretty low-level;
11701: some day it will hopefully be made unnecessary by an internals wordset
11702: that abstracts implementation details away completely.
1.1 anton 11703:
1.78 anton 11704: The terminology used here stems from indirect threaded Forth systems; in
11705: such a system, the XT of a word is represented by the CFA (code field
11706: address) of a word; the CFA points to a cell that contains the code
11707: address. The code address is the address of some machine code that
11708: performs the run-time action of invoking the word (e.g., the
11709: @code{dovar:} routine pushes the address of the body of the word (a
11710: variable) on the stack
11711: ).
1.1 anton 11712:
1.78 anton 11713: @cindex code address
11714: @cindex code field address
11715: In an indirect threaded Forth, you can get the code address of @i{name}
11716: with @code{' @i{name} @@}; in Gforth you can get it with @code{' @i{name}
11717: >code-address}, independent of the threading method.
1.1 anton 11718:
1.78 anton 11719: doc-threading-method
11720: doc->code-address
11721: doc-code-address!
1.1 anton 11722:
1.78 anton 11723: @cindex @code{does>}-handler
11724: @cindex @code{does>}-code
11725: For a word defined with @code{DOES>}, the code address usually points to
11726: a jump instruction (the @dfn{does-handler}) that jumps to the dodoes
11727: routine (in Gforth on some platforms, it can also point to the dodoes
11728: routine itself). What you are typically interested in, though, is
11729: whether a word is a @code{DOES>}-defined word, and what Forth code it
11730: executes; @code{>does-code} tells you that.
1.1 anton 11731:
1.78 anton 11732: doc->does-code
1.1 anton 11733:
1.78 anton 11734: To create a @code{DOES>}-defined word with the following basic words,
11735: you have to set up a @code{DOES>}-handler with @code{does-handler!};
11736: @code{/does-handler} aus behind you have to place your executable Forth
11737: code. Finally you have to create a word and modify its behaviour with
11738: @code{does-handler!}.
1.1 anton 11739:
1.78 anton 11740: doc-does-code!
11741: doc-does-handler!
11742: doc-/does-handler
1.1 anton 11743:
1.78 anton 11744: The code addresses produced by various defining words are produced by
11745: the following words:
1.1 anton 11746:
1.78 anton 11747: doc-docol:
11748: doc-docon:
11749: doc-dovar:
11750: doc-douser:
11751: doc-dodefer:
11752: doc-dofield:
1.1 anton 11753:
1.26 crook 11754: @c -------------------------------------------------------------
1.78 anton 11755: @node Passing Commands to the OS, Keeping track of Time, Threading Words, Words
1.21 crook 11756: @section Passing Commands to the Operating System
11757: @cindex operating system - passing commands
11758: @cindex shell commands
11759:
11760: Gforth allows you to pass an arbitrary string to the host operating
11761: system shell (if such a thing exists) for execution.
11762:
1.44 crook 11763:
1.21 crook 11764: doc-sh
11765: doc-system
11766: doc-$?
1.23 crook 11767: doc-getenv
1.21 crook 11768:
1.44 crook 11769:
1.26 crook 11770: @c -------------------------------------------------------------
1.47 crook 11771: @node Keeping track of Time, Miscellaneous Words, Passing Commands to the OS, Words
11772: @section Keeping track of Time
11773: @cindex time-related words
11774:
11775: doc-ms
11776: doc-time&date
1.79 anton 11777: doc-utime
11778: doc-cputime
1.47 crook 11779:
11780:
11781: @c -------------------------------------------------------------
11782: @node Miscellaneous Words, , Keeping track of Time, Words
1.21 crook 11783: @section Miscellaneous Words
11784: @cindex miscellaneous words
11785:
1.29 crook 11786: @comment TODO find homes for these
11787:
1.26 crook 11788: These section lists the ANS Forth words that are not documented
1.21 crook 11789: elsewhere in this manual. Ultimately, they all need proper homes.
11790:
11791: doc-[compile]
1.68 anton 11792: doc-quit
1.44 crook 11793:
1.26 crook 11794: The following ANS Forth words are not currently supported by Gforth
1.27 crook 11795: (@pxref{ANS conformance}):
1.21 crook 11796:
11797: @code{EDITOR}
11798: @code{EMIT?}
11799: @code{FORGET}
11800:
1.24 anton 11801: @c ******************************************************************
11802: @node Error messages, Tools, Words, Top
11803: @chapter Error messages
11804: @cindex error messages
11805: @cindex backtrace
11806:
11807: A typical Gforth error message looks like this:
11808:
11809: @example
11810: in file included from :-1
11811: in file included from ./yyy.fs:1
11812: ./xxx.fs:4: Invalid memory address
11813: bar
11814: ^^^
1.79 anton 11815: Backtrace:
1.25 anton 11816: $400E664C @@
11817: $400E6664 foo
1.24 anton 11818: @end example
11819:
11820: The message identifying the error is @code{Invalid memory address}. The
11821: error happened when text-interpreting line 4 of the file
11822: @file{./xxx.fs}. This line is given (it contains @code{bar}), and the
11823: word on the line where the error happened, is pointed out (with
11824: @code{^^^}).
11825:
11826: The file containing the error was included in line 1 of @file{./yyy.fs},
11827: and @file{yyy.fs} was included from a non-file (in this case, by giving
11828: @file{yyy.fs} as command-line parameter to Gforth).
11829:
11830: At the end of the error message you find a return stack dump that can be
11831: interpreted as a backtrace (possibly empty). On top you find the top of
11832: the return stack when the @code{throw} happened, and at the bottom you
11833: find the return stack entry just above the return stack of the topmost
11834: text interpreter.
11835:
11836: To the right of most return stack entries you see a guess for the word
11837: that pushed that return stack entry as its return address. This gives a
11838: backtrace. In our case we see that @code{bar} called @code{foo}, and
11839: @code{foo} called @code{@@} (and @code{@@} had an @emph{Invalid memory
11840: address} exception).
11841:
11842: Note that the backtrace is not perfect: We don't know which return stack
11843: entries are return addresses (so we may get false positives); and in
11844: some cases (e.g., for @code{abort"}) we cannot determine from the return
11845: address the word that pushed the return address, so for some return
11846: addresses you see no names in the return stack dump.
1.25 anton 11847:
11848: @cindex @code{catch} and backtraces
11849: The return stack dump represents the return stack at the time when a
11850: specific @code{throw} was executed. In programs that make use of
11851: @code{catch}, it is not necessarily clear which @code{throw} should be
11852: used for the return stack dump (e.g., consider one @code{throw} that
11853: indicates an error, which is caught, and during recovery another error
1.42 anton 11854: happens; which @code{throw} should be used for the stack dump?). Gforth
1.25 anton 11855: presents the return stack dump for the first @code{throw} after the last
11856: executed (not returned-to) @code{catch}; this works well in the usual
11857: case.
11858:
11859: @cindex @code{gforth-fast} and backtraces
11860: @cindex @code{gforth-fast}, difference from @code{gforth}
11861: @cindex backtraces with @code{gforth-fast}
11862: @cindex return stack dump with @code{gforth-fast}
1.79 anton 11863: @code{Gforth} is able to do a return stack dump for throws generated
1.25 anton 11864: from primitives (e.g., invalid memory address, stack empty etc.);
11865: @code{gforth-fast} is only able to do a return stack dump from a
11866: directly called @code{throw} (including @code{abort} etc.). This is the
1.30 anton 11867: only difference (apart from a speed factor of between 1.15 (K6-2) and
1.78 anton 11868: 2 (21264)) between @code{gforth} and @code{gforth-fast}. Given an
1.30 anton 11869: exception caused by a primitive in @code{gforth-fast}, you will
11870: typically see no return stack dump at all; however, if the exception is
11871: caught by @code{catch} (e.g., for restoring some state), and then
11872: @code{throw}n again, the return stack dump will be for the first such
11873: @code{throw}.
1.2 jwilke 11874:
1.5 anton 11875: @c ******************************************************************
1.24 anton 11876: @node Tools, ANS conformance, Error messages, Top
1.1 anton 11877: @chapter Tools
11878:
11879: @menu
11880: * ANS Report:: Report the words used, sorted by wordset.
11881: @end menu
11882:
11883: See also @ref{Emacs and Gforth}.
11884:
11885: @node ANS Report, , Tools, Tools
11886: @section @file{ans-report.fs}: Report the words used, sorted by wordset
11887: @cindex @file{ans-report.fs}
11888: @cindex report the words used in your program
11889: @cindex words used in your program
11890:
11891: If you want to label a Forth program as ANS Forth Program, you must
11892: document which wordsets the program uses; for extension wordsets, it is
11893: helpful to list the words the program requires from these wordsets
11894: (because Forth systems are allowed to provide only some words of them).
11895:
11896: The @file{ans-report.fs} tool makes it easy for you to determine which
11897: words from which wordset and which non-ANS words your application
11898: uses. You simply have to include @file{ans-report.fs} before loading the
11899: program you want to check. After loading your program, you can get the
11900: report with @code{print-ans-report}. A typical use is to run this as
11901: batch job like this:
11902: @example
11903: gforth ans-report.fs myprog.fs -e "print-ans-report bye"
11904: @end example
11905:
11906: The output looks like this (for @file{compat/control.fs}):
11907: @example
11908: The program uses the following words
11909: from CORE :
11910: : POSTPONE THEN ; immediate ?dup IF 0=
11911: from BLOCK-EXT :
11912: \
11913: from FILE :
11914: (
11915: @end example
11916:
11917: @subsection Caveats
11918:
11919: Note that @file{ans-report.fs} just checks which words are used, not whether
11920: they are used in an ANS Forth conforming way!
11921:
11922: Some words are defined in several wordsets in the
11923: standard. @file{ans-report.fs} reports them for only one of the
11924: wordsets, and not necessarily the one you expect. It depends on usage
11925: which wordset is the right one to specify. E.g., if you only use the
11926: compilation semantics of @code{S"}, it is a Core word; if you also use
11927: its interpretation semantics, it is a File word.
11928:
11929: @c ******************************************************************
1.65 anton 11930: @node ANS conformance, Standard vs Extensions, Tools, Top
1.1 anton 11931: @chapter ANS conformance
11932: @cindex ANS conformance of Gforth
11933:
11934: To the best of our knowledge, Gforth is an
11935:
11936: ANS Forth System
11937: @itemize @bullet
11938: @item providing the Core Extensions word set
11939: @item providing the Block word set
11940: @item providing the Block Extensions word set
11941: @item providing the Double-Number word set
11942: @item providing the Double-Number Extensions word set
11943: @item providing the Exception word set
11944: @item providing the Exception Extensions word set
11945: @item providing the Facility word set
1.40 anton 11946: @item providing @code{EKEY}, @code{EKEY>CHAR}, @code{EKEY?}, @code{MS} and @code{TIME&DATE} from the Facility Extensions word set
1.1 anton 11947: @item providing the File Access word set
11948: @item providing the File Access Extensions word set
11949: @item providing the Floating-Point word set
11950: @item providing the Floating-Point Extensions word set
11951: @item providing the Locals word set
11952: @item providing the Locals Extensions word set
11953: @item providing the Memory-Allocation word set
11954: @item providing the Memory-Allocation Extensions word set (that one's easy)
11955: @item providing the Programming-Tools word set
11956: @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
11957: @item providing the Search-Order word set
11958: @item providing the Search-Order Extensions word set
11959: @item providing the String word set
11960: @item providing the String Extensions word set (another easy one)
11961: @end itemize
11962:
11963: @cindex system documentation
11964: In addition, ANS Forth systems are required to document certain
11965: implementation choices. This chapter tries to meet these
11966: requirements. In many cases it gives a way to ask the system for the
11967: information instead of providing the information directly, in
11968: particular, if the information depends on the processor, the operating
11969: system or the installation options chosen, or if they are likely to
11970: change during the maintenance of Gforth.
11971:
11972: @comment The framework for the rest has been taken from pfe.
11973:
11974: @menu
11975: * The Core Words::
11976: * The optional Block word set::
11977: * The optional Double Number word set::
11978: * The optional Exception word set::
11979: * The optional Facility word set::
11980: * The optional File-Access word set::
11981: * The optional Floating-Point word set::
11982: * The optional Locals word set::
11983: * The optional Memory-Allocation word set::
11984: * The optional Programming-Tools word set::
11985: * The optional Search-Order word set::
11986: @end menu
11987:
11988:
11989: @c =====================================================================
11990: @node The Core Words, The optional Block word set, ANS conformance, ANS conformance
11991: @comment node-name, next, previous, up
11992: @section The Core Words
11993: @c =====================================================================
11994: @cindex core words, system documentation
11995: @cindex system documentation, core words
11996:
11997: @menu
11998: * core-idef:: Implementation Defined Options
11999: * core-ambcond:: Ambiguous Conditions
12000: * core-other:: Other System Documentation
12001: @end menu
12002:
12003: @c ---------------------------------------------------------------------
12004: @node core-idef, core-ambcond, The Core Words, The Core Words
12005: @subsection Implementation Defined Options
12006: @c ---------------------------------------------------------------------
12007: @cindex core words, implementation-defined options
12008: @cindex implementation-defined options, core words
12009:
12010:
12011: @table @i
12012: @item (Cell) aligned addresses:
12013: @cindex cell-aligned addresses
12014: @cindex aligned addresses
12015: processor-dependent. Gforth's alignment words perform natural alignment
12016: (e.g., an address aligned for a datum of size 8 is divisible by
12017: 8). Unaligned accesses usually result in a @code{-23 THROW}.
12018:
12019: @item @code{EMIT} and non-graphic characters:
12020: @cindex @code{EMIT} and non-graphic characters
12021: @cindex non-graphic characters and @code{EMIT}
12022: The character is output using the C library function (actually, macro)
12023: @code{putc}.
12024:
12025: @item character editing of @code{ACCEPT} and @code{EXPECT}:
12026: @cindex character editing of @code{ACCEPT} and @code{EXPECT}
12027: @cindex editing in @code{ACCEPT} and @code{EXPECT}
12028: @cindex @code{ACCEPT}, editing
12029: @cindex @code{EXPECT}, editing
12030: This is modeled on the GNU readline library (@pxref{Readline
12031: Interaction, , Command Line Editing, readline, The GNU Readline
12032: Library}) with Emacs-like key bindings. @kbd{Tab} deviates a little by
12033: producing a full word completion every time you type it (instead of
1.28 crook 12034: producing the common prefix of all completions). @xref{Command-line editing}.
1.1 anton 12035:
12036: @item character set:
12037: @cindex character set
12038: The character set of your computer and display device. Gforth is
12039: 8-bit-clean (but some other component in your system may make trouble).
12040:
12041: @item Character-aligned address requirements:
12042: @cindex character-aligned address requirements
12043: installation-dependent. Currently a character is represented by a C
12044: @code{unsigned char}; in the future we might switch to @code{wchar_t}
12045: (Comments on that requested).
12046:
12047: @item character-set extensions and matching of names:
12048: @cindex character-set extensions and matching of names
1.26 crook 12049: @cindex case-sensitivity for name lookup
12050: @cindex name lookup, case-sensitivity
12051: @cindex locale and case-sensitivity
1.21 crook 12052: Any character except the ASCII NUL character can be used in a
1.1 anton 12053: name. Matching is case-insensitive (except in @code{TABLE}s). The
1.47 crook 12054: matching is performed using the C library function @code{strncasecmp}, whose
1.1 anton 12055: function is probably influenced by the locale. E.g., the @code{C} locale
12056: does not know about accents and umlauts, so they are matched
12057: case-sensitively in that locale. For portability reasons it is best to
12058: write programs such that they work in the @code{C} locale. Then one can
12059: use libraries written by a Polish programmer (who might use words
12060: containing ISO Latin-2 encoded characters) and by a French programmer
12061: (ISO Latin-1) in the same program (of course, @code{WORDS} will produce
12062: funny results for some of the words (which ones, depends on the font you
12063: are using)). Also, the locale you prefer may not be available in other
12064: operating systems. Hopefully, Unicode will solve these problems one day.
12065:
12066: @item conditions under which control characters match a space delimiter:
12067: @cindex space delimiters
12068: @cindex control characters as delimiters
12069: If @code{WORD} is called with the space character as a delimiter, all
12070: white-space characters (as identified by the C macro @code{isspace()})
12071: are delimiters. @code{PARSE}, on the other hand, treats space like other
1.44 crook 12072: delimiters. @code{SWORD} treats space like @code{WORD}, but behaves
1.79 anton 12073: like @code{PARSE} otherwise. @code{Name}, which is used by the outer
1.1 anton 12074: interpreter (aka text interpreter) by default, treats all white-space
12075: characters as delimiters.
12076:
1.26 crook 12077: @item format of the control-flow stack:
12078: @cindex control-flow stack, format
12079: The data stack is used as control-flow stack. The size of a control-flow
1.1 anton 12080: stack item in cells is given by the constant @code{cs-item-size}. At the
12081: time of this writing, an item consists of a (pointer to a) locals list
12082: (third), an address in the code (second), and a tag for identifying the
12083: item (TOS). The following tags are used: @code{defstart},
12084: @code{live-orig}, @code{dead-orig}, @code{dest}, @code{do-dest},
12085: @code{scopestart}.
12086:
12087: @item conversion of digits > 35
12088: @cindex digits > 35
12089: The characters @code{[\]^_'} are the digits with the decimal value
12090: 36@minus{}41. There is no way to input many of the larger digits.
12091:
12092: @item display after input terminates in @code{ACCEPT} and @code{EXPECT}:
12093: @cindex @code{EXPECT}, display after end of input
12094: @cindex @code{ACCEPT}, display after end of input
12095: The cursor is moved to the end of the entered string. If the input is
12096: terminated using the @kbd{Return} key, a space is typed.
