Annotation of gforth/gforth.ds, revision 1.24

1.1       anton       1: \input texinfo   @c -*-texinfo-*-
                      2: @comment The source is gforth.ds, from which gforth.texi is generated
                      3: @comment %**start of header (This is for running Texinfo on a region.)
1.4       anton       4: @setfilename
1.17      anton       5: @settitle Gforth Manual
1.4       anton       6: @comment @setchapternewpage odd
1.1       anton       7: @comment %**end of header (This is for running Texinfo on a region.)
                      9: @ifinfo
1.17      anton      10: This file documents Gforth 0.1
1.1       anton      11: 
1.21      anton      12: Copyright @copyright{} 1995 Free Software Foundation, Inc.
1.1       anton      13: 
                     14:      Permission is granted to make and distribute verbatim copies of
                     15:      this manual provided the copyright notice and this permission notice
                     16:      are preserved on all copies.
1.4       anton      18: @ignore
1.1       anton      19:      Permission is granted to process this file through TeX and print the
                     20:      results, provided the printed document carries a copying permission
                     21:      notice identical to this one except for the removal of this paragraph
                     22:      (this paragraph not being relevant to the printed manual).
1.4       anton      24: @end ignore
1.1       anton      25:      Permission is granted to copy and distribute modified versions of this
                     26:      manual under the conditions for verbatim copying, provided also that the
                     27:      sections entitled "Distribution" and "General Public License" are
                     28:      included exactly as in the original, and provided that the entire
                     29:      resulting derived work is distributed under the terms of a permission
                     30:      notice identical to this one.
                     32:      Permission is granted to copy and distribute translations of this manual
                     33:      into another language, under the above conditions for modified versions,
                     34:      except that the sections entitled "Distribution" and "General Public
                     35:      License" may be included in a translation approved by the author instead
                     36:      of in the original English.
                     37: @end ifinfo
1.24    ! anton      39: @finalout
1.1       anton      40: @titlepage
                     41: @sp 10
1.17      anton      42: @center @titlefont{Gforth Manual}
1.1       anton      43: @sp 2
1.17      anton      44: @center for version 0.1
1.1       anton      45: @sp 2
                     46: @center Anton Ertl
1.17      anton      47: @sp 3
                     48: @center This manual is under construction
1.1       anton      49: 
                     50: @comment  The following two commands start the copyright page.
                     51: @page
                     52: @vskip 0pt plus 1filll
1.21      anton      53: Copyright @copyright{} 1995 Free Software Foundation, Inc.
1.1       anton      54: 
                     55: @comment !! Published by ... or You can get a copy of this manual ...
                     57:      Permission is granted to make and distribute verbatim copies of
                     58:      this manual provided the copyright notice and this permission notice
                     59:      are preserved on all copies.
                     61:      Permission is granted to copy and distribute modified versions of this
                     62:      manual under the conditions for verbatim copying, provided also that the
                     63:      sections entitled "Distribution" and "General Public License" are
                     64:      included exactly as in the original, and provided that the entire
                     65:      resulting derived work is distributed under the terms of a permission
                     66:      notice identical to this one.
                     68:      Permission is granted to copy and distribute translations of this manual
                     69:      into another language, under the above conditions for modified versions,
                     70:      except that the sections entitled "Distribution" and "General Public
                     71:      License" may be included in a translation approved by the author instead
                     72:      of in the original English.
                     73: @end titlepage
                     76: @node Top, License, (dir), (dir)
                     77: @ifinfo
1.17      anton      78: Gforth is a free implementation of ANS Forth available on many
1.21      anton      79: personal machines. This manual corresponds to version 0.1.
1.1       anton      80: @end ifinfo
                     82: @menu
1.4       anton      83: * License::                     
1.17      anton      84: * Goals::                       About the Gforth Project
1.4       anton      85: * Other Books::                 Things you might want to read
1.17      anton      86: * Invocation::                  Starting Gforth
                     87: * Words::                       Forth words available in Gforth
1.4       anton      88: * ANS conformance::             Implementation-defined options etc.
1.17      anton      89: * Model::                       The abstract machine of Gforth
                     90: * Emacs and Gforth::            The Gforth Mode
1.4       anton      91: * Internals::                   Implementation details
                     92: * Bugs::                        How to report them
1.17      anton      93: * Pedigree::                    Ancestors of Gforth
1.4       anton      94: * Word Index::                  An item for each Forth word
                     95: * Node Index::                  An item for each node
1.1       anton      96: @end menu
                     98: @node License, Goals, Top, Top
1.20      pazsan     99: @unnumbered GNU GENERAL PUBLIC LICENSE
                    100: @center Version 2, June 1991
                    102: @display
                    103: Copyright @copyright{} 1989, 1991 Free Software Foundation, Inc.
                    104: 675 Mass Ave, Cambridge, MA 02139, USA
                    106: Everyone is permitted to copy and distribute verbatim copies
                    107: of this license document, but changing it is not allowed.
                    108: @end display
                    110: @unnumberedsec Preamble
                    112:   The licenses for most software are designed to take away your
                    113: freedom to share and change it.  By contrast, the GNU General Public
                    114: License is intended to guarantee your freedom to share and change free
                    115: software---to make sure the software is free for all its users.  This
                    116: General Public License applies to most of the Free Software
                    117: Foundation's software and to any other program whose authors commit to
                    118: using it.  (Some other Free Software Foundation software is covered by
                    119: the GNU Library General Public License instead.)  You can apply it to
                    120: your programs, too.
                    122:   When we speak of free software, we are referring to freedom, not
                    123: price.  Our General Public Licenses are designed to make sure that you
                    124: have the freedom to distribute copies of free software (and charge for
                    125: this service if you wish), that you receive source code or can get it
                    126: if you want it, that you can change the software or use pieces of it
                    127: in new free programs; and that you know you can do these things.
                    129:   To protect your rights, we need to make restrictions that forbid
                    130: anyone to deny you these rights or to ask you to surrender the rights.
                    131: These restrictions translate to certain responsibilities for you if you
                    132: distribute copies of the software, or if you modify it.
                    134:   For example, if you distribute copies of such a program, whether
                    135: gratis or for a fee, you must give the recipients all the rights that
                    136: you have.  You must make sure that they, too, receive or can get the
                    137: source code.  And you must show them these terms so they know their
                    138: rights.
                    140:   We protect your rights with two steps: (1) copyright the software, and
                    141: (2) offer you this license which gives you legal permission to copy,
                    142: distribute and/or modify the software.
                    144:   Also, for each author's protection and ours, we want to make certain
                    145: that everyone understands that there is no warranty for this free
                    146: software.  If the software is modified by someone else and passed on, we
                    147: want its recipients to know that what they have is not the original, so
                    148: that any problems introduced by others will not reflect on the original
                    149: authors' reputations.
                    151:   Finally, any free program is threatened constantly by software
                    152: patents.  We wish to avoid the danger that redistributors of a free
                    153: program will individually obtain patent licenses, in effect making the
                    154: program proprietary.  To prevent this, we have made it clear that any
                    155: patent must be licensed for everyone's free use or not licensed at all.
                    157:   The precise terms and conditions for copying, distribution and
                    158: modification follow.
                    160: @iftex
                    162: @end iftex
                    163: @ifinfo
                    165: @end ifinfo
                    167: @enumerate 0
                    168: @item
                    169: This License applies to any program or other work which contains
                    170: a notice placed by the copyright holder saying it may be distributed
                    171: under the terms of this General Public License.  The ``Program'', below,
                    172: refers to any such program or work, and a ``work based on the Program''
                    173: means either the Program or any derivative work under copyright law:
                    174: that is to say, a work containing the Program or a portion of it,
                    175: either verbatim or with modifications and/or translated into another
                    176: language.  (Hereinafter, translation is included without limitation in
                    177: the term ``modification''.)  Each licensee is addressed as ``you''.
                    179: Activities other than copying, distribution and modification are not
                    180: covered by this License; they are outside its scope.  The act of
                    181: running the Program is not restricted, and the output from the Program
                    182: is covered only if its contents constitute a work based on the
                    183: Program (independent of having been made by running the Program).
                    184: Whether that is true depends on what the Program does.
                    186: @item
                    187: You may copy and distribute verbatim copies of the Program's
                    188: source code as you receive it, in any medium, provided that you
                    189: conspicuously and appropriately publish on each copy an appropriate
                    190: copyright notice and disclaimer of warranty; keep intact all the
                    191: notices that refer to this License and to the absence of any warranty;
                    192: and give any other recipients of the Program a copy of this License
                    193: along with the Program.
                    195: You may charge a fee for the physical act of transferring a copy, and
                    196: you may at your option offer warranty protection in exchange for a fee.
                    198: @item
                    199: You may modify your copy or copies of the Program or any portion
                    200: of it, thus forming a work based on the Program, and copy and
                    201: distribute such modifications or work under the terms of Section 1
                    202: above, provided that you also meet all of these conditions:
                    204: @enumerate a
                    205: @item
                    206: You must cause the modified files to carry prominent notices
                    207: stating that you changed the files and the date of any change.
                    209: @item
                    210: You must cause any work that you distribute or publish, that in
                    211: whole or in part contains or is derived from the Program or any
                    212: part thereof, to be licensed as a whole at no charge to all third
                    213: parties under the terms of this License.
                    215: @item
                    216: If the modified program normally reads commands interactively
                    217: when run, you must cause it, when started running for such
                    218: interactive use in the most ordinary way, to print or display an
                    219: announcement including an appropriate copyright notice and a
                    220: notice that there is no warranty (or else, saying that you provide
                    221: a warranty) and that users may redistribute the program under
                    222: these conditions, and telling the user how to view a copy of this
                    223: License.  (Exception: if the Program itself is interactive but
                    224: does not normally print such an announcement, your work based on
                    225: the Program is not required to print an announcement.)
                    226: @end enumerate
                    228: These requirements apply to the modified work as a whole.  If
                    229: identifiable sections of that work are not derived from the Program,
                    230: and can be reasonably considered independent and separate works in
                    231: themselves, then this License, and its terms, do not apply to those
                    232: sections when you distribute them as separate works.  But when you
                    233: distribute the same sections as part of a whole which is a work based
                    234: on the Program, the distribution of the whole must be on the terms of
                    235: this License, whose permissions for other licensees extend to the
                    236: entire whole, and thus to each and every part regardless of who wrote it.
                    238: Thus, it is not the intent of this section to claim rights or contest
                    239: your rights to work written entirely by you; rather, the intent is to
                    240: exercise the right to control the distribution of derivative or
                    241: collective works based on the Program.
                    243: In addition, mere aggregation of another work not based on the Program
                    244: with the Program (or with a work based on the Program) on a volume of
                    245: a storage or distribution medium does not bring the other work under
                    246: the scope of this License.
                    248: @item
                    249: You may copy and distribute the Program (or a work based on it,
                    250: under Section 2) in object code or executable form under the terms of
                    251: Sections 1 and 2 above provided that you also do one of the following:
                    253: @enumerate a
                    254: @item
                    255: Accompany it with the complete corresponding machine-readable
                    256: source code, which must be distributed under the terms of Sections
                    257: 1 and 2 above on a medium customarily used for software interchange; or,
                    259: @item
                    260: Accompany it with a written offer, valid for at least three
                    261: years, to give any third party, for a charge no more than your
                    262: cost of physically performing source distribution, a complete
                    263: machine-readable copy of the corresponding source code, to be
                    264: distributed under the terms of Sections 1 and 2 above on a medium
                    265: customarily used for software interchange; or,
                    267: @item
                    268: Accompany it with the information you received as to the offer
                    269: to distribute corresponding source code.  (This alternative is
                    270: allowed only for noncommercial distribution and only if you
                    271: received the program in object code or executable form with such
                    272: an offer, in accord with Subsection b above.)
                    273: @end enumerate
                    275: The source code for a work means the preferred form of the work for
                    276: making modifications to it.  For an executable work, complete source
                    277: code means all the source code for all modules it contains, plus any
                    278: associated interface definition files, plus the scripts used to
                    279: control compilation and installation of the executable.  However, as a
                    280: special exception, the source code distributed need not include
                    281: anything that is normally distributed (in either source or binary
                    282: form) with the major components (compiler, kernel, and so on) of the
                    283: operating system on which the executable runs, unless that component
                    284: itself accompanies the executable.
                    286: If distribution of executable or object code is made by offering
                    287: access to copy from a designated place, then offering equivalent
                    288: access to copy the source code from the same place counts as
                    289: distribution of the source code, even though third parties are not
                    290: compelled to copy the source along with the object code.
                    292: @item
                    293: You may not copy, modify, sublicense, or distribute the Program
                    294: except as expressly provided under this License.  Any attempt
                    295: otherwise to copy, modify, sublicense or distribute the Program is
                    296: void, and will automatically terminate your rights under this License.
                    297: However, parties who have received copies, or rights, from you under
                    298: this License will not have their licenses terminated so long as such
                    299: parties remain in full compliance.
                    301: @item
                    302: You are not required to accept this License, since you have not
                    303: signed it.  However, nothing else grants you permission to modify or
                    304: distribute the Program or its derivative works.  These actions are
                    305: prohibited by law if you do not accept this License.  Therefore, by
                    306: modifying or distributing the Program (or any work based on the
                    307: Program), you indicate your acceptance of this License to do so, and
                    308: all its terms and conditions for copying, distributing or modifying
                    309: the Program or works based on it.
                    311: @item
                    312: Each time you redistribute the Program (or any work based on the
                    313: Program), the recipient automatically receives a license from the
                    314: original licensor to copy, distribute or modify the Program subject to
                    315: these terms and conditions.  You may not impose any further
                    316: restrictions on the recipients' exercise of the rights granted herein.
                    317: You are not responsible for enforcing compliance by third parties to
                    318: this License.
                    320: @item
                    321: If, as a consequence of a court judgment or allegation of patent
                    322: infringement or for any other reason (not limited to patent issues),
                    323: conditions are imposed on you (whether by court order, agreement or
                    324: otherwise) that contradict the conditions of this License, they do not
                    325: excuse you from the conditions of this License.  If you cannot
                    326: distribute so as to satisfy simultaneously your obligations under this
                    327: License and any other pertinent obligations, then as a consequence you
                    328: may not distribute the Program at all.  For example, if a patent
                    329: license would not permit royalty-free redistribution of the Program by
                    330: all those who receive copies directly or indirectly through you, then
                    331: the only way you could satisfy both it and this License would be to
                    332: refrain entirely from distribution of the Program.
                    334: If any portion of this section is held invalid or unenforceable under
                    335: any particular circumstance, the balance of the section is intended to
                    336: apply and the section as a whole is intended to apply in other
                    337: circumstances.
                    339: It is not the purpose of this section to induce you to infringe any
                    340: patents or other property right claims or to contest validity of any
                    341: such claims; this section has the sole purpose of protecting the
                    342: integrity of the free software distribution system, which is
                    343: implemented by public license practices.  Many people have made
                    344: generous contributions to the wide range of software distributed
                    345: through that system in reliance on consistent application of that
                    346: system; it is up to the author/donor to decide if he or she is willing
                    347: to distribute software through any other system and a licensee cannot
                    348: impose that choice.
                    350: This section is intended to make thoroughly clear what is believed to
                    351: be a consequence of the rest of this License.
                    353: @item
                    354: If the distribution and/or use of the Program is restricted in
                    355: certain countries either by patents or by copyrighted interfaces, the
                    356: original copyright holder who places the Program under this License
                    357: may add an explicit geographical distribution limitation excluding
                    358: those countries, so that distribution is permitted only in or among
                    359: countries not thus excluded.  In such case, this License incorporates
                    360: the limitation as if written in the body of this License.
                    362: @item
                    363: The Free Software Foundation may publish revised and/or new versions
                    364: of the General Public License from time to time.  Such new versions will
                    365: be similar in spirit to the present version, but may differ in detail to
                    366: address new problems or concerns.
                    368: Each version is given a distinguishing version number.  If the Program
                    369: specifies a version number of this License which applies to it and ``any
                    370: later version'', you have the option of following the terms and conditions
                    371: either of that version or of any later version published by the Free
                    372: Software Foundation.  If the Program does not specify a version number of
                    373: this License, you may choose any version ever published by the Free Software
                    374: Foundation.
                    376: @item
                    377: If you wish to incorporate parts of the Program into other free
                    378: programs whose distribution conditions are different, write to the author
                    379: to ask for permission.  For software which is copyrighted by the Free
                    380: Software Foundation, write to the Free Software Foundation; we sometimes
                    381: make exceptions for this.  Our decision will be guided by the two goals
                    382: of preserving the free status of all derivatives of our free software and
                    383: of promoting the sharing and reuse of software generally.
                    385: @iftex
                    386: @heading NO WARRANTY
                    387: @end iftex
                    388: @ifinfo
                    389: @center NO WARRANTY
                    390: @end ifinfo
                    392: @item
                    401: REPAIR OR CORRECTION.
                    403: @item
                    412: POSSIBILITY OF SUCH DAMAGES.
                    413: @end enumerate
                    415: @iftex
                    416: @heading END OF TERMS AND CONDITIONS
                    417: @end iftex
                    418: @ifinfo
                    419: @center END OF TERMS AND CONDITIONS
                    420: @end ifinfo
                    422: @page
                    423: @unnumberedsec How to Apply These Terms to Your New Programs
                    425:   If you develop a new program, and you want it to be of the greatest
                    426: possible use to the public, the best way to achieve this is to make it
                    427: free software which everyone can redistribute and change under these terms.
                    429:   To do so, attach the following notices to the program.  It is safest
                    430: to attach them to the start of each source file to most effectively
                    431: convey the exclusion of warranty; and each file should have at least
                    432: the ``copyright'' line and a pointer to where the full notice is found.
                    434: @smallexample
                    435: @var{one line to give the program's name and a brief idea of what it does.}
                    436: Copyright (C) 19@var{yy}  @var{name of author}
                    438: This program is free software; you can redistribute it and/or modify 
                    439: it under the terms of the GNU General Public License as published by 
                    440: the Free Software Foundation; either version 2 of the License, or 
                    441: (at your option) any later version.
                    443: This program is distributed in the hope that it will be useful,
                    444: but WITHOUT ANY WARRANTY; without even the implied warranty of
                    446: GNU General Public License for more details.
                    448: You should have received a copy of the GNU General Public License
                    449: along with this program; if not, write to the Free Software
                    450: Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
                    451: @end smallexample
                    453: Also add information on how to contact you by electronic and paper mail.
                    455: If the program is interactive, make it output a short notice like this
                    456: when it starts in an interactive mode:
                    458: @smallexample
                    459: Gnomovision version 69, Copyright (C) 19@var{yy} @var{name of author}
                    460: Gnomovision comes with ABSOLUTELY NO WARRANTY; for details
                    461: type `show w'.  
                    462: This is free software, and you are welcome to redistribute it 
                    463: under certain conditions; type `show c' for details.
                    464: @end smallexample
                    466: The hypothetical commands @samp{show w} and @samp{show c} should show
                    467: the appropriate parts of the General Public License.  Of course, the
                    468: commands you use may be called something other than @samp{show w} and
                    469: @samp{show c}; they could even be mouse-clicks or menu items---whatever
                    470: suits your program.
                    472: You should also get your employer (if you work as a programmer) or your
                    473: school, if any, to sign a ``copyright disclaimer'' for the program, if
                    474: necessary.  Here is a sample; alter the names:
                    476: @smallexample
                    477: Yoyodyne, Inc., hereby disclaims all copyright interest in the program
                    478: `Gnomovision' (which makes passes at compilers) written by James Hacker.
                    480: @var{signature of Ty Coon}, 1 April 1989
                    481: Ty Coon, President of Vice
                    482: @end smallexample
                    484: This General Public License does not permit incorporating your program into
                    485: proprietary programs.  If your program is a subroutine library, you may
                    486: consider it more useful to permit linking proprietary applications with the
                    487: library.  If this is what you want to do, use the GNU Library General
                    488: Public License instead of this License.
1.1       anton     489: 
                    490: @iftex
1.23      pazsan    491: @node    Preface
                    492: @comment node-name,     next,           previous, up
1.1       anton     493: @unnumbered Preface
1.23      pazsan    494: @cindex Preface
1.17      anton     495: This manual documents Gforth. The reader is expected to know
1.1       anton     496: Forth. This manual is primarily a reference manual. @xref{Other Books}
                    497: for introductory material.
                    498: @end iftex
                    500: @node    Goals, Other Books, License, Top
                    501: @comment node-name,     next,           previous, up
1.17      anton     502: @chapter Goals of Gforth
1.1       anton     503: @cindex Goals
1.17      anton     504: The goal of the Gforth Project is to develop a standard model for
1.1       anton     505: ANSI Forth. This can be split into several subgoals:
                    507: @itemize @bullet
                    508: @item
1.17      anton     509: Gforth should conform to the ANSI Forth standard.
