Diff for /gforth/Attic/gforth.ds between versions 1.3 and 1.31

version 1.3, 1994/11/23 16:54:39 version 1.31, 1996/02/09 17:34:09
Line 1 Line 1
 \input texinfo   @c -*-texinfo-*-  \input texinfo   @c -*-texinfo-*-
 @comment The source is gforth.ds, from which gforth.texi is generated  @comment The source is gforth.ds, from which gforth.texi is generated
 @comment %**start of header (This is for running Texinfo on a region.)  @comment %**start of header (This is for running Texinfo on a region.)
 @setfilename gforth-info  @setfilename gforth.info
 @settitle GNU Forth Manual  @settitle Gforth Manual
 @setchapternewpage odd  @comment @setchapternewpage odd
 @comment %**end of header (This is for running Texinfo on a region.)  @comment %**end of header (This is for running Texinfo on a region.)
   
 @ifinfo  @ifinfo
 This file documents GNU Forth 0.0  This file documents Gforth 0.2
   
 Copyright @copyright{} 1994 GNU Forth Development Group  Copyright @copyright{} 1995 Free Software Foundation, Inc.
   
      Permission is granted to make and distribute verbatim copies of       Permission is granted to make and distribute verbatim copies of
      this manual provided the copyright notice and this permission notice       this manual provided the copyright notice and this permission notice
      are preserved on all copies.       are preserved on all copies.
             
      @ignore  @ignore
      Permission is granted to process this file through TeX and print the       Permission is granted to process this file through TeX and print the
      results, provided the printed document carries a copying permission       results, provided the printed document carries a copying permission
      notice identical to this one except for the removal of this paragraph       notice identical to this one except for the removal of this paragraph
      (this paragraph not being relevant to the printed manual).       (this paragraph not being relevant to the printed manual).
             
      @end ignore  @end ignore
      Permission is granted to copy and distribute modified versions of this       Permission is granted to copy and distribute modified versions of this
      manual under the conditions for verbatim copying, provided also that the       manual under the conditions for verbatim copying, provided also that the
      sections entitled "Distribution" and "General Public License" are       sections entitled "Distribution" and "General Public License" are
Line 36  Copyright @copyright{} 1994 GNU Forth De Line 36  Copyright @copyright{} 1994 GNU Forth De
      of in the original English.       of in the original English.
 @end ifinfo  @end ifinfo
   
   @finalout
 @titlepage  @titlepage
 @sp 10  @sp 10
 @center @titlefont{GNU Forth Manual}  @center @titlefont{Gforth Manual}
 @sp 2  @sp 2
 @center for version 0.0  @center for version 0.2
 @sp 2  @sp 2
 @center Anton Ertl  @center Anton Ertl
   @center Bernd Paysan
   @sp 3
   @center This manual is under construction
   
 @comment  The following two commands start the copyright page.  @comment  The following two commands start the copyright page.
 @page  @page
 @vskip 0pt plus 1filll  @vskip 0pt plus 1filll
 Copyright @copyright{} 1994 GNU Forth Development Group  Copyright @copyright{} 1995 Free Software Foundation, Inc.
   
 @comment !! Published by ... or You can get a copy of this manual ...  @comment !! Published by ... or You can get a copy of this manual ...
   
Line 72  Copyright @copyright{} 1994 GNU Forth De Line 76  Copyright @copyright{} 1994 GNU Forth De
   
 @node Top, License, (dir), (dir)  @node Top, License, (dir), (dir)
 @ifinfo  @ifinfo
 GNU Forth is a free implementation of ANS Forth available on many  Gforth is a free implementation of ANS Forth available on many
 personal machines. This manual corresponds to version 0.0.  personal machines. This manual corresponds to version 0.2.
 @end ifinfo  @end ifinfo
   
 @menu  @menu
 * License::               * License::                     
 * Goals::               About the GNU Forth Project  * Goals::                       About the Gforth Project
 * Other Books::         Things you might want to read  * Other Books::                 Things you might want to read
 * Invocation::          Starting GNU Forth  * Invocation::                  Starting Gforth
 * Words::               Forth words available in GNU Forth  * Words::                       Forth words available in Gforth
 * ANS conformance::     Implementation-defined options etc.  * ANS conformance::             Implementation-defined options etc.
 * Model::               The abstract machine of GNU Forth  * Model::                       The abstract machine of Gforth
 * Emacs and GForth::    The GForth Mode  * Emacs and Gforth::            The Gforth Mode
 * Internals::           Implementation details  * Internals::                   Implementation details
 * Bugs::                How to report them  * Bugs::                        How to report them
 * Pedigree::            Ancestors of GNU Forth  * Origin::                      Authors and ancestors of Gforth
 * Word Index::          An item for each Forth word  * Word Index::                  An item for each Forth word
 * Node Index::          An item for each node  * Node Index::                  An item for each node
 @end menu  @end menu
   
 @node License, Goals, Top, Top  @node License, Goals, Top, Top
 @unnumbered License  @unnumbered GNU GENERAL PUBLIC LICENSE
 !! Insert GPL here  @center Version 2, June 1991
   
   @display
   Copyright @copyright{} 1989, 1991 Free Software Foundation, Inc.
   675 Mass Ave, Cambridge, MA 02139, USA
   
   Everyone is permitted to copy and distribute verbatim copies
   of this license document, but changing it is not allowed.
   @end display
   
   @unnumberedsec Preamble
   
     The licenses for most software are designed to take away your
   freedom to share and change it.  By contrast, the GNU General Public
   License is intended to guarantee your freedom to share and change free
   software---to make sure the software is free for all its users.  This
   General Public License applies to most of the Free Software
   Foundation's software and to any other program whose authors commit to
   using it.  (Some other Free Software Foundation software is covered by
   the GNU Library General Public License instead.)  You can apply it to
   your programs, too.
   
     When we speak of free software, we are referring to freedom, not
   price.  Our General Public Licenses are designed to make sure that you
   have the freedom to distribute copies of free software (and charge for
   this service if you wish), that you receive source code or can get it
   if you want it, that you can change the software or use pieces of it
   in new free programs; and that you know you can do these things.
   
     To protect your rights, we need to make restrictions that forbid
   anyone to deny you these rights or to ask you to surrender the rights.
   These restrictions translate to certain responsibilities for you if you
   distribute copies of the software, or if you modify it.
   
     For example, if you distribute copies of such a program, whether
   gratis or for a fee, you must give the recipients all the rights that
   you have.  You must make sure that they, too, receive or can get the
   source code.  And you must show them these terms so they know their
   rights.
   
     We protect your rights with two steps: (1) copyright the software, and
   (2) offer you this license which gives you legal permission to copy,
   distribute and/or modify the software.
   
     Also, for each author's protection and ours, we want to make certain
   that everyone understands that there is no warranty for this free
   software.  If the software is modified by someone else and passed on, we
   want its recipients to know that what they have is not the original, so
   that any problems introduced by others will not reflect on the original
   authors' reputations.
   
     Finally, any free program is threatened constantly by software
   patents.  We wish to avoid the danger that redistributors of a free
   program will individually obtain patent licenses, in effect making the
   program proprietary.  To prevent this, we have made it clear that any
   patent must be licensed for everyone's free use or not licensed at all.
   
     The precise terms and conditions for copying, distribution and
   modification follow.
   
 @iftex  @iftex
   @unnumberedsec TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
   @end iftex
   @ifinfo
   @center TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
   @end ifinfo
   
   @enumerate 0
   @item
   This License applies to any program or other work which contains
   a notice placed by the copyright holder saying it may be distributed
   under the terms of this General Public License.  The ``Program'', below,
   refers to any such program or work, and a ``work based on the Program''
   means either the Program or any derivative work under copyright law:
   that is to say, a work containing the Program or a portion of it,
   either verbatim or with modifications and/or translated into another
   language.  (Hereinafter, translation is included without limitation in
   the term ``modification''.)  Each licensee is addressed as ``you''.
   
   Activities other than copying, distribution and modification are not
   covered by this License; they are outside its scope.  The act of
   running the Program is not restricted, and the output from the Program
   is covered only if its contents constitute a work based on the
   Program (independent of having been made by running the Program).
   Whether that is true depends on what the Program does.
   
   @item
   You may copy and distribute verbatim copies of the Program's
   source code as you receive it, in any medium, provided that you
   conspicuously and appropriately publish on each copy an appropriate
   copyright notice and disclaimer of warranty; keep intact all the
   notices that refer to this License and to the absence of any warranty;
   and give any other recipients of the Program a copy of this License
   along with the Program.
   
   You may charge a fee for the physical act of transferring a copy, and
   you may at your option offer warranty protection in exchange for a fee.
   
   @item
   You may modify your copy or copies of the Program or any portion
   of it, thus forming a work based on the Program, and copy and
   distribute such modifications or work under the terms of Section 1
   above, provided that you also meet all of these conditions:
   
   @enumerate a
   @item
   You must cause the modified files to carry prominent notices
   stating that you changed the files and the date of any change.
   
   @item
   You must cause any work that you distribute or publish, that in
   whole or in part contains or is derived from the Program or any
   part thereof, to be licensed as a whole at no charge to all third
   parties under the terms of this License.
   
   @item
   If the modified program normally reads commands interactively
   when run, you must cause it, when started running for such
   interactive use in the most ordinary way, to print or display an
   announcement including an appropriate copyright notice and a
   notice that there is no warranty (or else, saying that you provide
   a warranty) and that users may redistribute the program under
   these conditions, and telling the user how to view a copy of this
   License.  (Exception: if the Program itself is interactive but
   does not normally print such an announcement, your work based on
   the Program is not required to print an announcement.)
   @end enumerate
   
   These requirements apply to the modified work as a whole.  If
   identifiable sections of that work are not derived from the Program,
   and can be reasonably considered independent and separate works in
   themselves, then this License, and its terms, do not apply to those
   sections when you distribute them as separate works.  But when you
   distribute the same sections as part of a whole which is a work based
   on the Program, the distribution of the whole must be on the terms of
   this License, whose permissions for other licensees extend to the
   entire whole, and thus to each and every part regardless of who wrote it.
   