12097:
12098: @item exception abort sequence of @code{ABORT"}:
12099: @cindex exception abort sequence of @code{ABORT"}
12100: @cindex @code{ABORT"}, exception abort sequence
12101: The error string is stored into the variable @code{"error} and a
12102: @code{-2 throw} is performed.
12103:
12104: @item input line terminator:
12105: @cindex input line terminator
12106: @cindex line terminator on input
1.26 crook 12107: @cindex newline character on input
1.1 anton 12108: For interactive input, @kbd{C-m} (CR) and @kbd{C-j} (LF) terminate
12109: lines. One of these characters is typically produced when you type the
12110: @kbd{Enter} or @kbd{Return} key.
12111:
12112: @item maximum size of a counted string:
12113: @cindex maximum size of a counted string
12114: @cindex counted string, maximum size
12115: @code{s" /counted-string" environment? drop .}. Currently 255 characters
1.79 anton 12116: on all platforms, but this may change.
1.1 anton 12117:
12118: @item maximum size of a parsed string:
12119: @cindex maximum size of a parsed string
12120: @cindex parsed string, maximum size
12121: Given by the constant @code{/line}. Currently 255 characters.
12122:
12123: @item maximum size of a definition name, in characters:
12124: @cindex maximum size of a definition name, in characters
12125: @cindex name, maximum length
12126: 31
12127:
12128: @item maximum string length for @code{ENVIRONMENT?}, in characters:
12129: @cindex maximum string length for @code{ENVIRONMENT?}, in characters
12130: @cindex @code{ENVIRONMENT?} string length, maximum
12131: 31
12132:
12133: @item method of selecting the user input device:
12134: @cindex user input device, method of selecting
12135: The user input device is the standard input. There is currently no way to
12136: change it from within Gforth. However, the input can typically be
12137: redirected in the command line that starts Gforth.
12138:
12139: @item method of selecting the user output device:
12140: @cindex user output device, method of selecting
12141: @code{EMIT} and @code{TYPE} output to the file-id stored in the value
1.10 anton 12142: @code{outfile-id} (@code{stdout} by default). Gforth uses unbuffered
12143: output when the user output device is a terminal, otherwise the output
12144: is buffered.
1.1 anton 12145:
12146: @item methods of dictionary compilation:
12147: What are we expected to document here?
12148:
12149: @item number of bits in one address unit:
12150: @cindex number of bits in one address unit
12151: @cindex address unit, size in bits
12152: @code{s" address-units-bits" environment? drop .}. 8 in all current
1.79 anton 12153: platforms.
1.1 anton 12154:
12155: @item number representation and arithmetic:
12156: @cindex number representation and arithmetic
1.79 anton 12157: Processor-dependent. Binary two's complement on all current platforms.
1.1 anton 12158:
12159: @item ranges for integer types:
12160: @cindex ranges for integer types
12161: @cindex integer types, ranges
12162: Installation-dependent. Make environmental queries for @code{MAX-N},
12163: @code{MAX-U}, @code{MAX-D} and @code{MAX-UD}. The lower bounds for
12164: unsigned (and positive) types is 0. The lower bound for signed types on
12165: two's complement and one's complement machines machines can be computed
12166: by adding 1 to the upper bound.
12167:
12168: @item read-only data space regions:
12169: @cindex read-only data space regions
12170: @cindex data-space, read-only regions
12171: The whole Forth data space is writable.
12172:
12173: @item size of buffer at @code{WORD}:
12174: @cindex size of buffer at @code{WORD}
12175: @cindex @code{WORD} buffer size
12176: @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
12177: shared with the pictured numeric output string. If overwriting
12178: @code{PAD} is acceptable, it is as large as the remaining dictionary
12179: space, although only as much can be sensibly used as fits in a counted
12180: string.
12181:
12182: @item size of one cell in address units:
12183: @cindex cell size
12184: @code{1 cells .}.
12185:
12186: @item size of one character in address units:
12187: @cindex char size
1.79 anton 12188: @code{1 chars .}. 1 on all current platforms.
1.1 anton 12189:
12190: @item size of the keyboard terminal buffer:
12191: @cindex size of the keyboard terminal buffer
12192: @cindex terminal buffer, size
12193: Varies. You can determine the size at a specific time using @code{lp@@
12194: tib - .}. It is shared with the locals stack and TIBs of files that
12195: include the current file. You can change the amount of space for TIBs
12196: and locals stack at Gforth startup with the command line option
12197: @code{-l}.
12198:
12199: @item size of the pictured numeric output buffer:
12200: @cindex size of the pictured numeric output buffer
12201: @cindex pictured numeric output buffer, size
12202: @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
12203: shared with @code{WORD}.
12204:
12205: @item size of the scratch area returned by @code{PAD}:
12206: @cindex size of the scratch area returned by @code{PAD}
12207: @cindex @code{PAD} size
12208: The remainder of dictionary space. @code{unused pad here - - .}.
12209:
12210: @item system case-sensitivity characteristics:
12211: @cindex case-sensitivity characteristics
1.26 crook 12212: Dictionary searches are case-insensitive (except in
1.1 anton 12213: @code{TABLE}s). However, as explained above under @i{character-set
12214: extensions}, the matching for non-ASCII characters is determined by the
12215: locale you are using. In the default @code{C} locale all non-ASCII
12216: characters are matched case-sensitively.
12217:
12218: @item system prompt:
12219: @cindex system prompt
12220: @cindex prompt
12221: @code{ ok} in interpret state, @code{ compiled} in compile state.
12222:
12223: @item division rounding:
12224: @cindex division rounding
12225: installation dependent. @code{s" floored" environment? drop .}. We leave
12226: the choice to @code{gcc} (what to use for @code{/}) and to you (whether
12227: to use @code{fm/mod}, @code{sm/rem} or simply @code{/}).
12228:
12229: @item values of @code{STATE} when true:
12230: @cindex @code{STATE} values
12231: -1.
12232:
12233: @item values returned after arithmetic overflow:
12234: On two's complement machines, arithmetic is performed modulo
12235: 2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
12236: arithmetic (with appropriate mapping for signed types). Division by zero
12237: typically results in a @code{-55 throw} (Floating-point unidentified
1.80 anton 12238: fault) or @code{-10 throw} (divide by zero).
1.1 anton 12239:
12240: @item whether the current definition can be found after @t{DOES>}:
12241: @cindex @t{DOES>}, visibility of current definition
12242: No.
12243:
12244: @end table
12245:
12246: @c ---------------------------------------------------------------------
12247: @node core-ambcond, core-other, core-idef, The Core Words
12248: @subsection Ambiguous conditions
12249: @c ---------------------------------------------------------------------
12250: @cindex core words, ambiguous conditions
12251: @cindex ambiguous conditions, core words
12252:
12253: @table @i
12254:
12255: @item a name is neither a word nor a number:
12256: @cindex name not found
1.26 crook 12257: @cindex undefined word
1.80 anton 12258: @code{-13 throw} (Undefined word).
1.1 anton 12259:
12260: @item a definition name exceeds the maximum length allowed:
1.26 crook 12261: @cindex word name too long
1.1 anton 12262: @code{-19 throw} (Word name too long)
12263:
12264: @item addressing a region not inside the various data spaces of the forth system:
12265: @cindex Invalid memory address
1.32 anton 12266: The stacks, code space and header space are accessible. Machine code space is
1.1 anton 12267: typically readable. Accessing other addresses gives results dependent on
12268: the operating system. On decent systems: @code{-9 throw} (Invalid memory
12269: address).
12270:
12271: @item argument type incompatible with parameter:
1.26 crook 12272: @cindex argument type mismatch
1.1 anton 12273: This is usually not caught. Some words perform checks, e.g., the control
12274: flow words, and issue a @code{ABORT"} or @code{-12 THROW} (Argument type
12275: mismatch).
12276:
12277: @item attempting to obtain the execution token of a word with undefined execution semantics:
12278: @cindex Interpreting a compile-only word, for @code{'} etc.
12279: @cindex execution token of words with undefined execution semantics
12280: @code{-14 throw} (Interpreting a compile-only word). In some cases, you
12281: get an execution token for @code{compile-only-error} (which performs a
12282: @code{-14 throw} when executed).
12283:
12284: @item dividing by zero:
12285: @cindex dividing by zero
12286: @cindex floating point unidentified fault, integer division
1.80 anton 12287: On some platforms, this produces a @code{-10 throw} (Division by
1.24 anton 12288: zero); on other systems, this typically results in a @code{-55 throw}
12289: (Floating-point unidentified fault).
1.1 anton 12290:
12291: @item insufficient data stack or return stack space:
12292: @cindex insufficient data stack or return stack space
12293: @cindex stack overflow
1.26 crook 12294: @cindex address alignment exception, stack overflow
1.1 anton 12295: @cindex Invalid memory address, stack overflow
12296: Depending on the operating system, the installation, and the invocation
12297: of Gforth, this is either checked by the memory management hardware, or
1.24 anton 12298: it is not checked. If it is checked, you typically get a @code{-3 throw}
12299: (Stack overflow), @code{-5 throw} (Return stack overflow), or @code{-9
12300: throw} (Invalid memory address) (depending on the platform and how you
12301: achieved the overflow) as soon as the overflow happens. If it is not
12302: checked, overflows typically result in mysterious illegal memory
12303: accesses, producing @code{-9 throw} (Invalid memory address) or
12304: @code{-23 throw} (Address alignment exception); they might also destroy
12305: the internal data structure of @code{ALLOCATE} and friends, resulting in
12306: various errors in these words.
1.1 anton 12307:
12308: @item insufficient space for loop control parameters:
12309: @cindex insufficient space for loop control parameters
1.80 anton 12310: Like other return stack overflows.
1.1 anton 12311:
12312: @item insufficient space in the dictionary:
12313: @cindex insufficient space in the dictionary
12314: @cindex dictionary overflow
1.12 anton 12315: If you try to allot (either directly with @code{allot}, or indirectly
12316: with @code{,}, @code{create} etc.) more memory than available in the
12317: dictionary, you get a @code{-8 throw} (Dictionary overflow). If you try
12318: to access memory beyond the end of the dictionary, the results are
12319: similar to stack overflows.
1.1 anton 12320:
12321: @item interpreting a word with undefined interpretation semantics:
12322: @cindex interpreting a word with undefined interpretation semantics
12323: @cindex Interpreting a compile-only word
12324: For some words, we have defined interpretation semantics. For the
12325: others: @code{-14 throw} (Interpreting a compile-only word).
12326:
12327: @item modifying the contents of the input buffer or a string literal:
12328: @cindex modifying the contents of the input buffer or a string literal
12329: These are located in writable memory and can be modified.
12330:
12331: @item overflow of the pictured numeric output string:
12332: @cindex overflow of the pictured numeric output string
12333: @cindex pictured numeric output string, overflow
1.24 anton 12334: @code{-17 throw} (Pictured numeric ouput string overflow).
1.1 anton 12335:
12336: @item parsed string overflow:
12337: @cindex parsed string overflow
12338: @code{PARSE} cannot overflow. @code{WORD} does not check for overflow.
12339:
12340: @item producing a result out of range:
12341: @cindex result out of range
12342: On two's complement machines, arithmetic is performed modulo
12343: 2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
12344: arithmetic (with appropriate mapping for signed types). Division by zero
1.24 anton 12345: typically results in a @code{-10 throw} (divide by zero) or @code{-55
12346: throw} (floating point unidentified fault). @code{convert} and
12347: @code{>number} currently overflow silently.
1.1 anton 12348:
12349: @item reading from an empty data or return stack:
12350: @cindex stack empty
12351: @cindex stack underflow
1.24 anton 12352: @cindex return stack underflow
1.1 anton 12353: The data stack is checked by the outer (aka text) interpreter after
12354: every word executed. If it has underflowed, a @code{-4 throw} (Stack
12355: underflow) is performed. Apart from that, stacks may be checked or not,
1.24 anton 12356: depending on operating system, installation, and invocation. If they are
12357: caught by a check, they typically result in @code{-4 throw} (Stack
12358: underflow), @code{-6 throw} (Return stack underflow) or @code{-9 throw}
12359: (Invalid memory address), depending on the platform and which stack
12360: underflows and by how much. Note that even if the system uses checking
12361: (through the MMU), your program may have to underflow by a significant
12362: number of stack items to trigger the reaction (the reason for this is
12363: that the MMU, and therefore the checking, works with a page-size
12364: granularity). If there is no checking, the symptoms resulting from an
12365: underflow are similar to those from an overflow. Unbalanced return
1.80 anton 12366: stack errors can result in a variety of symptoms, including @code{-9 throw}
1.24 anton 12367: (Invalid memory address) and Illegal Instruction (typically @code{-260
12368: throw}).
1.1 anton 12369:
12370: @item unexpected end of the input buffer, resulting in an attempt to use a zero-length string as a name:
12371: @cindex unexpected end of the input buffer
12372: @cindex zero-length string as a name
12373: @cindex Attempt to use zero-length string as a name
12374: @code{Create} and its descendants perform a @code{-16 throw} (Attempt to
12375: use zero-length string as a name). Words like @code{'} probably will not
12376: find what they search. Note that it is possible to create zero-length
12377: names with @code{nextname} (should it not?).
12378:
12379: @item @code{>IN} greater than input buffer:
12380: @cindex @code{>IN} greater than input buffer
12381: The next invocation of a parsing word returns a string with length 0.
12382:
12383: @item @code{RECURSE} appears after @code{DOES>}:
12384: @cindex @code{RECURSE} appears after @code{DOES>}
12385: Compiles a recursive call to the defining word, not to the defined word.
12386:
12387: @item argument input source different than current input source for @code{RESTORE-INPUT}:
12388: @cindex argument input source different than current input source for @code{RESTORE-INPUT}
1.26 crook 12389: @cindex argument type mismatch, @code{RESTORE-INPUT}
1.1 anton 12390: @cindex @code{RESTORE-INPUT}, Argument type mismatch
12391: @code{-12 THROW}. Note that, once an input file is closed (e.g., because
12392: the end of the file was reached), its source-id may be
12393: reused. Therefore, restoring an input source specification referencing a
12394: closed file may lead to unpredictable results instead of a @code{-12
12395: THROW}.
12396:
12397: In the future, Gforth may be able to restore input source specifications
12398: from other than the current input source.
12399:
12400: @item data space containing definitions gets de-allocated:
12401: @cindex data space containing definitions gets de-allocated
12402: Deallocation with @code{allot} is not checked. This typically results in
12403: memory access faults or execution of illegal instructions.
12404:
12405: @item data space read/write with incorrect alignment:
12406: @cindex data space read/write with incorrect alignment
12407: @cindex alignment faults
1.26 crook 12408: @cindex address alignment exception
1.1 anton 12409: Processor-dependent. Typically results in a @code{-23 throw} (Address
1.12 anton 12410: alignment exception). Under Linux-Intel on a 486 or later processor with
1.1 anton 12411: alignment turned on, incorrect alignment results in a @code{-9 throw}
12412: (Invalid memory address). There are reportedly some processors with
1.12 anton 12413: alignment restrictions that do not report violations.
1.1 anton 12414:
12415: @item data space pointer not properly aligned, @code{,}, @code{C,}:
12416: @cindex data space pointer not properly aligned, @code{,}, @code{C,}
12417: Like other alignment errors.
12418:
12419: @item less than u+2 stack items (@code{PICK} and @code{ROLL}):
12420: Like other stack underflows.
12421:
12422: @item loop control parameters not available:
12423: @cindex loop control parameters not available
12424: Not checked. The counted loop words simply assume that the top of return
12425: stack items are loop control parameters and behave accordingly.
12426:
12427: @item most recent definition does not have a name (@code{IMMEDIATE}):
12428: @cindex most recent definition does not have a name (@code{IMMEDIATE})
12429: @cindex last word was headerless
12430: @code{abort" last word was headerless"}.
12431:
12432: @item name not defined by @code{VALUE} used by @code{TO}:
12433: @cindex name not defined by @code{VALUE} used by @code{TO}
12434: @cindex @code{TO} on non-@code{VALUE}s
12435: @cindex Invalid name argument, @code{TO}
12436: @code{-32 throw} (Invalid name argument) (unless name is a local or was
12437: defined by @code{CONSTANT}; in the latter case it just changes the constant).
12438:
12439: @item name not found (@code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]}):
12440: @cindex name not found (@code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]})
1.26 crook 12441: @cindex undefined word, @code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]}
1.1 anton 12442: @code{-13 throw} (Undefined word)
12443:
12444: @item parameters are not of the same type (@code{DO}, @code{?DO}, @code{WITHIN}):
12445: @cindex parameters are not of the same type (@code{DO}, @code{?DO}, @code{WITHIN})
12446: Gforth behaves as if they were of the same type. I.e., you can predict
12447: the behaviour by interpreting all parameters as, e.g., signed.
12448:
12449: @item @code{POSTPONE} or @code{[COMPILE]} applied to @code{TO}:
12450: @cindex @code{POSTPONE} or @code{[COMPILE]} applied to @code{TO}
12451: Assume @code{: X POSTPONE TO ; IMMEDIATE}. @code{X} performs the
12452: compilation semantics of @code{TO}.
12453:
12454: @item String longer than a counted string returned by @code{WORD}:
1.26 crook 12455: @cindex string longer than a counted string returned by @code{WORD}
1.1 anton 12456: @cindex @code{WORD}, string overflow
12457: Not checked. The string will be ok, but the count will, of course,
12458: contain only the least significant bits of the length.
12459:
12460: @item u greater than or equal to the number of bits in a cell (@code{LSHIFT}, @code{RSHIFT}):
12461: @cindex @code{LSHIFT}, large shift counts
12462: @cindex @code{RSHIFT}, large shift counts
12463: Processor-dependent. Typical behaviours are returning 0 and using only
12464: the low bits of the shift count.
12465:
12466: @item word not defined via @code{CREATE}:
12467: @cindex @code{>BODY} of non-@code{CREATE}d words
12468: @code{>BODY} produces the PFA of the word no matter how it was defined.