1.1       anton     510: @item
                    511: It should be a model, i.e. it should define all the
                    512: implementation-dependent things.
                    513: @item
                    514: It should become standard, i.e. widely accepted and used. This goal
                    515: is the most difficult one.
                    516: @end itemize
1.17      anton     518: To achieve these goals Gforth should be
1.1       anton     519: @itemize @bullet
                    520: @item
                    521: Similar to previous models (fig-Forth, F83)
                    522: @item
                    523: Powerful. It should provide for all the things that are considered
                    524: necessary today and even some that are not yet considered necessary.
                    525: @item
                    526: Efficient. It should not get the reputation of being exceptionally
                    527: slow.
                    528: @item
                    529: Free.
                    530: @item
                    531: Available on many machines/easy to port.
                    532: @end itemize
1.17      anton     534: Have we achieved these goals? Gforth conforms to the ANS Forth
                    535: standard. It may be considered a model, but we have not yet documented
1.1       anton     536: which parts of the model are stable and which parts we are likely to
1.17      anton     537: change. It certainly has not yet become a de facto standard. It has some
                    538: similarities and some differences to previous models. It has some
                    539: powerful features, but not yet everything that we envisioned. We
                    540: certainly have achieved our execution speed goals (@pxref{Performance}).
                    541: It is free and available on many machines.
1.1       anton     542: 
                    543: @node Other Books, Invocation, Goals, Top
                    544: @chapter Other books on ANS Forth
                    546: As the standard is relatively new, there are not many books out yet. It
1.17      anton     547: is not recommended to learn Forth by using Gforth and a book that is
1.1       anton     548: not written for ANS Forth, as you will not know your mistakes from the
                    549: deviations of the book.
                    551: There is, of course, the standard, the definite reference if you want to
1.19      anton     552: write ANS Forth programs. It is available in printed form from the
                    553: National Standards Institute Sales Department (Tel.: USA (212) 642-4900;
                    554: Fax.: USA (212) 302-1286) as document @cite{X3.215-1994} for about $200. You
                    555: can also get it from Global Engineering Documents (Tel.: USA (800)
                    556: 854-7179; Fax.: (303) 843-9880) for about $300.
                    558: @cite{dpANS6}, the last draft of the standard, which was then submitted to ANSI
                    559: for publication is available electronically and for free in some MS Word
                    560: format, and it has been converted to HTML. Some pointers to these
                    561: versions can be found through
1.24    ! anton     562: @*@file{}.
1.1       anton     563: 
1.21      anton     564: @cite{Forth: The new model} by Jack Woehr (Prentice-Hall, 1993) is an
1.1       anton     565: introductory book based on a draft version of the standard. It does not
                    566: cover the whole standard. It also contains interesting background
                    567: information (Jack Woehr was in the ANS Forth Technical Committe). It is
                    568: not appropriate for complete newbies, but programmers experienced in
                    569: other languages should find it ok.
                    571: @node Invocation, Words, Other Books, Top
                    572: @chapter Invocation
                    574: You will usually just say @code{gforth}. In many other cases the default
1.17      anton     575: Gforth image will be invoked like this:
1.1       anton     576: 
                    577: @example
                    578: gforth [files] [-e forth-code]
                    579: @end example
                    581: executing the contents of the files and the Forth code in the order they
                    582: are given.
                    584: In general, the command line looks like this:
                    586: @example
                    587: gforth [initialization options] [image-specific options]
                    588: @end example
                    590: The initialization options must come before the rest of the command
                    591: line. They are:
                    593: @table @code
                    594: @item --image-file @var{file}
1.20      pazsan    595: @item -i @var{file}
1.1       anton     596: Loads the Forth image @var{file} instead of the default
                    597: @file{}.
                    599: @item --path @var{path}
1.20      pazsan    600: @item -p @var{path}
1.1       anton     601: Uses @var{path} for searching the image file and Forth source code
                    602: files instead of the default in the environment variable
                    603: @code{GFORTHPATH} or the path specified at installation time (typically
                    604: @file{/usr/local/lib/gforth:.}). A path is given as a @code{:}-separated
                    605: list.
                    607: @item --dictionary-size @var{size}
                    608: @item -m @var{size}
                    609: Allocate @var{size} space for the Forth dictionary space instead of
                    610: using the default specified in the image (typically 256K). The
                    611: @var{size} specification consists of an integer and a unit (e.g.,
                    612: @code{4M}). The unit can be one of @code{b} (bytes), @code{e} (element
                    613: size, in this case Cells), @code{k} (kilobytes), and @code{M}
                    614: (Megabytes). If no unit is specified, @code{e} is used.
                    616: @item --data-stack-size @var{size}
                    617: @item -d @var{size}
                    618: Allocate @var{size} space for the data stack instead of using the
                    619: default specified in the image (typically 16K).
                    621: @item --return-stack-size @var{size}
                    622: @item -r @var{size}
                    623: Allocate @var{size} space for the return stack instead of using the
                    624: default specified in the image (typically 16K).
                    626: @item --fp-stack-size @var{size}
                    627: @item -f @var{size}
                    628: Allocate @var{size} space for the floating point stack instead of
                    629: using the default specified in the image (typically 16K). In this case
                    630: the unit specifier @code{e} refers to floating point numbers.
                    632: @item --locals-stack-size @var{size}
                    633: @item -l @var{size}
                    634: Allocate @var{size} space for the locals stack instead of using the
                    635: default specified in the image (typically 16K).
                    637: @end table
                    639: As explained above, the image-specific command-line arguments for the
                    640: default image @file{} consist of a sequence of filenames and
                    641: @code{-e @var{forth-code}} options that are interpreted in the seqence
                    642: in which they are given. The @code{-e @var{forth-code}} or
                    643: @code{--evaluate @var{forth-code}} option evaluates the forth
                    644: code. This option takes only one argument; if you want to evaluate more
                    645: Forth words, you have to quote them or use several @code{-e}s. To exit
                    646: after processing the command line (instead of entering interactive mode)
                    647: append @code{-e bye} to the command line.
1.22      anton     649: If you have several versions of Gforth installed, @code{gforth} will
                    650: invoke the version that was installed last. @code{gforth-@var{version}}
                    651: invokes a specific version. You may want to use the option
                    652: @code{--path}, if your environment contains the variable
                    653: @code{GFORTHPATH}.
1.1       anton     655: Not yet implemented:
                    656: On startup the system first executes the system initialization file
                    657: (unless the option @code{--no-init-file} is given; note that the system
                    658: resulting from using this option may not be ANS Forth conformant). Then
                    659: the user initialization file @file{.gforth.fs} is executed, unless the
                    660: option @code{--no-rc} is given; this file is first searched in @file{.},
                    661: then in @file{~}, then in the normal path (see above).
1.4       anton     663: @node Words, ANS conformance, Invocation, Top
1.1       anton     664: @chapter Forth Words
                    666: @menu
1.4       anton     667: * Notation::                    
                    668: * Arithmetic::                  
                    669: * Stack Manipulation::          
                    670: * Memory access::               
                    671: * Control Structures::          
                    672: * Locals::                      
                    673: * Defining Words::              
                    674: * Wordlists::                   
                    675: * Files::                       
                    676: * Blocks::                      
                    677: * Other I/O::                   
                    678: * Programming Tools::           
1.18      anton     679: * Assembler and Code words::    
1.4       anton     680: * Threading Words::             
1.1       anton     681: @end menu
                    683: @node Notation, Arithmetic, Words, Words
                    684: @section Notation
                    686: The Forth words are described in this section in the glossary notation
                    687: that has become a de-facto standard for Forth texts, i.e.
1.4       anton     689: @format
1.1       anton     690: @var{word}     @var{Stack effect}   @var{wordset}   @var{pronunciation}
1.4       anton     691: @end format
1.1       anton     692: @var{Description}
                    694: @table @var
                    695: @item word
1.17      anton     696: The name of the word. BTW, Gforth is case insensitive, so you can
1.14      anton     697: type the words in in lower case (However, @pxref{core-idef}).
1.1       anton     698: 
                    699: @item Stack effect
                    700: The stack effect is written in the notation @code{@var{before} --
                    701: @var{after}}, where @var{before} and @var{after} describe the top of
                    702: stack entries before and after the execution of the word. The rest of
                    703: the stack is not touched by the word. The top of stack is rightmost,
1.17      anton     704: i.e., a stack sequence is written as it is typed in. Note that Gforth
1.1       anton     705: uses a separate floating point stack, but a unified stack
                    706: notation. Also, return stack effects are not shown in @var{stack
                    707: effect}, but in @var{Description}. The name of a stack item describes
                    708: the type and/or the function of the item. See below for a discussion of
                    709: the types.
1.19      anton     711: All words have two stack effects: A compile-time stack effect and a
                    712: run-time stack effect. The compile-time stack-effect of most words is
                    713: @var{ -- }. If the compile-time stack-effect of a word deviates from
                    714: this standard behaviour, or the word does other unusual things at
                    715: compile time, both stack effects are shown; otherwise only the run-time
                    716: stack effect is shown.
1.1       anton     718: @item pronunciation
                    719: How the word is pronounced
                    721: @item wordset
                    722: The ANS Forth standard is divided into several wordsets. A standard
                    723: system need not support all of them. So, the fewer wordsets your program
                    724: uses the more portable it will be in theory. However, we suspect that
                    725: most ANS Forth systems on personal machines will feature all
                    726: wordsets. Words that are not defined in the ANS standard have
1.19      anton     727: @code{gforth} or @code{gforth-internal} as wordset. @code{gforth}
                    728: describes words that will work in future releases of Gforth;
                    729: @code{gforth-internal} words are more volatile. Environmental query
                    730: strings are also displayed like words; you can recognize them by the
                    731: @code{environment} in the wordset field.
1.1       anton     732: 
                    733: @item Description
                    734: A description of the behaviour of the word.
                    735: @end table
1.4       anton     737: The type of a stack item is specified by the character(s) the name
                    738: starts with:
1.1       anton     739: 
                    740: @table @code
                    741: @item f
                    742: Bool, i.e. @code{false} or @code{true}.
                    743: @item c
                    744: Char
                    745: @item w
                    746: Cell, can contain an integer or an address
                    747: @item n
                    748: signed integer
                    749: @item u
                    750: unsigned integer
                    751: @item d
                    752: double sized signed integer
                    753: @item ud
                    754: double sized unsigned integer
                    755: @item r
                    756: Float
                    757: @item a_
                    758: Cell-aligned address
                    759: @item c_
                    760: Char-aligned address (note that a Char is two bytes in Windows NT)
                    761: @item f_
                    762: Float-aligned address
                    763: @item df_
                    764: Address aligned for IEEE double precision float
                    765: @item sf_
                    766: Address aligned for IEEE single precision float
                    767: @item xt
                    768: Execution token, same size as Cell
                    769: @item wid
                    770: Wordlist ID, same size as Cell
                    771: @item f83name
                    772: Pointer to a name structure
                    773: @end table
1.4       anton     775: @node Arithmetic, Stack Manipulation, Notation, Words
1.1       anton     776: @section Arithmetic
                    777: Forth arithmetic is not checked, i.e., you will not hear about integer
                    778: overflow on addition or multiplication, you may hear about division by
                    779: zero if you are lucky. The operator is written after the operands, but
                    780: the operands are still in the original order. I.e., the infix @code{2-1}
                    781: corresponds to @code{2 1 -}. Forth offers a variety of division
                    782: operators. If you perform division with potentially negative operands,
                    783: you do not want to use @code{/} or @code{/mod} with its undefined
                    784: behaviour, but rather @code{fm/mod} or @code{sm/mod} (probably the
1.4       anton     785: former, @pxref{Mixed precision}).
                    787: @menu
                    788: * Single precision::            
                    789: * Bitwise operations::          
                    790: * Mixed precision::             operations with single and double-cell integers
                    791: * Double precision::            Double-cell integer arithmetic
                    792: * Floating Point::              
                    793: @end menu
1.1       anton     794: 
1.4       anton     795: @node Single precision, Bitwise operations, Arithmetic, Arithmetic
1.1       anton     796: @subsection Single precision
                    797: doc-+
                    798: doc--
                    799: doc-*
                    800: doc-/
                    801: doc-mod
                    802: doc-/mod
                    803: doc-negate
                    804: doc-abs
                    805: doc-min
                    806: doc-max
1.4       anton     808: @node Bitwise operations, Mixed precision, Single precision, Arithmetic
1.1       anton     809: @subsection Bitwise operations
                    810: doc-and
                    811: doc-or
                    812: doc-xor
                    813: doc-invert
                    814: doc-2*
                    815: doc-2/
1.4       anton     817: @node Mixed precision, Double precision, Bitwise operations, Arithmetic
1.1       anton     818: @subsection Mixed precision
                    819: doc-m+
                    820: doc-*/
                    821: doc-*/mod
                    822: doc-m*
                    823: doc-um*
                    824: doc-m*/
                    825: doc-um/mod
                    826: doc-fm/mod
                    827: doc-sm/rem
1.4       anton     829: @node Double precision, Floating Point, Mixed precision, Arithmetic
1.1       anton     830: @subsection Double precision
1.16      anton     831: 
                    832: The outer (aka text) interpreter converts numbers containing a dot into
                    833: a double precision number. Note that only numbers with the dot as last
                    834: character are standard-conforming.
1.1       anton     836: doc-d+
                    837: doc-d-
                    838: doc-dnegate
                    839: doc-dabs
                    840: doc-dmin
                    841: doc-dmax
1.4       anton     843: @node Floating Point,  , Double precision, Arithmetic
                    844: @subsection Floating Point
1.16      anton     845: 
                    846: The format of floating point numbers recognized by the outer (aka text)
                    847: interpreter is: a signed decimal number, possibly containing a decimal
                    848: point (@code{.}), followed by @code{E} or @code{e}, optionally followed
                    849: by a signed integer (the exponent). E.g., @code{1e} ist the same as
                    850: @code{+1.0e+1}. Note that a number without @code{e}
                    851: is not interpreted as floating-point number, but as double (if the
                    852: number contains a @code{.}) or single precision integer. Also,
                    853: conversions between string and floating point numbers always use base
                    854: 10, irrespective of the value of @code{BASE}. If @code{BASE} contains a
                    855: value greater then 14, the @code{E} may be interpreted as digit and the
                    856: number will be interpreted as integer, unless it has a signed exponent
                    857: (both @code{+} and @code{-} are allowed as signs).
1.4       anton     858: 
                    859: Angles in floating point operations are given in radians (a full circle
1.17      anton     860: has 2 pi radians). Note, that Gforth has a separate floating point
1.4       anton     861: stack, but we use the unified notation.
                    863: Floating point numbers have a number of unpleasant surprises for the
                    864: unwary (e.g., floating point addition is not associative) and even a few
                    865: for the wary. You should not use them unless you know what you are doing
                    866: or you don't care that the results you get are totally bogus. If you
                    867: want to learn about the problems of floating point numbers (and how to
1.11      anton     868: avoid them), you might start with @cite{David Goldberg, What Every
1.6       anton     869: Computer Scientist Should Know About Floating-Point Arithmetic, ACM
                    870: Computing Surveys 23(1):5@minus{}48, March 1991}.
1.4       anton     871: 
                    872: doc-f+
                    873: doc-f-
                    874: doc-f*
                    875: doc-f/
                    876: doc-fnegate
                    877: doc-fabs
                    878: doc-fmax
                    879: doc-fmin
                    880: doc-floor
                    881: doc-fround
                    882: doc-f**
                    883: doc-fsqrt
                    884: doc-fexp
                    885: doc-fexpm1
                    886: doc-fln
                    887: doc-flnp1
                    888: doc-flog
1.6       anton     889: doc-falog
1.4       anton     890: doc-fsin
                    891: doc-fcos
                    892: doc-fsincos
                    893: doc-ftan
                    894: doc-fasin
                    895: doc-facos
                    896: doc-fatan
                    897: doc-fatan2
                    898: doc-fsinh
                    899: doc-fcosh
                    900: doc-ftanh
                    901: doc-fasinh
                    902: doc-facosh
                    903: doc-fatanh
                    905: @node Stack Manipulation, Memory access, Arithmetic, Words
1.1       anton     906: @section Stack Manipulation
1.17      anton     908: Gforth has a data stack (aka parameter stack) for characters, cells,
1.1       anton     909: addresses, and double cells, a floating point stack for floating point
                    910: numbers, a return stack for storing the return addresses of colon
                    911: definitions and other data, and a locals stack for storing local
                    912: variables. Note that while every sane Forth has a separate floating
                    913: point stack, this is not strictly required; an ANS Forth system could
                    914: theoretically keep floating point numbers on the data stack. As an
                    915: additional difficulty, you don't know how many cells a floating point
                    916: number takes. It is reportedly possible to write words in a way that
                    917: they work also for a unified stack model, but we do not recommend trying
1.4       anton     918: it. Instead, just say that your program has an environmental dependency
                    919: on a separate FP stack.
                    921: Also, a Forth system is allowed to keep the local variables on the
1.1       anton     922: return stack. This is reasonable, as local variables usually eliminate
                    923: the need to use the return stack explicitly. So, if you want to produce
                    924: a standard complying program and if you are using local variables in a
                    925: word, forget about return stack manipulations in that word (see the
                    926: standard document for the exact rules).
1.4       anton     928: @menu
                    929: * Data stack::                  
                    930: * Floating point stack::        
                    931: * Return stack::                
                    932: * Locals stack::                
                    933: * Stack pointer manipulation::  
                    934: @end menu
                    936: @node Data stack, Floating point stack, Stack Manipulation, Stack Manipulation
1.1       anton     937: @subsection Data stack
                    938: doc-drop
                    939: doc-nip
                    940: doc-dup
                    941: doc-over
                    942: doc-tuck
                    943: doc-swap
                    944: doc-rot
                    945: doc--rot
                    946: doc-?dup
                    947: doc-pick
                    948: doc-roll
                    949: doc-2drop
                    950: doc-2nip
                    951: doc-2dup
                    952: doc-2over
                    953: doc-2tuck
                    954: doc-2swap
                    955: doc-2rot
1.4       anton     957: @node Floating point stack, Return stack, Data stack, Stack Manipulation
1.1       anton     958: @subsection Floating point stack
                    959: doc-fdrop
                    960: doc-fnip
                    961: doc-fdup
                    962: doc-fover
                    963: doc-ftuck
                    964: doc-fswap
                    965: doc-frot
1.4       anton     967: @node Return stack, Locals stack, Floating point stack, Stack Manipulation
1.1       anton     968: @subsection Return stack
                    969: doc->r
                    970: doc-r>
                    971: doc-r@
                    972: doc-rdrop
                    973: doc-2>r
                    974: doc-2r>
                    975: doc-2r@
                    976: doc-2rdrop
1.4       anton     978: @node Locals stack, Stack pointer manipulation, Return stack, Stack Manipulation
1.1       anton     979: @subsection Locals stack
1.4       anton     981: @node Stack pointer manipulation,  , Locals stack, Stack Manipulation
1.1       anton     982: @subsection Stack pointer manipulation
                    983: doc-sp@
                    984: doc-sp!
                    985: doc-fp@
                    986: doc-fp!
                    987: doc-rp@
                    988: doc-rp!
                    989: doc-lp@
                    990: doc-lp!
1.4       anton     992: @node Memory access, Control Structures, Stack Manipulation, Words
1.1       anton     993: @section Memory access
1.4       anton     995: @menu
                    996: * Stack-Memory transfers::      
                    997: * Address arithmetic::          
                    998: * Memory block access::         
                    999: @end menu
                   1001: @node Stack-Memory transfers, Address arithmetic, Memory access, Memory access
1.1       anton    1002: @subsection Stack-Memory transfers
                   1004: doc-@
                   1005: doc-!
                   1006: doc-+!
                   1007: doc-c@
                   1008: doc-c!
                   1009: doc-2@
                   1010: doc-2!
                   1011: doc-f@
                   1012: doc-f!
                   1013: doc-sf@
                   1014: doc-sf!
                   1015: doc-df@
                   1016: doc-df!
1.4       anton    1018: @node Address arithmetic, Memory block access, Stack-Memory transfers, Memory access
1.1       anton    1019: @subsection Address arithmetic
                   1021: ANS Forth does not specify the sizes of the data types. Instead, it
                   1022: offers a number of words for computing sizes and doing address
                   1023: arithmetic. Basically, address arithmetic is performed in terms of
                   1024: address units (aus); on most systems the address unit is one byte. Note
                   1025: that a character may have more than one au, so @code{chars} is no noop
                   1026: (on systems where it is a noop, it compiles to nothing).