   Thus, it is not the intent of this section to claim rights or contest
   your rights to work written entirely by you; rather, the intent is to
   exercise the right to control the distribution of derivative or
   collective works based on the Program.
   
   In addition, mere aggregation of another work not based on the Program
   with the Program (or with a work based on the Program) on a volume of
   a storage or distribution medium does not bring the other work under
   the scope of this License.
   
   @item
   You may copy and distribute the Program (or a work based on it,
   under Section 2) in object code or executable form under the terms of
   Sections 1 and 2 above provided that you also do one of the following:
   
   @enumerate a
   @item
   Accompany it with the complete corresponding machine-readable
   source code, which must be distributed under the terms of Sections
   1 and 2 above on a medium customarily used for software interchange; or,
   
   @item
   Accompany it with a written offer, valid for at least three
   years, to give any third party, for a charge no more than your
   cost of physically performing source distribution, a complete
   machine-readable copy of the corresponding source code, to be
   distributed under the terms of Sections 1 and 2 above on a medium
   customarily used for software interchange; or,
   
   @item
   Accompany it with the information you received as to the offer
   to distribute corresponding source code.  (This alternative is
   allowed only for noncommercial distribution and only if you
   received the program in object code or executable form with such
   an offer, in accord with Subsection b above.)
   @end enumerate
   
   The source code for a work means the preferred form of the work for
   making modifications to it.  For an executable work, complete source
   code means all the source code for all modules it contains, plus any
   associated interface definition files, plus the scripts used to
   control compilation and installation of the executable.  However, as a
   special exception, the source code distributed need not include
   anything that is normally distributed (in either source or binary
   form) with the major components (compiler, kernel, and so on) of the
   operating system on which the executable runs, unless that component
   itself accompanies the executable.
   
   If distribution of executable or object code is made by offering
   access to copy from a designated place, then offering equivalent
   access to copy the source code from the same place counts as
   distribution of the source code, even though third parties are not
   compelled to copy the source along with the object code.
   
   @item
   You may not copy, modify, sublicense, or distribute the Program
   except as expressly provided under this License.  Any attempt
   otherwise to copy, modify, sublicense or distribute the Program is
   void, and will automatically terminate your rights under this License.
   However, parties who have received copies, or rights, from you under
   this License will not have their licenses terminated so long as such
   parties remain in full compliance.
   
   @item
   You are not required to accept this License, since you have not
   signed it.  However, nothing else grants you permission to modify or
   distribute the Program or its derivative works.  These actions are
   prohibited by law if you do not accept this License.  Therefore, by
   modifying or distributing the Program (or any work based on the
   Program), you indicate your acceptance of this License to do so, and
   all its terms and conditions for copying, distributing or modifying
   the Program or works based on it.
   
   @item
   Each time you redistribute the Program (or any work based on the
   Program), the recipient automatically receives a license from the
   original licensor to copy, distribute or modify the Program subject to
   these terms and conditions.  You may not impose any further
   restrictions on the recipients' exercise of the rights granted herein.
   You are not responsible for enforcing compliance by third parties to
   this License.
   
   @item
   If, as a consequence of a court judgment or allegation of patent
   infringement or for any other reason (not limited to patent issues),
   conditions are imposed on you (whether by court order, agreement or
   otherwise) that contradict the conditions of this License, they do not
   excuse you from the conditions of this License.  If you cannot
   distribute so as to satisfy simultaneously your obligations under this
   License and any other pertinent obligations, then as a consequence you
   may not distribute the Program at all.  For example, if a patent
   license would not permit royalty-free redistribution of the Program by
   all those who receive copies directly or indirectly through you, then
   the only way you could satisfy both it and this License would be to
   refrain entirely from distribution of the Program.
   
   If any portion of this section is held invalid or unenforceable under
   any particular circumstance, the balance of the section is intended to
   apply and the section as a whole is intended to apply in other
   circumstances.
   
   It is not the purpose of this section to induce you to infringe any
   patents or other property right claims or to contest validity of any
   such claims; this section has the sole purpose of protecting the
   integrity of the free software distribution system, which is
   implemented by public license practices.  Many people have made
   generous contributions to the wide range of software distributed
   through that system in reliance on consistent application of that
   system; it is up to the author/donor to decide if he or she is willing
   to distribute software through any other system and a licensee cannot
   impose that choice.
   
   This section is intended to make thoroughly clear what is believed to
   be a consequence of the rest of this License.
   
   @item
   If the distribution and/or use of the Program is restricted in
   certain countries either by patents or by copyrighted interfaces, the
   original copyright holder who places the Program under this License
   may add an explicit geographical distribution limitation excluding
   those countries, so that distribution is permitted only in or among
   countries not thus excluded.  In such case, this License incorporates
   the limitation as if written in the body of this License.
   
   @item
   The Free Software Foundation may publish revised and/or new versions
   of the General Public License from time to time.  Such new versions will
   be similar in spirit to the present version, but may differ in detail to
   address new problems or concerns.
   
   Each version is given a distinguishing version number.  If the Program
   specifies a version number of this License which applies to it and ``any
   later version'', you have the option of following the terms and conditions
   either of that version or of any later version published by the Free
   Software Foundation.  If the Program does not specify a version number of
   this License, you may choose any version ever published by the Free Software
   Foundation.
   
   @item
   If you wish to incorporate parts of the Program into other free
   programs whose distribution conditions are different, write to the author
   to ask for permission.  For software which is copyrighted by the Free
   Software Foundation, write to the Free Software Foundation; we sometimes
   make exceptions for this.  Our decision will be guided by the two goals
   of preserving the free status of all derivatives of our free software and
   of promoting the sharing and reuse of software generally.
   
   @iftex
   @heading NO WARRANTY
   @end iftex
   @ifinfo
   @center NO WARRANTY
   @end ifinfo
   
   @item
   BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY
   FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW.  EXCEPT WHEN
   OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES
   PROVIDE THE PROGRAM ``AS IS'' WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED
   OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
   MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.  THE ENTIRE RISK AS
   TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU.  SHOULD THE
   PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING,
   REPAIR OR CORRECTION.
   
   @item
   IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
   WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR
   REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES,
   INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING
   OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED
   TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY
   YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER
   PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE
   POSSIBILITY OF SUCH DAMAGES.
   @end enumerate
   
   @iftex
   @heading END OF TERMS AND CONDITIONS
   @end iftex
   @ifinfo
   @center END OF TERMS AND CONDITIONS
   @end ifinfo
   
   @page
   @unnumberedsec How to Apply These Terms to Your New Programs
   
     If you develop a new program, and you want it to be of the greatest
   possible use to the public, the best way to achieve this is to make it
   free software which everyone can redistribute and change under these terms.
   
     To do so, attach the following notices to the program.  It is safest
   to attach them to the start of each source file to most effectively
   convey the exclusion of warranty; and each file should have at least
   the ``copyright'' line and a pointer to where the full notice is found.
   
   @smallexample
   @var{one line to give the program's name and a brief idea of what it does.}
   Copyright (C) 19@var{yy}  @var{name of author}
   
   This program is free software; you can redistribute it and/or modify 
   it under the terms of the GNU General Public License as published by 
   the Free Software Foundation; either version 2 of the License, or 
   (at your option) any later version.
   
   This program is distributed in the hope that it will be useful,
   but WITHOUT ANY WARRANTY; without even the implied warranty of
   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
   GNU General Public License for more details.
   
   You should have received a copy of the GNU General Public License
   along with this program; if not, write to the Free Software
   Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
   @end smallexample
   
   Also add information on how to contact you by electronic and paper mail.
   
   If the program is interactive, make it output a short notice like this
   when it starts in an interactive mode:
   
   @smallexample
   Gnomovision version 69, Copyright (C) 19@var{yy} @var{name of author}
   Gnomovision comes with ABSOLUTELY NO WARRANTY; for details
   type `show w'.  
   This is free software, and you are welcome to redistribute it 
   under certain conditions; type `show c' for details.
   @end smallexample
   
   The hypothetical commands @samp{show w} and @samp{show c} should show
   the appropriate parts of the General Public License.  Of course, the
   commands you use may be called something other than @samp{show w} and
   @samp{show c}; they could even be mouse-clicks or menu items---whatever
   suits your program.
   
   You should also get your employer (if you work as a programmer) or your
   school, if any, to sign a ``copyright disclaimer'' for the program, if
   necessary.  Here is a sample; alter the names:
   
   @smallexample
   Yoyodyne, Inc., hereby disclaims all copyright interest in the program
   `Gnomovision' (which makes passes at compilers) written by James Hacker.
   
   @var{signature of Ty Coon}, 1 April 1989
   Ty Coon, President of Vice
   @end smallexample
   
   This General Public License does not permit incorporating your program into
   proprietary programs.  If your program is a subroutine library, you may
   consider it more useful to permit linking proprietary applications with the
   library.  If this is what you want to do, use the GNU Library General
   Public License instead of this License.
   