12469:
12470: @cindex @code{DOES>} of non-@code{CREATE}d words
12471: @code{DOES>} changes the execution semantics of the last defined word no
12472: matter how it was defined. E.g., @code{CONSTANT DOES>} is equivalent to
12473: @code{CREATE , DOES>}.
12474:
12475: @item words improperly used outside @code{<#} and @code{#>}:
12476: Not checked. As usual, you can expect memory faults.
12477:
12478: @end table
12479:
12480:
12481: @c ---------------------------------------------------------------------
12482: @node core-other, , core-ambcond, The Core Words
12483: @subsection Other system documentation
12484: @c ---------------------------------------------------------------------
12485: @cindex other system documentation, core words
12486: @cindex core words, other system documentation
12487:
12488: @table @i
12489: @item nonstandard words using @code{PAD}:
12490: @cindex @code{PAD} use by nonstandard words
12491: None.
12492:
12493: @item operator's terminal facilities available:
12494: @cindex operator's terminal facilities available
1.80 anton 12495: After processing the OS's command line, Gforth goes into interactive mode,
1.1 anton 12496: and you can give commands to Gforth interactively. The actual facilities
12497: available depend on how you invoke Gforth.
12498:
12499: @item program data space available:
12500: @cindex program data space available
12501: @cindex data space available
12502: @code{UNUSED .} gives the remaining dictionary space. The total
12503: dictionary space can be specified with the @code{-m} switch
12504: (@pxref{Invoking Gforth}) when Gforth starts up.
12505:
12506: @item return stack space available:
12507: @cindex return stack space available
12508: You can compute the total return stack space in cells with
12509: @code{s" RETURN-STACK-CELLS" environment? drop .}. You can specify it at
12510: startup time with the @code{-r} switch (@pxref{Invoking Gforth}).
12511:
12512: @item stack space available:
12513: @cindex stack space available
12514: You can compute the total data stack space in cells with
12515: @code{s" STACK-CELLS" environment? drop .}. You can specify it at
12516: startup time with the @code{-d} switch (@pxref{Invoking Gforth}).
12517:
12518: @item system dictionary space required, in address units:
12519: @cindex system dictionary space required, in address units
12520: Type @code{here forthstart - .} after startup. At the time of this
12521: writing, this gives 80080 (bytes) on a 32-bit system.
12522: @end table
12523:
12524:
12525: @c =====================================================================
12526: @node The optional Block word set, The optional Double Number word set, The Core Words, ANS conformance
12527: @section The optional Block word set
12528: @c =====================================================================
12529: @cindex system documentation, block words
12530: @cindex block words, system documentation
12531:
12532: @menu
12533: * block-idef:: Implementation Defined Options
12534: * block-ambcond:: Ambiguous Conditions
12535: * block-other:: Other System Documentation
12536: @end menu
12537:
12538:
12539: @c ---------------------------------------------------------------------
12540: @node block-idef, block-ambcond, The optional Block word set, The optional Block word set
12541: @subsection Implementation Defined Options
12542: @c ---------------------------------------------------------------------
12543: @cindex implementation-defined options, block words
12544: @cindex block words, implementation-defined options
12545:
12546: @table @i
12547: @item the format for display by @code{LIST}:
12548: @cindex @code{LIST} display format
12549: First the screen number is displayed, then 16 lines of 64 characters,
12550: each line preceded by the line number.
12551:
12552: @item the length of a line affected by @code{\}:
12553: @cindex length of a line affected by @code{\}
12554: @cindex @code{\}, line length in blocks
12555: 64 characters.
12556: @end table
12557:
12558:
12559: @c ---------------------------------------------------------------------
12560: @node block-ambcond, block-other, block-idef, The optional Block word set
12561: @subsection Ambiguous conditions
12562: @c ---------------------------------------------------------------------
12563: @cindex block words, ambiguous conditions
12564: @cindex ambiguous conditions, block words
12565:
12566: @table @i
12567: @item correct block read was not possible:
12568: @cindex block read not possible
12569: Typically results in a @code{throw} of some OS-derived value (between
12570: -512 and -2048). If the blocks file was just not long enough, blanks are
12571: supplied for the missing portion.
12572:
12573: @item I/O exception in block transfer:
12574: @cindex I/O exception in block transfer
12575: @cindex block transfer, I/O exception
12576: Typically results in a @code{throw} of some OS-derived value (between
12577: -512 and -2048).
12578:
12579: @item invalid block number:
12580: @cindex invalid block number
12581: @cindex block number invalid
12582: @code{-35 throw} (Invalid block number)
12583:
12584: @item a program directly alters the contents of @code{BLK}:
12585: @cindex @code{BLK}, altering @code{BLK}
12586: The input stream is switched to that other block, at the same
12587: position. If the storing to @code{BLK} happens when interpreting
12588: non-block input, the system will get quite confused when the block ends.
12589:
12590: @item no current block buffer for @code{UPDATE}:
12591: @cindex @code{UPDATE}, no current block buffer
12592: @code{UPDATE} has no effect.
12593:
12594: @end table
12595:
12596: @c ---------------------------------------------------------------------
12597: @node block-other, , block-ambcond, The optional Block word set
12598: @subsection Other system documentation
12599: @c ---------------------------------------------------------------------
12600: @cindex other system documentation, block words
12601: @cindex block words, other system documentation
12602:
12603: @table @i
12604: @item any restrictions a multiprogramming system places on the use of buffer addresses:
12605: No restrictions (yet).
12606:
12607: @item the number of blocks available for source and data:
12608: depends on your disk space.
12609:
12610: @end table
12611:
12612:
12613: @c =====================================================================
12614: @node The optional Double Number word set, The optional Exception word set, The optional Block word set, ANS conformance
12615: @section The optional Double Number word set
12616: @c =====================================================================
12617: @cindex system documentation, double words
12618: @cindex double words, system documentation
12619:
12620: @menu
12621: * double-ambcond:: Ambiguous Conditions
12622: @end menu
12623:
12624:
12625: @c ---------------------------------------------------------------------
12626: @node double-ambcond, , The optional Double Number word set, The optional Double Number word set
12627: @subsection Ambiguous conditions
12628: @c ---------------------------------------------------------------------
12629: @cindex double words, ambiguous conditions
12630: @cindex ambiguous conditions, double words
12631:
12632: @table @i
1.29 crook 12633: @item @i{d} outside of range of @i{n} in @code{D>S}:
12634: @cindex @code{D>S}, @i{d} out of range of @i{n}
12635: The least significant cell of @i{d} is produced.
1.1 anton 12636:
12637: @end table
12638:
12639:
12640: @c =====================================================================
12641: @node The optional Exception word set, The optional Facility word set, The optional Double Number word set, ANS conformance
12642: @section The optional Exception word set
12643: @c =====================================================================
12644: @cindex system documentation, exception words
12645: @cindex exception words, system documentation
12646:
12647: @menu
12648: * exception-idef:: Implementation Defined Options
12649: @end menu
12650:
12651:
12652: @c ---------------------------------------------------------------------
12653: @node exception-idef, , The optional Exception word set, The optional Exception word set
12654: @subsection Implementation Defined Options
12655: @c ---------------------------------------------------------------------
12656: @cindex implementation-defined options, exception words
12657: @cindex exception words, implementation-defined options
12658:
12659: @table @i
12660: @item @code{THROW}-codes used in the system:
12661: @cindex @code{THROW}-codes used in the system
12662: The codes -256@minus{}-511 are used for reporting signals. The mapping
1.29 crook 12663: from OS signal numbers to throw codes is -256@minus{}@i{signal}. The
1.1 anton 12664: codes -512@minus{}-2047 are used for OS errors (for file and memory
12665: allocation operations). The mapping from OS error numbers to throw codes
12666: is -512@minus{}@code{errno}. One side effect of this mapping is that
12667: undefined OS errors produce a message with a strange number; e.g.,
12668: @code{-1000 THROW} results in @code{Unknown error 488} on my system.
12669: @end table
12670:
12671: @c =====================================================================
12672: @node The optional Facility word set, The optional File-Access word set, The optional Exception word set, ANS conformance
12673: @section The optional Facility word set
12674: @c =====================================================================
12675: @cindex system documentation, facility words
12676: @cindex facility words, system documentation
12677:
12678: @menu
12679: * facility-idef:: Implementation Defined Options
12680: * facility-ambcond:: Ambiguous Conditions
12681: @end menu
12682:
12683:
12684: @c ---------------------------------------------------------------------
12685: @node facility-idef, facility-ambcond, The optional Facility word set, The optional Facility word set
12686: @subsection Implementation Defined Options
12687: @c ---------------------------------------------------------------------
12688: @cindex implementation-defined options, facility words
12689: @cindex facility words, implementation-defined options
12690:
12691: @table @i
12692: @item encoding of keyboard events (@code{EKEY}):
12693: @cindex keyboard events, encoding in @code{EKEY}
12694: @cindex @code{EKEY}, encoding of keyboard events
1.40 anton 12695: Keys corresponding to ASCII characters are encoded as ASCII characters.
1.41 anton 12696: Other keys are encoded with the constants @code{k-left}, @code{k-right},
12697: @code{k-up}, @code{k-down}, @code{k-home}, @code{k-end}, @code{k1},
12698: @code{k2}, @code{k3}, @code{k4}, @code{k5}, @code{k6}, @code{k7},
12699: @code{k8}, @code{k9}, @code{k10}, @code{k11}, @code{k12}.
1.40 anton 12700:
1.1 anton 12701:
12702: @item duration of a system clock tick:
12703: @cindex duration of a system clock tick
12704: @cindex clock tick duration
12705: System dependent. With respect to @code{MS}, the time is specified in
12706: microseconds. How well the OS and the hardware implement this, is
12707: another question.
12708:
12709: @item repeatability to be expected from the execution of @code{MS}:
12710: @cindex repeatability to be expected from the execution of @code{MS}
12711: @cindex @code{MS}, repeatability to be expected
12712: System dependent. On Unix, a lot depends on load. If the system is
12713: lightly loaded, and the delay is short enough that Gforth does not get
12714: swapped out, the performance should be acceptable. Under MS-DOS and
12715: other single-tasking systems, it should be good.
12716:
12717: @end table
12718:
12719:
12720: @c ---------------------------------------------------------------------
12721: @node facility-ambcond, , facility-idef, The optional Facility word set
12722: @subsection Ambiguous conditions
12723: @c ---------------------------------------------------------------------
12724: @cindex facility words, ambiguous conditions
12725: @cindex ambiguous conditions, facility words
12726:
12727: @table @i
12728: @item @code{AT-XY} can't be performed on user output device:
12729: @cindex @code{AT-XY} can't be performed on user output device
12730: Largely terminal dependent. No range checks are done on the arguments.
12731: No errors are reported. You may see some garbage appearing, you may see
12732: simply nothing happen.
12733:
12734: @end table
12735:
12736:
12737: @c =====================================================================
12738: @node The optional File-Access word set, The optional Floating-Point word set, The optional Facility word set, ANS conformance
12739: @section The optional File-Access word set
12740: @c =====================================================================
12741: @cindex system documentation, file words
12742: @cindex file words, system documentation
12743:
12744: @menu
12745: * file-idef:: Implementation Defined Options
12746: * file-ambcond:: Ambiguous Conditions
12747: @end menu
12748:
12749: @c ---------------------------------------------------------------------
12750: @node file-idef, file-ambcond, The optional File-Access word set, The optional File-Access word set
12751: @subsection Implementation Defined Options
12752: @c ---------------------------------------------------------------------
12753: @cindex implementation-defined options, file words
12754: @cindex file words, implementation-defined options
12755:
12756: @table @i
12757: @item file access methods used:
12758: @cindex file access methods used
12759: @code{R/O}, @code{R/W} and @code{BIN} work as you would
12760: expect. @code{W/O} translates into the C file opening mode @code{w} (or
12761: @code{wb}): The file is cleared, if it exists, and created, if it does
12762: not (with both @code{open-file} and @code{create-file}). Under Unix
12763: @code{create-file} creates a file with 666 permissions modified by your
12764: umask.
12765:
12766: @item file exceptions:
12767: @cindex file exceptions
12768: The file words do not raise exceptions (except, perhaps, memory access
12769: faults when you pass illegal addresses or file-ids).
12770:
12771: @item file line terminator:
12772: @cindex file line terminator
12773: System-dependent. Gforth uses C's newline character as line
12774: terminator. What the actual character code(s) of this are is
12775: system-dependent.
12776:
12777: @item file name format:
12778: @cindex file name format
12779: System dependent. Gforth just uses the file name format of your OS.
12780:
12781: @item information returned by @code{FILE-STATUS}:
12782: @cindex @code{FILE-STATUS}, returned information
12783: @code{FILE-STATUS} returns the most powerful file access mode allowed
12784: for the file: Either @code{R/O}, @code{W/O} or @code{R/W}. If the file
12785: cannot be accessed, @code{R/O BIN} is returned. @code{BIN} is applicable
12786: along with the returned mode.
12787:
12788: @item input file state after an exception when including source:
12789: @cindex exception when including source
12790: All files that are left via the exception are closed.
12791:
1.29 crook 12792: @item @i{ior} values and meaning:
12793: @cindex @i{ior} values and meaning
1.68 anton 12794: @cindex @i{wior} values and meaning
1.29 crook 12795: The @i{ior}s returned by the file and memory allocation words are
1.1 anton 12796: intended as throw codes. They typically are in the range
12797: -512@minus{}-2047 of OS errors. The mapping from OS error numbers to
1.29 crook 12798: @i{ior}s is -512@minus{}@i{errno}.
1.1 anton 12799:
12800: @item maximum depth of file input nesting:
12801: @cindex maximum depth of file input nesting
12802: @cindex file input nesting, maximum depth
12803: limited by the amount of return stack, locals/TIB stack, and the number
12804: of open files available. This should not give you troubles.
12805:
12806: @item maximum size of input line:
12807: @cindex maximum size of input line
12808: @cindex input line size, maximum
12809: @code{/line}. Currently 255.
12810:
12811: @item methods of mapping block ranges to files:
12812: @cindex mapping block ranges to files
12813: @cindex files containing blocks
12814: @cindex blocks in files
12815: By default, blocks are accessed in the file @file{blocks.fb} in the
12816: current working directory. The file can be switched with @code{USE}.
12817:
12818: @item number of string buffers provided by @code{S"}:
12819: @cindex @code{S"}, number of string buffers
12820: 1
12821:
12822: @item size of string buffer used by @code{S"}:
12823: @cindex @code{S"}, size of string buffer
12824: @code{/line}. currently 255.
12825:
12826: @end table
12827:
12828: @c ---------------------------------------------------------------------
12829: @node file-ambcond, , file-idef, The optional File-Access word set
12830: @subsection Ambiguous conditions
12831: @c ---------------------------------------------------------------------
12832: @cindex file words, ambiguous conditions
12833: @cindex ambiguous conditions, file words
12834:
12835: @table @i
12836: @item attempting to position a file outside its boundaries:
12837: @cindex @code{REPOSITION-FILE}, outside the file's boundaries
12838: @code{REPOSITION-FILE} is performed as usual: Afterwards,
12839: @code{FILE-POSITION} returns the value given to @code{REPOSITION-FILE}.
12840:
12841: @item attempting to read from file positions not yet written:
12842: @cindex reading from file positions not yet written
12843: End-of-file, i.e., zero characters are read and no error is reported.
12844:
1.29 crook 12845: @item @i{file-id} is invalid (@code{INCLUDE-FILE}):
12846: @cindex @code{INCLUDE-FILE}, @i{file-id} is invalid
1.1 anton 12847: An appropriate exception may be thrown, but a memory fault or other
12848: problem is more probable.
12849:
1.29 crook 12850: @item I/O exception reading or closing @i{file-id} (@code{INCLUDE-FILE}, @code{INCLUDED}):
12851: @cindex @code{INCLUDE-FILE}, I/O exception reading or closing @i{file-id}
12852: @cindex @code{INCLUDED}, I/O exception reading or closing @i{file-id}
12853: The @i{ior} produced by the operation, that discovered the problem, is
1.1 anton 12854: thrown.
12855:
12856: @item named file cannot be opened (@code{INCLUDED}):
12857: @cindex @code{INCLUDED}, named file cannot be opened
1.29 crook 12858: The @i{ior} produced by @code{open-file} is thrown.
1.1 anton 12859:
12860: @item requesting an unmapped block number:
12861: @cindex unmapped block numbers
12862: There are no unmapped legal block numbers. On some operating systems,
12863: writing a block with a large number may overflow the file system and
12864: have an error message as consequence.
12865:
12866: @item using @code{source-id} when @code{blk} is non-zero:
12867: @cindex @code{SOURCE-ID}, behaviour when @code{BLK} is non-zero
12868: @code{source-id} performs its function. Typically it will give the id of
12869: the source which loaded the block. (Better ideas?)
12870:
12871: @end table
12872:
12873:
12874: @c =====================================================================
12875: @node The optional Floating-Point word set, The optional Locals word set, The optional File-Access word set, ANS conformance
12876: @section The optional Floating-Point word set
12877: @c =====================================================================
12878: @cindex system documentation, floating-point words
12879: @cindex floating-point words, system documentation
12880:
12881: @menu
12882: * floating-idef:: Implementation Defined Options
12883: * floating-ambcond:: Ambiguous Conditions
12884: @end menu
12885:
12886:
12887: @c ---------------------------------------------------------------------
12888: @node floating-idef, floating-ambcond, The optional Floating-Point word set, The optional Floating-Point word set
12889: @subsection Implementation Defined Options
12890: @c ---------------------------------------------------------------------
12891: @cindex implementation-defined options, floating-point words
12892: @cindex floating-point words, implementation-defined options
12893:
12894: @table @i
12895: @item format and range of floating point numbers:
12896: @cindex format and range of floating point numbers
12897: @cindex floating point numbers, format and range
12898: System-dependent; the @code{double} type of C.