                   1028: ANS Forth also defines words for aligning addresses for specific
                   1029: addresses. Many computers require that accesses to specific data types
                   1030: must only occur at specific addresses; e.g., that cells may only be
                   1031: accessed at addresses divisible by 4. Even if a machine allows unaligned
                   1032: accesses, it can usually perform aligned accesses faster. 
1.17      anton    1034: For the performance-conscious: alignment operations are usually only
1.1       anton    1035: necessary during the definition of a data structure, not during the
                   1036: (more frequent) accesses to it.
                   1038: ANS Forth defines no words for character-aligning addresses. This is not
                   1039: an oversight, but reflects the fact that addresses that are not
                   1040: char-aligned have no use in the standard and therefore will not be
                   1041: created.
                   1043: The standard guarantees that addresses returned by @code{CREATE}d words
1.17      anton    1044: are cell-aligned; in addition, Gforth guarantees that these addresses
1.1       anton    1045: are aligned for all purposes.
1.9       anton    1047: Note that the standard defines a word @code{char}, which has nothing to
                   1048: do with address arithmetic.
1.1       anton    1050: doc-chars
                   1051: doc-char+
                   1052: doc-cells
                   1053: doc-cell+
                   1054: doc-align
                   1055: doc-aligned
                   1056: doc-floats
                   1057: doc-float+
                   1058: doc-falign
                   1059: doc-faligned
                   1060: doc-sfloats
                   1061: doc-sfloat+
                   1062: doc-sfalign
                   1063: doc-sfaligned
                   1064: doc-dfloats
                   1065: doc-dfloat+
                   1066: doc-dfalign
                   1067: doc-dfaligned
1.10      anton    1068: doc-maxalign
                   1069: doc-maxaligned
                   1070: doc-cfalign
                   1071: doc-cfaligned
1.1       anton    1072: doc-address-unit-bits
1.4       anton    1074: @node Memory block access,  , Address arithmetic, Memory access
1.1       anton    1075: @subsection Memory block access
                   1077: doc-move
                   1078: doc-erase
                   1080: While the previous words work on address units, the rest works on
                   1081: characters.
                   1083: doc-cmove
                   1084: doc-cmove>
                   1085: doc-fill
                   1086: doc-blank
1.4       anton    1088: @node Control Structures, Locals, Memory access, Words
1.1       anton    1089: @section Control Structures
                   1091: Control structures in Forth cannot be used in interpret state, only in
                   1092: compile state, i.e., in a colon definition. We do not like this
                   1093: limitation, but have not seen a satisfying way around it yet, although
                   1094: many schemes have been proposed.
1.4       anton    1096: @menu
                   1097: * Selection::                   
                   1098: * Simple Loops::                
                   1099: * Counted Loops::               
                   1100: * Arbitrary control structures::  
                   1101: * Calls and returns::           
                   1102: * Exception Handling::          
                   1103: @end menu
                   1105: @node Selection, Simple Loops, Control Structures, Control Structures
1.1       anton    1106: @subsection Selection
                   1108: @example
                   1109: @var{flag}
                   1110: IF
                   1111:   @var{code}
                   1112: ENDIF
                   1113: @end example
                   1114: or
                   1115: @example
                   1116: @var{flag}
                   1117: IF
                   1118:   @var{code1}
                   1119: ELSE
                   1120:   @var{code2}
                   1121: ENDIF
                   1122: @end example
1.4       anton    1124: You can use @code{THEN} instead of @code{ENDIF}. Indeed, @code{THEN} is
1.1       anton    1125: standard, and @code{ENDIF} is not, although it is quite popular. We
                   1126: recommend using @code{ENDIF}, because it is less confusing for people
                   1127: who also know other languages (and is not prone to reinforcing negative
                   1128: prejudices against Forth in these people). Adding @code{ENDIF} to a
                   1129: system that only supplies @code{THEN} is simple:
                   1130: @example
                   1131: : endif   POSTPONE then ; immediate
                   1132: @end example
                   1134: [According to @cite{Webster's New Encyclopedic Dictionary}, @dfn{then
                   1135: (adv.)}  has the following meanings:
                   1136: @quotation
                   1137: ... 2b: following next after in order ... 3d: as a necessary consequence
                   1138: (if you were there, then you saw them).
                   1139: @end quotation
                   1140: Forth's @code{THEN} has the meaning 2b, whereas @code{THEN} in Pascal
                   1141: and many other programming languages has the meaning 3d.]
                   1143: We also provide the words @code{?dup-if} and @code{?dup-0=-if}, so you
                   1144: can avoid using @code{?dup}.
                   1146: @example
                   1147: @var{n}
                   1148: CASE
                   1149:   @var{n1} OF @var{code1} ENDOF
                   1150:   @var{n2} OF @var{code2} ENDOF
1.4       anton    1151:   @dots{}
1.1       anton    1152: ENDCASE
                   1153: @end example
                   1155: Executes the first @var{codei}, where the @var{ni} is equal to
                   1156: @var{n}. A default case can be added by simply writing the code after
                   1157: the last @code{ENDOF}. It may use @var{n}, which is on top of the stack,
                   1158: but must not consume it.
1.4       anton    1160: @node Simple Loops, Counted Loops, Selection, Control Structures
1.1       anton    1161: @subsection Simple Loops
                   1163: @example
                   1164: BEGIN
                   1165:   @var{code1}
                   1166:   @var{flag}
                   1167: WHILE
                   1168:   @var{code2}
                   1169: REPEAT
                   1170: @end example
                   1172: @var{code1} is executed and @var{flag} is computed. If it is true,
                   1173: @var{code2} is executed and the loop is restarted; If @var{flag} is false, execution continues after the @code{REPEAT}.
                   1175: @example
                   1176: BEGIN
                   1177:   @var{code}
                   1178:   @var{flag}
                   1179: UNTIL
                   1180: @end example
                   1182: @var{code} is executed. The loop is restarted if @code{flag} is false.
                   1184: @example
                   1185: BEGIN
                   1186:   @var{code}
                   1187: AGAIN
                   1188: @end example
                   1190: This is an endless loop.
1.4       anton    1192: @node Counted Loops, Arbitrary control structures, Simple Loops, Control Structures
1.1       anton    1193: @subsection Counted Loops
                   1195: The basic counted loop is:
                   1196: @example
                   1197: @var{limit} @var{start}
                   1198: ?DO
                   1199:   @var{body}
                   1200: LOOP
                   1201: @end example
                   1203: This performs one iteration for every integer, starting from @var{start}
                   1204: and up to, but excluding @var{limit}. The counter, aka index, can be
                   1205: accessed with @code{i}. E.g., the loop
                   1206: @example
                   1207: 10 0 ?DO
                   1208:   i .
                   1209: LOOP
                   1210: @end example
                   1211: prints
                   1212: @example
                   1213: 0 1 2 3 4 5 6 7 8 9
                   1214: @end example
                   1215: The index of the innermost loop can be accessed with @code{i}, the index
                   1216: of the next loop with @code{j}, and the index of the third loop with
                   1217: @code{k}.
                   1219: The loop control data are kept on the return stack, so there are some
                   1220: restrictions on mixing return stack accesses and counted loop
                   1221: words. E.g., if you put values on the return stack outside the loop, you
                   1222: cannot read them inside the loop. If you put values on the return stack
                   1223: within a loop, you have to remove them before the end of the loop and
                   1224: before accessing the index of the loop.
                   1226: There are several variations on the counted loop:
                   1228: @code{LEAVE} leaves the innermost counted loop immediately.
1.18      anton    1230: If @var{start} is greater than @var{limit}, a @code{?DO} loop is entered
                   1231: (and @code{LOOP} iterates until they become equal by wrap-around
                   1232: arithmetic). This behaviour is usually not what you want. Therefore,
                   1233: Gforth offers @code{+DO} and @code{U+DO} (as replacements for
                   1234: @code{?DO}), which do not enter the loop if @var{start} is greater than
                   1235: @var{limit}; @code{+DO} is for signed loop parameters, @code{U+DO} for
                   1236: unsigned loop parameters. These words can be implemented easily on
                   1237: standard systems, so using them does not make your programs hard to
                   1238: port; e.g.:
                   1239: @example
                   1240: : +DO ( compile-time: -- do-sys; run-time: n1 n2 -- )
                   1241:     POSTPONE over POSTPONE min POSTPONE ?DO ; immediate
                   1242: @end example
1.1       anton    1244: @code{LOOP} can be replaced with @code{@var{n} +LOOP}; this updates the
                   1245: index by @var{n} instead of by 1. The loop is terminated when the border
                   1246: between @var{limit-1} and @var{limit} is crossed. E.g.:
1.18      anton    1248: @code{4 0 +DO  i .  2 +LOOP}   prints @code{0 2}
1.1       anton    1249: 
1.18      anton    1250: @code{4 1 +DO  i .  2 +LOOP}   prints @code{1 3}
1.1       anton    1251: 
                   1252: The behaviour of @code{@var{n} +LOOP} is peculiar when @var{n} is negative:
1.2       anton    1254: @code{-1 0 ?DO  i .  -1 +LOOP}  prints @code{0 -1}
1.1       anton    1255: 
1.2       anton    1256: @code{ 0 0 ?DO  i .  -1 +LOOP}  prints nothing
1.1       anton    1257: 
1.18      anton    1258: Therefore we recommend avoiding @code{@var{n} +LOOP} with negative
                   1259: @var{n}. One alternative is @code{@var{u} -LOOP}, which reduces the
                   1260: index by @var{u} each iteration. The loop is terminated when the border
                   1261: between @var{limit+1} and @var{limit} is crossed. Gforth also provides
                   1262: @code{-DO} and @code{U-DO} for down-counting loops. E.g.:
1.1       anton    1263: 
1.18      anton    1264: @code{-2 0 -DO  i .  1 -LOOP}  prints @code{0 -1}
1.1       anton    1265: 
1.18      anton    1266: @code{-1 0 -DO  i .  1 -LOOP}  prints @code{0}
1.1       anton    1267: 
1.18      anton    1268: @code{ 0 0 -DO  i .  1 -LOOP}  prints nothing
1.1       anton    1269: 
1.18      anton    1270: Another alternative is @code{@var{n} S+LOOP}, where the negative
                   1271: case behaves symmetrical to the positive case:
1.1       anton    1272: 
1.18      anton    1273: @code{-2 0 -DO  i .  -1 S+LOOP}  prints @code{0 -1}
                   1275: The loop is terminated when the border between @var{limit@minus{}sgn(n)}
                   1276: and @var{limit} is crossed. Unfortunately, neither @code{-LOOP} nor
                   1277: @code{S+LOOP} are part of the ANS Forth standard, and they are not easy
                   1278: to implement using standard words. If you want to write standard
                   1279: programs, just avoid counting down.
                   1281: @code{?DO} can also be replaced by @code{DO}. @code{DO} always enters
                   1282: the loop, independent of the loop parameters. Do not use @code{DO}, even
                   1283: if you know that the loop is entered in any case. Such knowledge tends
                   1284: to become invalid during maintenance of a program, and then the
                   1285: @code{DO} will make trouble.
1.1       anton    1286: 
                   1287: @code{UNLOOP} is used to prepare for an abnormal loop exit, e.g., via
                   1288: @code{EXIT}. @code{UNLOOP} removes the loop control parameters from the
                   1289: return stack so @code{EXIT} can get to its return address.
                   1291: Another counted loop is
                   1292: @example
                   1293: @var{n}
                   1294: FOR
                   1295:   @var{body}
                   1296: NEXT
                   1297: @end example
                   1298: This is the preferred loop of native code compiler writers who are too
1.17      anton    1299: lazy to optimize @code{?DO} loops properly. In Gforth, this loop
1.1       anton    1300: iterates @var{n+1} times; @code{i} produces values starting with @var{n}
                   1301: and ending with 0. Other Forth systems may behave differently, even if
                   1302: they support @code{FOR} loops.
1.4       anton    1304: @node Arbitrary control structures, Calls and returns, Counted Loops, Control Structures
1.2       anton    1305: @subsection Arbitrary control structures
                   1307: ANS Forth permits and supports using control structures in a non-nested
                   1308: way. Information about incomplete control structures is stored on the
                   1309: control-flow stack. This stack may be implemented on the Forth data
1.17      anton    1310: stack, and this is what we have done in Gforth.
1.2       anton    1311: 
                   1312: An @i{orig} entry represents an unresolved forward branch, a @i{dest}
                   1313: entry represents a backward branch target. A few words are the basis for
                   1314: building any control structure possible (except control structures that
                   1315: need storage, like calls, coroutines, and backtracking).
1.3       anton    1317: doc-if
                   1318: doc-ahead
                   1319: doc-then
                   1320: doc-begin
                   1321: doc-until
                   1322: doc-again
                   1323: doc-cs-pick
                   1324: doc-cs-roll
1.2       anton    1325: 
1.17      anton    1326: On many systems control-flow stack items take one word, in Gforth they
1.2       anton    1327: currently take three (this may change in the future). Therefore it is a
                   1328: really good idea to manipulate the control flow stack with
                   1329: @code{cs-pick} and @code{cs-roll}, not with data stack manipulation
                   1330: words.
                   1332: Some standard control structure words are built from these words:
1.3       anton    1334: doc-else
                   1335: doc-while
                   1336: doc-repeat
1.2       anton    1337: 
                   1338: Counted loop words constitute a separate group of words:
1.3       anton    1340: doc-?do
1.18      anton    1341: doc-+do
                   1342: doc-u+do
                   1343: doc--do
                   1344: doc-u-do
1.3       anton    1345: doc-do
                   1346: doc-for
                   1347: doc-loop
                   1348: doc-s+loop
                   1349: doc-+loop
1.18      anton    1350: doc--loop
1.3       anton    1351: doc-next
                   1352: doc-leave
                   1353: doc-?leave
                   1354: doc-unloop
1.10      anton    1355: doc-done
1.2       anton    1356: 
                   1357: The standard does not allow using @code{cs-pick} and @code{cs-roll} on
                   1358: @i{do-sys}. Our system allows it, but it's your job to ensure that for
                   1359: every @code{?DO} etc. there is exactly one @code{UNLOOP} on any path
1.3       anton    1360: through the definition (@code{LOOP} etc. compile an @code{UNLOOP} on the
                   1361: fall-through path). Also, you have to ensure that all @code{LEAVE}s are
1.7       pazsan   1362: resolved (by using one of the loop-ending words or @code{DONE}).
1.2       anton    1363: 
                   1364: Another group of control structure words are
1.3       anton    1366: doc-case
                   1367: doc-endcase
                   1368: doc-of
                   1369: doc-endof
1.2       anton    1370: 
                   1371: @i{case-sys} and @i{of-sys} cannot be processed using @code{cs-pick} and
                   1372: @code{cs-roll}.
1.3       anton    1374: @subsubsection Programming Style
                   1376: In order to ensure readability we recommend that you do not create
                   1377: arbitrary control structures directly, but define new control structure
                   1378: words for the control structure you want and use these words in your
                   1379: program.
                   1381: E.g., instead of writing
                   1383: @example
                   1384: begin
                   1385:   ...
                   1386: if [ 1 cs-roll ]
                   1387:   ...
                   1388: again then
                   1389: @end example
                   1391: we recommend defining control structure words, e.g.,
                   1393: @example
                   1394: : while ( dest -- orig dest )
                   1395:  POSTPONE if
                   1396:  1 cs-roll ; immediate
                   1398: : repeat ( orig dest -- )
                   1399:  POSTPONE again
                   1400:  POSTPONE then ; immediate
                   1401: @end example
                   1403: and then using these to create the control structure:
                   1405: @example
                   1406: begin
                   1407:   ...
                   1408: while
                   1409:   ...
                   1410: repeat
                   1411: @end example
                   1413: That's much easier to read, isn't it? Of course, @code{BEGIN} and
                   1414: @code{WHILE} are predefined, so in this example it would not be
                   1415: necessary to define them.
1.4       anton    1417: @node Calls and returns, Exception Handling, Arbitrary control structures, Control Structures
1.3       anton    1418: @subsection Calls and returns
                   1420: A definition can be called simply be writing the name of the
1.17      anton    1421: definition. When the end of the definition is reached, it returns. An
                   1422: earlier return can be forced using
1.3       anton    1423: 
                   1424: doc-exit
                   1426: Don't forget to clean up the return stack and @code{UNLOOP} any
                   1427: outstanding @code{?DO}...@code{LOOP}s before @code{EXIT}ing. The
                   1428: primitive compiled by @code{EXIT} is
                   1430: doc-;s
1.4       anton    1432: @node Exception Handling,  , Calls and returns, Control Structures
1.3       anton    1433: @subsection Exception Handling
                   1435: doc-catch
                   1436: doc-throw
1.4       anton    1438: @node Locals, Defining Words, Control Structures, Words
1.1       anton    1439: @section Locals
1.2       anton    1441: Local variables can make Forth programming more enjoyable and Forth
                   1442: programs easier to read. Unfortunately, the locals of ANS Forth are
                   1443: laden with restrictions. Therefore, we provide not only the ANS Forth
                   1444: locals wordset, but also our own, more powerful locals wordset (we
                   1445: implemented the ANS Forth locals wordset through our locals wordset).
1.24    ! anton    1447: The ideas in this section have also been published in the paper
        !          1448: @cite{Automatic Scoping of Local Variables} by M. Anton Ertl, presented
        !          1449: at EuroForth '94; it is available at
        !          1450: @*@file{}.
        !          1451: 
1.2       anton    1452: @menu
1.17      anton    1453: * Gforth locals::               
1.4       anton    1454: * ANS Forth locals::            
1.2       anton    1455: @end menu
1.17      anton    1457: @node Gforth locals, ANS Forth locals, Locals, Locals
                   1458: @subsection Gforth locals
1.2       anton    1459: 
                   1460: Locals can be defined with
                   1462: @example
                   1463: @{ local1 local2 ... -- comment @}
                   1464: @end example
                   1465: or
                   1466: @example
                   1467: @{ local1 local2 ... @}
                   1468: @end example
                   1470: E.g.,
                   1471: @example
                   1472: : max @{ n1 n2 -- n3 @}
                   1473:  n1 n2 > if
                   1474:    n1
                   1475:  else
                   1476:    n2
                   1477:  endif ;
                   1478: @end example
                   1480: The similarity of locals definitions with stack comments is intended. A
                   1481: locals definition often replaces the stack comment of a word. The order
                   1482: of the locals corresponds to the order in a stack comment and everything
                   1483: after the @code{--} is really a comment.
                   1485: This similarity has one disadvantage: It is too easy to confuse locals
                   1486: declarations with stack comments, causing bugs and making them hard to
                   1487: find. However, this problem can be avoided by appropriate coding
                   1488: conventions: Do not use both notations in the same program. If you do,
                   1489: they should be distinguished using additional means, e.g. by position.
                   1491: The name of the local may be preceded by a type specifier, e.g.,
                   1492: @code{F:} for a floating point value:
                   1494: @example
                   1495: : CX* @{ F: Ar F: Ai F: Br F: Bi -- Cr Ci @}
                   1496: \ complex multiplication
                   1497:  Ar Br f* Ai Bi f* f-
                   1498:  Ar Bi f* Ai Br f* f+ ;
                   1499: @end example
1.17      anton    1501: Gforth currently supports cells (@code{W:}, @code{W^}), doubles
1.2       anton    1502: (@code{D:}, @code{D^}), floats (@code{F:}, @code{F^}) and characters
                   1503: (@code{C:}, @code{C^}) in two flavours: a value-flavoured local (defined
                   1504: with @code{W:}, @code{D:} etc.) produces its value and can be changed
                   1505: with @code{TO}. A variable-flavoured local (defined with @code{W^} etc.)
                   1506: produces its address (which becomes invalid when the variable's scope is
                   1507: left). E.g., the standard word @code{emit} can be defined in therms of
                   1508: @code{type} like this:
                   1510: @example
                   1511: : emit @{ C^ char* -- @}
                   1512:     char* 1 type ;
                   1513: @end example
                   1515: A local without type specifier is a @code{W:} local. Both flavours of
                   1516: locals are initialized with values from the data or FP stack.
                   1518: Currently there is no way to define locals with user-defined data
                   1519: structures, but we are working on it.
1.17      anton    1521: Gforth allows defining locals everywhere in a colon definition. This
1.7       pazsan   1522: poses the following questions:
1.2       anton    1523: 
1.4       anton    1524: @menu
                   1525: * Where are locals visible by name?::  
1.14      anton    1526: * How long do locals live?::    
1.4       anton    1527: * Programming Style::           
                   1528: * Implementation::              
                   1529: @end menu
1.17      anton    1531: @node Where are locals visible by name?, How long do locals live?, Gforth locals, Gforth locals
1.2       anton    1532: @subsubsection Where are locals visible by name?