   @iftex
   @node    Preface
   @comment node-name,     next,           previous, up
 @unnumbered Preface  @unnumbered Preface
 This manual documents GNU Forth. The reader is expected to know  @cindex Preface
   This manual documents Gforth. The reader is expected to know
 Forth. This manual is primarily a reference manual. @xref{Other Books}  Forth. This manual is primarily a reference manual. @xref{Other Books}
 for introductory material.  for introductory material.
 @end iftex  @end iftex
   
 @node    Goals, Other Books, License, Top  @node    Goals, Other Books, License, Top
 @comment node-name,     next,           previous, up  @comment node-name,     next,           previous, up
 @chapter Goals of GNU Forth  @chapter Goals of Gforth
 @cindex Goals  @cindex Goals
 The goal of the GNU Forth Project is to develop a standard model for  The goal of the Gforth Project is to develop a standard model for
 ANSI Forth. This can be split into several subgoals:  ANSI Forth. This can be split into several subgoals:
   
 @itemize @bullet  @itemize @bullet
 @item  @item
 GNU Forth should conform to the ANSI Forth standard.  Gforth should conform to the ANSI Forth standard.
 @item  @item
 It should be a model, i.e. it should define all the  It should be a model, i.e. it should define all the
 implementation-dependent things.  implementation-dependent things.
Line 121  It should become standard, i.e. widely a Line 516  It should become standard, i.e. widely a
 is the most difficult one.  is the most difficult one.
 @end itemize  @end itemize
   
 To achieve these goals GNU Forth should be  To achieve these goals Gforth should be
 @itemize @bullet  @itemize @bullet
 @item  @item
 Similar to previous models (fig-Forth, F83)  Similar to previous models (fig-Forth, F83)
Line 137  Free. Line 532  Free.
 Available on many machines/easy to port.  Available on many machines/easy to port.
 @end itemize  @end itemize
   
 Have we achieved these goals? GNU Forth conforms to the ANS Forth  Have we achieved these goals? Gforth conforms to the ANS Forth
 standard; it may be considered a model, but we have not yet documented  standard. It may be considered a model, but we have not yet documented
 which parts of the model are stable and which parts we are likely to  which parts of the model are stable and which parts we are likely to
 change; it certainly has not yet become a de facto standard. It has some  change. It certainly has not yet become a de facto standard. It has some
 similarities and some differences to previous models; It has some  similarities and some differences to previous models. It has some
 powerful features, but not yet everything that we envisioned; on RISCs  powerful features, but not yet everything that we envisioned. We
 it is as fast as interpreters programmed in assembly, on  certainly have achieved our execution speed goals (@pxref{Performance}).
 register-starved machines it is not so fast, but still faster than any  It is free and available on many machines.
 other C-based interpretive implementation; it is free and available on  
 many machines.  
   
 @node Other Books, Invocation, Goals, Top  @node Other Books, Invocation, Goals, Top
 @chapter Other books on ANS Forth  @chapter Other books on ANS Forth
   
 As the standard is relatively new, there are not many books out yet. It  As the standard is relatively new, there are not many books out yet. It
 is not recommended to learn Forth by using GNU Forth and a book that is  is not recommended to learn Forth by using Gforth and a book that is
 not written for ANS Forth, as you will not know your mistakes from the  not written for ANS Forth, as you will not know your mistakes from the
 deviations of the book.  deviations of the book.
   
 There is, of course, the standard, the definite reference if you want to  There is, of course, the standard, the definite reference if you want to
 write ANS Forth programs. It will be available in printed form from  write ANS Forth programs. It is available in printed form from the
 Global Engineering Documents !! somtime in spring or summer 1994. If you  National Standards Institute Sales Department (Tel.: USA (212) 642-4900;
 are lucky, you can still get dpANS6 (the draft that was approved as  Fax.: USA (212) 302-1286) as document @cite{X3.215-1994} for about $200. You
 standard) by aftp from ftp.uu.net:/vendor/minerva/x3j14.  can also get it from Global Engineering Documents (Tel.: USA (800)
   854-7179; Fax.: (303) 843-9880) for about $300.
   
   @cite{dpANS6}, the last draft of the standard, which was then submitted to ANSI
   for publication is available electronically and for free in some MS Word
   format, and it has been converted to HTML. Some pointers to these
   versions can be found through
   @*@file{http://www.complang.tuwien.ac.at/projects/forth.html}.
   
 @cite{Forth: The new model} by Jack Woehr (!! Publisher) is an  @cite{Forth: The new model} by Jack Woehr (Prentice-Hall, 1993) is an
 introductory book based on a draft version of the standard. It does not  introductory book based on a draft version of the standard. It does not
 cover the whole standard. It also contains interesting background  cover the whole standard. It also contains interesting background
 information (Jack Woehr was in the ANS Forth Technical Committe). It is  information (Jack Woehr was in the ANS Forth Technical Committe). It is
Line 173  other languages should find it ok. Line 573  other languages should find it ok.
 @chapter Invocation  @chapter Invocation
   
 You will usually just say @code{gforth}. In many other cases the default  You will usually just say @code{gforth}. In many other cases the default
 GNU Forth image will be invoked like this:  Gforth image will be invoked like this:
   
 @example  @example
 gforth [files] [-e forth-code]  gforth [files] [-e forth-code]
Line 193  line. They are: Line 593  line. They are:
   
 @table @code  @table @code
 @item --image-file @var{file}  @item --image-file @var{file}
   @item -i @var{file}
 Loads the Forth image @var{file} instead of the default  Loads the Forth image @var{file} instead of the default
 @file{gforth.fi}.  @file{gforth.fi}.
   
 @item --path @var{path}  @item --path @var{path}
   @item -p @var{path}
 Uses @var{path} for searching the image file and Forth source code  Uses @var{path} for searching the image file and Forth source code
 files instead of the default in the environment variable  files instead of the default in the environment variable
 @code{GFORTHPATH} or the path specified at installation time (typically  @code{GFORTHPATH} or the path specified at installation time (typically
Line 245  Forth words, you have to quote them or u Line 647  Forth words, you have to quote them or u
 after processing the command line (instead of entering interactive mode)  after processing the command line (instead of entering interactive mode)
 append @code{-e bye} to the command line.  append @code{-e bye} to the command line.
   
   If you have several versions of Gforth installed, @code{gforth} will
   invoke the version that was installed last. @code{gforth-@var{version}}
   invokes a specific version. You may want to use the option
   @code{--path}, if your environment contains the variable
   @code{GFORTHPATH}.
   
 Not yet implemented:  Not yet implemented:
 On startup the system first executes the system initialization file  On startup the system first executes the system initialization file
 (unless the option @code{--no-init-file} is given; note that the system  (unless the option @code{--no-init-file} is given; note that the system
Line 253  the user initialization file @file{.gfor Line 661  the user initialization file @file{.gfor
 option @code{--no-rc} is given; this file is first searched in @file{.},  option @code{--no-rc} is given; this file is first searched in @file{.},
 then in @file{~}, then in the normal path (see above).  then in @file{~}, then in the normal path (see above).
   
 @node Words,  , Invocation, Top  @node Words, ANS conformance, Invocation, Top
 @chapter Forth Words  @chapter Forth Words
   
 @menu  @menu
 * Notation::  * Notation::                    
 * Arithmetic::  * Arithmetic::                  
 * Stack Manipulation::  * Stack Manipulation::          
 * Memory access::  * Memory access::               
 * Control Structures::  * Control Structures::          
 * Local Variables::  * Locals::                      
 * Defining Words::  * Defining Words::              
 * Vocabularies::  * Wordlists::                   
 * Files::  * Files::                       
 * Blocks::  * Blocks::                      
 * Other I/O::  * Other I/O::                   
 * Programming Tools::  * Programming Tools::           
   * Assembler and Code words::    
   * Threading Words::             
 @end menu  @end menu
   
 @node Notation, Arithmetic, Words, Words  @node Notation, Arithmetic, Words, Words
Line 277  then in @file{~}, then in the normal pat Line 687  then in @file{~}, then in the normal pat
 The Forth words are described in this section in the glossary notation  The Forth words are described in this section in the glossary notation
 that has become a de-facto standard for Forth texts, i.e.  that has become a de-facto standard for Forth texts, i.e.
   
 @quotation  @format
 @var{word}     @var{Stack effect}   @var{wordset}   @var{pronunciation}  @var{word}     @var{Stack effect}   @var{wordset}   @var{pronunciation}
   @end format
 @var{Description}  @var{Description}
 @end quotation  
   
 @table @var  @table @var
 @item word  @item word
 The name of the word. BTW, GNU Forth is case insensitive, so you can  The name of the word. BTW, Gforth is case insensitive, so you can
 type the words in in lower case.  type the words in in lower case (However, @pxref{core-idef}).
   
 @item Stack effect  @item Stack effect
 The stack effect is written in the notation @code{@var{before} --  The stack effect is written in the notation @code{@var{before} --
 @var{after}}, where @var{before} and @var{after} describe the top of  @var{after}}, where @var{before} and @var{after} describe the top of
 stack entries before and after the execution of the word. The rest of  stack entries before and after the execution of the word. The rest of
 the stack is not touched by the word. The top of stack is rightmost,  the stack is not touched by the word. The top of stack is rightmost,
 i.e., a stack sequence is written as it is typed in. Note that GNU Forth  i.e., a stack sequence is written as it is typed in. Note that Gforth
 uses a separate floating point stack, but a unified stack  uses a separate floating point stack, but a unified stack
 notation. Also, return stack effects are not shown in @var{stack  notation. Also, return stack effects are not shown in @var{stack
 effect}, but in @var{Description}. The name of a stack item describes  effect}, but in @var{Description}. The name of a stack item describes
 the type and/or the function of the item. See below for a discussion of  the type and/or the function of the item. See below for a discussion of
 the types.  the types.
   
   All words have two stack effects: A compile-time stack effect and a
   run-time stack effect. The compile-time stack-effect of most words is
   @var{ -- }. If the compile-time stack-effect of a word deviates from
   this standard behaviour, or the word does other unusual things at
   compile time, both stack effects are shown; otherwise only the run-time
   stack effect is shown.
   
 @item pronunciation  @item pronunciation
 How the word is pronounced  How the word is pronounced
   
Line 308  system need not support all of them. So, Line 725  system need not support all of them. So,
 uses the more portable it will be in theory. However, we suspect that  uses the more portable it will be in theory. However, we suspect that
 most ANS Forth systems on personal machines will feature all  most ANS Forth systems on personal machines will feature all
 wordsets. Words that are not defined in the ANS standard have  wordsets. Words that are not defined in the ANS standard have
 @code{gforth} as wordset.  @code{gforth} or @code{gforth-internal} as wordset. @code{gforth}
   describes words that will work in future releases of Gforth;
   @code{gforth-internal} words are more volatile. Environmental query
   strings are also displayed like words; you can recognize them by the
   @code{environment} in the wordset field.
   