12899:
1.29 crook 12900: @item results of @code{REPRESENT} when @i{float} is out of range:
12901: @cindex @code{REPRESENT}, results when @i{float} is out of range
1.1 anton 12902: System dependent; @code{REPRESENT} is implemented using the C library
12903: function @code{ecvt()} and inherits its behaviour in this respect.
12904:
12905: @item rounding or truncation of floating-point numbers:
12906: @cindex rounding of floating-point numbers
12907: @cindex truncation of floating-point numbers
12908: @cindex floating-point numbers, rounding or truncation
12909: System dependent; the rounding behaviour is inherited from the hosting C
12910: compiler. IEEE-FP-based (i.e., most) systems by default round to
12911: nearest, and break ties by rounding to even (i.e., such that the last
12912: bit of the mantissa is 0).
12913:
12914: @item size of floating-point stack:
12915: @cindex floating-point stack size
12916: @code{s" FLOATING-STACK" environment? drop .} gives the total size of
12917: the floating-point stack (in floats). You can specify this on startup
12918: with the command-line option @code{-f} (@pxref{Invoking Gforth}).
12919:
12920: @item width of floating-point stack:
12921: @cindex floating-point stack width
12922: @code{1 floats}.
12923:
12924: @end table
12925:
12926:
12927: @c ---------------------------------------------------------------------
12928: @node floating-ambcond, , floating-idef, The optional Floating-Point word set
12929: @subsection Ambiguous conditions
12930: @c ---------------------------------------------------------------------
12931: @cindex floating-point words, ambiguous conditions
12932: @cindex ambiguous conditions, floating-point words
12933:
12934: @table @i
12935: @item @code{df@@} or @code{df!} used with an address that is not double-float aligned:
12936: @cindex @code{df@@} or @code{df!} used with an address that is not double-float aligned
12937: System-dependent. Typically results in a @code{-23 THROW} like other
12938: alignment violations.
12939:
12940: @item @code{f@@} or @code{f!} used with an address that is not float aligned:
12941: @cindex @code{f@@} used with an address that is not float aligned
12942: @cindex @code{f!} used with an address that is not float aligned
12943: System-dependent. Typically results in a @code{-23 THROW} like other
12944: alignment violations.
12945:
12946: @item floating-point result out of range:
12947: @cindex floating-point result out of range
1.80 anton 12948: System-dependent. Can result in a @code{-43 throw} (floating point
12949: overflow), @code{-54 throw} (floating point underflow), @code{-41 throw}
12950: (floating point inexact result), @code{-55 THROW} (Floating-point
1.1 anton 12951: unidentified fault), or can produce a special value representing, e.g.,
12952: Infinity.
12953:
12954: @item @code{sf@@} or @code{sf!} used with an address that is not single-float aligned:
12955: @cindex @code{sf@@} or @code{sf!} used with an address that is not single-float aligned
12956: System-dependent. Typically results in an alignment fault like other
12957: alignment violations.
12958:
1.35 anton 12959: @item @code{base} is not decimal (@code{REPRESENT}, @code{F.}, @code{FE.}, @code{FS.}):
12960: @cindex @code{base} is not decimal (@code{REPRESENT}, @code{F.}, @code{FE.}, @code{FS.})
1.1 anton 12961: The floating-point number is converted into decimal nonetheless.
12962:
12963: @item Both arguments are equal to zero (@code{FATAN2}):
12964: @cindex @code{FATAN2}, both arguments are equal to zero
12965: System-dependent. @code{FATAN2} is implemented using the C library
12966: function @code{atan2()}.
12967:
1.29 crook 12968: @item Using @code{FTAN} on an argument @i{r1} where cos(@i{r1}) is zero:
12969: @cindex @code{FTAN} on an argument @i{r1} where cos(@i{r1}) is zero
12970: System-dependent. Anyway, typically the cos of @i{r1} will not be zero
1.1 anton 12971: because of small errors and the tan will be a very large (or very small)
12972: but finite number.
12973:
1.29 crook 12974: @item @i{d} cannot be presented precisely as a float in @code{D>F}:
12975: @cindex @code{D>F}, @i{d} cannot be presented precisely as a float
1.1 anton 12976: The result is rounded to the nearest float.
12977:
12978: @item dividing by zero:
12979: @cindex dividing by zero, floating-point
12980: @cindex floating-point dividing by zero
12981: @cindex floating-point unidentified fault, FP divide-by-zero
1.80 anton 12982: Platform-dependent; can produce an Infinity, NaN, @code{-42 throw}
12983: (floating point divide by zero) or @code{-55 throw} (Floating-point
12984: unidentified fault).
1.1 anton 12985:
12986: @item exponent too big for conversion (@code{DF!}, @code{DF@@}, @code{SF!}, @code{SF@@}):
12987: @cindex exponent too big for conversion (@code{DF!}, @code{DF@@}, @code{SF!}, @code{SF@@})
12988: System dependent. On IEEE-FP based systems the number is converted into
12989: an infinity.
12990:
1.29 crook 12991: @item @i{float}<1 (@code{FACOSH}):
12992: @cindex @code{FACOSH}, @i{float}<1
1.1 anton 12993: @cindex floating-point unidentified fault, @code{FACOSH}
1.80 anton 12994: Platform-dependent; on IEEE-FP systems typically produces a NaN.
1.1 anton 12995:
1.29 crook 12996: @item @i{float}=<-1 (@code{FLNP1}):
12997: @cindex @code{FLNP1}, @i{float}=<-1
1.1 anton 12998: @cindex floating-point unidentified fault, @code{FLNP1}
1.80 anton 12999: Platform-dependent; on IEEE-FP systems typically produces a NaN (or a
13000: negative infinity for @i{float}=-1).
1.1 anton 13001:
1.29 crook 13002: @item @i{float}=<0 (@code{FLN}, @code{FLOG}):
13003: @cindex @code{FLN}, @i{float}=<0
13004: @cindex @code{FLOG}, @i{float}=<0
1.1 anton 13005: @cindex floating-point unidentified fault, @code{FLN} or @code{FLOG}
1.80 anton 13006: Platform-dependent; on IEEE-FP systems typically produces a NaN (or a
13007: negative infinity for @i{float}=0).
1.1 anton 13008:
1.29 crook 13009: @item @i{float}<0 (@code{FASINH}, @code{FSQRT}):
13010: @cindex @code{FASINH}, @i{float}<0
13011: @cindex @code{FSQRT}, @i{float}<0
1.1 anton 13012: @cindex floating-point unidentified fault, @code{FASINH} or @code{FSQRT}
1.80 anton 13013: Platform-dependent; for @code{fsqrt} this typically gives a NaN, for
13014: @code{fasinh} some platforms produce a NaN, others a number (bug in the
13015: C library?).
1.1 anton 13016:
1.29 crook 13017: @item |@i{float}|>1 (@code{FACOS}, @code{FASIN}, @code{FATANH}):
13018: @cindex @code{FACOS}, |@i{float}|>1
13019: @cindex @code{FASIN}, |@i{float}|>1
13020: @cindex @code{FATANH}, |@i{float}|>1
1.1 anton 13021: @cindex floating-point unidentified fault, @code{FACOS}, @code{FASIN} or @code{FATANH}
1.80 anton 13022: Platform-dependent; IEEE-FP systems typically produce a NaN.
1.1 anton 13023:
1.29 crook 13024: @item integer part of float cannot be represented by @i{d} in @code{F>D}:
13025: @cindex @code{F>D}, integer part of float cannot be represented by @i{d}
1.1 anton 13026: @cindex floating-point unidentified fault, @code{F>D}
1.80 anton 13027: Platform-dependent; typically, some double number is produced and no
13028: error is reported.
1.1 anton 13029:
13030: @item string larger than pictured numeric output area (@code{f.}, @code{fe.}, @code{fs.}):
13031: @cindex string larger than pictured numeric output area (@code{f.}, @code{fe.}, @code{fs.})
1.80 anton 13032: @code{Precision} characters of the numeric output area are used. If
13033: @code{precision} is too high, these words will smash the data or code
13034: close to @code{here}.
1.1 anton 13035: @end table
13036:
13037: @c =====================================================================
13038: @node The optional Locals word set, The optional Memory-Allocation word set, The optional Floating-Point word set, ANS conformance
13039: @section The optional Locals word set
13040: @c =====================================================================
13041: @cindex system documentation, locals words
13042: @cindex locals words, system documentation
13043:
13044: @menu
13045: * locals-idef:: Implementation Defined Options
13046: * locals-ambcond:: Ambiguous Conditions
13047: @end menu
13048:
13049:
13050: @c ---------------------------------------------------------------------
13051: @node locals-idef, locals-ambcond, The optional Locals word set, The optional Locals word set
13052: @subsection Implementation Defined Options
13053: @c ---------------------------------------------------------------------
13054: @cindex implementation-defined options, locals words
13055: @cindex locals words, implementation-defined options
13056:
13057: @table @i
13058: @item maximum number of locals in a definition:
13059: @cindex maximum number of locals in a definition
13060: @cindex locals, maximum number in a definition
13061: @code{s" #locals" environment? drop .}. Currently 15. This is a lower
13062: bound, e.g., on a 32-bit machine there can be 41 locals of up to 8
13063: characters. The number of locals in a definition is bounded by the size
13064: of locals-buffer, which contains the names of the locals.
13065:
13066: @end table
13067:
13068:
13069: @c ---------------------------------------------------------------------
13070: @node locals-ambcond, , locals-idef, The optional Locals word set
13071: @subsection Ambiguous conditions
13072: @c ---------------------------------------------------------------------
13073: @cindex locals words, ambiguous conditions
13074: @cindex ambiguous conditions, locals words
13075:
13076: @table @i
13077: @item executing a named local in interpretation state:
13078: @cindex local in interpretation state
13079: @cindex Interpreting a compile-only word, for a local
13080: Locals have no interpretation semantics. If you try to perform the
13081: interpretation semantics, you will get a @code{-14 throw} somewhere
13082: (Interpreting a compile-only word). If you perform the compilation
13083: semantics, the locals access will be compiled (irrespective of state).
13084:
1.29 crook 13085: @item @i{name} not defined by @code{VALUE} or @code{(LOCAL)} (@code{TO}):
1.1 anton 13086: @cindex name not defined by @code{VALUE} or @code{(LOCAL)} used by @code{TO}
13087: @cindex @code{TO} on non-@code{VALUE}s and non-locals
13088: @cindex Invalid name argument, @code{TO}
13089: @code{-32 throw} (Invalid name argument)
13090:
13091: @end table
13092:
13093:
13094: @c =====================================================================
13095: @node The optional Memory-Allocation word set, The optional Programming-Tools word set, The optional Locals word set, ANS conformance
13096: @section The optional Memory-Allocation word set
13097: @c =====================================================================
13098: @cindex system documentation, memory-allocation words
13099: @cindex memory-allocation words, system documentation
13100:
13101: @menu
13102: * memory-idef:: Implementation Defined Options
13103: @end menu
13104:
13105:
13106: @c ---------------------------------------------------------------------
13107: @node memory-idef, , The optional Memory-Allocation word set, The optional Memory-Allocation word set
13108: @subsection Implementation Defined Options
13109: @c ---------------------------------------------------------------------
13110: @cindex implementation-defined options, memory-allocation words
13111: @cindex memory-allocation words, implementation-defined options
13112:
13113: @table @i
1.29 crook 13114: @item values and meaning of @i{ior}:
13115: @cindex @i{ior} values and meaning
13116: The @i{ior}s returned by the file and memory allocation words are
1.1 anton 13117: intended as throw codes. They typically are in the range
13118: -512@minus{}-2047 of OS errors. The mapping from OS error numbers to
1.29 crook 13119: @i{ior}s is -512@minus{}@i{errno}.
1.1 anton 13120:
13121: @end table
13122:
13123: @c =====================================================================
13124: @node The optional Programming-Tools word set, The optional Search-Order word set, The optional Memory-Allocation word set, ANS conformance
13125: @section The optional Programming-Tools word set
13126: @c =====================================================================
13127: @cindex system documentation, programming-tools words
13128: @cindex programming-tools words, system documentation
13129:
13130: @menu
13131: * programming-idef:: Implementation Defined Options
13132: * programming-ambcond:: Ambiguous Conditions
13133: @end menu
13134:
13135:
13136: @c ---------------------------------------------------------------------
13137: @node programming-idef, programming-ambcond, The optional Programming-Tools word set, The optional Programming-Tools word set
13138: @subsection Implementation Defined Options
13139: @c ---------------------------------------------------------------------
13140: @cindex implementation-defined options, programming-tools words
13141: @cindex programming-tools words, implementation-defined options
13142:
13143: @table @i
13144: @item ending sequence for input following @code{;CODE} and @code{CODE}:
13145: @cindex @code{;CODE} ending sequence
13146: @cindex @code{CODE} ending sequence
13147: @code{END-CODE}
13148:
13149: @item manner of processing input following @code{;CODE} and @code{CODE}:
13150: @cindex @code{;CODE}, processing input
13151: @cindex @code{CODE}, processing input
13152: The @code{ASSEMBLER} vocabulary is pushed on the search order stack, and
13153: the input is processed by the text interpreter, (starting) in interpret
13154: state.
13155:
13156: @item search order capability for @code{EDITOR} and @code{ASSEMBLER}:
13157: @cindex @code{ASSEMBLER}, search order capability
13158: The ANS Forth search order word set.
13159:
13160: @item source and format of display by @code{SEE}:
13161: @cindex @code{SEE}, source and format of output
1.80 anton 13162: The source for @code{see} is the executable code used by the inner
1.1 anton 13163: interpreter. The current @code{see} tries to output Forth source code
1.80 anton 13164: (and on some platforms, assembly code for primitives) as well as
13165: possible.
1.1 anton 13166:
13167: @end table
13168:
13169: @c ---------------------------------------------------------------------
13170: @node programming-ambcond, , programming-idef, The optional Programming-Tools word set
13171: @subsection Ambiguous conditions
13172: @c ---------------------------------------------------------------------
13173: @cindex programming-tools words, ambiguous conditions
13174: @cindex ambiguous conditions, programming-tools words
13175:
13176: @table @i
13177:
1.21 crook 13178: @item deleting the compilation word list (@code{FORGET}):
13179: @cindex @code{FORGET}, deleting the compilation word list
1.1 anton 13180: Not implemented (yet).
13181:
1.29 crook 13182: @item fewer than @i{u}+1 items on the control-flow stack (@code{CS-PICK}, @code{CS-ROLL}):
13183: @cindex @code{CS-PICK}, fewer than @i{u}+1 items on the control flow-stack
13184: @cindex @code{CS-ROLL}, fewer than @i{u}+1 items on the control flow-stack
1.1 anton 13185: @cindex control-flow stack underflow
13186: This typically results in an @code{abort"} with a descriptive error
13187: message (may change into a @code{-22 throw} (Control structure mismatch)
13188: in the future). You may also get a memory access error. If you are
13189: unlucky, this ambiguous condition is not caught.
13190:
1.29 crook 13191: @item @i{name} can't be found (@code{FORGET}):
13192: @cindex @code{FORGET}, @i{name} can't be found
1.1 anton 13193: Not implemented (yet).
13194:
1.29 crook 13195: @item @i{name} not defined via @code{CREATE}:
13196: @cindex @code{;CODE}, @i{name} not defined via @code{CREATE}
1.1 anton 13197: @code{;CODE} behaves like @code{DOES>} in this respect, i.e., it changes
13198: the execution semantics of the last defined word no matter how it was
13199: defined.
13200:
13201: @item @code{POSTPONE} applied to @code{[IF]}:
13202: @cindex @code{POSTPONE} applied to @code{[IF]}
13203: @cindex @code{[IF]} and @code{POSTPONE}
13204: After defining @code{: X POSTPONE [IF] ; IMMEDIATE}. @code{X} is
13205: equivalent to @code{[IF]}.
13206:
13207: @item reaching the end of the input source before matching @code{[ELSE]} or @code{[THEN]}:
13208: @cindex @code{[IF]}, end of the input source before matching @code{[ELSE]} or @code{[THEN]}
13209: Continue in the same state of conditional compilation in the next outer
13210: input source. Currently there is no warning to the user about this.
13211:
13212: @item removing a needed definition (@code{FORGET}):
13213: @cindex @code{FORGET}, removing a needed definition
13214: Not implemented (yet).
13215:
13216: @end table
13217:
13218:
13219: @c =====================================================================
13220: @node The optional Search-Order word set, , The optional Programming-Tools word set, ANS conformance
13221: @section The optional Search-Order word set
13222: @c =====================================================================
13223: @cindex system documentation, search-order words
13224: @cindex search-order words, system documentation
13225:
13226: @menu
13227: * search-idef:: Implementation Defined Options
13228: * search-ambcond:: Ambiguous Conditions
13229: @end menu
13230:
13231:
13232: @c ---------------------------------------------------------------------
13233: @node search-idef, search-ambcond, The optional Search-Order word set, The optional Search-Order word set
13234: @subsection Implementation Defined Options
13235: @c ---------------------------------------------------------------------
13236: @cindex implementation-defined options, search-order words
13237: @cindex search-order words, implementation-defined options
13238:
13239: @table @i
13240: @item maximum number of word lists in search order:
13241: @cindex maximum number of word lists in search order
13242: @cindex search order, maximum depth
13243: @code{s" wordlists" environment? drop .}. Currently 16.
13244:
13245: @item minimum search order:
13246: @cindex minimum search order
13247: @cindex search order, minimum
13248: @code{root root}.