                   1534: Basically, the answer is that locals are visible where you would expect
                   1535: it in block-structured languages, and sometimes a little longer. If you
                   1536: want to restrict the scope of a local, enclose its definition in
                   1537: @code{SCOPE}...@code{ENDSCOPE}.
                   1539: doc-scope
                   1540: doc-endscope
                   1542: These words behave like control structure words, so you can use them
                   1543: with @code{CS-PICK} and @code{CS-ROLL} to restrict the scope in
                   1544: arbitrary ways.
                   1546: If you want a more exact answer to the visibility question, here's the
                   1547: basic principle: A local is visible in all places that can only be
                   1548: reached through the definition of the local@footnote{In compiler
                   1549: construction terminology, all places dominated by the definition of the
                   1550: local.}. In other words, it is not visible in places that can be reached
                   1551: without going through the definition of the local. E.g., locals defined
                   1552: in @code{IF}...@code{ENDIF} are visible until the @code{ENDIF}, locals
                   1553: defined in @code{BEGIN}...@code{UNTIL} are visible after the
                   1554: @code{UNTIL} (until, e.g., a subsequent @code{ENDSCOPE}).
                   1556: The reasoning behind this solution is: We want to have the locals
                   1557: visible as long as it is meaningful. The user can always make the
                   1558: visibility shorter by using explicit scoping. In a place that can
                   1559: only be reached through the definition of a local, the meaning of a
                   1560: local name is clear. In other places it is not: How is the local
                   1561: initialized at the control flow path that does not contain the
                   1562: definition? Which local is meant, if the same name is defined twice in
                   1563: two independent control flow paths?
                   1565: This should be enough detail for nearly all users, so you can skip the
                   1566: rest of this section. If you relly must know all the gory details and
                   1567: options, read on.
                   1569: In order to implement this rule, the compiler has to know which places
                   1570: are unreachable. It knows this automatically after @code{AHEAD},
                   1571: @code{AGAIN}, @code{EXIT} and @code{LEAVE}; in other cases (e.g., after
                   1572: most @code{THROW}s), you can use the word @code{UNREACHABLE} to tell the
                   1573: compiler that the control flow never reaches that place. If
                   1574: @code{UNREACHABLE} is not used where it could, the only consequence is
                   1575: that the visibility of some locals is more limited than the rule above
                   1576: says. If @code{UNREACHABLE} is used where it should not (i.e., if you
                   1577: lie to the compiler), buggy code will be produced.
                   1579: Another problem with this rule is that at @code{BEGIN}, the compiler
1.3       anton    1580: does not know which locals will be visible on the incoming
                   1581: back-edge. All problems discussed in the following are due to this
                   1582: ignorance of the compiler (we discuss the problems using @code{BEGIN}
                   1583: loops as examples; the discussion also applies to @code{?DO} and other
1.2       anton    1584: loops). Perhaps the most insidious example is:
                   1585: @example
                   1586: AHEAD
                   1587: BEGIN
                   1588:   x
                   1589: [ 1 CS-ROLL ] THEN
1.4       anton    1590:   @{ x @}
1.2       anton    1591:   ...
                   1592: UNTIL
                   1593: @end example
                   1595: This should be legal according to the visibility rule. The use of
                   1596: @code{x} can only be reached through the definition; but that appears
                   1597: textually below the use.
                   1599: From this example it is clear that the visibility rules cannot be fully
                   1600: implemented without major headaches. Our implementation treats common
                   1601: cases as advertised and the exceptions are treated in a safe way: The
                   1602: compiler makes a reasonable guess about the locals visible after a
                   1603: @code{BEGIN}; if it is too pessimistic, the
                   1604: user will get a spurious error about the local not being defined; if the
                   1605: compiler is too optimistic, it will notice this later and issue a
                   1606: warning. In the case above the compiler would complain about @code{x}
                   1607: being undefined at its use. You can see from the obscure examples in
                   1608: this section that it takes quite unusual control structures to get the
                   1609: compiler into trouble, and even then it will often do fine.
                   1611: If the @code{BEGIN} is reachable from above, the most optimistic guess
                   1612: is that all locals visible before the @code{BEGIN} will also be
                   1613: visible after the @code{BEGIN}. This guess is valid for all loops that
                   1614: are entered only through the @code{BEGIN}, in particular, for normal
                   1615: @code{BEGIN}...@code{WHILE}...@code{REPEAT} and
                   1616: @code{BEGIN}...@code{UNTIL} loops and it is implemented in our
                   1617: compiler. When the branch to the @code{BEGIN} is finally generated by
                   1618: @code{AGAIN} or @code{UNTIL}, the compiler checks the guess and
                   1619: warns the user if it was too optimisitic:
                   1620: @example
                   1621: IF
1.4       anton    1622:   @{ x @}
1.2       anton    1623: BEGIN
                   1624:   \ x ? 
                   1625: [ 1 cs-roll ] THEN
                   1626:   ...
                   1627: UNTIL
                   1628: @end example
                   1630: Here, @code{x} lives only until the @code{BEGIN}, but the compiler
                   1631: optimistically assumes that it lives until the @code{THEN}. It notices
                   1632: this difference when it compiles the @code{UNTIL} and issues a
                   1633: warning. The user can avoid the warning, and make sure that @code{x}
                   1634: is not used in the wrong area by using explicit scoping:
                   1635: @example
                   1636: IF
                   1637:   SCOPE
1.4       anton    1638:   @{ x @}
1.2       anton    1639:   ENDSCOPE
                   1640: BEGIN
                   1641: [ 1 cs-roll ] THEN
                   1642:   ...
                   1643: UNTIL
                   1644: @end example
                   1646: Since the guess is optimistic, there will be no spurious error messages
                   1647: about undefined locals.
                   1649: If the @code{BEGIN} is not reachable from above (e.g., after
                   1650: @code{AHEAD} or @code{EXIT}), the compiler cannot even make an
                   1651: optimistic guess, as the locals visible after the @code{BEGIN} may be
                   1652: defined later. Therefore, the compiler assumes that no locals are
1.17      anton    1653: visible after the @code{BEGIN}. However, the user can use
1.2       anton    1654: @code{ASSUME-LIVE} to make the compiler assume that the same locals are
1.17      anton    1655: visible at the BEGIN as at the point where the top control-flow stack
                   1656: item was created.
1.2       anton    1657: 
                   1658: doc-assume-live
                   1660: E.g.,
                   1661: @example
1.4       anton    1662: @{ x @}
1.2       anton    1663: AHEAD
                   1664: ASSUME-LIVE
                   1665: BEGIN
                   1666:   x
                   1667: [ 1 CS-ROLL ] THEN
                   1668:   ...
                   1669: UNTIL
                   1670: @end example
                   1672: Other cases where the locals are defined before the @code{BEGIN} can be
                   1673: handled by inserting an appropriate @code{CS-ROLL} before the
                   1674: @code{ASSUME-LIVE} (and changing the control-flow stack manipulation
                   1675: behind the @code{ASSUME-LIVE}).
                   1677: Cases where locals are defined after the @code{BEGIN} (but should be
                   1678: visible immediately after the @code{BEGIN}) can only be handled by
                   1679: rearranging the loop. E.g., the ``most insidious'' example above can be
                   1680: arranged into:
                   1681: @example
                   1682: BEGIN
1.4       anton    1683:   @{ x @}
1.2       anton    1684:   ... 0=
                   1685: WHILE
                   1686:   x
                   1687: REPEAT
                   1688: @end example
1.17      anton    1690: @node How long do locals live?, Programming Style, Where are locals visible by name?, Gforth locals
1.2       anton    1691: @subsubsection How long do locals live?
                   1693: The right answer for the lifetime question would be: A local lives at
                   1694: least as long as it can be accessed. For a value-flavoured local this
                   1695: means: until the end of its visibility. However, a variable-flavoured
                   1696: local could be accessed through its address far beyond its visibility
                   1697: scope. Ultimately, this would mean that such locals would have to be
                   1698: garbage collected. Since this entails un-Forth-like implementation
                   1699: complexities, I adopted the same cowardly solution as some other
                   1700: languages (e.g., C): The local lives only as long as it is visible;
                   1701: afterwards its address is invalid (and programs that access it
                   1702: afterwards are erroneous).
1.17      anton    1704: @node Programming Style, Implementation, How long do locals live?, Gforth locals
1.2       anton    1705: @subsubsection Programming Style
                   1707: The freedom to define locals anywhere has the potential to change
                   1708: programming styles dramatically. In particular, the need to use the
                   1709: return stack for intermediate storage vanishes. Moreover, all stack
                   1710: manipulations (except @code{PICK}s and @code{ROLL}s with run-time
                   1711: determined arguments) can be eliminated: If the stack items are in the
                   1712: wrong order, just write a locals definition for all of them; then
                   1713: write the items in the order you want.
                   1715: This seems a little far-fetched and eliminating stack manipulations is
1.4       anton    1716: unlikely to become a conscious programming objective. Still, the number
                   1717: of stack manipulations will be reduced dramatically if local variables
1.17      anton    1718: are used liberally (e.g., compare @code{max} in @ref{Gforth locals} with
1.4       anton    1719: a traditional implementation of @code{max}).
1.2       anton    1720: 
                   1721: This shows one potential benefit of locals: making Forth programs more
                   1722: readable. Of course, this benefit will only be realized if the
                   1723: programmers continue to honour the principle of factoring instead of
                   1724: using the added latitude to make the words longer.
                   1726: Using @code{TO} can and should be avoided.  Without @code{TO},
                   1727: every value-flavoured local has only a single assignment and many
                   1728: advantages of functional languages apply to Forth. I.e., programs are
                   1729: easier to analyse, to optimize and to read: It is clear from the
                   1730: definition what the local stands for, it does not turn into something
                   1731: different later.
                   1733: E.g., a definition using @code{TO} might look like this:
                   1734: @example
                   1735: : strcmp @{ addr1 u1 addr2 u2 -- n @}
                   1736:  u1 u2 min 0
                   1737:  ?do
                   1738:    addr1 c@ addr2 c@ - ?dup
                   1739:    if
                   1740:      unloop exit
                   1741:    then
                   1742:    addr1 char+ TO addr1
                   1743:    addr2 char+ TO addr2
                   1744:  loop
                   1745:  u1 u2 - ;
                   1746: @end example
                   1747: Here, @code{TO} is used to update @code{addr1} and @code{addr2} at
                   1748: every loop iteration. @code{strcmp} is a typical example of the
                   1749: readability problems of using @code{TO}. When you start reading
                   1750: @code{strcmp}, you think that @code{addr1} refers to the start of the
                   1751: string. Only near the end of the loop you realize that it is something
                   1752: else.
                   1754: This can be avoided by defining two locals at the start of the loop that
                   1755: are initialized with the right value for the current iteration.
                   1756: @example
                   1757: : strcmp @{ addr1 u1 addr2 u2 -- n @}
                   1758:  addr1 addr2
                   1759:  u1 u2 min 0 
                   1760:  ?do @{ s1 s2 @}
                   1761:    s1 c@ s2 c@ - ?dup 
                   1762:    if
                   1763:      unloop exit
                   1764:    then
                   1765:    s1 char+ s2 char+
                   1766:  loop
                   1767:  2drop
                   1768:  u1 u2 - ;
                   1769: @end example
                   1770: Here it is clear from the start that @code{s1} has a different value
                   1771: in every loop iteration.
1.17      anton    1773: @node Implementation,  , Programming Style, Gforth locals
1.2       anton    1774: @subsubsection Implementation
1.17      anton    1776: Gforth uses an extra locals stack. The most compelling reason for
1.2       anton    1777: this is that the return stack is not float-aligned; using an extra stack
                   1778: also eliminates the problems and restrictions of using the return stack
                   1779: as locals stack. Like the other stacks, the locals stack grows toward
                   1780: lower addresses. A few primitives allow an efficient implementation:
                   1782: doc-@local#
                   1783: doc-f@local#
                   1784: doc-laddr#
                   1785: doc-lp+!#
                   1786: doc-lp!
                   1787: doc->l
                   1788: doc-f>l
                   1790: In addition to these primitives, some specializations of these
                   1791: primitives for commonly occurring inline arguments are provided for
                   1792: efficiency reasons, e.g., @code{@@local0} as specialization of
                   1793: @code{@@local#} for the inline argument 0. The following compiling words
                   1794: compile the right specialized version, or the general version, as
                   1795: appropriate:
1.12      anton    1797: doc-compile-@local
                   1798: doc-compile-f@local
1.2       anton    1799: doc-compile-lp+!
                   1801: Combinations of conditional branches and @code{lp+!#} like
                   1802: @code{?branch-lp+!#} (the locals pointer is only changed if the branch
                   1803: is taken) are provided for efficiency and correctness in loops.
                   1805: A special area in the dictionary space is reserved for keeping the
                   1806: local variable names. @code{@{} switches the dictionary pointer to this
                   1807: area and @code{@}} switches it back and generates the locals
                   1808: initializing code. @code{W:} etc.@ are normal defining words. This
                   1809: special area is cleared at the start of every colon definition.
1.17      anton    1811: A special feature of Gforth's dictionary is used to implement the
1.2       anton    1812: definition of locals without type specifiers: every wordlist (aka
                   1813: vocabulary) has its own methods for searching
1.4       anton    1814: etc. (@pxref{Wordlists}). For the present purpose we defined a wordlist
1.2       anton    1815: with a special search method: When it is searched for a word, it
                   1816: actually creates that word using @code{W:}. @code{@{} changes the search
                   1817: order to first search the wordlist containing @code{@}}, @code{W:} etc.,
                   1818: and then the wordlist for defining locals without type specifiers.
                   1820: The lifetime rules support a stack discipline within a colon
                   1821: definition: The lifetime of a local is either nested with other locals
                   1822: lifetimes or it does not overlap them.
                   1824: At @code{BEGIN}, @code{IF}, and @code{AHEAD} no code for locals stack
                   1825: pointer manipulation is generated. Between control structure words
                   1826: locals definitions can push locals onto the locals stack. @code{AGAIN}
                   1827: is the simplest of the other three control flow words. It has to
                   1828: restore the locals stack depth of the corresponding @code{BEGIN}
                   1829: before branching. The code looks like this:
                   1830: @format
                   1831: @code{lp+!#} current-locals-size @minus{} dest-locals-size
                   1832: @code{branch} <begin>
                   1833: @end format
                   1835: @code{UNTIL} is a little more complicated: If it branches back, it
                   1836: must adjust the stack just like @code{AGAIN}. But if it falls through,
                   1837: the locals stack must not be changed. The compiler generates the
                   1838: following code:
                   1839: @format
                   1840: @code{?branch-lp+!#} <begin> current-locals-size @minus{} dest-locals-size
                   1841: @end format
                   1842: The locals stack pointer is only adjusted if the branch is taken.
                   1844: @code{THEN} can produce somewhat inefficient code:
                   1845: @format
                   1846: @code{lp+!#} current-locals-size @minus{} orig-locals-size
                   1847: <orig target>:
                   1848: @code{lp+!#} orig-locals-size @minus{} new-locals-size
                   1849: @end format
                   1850: The second @code{lp+!#} adjusts the locals stack pointer from the
1.4       anton    1851: level at the @var{orig} point to the level after the @code{THEN}. The
1.2       anton    1852: first @code{lp+!#} adjusts the locals stack pointer from the current
                   1853: level to the level at the orig point, so the complete effect is an
                   1854: adjustment from the current level to the right level after the
                   1855: @code{THEN}.
                   1857: In a conventional Forth implementation a dest control-flow stack entry
                   1858: is just the target address and an orig entry is just the address to be
                   1859: patched. Our locals implementation adds a wordlist to every orig or dest
                   1860: item. It is the list of locals visible (or assumed visible) at the point
                   1861: described by the entry. Our implementation also adds a tag to identify
                   1862: the kind of entry, in particular to differentiate between live and dead
                   1863: (reachable and unreachable) orig entries.
                   1865: A few unusual operations have to be performed on locals wordlists:
                   1867: doc-common-list
                   1868: doc-sub-list?
                   1869: doc-list-size
                   1871: Several features of our locals wordlist implementation make these
                   1872: operations easy to implement: The locals wordlists are organised as
                   1873: linked lists; the tails of these lists are shared, if the lists
                   1874: contain some of the same locals; and the address of a name is greater
                   1875: than the address of the names behind it in the list.
                   1877: Another important implementation detail is the variable
                   1878: @code{dead-code}. It is used by @code{BEGIN} and @code{THEN} to
                   1879: determine if they can be reached directly or only through the branch
                   1880: that they resolve. @code{dead-code} is set by @code{UNREACHABLE},
                   1881: @code{AHEAD}, @code{EXIT} etc., and cleared at the start of a colon
                   1882: definition, by @code{BEGIN} and usually by @code{THEN}.
                   1884: Counted loops are similar to other loops in most respects, but
                   1885: @code{LEAVE} requires special attention: It performs basically the same
                   1886: service as @code{AHEAD}, but it does not create a control-flow stack
                   1887: entry. Therefore the information has to be stored elsewhere;
                   1888: traditionally, the information was stored in the target fields of the
                   1889: branches created by the @code{LEAVE}s, by organizing these fields into a
                   1890: linked list. Unfortunately, this clever trick does not provide enough
                   1891: space for storing our extended control flow information. Therefore, we
                   1892: introduce another stack, the leave stack. It contains the control-flow
                   1893: stack entries for all unresolved @code{LEAVE}s.
                   1895: Local names are kept until the end of the colon definition, even if
                   1896: they are no longer visible in any control-flow path. In a few cases
                   1897: this may lead to increased space needs for the locals name area, but
                   1898: usually less than reclaiming this space would cost in code size.
1.17      anton    1901: @node ANS Forth locals,  , Gforth locals, Locals
1.2       anton    1902: @subsection ANS Forth locals
                   1904: The ANS Forth locals wordset does not define a syntax for locals, but
                   1905: words that make it possible to define various syntaxes. One of the
1.17      anton    1906: possible syntaxes is a subset of the syntax we used in the Gforth locals
1.2       anton    1907: wordset, i.e.:
                   1909: @example
                   1910: @{ local1 local2 ... -- comment @}
                   1911: @end example
                   1912: or
                   1913: @example
                   1914: @{ local1 local2 ... @}
                   1915: @end example
                   1917: The order of the locals corresponds to the order in a stack comment. The
                   1918: restrictions are:
1.1       anton    1919: 
1.2       anton    1920: @itemize @bullet
                   1921: @item
1.17      anton    1922: Locals can only be cell-sized values (no type specifiers are allowed).
1.2       anton    1923: @item
                   1924: Locals can be defined only outside control structures.
                   1925: @item
                   1926: Locals can interfere with explicit usage of the return stack. For the
                   1927: exact (and long) rules, see the standard. If you don't use return stack
1.17      anton    1928: accessing words in a definition using locals, you will be all right. The
1.2       anton    1929: purpose of this rule is to make locals implementation on the return
                   1930: stack easier.
                   1931: @item
                   1932: The whole definition must be in one line.
                   1933: @end itemize
                   1935: Locals defined in this way behave like @code{VALUE}s
1.4       anton    1936: (@xref{Values}). I.e., they are initialized from the stack. Using their
1.2       anton    1937: name produces their value. Their value can be changed using @code{TO}.
1.17      anton    1939: Since this syntax is supported by Gforth directly, you need not do
1.2       anton    1940: anything to use it. If you want to port a program using this syntax to
                   1941: another ANS Forth system, use @file{anslocal.fs} to implement the syntax
                   1942: on the other system.
                   1944: Note that a syntax shown in the standard, section A.13 looks
                   1945: similar, but is quite different in having the order of locals
                   1946: reversed. Beware!
                   1948: The ANS Forth locals wordset itself consists of the following word
                   1950: doc-(local)
                   1952: The ANS Forth locals extension wordset defines a syntax, but it is so
                   1953: awful that we strongly recommend not to use it. We have implemented this
1.17      anton    1954: syntax to make porting to Gforth easy, but do not document it here. The
1.2       anton    1955: problem with this syntax is that the locals are defined in an order
                   1956: reversed with respect to the standard stack comment notation, making
                   1957: programs harder to read, and easier to misread and miswrite. The only
                   1958: merit of this syntax is that it is easy to implement using the ANS Forth
                   1959: locals wordset.
1.3       anton    1960: 
1.4       anton    1961: @node Defining Words, Wordlists, Locals, Words
                   1962: @section Defining Words
1.14      anton    1964: @menu
                   1965: * Values::                      
                   1966: @end menu
1.4       anton    1968: @node Values,  , Defining Words, Defining Words
                   1969: @subsection Values
                   1971: @node Wordlists, Files, Defining Words, Words
                   1972: @section Wordlists
                   1974: @node Files, Blocks, Wordlists, Words
                   1975: @section Files
                   1977: @node Blocks, Other I/O, Files, Words
                   1978: @section Blocks
                   1980: @node Other I/O, Programming Tools, Blocks, Words
                   1981: @section Other I/O
1.18      anton    1983: @node Programming Tools, Assembler and Code words, Other I/O, Words
1.4       anton    1984: @section Programming Tools
1.5       anton    1986: @menu
                   1987: * Debugging::                   Simple and quick.