 @item Description  @item Description
 A description of the behaviour of the word.  A description of the behaviour of the word.
 @end table  @end table
   
 The name of a stack item corresponds in the following way with its type:  The type of a stack item is specified by the character(s) the name
   starts with:
   
 @table @code  @table @code
 @item name starts with  
 Type  
 @item f  @item f
 Bool, i.e. @code{false} or @code{true}.  Bool, i.e. @code{false} or @code{true}.
 @item c  @item c
Line 353  Wordlist ID, same size as Cell Line 773  Wordlist ID, same size as Cell
 Pointer to a name structure  Pointer to a name structure
 @end table  @end table
   
 @node Arithmetic,  , Notation, Words  @node Arithmetic, Stack Manipulation, Notation, Words
 @section Arithmetic  @section Arithmetic
 Forth arithmetic is not checked, i.e., you will not hear about integer  Forth arithmetic is not checked, i.e., you will not hear about integer
 overflow on addition or multiplication, you may hear about division by  overflow on addition or multiplication, you may hear about division by
Line 363  corresponds to @code{2 1 -}. Forth offer Line 783  corresponds to @code{2 1 -}. Forth offer
 operators. If you perform division with potentially negative operands,  operators. If you perform division with potentially negative operands,
 you do not want to use @code{/} or @code{/mod} with its undefined  you do not want to use @code{/} or @code{/mod} with its undefined
 behaviour, but rather @code{fm/mod} or @code{sm/mod} (probably the  behaviour, but rather @code{fm/mod} or @code{sm/mod} (probably the
 former).  former, @pxref{Mixed precision}).
   
   @menu
   * Single precision::            
   * Bitwise operations::          
   * Mixed precision::             operations with single and double-cell integers
   * Double precision::            Double-cell integer arithmetic
   * Floating Point::              
   @end menu
   
   @node Single precision, Bitwise operations, Arithmetic, Arithmetic
 @subsection Single precision  @subsection Single precision
 doc-+  doc-+
 doc--  doc--
Line 377  doc-abs Line 806  doc-abs
 doc-min  doc-min
 doc-max  doc-max
   
   @node Bitwise operations, Mixed precision, Single precision, Arithmetic
 @subsection Bitwise operations  @subsection Bitwise operations
 doc-and  doc-and
 doc-or  doc-or
Line 385  doc-invert Line 815  doc-invert
 doc-2*  doc-2*
 doc-2/  doc-2/
   
   @node Mixed precision, Double precision, Bitwise operations, Arithmetic
 @subsection Mixed precision  @subsection Mixed precision
 doc-m+  doc-m+
 doc-*/  doc-*/
Line 396  doc-um/mod Line 827  doc-um/mod
 doc-fm/mod  doc-fm/mod
 doc-sm/rem  doc-sm/rem
   
   @node Double precision, Floating Point, Mixed precision, Arithmetic
 @subsection Double precision  @subsection Double precision
   
   The outer (aka text) interpreter converts numbers containing a dot into
   a double precision number. Note that only numbers with the dot as last
   character are standard-conforming.
   
 doc-d+  doc-d+
 doc-d-  doc-d-
 doc-dnegate  doc-dnegate
Line 404  doc-dabs Line 841  doc-dabs
 doc-dmin  doc-dmin
 doc-dmax  doc-dmax
   
 @node Stack Manipulation,,,  @node Floating Point,  , Double precision, Arithmetic
   @subsection Floating Point
   
   The format of floating point numbers recognized by the outer (aka text)
   interpreter is: a signed decimal number, possibly containing a decimal
   point (@code{.}), followed by @code{E} or @code{e}, optionally followed
   by a signed integer (the exponent). E.g., @code{1e} ist the same as
   @code{+1.0e+1}. Note that a number without @code{e}
   is not interpreted as floating-point number, but as double (if the
   number contains a @code{.}) or single precision integer. Also,
   conversions between string and floating point numbers always use base
   10, irrespective of the value of @code{BASE}. If @code{BASE} contains a
   value greater then 14, the @code{E} may be interpreted as digit and the
   number will be interpreted as integer, unless it has a signed exponent
   (both @code{+} and @code{-} are allowed as signs).
   
   Angles in floating point operations are given in radians (a full circle
   has 2 pi radians). Note, that Gforth has a separate floating point
   stack, but we use the unified notation.
   
   Floating point numbers have a number of unpleasant surprises for the
   unwary (e.g., floating point addition is not associative) and even a few
   for the wary. You should not use them unless you know what you are doing
   or you don't care that the results you get are totally bogus. If you
   want to learn about the problems of floating point numbers (and how to
   avoid them), you might start with @cite{David Goldberg, What Every
   Computer Scientist Should Know About Floating-Point Arithmetic, ACM
   Computing Surveys 23(1):5@minus{}48, March 1991}.
   
   doc-f+
   doc-f-
   doc-f*
   doc-f/
   doc-fnegate
   doc-fabs
   doc-fmax
   doc-fmin
   doc-floor
   doc-fround
   doc-f**
   doc-fsqrt
   doc-fexp
   doc-fexpm1
   doc-fln
   doc-flnp1
   doc-flog
   doc-falog
   doc-fsin
   doc-fcos
   doc-fsincos
   doc-ftan
   doc-fasin
   doc-facos
   doc-fatan
   doc-fatan2
   doc-fsinh
   doc-fcosh
   doc-ftanh
   doc-fasinh
   doc-facosh
   doc-fatanh
   
   @node Stack Manipulation, Memory access, Arithmetic, Words
 @section Stack Manipulation  @section Stack Manipulation
   
 gforth has a data stack (aka parameter stack) for characters, cells,  Gforth has a data stack (aka parameter stack) for characters, cells,
 addresses, and double cells, a floating point stack for floating point  addresses, and double cells, a floating point stack for floating point
 numbers, a return stack for storing the return addresses of colon  numbers, a return stack for storing the return addresses of colon
 definitions and other data, and a locals stack for storing local  definitions and other data, and a locals stack for storing local
Line 417  theoretically keep floating point number Line 916  theoretically keep floating point number
 additional difficulty, you don't know how many cells a floating point  additional difficulty, you don't know how many cells a floating point
 number takes. It is reportedly possible to write words in a way that  number takes. It is reportedly possible to write words in a way that
 they work also for a unified stack model, but we do not recommend trying  they work also for a unified stack model, but we do not recommend trying
 it. Also, a Forth system is allowed to keep the local variables on the  it. Instead, just say that your program has an environmental dependency
   on a separate FP stack.
   
   Also, a Forth system is allowed to keep the local variables on the
 return stack. This is reasonable, as local variables usually eliminate  return stack. This is reasonable, as local variables usually eliminate
 the need to use the return stack explicitly. So, if you want to produce  the need to use the return stack explicitly. So, if you want to produce
 a standard complying program and if you are using local variables in a  a standard complying program and if you are using local variables in a
 word, forget about return stack manipulations in that word (see the  word, forget about return stack manipulations in that word (see the
 standard document for the exact rules).  standard document for the exact rules).
   
   @menu
   * Data stack::                  
   * Floating point stack::        
   * Return stack::                
   * Locals stack::                
   * Stack pointer manipulation::  
   @end menu
   
   @node Data stack, Floating point stack, Stack Manipulation, Stack Manipulation
 @subsection Data stack  @subsection Data stack
 doc-drop  doc-drop
 doc-nip  doc-nip
Line 444  doc-2tuck Line 955  doc-2tuck
 doc-2swap  doc-2swap
 doc-2rot  doc-2rot
   
   @node Floating point stack, Return stack, Data stack, Stack Manipulation
 @subsection Floating point stack  @subsection Floating point stack
 doc-fdrop  doc-fdrop
 doc-fnip  doc-fnip
Line 453  doc-ftuck Line 965  doc-ftuck
 doc-fswap  doc-fswap
 doc-frot  doc-frot
   
   @node Return stack, Locals stack, Floating point stack, Stack Manipulation
 @subsection Return stack  @subsection Return stack
 doc->r  doc->r
 doc-r>  doc-r>
Line 463  doc-2r> Line 976  doc-2r>
 doc-2r@  doc-2r@
 doc-2rdrop  doc-2rdrop
   
   @node Locals stack, Stack pointer manipulation, Return stack, Stack Manipulation
 @subsection Locals stack  @subsection Locals stack
   
   @node Stack pointer manipulation,  , Locals stack, Stack Manipulation
 @subsection Stack pointer manipulation  @subsection Stack pointer manipulation
 doc-sp@  doc-sp@
 doc-sp!  doc-sp!
Line 475  doc-rp! Line 990  doc-rp!
 doc-lp@  doc-lp@
 doc-lp!  doc-lp!
   
 @node Memory access  @node Memory access, Control Structures, Stack Manipulation, Words
 @section Memory access  @section Memory access
   
   @menu
   * Stack-Memory transfers::      
   * Address arithmetic::          
   * Memory block access::         
   @end menu
   
   @node Stack-Memory transfers, Address arithmetic, Memory access, Memory access
 @subsection Stack-Memory transfers  @subsection Stack-Memory transfers
   
 doc-@  doc-@
Line 494  doc-sf! Line 1016  doc-sf!
 doc-df@  doc-df@
 doc-df!  doc-df!
   
   @node Address arithmetic, Memory block access, Stack-Memory transfers, Memory access
 @subsection Address arithmetic  @subsection Address arithmetic
   
 ANS Forth does not specify the sizes of the data types. Instead, it  ANS Forth does not specify the sizes of the data types. Instead, it
Line 509  must only occur at specific addresses; e Line 1032  must only occur at specific addresses; e
 accessed at addresses divisible by 4. Even if a machine allows unaligned  accessed at addresses divisible by 4. Even if a machine allows unaligned
 accesses, it can usually perform aligned accesses faster.   accesses, it can usually perform aligned accesses faster. 
   
 For the performance-concious: alignment operations are usually only  For the performance-conscious: alignment operations are usually only
 necessary during the definition of a data structure, not during the  necessary during the definition of a data structure, not during the
 (more frequent) accesses to it.  (more frequent) accesses to it.
   
Line 519  char-aligned have no use in the standard Line 1042  char-aligned have no use in the standard
 created.  created.
   