13249:
13250: @end table
13251:
13252: @c ---------------------------------------------------------------------
13253: @node search-ambcond, , search-idef, The optional Search-Order word set
13254: @subsection Ambiguous conditions
13255: @c ---------------------------------------------------------------------
13256: @cindex search-order words, ambiguous conditions
13257: @cindex ambiguous conditions, search-order words
13258:
13259: @table @i
1.21 crook 13260: @item changing the compilation word list (during compilation):
13261: @cindex changing the compilation word list (during compilation)
13262: @cindex compilation word list, change before definition ends
13263: The word is entered into the word list that was the compilation word list
1.1 anton 13264: at the start of the definition. Any changes to the name field (e.g.,
13265: @code{immediate}) or the code field (e.g., when executing @code{DOES>})
13266: are applied to the latest defined word (as reported by @code{last} or
1.21 crook 13267: @code{lastxt}), if possible, irrespective of the compilation word list.
1.1 anton 13268:
13269: @item search order empty (@code{previous}):
13270: @cindex @code{previous}, search order empty
1.26 crook 13271: @cindex vocstack empty, @code{previous}
1.1 anton 13272: @code{abort" Vocstack empty"}.
13273:
13274: @item too many word lists in search order (@code{also}):
13275: @cindex @code{also}, too many word lists in search order
1.26 crook 13276: @cindex vocstack full, @code{also}
1.1 anton 13277: @code{abort" Vocstack full"}.
13278:
13279: @end table
13280:
13281: @c ***************************************************************
1.65 anton 13282: @node Standard vs Extensions, Model, ANS conformance, Top
13283: @chapter Should I use Gforth extensions?
13284: @cindex Gforth extensions
13285:
13286: As you read through the rest of this manual, you will see documentation
13287: for @i{Standard} words, and documentation for some appealing Gforth
13288: @i{extensions}. You might ask yourself the question: @i{``Should I
13289: restrict myself to the standard, or should I use the extensions?''}
13290:
13291: The answer depends on the goals you have for the program you are working
13292: on:
13293:
13294: @itemize @bullet
13295:
13296: @item Is it just for yourself or do you want to share it with others?
13297:
13298: @item
13299: If you want to share it, do the others all use Gforth?
13300:
13301: @item
13302: If it is just for yourself, do you want to restrict yourself to Gforth?
13303:
13304: @end itemize
13305:
13306: If restricting the program to Gforth is ok, then there is no reason not
13307: to use extensions. It is still a good idea to keep to the standard
13308: where it is easy, in case you want to reuse these parts in another
13309: program that you want to be portable.
13310:
13311: If you want to be able to port the program to other Forth systems, there
13312: are the following points to consider:
13313:
13314: @itemize @bullet
13315:
13316: @item
13317: Most Forth systems that are being maintained support the ANS Forth
13318: standard. So if your program complies with the standard, it will be
13319: portable among many systems.
13320:
13321: @item
13322: A number of the Gforth extensions can be implemented in ANS Forth using
13323: public-domain files provided in the @file{compat/} directory. These are
13324: mentioned in the text in passing. There is no reason not to use these
13325: extensions, your program will still be ANS Forth compliant; just include
13326: the appropriate compat files with your program.
13327:
13328: @item
13329: The tool @file{ans-report.fs} (@pxref{ANS Report}) makes it easy to
13330: analyse your program and determine what non-Standard words it relies
13331: upon. However, it does not check whether you use standard words in a
13332: non-standard way.
13333:
13334: @item
13335: Some techniques are not standardized by ANS Forth, and are hard or
13336: impossible to implement in a standard way, but can be implemented in
13337: most Forth systems easily, and usually in similar ways (e.g., accessing
13338: word headers). Forth has a rich historical precedent for programmers
13339: taking advantage of implementation-dependent features of their tools
13340: (for example, relying on a knowledge of the dictionary
13341: structure). Sometimes these techniques are necessary to extract every
13342: last bit of performance from the hardware, sometimes they are just a
13343: programming shorthand.
13344:
13345: @item
13346: Does using a Gforth extension save more work than the porting this part
13347: to other Forth systems (if any) will cost?
13348:
13349: @item
13350: Is the additional functionality worth the reduction in portability and
13351: the additional porting problems?
13352:
13353: @end itemize
13354:
13355: In order to perform these consideratios, you need to know what's
13356: standard and what's not. This manual generally states if something is
1.81 ! anton 13357: non-standard, but the authoritative source is the
! 13358: @uref{http://www.taygeta.com/forth/dpans.html,standard document}.
1.65 anton 13359: Appendix A of the Standard (@var{Rationale}) provides a valuable insight
13360: into the thought processes of the technical committee.
13361:
13362: Note also that portability between Forth systems is not the only
13363: portability issue; there is also the issue of portability between
13364: different platforms (processor/OS combinations).
13365:
13366: @c ***************************************************************
13367: @node Model, Integrating Gforth, Standard vs Extensions, Top
1.1 anton 13368: @chapter Model
13369:
13370: This chapter has yet to be written. It will contain information, on
13371: which internal structures you can rely.
13372:
13373: @c ***************************************************************
13374: @node Integrating Gforth, Emacs and Gforth, Model, Top
13375: @chapter Integrating Gforth into C programs
13376:
13377: This is not yet implemented.
13378:
13379: Several people like to use Forth as scripting language for applications
13380: that are otherwise written in C, C++, or some other language.
13381:
13382: The Forth system ATLAST provides facilities for embedding it into
13383: applications; unfortunately it has several disadvantages: most
13384: importantly, it is not based on ANS Forth, and it is apparently dead
13385: (i.e., not developed further and not supported). The facilities
1.21 crook 13386: provided by Gforth in this area are inspired by ATLAST's facilities, so
1.1 anton 13387: making the switch should not be hard.
13388:
13389: We also tried to design the interface such that it can easily be
13390: implemented by other Forth systems, so that we may one day arrive at a
13391: standardized interface. Such a standard interface would allow you to
13392: replace the Forth system without having to rewrite C code.
13393:
13394: You embed the Gforth interpreter by linking with the library
13395: @code{libgforth.a} (give the compiler the option @code{-lgforth}). All
13396: global symbols in this library that belong to the interface, have the
13397: prefix @code{forth_}. (Global symbols that are used internally have the
13398: prefix @code{gforth_}).
13399:
13400: You can include the declarations of Forth types and the functions and
13401: variables of the interface with @code{#include <forth.h>}.
13402:
13403: Types.
13404:
13405: Variables.
13406:
13407: Data and FP Stack pointer. Area sizes.
13408:
13409: functions.
13410:
13411: forth_init(imagefile)
13412: forth_evaluate(string) exceptions?
13413: forth_goto(address) (or forth_execute(xt)?)
13414: forth_continue() (a corountining mechanism)
13415:
13416: Adding primitives.
13417:
13418: No checking.
13419:
13420: Signals?
13421:
13422: Accessing the Stacks
13423:
1.26 crook 13424: @c ******************************************************************
1.1 anton 13425: @node Emacs and Gforth, Image Files, Integrating Gforth, Top
13426: @chapter Emacs and Gforth
13427: @cindex Emacs and Gforth
13428:
13429: @cindex @file{gforth.el}
13430: @cindex @file{forth.el}
13431: @cindex Rydqvist, Goran
13432: @cindex comment editing commands
13433: @cindex @code{\}, editing with Emacs
13434: @cindex debug tracer editing commands
13435: @cindex @code{~~}, removal with Emacs
13436: @cindex Forth mode in Emacs
13437: Gforth comes with @file{gforth.el}, an improved version of
13438: @file{forth.el} by Goran Rydqvist (included in the TILE package). The
1.26 crook 13439: improvements are:
13440:
13441: @itemize @bullet
13442: @item
13443: A better (but still not perfect) handling of indentation.
13444: @item
13445: Comment paragraph filling (@kbd{M-q})
13446: @item
13447: Commenting (@kbd{C-x \}) and uncommenting (@kbd{C-u C-x \}) of regions
13448: @item
13449: Removal of debugging tracers (@kbd{C-x ~}, @pxref{Debugging}).
1.41 anton 13450: @item
13451: Support of the @code{info-lookup} feature for looking up the
13452: documentation of a word.
1.26 crook 13453: @end itemize
13454:
13455: I left the stuff I do not use alone, even though some of it only makes
13456: sense for TILE. To get a description of these features, enter Forth mode
13457: and type @kbd{C-h m}.
1.1 anton 13458:
13459: @cindex source location of error or debugging output in Emacs
13460: @cindex error output, finding the source location in Emacs
13461: @cindex debugging output, finding the source location in Emacs
13462: In addition, Gforth supports Emacs quite well: The source code locations
13463: given in error messages, debugging output (from @code{~~}) and failed
13464: assertion messages are in the right format for Emacs' compilation mode
13465: (@pxref{Compilation, , Running Compilations under Emacs, emacs, Emacs
13466: Manual}) so the source location corresponding to an error or other
13467: message is only a few keystrokes away (@kbd{C-x `} for the next error,
13468: @kbd{C-c C-c} for the error under the cursor).
13469:
13470: @cindex @file{TAGS} file
13471: @cindex @file{etags.fs}
13472: @cindex viewing the source of a word in Emacs
1.43 anton 13473: @cindex @code{require}, placement in files
13474: @cindex @code{include}, placement in files
13475: Also, if you @code{require} @file{etags.fs}, a new @file{TAGS} file will
1.26 crook 13476: be produced (@pxref{Tags, , Tags Tables, emacs, Emacs Manual}) that
1.1 anton 13477: contains the definitions of all words defined afterwards. You can then
13478: find the source for a word using @kbd{M-.}. Note that emacs can use
13479: several tags files at the same time (e.g., one for the Gforth sources
13480: and one for your program, @pxref{Select Tags Table,,Selecting a Tags
13481: Table,emacs, Emacs Manual}). The TAGS file for the preloaded words is
13482: @file{$(datadir)/gforth/$(VERSION)/TAGS} (e.g.,
1.43 anton 13483: @file{/usr/local/share/gforth/0.2.0/TAGS}). To get the best behaviour
13484: with @file{etags.fs}, you should avoid putting definitions both before
13485: and after @code{require} etc., otherwise you will see the same file
13486: visited several times by commands like @code{tags-search}.
1.1 anton 13487:
1.41 anton 13488: @cindex viewing the documentation of a word in Emacs
13489: @cindex context-sensitive help
13490: Moreover, for words documented in this manual, you can look up the
13491: glossary entry quickly by using @kbd{C-h TAB}
1.80 anton 13492: (@code{info-lookup-symbol}, @pxref{Documentation, ,Documentation
1.41 anton 13493: Commands, emacs, Emacs Manual}). This feature requires Emacs 20.3 or
1.42 anton 13494: later and does not work for words containing @code{:}.
1.41 anton 13495:
13496:
1.1 anton 13497: @cindex @file{.emacs}
13498: To get all these benefits, add the following lines to your @file{.emacs}
13499: file:
13500:
13501: @example
13502: (autoload 'forth-mode "gforth.el")
13503: (setq auto-mode-alist (cons '("\\.fs\\'" . forth-mode) auto-mode-alist))
13504: @end example
13505:
1.26 crook 13506: @c ******************************************************************
1.1 anton 13507: @node Image Files, Engine, Emacs and Gforth, Top
13508: @chapter Image Files
1.26 crook 13509: @cindex image file
13510: @cindex @file{.fi} files
1.1 anton 13511: @cindex precompiled Forth code
13512: @cindex dictionary in persistent form
13513: @cindex persistent form of dictionary
13514:
13515: An image file is a file containing an image of the Forth dictionary,
13516: i.e., compiled Forth code and data residing in the dictionary. By
13517: convention, we use the extension @code{.fi} for image files.
13518:
13519: @menu
1.18 anton 13520: * Image Licensing Issues:: Distribution terms for images.
13521: * Image File Background:: Why have image files?
1.67 anton 13522: * Non-Relocatable Image Files:: don't always work.
1.18 anton 13523: * Data-Relocatable Image Files:: are better.
1.67 anton 13524: * Fully Relocatable Image Files:: better yet.
1.18 anton 13525: * Stack and Dictionary Sizes:: Setting the default sizes for an image.
1.29 crook 13526: * Running Image Files:: @code{gforth -i @i{file}} or @i{file}.
1.18 anton 13527: * Modifying the Startup Sequence:: and turnkey applications.
1.1 anton 13528: @end menu
13529:
1.18 anton 13530: @node Image Licensing Issues, Image File Background, Image Files, Image Files
13531: @section Image Licensing Issues
13532: @cindex license for images
13533: @cindex image license
13534:
13535: An image created with @code{gforthmi} (@pxref{gforthmi}) or
13536: @code{savesystem} (@pxref{Non-Relocatable Image Files}) includes the
13537: original image; i.e., according to copyright law it is a derived work of
13538: the original image.
13539:
13540: Since Gforth is distributed under the GNU GPL, the newly created image
13541: falls under the GNU GPL, too. In particular, this means that if you
13542: distribute the image, you have to make all of the sources for the image
13543: available, including those you wrote. For details see @ref{License, ,
13544: GNU General Public License (Section 3)}.
13545:
13546: If you create an image with @code{cross} (@pxref{cross.fs}), the image
13547: contains only code compiled from the sources you gave it; if none of
13548: these sources is under the GPL, the terms discussed above do not apply
13549: to the image. However, if your image needs an engine (a gforth binary)
13550: that is under the GPL, you should make sure that you distribute both in
13551: a way that is at most a @emph{mere aggregation}, if you don't want the
13552: terms of the GPL to apply to the image.
13553:
13554: @node Image File Background, Non-Relocatable Image Files, Image Licensing Issues, Image Files
1.1 anton 13555: @section Image File Background
13556: @cindex image file background
13557:
1.80 anton 13558: Gforth consists not only of primitives (in the engine), but also of
1.1 anton 13559: definitions written in Forth. Since the Forth compiler itself belongs to
13560: those definitions, it is not possible to start the system with the
1.80 anton 13561: engine and the Forth source alone. Therefore we provide the Forth
1.26 crook 13562: code as an image file in nearly executable form. When Gforth starts up,
13563: a C routine loads the image file into memory, optionally relocates the
13564: addresses, then sets up the memory (stacks etc.) according to
13565: information in the image file, and (finally) starts executing Forth
13566: code.
1.1 anton 13567:
13568: The image file variants represent different compromises between the
13569: goals of making it easy to generate image files and making them
13570: portable.
13571:
13572: @cindex relocation at run-time
1.26 crook 13573: Win32Forth 3.4 and Mitch Bradley's @code{cforth} use relocation at
1.1 anton 13574: run-time. This avoids many of the complications discussed below (image
13575: files are data relocatable without further ado), but costs performance
13576: (one addition per memory access).
13577:
13578: @cindex relocation at load-time
1.26 crook 13579: By contrast, the Gforth loader performs relocation at image load time. The
13580: loader also has to replace tokens that represent primitive calls with the
1.1 anton 13581: appropriate code-field addresses (or code addresses in the case of
13582: direct threading).
13583:
13584: There are three kinds of image files, with different degrees of
13585: relocatability: non-relocatable, data-relocatable, and fully relocatable
13586: image files.
13587:
13588: @cindex image file loader
13589: @cindex relocating loader
13590: @cindex loader for image files
13591: These image file variants have several restrictions in common; they are
13592: caused by the design of the image file loader:
13593:
13594: @itemize @bullet
13595: @item
13596: There is only one segment; in particular, this means, that an image file
13597: cannot represent @code{ALLOCATE}d memory chunks (and pointers to
1.26 crook 13598: them). The contents of the stacks are not represented, either.
1.1 anton 13599:
13600: @item
13601: The only kinds of relocation supported are: adding the same offset to
13602: all cells that represent data addresses; and replacing special tokens
13603: with code addresses or with pieces of machine code.
13604:
13605: If any complex computations involving addresses are performed, the
13606: results cannot be represented in the image file. Several applications that
13607: use such computations come to mind:
13608: @itemize @minus
13609: @item
13610: Hashing addresses (or data structures which contain addresses) for table
13611: lookup. If you use Gforth's @code{table}s or @code{wordlist}s for this
13612: purpose, you will have no problem, because the hash tables are
13613: recomputed automatically when the system is started. If you use your own
13614: hash tables, you will have to do something similar.
13615:
13616: @item
13617: There's a cute implementation of doubly-linked lists that uses
13618: @code{XOR}ed addresses. You could represent such lists as singly-linked
13619: in the image file, and restore the doubly-linked representation on
13620: startup.@footnote{In my opinion, though, you should think thrice before
13621: using a doubly-linked list (whatever implementation).}
13622:
13623: @item
13624: The code addresses of run-time routines like @code{docol:} cannot be
13625: represented in the image file (because their tokens would be replaced by
13626: machine code in direct threaded implementations). As a workaround,
13627: compute these addresses at run-time with @code{>code-address} from the
13628: executions tokens of appropriate words (see the definitions of
1.80 anton 13629: @code{docol:} and friends in @file{kernel/getdoers.fs}).
1.1 anton 13630:
13631: @item
13632: On many architectures addresses are represented in machine code in some
13633: shifted or mangled form. You cannot put @code{CODE} words that contain
13634: absolute addresses in this form in a relocatable image file. Workarounds
13635: are representing the address in some relative form (e.g., relative to
13636: the CFA, which is present in some register), or loading the address from
13637: a place where it is stored in a non-mangled form.
13638: @end itemize
13639: @end itemize
13640:
13641: @node Non-Relocatable Image Files, Data-Relocatable Image Files, Image File Background, Image Files
13642: @section Non-Relocatable Image Files
13643: @cindex non-relocatable image files
1.26 crook 13644: @cindex image file, non-relocatable
1.1 anton 13645:
13646: These files are simple memory dumps of the dictionary. They are specific
13647: to the executable (i.e., @file{gforth} file) they were created
13648: with. What's worse, they are specific to the place on which the
13649: dictionary resided when the image was created. Now, there is no
13650: guarantee that the dictionary will reside at the same place the next
13651: time you start Gforth, so there's no guarantee that a non-relocatable
13652: image will work the next time (Gforth will complain instead of crashing,
13653: though).