                   1988: * Assertions::                  Making your programs self-checking.
                   1989: @end menu
                   1991: @node Debugging, Assertions, Programming Tools, Programming Tools
1.4       anton    1992: @subsection Debugging
                   1994: The simple debugging aids provided in @file{debugging.fs}
                   1995: are meant to support a different style of debugging than the
                   1996: tracing/stepping debuggers used in languages with long turn-around
                   1997: times.
                   1999: A much better (faster) way in fast-compilig languages is to add
                   2000: printing code at well-selected places, let the program run, look at
                   2001: the output, see where things went wrong, add more printing code, etc.,
                   2002: until the bug is found.
                   2004: The word @code{~~} is easy to insert. It just prints debugging
                   2005: information (by default the source location and the stack contents). It
                   2006: is also easy to remove (@kbd{C-x ~} in the Emacs Forth mode to
                   2007: query-replace them with nothing). The deferred words
                   2008: @code{printdebugdata} and @code{printdebugline} control the output of
                   2009: @code{~~}. The default source location output format works well with
                   2010: Emacs' compilation mode, so you can step through the program at the
1.5       anton    2011: source level using @kbd{C-x `} (the advantage over a stepping debugger
                   2012: is that you can step in any direction and you know where the crash has
                   2013: happened or where the strange data has occurred).
1.4       anton    2014: 
                   2015: Note that the default actions clobber the contents of the pictured
                   2016: numeric output string, so you should not use @code{~~}, e.g., between
                   2017: @code{<#} and @code{#>}.
                   2019: doc-~~
                   2020: doc-printdebugdata
                   2021: doc-printdebugline
1.5       anton    2023: @node Assertions,  , Debugging, Programming Tools
1.4       anton    2024: @subsection Assertions
1.5       anton    2026: It is a good idea to make your programs self-checking, in particular, if
                   2027: you use an assumption (e.g., that a certain field of a data structure is
1.17      anton    2028: never zero) that may become wrong during maintenance. Gforth supports
1.5       anton    2029: assertions for this purpose. They are used like this:
                   2031: @example
                   2032: assert( @var{flag} )
                   2033: @end example
                   2035: The code between @code{assert(} and @code{)} should compute a flag, that
                   2036: should be true if everything is alright and false otherwise. It should
                   2037: not change anything else on the stack. The overall stack effect of the
                   2038: assertion is @code{( -- )}. E.g.
                   2040: @example
                   2041: assert( 1 1 + 2 = ) \ what we learn in school
                   2042: assert( dup 0<> ) \ assert that the top of stack is not zero
                   2043: assert( false ) \ this code should not be reached
                   2044: @end example
                   2046: The need for assertions is different at different times. During
                   2047: debugging, we want more checking, in production we sometimes care more
                   2048: for speed. Therefore, assertions can be turned off, i.e., the assertion
                   2049: becomes a comment. Depending on the importance of an assertion and the
                   2050: time it takes to check it, you may want to turn off some assertions and
1.17      anton    2051: keep others turned on. Gforth provides several levels of assertions for
1.5       anton    2052: this purpose:
                   2054: doc-assert0(
                   2055: doc-assert1(
                   2056: doc-assert2(
                   2057: doc-assert3(
                   2058: doc-assert(
                   2059: doc-)
                   2061: @code{Assert(} is the same as @code{assert1(}. The variable
                   2062: @code{assert-level} specifies the highest assertions that are turned
                   2063: on. I.e., at the default @code{assert-level} of one, @code{assert0(} and
                   2064: @code{assert1(} assertions perform checking, while @code{assert2(} and
                   2065: @code{assert3(} assertions are treated as comments.
                   2067: Note that the @code{assert-level} is evaluated at compile-time, not at
                   2068: run-time. I.e., you cannot turn assertions on or off at run-time, you
                   2069: have to set the @code{assert-level} appropriately before compiling a
                   2070: piece of code. You can compile several pieces of code at several
                   2071: @code{assert-level}s (e.g., a trusted library at level 1 and newly
                   2072: written code at level 3).
                   2074: doc-assert-level
                   2076: If an assertion fails, a message compatible with Emacs' compilation mode
                   2077: is produced and the execution is aborted (currently with @code{ABORT"}.
                   2078: If there is interest, we will introduce a special throw code. But if you
                   2079: intend to @code{catch} a specific condition, using @code{throw} is
                   2080: probably more appropriate than an assertion).
1.18      anton    2082: @node Assembler and Code words, Threading Words, Programming Tools, Words
                   2083: @section Assembler and Code words
                   2085: Gforth provides some words for defining primitives (words written in
                   2086: machine code), and for defining the the machine-code equivalent of
                   2087: @code{DOES>}-based defining words. However, the machine-independent
                   2088: nature of Gforth poses a few problems: First of all. Gforth runs on
                   2089: several architectures, so it can provide no standard assembler. What's
                   2090: worse is that the register allocation not only depends on the processor,
                   2091: but also on the gcc version and options used.
                   2093: The words Gforth offers encapsulate some system dependences (e.g., the
                   2094: header structure), so a system-independent assembler may be used in
                   2095: Gforth. If you do not have an assembler, you can compile machine code
                   2096: directly with @code{,} and @code{c,}.
                   2098: doc-assembler
                   2099: doc-code
                   2100: doc-end-code
                   2101: doc-;code
                   2102: doc-flush-icache
                   2104: If @code{flush-icache} does not work correctly, @code{code} words
                   2105: etc. will not work (reliably), either.
                   2107: These words are rarely used. Therefore they reside in @code{code.fs},
                   2108: which is usually not loaded (except @code{flush-icache}, which is always
1.19      anton    2109: present). You can load them with @code{require code.fs}.
1.18      anton    2110: 
                   2111: Another option for implementing normal and defining words efficiently
                   2112: is: adding the wanted functionality to the source of Gforth. For normal
                   2113: words you just have to edit @file{primitives}, defining words (for fast
                   2114: defined words) probably require changes in @file{engine.c},
                   2115: @file{kernal.fs}, @file{prims2x.fs}, and possibly @file{cross.fs}.
                   2118: @node Threading Words,  , Assembler and Code words, Words
1.4       anton    2119: @section Threading Words
                   2121: These words provide access to code addresses and other threading stuff
1.17      anton    2122: in Gforth (and, possibly, other interpretive Forths). It more or less
1.4       anton    2123: abstracts away the differences between direct and indirect threading
                   2124: (and, for direct threading, the machine dependences). However, at
                   2125: present this wordset is still inclomplete. It is also pretty low-level;
                   2126: some day it will hopefully be made unnecessary by an internals words set
                   2127: that abstracts implementation details away completely.
                   2129: doc->code-address
                   2130: doc->does-code
                   2131: doc-code-address!
                   2132: doc-does-code!
                   2133: doc-does-handler!
                   2134: doc-/does-handler
1.18      anton    2136: The code addresses produced by various defining words are produced by
                   2137: the following words:
1.14      anton    2138: 
1.18      anton    2139: doc-docol:
                   2140: doc-docon:
                   2141: doc-dovar:
                   2142: doc-douser:
                   2143: doc-dodefer:
                   2144: doc-dofield:
                   2146: Currently there is no installation-independent way for recogizing words
                   2147: defined by a @code{CREATE}...@code{DOES>} word; however, once you know
                   2148: that a word is defined by a @code{CREATE}...@code{DOES>} word, you can
                   2149: use @code{>DOES-CODE}.
1.14      anton    2150: 
1.4       anton    2151: @node ANS conformance, Model, Words, Top
                   2152: @chapter ANS conformance
1.17      anton    2154: To the best of our knowledge, Gforth is an
1.14      anton    2155: 
1.15      anton    2156: ANS Forth System
                   2157: @itemize
                   2158: @item providing the Core Extensions word set
                   2159: @item providing the Block word set
                   2160: @item providing the Block Extensions word set
                   2161: @item providing the Double-Number word set
                   2162: @item providing the Double-Number Extensions word set
                   2163: @item providing the Exception word set
                   2164: @item providing the Exception Extensions word set
                   2165: @item providing the Facility word set
                   2166: @item providing @code{MS} and @code{TIME&DATE} from the Facility Extensions word set
                   2167: @item providing the File Access word set
                   2168: @item providing the File Access Extensions word set
                   2169: @item providing the Floating-Point word set
                   2170: @item providing the Floating-Point Extensions word set
                   2171: @item providing the Locals word set
                   2172: @item providing the Locals Extensions word set
                   2173: @item providing the Memory-Allocation word set
                   2174: @item providing the Memory-Allocation Extensions word set (that one's easy)
                   2175: @item providing the Programming-Tools word set
1.18      anton    2176: @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
1.15      anton    2177: @item providing the Search-Order word set
                   2178: @item providing the Search-Order Extensions word set
                   2179: @item providing the String word set
                   2180: @item providing the String Extensions word set (another easy one)
                   2181: @end itemize
                   2183: In addition, ANS Forth systems are required to document certain
                   2184: implementation choices. This chapter tries to meet these
                   2185: requirements. In many cases it gives a way to ask the system for the
                   2186: information instead of providing the information directly, in
                   2187: particular, if the information depends on the processor, the operating
                   2188: system or the installation options chosen, or if they are likely to
1.17      anton    2189: change during the maintenance of Gforth.
1.15      anton    2190: 
1.14      anton    2191: @comment The framework for the rest has been taken from pfe.
                   2193: @menu
                   2194: * The Core Words::              
                   2195: * The optional Block word set::  
                   2196: * The optional Double Number word set::  
                   2197: * The optional Exception word set::  
                   2198: * The optional Facility word set::  
                   2199: * The optional File-Access word set::  
                   2200: * The optional Floating-Point word set::  
                   2201: * The optional Locals word set::  
                   2202: * The optional Memory-Allocation word set::  
                   2203: * The optional Programming-Tools word set::  
                   2204: * The optional Search-Order word set::  
                   2205: @end menu
                   2208: @c =====================================================================
                   2209: @node The Core Words, The optional Block word set, ANS conformance, ANS conformance
                   2210: @comment  node-name,  next,  previous,  up
                   2211: @section The Core Words
                   2212: @c =====================================================================
                   2214: @menu
1.15      anton    2215: * core-idef::                   Implementation Defined Options                   
                   2216: * core-ambcond::                Ambiguous Conditions                
                   2217: * core-other::                  Other System Documentation                  
1.14      anton    2218: @end menu
                   2220: @c ---------------------------------------------------------------------
                   2221: @node core-idef, core-ambcond, The Core Words, The Core Words
                   2222: @subsection Implementation Defined Options
                   2223: @c ---------------------------------------------------------------------
                   2225: @table @i
                   2227: @item (Cell) aligned addresses:
1.17      anton    2228: processor-dependent. Gforth's alignment words perform natural alignment
1.14      anton    2229: (e.g., an address aligned for a datum of size 8 is divisible by
                   2230: 8). Unaligned accesses usually result in a @code{-23 THROW}.
                   2232: @item @code{EMIT} and non-graphic characters:
                   2233: The character is output using the C library function (actually, macro)
                   2234: @code{putchar}.
                   2236: @item character editing of @code{ACCEPT} and @code{EXPECT}:
                   2237: This is modeled on the GNU readline library (@pxref{Readline
                   2238: Interaction, , Command Line Editing, readline, The GNU Readline
                   2239: Library}) with Emacs-like key bindings. @kbd{Tab} deviates a little by
                   2240: producing a full word completion every time you type it (instead of
                   2241: producing the common prefix of all completions).
                   2243: @item character set:
                   2244: The character set of your computer and display device. Gforth is
                   2245: 8-bit-clean (but some other component in your system may make trouble).
                   2247: @item Character-aligned address requirements:
                   2248: installation-dependent. Currently a character is represented by a C
                   2249: @code{unsigned char}; in the future we might switch to @code{wchar_t}
                   2250: (Comments on that requested).
                   2252: @item character-set extensions and matching of names:
1.17      anton    2253: Any character except the ASCII NUL charcter can be used in a
                   2254: name. Matching is case-insensitive. The matching is performed using the
                   2255: C function @code{strncasecmp}, whose function is probably influenced by
                   2256: the locale. E.g., the @code{C} locale does not know about accents and
1.14      anton    2257: umlauts, so they are matched case-sensitively in that locale. For
                   2258: portability reasons it is best to write programs such that they work in
                   2259: the @code{C} locale. Then one can use libraries written by a Polish
                   2260: programmer (who might use words containing ISO Latin-2 encoded
                   2261: characters) and by a French programmer (ISO Latin-1) in the same program
                   2262: (of course, @code{WORDS} will produce funny results for some of the
                   2263: words (which ones, depends on the font you are using)). Also, the locale
                   2264: you prefer may not be available in other operating systems. Hopefully,
                   2265: Unicode will solve these problems one day.
                   2267: @item conditions under which control characters match a space delimiter:
                   2268: If @code{WORD} is called with the space character as a delimiter, all
                   2269: white-space characters (as identified by the C macro @code{isspace()})
                   2270: are delimiters. @code{PARSE}, on the other hand, treats space like other
                   2271: delimiters. @code{PARSE-WORD} treats space like @code{WORD}, but behaves
                   2272: like @code{PARSE} otherwise. @code{(NAME)}, which is used by the outer
                   2273: interpreter (aka text interpreter) by default, treats all white-space
                   2274: characters as delimiters.
                   2276: @item format of the control flow stack:
                   2277: The data stack is used as control flow stack. The size of a control flow
                   2278: stack item in cells is given by the constant @code{cs-item-size}. At the
                   2279: time of this writing, an item consists of a (pointer to a) locals list
                   2280: (third), an address in the code (second), and a tag for identifying the
                   2281: item (TOS). The following tags are used: @code{defstart},
                   2282: @code{live-orig}, @code{dead-orig}, @code{dest}, @code{do-dest},
                   2283: @code{scopestart}.
                   2285: @item conversion of digits > 35
                   2286: The characters @code{[\]^_'} are the digits with the decimal value
                   2287: 36@minus{}41. There is no way to input many of the larger digits.
                   2289: @item display after input terminates in @code{ACCEPT} and @code{EXPECT}:
                   2290: The cursor is moved to the end of the entered string. If the input is
                   2291: terminated using the @kbd{Return} key, a space is typed.
                   2293: @item exception abort sequence of @code{ABORT"}:
                   2294: The error string is stored into the variable @code{"error} and a
                   2295: @code{-2 throw} is performed.
                   2297: @item input line terminator:
                   2298: For interactive input, @kbd{C-m} and @kbd{C-j} terminate lines. One of
                   2299: these characters is typically produced when you type the @kbd{Enter} or
                   2300: @kbd{Return} key.
                   2302: @item maximum size of a counted string:
                   2303: @code{s" /counted-string" environment? drop .}. Currently 255 characters
                   2304: on all ports, but this may change.
                   2306: @item maximum size of a parsed string:
                   2307: Given by the constant @code{/line}. Currently 255 characters.
                   2309: @item maximum size of a definition name, in characters:
                   2310: 31
                   2312: @item maximum string length for @code{ENVIRONMENT?}, in characters:
                   2313: 31
                   2315: @item method of selecting the user input device:
1.17      anton    2316: The user input device is the standard input. There is currently no way to
                   2317: change it from within Gforth. However, the input can typically be
                   2318: redirected in the command line that starts Gforth.
1.14      anton    2319: 
                   2320: @item method of selecting the user output device:
                   2321: The user output device is the standard output. It cannot be redirected
1.17      anton    2322: from within Gforth, but typically from the command line that starts
                   2323: Gforth. Gforth uses buffered output, so output on a terminal does not
1.14      anton    2324: become visible before the next newline or buffer overflow. Output on
                   2325: non-terminals is invisible until the buffer overflows.
                   2327: @item methods of dictionary compilation:
1.17      anton    2328: What are we expected to document here?
1.14      anton    2329: 
                   2330: @item number of bits in one address unit:
                   2331: @code{s" address-units-bits" environment? drop .}. 8 in all current
                   2332: ports.
                   2334: @item number representation and arithmetic:
                   2335: Processor-dependent. Binary two's complement on all current ports.
                   2337: @item ranges for integer types:
                   2338: Installation-dependent. Make environmental queries for @code{MAX-N},
                   2339: @code{MAX-U}, @code{MAX-D} and @code{MAX-UD}. The lower bounds for
                   2340: unsigned (and positive) types is 0. The lower bound for signed types on
                   2341: two's complement and one's complement machines machines can be computed
                   2342: by adding 1 to the upper bound.
                   2344: @item read-only data space regions:
                   2345: The whole Forth data space is writable.
                   2347: @item size of buffer at @code{WORD}:
                   2348: @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
                   2349: shared with the pictured numeric output string. If overwriting
                   2350: @code{PAD} is acceptable, it is as large as the remaining dictionary
                   2351: space, although only as much can be sensibly used as fits in a counted
                   2352: string.
                   2354: @item size of one cell in address units:
                   2355: @code{1 cells .}.
                   2357: @item size of one character in address units:
                   2358: @code{1 chars .}. 1 on all current ports.
                   2360: @item size of the keyboard terminal buffer:
                   2361: Varies. You can determine the size at a specific time using @code{lp@
                   2362: tib - .}. It is shared with the locals stack and TIBs of files that
                   2363: include the current file. You can change the amount of space for TIBs
1.17      anton    2364: and locals stack at Gforth startup with the command line option
1.14      anton    2365: @code{-l}.
                   2367: @item size of the pictured numeric output buffer:
                   2368: @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
                   2369: shared with @code{WORD}.
                   2371: @item size of the scratch area returned by @code{PAD}:
                   2372: The remainder of dictionary space. You can even use the unused part of
                   2373: the data stack space. The current size can be computed with @code{sp@
                   2374: pad - .}.
                   2376: @item system case-sensitivity characteristics:
                   2377: Dictionary searches are case insensitive. However, as explained above
                   2378: under @i{character-set extensions}, the matching for non-ASCII
                   2379: characters is determined by the locale you are using. In the default
                   2380: @code{C} locale all non-ASCII characters are matched case-sensitively.
                   2382: @item system prompt:
                   2383: @code{ ok} in interpret state, @code{ compiled} in compile state.
                   2385: @item division rounding:
                   2386: installation dependent. @code{s" floored" environment? drop .}. We leave
                   2387: the choice to gcc (what to use for @code{/}) and to you (whether to use
                   2388: @code{fm/mod}, @code{sm/rem} or simply @code{/}).
                   2390: @item values of @code{STATE} when true:
                   2391: -1.
                   2393: @item values returned after arithmetic overflow:
                   2394: On two's complement machines, arithmetic is performed modulo
                   2395: 2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
                   2396: arithmetic (with appropriate mapping for signed types). Division by zero
                   2397: typically results in a @code{-55 throw} (floatingpoint unidentified
                   2398: fault), although a @code{-10 throw} (divide by zero) would be more
                   2399: appropriate.
                   2401: @item whether the current definition can be found after @t{DOES>}:
                   2402: No.
                   2404: @end table
                   2406: @c ---------------------------------------------------------------------
                   2407: @node core-ambcond, core-other, core-idef, The Core Words
                   2408: @subsection Ambiguous conditions
                   2409: @c ---------------------------------------------------------------------
                   2411: @table @i
                   2413: @item a name is neither a word nor a number:
                   2414: @code{-13 throw} (Undefined word)
                   2416: @item a definition name exceeds the maximum length allowed:
                   2417: @code{-19 throw} (Word name too long)
                   2419: @item addressing a region not inside the various data spaces of the forth system:
                   2420: The stacks, code space and name space are accessible. Machine code space is
                   2421: typically readable. Accessing other addresses gives results dependent on
                   2422: the operating system. On decent systems: @code{-9 throw} (Invalid memory
                   2423: address).
                   2425: @item argument type incompatible with parameter:
                   2426: This is usually not caught. Some words perform checks, e.g., the control
                   2427: flow words, and issue a @code{ABORT"} or @code{-12 THROW} (Argument type
                   2428: mismatch).
                   2430: @item attempting to obtain the execution token of a word with undefined execution semantics:
                   2431: You get an execution token representing the compilation semantics
                   2432: instead.
                   2434: @item dividing by zero:
                   2435: typically results in a @code{-55 throw} (floating point unidentified
                   2436: fault), although a @code{-10 throw} (divide by zero) would be more
                   2437: appropriate.