 The standard guarantees that addresses returned by @code{CREATE}d words  The standard guarantees that addresses returned by @code{CREATE}d words
 are cell-aligned; in addition, gforth guarantees that these addresses  are cell-aligned; in addition, Gforth guarantees that these addresses
 are aligned for all purposes.  are aligned for all purposes.
   
   Note that the standard defines a word @code{char}, which has nothing to
   do with address arithmetic.
   
 doc-chars  doc-chars
 doc-char+  doc-char+
 doc-cells  doc-cells
Line 540  doc-dfloats Line 1066  doc-dfloats
 doc-dfloat+  doc-dfloat+
 doc-dfalign  doc-dfalign
 doc-dfaligned  doc-dfaligned
   doc-maxalign
   doc-maxaligned
   doc-cfalign
   doc-cfaligned
 doc-address-unit-bits  doc-address-unit-bits
   
   @node Memory block access,  , Address arithmetic, Memory access
 @subsection Memory block access  @subsection Memory block access
   
 doc-move  doc-move
Line 555  doc-cmove> Line 1086  doc-cmove>
 doc-fill  doc-fill
 doc-blank  doc-blank
   
 @node Control Structures  @node Control Structures, Locals, Memory access, Words
 @section Control Structures  @section Control Structures
   
 Control structures in Forth cannot be used in interpret state, only in  Control structures in Forth cannot be used in interpret state, only in
Line 563  compile state, i.e., in a colon definiti Line 1094  compile state, i.e., in a colon definiti
 limitation, but have not seen a satisfying way around it yet, although  limitation, but have not seen a satisfying way around it yet, although
 many schemes have been proposed.  many schemes have been proposed.
   
   @menu
   * Selection::                   
   * Simple Loops::                
   * Counted Loops::               
   * Arbitrary control structures::  
   * Calls and returns::           
   * Exception Handling::          
   @end menu
   
   @node Selection, Simple Loops, Control Structures, Control Structures
 @subsection Selection  @subsection Selection
   
 @example  @example
Line 581  ELSE Line 1122  ELSE
 ENDIF  ENDIF
 @end example  @end example
   
 You can use @code{THEN} instead of {ENDIF}. Indeed, @code{THEN} is  You can use @code{THEN} instead of @code{ENDIF}. Indeed, @code{THEN} is
 standard, and @code{ENDIF} is not, although it is quite popular. We  standard, and @code{ENDIF} is not, although it is quite popular. We
 recommend using @code{ENDIF}, because it is less confusing for people  recommend using @code{ENDIF}, because it is less confusing for people
 who also know other languages (and is not prone to reinforcing negative  who also know other languages (and is not prone to reinforcing negative
Line 600  system that only supplies @code{THEN} is Line 1141  system that only supplies @code{THEN} is
 Forth's @code{THEN} has the meaning 2b, whereas @code{THEN} in Pascal  Forth's @code{THEN} has the meaning 2b, whereas @code{THEN} in Pascal
 and many other programming languages has the meaning 3d.]  and many other programming languages has the meaning 3d.]
   
 We also provide the words @code{?dup-if} and @code{?dup-0=-if}, so you  Gforth also provides the words @code{?dup-if} and @code{?dup-0=-if}, so
 can avoid using @code{?dup}.  you can avoid using @code{?dup}. Using these alternatives is also more
   efficient than using @code{?dup}. Definitions in plain standard Forth
   for @code{ENDIF}, @code{?DUP-IF} and @code{?DUP-0=-IF} are provided in
   @file{compat/control.fs}.
   
 @example  @example
 @var{n}  @var{n}
 CASE  CASE
   @var{n1} OF @var{code1} ENDOF    @var{n1} OF @var{code1} ENDOF
   @var{n2} OF @var{code2} ENDOF    @var{n2} OF @var{code2} ENDOF
   @dots    @dots{}
 ENDCASE  ENDCASE
 @end example  @end example
   
Line 617  Executes the first @var{codei}, where th Line 1161  Executes the first @var{codei}, where th
 the last @code{ENDOF}. It may use @var{n}, which is on top of the stack,  the last @code{ENDOF}. It may use @var{n}, which is on top of the stack,
 but must not consume it.  but must not consume it.
   
   @node Simple Loops, Counted Loops, Selection, Control Structures
 @subsection Simple Loops  @subsection Simple Loops
   
 @example  @example
Line 648  AGAIN Line 1193  AGAIN
   
 This is an endless loop.  This is an endless loop.
   
   @node Counted Loops, Arbitrary control structures, Simple Loops, Control Structures
 @subsection Counted Loops  @subsection Counted Loops
   
 The basic counted loop is:  The basic counted loop is:
Line 685  There are several variations on the coun Line 1231  There are several variations on the coun
   
 @code{LEAVE} leaves the innermost counted loop immediately.  @code{LEAVE} leaves the innermost counted loop immediately.
   
   If @var{start} is greater than @var{limit}, a @code{?DO} loop is entered
   (and @code{LOOP} iterates until they become equal by wrap-around
   arithmetic). This behaviour is usually not what you want. Therefore,
   Gforth offers @code{+DO} and @code{U+DO} (as replacements for
   @code{?DO}), which do not enter the loop if @var{start} is greater than
   @var{limit}; @code{+DO} is for signed loop parameters, @code{U+DO} for
   unsigned loop parameters.
   
 @code{LOOP} can be replaced with @code{@var{n} +LOOP}; this updates the  @code{LOOP} can be replaced with @code{@var{n} +LOOP}; this updates the
 index by @var{n} instead of by 1. The loop is terminated when the border  index by @var{n} instead of by 1. The loop is terminated when the border
 between @var{limit-1} and @var{limit} is crossed. E.g.:  between @var{limit-1} and @var{limit} is crossed. E.g.:
   
 @code{4 0 ?DO  i .  2 +LOOP}   prints @code{0 2}  @code{4 0 +DO  i .  2 +LOOP}   prints @code{0 2}
   
 @code{4 1 ?DO  i .  2 +LOOP}   prints @code{1 3}  @code{4 1 +DO  i .  2 +LOOP}   prints @code{1 3}
   
 The behaviour of @code{@var{n} +LOOP} is peculiar when @var{n} is negative:  The behaviour of @code{@var{n} +LOOP} is peculiar when @var{n} is negative:
   
Line 699  The behaviour of @code{@var{n} +LOOP} is Line 1253  The behaviour of @code{@var{n} +LOOP} is
   
 @code{ 0 0 ?DO  i .  -1 +LOOP}  prints nothing  @code{ 0 0 ?DO  i .  -1 +LOOP}  prints nothing
   
 Therefore we recommend avoiding using @code{@var{n} +LOOP} with negative  Therefore we recommend avoiding @code{@var{n} +LOOP} with negative
 @var{n}. One alternative is @code{@var{n} S+LOOP}, where the negative  @var{n}. One alternative is @code{@var{u} -LOOP}, which reduces the
 case behaves symmetrical to the positive case:  index by @var{u} each iteration. The loop is terminated when the border
   between @var{limit+1} and @var{limit} is crossed. Gforth also provides
 @code{-2 0 ?DO  i .  -1 +LOOP}  prints @code{0 -1}  @code{-DO} and @code{U-DO} for down-counting loops. E.g.:
   
 @code{-1 0 ?DO  i .  -1 +LOOP}  prints @code{0}  @code{-2 0 -DO  i .  1 -LOOP}  prints @code{0 -1}
   
 @code{ 0 0 ?DO  i .  -1 +LOOP}  prints nothing  @code{-1 0 -DO  i .  1 -LOOP}  prints @code{0}
   
 The loop is terminated when the border between @var{limit@minus{}sgn(n)} and  @code{ 0 0 -DO  i .  1 -LOOP}  prints nothing
 @var{limit} is crossed. However, @code{S+LOOP} is not part of the ANS  
 Forth standard.  Unfortunately, @code{+DO}, @code{U+DO}, @code{-DO}, @code{U-DO} and
   @code{-LOOP} are not in the ANS Forth standard. However, an
 @code{?DO} can be replaced by @code{DO}. @code{DO} enters the loop even  implementation for these words that uses only standard words is provided
 when the start and the limit value are equal. We do not recommend using  in @file{compat/loops.fs}.
 @code{DO}. It will just give you maintenance troubles.  
   @code{?DO} can also be replaced by @code{DO}. @code{DO} always enters
   the loop, independent of the loop parameters. Do not use @code{DO}, even
   if you know that the loop is entered in any case. Such knowledge tends
   to become invalid during maintenance of a program, and then the
   @code{DO} will make trouble.
   
 @code{UNLOOP} is used to prepare for an abnormal loop exit, e.g., via  @code{UNLOOP} is used to prepare for an abnormal loop exit, e.g., via
 @code{EXIT}. @code{UNLOOP} removes the loop control parameters from the  @code{EXIT}. @code{UNLOOP} removes the loop control parameters from the
Line 729  FOR Line 1288  FOR
 NEXT  NEXT
 @end example  @end example
 This is the preferred loop of native code compiler writers who are too  This is the preferred loop of native code compiler writers who are too
 lazy to optimize @code{?DO} loops properly. In GNU Forth, this loop  lazy to optimize @code{?DO} loops properly. In Gforth, this loop
 iterates @var{n+1} times; @code{i} produces values starting with @var{n}  iterates @var{n+1} times; @code{i} produces values starting with @var{n}
 and ending with 0. Other Forth systems may behave differently, even if  and ending with 0. Other Forth systems may behave differently, even if
 they support @code{FOR} loops.  they support @code{FOR} loops. To avoid problems, don't use @code{FOR}
   loops.
   
   @node Arbitrary control structures, Calls and returns, Counted Loops, Control Structures
 @subsection Arbitrary control structures  @subsection Arbitrary control structures
   
 ANS Forth permits and supports using control structures in a non-nested  ANS Forth permits and supports using control structures in a non-nested
 way. Information about incomplete control structures is stored on the  way. Information about incomplete control structures is stored on the
 control-flow stack. This stack may be implemented on the Forth data  control-flow stack. This stack may be implemented on the Forth data
 stack, and this is what we have done in gforth.  stack, and this is what we have done in Gforth.
   