13654:
13655: You can create a non-relocatable image file with
13656:
1.44 crook 13657:
1.1 anton 13658: doc-savesystem
13659:
1.44 crook 13660:
1.1 anton 13661: @node Data-Relocatable Image Files, Fully Relocatable Image Files, Non-Relocatable Image Files, Image Files
13662: @section Data-Relocatable Image Files
13663: @cindex data-relocatable image files
1.26 crook 13664: @cindex image file, data-relocatable
1.1 anton 13665:
13666: These files contain relocatable data addresses, but fixed code addresses
13667: (instead of tokens). They are specific to the executable (i.e.,
13668: @file{gforth} file) they were created with. For direct threading on some
13669: architectures (e.g., the i386), data-relocatable images do not work. You
13670: get a data-relocatable image, if you use @file{gforthmi} with a
13671: Gforth binary that is not doubly indirect threaded (@pxref{Fully
13672: Relocatable Image Files}).
13673:
13674: @node Fully Relocatable Image Files, Stack and Dictionary Sizes, Data-Relocatable Image Files, Image Files
13675: @section Fully Relocatable Image Files
13676: @cindex fully relocatable image files
1.26 crook 13677: @cindex image file, fully relocatable
1.1 anton 13678:
13679: @cindex @file{kern*.fi}, relocatability
13680: @cindex @file{gforth.fi}, relocatability
13681: These image files have relocatable data addresses, and tokens for code
13682: addresses. They can be used with different binaries (e.g., with and
13683: without debugging) on the same machine, and even across machines with
13684: the same data formats (byte order, cell size, floating point
13685: format). However, they are usually specific to the version of Gforth
13686: they were created with. The files @file{gforth.fi} and @file{kernl*.fi}
13687: are fully relocatable.
13688:
13689: There are two ways to create a fully relocatable image file:
13690:
13691: @menu
1.29 crook 13692: * gforthmi:: The normal way
1.1 anton 13693: * cross.fs:: The hard way
13694: @end menu
13695:
13696: @node gforthmi, cross.fs, Fully Relocatable Image Files, Fully Relocatable Image Files
13697: @subsection @file{gforthmi}
13698: @cindex @file{comp-i.fs}
13699: @cindex @file{gforthmi}
13700:
13701: You will usually use @file{gforthmi}. If you want to create an
1.29 crook 13702: image @i{file} that contains everything you would load by invoking
13703: Gforth with @code{gforth @i{options}}, you simply say:
1.1 anton 13704: @example
1.29 crook 13705: gforthmi @i{file} @i{options}
1.1 anton 13706: @end example
13707:
13708: E.g., if you want to create an image @file{asm.fi} that has the file
13709: @file{asm.fs} loaded in addition to the usual stuff, you could do it
13710: like this:
13711:
13712: @example
13713: gforthmi asm.fi asm.fs
13714: @end example
13715:
1.27 crook 13716: @file{gforthmi} is implemented as a sh script and works like this: It
13717: produces two non-relocatable images for different addresses and then
13718: compares them. Its output reflects this: first you see the output (if
1.62 crook 13719: any) of the two Gforth invocations that produce the non-relocatable image
1.27 crook 13720: files, then you see the output of the comparing program: It displays the
13721: offset used for data addresses and the offset used for code addresses;
1.1 anton 13722: moreover, for each cell that cannot be represented correctly in the
1.44 crook 13723: image files, it displays a line like this:
1.1 anton 13724:
13725: @example
13726: 78DC BFFFFA50 BFFFFA40
13727: @end example
13728:
13729: This means that at offset $78dc from @code{forthstart}, one input image
13730: contains $bffffa50, and the other contains $bffffa40. Since these cells
13731: cannot be represented correctly in the output image, you should examine
13732: these places in the dictionary and verify that these cells are dead
13733: (i.e., not read before they are written).
1.39 anton 13734:
13735: @cindex --application, @code{gforthmi} option
13736: If you insert the option @code{--application} in front of the image file
13737: name, you will get an image that uses the @code{--appl-image} option
13738: instead of the @code{--image-file} option (@pxref{Invoking
13739: Gforth}). When you execute such an image on Unix (by typing the image
13740: name as command), the Gforth engine will pass all options to the image
13741: instead of trying to interpret them as engine options.
1.1 anton 13742:
1.27 crook 13743: If you type @file{gforthmi} with no arguments, it prints some usage
13744: instructions.
13745:
1.1 anton 13746: @cindex @code{savesystem} during @file{gforthmi}
13747: @cindex @code{bye} during @file{gforthmi}
13748: @cindex doubly indirect threaded code
1.44 crook 13749: @cindex environment variables
13750: @cindex @code{GFORTHD} -- environment variable
13751: @cindex @code{GFORTH} -- environment variable
1.1 anton 13752: @cindex @code{gforth-ditc}
1.29 crook 13753: There are a few wrinkles: After processing the passed @i{options}, the
1.1 anton 13754: words @code{savesystem} and @code{bye} must be visible. A special doubly
13755: indirect threaded version of the @file{gforth} executable is used for
1.62 crook 13756: creating the non-relocatable images; you can pass the exact filename of
1.1 anton 13757: this executable through the environment variable @code{GFORTHD}
13758: (default: @file{gforth-ditc}); if you pass a version that is not doubly
13759: indirect threaded, you will not get a fully relocatable image, but a
1.27 crook 13760: data-relocatable image (because there is no code address offset). The
13761: normal @file{gforth} executable is used for creating the relocatable
13762: image; you can pass the exact filename of this executable through the
13763: environment variable @code{GFORTH}.
1.1 anton 13764:
13765: @node cross.fs, , gforthmi, Fully Relocatable Image Files
13766: @subsection @file{cross.fs}
13767: @cindex @file{cross.fs}
13768: @cindex cross-compiler
13769: @cindex metacompiler
1.47 crook 13770: @cindex target compiler
1.1 anton 13771:
13772: You can also use @code{cross}, a batch compiler that accepts a Forth-like
1.47 crook 13773: programming language (@pxref{Cross Compiler}).
1.1 anton 13774:
1.47 crook 13775: @code{cross} allows you to create image files for machines with
1.1 anton 13776: different data sizes and data formats than the one used for generating
13777: the image file. You can also use it to create an application image that
13778: does not contain a Forth compiler. These features are bought with
13779: restrictions and inconveniences in programming. E.g., addresses have to
13780: be stored in memory with special words (@code{A!}, @code{A,}, etc.) in
13781: order to make the code relocatable.
13782:
13783:
13784: @node Stack and Dictionary Sizes, Running Image Files, Fully Relocatable Image Files, Image Files
13785: @section Stack and Dictionary Sizes
13786: @cindex image file, stack and dictionary sizes
13787: @cindex dictionary size default
13788: @cindex stack size default
13789:
13790: If you invoke Gforth with a command line flag for the size
13791: (@pxref{Invoking Gforth}), the size you specify is stored in the
13792: dictionary. If you save the dictionary with @code{savesystem} or create
13793: an image with @file{gforthmi}, this size will become the default
13794: for the resulting image file. E.g., the following will create a
1.21 crook 13795: fully relocatable version of @file{gforth.fi} with a 1MB dictionary:
1.1 anton 13796:
13797: @example
13798: gforthmi gforth.fi -m 1M
13799: @end example
13800:
13801: In other words, if you want to set the default size for the dictionary
13802: and the stacks of an image, just invoke @file{gforthmi} with the
13803: appropriate options when creating the image.
13804:
13805: @cindex stack size, cache-friendly
13806: Note: For cache-friendly behaviour (i.e., good performance), you should
13807: make the sizes of the stacks modulo, say, 2K, somewhat different. E.g.,
13808: the default stack sizes are: data: 16k (mod 2k=0); fp: 15.5k (mod
13809: 2k=1.5k); return: 15k(mod 2k=1k); locals: 14.5k (mod 2k=0.5k).
13810:
13811: @node Running Image Files, Modifying the Startup Sequence, Stack and Dictionary Sizes, Image Files
13812: @section Running Image Files
13813: @cindex running image files
13814: @cindex invoking image files
13815: @cindex image file invocation
13816:
13817: @cindex -i, invoke image file
13818: @cindex --image file, invoke image file
1.29 crook 13819: You can invoke Gforth with an image file @i{image} instead of the
1.1 anton 13820: default @file{gforth.fi} with the @code{-i} flag (@pxref{Invoking Gforth}):
13821: @example
1.29 crook 13822: gforth -i @i{image}
1.1 anton 13823: @end example
13824:
13825: @cindex executable image file
1.26 crook 13826: @cindex image file, executable
1.1 anton 13827: If your operating system supports starting scripts with a line of the
13828: form @code{#! ...}, you just have to type the image file name to start
13829: Gforth with this image file (note that the file extension @code{.fi} is
1.29 crook 13830: just a convention). I.e., to run Gforth with the image file @i{image},
13831: you can just type @i{image} instead of @code{gforth -i @i{image}}.
1.27 crook 13832: This works because every @code{.fi} file starts with a line of this
13833: format:
13834:
13835: @example
13836: #! /usr/local/bin/gforth-0.4.0 -i
13837: @end example
13838:
13839: The file and pathname for the Gforth engine specified on this line is
13840: the specific Gforth executable that it was built against; i.e. the value
13841: of the environment variable @code{GFORTH} at the time that
13842: @file{gforthmi} was executed.
1.1 anton 13843:
1.27 crook 13844: You can make use of the same shell capability to make a Forth source
13845: file into an executable. For example, if you place this text in a file:
1.26 crook 13846:
13847: @example
13848: #! /usr/local/bin/gforth
13849:
13850: ." Hello, world" CR
13851: bye
13852: @end example
13853:
13854: @noindent
1.27 crook 13855: and then make the file executable (chmod +x in Unix), you can run it
1.26 crook 13856: directly from the command line. The sequence @code{#!} is used in two
13857: ways; firstly, it is recognised as a ``magic sequence'' by the operating
1.29 crook 13858: system@footnote{The Unix kernel actually recognises two types of files:
13859: executable files and files of data, where the data is processed by an
13860: interpreter that is specified on the ``interpreter line'' -- the first
13861: line of the file, starting with the sequence #!. There may be a small
13862: limit (e.g., 32) on the number of characters that may be specified on
13863: the interpreter line.} secondly it is treated as a comment character by
13864: Gforth. Because of the second usage, a space is required between
1.80 anton 13865: @code{#!} and the path to the executable (moreover, some Unixes
13866: require the sequence @code{#! /}).
1.27 crook 13867:
13868: The disadvantage of this latter technique, compared with using
1.80 anton 13869: @file{gforthmi}, is that it is slightly slower; the Forth source code is
13870: compiled on-the-fly, each time the program is invoked.
1.26 crook 13871:
1.1 anton 13872: doc-#!
13873:
1.44 crook 13874:
1.1 anton 13875: @node Modifying the Startup Sequence, , Running Image Files, Image Files
13876: @section Modifying the Startup Sequence
13877: @cindex startup sequence for image file
13878: @cindex image file initialization sequence
13879: @cindex initialization sequence of image file
13880:
13881: You can add your own initialization to the startup sequence through the
1.26 crook 13882: deferred word @code{'cold}. @code{'cold} is invoked just before the
1.80 anton 13883: image-specific command line processing (i.e., loading files and
1.26 crook 13884: evaluating (@code{-e}) strings) starts.
1.1 anton 13885:
13886: A sequence for adding your initialization usually looks like this:
13887:
13888: @example
13889: :noname
13890: Defers 'cold \ do other initialization stuff (e.g., rehashing wordlists)
13891: ... \ your stuff
13892: ; IS 'cold
13893: @end example
13894:
13895: @cindex turnkey image files
1.26 crook 13896: @cindex image file, turnkey applications
1.1 anton 13897: You can make a turnkey image by letting @code{'cold} execute a word
13898: (your turnkey application) that never returns; instead, it exits Gforth
13899: via @code{bye} or @code{throw}.
13900:
13901: @cindex command-line arguments, access
13902: @cindex arguments on the command line, access
13903: You can access the (image-specific) command-line arguments through the
1.26 crook 13904: variables @code{argc} and @code{argv}. @code{arg} provides convenient
1.1 anton 13905: access to @code{argv}.
13906:
1.26 crook 13907: If @code{'cold} exits normally, Gforth processes the command-line
13908: arguments as files to be loaded and strings to be evaluated. Therefore,
13909: @code{'cold} should remove the arguments it has used in this case.
13910:
1.44 crook 13911:
13912:
1.26 crook 13913: doc-'cold
1.1 anton 13914: doc-argc
13915: doc-argv
13916: doc-arg
13917:
13918:
1.44 crook 13919:
1.1 anton 13920: @c ******************************************************************
1.13 pazsan 13921: @node Engine, Binding to System Library, Image Files, Top
1.1 anton 13922: @chapter Engine
13923: @cindex engine
13924: @cindex virtual machine
13925:
1.26 crook 13926: Reading this chapter is not necessary for programming with Gforth. It
1.1 anton 13927: may be helpful for finding your way in the Gforth sources.
13928:
1.66 anton 13929: The ideas in this section have also been published in Bernd Paysan,
13930: @cite{ANS fig/GNU/??? Forth} (in German), Forth-Tagung '93 and M. Anton
13931: Ertl, @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl93.ps.Z, A
13932: Portable Forth Engine}}, EuroForth '93.
1.1 anton 13933:
13934: @menu
13935: * Portability::
13936: * Threading::
13937: * Primitives::
13938: * Performance::
13939: @end menu
13940:
13941: @node Portability, Threading, Engine, Engine
13942: @section Portability
13943: @cindex engine portability
13944:
1.26 crook 13945: An important goal of the Gforth Project is availability across a wide
13946: range of personal machines. fig-Forth, and, to a lesser extent, F83,
13947: achieved this goal by manually coding the engine in assembly language
13948: for several then-popular processors. This approach is very
13949: labor-intensive and the results are short-lived due to progress in
13950: computer architecture.
1.1 anton 13951:
13952: @cindex C, using C for the engine
13953: Others have avoided this problem by coding in C, e.g., Mitch Bradley
13954: (cforth), Mikael Patel (TILE) and Dirk Zoller (pfe). This approach is
13955: particularly popular for UNIX-based Forths due to the large variety of
13956: architectures of UNIX machines. Unfortunately an implementation in C
13957: does not mix well with the goals of efficiency and with using
13958: traditional techniques: Indirect or direct threading cannot be expressed
13959: in C, and switch threading, the fastest technique available in C, is
13960: significantly slower. Another problem with C is that it is very
13961: cumbersome to express double integer arithmetic.
13962:
13963: @cindex GNU C for the engine
13964: @cindex long long
13965: Fortunately, there is a portable language that does not have these
13966: limitations: GNU C, the version of C processed by the GNU C compiler
13967: (@pxref{C Extensions, , Extensions to the C Language Family, gcc.info,
13968: GNU C Manual}). Its labels as values feature (@pxref{Labels as Values, ,
13969: Labels as Values, gcc.info, GNU C Manual}) makes direct and indirect
13970: threading possible, its @code{long long} type (@pxref{Long Long, ,
13971: Double-Word Integers, gcc.info, GNU C Manual}) corresponds to Forth's
13972: double numbers@footnote{Unfortunately, long longs are not implemented
13973: properly on all machines (e.g., on alpha-osf1, long longs are only 64
13974: bits, the same size as longs (and pointers), but they should be twice as
1.4 anton 13975: long according to @pxref{Long Long, , Double-Word Integers, gcc.info, GNU
1.1 anton 13976: C Manual}). So, we had to implement doubles in C after all. Still, on
13977: most machines we can use long longs and achieve better performance than
13978: with the emulation package.}. GNU C is available for free on all
13979: important (and many unimportant) UNIX machines, VMS, 80386s running
13980: MS-DOS, the Amiga, and the Atari ST, so a Forth written in GNU C can run
13981: on all these machines.
13982:
13983: Writing in a portable language has the reputation of producing code that
13984: is slower than assembly. For our Forth engine we repeatedly looked at
13985: the code produced by the compiler and eliminated most compiler-induced
13986: inefficiencies by appropriate changes in the source code.
13987:
13988: @cindex explicit register declarations
13989: @cindex --enable-force-reg, configuration flag
13990: @cindex -DFORCE_REG
13991: However, register allocation cannot be portably influenced by the
13992: programmer, leading to some inefficiencies on register-starved
13993: machines. We use explicit register declarations (@pxref{Explicit Reg
13994: Vars, , Variables in Specified Registers, gcc.info, GNU C Manual}) to
13995: improve the speed on some machines. They are turned on by using the
13996: configuration flag @code{--enable-force-reg} (@code{gcc} switch
13997: @code{-DFORCE_REG}). Unfortunately, this feature not only depends on the
13998: machine, but also on the compiler version: On some machines some
13999: compiler versions produce incorrect code when certain explicit register
14000: declarations are used. So by default @code{-DFORCE_REG} is not used.
14001:
14002: @node Threading, Primitives, Portability, Engine
14003: @section Threading
14004: @cindex inner interpreter implementation
14005: @cindex threaded code implementation
14006:
14007: @cindex labels as values
14008: GNU C's labels as values extension (available since @code{gcc-2.0},
14009: @pxref{Labels as Values, , Labels as Values, gcc.info, GNU C Manual})
1.29 crook 14010: makes it possible to take the address of @i{label} by writing
14011: @code{&&@i{label}}. This address can then be used in a statement like
14012: @code{goto *@i{address}}. I.e., @code{goto *&&x} is the same as
1.1 anton 14013: @code{goto x}.
14014:
1.26 crook 14015: @cindex @code{NEXT}, indirect threaded
1.1 anton 14016: @cindex indirect threaded inner interpreter
14017: @cindex inner interpreter, indirect threaded
1.26 crook 14018: With this feature an indirect threaded @code{NEXT} looks like:
1.1 anton 14019: @example
14020: cfa = *ip++;
14021: ca = *cfa;
14022: goto *ca;
14023: @end example
14024: @cindex instruction pointer
14025: For those unfamiliar with the names: @code{ip} is the Forth instruction
14026: pointer; the @code{cfa} (code-field address) corresponds to ANS Forths
14027: execution token and points to the code field of the next word to be
14028: executed; The @code{ca} (code address) fetched from there points to some
14029: executable code, e.g., a primitive or the colon definition handler
14030: @code{docol}.