                   2439: @item insufficient data stack or return stack space:
                   2440: Not checked. This typically results in mysterious illegal memory
                   2441: accesses, producing @code{-9 throw} (Invalid memory address) or
                   2442: @code{-23 throw} (Address alignment exception).
                   2444: @item insufficient space for loop control parameters:
                   2445: like other return stack overflows.
                   2447: @item insufficient space in the dictionary:
                   2448: Not checked. Similar results as stack overflows. However, typically the
                   2449: error appears at a different place when one inserts or removes code.
                   2451: @item interpreting a word with undefined interpretation semantics:
                   2452: For some words, we defined interpretation semantics. For the others:
                   2453: @code{-14 throw} (Interpreting a compile-only word). Note that this is
                   2454: checked only by the outer (aka text) interpreter; if the word is
                   2455: @code{execute}d in some other way, it will typically perform it's
                   2456: compilation semantics even in interpret state. (We could change @code{'}
                   2457: and relatives not to give the xt of such words, but we think that would
                   2458: be too restrictive).
                   2460: @item modifying the contents of the input buffer or a string literal:
                   2461: These are located in writable memory and can be modified.
                   2463: @item overflow of the pictured numeric output string:
                   2464: Not checked.
                   2466: @item parsed string overflow:
                   2467: @code{PARSE} cannot overflow. @code{WORD} does not check for overflow.
                   2469: @item producing a result out of range:
                   2470: On two's complement machines, arithmetic is performed modulo
                   2471: 2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
                   2472: arithmetic (with appropriate mapping for signed types). Division by zero
                   2473: typically results in a @code{-55 throw} (floatingpoint unidentified
                   2474: fault), although a @code{-10 throw} (divide by zero) would be more
                   2475: appropriate. @code{convert} and @code{>number} currently overflow
                   2476: silently.
                   2478: @item reading from an empty data or return stack:
                   2479: The data stack is checked by the outer (aka text) interpreter after
                   2480: every word executed. If it has underflowed, a @code{-4 throw} (Stack
                   2481: underflow) is performed. Apart from that, the stacks are not checked and
                   2482: underflows can result in similar behaviour as overflows (of adjacent
                   2483: stacks).
                   2485: @item unexepected end of the input buffer, resulting in an attempt to use a zero-length string as a name:
                   2486: @code{Create} and its descendants perform a @code{-16 throw} (Attempt to
                   2487: use zero-length string as a name). Words like @code{'} probably will not
                   2488: find what they search. Note that it is possible to create zero-length
                   2489: names with @code{nextname} (should it not?).
                   2491: @item @code{>IN} greater than input buffer:
                   2492: The next invocation of a parsing word returns a string wih length 0.
                   2494: @item @code{RECURSE} appears after @code{DOES>}:
                   2495: Compiles a recursive call to the defining word not to the defined word.
                   2497: @item argument input source different than current input source for @code{RESTORE-INPUT}:
                   2498: !!???If the argument input source is a valid input source then it gets
1.19      anton    2499: restored. Otherwise causes @code{-12 THROW}, which, unless caught, issues
1.14      anton    2500: the message "argument type mismatch" and aborts.
                   2502: @item data space containing definitions gets de-allocated:
                   2503: Deallocation with @code{allot} is not checked. This typically resuls in
                   2504: memory access faults or execution of illegal instructions.
                   2506: @item data space read/write with incorrect alignment:
                   2507: Processor-dependent. Typically results in a @code{-23 throw} (Address
                   2508: alignment exception). Under Linux on a 486 or later processor with
                   2509: alignment turned on, incorrect alignment results in a @code{-9 throw}
                   2510: (Invalid memory address). There are reportedly some processors with
                   2511: alignment restrictions that do not report them.
                   2513: @item data space pointer not properly aligned, @code{,}, @code{C,}:
                   2514: Like other alignment errors.
                   2516: @item less than u+2 stack items (@code{PICK} and @code{ROLL}):
                   2517: Not checked. May cause an illegal memory access.
                   2519: @item loop control parameters not available:
                   2520: Not checked. The counted loop words simply assume that the top of return
                   2521: stack items are loop control parameters and behave accordingly.
                   2523: @item most recent definition does not have a name (@code{IMMEDIATE}):
                   2524: @code{abort" last word was headerless"}.
                   2526: @item name not defined by @code{VALUE} used by @code{TO}:
                   2527: @code{-32 throw} (Invalid name argument)
1.15      anton    2529: @item name not found (@code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]}):
1.14      anton    2530: @code{-13 throw} (Undefined word)
                   2532: @item parameters are not of the same type (@code{DO}, @code{?DO}, @code{WITHIN}):
                   2533: Gforth behaves as if they were of the same type. I.e., you can predict
                   2534: the behaviour by interpreting all parameters as, e.g., signed.
                   2536: @item @code{POSTPONE} or @code{[COMPILE]} applied to @code{TO}:
                   2537: Assume @code{: X POSTPONE TO ; IMMEDIATE}. @code{X} is equivalent to
                   2538: @code{TO}.
                   2540: @item String longer than a counted string returned by @code{WORD}:
                   2541: Not checked. The string will be ok, but the count will, of course,
                   2542: contain only the least significant bits of the length.
1.15      anton    2544: @item u greater than or equal to the number of bits in a cell (@code{LSHIFT}, @code{RSHIFT}):
1.14      anton    2545: Processor-dependent. Typical behaviours are returning 0 and using only
                   2546: the low bits of the shift count.
                   2548: @item word not defined via @code{CREATE}:
                   2549: @code{>BODY} produces the PFA of the word no matter how it was defined.
                   2551: @code{DOES>} changes the execution semantics of the last defined word no
                   2552: matter how it was defined. E.g., @code{CONSTANT DOES>} is equivalent to
                   2553: @code{CREATE , DOES>}.
                   2555: @item words improperly used outside @code{<#} and @code{#>}:
                   2556: Not checked. As usual, you can expect memory faults.
                   2558: @end table
                   2561: @c ---------------------------------------------------------------------
                   2562: @node core-other,  , core-ambcond, The Core Words
                   2563: @subsection Other system documentation
                   2564: @c ---------------------------------------------------------------------
                   2566: @table @i
                   2568: @item nonstandard words using @code{PAD}:
                   2569: None.
                   2571: @item operator's terminal facilities available:
                   2572: !!??
                   2574: @item program data space available:
                   2575: @code{sp@ here - .} gives the space remaining for dictionary and data
                   2576: stack together.
                   2578: @item return stack space available:
                   2579: !!??
                   2581: @item stack space available:
                   2582: @code{sp@ here - .} gives the space remaining for dictionary and data
                   2583: stack together.
                   2585: @item system dictionary space required, in address units:
                   2586: Type @code{here forthstart - .} after startup. At the time of this
                   2587: writing, this gives 70108 (bytes) on a 32-bit system.
                   2588: @end table
                   2591: @c =====================================================================
                   2592: @node The optional Block word set, The optional Double Number word set, The Core Words, ANS conformance
                   2593: @section The optional Block word set
                   2594: @c =====================================================================
                   2596: @menu
1.15      anton    2597: * block-idef::                  Implementation Defined Options                  
                   2598: * block-ambcond::               Ambiguous Conditions               
                   2599: * block-other::                 Other System Documentation                 
1.14      anton    2600: @end menu
                   2603: @c ---------------------------------------------------------------------
                   2604: @node block-idef, block-ambcond, The optional Block word set, The optional Block word set
                   2605: @subsection Implementation Defined Options
                   2606: @c ---------------------------------------------------------------------
                   2608: @table @i
                   2610: @item the format for display by @code{LIST}:
                   2611: First the screen number is displayed, then 16 lines of 64 characters,
                   2612: each line preceded by the line number.
                   2614: @item the length of a line affected by @code{\}:
                   2615: 64 characters.
                   2616: @end table
                   2619: @c ---------------------------------------------------------------------
                   2620: @node block-ambcond, block-other, block-idef, The optional Block word set
                   2621: @subsection Ambiguous conditions
                   2622: @c ---------------------------------------------------------------------
                   2624: @table @i
                   2626: @item correct block read was not possible:
                   2627: Typically results in a @code{throw} of some OS-derived value (between
                   2628: -512 and -2048). If the blocks file was just not long enough, blanks are
                   2629: supplied for the missing portion.
                   2631: @item I/O exception in block transfer:
                   2632: Typically results in a @code{throw} of some OS-derived value (between
                   2633: -512 and -2048).
                   2635: @item invalid block number:
                   2636: @code{-35 throw} (Invalid block number)
                   2638: @item a program directly alters the contents of @code{BLK}:
                   2639: The input stream is switched to that other block, at the same
                   2640: position. If the storing to @code{BLK} happens when interpreting
                   2641: non-block input, the system will get quite confused when the block ends.
                   2643: @item no current block buffer for @code{UPDATE}:
                   2644: @code{UPDATE} has no effect.
                   2646: @end table
                   2649: @c ---------------------------------------------------------------------
                   2650: @node block-other,  , block-ambcond, The optional Block word set
                   2651: @subsection Other system documentation
                   2652: @c ---------------------------------------------------------------------
                   2654: @table @i
                   2656: @item any restrictions a multiprogramming system places on the use of buffer addresses:
                   2657: No restrictions (yet).
                   2659: @item the number of blocks available for source and data:
                   2660: depends on your disk space.
                   2662: @end table
                   2665: @c =====================================================================
                   2666: @node The optional Double Number word set, The optional Exception word set, The optional Block word set, ANS conformance
                   2667: @section The optional Double Number word set
                   2668: @c =====================================================================
                   2670: @menu
1.15      anton    2671: * double-ambcond::              Ambiguous Conditions              
1.14      anton    2672: @end menu
                   2675: @c ---------------------------------------------------------------------
1.15      anton    2676: @node double-ambcond,  , The optional Double Number word set, The optional Double Number word set
1.14      anton    2677: @subsection Ambiguous conditions
                   2678: @c ---------------------------------------------------------------------
                   2680: @table @i
1.15      anton    2682: @item @var{d} outside of range of @var{n} in @code{D>S}:
1.14      anton    2683: The least significant cell of @var{d} is produced.
                   2685: @end table
                   2688: @c =====================================================================
                   2689: @node The optional Exception word set, The optional Facility word set, The optional Double Number word set, ANS conformance
                   2690: @section The optional Exception word set
                   2691: @c =====================================================================
                   2693: @menu
1.15      anton    2694: * exception-idef::              Implementation Defined Options              
1.14      anton    2695: @end menu
                   2698: @c ---------------------------------------------------------------------
1.15      anton    2699: @node exception-idef,  , The optional Exception word set, The optional Exception word set
1.14      anton    2700: @subsection Implementation Defined Options
                   2701: @c ---------------------------------------------------------------------
                   2703: @table @i
                   2704: @item @code{THROW}-codes used in the system:
                   2705: The codes -256@minus{}-511 are used for reporting signals (see
                   2706: @file{errore.fs}). The codes -512@minus{}-2047 are used for OS errors
                   2707: (for file and memory allocation operations). The mapping from OS error
                   2708: numbers to throw code is -512@minus{}@var{errno}. One side effect of
                   2709: this mapping is that undefined OS errors produce a message with a
                   2710: strange number; e.g., @code{-1000 THROW} results in @code{Unknown error
                   2711: 488} on my system.
                   2712: @end table
                   2714: @c =====================================================================
                   2715: @node The optional Facility word set, The optional File-Access word set, The optional Exception word set, ANS conformance
                   2716: @section The optional Facility word set
                   2717: @c =====================================================================
                   2719: @menu
1.15      anton    2720: * facility-idef::               Implementation Defined Options               
                   2721: * facility-ambcond::            Ambiguous Conditions            
1.14      anton    2722: @end menu
                   2725: @c ---------------------------------------------------------------------
                   2726: @node facility-idef, facility-ambcond, The optional Facility word set, The optional Facility word set
                   2727: @subsection Implementation Defined Options
                   2728: @c ---------------------------------------------------------------------
                   2730: @table @i
                   2732: @item encoding of keyboard events (@code{EKEY}):
                   2733: Not yet implemeted.
                   2735: @item duration of a system clock tick
                   2736: System dependent. With respect to @code{MS}, the time is specified in
                   2737: microseconds. How well the OS and the hardware implement this, is
                   2738: another question.
                   2740: @item repeatability to be expected from the execution of @code{MS}:
                   2741: System dependent. On Unix, a lot depends on load. If the system is
1.17      anton    2742: lightly loaded, and the delay is short enough that Gforth does not get
1.14      anton    2743: swapped out, the performance should be acceptable. Under MS-DOS and
                   2744: other single-tasking systems, it should be good.
                   2746: @end table
                   2749: @c ---------------------------------------------------------------------
1.15      anton    2750: @node facility-ambcond,  , facility-idef, The optional Facility word set
1.14      anton    2751: @subsection Ambiguous conditions
                   2752: @c ---------------------------------------------------------------------
                   2754: @table @i
                   2756: @item @code{AT-XY} can't be performed on user output device:
                   2757: Largely terminal dependant. No range checks are done on the arguments.
                   2758: No errors are reported. You may see some garbage appearing, you may see
                   2759: simply nothing happen.
                   2761: @end table
                   2764: @c =====================================================================
                   2765: @node The optional File-Access word set, The optional Floating-Point word set, The optional Facility word set, ANS conformance
                   2766: @section The optional File-Access word set
                   2767: @c =====================================================================
                   2769: @menu
1.15      anton    2770: * file-idef::                   Implementation Defined Options                   
                   2771: * file-ambcond::                Ambiguous Conditions                
1.14      anton    2772: @end menu
                   2775: @c ---------------------------------------------------------------------
                   2776: @node file-idef, file-ambcond, The optional File-Access word set, The optional File-Access word set
                   2777: @subsection Implementation Defined Options
                   2778: @c ---------------------------------------------------------------------
                   2780: @table @i
                   2782: @item File access methods used:
                   2783: @code{R/O}, @code{R/W} and @code{BIN} work as you would
                   2784: expect. @code{W/O} translates into the C file opening mode @code{w} (or
                   2785: @code{wb}): The file is cleared, if it exists, and created, if it does
1.15      anton    2786: not (both with @code{open-file} and @code{create-file}).  Under Unix
1.14      anton    2787: @code{create-file} creates a file with 666 permissions modified by your
                   2788: umask.
                   2790: @item file exceptions:
                   2791: The file words do not raise exceptions (except, perhaps, memory access
                   2792: faults when you pass illegal addresses or file-ids).
                   2794: @item file line terminator:
                   2795: System-dependent. Gforth uses C's newline character as line
                   2796: terminator. What the actual character code(s) of this are is
                   2797: system-dependent.
                   2799: @item file name format
                   2800: System dependent. Gforth just uses the file name format of your OS.
                   2802: @item information returned by @code{FILE-STATUS}:
                   2803: @code{FILE-STATUS} returns the most powerful file access mode allowed
                   2804: for the file: Either @code{R/O}, @code{W/O} or @code{R/W}. If the file
                   2805: cannot be accessed, @code{R/O BIN} is returned. @code{BIN} is applicable
                   2806: along with the retured mode.
                   2808: @item input file state after an exception when including source:
                   2809: All files that are left via the exception are closed.
                   2811: @item @var{ior} values and meaning:
1.15      anton    2812: The @var{ior}s returned by the file and memory allocation words are
                   2813: intended as throw codes. They typically are in the range
                   2814: -512@minus{}-2047 of OS errors.  The mapping from OS error numbers to
                   2815: @var{ior}s is -512@minus{}@var{errno}.
1.14      anton    2816: 
                   2817: @item maximum depth of file input nesting:
                   2818: limited by the amount of return stack, locals/TIB stack, and the number
                   2819: of open files available. This should not give you troubles.
                   2821: @item maximum size of input line:
                   2822: @code{/line}. Currently 255.
                   2824: @item methods of mapping block ranges to files:
                   2825: Currently, the block words automatically access the file
                   2826: @file{blocks.fb} in the currend working directory. More sophisticated
                   2827: methods could be implemented if there is demand (and a volunteer).
                   2829: @item number of string buffers provided by @code{S"}:
                   2830: 1
                   2832: @item size of string buffer used by @code{S"}:
                   2833: @code{/line}. currently 255.
                   2835: @end table
                   2837: @c ---------------------------------------------------------------------
1.15      anton    2838: @node file-ambcond,  , file-idef, The optional File-Access word set
1.14      anton    2839: @subsection Ambiguous conditions
                   2840: @c ---------------------------------------------------------------------
                   2842: @table @i
                   2844: @item attempting to position a file outside it's boundaries:
                   2845: @code{REPOSITION-FILE} is performed as usual: Afterwards,
                   2846: @code{FILE-POSITION} returns the value given to @code{REPOSITION-FILE}.
                   2848: @item attempting to read from file positions not yet written:
                   2849: End-of-file, i.e., zero characters are read and no error is reported.
                   2851: @item @var{file-id} is invalid (@code{INCLUDE-FILE}):
                   2852: An appropriate exception may be thrown, but a memory fault or other
                   2853: problem is more probable.
                   2855: @item I/O exception reading or closing @var{file-id} (@code{include-file}, @code{included}):
                   2856: The @var{ior} produced by the operation, that discovered the problem, is
                   2857: thrown.
                   2859: @item named file cannot be opened (@code{included}):
                   2860: The @var{ior} produced by @code{open-file} is thrown.
                   2862: @item requesting an unmapped block number:
                   2863: There are no unmapped legal block numbers. On some operating systems,
                   2864: writing a block with a large number may overflow the file system and
                   2865: have an error message as consequence.
                   2867: @item using @code{source-id} when @code{blk} is non-zero:
                   2868: @code{source-id} performs its function. Typically it will give the id of
                   2869: the source which loaded the block. (Better ideas?)
                   2871: @end table
                   2874: @c =====================================================================
                   2875: @node  The optional Floating-Point word set, The optional Locals word set, The optional File-Access word set, ANS conformance
1.15      anton    2876: @section The optional Floating-Point word set
1.14      anton    2877: @c =====================================================================
                   2879: @menu
1.15      anton    2880: * floating-idef::               Implementation Defined Options
                   2881: * floating-ambcond::            Ambiguous Conditions            
1.14      anton    2882: @end menu
                   2885: @c ---------------------------------------------------------------------
                   2886: @node floating-idef, floating-ambcond, The optional Floating-Point word set, The optional Floating-Point word set
                   2887: @subsection Implementation Defined Options
                   2888: @c ---------------------------------------------------------------------
                   2890: @table @i
1.15      anton    2892: @item format and range of floating point numbers:
                   2893: System-dependent; the @code{double} type of C.
1.14      anton    2894: 
1.15      anton    2895: @item results of @code{REPRESENT} when @var{float} is out of range:
                   2896: System dependent; @code{REPRESENT} is implemented using the C library
                   2897: function @code{ecvt()} and inherits its behaviour in this respect.
1.14      anton    2898: 
1.15      anton    2899: @item rounding or truncation of floating-point numbers:
                   2900: What's the question?!!
1.14      anton    2901: 
1.15      anton    2902: @item size of floating-point stack:
                   2903: @code{s" FLOATING-STACK" environment? drop .}. Can be changed at startup
                   2904: with the command-line option @code{-f}.
1.14      anton    2905: 
1.15      anton    2906: @item width of floating-point stack:
                   2907: @code{1 floats}.
1.14      anton    2908: 
                   2909: @end table
                   2912: @c ---------------------------------------------------------------------
1.15      anton    2913: @node floating-ambcond,  , floating-idef, The optional Floating-Point word set
                   2914: @subsection Ambiguous conditions
1.14      anton    2915: @c ---------------------------------------------------------------------
                   2917: @table @i
1.15      anton    2919: @item @code{df@@} or @code{df!} used with an address that is not double-float  aligned:
                   2920: System-dependent. Typically results in an alignment fault like other
                   2921: alignment violations.
1.14      anton    2922: 
1.15      anton    2923: @item @code{f@@} or @code{f!} used with an address that is not float  aligned:
                   2924: System-dependent. Typically results in an alignment fault like other
                   2925: alignment violations.
1.14      anton    2926: 
1.15      anton    2927: @item Floating-point result out of range:
                   2928: System-dependent. Can result in a @code{-55 THROW} (Floating-point
                   2929: unidentified fault), or can produce a special value representing, e.g.,
                   2930: Infinity.
1.14      anton    2931: 
1.15      anton    2932: @item @code{sf@@} or @code{sf!} used with an address that is not single-float  aligned:
                   2933: System-dependent. Typically results in an alignment fault like other
                   2934: alignment violations.
1.14      anton    2935: 
1.15      anton    2936: @item BASE is not decimal (@code{REPRESENT}, @code{F.}, @code{FE.}, @code{FS.}):
                   2937: The floating-point number is converted into decimal nonetheless.