 An @i{orig} entry represents an unresolved forward branch, a @i{dest}  An @i{orig} entry represents an unresolved forward branch, a @i{dest}
 entry represents a backward branch target. A few words are the basis for  entry represents a backward branch target. A few words are the basis for
Line 755  doc-again Line 1316  doc-again
 doc-cs-pick  doc-cs-pick
 doc-cs-roll  doc-cs-roll
   
 On many systems control-flow stack items take one word, in gforth they  On many systems control-flow stack items take one word, in Gforth they
 currently take three (this may change in the future). Therefore it is a  currently take three (this may change in the future). Therefore it is a
 really good idea to manipulate the control flow stack with  really good idea to manipulate the control flow stack with
 @code{cs-pick} and @code{cs-roll}, not with data stack manipulation  @code{cs-pick} and @code{cs-roll}, not with data stack manipulation
Line 767  doc-else Line 1328  doc-else
 doc-while  doc-while
 doc-repeat  doc-repeat
   
   Gforth adds some more control-structure words:
   
   doc-endif
   doc-?dup-if
   doc-?dup-0=-if
   
 Counted loop words constitute a separate group of words:  Counted loop words constitute a separate group of words:
   
 doc-?do  doc-?do
   doc-+do
   doc-u+do
   doc--do
   doc-u-do
 doc-do  doc-do
 doc-for  doc-for
 doc-loop  doc-loop
 doc-s+loop  
 doc-+loop  doc-+loop
   doc--loop
 doc-next  doc-next
 doc-leave  doc-leave
 doc-?leave  doc-?leave
 doc-unloop  doc-unloop
 doc-undo  doc-done
   
 The standard does not allow using @code{cs-pick} and @code{cs-roll} on  The standard does not allow using @code{cs-pick} and @code{cs-roll} on
 @i{do-sys}. Our system allows it, but it's your job to ensure that for  @i{do-sys}. Our system allows it, but it's your job to ensure that for
 every @code{?DO} etc. there is exactly one @code{UNLOOP} on any path  every @code{?DO} etc. there is exactly one @code{UNLOOP} on any path
 through the definition (@code{LOOP} etc. compile an @code{UNLOOP} on the  through the definition (@code{LOOP} etc. compile an @code{UNLOOP} on the
 fall-through path). Also, you have to ensure that all @code{LEAVE}s are  fall-through path). Also, you have to ensure that all @code{LEAVE}s are
 resolved (by using one of the loop-ending words or @code{UNDO}).  resolved (by using one of the loop-ending words or @code{DONE}).
   
 Another group of control structure words are  Another group of control structure words are
   
Line 837  while Line 1408  while
 repeat  repeat
 @end example  @end example
   
 That's much easier to read, isn't it? Of course, @code{BEGIN} and  That's much easier to read, isn't it? Of course, @code{REPEAT} and
 @code{WHILE} are predefined, so in this example it would not be  @code{WHILE} are predefined, so in this example it would not be
 necessary to define them.  necessary to define them.
   
   @node Calls and returns, Exception Handling, Arbitrary control structures, Control Structures
 @subsection Calls and returns  @subsection Calls and returns
   
 A definition can be called simply be writing the name of the  A definition can be called simply be writing the name of the
 definition. When the end of the definition is reached, it returns. An earlier return can be forced using  definition. When the end of the definition is reached, it returns. An
   earlier return can be forced using
   
 doc-exit  doc-exit
   
Line 854  primitive compiled by @code{EXIT} is Line 1427  primitive compiled by @code{EXIT} is
   
 doc-;s  doc-;s
   
   @node Exception Handling,  , Calls and returns, Control Structures
 @subsection Exception Handling  @subsection Exception Handling
   
 doc-catch  doc-catch
 doc-throw  doc-throw
   
 @node Locals  @node Locals, Defining Words, Control Structures, Words
 @section Locals  @section Locals
   
 Local variables can make Forth programming more enjoyable and Forth  Local variables can make Forth programming more enjoyable and Forth
Line 868  laden with restrictions. Therefore, we p Line 1442  laden with restrictions. Therefore, we p
 locals wordset, but also our own, more powerful locals wordset (we  locals wordset, but also our own, more powerful locals wordset (we
 implemented the ANS Forth locals wordset through our locals wordset).  implemented the ANS Forth locals wordset through our locals wordset).
   
   The ideas in this section have also been published in the paper
   @cite{Automatic Scoping of Local Variables} by M. Anton Ertl, presented
   at EuroForth '94; it is available at
   @*@file{http://www.complang.tuwien.ac.at/papers/ertl94l.ps.gz}.
   
 @menu  @menu
   * Gforth locals::               
   * ANS Forth locals::            
 @end menu  @end menu
   
 @subsection gforth locals  @node Gforth locals, ANS Forth locals, Locals, Locals
   @subsection Gforth locals
   
 Locals can be defined with  Locals can be defined with
   
Line 914  The name of the local may be preceded by Line 1496  The name of the local may be preceded by
  Ar Bi f* Ai Br f* f+ ;   Ar Bi f* Ai Br f* f+ ;
 @end example  @end example
   
 GNU Forth currently supports cells (@code{W:}, @code{W^}), doubles  Gforth currently supports cells (@code{W:}, @code{W^}), doubles
 (@code{D:}, @code{D^}), floats (@code{F:}, @code{F^}) and characters  (@code{D:}, @code{D^}), floats (@code{F:}, @code{F^}) and characters
 (@code{C:}, @code{C^}) in two flavours: a value-flavoured local (defined  (@code{C:}, @code{C^}) in two flavours: a value-flavoured local (defined
 with @code{W:}, @code{D:} etc.) produces its value and can be changed  with @code{W:}, @code{D:} etc.) produces its value and can be changed
Line 934  locals are initialized with values from Line 1516  locals are initialized with values from
 Currently there is no way to define locals with user-defined data  Currently there is no way to define locals with user-defined data
 structures, but we are working on it.  structures, but we are working on it.
   
 GNU Forth allows defining locals everywhere in a colon definition. This poses the following questions:  Gforth allows defining locals everywhere in a colon definition. This
   poses the following questions:
   
   @menu
   * Where are locals visible by name?::  
   * How long do locals live?::    
   * Programming Style::           
   * Implementation::              
   @end menu
   
   @node Where are locals visible by name?, How long do locals live?, Gforth locals, Gforth locals
 @subsubsection Where are locals visible by name?  @subsubsection Where are locals visible by name?
   
 Basically, the answer is that locals are visible where you would expect  Basically, the answer is that locals are visible where you would expect
Line 994  AHEAD Line 1585  AHEAD
 BEGIN  BEGIN
   x    x
 [ 1 CS-ROLL ] THEN  [ 1 CS-ROLL ] THEN
   { x }    @{ x @}
   ...    ...
 UNTIL  UNTIL
 @end example  @end example
Line 1026  compiler. When the branch to the @code{B Line 1617  compiler. When the branch to the @code{B
 warns the user if it was too optimisitic:  warns the user if it was too optimisitic:
 @example  @example
 IF  IF
   { x }    @{ x @}
 BEGIN  BEGIN
   \ x ?     \ x ? 
 [ 1 cs-roll ] THEN  [ 1 cs-roll ] THEN
Line 1042  is not used in the wrong area by using e Line 1633  is not used in the wrong area by using e
 @example  @example
 IF  IF
   SCOPE    SCOPE
   { x }    @{ x @}
   ENDSCOPE    ENDSCOPE
 BEGIN  BEGIN
 [ 1 cs-roll ] THEN  [ 1 cs-roll ] THEN
Line 1057  If the @code{BEGIN} is not reachable fro Line 1648  If the @code{BEGIN} is not reachable fro
 @code{AHEAD} or @code{EXIT}), the compiler cannot even make an  @code{AHEAD} or @code{EXIT}), the compiler cannot even make an
 optimistic guess, as the locals visible after the @code{BEGIN} may be  optimistic guess, as the locals visible after the @code{BEGIN} may be
 defined later. Therefore, the compiler assumes that no locals are  defined later. Therefore, the compiler assumes that no locals are
 visible after the @code{BEGIN}. However, the useer can use  visible after the @code{BEGIN}. However, the user can use
 @code{ASSUME-LIVE} to make the compiler assume that the same locals are  @code{ASSUME-LIVE} to make the compiler assume that the same locals are
 visible at the BEGIN as at the point where the item was created.  visible at the BEGIN as at the point where the top control-flow stack
   item was created.
   
 doc-assume-live  doc-assume-live
   
 E.g.,  E.g.,
 @example  @example
 { x }  @{ x @}
 AHEAD  AHEAD
 ASSUME-LIVE  ASSUME-LIVE
 BEGIN  BEGIN
Line 1086  rearranging the loop. E.g., the ``most i Line 1678  rearranging the loop. E.g., the ``most i
 arranged into:  arranged into:
 @example  @example
 BEGIN  BEGIN
   { x }    @{ x @}
   ... 0=    ... 0=
 WHILE  WHILE
   x    x
 REPEAT  REPEAT
 @end example  @end example
   
   @node How long do locals live?, Programming Style, Where are locals visible by name?, Gforth locals
 @subsubsection How long do locals live?  @subsubsection How long do locals live?
   
 The right answer for the lifetime question would be: A local lives at  The right answer for the lifetime question would be: A local lives at
Line 1106  languages (e.g., C): The local lives onl Line 1699  languages (e.g., C): The local lives onl
 afterwards its address is invalid (and programs that access it  afterwards its address is invalid (and programs that access it
 afterwards are erroneous).  afterwards are erroneous).
   
   @node Programming Style, Implementation, How long do locals live?, Gforth locals
 @subsubsection Programming Style  @subsubsection Programming Style
   
 The freedom to define locals anywhere has the potential to change  The freedom to define locals anywhere has the potential to change
Line 1117  wrong order, just write a locals definit Line 1711  wrong order, just write a locals definit
 write the items in the order you want.  write the items in the order you want.
   