14031:
1.26 crook 14032: @cindex @code{NEXT}, direct threaded
1.1 anton 14033: @cindex direct threaded inner interpreter
14034: @cindex inner interpreter, direct threaded
14035: Direct threading is even simpler:
14036: @example
14037: ca = *ip++;
14038: goto *ca;
14039: @end example
14040:
14041: Of course we have packaged the whole thing neatly in macros called
1.26 crook 14042: @code{NEXT} and @code{NEXT1} (the part of @code{NEXT} after fetching the cfa).
1.1 anton 14043:
14044: @menu
14045: * Scheduling::
14046: * Direct or Indirect Threaded?::
14047: * DOES>::
14048: @end menu
14049:
14050: @node Scheduling, Direct or Indirect Threaded?, Threading, Threading
14051: @subsection Scheduling
14052: @cindex inner interpreter optimization
14053:
14054: There is a little complication: Pipelined and superscalar processors,
14055: i.e., RISC and some modern CISC machines can process independent
14056: instructions while waiting for the results of an instruction. The
14057: compiler usually reorders (schedules) the instructions in a way that
14058: achieves good usage of these delay slots. However, on our first tries
14059: the compiler did not do well on scheduling primitives. E.g., for
14060: @code{+} implemented as
14061: @example
14062: n=sp[0]+sp[1];
14063: sp++;
14064: sp[0]=n;
14065: NEXT;
14066: @end example
1.81 ! anton 14067: the @code{NEXT} comes strictly after the other code, i.e., there is
! 14068: nearly no scheduling. After a little thought the problem becomes clear:
! 14069: The compiler cannot know that @code{sp} and @code{ip} point to different
1.21 crook 14070: addresses (and the version of @code{gcc} we used would not know it even
14071: if it was possible), so it could not move the load of the cfa above the
14072: store to the TOS. Indeed the pointers could be the same, if code on or
14073: very near the top of stack were executed. In the interest of speed we
14074: chose to forbid this probably unused ``feature'' and helped the compiler
1.81 ! anton 14075: in scheduling: @code{NEXT} is divided into several parts:
! 14076: @code{NEXT_P0}, @code{NEXT_P1} and @code{NEXT_P2}). @code{+} now looks
! 14077: like:
1.1 anton 14078: @example
1.81 ! anton 14079: NEXT_P0;
1.1 anton 14080: n=sp[0]+sp[1];
14081: sp++;
14082: NEXT_P1;
14083: sp[0]=n;
14084: NEXT_P2;
14085: @end example
14086:
1.81 ! anton 14087: There are various schemes that distribute the different operations of
! 14088: NEXT between these parts in several ways; in general, different schemes
! 14089: perform best on different processors. We use a scheme for most
! 14090: architectures that performs well for most processors of this
! 14091: architecture; in the furture we may switch to benchmarking and chosing
! 14092: the scheme on installation time.
! 14093:
1.1 anton 14094:
14095: @node Direct or Indirect Threaded?, DOES>, Scheduling, Threading
14096: @subsection Direct or Indirect Threaded?
14097: @cindex threading, direct or indirect?
14098:
14099: @cindex -DDIRECT_THREADED
14100: Both! After packaging the nasty details in macro definitions we
14101: realized that we could switch between direct and indirect threading by
14102: simply setting a compilation flag (@code{-DDIRECT_THREADED}) and
14103: defining a few machine-specific macros for the direct-threading case.
14104: On the Forth level we also offer access words that hide the
14105: differences between the threading methods (@pxref{Threading Words}).
14106:
14107: Indirect threading is implemented completely machine-independently.
14108: Direct threading needs routines for creating jumps to the executable
1.21 crook 14109: code (e.g. to @code{docol} or @code{dodoes}). These routines are inherently
14110: machine-dependent, but they do not amount to many source lines. Therefore,
14111: even porting direct threading to a new machine requires little effort.
1.1 anton 14112:
14113: @cindex --enable-indirect-threaded, configuration flag
14114: @cindex --enable-direct-threaded, configuration flag
14115: The default threading method is machine-dependent. You can enforce a
14116: specific threading method when building Gforth with the configuration
14117: flag @code{--enable-direct-threaded} or
14118: @code{--enable-indirect-threaded}. Note that direct threading is not
14119: supported on all machines.
14120:
14121: @node DOES>, , Direct or Indirect Threaded?, Threading
14122: @subsection DOES>
14123: @cindex @code{DOES>} implementation
14124:
1.26 crook 14125: @cindex @code{dodoes} routine
14126: @cindex @code{DOES>}-code
1.1 anton 14127: One of the most complex parts of a Forth engine is @code{dodoes}, i.e.,
14128: the chunk of code executed by every word defined by a
14129: @code{CREATE}...@code{DOES>} pair. The main problem here is: How to find
14130: the Forth code to be executed, i.e. the code after the
1.26 crook 14131: @code{DOES>} (the @code{DOES>}-code)? There are two solutions:
1.1 anton 14132:
1.21 crook 14133: In fig-Forth the code field points directly to the @code{dodoes} and the
1.45 crook 14134: @code{DOES>}-code address is stored in the cell after the code address (i.e. at
1.29 crook 14135: @code{@i{CFA} cell+}). It may seem that this solution is illegal in
1.1 anton 14136: the Forth-79 and all later standards, because in fig-Forth this address
14137: lies in the body (which is illegal in these standards). However, by
14138: making the code field larger for all words this solution becomes legal
14139: again. We use this approach for the indirect threaded version and for
14140: direct threading on some machines. Leaving a cell unused in most words
14141: is a bit wasteful, but on the machines we are targeting this is hardly a
14142: problem. The other reason for having a code field size of two cells is
14143: to avoid having different image files for direct and indirect threaded
14144: systems (direct threaded systems require two-cell code fields on many
14145: machines).
14146:
1.26 crook 14147: @cindex @code{DOES>}-handler
1.1 anton 14148: The other approach is that the code field points or jumps to the cell
1.26 crook 14149: after @code{DOES>}. In this variant there is a jump to @code{dodoes} at
14150: this address (the @code{DOES>}-handler). @code{dodoes} can then get the
14151: @code{DOES>}-code address by computing the code address, i.e., the address of
1.45 crook 14152: the jump to @code{dodoes}, and add the length of that jump field. A variant of
1.1 anton 14153: this is to have a call to @code{dodoes} after the @code{DOES>}; then the
14154: return address (which can be found in the return register on RISCs) is
1.26 crook 14155: the @code{DOES>}-code address. Since the two cells available in the code field
1.1 anton 14156: are used up by the jump to the code address in direct threading on many
14157: architectures, we use this approach for direct threading on these
14158: architectures. We did not want to add another cell to the code field.
14159:
14160: @node Primitives, Performance, Threading, Engine
14161: @section Primitives
14162: @cindex primitives, implementation
14163: @cindex virtual machine instructions, implementation
14164:
14165: @menu
14166: * Automatic Generation::
14167: * TOS Optimization::
14168: * Produced code::
14169: @end menu
14170:
14171: @node Automatic Generation, TOS Optimization, Primitives, Primitives
14172: @subsection Automatic Generation
14173: @cindex primitives, automatic generation
14174:
14175: @cindex @file{prims2x.fs}
14176: Since the primitives are implemented in a portable language, there is no
14177: longer any need to minimize the number of primitives. On the contrary,
14178: having many primitives has an advantage: speed. In order to reduce the
14179: number of errors in primitives and to make programming them easier, we
14180: provide a tool, the primitive generator (@file{prims2x.fs}), that
14181: automatically generates most (and sometimes all) of the C code for a
14182: primitive from the stack effect notation. The source for a primitive
14183: has the following form:
14184:
14185: @cindex primitive source format
14186: @format
1.58 anton 14187: @i{Forth-name} ( @i{stack-effect} ) @i{category} [@i{pronounc.}]
1.29 crook 14188: [@code{""}@i{glossary entry}@code{""}]
14189: @i{C code}
1.1 anton 14190: [@code{:}
1.29 crook 14191: @i{Forth code}]
1.1 anton 14192: @end format
14193:
14194: The items in brackets are optional. The category and glossary fields
14195: are there for generating the documentation, the Forth code is there
14196: for manual implementations on machines without GNU C. E.g., the source
14197: for the primitive @code{+} is:
14198: @example
1.58 anton 14199: + ( n1 n2 -- n ) core plus
1.1 anton 14200: n = n1+n2;
14201: @end example
14202:
14203: This looks like a specification, but in fact @code{n = n1+n2} is C
14204: code. Our primitive generation tool extracts a lot of information from
14205: the stack effect notations@footnote{We use a one-stack notation, even
14206: though we have separate data and floating-point stacks; The separate
14207: notation can be generated easily from the unified notation.}: The number
14208: of items popped from and pushed on the stack, their type, and by what
14209: name they are referred to in the C code. It then generates a C code
14210: prelude and postlude for each primitive. The final C code for @code{+}
14211: looks like this:
14212:
14213: @example
1.46 pazsan 14214: I_plus: /* + ( n1 n2 -- n ) */ /* label, stack effect */
1.1 anton 14215: /* */ /* documentation */
1.81 ! anton 14216: NAME("+") /* debugging output (with -DDEBUG) */
1.1 anton 14217: @{
14218: DEF_CA /* definition of variable ca (indirect threading) */
14219: Cell n1; /* definitions of variables */
14220: Cell n2;
14221: Cell n;
1.81 ! anton 14222: NEXT_P0; /* NEXT part 0 */
1.1 anton 14223: n1 = (Cell) sp[1]; /* input */
14224: n2 = (Cell) TOS;
14225: sp += 1; /* stack adjustment */
14226: @{
14227: n = n1+n2; /* C code taken from the source */
14228: @}
14229: NEXT_P1; /* NEXT part 1 */
14230: TOS = (Cell)n; /* output */
14231: NEXT_P2; /* NEXT part 2 */
14232: @}
14233: @end example
14234:
14235: This looks long and inefficient, but the GNU C compiler optimizes quite
14236: well and produces optimal code for @code{+} on, e.g., the R3000 and the
14237: HP RISC machines: Defining the @code{n}s does not produce any code, and
14238: using them as intermediate storage also adds no cost.
14239:
1.26 crook 14240: There are also other optimizations that are not illustrated by this
14241: example: assignments between simple variables are usually for free (copy
1.1 anton 14242: propagation). If one of the stack items is not used by the primitive
14243: (e.g. in @code{drop}), the compiler eliminates the load from the stack
14244: (dead code elimination). On the other hand, there are some things that
14245: the compiler does not do, therefore they are performed by
14246: @file{prims2x.fs}: The compiler does not optimize code away that stores
14247: a stack item to the place where it just came from (e.g., @code{over}).
14248:
14249: While programming a primitive is usually easy, there are a few cases
14250: where the programmer has to take the actions of the generator into
14251: account, most notably @code{?dup}, but also words that do not (always)
1.26 crook 14252: fall through to @code{NEXT}.
1.1 anton 14253:
14254: @node TOS Optimization, Produced code, Automatic Generation, Primitives
14255: @subsection TOS Optimization
14256: @cindex TOS optimization for primitives
14257: @cindex primitives, keeping the TOS in a register
14258:
14259: An important optimization for stack machine emulators, e.g., Forth
14260: engines, is keeping one or more of the top stack items in
1.29 crook 14261: registers. If a word has the stack effect @i{in1}...@i{inx} @code{--}
14262: @i{out1}...@i{outy}, keeping the top @i{n} items in registers
1.1 anton 14263: @itemize @bullet
14264: @item
1.29 crook 14265: is better than keeping @i{n-1} items, if @i{x>=n} and @i{y>=n},
1.1 anton 14266: due to fewer loads from and stores to the stack.
1.29 crook 14267: @item is slower than keeping @i{n-1} items, if @i{x<>y} and @i{x<n} and
14268: @i{y<n}, due to additional moves between registers.
1.1 anton 14269: @end itemize
14270:
14271: @cindex -DUSE_TOS
14272: @cindex -DUSE_NO_TOS
14273: In particular, keeping one item in a register is never a disadvantage,
14274: if there are enough registers. Keeping two items in registers is a
14275: disadvantage for frequent words like @code{?branch}, constants,
14276: variables, literals and @code{i}. Therefore our generator only produces
14277: code that keeps zero or one items in registers. The generated C code
14278: covers both cases; the selection between these alternatives is made at
14279: C-compile time using the switch @code{-DUSE_TOS}. @code{TOS} in the C
14280: code for @code{+} is just a simple variable name in the one-item case,
14281: otherwise it is a macro that expands into @code{sp[0]}. Note that the
14282: GNU C compiler tries to keep simple variables like @code{TOS} in
14283: registers, and it usually succeeds, if there are enough registers.
14284:
14285: @cindex -DUSE_FTOS
14286: @cindex -DUSE_NO_FTOS
14287: The primitive generator performs the TOS optimization for the
14288: floating-point stack, too (@code{-DUSE_FTOS}). For floating-point
14289: operations the benefit of this optimization is even larger:
14290: floating-point operations take quite long on most processors, but can be
14291: performed in parallel with other operations as long as their results are
14292: not used. If the FP-TOS is kept in a register, this works. If
14293: it is kept on the stack, i.e., in memory, the store into memory has to
14294: wait for the result of the floating-point operation, lengthening the
14295: execution time of the primitive considerably.
14296:
14297: The TOS optimization makes the automatic generation of primitives a
14298: bit more complicated. Just replacing all occurrences of @code{sp[0]} by
14299: @code{TOS} is not sufficient. There are some special cases to
14300: consider:
14301: @itemize @bullet
14302: @item In the case of @code{dup ( w -- w w )} the generator must not
14303: eliminate the store to the original location of the item on the stack,
14304: if the TOS optimization is turned on.
14305: @item Primitives with stack effects of the form @code{--}
1.29 crook 14306: @i{out1}...@i{outy} must store the TOS to the stack at the start.
14307: Likewise, primitives with the stack effect @i{in1}...@i{inx} @code{--}
1.1 anton 14308: must load the TOS from the stack at the end. But for the null stack
14309: effect @code{--} no stores or loads should be generated.
14310: @end itemize
14311:
14312: @node Produced code, , TOS Optimization, Primitives
14313: @subsection Produced code
14314: @cindex primitives, assembly code listing
14315:
14316: @cindex @file{engine.s}
14317: To see what assembly code is produced for the primitives on your machine
14318: with your compiler and your flag settings, type @code{make engine.s} and
1.81 ! anton 14319: look at the resulting file @file{engine.s}. Alternatively, you can also
! 14320: disassemble the code of primitives with @code{see} on some architectures.
1.1 anton 14321:
14322: @node Performance, , Primitives, Engine
14323: @section Performance
14324: @cindex performance of some Forth interpreters
14325: @cindex engine performance
14326: @cindex benchmarking Forth systems
14327: @cindex Gforth performance
14328:
14329: On RISCs the Gforth engine is very close to optimal; i.e., it is usually
14330: impossible to write a significantly faster engine.
14331:
14332: On register-starved machines like the 386 architecture processors
14333: improvements are possible, because @code{gcc} does not utilize the
14334: registers as well as a human, even with explicit register declarations;
14335: e.g., Bernd Beuster wrote a Forth system fragment in assembly language
14336: and hand-tuned it for the 486; this system is 1.19 times faster on the
14337: Sieve benchmark on a 486DX2/66 than Gforth compiled with
1.40 anton 14338: @code{gcc-2.6.3} with @code{-DFORCE_REG}. The situation has improved
14339: with gcc-2.95 and gforth-0.4.9; now the most important virtual machine
14340: registers fit in real registers (and we can even afford to use the TOS
14341: optimization), resulting in a speedup of 1.14 on the sieve over the
14342: earlier results.
1.1 anton 14343:
14344: @cindex Win32Forth performance
14345: @cindex NT Forth performance
14346: @cindex eforth performance
14347: @cindex ThisForth performance
14348: @cindex PFE performance
14349: @cindex TILE performance
1.81 ! anton 14350: The potential advantage of assembly language implementations is not
! 14351: necessarily realized in complete Forth systems: We compared Gforth-0.4.9
! 14352: (direct threaded, compiled with @code{gcc-2.95.1} and
! 14353: @code{-DFORCE_REG}) with Win32Forth 1.2093 (newer versions are
! 14354: reportedly much faster), LMI's NT Forth (Beta, May 1994) and Eforth
! 14355: (with and without peephole (aka pinhole) optimization of the threaded
! 14356: code); all these systems were written in assembly language. We also
! 14357: compared Gforth with three systems written in C: PFE-0.9.14 (compiled
! 14358: with @code{gcc-2.6.3} with the default configuration for Linux:
! 14359: @code{-O2 -fomit-frame-pointer -DUSE_REGS -DUNROLL_NEXT}), ThisForth
! 14360: Beta (compiled with @code{gcc-2.6.3 -O3 -fomit-frame-pointer}; ThisForth
! 14361: employs peephole optimization of the threaded code) and TILE (compiled
! 14362: with @code{make opt}). We benchmarked Gforth, PFE, ThisForth and TILE on
! 14363: a 486DX2/66 under Linux. Kenneth O'Heskin kindly provided the results
! 14364: for Win32Forth and NT Forth on a 486DX2/66 with similar memory
! 14365: performance under Windows NT. Marcel Hendrix ported Eforth to Linux,
! 14366: then extended it to run the benchmarks, added the peephole optimizer,
! 14367: ran the benchmarks and reported the results.