1.14      anton    2938: 
1.15      anton    2939: @item Both arguments are equal to zero (@code{FATAN2}):
                   2940: System-dependent. @code{FATAN2} is implemented using the C library
                   2941: function @code{atan2()}.
1.14      anton    2942: 
1.15      anton    2943: @item Using ftan on an argument @var{r1} where cos(@var{r1}) is zero:
                   2944: System-dependent. Anyway, typically the cos of @var{r1} will not be zero
                   2945: because of small errors and the tan will be a very large (or very small)
                   2946: but finite number.
1.14      anton    2947: 
1.15      anton    2948: @item @var{d} cannot be presented precisely as a float in @code{D>F}:
                   2949: The result is rounded to the nearest float.
1.14      anton    2950: 
1.15      anton    2951: @item dividing by zero:
                   2952: @code{-55 throw} (Floating-point unidentified fault)
1.14      anton    2953: 
1.15      anton    2954: @item exponent too big for conversion (@code{DF!}, @code{DF@@}, @code{SF!}, @code{SF@@}):
                   2955: System dependent. On IEEE-FP based systems the number is converted into
                   2956: an infinity.
1.14      anton    2957: 
1.15      anton    2958: @item @var{float}<1 (@code{facosh}):
                   2959: @code{-55 throw} (Floating-point unidentified fault)
1.14      anton    2960: 
1.15      anton    2961: @item @var{float}=<-1 (@code{flnp1}):
                   2962: @code{-55 throw} (Floating-point unidentified fault). On IEEE-FP systems
                   2963: negative infinity is typically produced for @var{float}=-1.
1.14      anton    2964: 
1.15      anton    2965: @item @var{float}=<0 (@code{fln}, @code{flog}):
                   2966: @code{-55 throw} (Floating-point unidentified fault). On IEEE-FP systems
                   2967: negative infinity is typically produced for @var{float}=0.
1.14      anton    2968: 
1.15      anton    2969: @item @var{float}<0 (@code{fasinh}, @code{fsqrt}):
                   2970: @code{-55 throw} (Floating-point unidentified fault). @code{fasinh}
                   2971: produces values for these inputs on my Linux box (Bug in the C library?)
1.14      anton    2972: 
1.15      anton    2973: @item |@var{float}|>1 (@code{facos}, @code{fasin}, @code{fatanh}):
                   2974: @code{-55 throw} (Floating-point unidentified fault).
1.14      anton    2975: 
1.15      anton    2976: @item integer part of float cannot be represented by @var{d} in @code{f>d}:
                   2977: @code{-55 throw} (Floating-point unidentified fault).
1.14      anton    2978: 
1.15      anton    2979: @item string larger than pictured numeric output area (@code{f.}, @code{fe.}, @code{fs.}):
                   2980: This does not happen.
                   2981: @end table
1.14      anton    2982: 
                   2985: @c =====================================================================
1.15      anton    2986: @node  The optional Locals word set, The optional Memory-Allocation word set, The optional Floating-Point word set, ANS conformance
                   2987: @section The optional Locals word set
1.14      anton    2988: @c =====================================================================
                   2990: @menu
1.15      anton    2991: * locals-idef::                 Implementation Defined Options                 
                   2992: * locals-ambcond::              Ambiguous Conditions              
1.14      anton    2993: @end menu
                   2996: @c ---------------------------------------------------------------------
1.15      anton    2997: @node locals-idef, locals-ambcond, The optional Locals word set, The optional Locals word set
1.14      anton    2998: @subsection Implementation Defined Options
                   2999: @c ---------------------------------------------------------------------
                   3001: @table @i
1.15      anton    3003: @item maximum number of locals in a definition:
                   3004: @code{s" #locals" environment? drop .}. Currently 15. This is a lower
                   3005: bound, e.g., on a 32-bit machine there can be 41 locals of up to 8
                   3006: characters. The number of locals in a definition is bounded by the size
                   3007: of locals-buffer, which contains the names of the locals.
1.14      anton    3008: 
                   3009: @end table
                   3012: @c ---------------------------------------------------------------------
1.15      anton    3013: @node locals-ambcond,  , locals-idef, The optional Locals word set
1.14      anton    3014: @subsection Ambiguous conditions
                   3015: @c ---------------------------------------------------------------------
                   3017: @table @i
1.15      anton    3019: @item executing a named local in interpretation state:
                   3020: @code{-14 throw} (Interpreting a compile-only word).
1.14      anton    3021: 
1.15      anton    3022: @item @var{name} not defined by @code{VALUE} or @code{(LOCAL)} (@code{TO}):
                   3023: @code{-32 throw} (Invalid name argument)
1.14      anton    3024: 
                   3025: @end table
                   3028: @c =====================================================================
1.15      anton    3029: @node  The optional Memory-Allocation word set, The optional Programming-Tools word set, The optional Locals word set, ANS conformance
                   3030: @section The optional Memory-Allocation word set
1.14      anton    3031: @c =====================================================================
                   3033: @menu
1.15      anton    3034: * memory-idef::                 Implementation Defined Options                 
1.14      anton    3035: @end menu
                   3038: @c ---------------------------------------------------------------------
1.15      anton    3039: @node memory-idef,  , The optional Memory-Allocation word set, The optional Memory-Allocation word set
1.14      anton    3040: @subsection Implementation Defined Options
                   3041: @c ---------------------------------------------------------------------
                   3043: @table @i
1.15      anton    3045: @item values and meaning of @var{ior}:
                   3046: The @var{ior}s returned by the file and memory allocation words are
                   3047: intended as throw codes. They typically are in the range
                   3048: -512@minus{}-2047 of OS errors.  The mapping from OS error numbers to
                   3049: @var{ior}s is -512@minus{}@var{errno}.
1.14      anton    3050: 
                   3051: @end table
                   3053: @c =====================================================================
1.15      anton    3054: @node  The optional Programming-Tools word set, The optional Search-Order word set, The optional Memory-Allocation word set, ANS conformance
                   3055: @section The optional Programming-Tools word set
1.14      anton    3056: @c =====================================================================
                   3058: @menu
1.15      anton    3059: * programming-idef::            Implementation Defined Options            
                   3060: * programming-ambcond::         Ambiguous Conditions         
1.14      anton    3061: @end menu
                   3064: @c ---------------------------------------------------------------------
1.15      anton    3065: @node programming-idef, programming-ambcond, The optional Programming-Tools word set, The optional Programming-Tools word set
1.14      anton    3066: @subsection Implementation Defined Options
                   3067: @c ---------------------------------------------------------------------
                   3069: @table @i
1.15      anton    3071: @item ending sequence for input following @code{;code} and @code{code}:
                   3072: Not implemented (yet).
1.14      anton    3073: 
1.15      anton    3074: @item manner of processing input following @code{;code} and @code{code}:
                   3075: Not implemented (yet).
                   3077: @item search order capability for @code{EDITOR} and @code{ASSEMBLER}:
                   3078: Not implemented (yet). If they were implemented, they would use the
                   3079: search order wordset.
                   3081: @item source and format of display by @code{SEE}:
                   3082: The source for @code{see} is the intermediate code used by the inner
                   3083: interpreter.  The current @code{see} tries to output Forth source code
                   3084: as well as possible.
1.14      anton    3086: @end table
                   3088: @c ---------------------------------------------------------------------
1.15      anton    3089: @node programming-ambcond,  , programming-idef, The optional Programming-Tools word set
1.14      anton    3090: @subsection Ambiguous conditions
                   3091: @c ---------------------------------------------------------------------
                   3093: @table @i
1.15      anton    3095: @item deleting the compilation wordlist (@code{FORGET}):
                   3096: Not implemented (yet).
1.14      anton    3097: 
1.15      anton    3098: @item fewer than @var{u}+1 items on the control flow stack (@code{CS-PICK}, @code{CS-ROLL}):
                   3099: This typically results in an @code{abort"} with a descriptive error
                   3100: message (may change into a @code{-22 throw} (Control structure mismatch)
                   3101: in the future). You may also get a memory access error. If you are
                   3102: unlucky, this ambiguous condition is not caught.
                   3104: @item @var{name} can't be found (@code{forget}):
                   3105: Not implemented (yet).
1.14      anton    3106: 
1.15      anton    3107: @item @var{name} not defined via @code{CREATE}:
                   3108: @code{;code} is not implemented (yet). If it were, it would behave like
                   3109: @code{DOES>} in this respect, i.e., change the execution semantics of
                   3110: the last defined word no matter how it was defined.
1.14      anton    3111: 
1.15      anton    3112: @item @code{POSTPONE} applied to @code{[IF]}:
                   3113: After defining @code{: X POSTPONE [IF] ; IMMEDIATE}. @code{X} is
                   3114: equivalent to @code{[IF]}.
1.14      anton    3115: 
1.15      anton    3116: @item reaching the end of the input source before matching @code{[ELSE]} or @code{[THEN]}:
                   3117: Continue in the same state of conditional compilation in the next outer
                   3118: input source. Currently there is no warning to the user about this.
1.14      anton    3119: 
1.15      anton    3120: @item removing a needed definition (@code{FORGET}):
                   3121: Not implemented (yet).
1.14      anton    3122: 
                   3123: @end table
                   3126: @c =====================================================================
1.15      anton    3127: @node  The optional Search-Order word set,  , The optional Programming-Tools word set, ANS conformance
                   3128: @section The optional Search-Order word set
1.14      anton    3129: @c =====================================================================
                   3131: @menu
1.15      anton    3132: * search-idef::                 Implementation Defined Options                 
                   3133: * search-ambcond::              Ambiguous Conditions              
1.14      anton    3134: @end menu
                   3137: @c ---------------------------------------------------------------------
1.15      anton    3138: @node search-idef, search-ambcond, The optional Search-Order word set, The optional Search-Order word set
1.14      anton    3139: @subsection Implementation Defined Options
                   3140: @c ---------------------------------------------------------------------
                   3142: @table @i
1.15      anton    3144: @item maximum number of word lists in search order:
                   3145: @code{s" wordlists" environment? drop .}. Currently 16.
                   3147: @item minimum search order:
                   3148: @code{root root}.
1.14      anton    3149: 
                   3150: @end table
                   3152: @c ---------------------------------------------------------------------
1.15      anton    3153: @node search-ambcond,  , search-idef, The optional Search-Order word set
1.14      anton    3154: @subsection Ambiguous conditions
                   3155: @c ---------------------------------------------------------------------
                   3157: @table @i
1.15      anton    3159: @item changing the compilation wordlist (during compilation):
                   3160: The definition is put into the wordlist that is the compilation wordlist
                   3161: when @code{REVEAL} is executed (by @code{;}, @code{DOES>},
                   3162: @code{RECURSIVE}, etc.).
1.14      anton    3163: 
1.15      anton    3164: @item search order empty (@code{previous}):
                   3165: @code{abort" Vocstack empty"}.
1.14      anton    3166: 
1.15      anton    3167: @item too many word lists in search order (@code{also}):
                   3168: @code{abort" Vocstack full"}.
1.14      anton    3169: 
                   3170: @end table
1.13      anton    3171: 
1.17      anton    3173: @node Model, Emacs and Gforth, ANS conformance, Top
1.4       anton    3174: @chapter Model
1.17      anton    3176: @node Emacs and Gforth, Internals, Model, Top
                   3177: @chapter Emacs and Gforth
1.4       anton    3178: 
1.17      anton    3179: Gforth comes with @file{gforth.el}, an improved version of
1.4       anton    3180: @file{forth.el} by Goran Rydqvist (icluded in the TILE package). The
                   3181: improvements are a better (but still not perfect) handling of
                   3182: indentation. I have also added comment paragraph filling (@kbd{M-q}),
1.8       anton    3183: commenting (@kbd{C-x \}) and uncommenting (@kbd{C-u C-x \}) regions and
                   3184: removing debugging tracers (@kbd{C-x ~}, @pxref{Debugging}). I left the
                   3185: stuff I do not use alone, even though some of it only makes sense for
                   3186: TILE. To get a description of these features, enter Forth mode and type
                   3187: @kbd{C-h m}.
1.4       anton    3188: 
1.17      anton    3189: In addition, Gforth supports Emacs quite well: The source code locations
1.4       anton    3190: given in error messages, debugging output (from @code{~~}) and failed
                   3191: assertion messages are in the right format for Emacs' compilation mode
                   3192: (@pxref{Compilation, , Running Compilations under Emacs, emacs, Emacs
                   3193: Manual}) so the source location corresponding to an error or other
                   3194: message is only a few keystrokes away (@kbd{C-x `} for the next error,
                   3195: @kbd{C-c C-c} for the error under the cursor).
                   3197: Also, if you @code{include} @file{etags.fs}, a new @file{TAGS} file
                   3198: (@pxref{Tags, , Tags Tables, emacs, Emacs Manual}) will be produced that
                   3199: contains the definitions of all words defined afterwards. You can then
                   3200: find the source for a word using @kbd{M-.}. Note that emacs can use
1.17      anton    3201: several tags files at the same time (e.g., one for the Gforth sources
1.4       anton    3202: and one for your program).
                   3204: To get all these benefits, add the following lines to your @file{.emacs}
                   3205: file:
                   3207: @example
                   3208: (autoload 'forth-mode "gforth.el")
                   3209: (setq auto-mode-alist (cons '("\\.fs\\'" . forth-mode) auto-mode-alist))
                   3210: @end example
1.17      anton    3212: @node Internals, Bugs, Emacs and Gforth, Top
1.3       anton    3213: @chapter Internals
1.17      anton    3215: Reading this section is not necessary for programming with Gforth. It
                   3216: should be helpful for finding your way in the Gforth sources.
1.3       anton    3217: 
1.24    ! anton    3218: The ideas in this section have also been published in the papers
        !          3219: @cite{ANS fig/GNU/??? Forth} (in German) by Bernd Paysan, presented at
        !          3220: the Forth-Tagung '93 and @cite{A Portable Forth Engine} by M. Anton
        !          3221: Ertl, presented at EuroForth '93; the latter is available at
        !          3222: @*@file{}.
        !          3223: 
1.4       anton    3224: @menu
                   3225: * Portability::                 
                   3226: * Threading::                   
                   3227: * Primitives::                  
                   3228: * System Architecture::         
1.17      anton    3229: * Performance::                 
1.4       anton    3230: @end menu
                   3232: @node Portability, Threading, Internals, Internals
1.3       anton    3233: @section Portability
                   3235: One of the main goals of the effort is availability across a wide range
                   3236: of personal machines. fig-Forth, and, to a lesser extent, F83, achieved
                   3237: this goal by manually coding the engine in assembly language for several
                   3238: then-popular processors. This approach is very labor-intensive and the
                   3239: results are short-lived due to progress in computer architecture.
                   3241: Others have avoided this problem by coding in C, e.g., Mitch Bradley
                   3242: (cforth), Mikael Patel (TILE) and Dirk Zoller (pfe). This approach is
                   3243: particularly popular for UNIX-based Forths due to the large variety of
                   3244: architectures of UNIX machines. Unfortunately an implementation in C
                   3245: does not mix well with the goals of efficiency and with using
                   3246: traditional techniques: Indirect or direct threading cannot be expressed
                   3247: in C, and switch threading, the fastest technique available in C, is
                   3248: significantly slower. Another problem with C is that it's very
                   3249: cumbersome to express double integer arithmetic.
                   3251: Fortunately, there is a portable language that does not have these
                   3252: limitations: GNU C, the version of C processed by the GNU C compiler
                   3253: (@pxref{C Extensions, , Extensions to the C Language Family,,
                   3254: GNU C Manual}). Its labels as values feature (@pxref{Labels as Values, ,
                   3255: Labels as Values,, GNU C Manual}) makes direct and indirect
                   3256: threading possible, its @code{long long} type (@pxref{Long Long, ,
                   3257: Double-Word Integers,, GNU C Manual}) corresponds to Forths
                   3258: double numbers. GNU C is available for free on all important (and many
                   3259: unimportant) UNIX machines, VMS, 80386s running MS-DOS, the Amiga, and
                   3260: the Atari ST, so a Forth written in GNU C can run on all these
1.17      anton    3261: machines.
1.3       anton    3262: 
                   3263: Writing in a portable language has the reputation of producing code that
                   3264: is slower than assembly. For our Forth engine we repeatedly looked at
                   3265: the code produced by the compiler and eliminated most compiler-induced
                   3266: inefficiencies by appropriate changes in the source-code.
                   3268: However, register allocation cannot be portably influenced by the
                   3269: programmer, leading to some inefficiencies on register-starved
                   3270: machines. We use explicit register declarations (@pxref{Explicit Reg
                   3271: Vars, , Variables in Specified Registers,, GNU C Manual}) to
                   3272: improve the speed on some machines. They are turned on by using the
                   3273: @code{gcc} switch @code{-DFORCE_REG}. Unfortunately, this feature not
                   3274: only depends on the machine, but also on the compiler version: On some
                   3275: machines some compiler versions produce incorrect code when certain
                   3276: explicit register declarations are used. So by default
                   3277: @code{-DFORCE_REG} is not used.
1.4       anton    3279: @node Threading, Primitives, Portability, Internals
1.3       anton    3280: @section Threading
                   3282: GNU C's labels as values extension (available since @code{gcc-2.0},
                   3283: @pxref{Labels as Values, , Labels as Values,, GNU C Manual})
                   3284: makes it possible to take the address of @var{label} by writing
                   3285: @code{&&@var{label}}.  This address can then be used in a statement like
                   3286: @code{goto *@var{address}}. I.e., @code{goto *&&x} is the same as
                   3287: @code{goto x}.
                   3289: With this feature an indirect threaded NEXT looks like:
                   3290: @example
                   3291: cfa = *ip++;
                   3292: ca = *cfa;
                   3293: goto *ca;
                   3294: @end example
                   3295: For those unfamiliar with the names: @code{ip} is the Forth instruction
                   3296: pointer; the @code{cfa} (code-field address) corresponds to ANS Forths
                   3297: execution token and points to the code field of the next word to be
                   3298: executed; The @code{ca} (code address) fetched from there points to some
                   3299: executable code, e.g., a primitive or the colon definition handler
                   3300: @code{docol}.
                   3302: Direct threading is even simpler:
                   3303: @example
                   3304: ca = *ip++;
                   3305: goto *ca;
                   3306: @end example
                   3308: Of course we have packaged the whole thing neatly in macros called
                   3309: @code{NEXT} and @code{NEXT1} (the part of NEXT after fetching the cfa).
1.4       anton    3311: @menu
                   3312: * Scheduling::                  
                   3313: * Direct or Indirect Threaded?::  
                   3314: * DOES>::                       
                   3315: @end menu
                   3317: @node Scheduling, Direct or Indirect Threaded?, Threading, Threading
1.3       anton    3318: @subsection Scheduling
                   3320: There is a little complication: Pipelined and superscalar processors,
                   3321: i.e., RISC and some modern CISC machines can process independent
                   3322: instructions while waiting for the results of an instruction. The
                   3323: compiler usually reorders (schedules) the instructions in a way that
                   3324: achieves good usage of these delay slots. However, on our first tries
                   3325: the compiler did not do well on scheduling primitives. E.g., for
                   3326: @code{+} implemented as
                   3327: @example
                   3328: n=sp[0]+sp[1];
                   3329: sp++;
                   3330: sp[0]=n;
                   3331: NEXT;
                   3332: @end example
                   3333: the NEXT comes strictly after the other code, i.e., there is nearly no
                   3334: scheduling. After a little thought the problem becomes clear: The
                   3335: compiler cannot know that sp and ip point to different addresses (and
1.4       anton    3336: the version of @code{gcc} we used would not know it even if it was
                   3337: possible), so it could not move the load of the cfa above the store to
                   3338: the TOS. Indeed the pointers could be the same, if code on or very near
                   3339: the top of stack were executed. In the interest of speed we chose to
                   3340: forbid this probably unused ``feature'' and helped the compiler in
                   3341: scheduling: NEXT is divided into the loading part (@code{NEXT_P1}) and
                   3342: the goto part (@code{NEXT_P2}). @code{+} now looks like:
1.3       anton    3343: @example
                   3344: n=sp[0]+sp[1];
                   3345: sp++;
                   3346: NEXT_P1;
                   3347: sp[0]=n;
                   3348: NEXT_P2;
                   3349: @end example
1.4       anton    3350: This can be scheduled optimally by the compiler.
1.3       anton    3351: 
                   3352: This division can be turned off with the switch @code{-DCISC_NEXT}. This
                   3353: switch is on by default on machines that do not profit from scheduling
                   3354: (e.g., the 80386), in order to preserve registers.
1.4       anton    3356: @node Direct or Indirect Threaded?, DOES>, Scheduling, Threading
1.3       anton    3357: @subsection Direct or Indirect Threaded?