 This seems a little far-fetched and eliminating stack manipulations is  This seems a little far-fetched and eliminating stack manipulations is
 unlikely to become a conscious programming objective. Still, the  unlikely to become a conscious programming objective. Still, the number
 number of stack manipulations will be reduced dramatically if local  of stack manipulations will be reduced dramatically if local variables
 variables are used liberally (e.g., compare @code{max} in \sect{misc}  are used liberally (e.g., compare @code{max} in @ref{Gforth locals} with
 with a traditional implementation of @code{max}).  a traditional implementation of @code{max}).
   
 This shows one potential benefit of locals: making Forth programs more  This shows one potential benefit of locals: making Forth programs more
 readable. Of course, this benefit will only be realized if the  readable. Of course, this benefit will only be realized if the
Line 1139  E.g., a definition using @code{TO} might Line 1733  E.g., a definition using @code{TO} might
 : strcmp @{ addr1 u1 addr2 u2 -- n @}  : strcmp @{ addr1 u1 addr2 u2 -- n @}
  u1 u2 min 0   u1 u2 min 0
  ?do   ?do
    addr1 c@ addr2 c@ - ?dup     addr1 c@ addr2 c@ -
    if     ?dup-if
      unloop exit       unloop exit
    then     then
    addr1 char+ TO addr1     addr1 char+ TO addr1
Line 1162  are initialized with the right value for Line 1756  are initialized with the right value for
  addr1 addr2   addr1 addr2
  u1 u2 min 0    u1 u2 min 0 
  ?do @{ s1 s2 @}   ?do @{ s1 s2 @}
    s1 c@ s2 c@ - ?dup      s1 c@ s2 c@ -
    if     ?dup-if
      unloop exit       unloop exit
    then     then
    s1 char+ s2 char+     s1 char+ s2 char+
Line 1174  are initialized with the right value for Line 1768  are initialized with the right value for
 Here it is clear from the start that @code{s1} has a different value  Here it is clear from the start that @code{s1} has a different value
 in every loop iteration.  in every loop iteration.
   
   @node Implementation,  , Programming Style, Gforth locals
 @subsubsection Implementation  @subsubsection Implementation
   
 GNU Forth uses an extra locals stack. The most compelling reason for  Gforth uses an extra locals stack. The most compelling reason for
 this is that the return stack is not float-aligned; using an extra stack  this is that the return stack is not float-aligned; using an extra stack
 also eliminates the problems and restrictions of using the return stack  also eliminates the problems and restrictions of using the return stack
 as locals stack. Like the other stacks, the locals stack grows toward  as locals stack. Like the other stacks, the locals stack grows toward
Line 1197 local0 Line 1792 local0
 compile the right specialized version, or the general version, as  compile the right specialized version, or the general version, as
 appropriate:  appropriate:
   
 doc-compile-@@local  doc-compile-@local
 doc-compile-f@@local  doc-compile-f@local
 doc-compile-lp+!  doc-compile-lp+!
   
 Combinations of conditional branches and @code{lp+!#} like  Combinations of conditional branches and @code{lp+!#} like
Line 1211  area and @code{@}} switches it back and Line 1806  area and @code{@}} switches it back and
 initializing code. @code{W:} etc.@ are normal defining words. This  initializing code. @code{W:} etc.@ are normal defining words. This
 special area is cleared at the start of every colon definition.  special area is cleared at the start of every colon definition.
   
 A special feature of GNU Forths dictionary is used to implement the  A special feature of Gforth's dictionary is used to implement the
 definition of locals without type specifiers: every wordlist (aka  definition of locals without type specifiers: every wordlist (aka
 vocabulary) has its own methods for searching  vocabulary) has its own methods for searching
 etc. (@xref{dictionary}). For the present purpose we defined a wordlist  etc. (@pxref{Wordlists}). For the present purpose we defined a wordlist
 with a special search method: When it is searched for a word, it  with a special search method: When it is searched for a word, it
 actually creates that word using @code{W:}. @code{@{} changes the search  actually creates that word using @code{W:}. @code{@{} changes the search
 order to first search the wordlist containing @code{@}}, @code{W:} etc.,  order to first search the wordlist containing @code{@}}, @code{W:} etc.,
Line 1251  The locals stack pointer is only adjuste Line 1846  The locals stack pointer is only adjuste
 @code{lp+!#} orig-locals-size @minus{} new-locals-size  @code{lp+!#} orig-locals-size @minus{} new-locals-size
 @end format  @end format
 The second @code{lp+!#} adjusts the locals stack pointer from the  The second @code{lp+!#} adjusts the locals stack pointer from the
 level at the {\em orig} point to the level after the @code{THEN}. The  level at the @var{orig} point to the level after the @code{THEN}. The
 first @code{lp+!#} adjusts the locals stack pointer from the current  first @code{lp+!#} adjusts the locals stack pointer from the current
 level to the level at the orig point, so the complete effect is an  level to the level at the orig point, so the complete effect is an
 adjustment from the current level to the right level after the  adjustment from the current level to the right level after the
Line 1301  this may lead to increased space needs f Line 1896  this may lead to increased space needs f
 usually less than reclaiming this space would cost in code size.  usually less than reclaiming this space would cost in code size.
   
   
   @node ANS Forth locals,  , Gforth locals, Locals
 @subsection ANS Forth locals  @subsection ANS Forth locals
   
 The ANS Forth locals wordset does not define a syntax for locals, but  The ANS Forth locals wordset does not define a syntax for locals, but
 words that make it possible to define various syntaxes. One of the  words that make it possible to define various syntaxes. One of the
 possible syntaxes is a subset of the syntax we used in the gforth locals  possible syntaxes is a subset of the syntax we used in the Gforth locals
 wordset, i.e.:  wordset, i.e.:
   
 @example  @example
Line 1321  restrictions are: Line 1917  restrictions are:
   
 @itemize @bullet  @itemize @bullet
 @item  @item
 Locals can only be cell-sized values (no type specifers are allowed).  Locals can only be cell-sized values (no type specifiers are allowed).
 @item  @item
 Locals can be defined only outside control structures.  Locals can be defined only outside control structures.
 @item  @item
 Locals can interfere with explicit usage of the return stack. For the  Locals can interfere with explicit usage of the return stack. For the
 exact (and long) rules, see the standard. If you don't use return stack  exact (and long) rules, see the standard. If you don't use return stack
 accessing words in a definition using locals, you will we all right. The  accessing words in a definition using locals, you will be all right. The
 purpose of this rule is to make locals implementation on the return  purpose of this rule is to make locals implementation on the return
 stack easier.  stack easier.
 @item  @item
Line 1335  The whole definition must be in one line Line 1931  The whole definition must be in one line
 @end itemize  @end itemize
   
 Locals defined in this way behave like @code{VALUE}s  Locals defined in this way behave like @code{VALUE}s
 (@xref{values}). I.e., they are initialized from the stack. Using their  (@xref{Values}). I.e., they are initialized from the stack. Using their
 name produces their value. Their value can be changed using @code{TO}.  name produces their value. Their value can be changed using @code{TO}.
   
 Since this syntax is supported by gforth directly, you need not do  Since this syntax is supported by Gforth directly, you need not do
 anything to use it. If you want to port a program using this syntax to  anything to use it. If you want to port a program using this syntax to
 another ANS Forth system, use @file{anslocal.fs} to implement the syntax  another ANS Forth system, use @file{compat/anslocal.fs} to implement the
 on the other system.  syntax on the other system.
   
 Note that a syntax shown in the standard, section A.13 looks  Note that a syntax shown in the standard, section A.13 looks
 similar, but is quite different in having the order of locals  similar, but is quite different in having the order of locals
Line 1353  doc-(local) Line 1949  doc-(local)
   
 The ANS Forth locals extension wordset defines a syntax, but it is so  The ANS Forth locals extension wordset defines a syntax, but it is so
 awful that we strongly recommend not to use it. We have implemented this  awful that we strongly recommend not to use it. We have implemented this
 syntax to make porting to gforth easy, but do not document it here. The  syntax to make porting to Gforth easy, but do not document it here. The
 problem with this syntax is that the locals are defined in an order  problem with this syntax is that the locals are defined in an order
 reversed with respect to the standard stack comment notation, making  reversed with respect to the standard stack comment notation, making
 programs harder to read, and easier to misread and miswrite. The only  programs harder to read, and easier to misread and miswrite. The only
 merit of this syntax is that it is easy to implement using the ANS Forth  merit of this syntax is that it is easy to implement using the ANS Forth
 locals wordset.  locals wordset.
   
 @node Internals  @node Defining Words, Wordlists, Locals, Words
   @section Defining Words
   
   @menu
   * Values::                      
   @end menu
   
   @node Values,  , Defining Words, Defining Words
   @subsection Values
   
   @node Wordlists, Files, Defining Words, Words
   @section Wordlists
   
   @node Files, Blocks, Wordlists, Words
   @section Files
   
   @node Blocks, Other I/O, Files, Words
   @section Blocks
   
   @node Other I/O, Programming Tools, Blocks, Words
   @section Other I/O
   
   @node Programming Tools, Assembler and Code words, Other I/O, Words
   @section Programming Tools
   
   @menu
   * Debugging::                   Simple and quick.
   * Assertions::                  Making your programs self-checking.
   @end menu
   
   @node Debugging, Assertions, Programming Tools, Programming Tools
   @subsection Debugging
   
   The simple debugging aids provided in @file{debugging.fs}
   are meant to support a different style of debugging than the
   tracing/stepping debuggers used in languages with long turn-around
   times.
   
   A much better (faster) way in fast-compilig languages is to add
   printing code at well-selected places, let the program run, look at
   the output, see where things went wrong, add more printing code, etc.,
   until the bug is found.
   