1.40 anton 14368:
1.1 anton 14369: We used four small benchmarks: the ubiquitous Sieve; bubble-sorting and
14370: matrix multiplication come from the Stanford integer benchmarks and have
14371: been translated into Forth by Martin Fraeman; we used the versions
14372: included in the TILE Forth package, but with bigger data set sizes; and
14373: a recursive Fibonacci number computation for benchmarking calling
14374: performance. The following table shows the time taken for the benchmarks
14375: scaled by the time taken by Gforth (in other words, it shows the speedup
14376: factor that Gforth achieved over the other systems).
14377:
14378: @example
1.40 anton 14379: relative Win32- NT eforth This-
1.1 anton 14380: time Gforth Forth Forth eforth +opt PFE Forth TILE
1.81 ! anton 14381: sieve 1.00 1.60 1.32 1.60 0.98 1.82 3.67 9.91
! 14382: bubble 1.00 1.55 1.66 1.75 1.04 1.78 4.58
! 14383: matmul 1.00 1.71 1.57 1.69 0.86 1.83 4.74
! 14384: fib 1.00 1.76 1.54 1.41 1.00 2.01 3.45 4.96
1.1 anton 14385: @end example
14386:
1.26 crook 14387: You may be quite surprised by the good performance of Gforth when
14388: compared with systems written in assembly language. One important reason
14389: for the disappointing performance of these other systems is probably
14390: that they are not written optimally for the 486 (e.g., they use the
14391: @code{lods} instruction). In addition, Win32Forth uses a comfortable,
14392: but costly method for relocating the Forth image: like @code{cforth}, it
14393: computes the actual addresses at run time, resulting in two address
14394: computations per @code{NEXT} (@pxref{Image File Background}).
14395:
1.40 anton 14396: Only Eforth with the peephole optimizer performs comparable to
14397: Gforth. The speedups achieved with peephole optimization of threaded
14398: code are quite remarkable. Adding a peephole optimizer to Gforth should
14399: cause similar speedups.
1.1 anton 14400:
14401: The speedup of Gforth over PFE, ThisForth and TILE can be easily
14402: explained with the self-imposed restriction of the latter systems to
14403: standard C, which makes efficient threading impossible (however, the
1.4 anton 14404: measured implementation of PFE uses a GNU C extension: @pxref{Global Reg
1.1 anton 14405: Vars, , Defining Global Register Variables, gcc.info, GNU C Manual}).
14406: Moreover, current C compilers have a hard time optimizing other aspects
14407: of the ThisForth and the TILE source.
14408:
1.26 crook 14409: The performance of Gforth on 386 architecture processors varies widely
14410: with the version of @code{gcc} used. E.g., @code{gcc-2.5.8} failed to
14411: allocate any of the virtual machine registers into real machine
14412: registers by itself and would not work correctly with explicit register
1.40 anton 14413: declarations, giving a 1.5 times slower engine (on a 486DX2/66 running
1.26 crook 14414: the Sieve) than the one measured above.
1.1 anton 14415:
1.26 crook 14416: Note that there have been several releases of Win32Forth since the
14417: release presented here, so the results presented above may have little
1.40 anton 14418: predictive value for the performance of Win32Forth today (results for
14419: the current release on an i486DX2/66 are welcome).
1.1 anton 14420:
14421: @cindex @file{Benchres}
1.66 anton 14422: In
14423: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl&maierhofer95.ps.gz,
14424: Translating Forth to Efficient C}} by M. Anton Ertl and Martin
1.1 anton 14425: Maierhofer (presented at EuroForth '95), an indirect threaded version of
1.66 anton 14426: Gforth is compared with Win32Forth, NT Forth, PFE, ThisForth, and
14427: several native code systems; that version of Gforth is slower on a 486
14428: than the direct threaded version used here. You can find a newer version
14429: of these measurements at
1.47 crook 14430: @uref{http://www.complang.tuwien.ac.at/forth/performance.html}. You can
1.1 anton 14431: find numbers for Gforth on various machines in @file{Benchres}.
14432:
1.26 crook 14433: @c ******************************************************************
1.13 pazsan 14434: @node Binding to System Library, Cross Compiler, Engine, Top
1.14 pazsan 14435: @chapter Binding to System Library
1.13 pazsan 14436:
14437: @node Cross Compiler, Bugs, Binding to System Library, Top
1.14 pazsan 14438: @chapter Cross Compiler
1.47 crook 14439: @cindex @file{cross.fs}
14440: @cindex cross-compiler
14441: @cindex metacompiler
14442: @cindex target compiler
1.13 pazsan 14443:
1.46 pazsan 14444: The cross compiler is used to bootstrap a Forth kernel. Since Gforth is
14445: mostly written in Forth, including crucial parts like the outer
14446: interpreter and compiler, it needs compiled Forth code to get
14447: started. The cross compiler allows to create new images for other
14448: architectures, even running under another Forth system.
1.13 pazsan 14449:
14450: @menu
1.67 anton 14451: * Using the Cross Compiler::
14452: * How the Cross Compiler Works::
1.13 pazsan 14453: @end menu
14454:
1.21 crook 14455: @node Using the Cross Compiler, How the Cross Compiler Works, Cross Compiler, Cross Compiler
1.14 pazsan 14456: @section Using the Cross Compiler
1.46 pazsan 14457:
14458: The cross compiler uses a language that resembles Forth, but isn't. The
14459: main difference is that you can execute Forth code after definition,
14460: while you usually can't execute the code compiled by cross, because the
14461: code you are compiling is typically for a different computer than the
14462: one you are compiling on.
14463:
1.81 ! anton 14464: @c anton: This chapter is somewhat different from waht I would expect: I
! 14465: @c would expect an explanation of the cross language and how to create an
! 14466: @c application image with it. The section explains some aspects of
! 14467: @c creating a Gforth kernel.
! 14468:
1.46 pazsan 14469: The Makefile is already set up to allow you to create kernels for new
14470: architectures with a simple make command. The generic kernels using the
14471: GCC compiled virtual machine are created in the normal build process
14472: with @code{make}. To create a embedded Gforth executable for e.g. the
14473: 8086 processor (running on a DOS machine), type
14474:
14475: @example
14476: make kernl-8086.fi
14477: @end example
14478:
14479: This will use the machine description from the @file{arch/8086}
14480: directory to create a new kernel. A machine file may look like that:
14481:
14482: @example
14483: \ Parameter for target systems 06oct92py
14484:
14485: 4 Constant cell \ cell size in bytes
14486: 2 Constant cell<< \ cell shift to bytes
14487: 5 Constant cell>bit \ cell shift to bits
14488: 8 Constant bits/char \ bits per character
14489: 8 Constant bits/byte \ bits per byte [default: 8]
14490: 8 Constant float \ bytes per float
14491: 8 Constant /maxalign \ maximum alignment in bytes
14492: false Constant bigendian \ byte order
14493: ( true=big, false=little )
14494:
14495: include machpc.fs \ feature list
14496: @end example
14497:
14498: This part is obligatory for the cross compiler itself, the feature list
14499: is used by the kernel to conditionally compile some features in and out,
14500: depending on whether the target supports these features.
14501:
14502: There are some optional features, if you define your own primitives,
14503: have an assembler, or need special, nonstandard preparation to make the
1.81 ! anton 14504: boot process work. @code{asm-include} includes an assembler,
1.46 pazsan 14505: @code{prims-include} includes primitives, and @code{>boot} prepares for
14506: booting.
14507:
14508: @example
14509: : asm-include ." Include assembler" cr
14510: s" arch/8086/asm.fs" included ;
14511:
14512: : prims-include ." Include primitives" cr
14513: s" arch/8086/prim.fs" included ;
14514:
14515: : >boot ." Prepare booting" cr
14516: s" ' boot >body into-forth 1+ !" evaluate ;
14517: @end example
14518:
14519: These words are used as sort of macro during the cross compilation in
1.81 ! anton 14520: the file @file{kernel/main.fs}. Instead of using these macros, it would
1.46 pazsan 14521: be possible --- but more complicated --- to write a new kernel project
14522: file, too.
14523:
14524: @file{kernel/main.fs} expects the machine description file name on the
14525: stack; the cross compiler itself (@file{cross.fs}) assumes that either
14526: @code{mach-file} leaves a counted string on the stack, or
14527: @code{machine-file} leaves an address, count pair of the filename on the
14528: stack.
14529:
14530: The feature list is typically controlled using @code{SetValue}, generic
14531: files that are used by several projects can use @code{DefaultValue}
14532: instead. Both functions work like @code{Value}, when the value isn't
14533: defined, but @code{SetValue} works like @code{to} if the value is
14534: defined, and @code{DefaultValue} doesn't set anything, if the value is
14535: defined.
14536:
14537: @example
14538: \ generic mach file for pc gforth 03sep97jaw
14539:
14540: true DefaultValue NIL \ relocating
14541:
14542: >ENVIRON
14543:
14544: true DefaultValue file \ controls the presence of the
14545: \ file access wordset
14546: true DefaultValue OS \ flag to indicate a operating system
14547:
14548: true DefaultValue prims \ true: primitives are c-code
14549:
14550: true DefaultValue floating \ floating point wordset is present
14551:
14552: true DefaultValue glocals \ gforth locals are present
14553: \ will be loaded
14554: true DefaultValue dcomps \ double number comparisons
14555:
14556: true DefaultValue hash \ hashing primitives are loaded/present
14557:
14558: true DefaultValue xconds \ used together with glocals,
14559: \ special conditionals supporting gforths'
14560: \ local variables
14561: true DefaultValue header \ save a header information
14562:
14563: true DefaultValue backtrace \ enables backtrace code
14564:
14565: false DefaultValue ec
14566: false DefaultValue crlf
14567:
14568: cell 2 = [IF] &32 [ELSE] &256 [THEN] KB DefaultValue kernel-size
14569:
14570: &16 KB DefaultValue stack-size
14571: &15 KB &512 + DefaultValue fstack-size
14572: &15 KB DefaultValue rstack-size
14573: &14 KB &512 + DefaultValue lstack-size
14574: @end example
1.13 pazsan 14575:
1.48 anton 14576: @node How the Cross Compiler Works, , Using the Cross Compiler, Cross Compiler
1.14 pazsan 14577: @section How the Cross Compiler Works
1.13 pazsan 14578:
14579: @node Bugs, Origin, Cross Compiler, Top
1.21 crook 14580: @appendix Bugs
1.1 anton 14581: @cindex bug reporting
14582:
1.21 crook 14583: Known bugs are described in the file @file{BUGS} in the Gforth distribution.
1.1 anton 14584:
14585: If you find a bug, please send a bug report to
1.33 anton 14586: @email{bug-gforth@@gnu.org}. A bug report should include this
1.21 crook 14587: information:
14588:
14589: @itemize @bullet
14590: @item
1.81 ! anton 14591: A program (or a sequence of keyboard commands) that reproduces the bug.
! 14592: @item
! 14593: A description of what you think constitutes the buggy behaviour.
! 14594: @item
1.21 crook 14595: The Gforth version used (it is announced at the start of an
14596: interactive Gforth session).
14597: @item
14598: The machine and operating system (on Unix
14599: systems @code{uname -a} will report this information).
14600: @item
1.81 ! anton 14601: The installation options (you can find the configure options at the
! 14602: start of @file{config.status}) and configuration (@code{configure}
! 14603: output or @file{config.cache}).
1.21 crook 14604: @item
14605: A complete list of changes (if any) you (or your installer) have made to the
14606: Gforth sources.
14607: @end itemize
1.1 anton 14608:
14609: For a thorough guide on reporting bugs read @ref{Bug Reporting, , How
14610: to Report Bugs, gcc.info, GNU C Manual}.
14611:
14612:
1.21 crook 14613: @node Origin, Forth-related information, Bugs, Top
14614: @appendix Authors and Ancestors of Gforth
1.1 anton 14615:
14616: @section Authors and Contributors
14617: @cindex authors of Gforth
14618: @cindex contributors to Gforth
14619:
14620: The Gforth project was started in mid-1992 by Bernd Paysan and Anton
1.81 ! anton 14621: Ertl. The third major author was Jens Wilke. Neal Crook contributed a
! 14622: lot to the manual. Assemblers and disassemblers were contributed by
! 14623: Andrew McKewan, Christian Pirker, and Bernd Thallner. Lennart Benschop
! 14624: (who was one of Gforth's first users, in mid-1993) and Stuart Ramsden
! 14625: inspired us with their continuous feedback. Lennart Benshop contributed
1.1 anton 14626: @file{glosgen.fs}, while Stuart Ramsden has been working on automatic
14627: support for calling C libraries. Helpful comments also came from Paul
14628: Kleinrubatscher, Christian Pirker, Dirk Zoller, Marcel Hendrix, John
1.58 anton 14629: Wavrik, Barrie Stott, Marc de Groot, Jorge Acerada, Bruce Hoyt, and
14630: Robert Epprecht. Since the release of Gforth-0.2.1 there were also
14631: helpful comments from many others; thank you all, sorry for not listing
14632: you here (but digging through my mailbox to extract your names is on my
1.81 ! anton 14633: to-do list).
1.1 anton 14634:
14635: Gforth also owes a lot to the authors of the tools we used (GCC, CVS,
14636: and autoconf, among others), and to the creators of the Internet: Gforth
1.21 crook 14637: was developed across the Internet, and its authors did not meet
1.20 pazsan 14638: physically for the first 4 years of development.
1.1 anton 14639:
14640: @section Pedigree
1.26 crook 14641: @cindex pedigree of Gforth
1.1 anton 14642:
1.81 ! anton 14643: Gforth descends from bigFORTH (1993) and fig-Forth. Of course, a
! 14644: significant part of the design of Gforth was prescribed by ANS Forth.
1.1 anton 14645:
1.20 pazsan 14646: Bernd Paysan wrote bigFORTH, a descendent from TurboForth, an unreleased
1.1 anton 14647: 32 bit native code version of VolksForth for the Atari ST, written
14648: mostly by Dietrich Weineck.
14649:
1.81 ! anton 14650: VolksForth was written by Klaus Schleisiek, Bernd Pennemann, Georg
! 14651: Rehfeld and Dietrich Weineck for the C64 (called UltraForth there) in
! 14652: the mid-80s and ported to the Atari ST in 1986. It descends from F83.
1.1 anton 14653:
14654: Henry Laxen and Mike Perry wrote F83 as a model implementation of the
14655: Forth-83 standard. !! Pedigree? When?
14656:
14657: A team led by Bill Ragsdale implemented fig-Forth on many processors in
14658: 1979. Robert Selzer and Bill Ragsdale developed the original
14659: implementation of fig-Forth for the 6502 based on microForth.
14660:
14661: The principal architect of microForth was Dean Sanderson. microForth was
14662: FORTH, Inc.'s first off-the-shelf product. It was developed in 1976 for
14663: the 1802, and subsequently implemented on the 8080, the 6800 and the
14664: Z80.
14665:
14666: All earlier Forth systems were custom-made, usually by Charles Moore,
14667: who discovered (as he puts it) Forth during the late 60s. The first full
14668: Forth existed in 1971.
14669:
1.81 ! anton 14670: A part of the information in this section comes from
! 14671: @cite{@uref{http://www.forth.com/Content/History/History1.htm,The
! 14672: Evolution of Forth}} by Elizabeth D. Rather, Donald R. Colburn and
! 14673: Charles H. Moore, presented at the HOPL-II conference and preprinted in
! 14674: SIGPLAN Notices 28(3), 1993. You can find more historical and
! 14675: genealogical information about Forth there.
1.1 anton 14676:
1.81 ! anton 14677: @c ------------------------------------------------------------------
1.21 crook 14678: @node Forth-related information, Word Index, Origin, Top
14679: @appendix Other Forth-related information
14680: @cindex Forth-related information
14681:
1.81 ! anton 14682: @c anton: I threw most of this stuff out, because it can be found through
! 14683: @c the FAQ and the FAQ is more likely to be up-to-date.
1.21 crook 14684:
14685: @cindex comp.lang.forth
14686: @cindex frequently asked questions
1.81 ! anton 14687: There is an active news group (comp.lang.forth) discussing Forth
! 14688: (including Gforth) and Forth-related issues. Its
! 14689: @uref{http://www.complang.tuwien.ac.at/forth/faq/faq-general-2.html,FAQs}
! 14690: (frequently asked questions and their answers) contains a lot of
! 14691: information on Forth. You should read it before posting to
! 14692: comp.lang.forth.
1.21 crook 14693:
1.81 ! anton 14694: The ANS Forth standard is most usable in its
! 14695: @uref{http://www.taygeta.com/forth/dpans.html, HTML form}.
1.21 crook 14696:
1.81 ! anton 14697: @c ------------------------------------------------------------------
! 14698: @node Word Index, Concept Index, Forth-related information, Top
1.1 anton 14699: @unnumbered Word Index
14700:
1.26 crook 14701: This index is a list of Forth words that have ``glossary'' entries
14702: within this manual. Each word is listed with its stack effect and
14703: wordset.
1.1 anton 14704:
14705: @printindex fn
14706:
1.81 ! anton 14707: @c anton: the name index seems superfluous given the word and concept indices.
! 14708:
! 14709: @c @node Name Index, Concept Index, Word Index, Top
! 14710: @c @unnumbered Name Index
1.41 anton 14711:
1.81 ! anton 14712: @c This index is a list of Forth words that have ``glossary'' entries
! 14713: @c within this manual.
1.41 anton 14714:
1.81 ! anton 14715: @c @printindex ky
1.41 anton 14716:
1.81 ! anton 14717: @node Concept Index, , Word Index, Top
1.1 anton 14718: @unnumbered Concept and Word Index
14719:
1.26 crook 14720: Not all entries listed in this index are present verbatim in the
14721: text. This index also duplicates, in abbreviated form, all of the words
14722: listed in the Word Index (only the names are listed for the words here).
1.1 anton 14723:
14724: @printindex cp
14725:
14726: @contents
14727: @bye
1.81 ! anton 14728:
! 14729:
1.1 anton 14730:
FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>