                   3359: Both! After packaging the nasty details in macro definitions we
                   3360: realized that we could switch between direct and indirect threading by
                   3361: simply setting a compilation flag (@code{-DDIRECT_THREADED}) and
                   3362: defining a few machine-specific macros for the direct-threading case.
                   3363: On the Forth level we also offer access words that hide the
                   3364: differences between the threading methods (@pxref{Threading Words}).
                   3366: Indirect threading is implemented completely
                   3367: machine-independently. Direct threading needs routines for creating
                   3368: jumps to the executable code (e.g. to docol or dodoes). These routines
                   3369: are inherently machine-dependent, but they do not amount to many source
                   3370: lines. I.e., even porting direct threading to a new machine is a small
                   3371: effort.
1.4       anton    3373: @node DOES>,  , Direct or Indirect Threaded?, Threading
1.3       anton    3374: @subsection DOES>
                   3375: One of the most complex parts of a Forth engine is @code{dodoes}, i.e.,
                   3376: the chunk of code executed by every word defined by a
                   3377: @code{CREATE}...@code{DOES>} pair. The main problem here is: How to find
                   3378: the Forth code to be executed, i.e. the code after the @code{DOES>} (the
                   3379: DOES-code)? There are two solutions:
                   3381: In fig-Forth the code field points directly to the dodoes and the
                   3382: DOES-code address is stored in the cell after the code address
                   3383: (i.e. at cfa cell+). It may seem that this solution is illegal in the
                   3384: Forth-79 and all later standards, because in fig-Forth this address
                   3385: lies in the body (which is illegal in these standards). However, by
                   3386: making the code field larger for all words this solution becomes legal
                   3387: again. We use this approach for the indirect threaded version. Leaving
                   3388: a cell unused in most words is a bit wasteful, but on the machines we
                   3389: are targetting this is hardly a problem. The other reason for having a
                   3390: code field size of two cells is to avoid having different image files
1.4       anton    3391: for direct and indirect threaded systems (@pxref{System Architecture}).
1.3       anton    3392: 
                   3393: The other approach is that the code field points or jumps to the cell
                   3394: after @code{DOES}. In this variant there is a jump to @code{dodoes} at
                   3395: this address. @code{dodoes} can then get the DOES-code address by
                   3396: computing the code address, i.e., the address of the jump to dodoes,
                   3397: and add the length of that jump field. A variant of this is to have a
                   3398: call to @code{dodoes} after the @code{DOES>}; then the return address
                   3399: (which can be found in the return register on RISCs) is the DOES-code
                   3400: address. Since the two cells available in the code field are usually
                   3401: used up by the jump to the code address in direct threading, we use
                   3402: this approach for direct threading. We did not want to add another
                   3403: cell to the code field.
1.4       anton    3405: @node Primitives, System Architecture, Threading, Internals
1.3       anton    3406: @section Primitives
1.4       anton    3408: @menu
                   3409: * Automatic Generation::        
                   3410: * TOS Optimization::            
                   3411: * Produced code::               
                   3412: @end menu
                   3414: @node Automatic Generation, TOS Optimization, Primitives, Primitives
1.3       anton    3415: @subsection Automatic Generation
                   3417: Since the primitives are implemented in a portable language, there is no
                   3418: longer any need to minimize the number of primitives. On the contrary,
                   3419: having many primitives is an advantage: speed. In order to reduce the
                   3420: number of errors in primitives and to make programming them easier, we
                   3421: provide a tool, the primitive generator (@file{prims2x.fs}), that
                   3422: automatically generates most (and sometimes all) of the C code for a
                   3423: primitive from the stack effect notation.  The source for a primitive
                   3424: has the following form:
                   3426: @format
                   3427: @var{Forth-name}       @var{stack-effect}      @var{category}  [@var{pronounc.}]
                   3428: [@code{""}@var{glossary entry}@code{""}]
                   3429: @var{C code}
                   3430: [@code{:}
                   3431: @var{Forth code}]
                   3432: @end format
                   3434: The items in brackets are optional. The category and glossary fields
                   3435: are there for generating the documentation, the Forth code is there
                   3436: for manual implementations on machines without GNU C. E.g., the source
                   3437: for the primitive @code{+} is:
                   3438: @example
                   3439: +    n1 n2 -- n    core    plus
                   3440: n = n1+n2;
                   3441: @end example
                   3443: This looks like a specification, but in fact @code{n = n1+n2} is C
                   3444: code. Our primitive generation tool extracts a lot of information from
                   3445: the stack effect notations@footnote{We use a one-stack notation, even
                   3446: though we have separate data and floating-point stacks; The separate
                   3447: notation can be generated easily from the unified notation.}: The number
                   3448: of items popped from and pushed on the stack, their type, and by what
                   3449: name they are referred to in the C code. It then generates a C code
                   3450: prelude and postlude for each primitive. The final C code for @code{+}
                   3451: looks like this:
                   3453: @example
                   3454: I_plus:        /* + ( n1 n2 -- n ) */  /* label, stack effect */
                   3455: /*  */                          /* documentation */
1.4       anton    3456: @{
1.3       anton    3457: DEF_CA                          /* definition of variable ca (indirect threading) */
                   3458: Cell n1;                        /* definitions of variables */
                   3459: Cell n2;
                   3460: Cell n;
                   3461: n1 = (Cell) sp[1];              /* input */
                   3462: n2 = (Cell) TOS;
                   3463: sp += 1;                        /* stack adjustment */
                   3464: NAME("+")                       /* debugging output (with -DDEBUG) */
1.4       anton    3465: @{
1.3       anton    3466: n = n1+n2;                      /* C code taken from the source */
1.4       anton    3467: @}
1.3       anton    3468: NEXT_P1;                        /* NEXT part 1 */
                   3469: TOS = (Cell)n;                  /* output */
                   3470: NEXT_P2;                        /* NEXT part 2 */
1.4       anton    3471: @}
1.3       anton    3472: @end example
                   3474: This looks long and inefficient, but the GNU C compiler optimizes quite
                   3475: well and produces optimal code for @code{+} on, e.g., the R3000 and the
                   3476: HP RISC machines: Defining the @code{n}s does not produce any code, and
                   3477: using them as intermediate storage also adds no cost.
                   3479: There are also other optimizations, that are not illustrated by this
                   3480: example: Assignments between simple variables are usually for free (copy
                   3481: propagation). If one of the stack items is not used by the primitive
                   3482: (e.g.  in @code{drop}), the compiler eliminates the load from the stack
                   3483: (dead code elimination). On the other hand, there are some things that
                   3484: the compiler does not do, therefore they are performed by
                   3485: @file{prims2x.fs}: The compiler does not optimize code away that stores
                   3486: a stack item to the place where it just came from (e.g., @code{over}).
                   3488: While programming a primitive is usually easy, there are a few cases
                   3489: where the programmer has to take the actions of the generator into
                   3490: account, most notably @code{?dup}, but also words that do not (always)
                   3491: fall through to NEXT.
1.4       anton    3493: @node TOS Optimization, Produced code, Automatic Generation, Primitives
1.3       anton    3494: @subsection TOS Optimization
                   3496: An important optimization for stack machine emulators, e.g., Forth
                   3497: engines, is keeping  one or more of the top stack items in
1.4       anton    3498: registers.  If a word has the stack effect @var{in1}...@var{inx} @code{--}
                   3499: @var{out1}...@var{outy}, keeping the top @var{n} items in registers
1.3       anton    3500: @itemize
                   3501: @item
                   3502: is better than keeping @var{n-1} items, if @var{x>=n} and @var{y>=n},
                   3503: due to fewer loads from and stores to the stack.
                   3504: @item is slower than keeping @var{n-1} items, if @var{x<>y} and @var{x<n} and
                   3505: @var{y<n}, due to additional moves between registers.
                   3506: @end itemize
                   3508: In particular, keeping one item in a register is never a disadvantage,
                   3509: if there are enough registers. Keeping two items in registers is a
                   3510: disadvantage for frequent words like @code{?branch}, constants,
                   3511: variables, literals and @code{i}. Therefore our generator only produces
                   3512: code that keeps zero or one items in registers. The generated C code
                   3513: covers both cases; the selection between these alternatives is made at
                   3514: C-compile time using the switch @code{-DUSE_TOS}. @code{TOS} in the C
                   3515: code for @code{+} is just a simple variable name in the one-item case,
                   3516: otherwise it is a macro that expands into @code{sp[0]}. Note that the
                   3517: GNU C compiler tries to keep simple variables like @code{TOS} in
                   3518: registers, and it usually succeeds, if there are enough registers.
                   3520: The primitive generator performs the TOS optimization for the
                   3521: floating-point stack, too (@code{-DUSE_FTOS}). For floating-point
                   3522: operations the benefit of this optimization is even larger:
                   3523: floating-point operations take quite long on most processors, but can be
                   3524: performed in parallel with other operations as long as their results are
                   3525: not used. If the FP-TOS is kept in a register, this works. If
                   3526: it is kept on the stack, i.e., in memory, the store into memory has to
                   3527: wait for the result of the floating-point operation, lengthening the
                   3528: execution time of the primitive considerably.
                   3530: The TOS optimization makes the automatic generation of primitives a
                   3531: bit more complicated. Just replacing all occurrences of @code{sp[0]} by
                   3532: @code{TOS} is not sufficient. There are some special cases to
                   3533: consider:
                   3534: @itemize
                   3535: @item In the case of @code{dup ( w -- w w )} the generator must not
                   3536: eliminate the store to the original location of the item on the stack,
                   3537: if the TOS optimization is turned on.
1.4       anton    3538: @item Primitives with stack effects of the form @code{--}
                   3539: @var{out1}...@var{outy} must store the TOS to the stack at the start.
                   3540: Likewise, primitives with the stack effect @var{in1}...@var{inx} @code{--}
1.3       anton    3541: must load the TOS from the stack at the end. But for the null stack
                   3542: effect @code{--} no stores or loads should be generated.
                   3543: @end itemize
1.4       anton    3545: @node Produced code,  , TOS Optimization, Primitives
1.3       anton    3546: @subsection Produced code
                   3548: To see what assembly code is produced for the primitives on your machine
                   3549: with your compiler and your flag settings, type @code{make engine.s} and
1.4       anton    3550: look at the resulting file @file{engine.s}.
1.3       anton    3551: 
1.17      anton    3552: @node System Architecture, Performance, Primitives, Internals
1.3       anton    3553: @section System Architecture
                   3555: Our Forth system consists not only of primitives, but also of
                   3556: definitions written in Forth. Since the Forth compiler itself belongs
                   3557: to those definitions, it is not possible to start the system with the
                   3558: primitives and the Forth source alone. Therefore we provide the Forth
                   3559: code as an image file in nearly executable form. At the start of the
                   3560: system a C routine loads the image file into memory, sets up the
                   3561: memory (stacks etc.) according to information in the image file, and
                   3562: starts executing Forth code.
                   3564: The image file format is a compromise between the goals of making it
                   3565: easy to generate image files and making them portable. The easiest way
                   3566: to generate an image file is to just generate a memory dump. However,
                   3567: this kind of image file cannot be used on a different machine, or on
                   3568: the next version of the engine on the same machine, it even might not
                   3569: work with the same engine compiled by a different version of the C
                   3570: compiler. We would like to have as few versions of the image file as
                   3571: possible, because we do not want to distribute many versions of the
                   3572: same image file, and to make it easy for the users to use their image
                   3573: files on many machines. We currently need to create a different image
                   3574: file for machines with different cell sizes and different byte order
1.17      anton    3575: (little- or big-endian)@footnote{We are considering adding information to the
1.3       anton    3576: image file that enables the loader to change the byte order.}.
                   3578: Forth code that is going to end up in a portable image file has to
1.4       anton    3579: comply to some restrictions: addresses have to be stored in memory with
                   3580: special words (@code{A!}, @code{A,}, etc.) in order to make the code
                   3581: relocatable. Cells, floats, etc., have to be stored at the natural
                   3582: alignment boundaries@footnote{E.g., store floats (8 bytes) at an address
                   3583: dividable by~8. This happens automatically in our system when you use
                   3584: the ANS Forth alignment words.}, in order to avoid alignment faults on
                   3585: machines with stricter alignment. The image file is produced by a
                   3586: metacompiler (@file{cross.fs}).
1.3       anton    3587: 
                   3588: So, unlike the image file of Mitch Bradleys @code{cforth}, our image
                   3589: file is not directly executable, but has to undergo some manipulations
                   3590: during loading. Address relocation is performed at image load-time, not
                   3591: at run-time. The loader also has to replace tokens standing for
                   3592: primitive calls with the appropriate code-field addresses (or code
                   3593: addresses in the case of direct threading).
1.4       anton    3594: 
1.17      anton    3595: @node  Performance,  , System Architecture, Internals
                   3596: @section Performance
                   3598: On RISCs the Gforth engine is very close to optimal; i.e., it is usually
                   3599: impossible to write a significantly faster engine.
                   3601: On register-starved machines like the 386 architecture processors
                   3602: improvements are possible, because @code{gcc} does not utilize the
                   3603: registers as well as a human, even with explicit register declarations;
                   3604: e.g., Bernd Beuster wrote a Forth system fragment in assembly language
                   3605: and hand-tuned it for the 486; this system is 1.19 times faster on the
                   3606: Sieve benchmark on a 486DX2/66 than Gforth compiled with
                   3607: @code{gcc-2.6.3} with @code{-DFORCE_REG}.
                   3609: However, this potential advantage of assembly language implementations
                   3610: is not necessarily realized in complete Forth systems: We compared
                   3611: Gforth (compiled with @code{gcc-2.6.3} and @code{-DFORCE_REG}) with
1.18      anton    3612: Win32Forth 1.2093 and LMI's NT Forth (Beta, May 1994), two systems
                   3613: written in assembly, and with two systems written in C: PFE-0.9.11
                   3614: (compiled with @code{gcc-2.6.3} with the default configuration for
                   3615: Linux: @code{-O2 -fomit-frame-pointer -DUSE_REGS}) and ThisForth Beta
                   3616: (compiled with gcc-2.6.3 -O3 -fomit-frame-pointer). We benchmarked
                   3617: Gforth, PFE and ThisForth on a 486DX2/66 under Linux. Kenneth O'Heskin
                   3618: kindly provided the results for Win32Forth and NT Forth on a 486DX2/66
                   3619: with similar memory performance under Windows NT.
1.17      anton    3620:  
                   3621: We used four small benchmarks: the ubiquitous Sieve; bubble-sorting and
                   3622: matrix multiplication come from the Stanford integer benchmarks and have
                   3623: been translated into Forth by Martin Fraeman; we used the versions
                   3624: included in the TILE Forth package; and a recursive Fibonacci number
1.24    ! anton    3625: computation for benchmarking calling performance. The following table shows
1.17      anton    3626: the time taken for the benchmarks scaled by the time taken by Gforth (in
                   3627: other words, it shows the speedup factor that Gforth achieved over the
                   3628: other systems).
                   3630: @example
                   3631: relative             Win32-        NT               This-
                   3632:   time     Gforth     Forth     Forth       PFE     Forth
                   3633: sieve        1.00      1.30      1.07      1.67      2.98
                   3634: bubble       1.00      1.30      1.40      1.66
                   3635: matmul       1.00      1.40      1.29      2.24
                   3636: fib          1.00      1.44      1.26      1.82      2.82
                   3637: @end example
                   3639: You may find the good performance of Gforth compared with the systems
                   3640: written in assembly language quite surprising. One important reason for
                   3641: the disappointing performance of these systems is probably that they are
                   3642: not written optimally for the 486 (e.g., they use the @code{lods}
                   3643: instruction). In addition, Win32Forth uses a comfortable, but costly
                   3644: method for relocating the Forth image: like @code{cforth}, it computes
                   3645: the actual addresses at run time, resulting in two address computations
                   3646: per NEXT (@pxref{System Architecture}).
                   3648: The speedup of Gforth over PFE and ThisForth can be easily explained
                   3649: with the self-imposed restriction to standard C (although the measured
                   3650: implementation of PFE uses a GNU C extension: global register
                   3651: variables), which makes efficient threading impossible.  Moreover,
                   3652: current C compilers have a hard time optimizing other aspects of the
                   3653: ThisForth source.
                   3655: Note that the performance of Gforth on 386 architecture processors
                   3656: varies widely with the version of @code{gcc} used. E.g., @code{gcc-2.5.8}
                   3657: failed to allocate any of the virtual machine registers into real
                   3658: machine registers by itself and would not work correctly with explicit
                   3659: register declarations, giving a 1.3 times slower engine (on a 486DX2/66
                   3660: running the Sieve) than the one measured above.
1.24    ! anton    3662: The numbers in this section have also been published in the paper
        !          3663: @cite{Translating Forth to Efficient C} by M. Anton Ertl and Martin
        !          3664: Maierhofer, presented at EuroForth '95. It is available at
        !          3665: @*@file{};
        !          3666: it also contains numbers for some native code systems. You can find
        !          3667: numbers for Gforth on various machines in @file{Benchres}.
        !          3668: 
1.4       anton    3669: @node Bugs, Pedigree, Internals, Top
                   3670: @chapter Bugs
1.17      anton    3672: Known bugs are described in the file BUGS in the Gforth distribution.
1.24    ! anton    3674: If you find a bug, please send a bug report to
        !          3675: @code{}. A bug report should
1.17      anton    3676: describe the Gforth version used (it is announced at the start of an
                   3677: interactive Gforth session), the machine and operating system (on Unix
                   3678: systems you can use @code{uname -a} to produce this information), the
1.24    ! anton    3679: installation options (send the @code{config.status} file), and a
        !          3680: complete list of changes you (or your installer) have made to the Gforth
        !          3681: sources (if any); it should contain a program (or a sequence of keyboard
        !          3682: commands) that reproduces the bug and a description of what you think
        !          3683: constitutes the buggy behaviour.
1.17      anton    3684: 
                   3685: For a thorough guide on reporting bugs read @ref{Bug Reporting, , How
                   3686: to Report Bugs,, GNU C Manual}.
1.4       anton    3689: @node Pedigree, Word Index, Bugs, Top
                   3690: @chapter Pedigree
1.17      anton    3692: Gforth descends from BigForth (1993) and fig-Forth. Gforth and PFE (by
1.24    ! anton    3693: Dirk Zoller) will cross-fertilize each other. Of course, a significant
        !          3694: part of the design of Gforth was prescribed by ANS Forth.
1.17      anton    3695: 
1.23      pazsan   3696: Bernd Paysan wrote BigForth, a descendent from TurboForth, an unreleased
                   3697: 32 bit native code version of VolksForth for the Atari ST, written
                   3698: mostly by Dietrich Weineck.
                   3700: VolksForth descends from F83. It was written by Klaus Schleisiek, Bernd
                   3701: Pennemann, Georg Rehfeld and Dietrich Weineck for the C64 (called
1.24    ! anton    3702: UltraForth there) in the mid-80s and ported to the Atari ST in 1986.
1.17      anton    3703: 
                   3704: Laxen and Perry wrote F83 as a model implementation of the
                   3705: Forth-83 standard. !! Pedigree? When?
                   3707: A team led by Bill Ragsdale implemented fig-Forth on many processors in
1.24    ! anton    3708: 1979. Robert Selzer and Bill Ragsdale developed the original
        !          3709: implementation of fig-Forth for the 6502 based on microForth.
        !          3710: 
        !          3711: The principal architect of microForth was Dean Sanderson. microForth was
        !          3712: FORTH, Inc.'s first off-the-shelf product. It was developped in 1976 for
        !          3713: the 1802, and subsequently implemented on the 8080, the 6800 and the
        !          3714: Z80.
1.17      anton    3715: 
1.24    ! anton    3716: All earlier Forth systems were custom-made, usually by Charles Moore,
        !          3717: who discovered (as he puts it) Forth in the late 60s.
1.17      anton    3718: 
                   3719: A part of the information in this section comes from @cite{The Evolution
                   3720: of Forth} by Elizabeth D. Rather, Donald R. Colburn and Charles
                   3721: H. Moore, presented at the HOPL-II conference and preprinted in SIGPLAN
                   3722: Notices 28(3), 1993.  You can find more historical and genealogical
                   3723: information about Forth there.
1.4       anton    3725: @node Word Index, Node Index, Pedigree, Top
                   3726: @chapter Word Index
1.18      anton    3728: This index is as incomplete as the manual. Each word is listed with
                   3729: stack effect and wordset.
1.17      anton    3730: 
                   3731: @printindex fn
1.4       anton    3733: @node Node Index,  , Word Index, Top
                   3734: @chapter Node Index
1.17      anton    3735: 
                   3736: This index is even less complete than the manual.
1.1       anton    3737: 
                   3738: @contents
                   3739: @bye

FreeBSD-CVSweb <>