   The word @code{~~} is easy to insert. It just prints debugging
   information (by default the source location and the stack contents). It
   is also easy to remove (@kbd{C-x ~} in the Emacs Forth mode to
   query-replace them with nothing). The deferred words
   @code{printdebugdata} and @code{printdebugline} control the output of
   @code{~~}. The default source location output format works well with
   Emacs' compilation mode, so you can step through the program at the
   source level using @kbd{C-x `} (the advantage over a stepping debugger
   is that you can step in any direction and you know where the crash has
   happened or where the strange data has occurred).
   
   Note that the default actions clobber the contents of the pictured
   numeric output string, so you should not use @code{~~}, e.g., between
   @code{<#} and @code{#>}.
   
   doc-~~
   doc-printdebugdata
   doc-printdebugline
   
   @node Assertions,  , Debugging, Programming Tools
   @subsection Assertions
   
   It is a good idea to make your programs self-checking, in particular, if
   you use an assumption (e.g., that a certain field of a data structure is
   never zero) that may become wrong during maintenance. Gforth supports
   assertions for this purpose. They are used like this:
   
   @example
   assert( @var{flag} )
   @end example
   
   The code between @code{assert(} and @code{)} should compute a flag, that
   should be true if everything is alright and false otherwise. It should
   not change anything else on the stack. The overall stack effect of the
   assertion is @code{( -- )}. E.g.
   
   @example
   assert( 1 1 + 2 = ) \ what we learn in school
   assert( dup 0<> ) \ assert that the top of stack is not zero
   assert( false ) \ this code should not be reached
   @end example
   
   The need for assertions is different at different times. During
   debugging, we want more checking, in production we sometimes care more
   for speed. Therefore, assertions can be turned off, i.e., the assertion
   becomes a comment. Depending on the importance of an assertion and the
   time it takes to check it, you may want to turn off some assertions and
   keep others turned on. Gforth provides several levels of assertions for
   this purpose:
   
   doc-assert0(
   doc-assert1(
   doc-assert2(
   doc-assert3(
   doc-assert(
   doc-)
   
   @code{Assert(} is the same as @code{assert1(}. The variable
   @code{assert-level} specifies the highest assertions that are turned
   on. I.e., at the default @code{assert-level} of one, @code{assert0(} and
   @code{assert1(} assertions perform checking, while @code{assert2(} and
   @code{assert3(} assertions are treated as comments.
   
   Note that the @code{assert-level} is evaluated at compile-time, not at
   run-time. I.e., you cannot turn assertions on or off at run-time, you
   have to set the @code{assert-level} appropriately before compiling a
   piece of code. You can compile several pieces of code at several
   @code{assert-level}s (e.g., a trusted library at level 1 and newly
   written code at level 3).
   
   doc-assert-level
   
   If an assertion fails, a message compatible with Emacs' compilation mode
   is produced and the execution is aborted (currently with @code{ABORT"}.
   If there is interest, we will introduce a special throw code. But if you
   intend to @code{catch} a specific condition, using @code{throw} is
   probably more appropriate than an assertion).
   
   @node Assembler and Code words, Threading Words, Programming Tools, Words
   @section Assembler and Code words
   
   Gforth provides some words for defining primitives (words written in
   machine code), and for defining the the machine-code equivalent of
   @code{DOES>}-based defining words. However, the machine-independent
   nature of Gforth poses a few problems: First of all. Gforth runs on
   several architectures, so it can provide no standard assembler. What's
   worse is that the register allocation not only depends on the processor,
   but also on the @code{gcc} version and options used.
   
   The words that Gforth offers encapsulate some system dependences (e.g., the
   header structure), so a system-independent assembler may be used in
   Gforth. If you do not have an assembler, you can compile machine code
   directly with @code{,} and @code{c,}.
   
   doc-assembler
   doc-code
   doc-end-code
   doc-;code
   doc-flush-icache
   
   If @code{flush-icache} does not work correctly, @code{code} words
   etc. will not work (reliably), either.
   
   These words are rarely used. Therefore they reside in @code{code.fs},
   which is usually not loaded (except @code{flush-icache}, which is always
   present). You can load them with @code{require code.fs}.
   
   In the assembly code you will want to refer to the inner interpreter's
   registers (e.g., the data stack pointer) and you may want to use other
   registers for temporary storage. Unfortunately, the register allocation
   is installation-dependent.
   
   The easiest solution is to use explicit register declarations
   (@pxref{Explicit Reg Vars, , Variables in Specified Registers, gcc.info,
   GNU C Manual}) for all of the inner interpreter's registers: You have to
   compile Gforth with @code{-DFORCE_REG} (configure option
   @code{--enable-force-reg}) and the appropriate declarations must be
   present in the @code{machine.h} file (see @code{mips.h} for an example;
   you can find a full list of all declarable register symbols with
   @code{grep register engine.c}). If you give explicit registers to all
   variables that are declared at the beginning of @code{engine()}, you
   should be able to use the other caller-saved registers for temporary
   storage. Alternatively, you can use the @code{gcc} option
   @code{-ffixed-REG} (@pxref{Code Gen Options, , Options for Code
   Generation Conventions, gcc.info, GNU C Manual}) to reserve a register
   (however, this restriction on register allocation may slow Gforth
   significantly).
   
   If this solution is not viable (e.g., because @code{gcc} does not allow
   you to explicitly declare all the registers you need), you have to find
   out by looking at the code where the inner interpreter's registers
   reside and which registers can be used for temporary storage. You can
   get an assembly listing of the engine's code with @code{make engine.s}.
   
   In any case, it is good practice to abstract your assembly code from the
   actual register allocation. E.g., if the data stack pointer resides in
   register @code{$17}, create an alias for this register called @code{sp},
   and use that in your assembly code.
   
   Another option for implementing normal and defining words efficiently
   is: adding the wanted functionality to the source of Gforth. For normal
   words you just have to edit @file{primitives}, defining words (for fast
   defined words) may require changes in @file{engine.c},
   @file{kernal.fs}, @file{prims2x.fs}, and possibly @file{cross.fs}.
   
   
   @node Threading Words,  , Assembler and Code words, Words
   @section Threading Words
   
   These words provide access to code addresses and other threading stuff
   in Gforth (and, possibly, other interpretive Forths). It more or less
   abstracts away the differences between direct and indirect threading
   (and, for direct threading, the machine dependences). However, at
   present this wordset is still inclomplete. It is also pretty low-level;
   some day it will hopefully be made unnecessary by an internals words set
   that abstracts implementation details away completely.
   
   doc->code-address
   doc->does-code
   doc-code-address!
   doc-does-code!
   doc-does-handler!
   doc-/does-handler
   
   The code addresses produced by various defining words are produced by
   the following words:
   
   doc-docol:
   doc-docon:
   doc-dovar:
   doc-douser:
   doc-dodefer:
   doc-dofield:
   
   Currently there is no installation-independent way for recogizing words
   defined by a @code{CREATE}...@code{DOES>} word; however, once you know
   that a word is defined by a @code{CREATE}...@code{DOES>} word, you can
   use @code{>DOES-CODE}.
   
   @node ANS conformance, Model, Words, Top
   @chapter ANS conformance
   
   To the best of our knowledge, Gforth is an
   
   ANS Forth System
   @itemize
   @item providing the Core Extensions word set
   @item providing the Block word set
   @item providing the Block Extensions word set
   @item providing the Double-Number word set
   @item providing the Double-Number Extensions word set
   @item providing the Exception word set
   @item providing the Exception Extensions word set
   @item providing the Facility word set
   @item providing @code{MS} and @code{TIME&DATE} from the Facility Extensions word set
   @item providing the File Access word set
   @item providing the File Access Extensions word set
   @item providing the Floating-Point word set
   @item providing the Floating-Point Extensions word set
   @item providing the Locals word set
   @item providing the Locals Extensions word set
   @item providing the Memory-Allocation word set
   @item providing the Memory-Allocation Extensions word set (that one's easy)
   @item providing the Programming-Tools word set
   @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
   @item providing the Search-Order word set
   @item providing the Search-Order Extensions word set
   @item providing the String word set
   @item providing the String Extensions word set (another easy one)
   @end itemize
   
   In addition, ANS Forth systems are required to document certain
   implementation choices. This chapter tries to meet these
   requirements. In many cases it gives a way to ask the system for the
   information instead of providing the information directly, in
   particular, if the information depends on the processor, the operating
   system or the installation options chosen, or if they are likely to
   change during the maintenance of Gforth.
   
   @comment The framework for the rest has been taken from pfe.
   
   @menu
   * The Core Words::              
   * The optional Block word set::  
   * The optional Double Number word set::  
   * The optional Exception word set::  
   * The optional Facility word set::  
   * The optional File-Access word set::  
   * The optional Floating-Point word set::  
   * The optional Locals word set::  
   * The optional Memory-Allocation word set::  
   * The optional Programming-Tools word set::  
   * The optional Search-Order word set::  
   @end menu
   
   
   @c =====================================================================
   @node The Core Words, The optional Block word set, ANS conformance, ANS conformance
   @comment  node-name,  next,  previous,  up
   @section The Core Words
   @c =====================================================================
   
   @menu
   * core-idef::                   Implementation Defined Options                   
   * core-ambcond::                Ambiguous Conditions                
   * core-other::                  Other System Documentation                  
   @end menu
   
   @c ---------------------------------------------------------------------
   @node core-idef, core-ambcond, The Core Words, The Core Words
   @subsection Implementation Defined Options
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item (Cell) aligned addresses:
   processor-dependent. Gforth's alignment words perform natural alignment
   (e.g., an address aligned for a datum of size 8 is divisible by
   8). Unaligned accesses usually result in a @code{-23 THROW}.
   
   @item @code{EMIT} and non-graphic characters:
   The character is output using the C library function (actually, macro)
   @code{putchar}.
   
   @item character editing of @code{ACCEPT} and @code{EXPECT}:
   This is modeled on the GNU readline library (@pxref{Readline
   Interaction, , Command Line Editing, readline, The GNU Readline
   Library}) with Emacs-like key bindings. @kbd{Tab} deviates a little by
   producing a full word completion every time you type it (instead of
   producing the common prefix of all completions).
   
   @item character set:
   The character set of your computer and display device. Gforth is
   8-bit-clean (but some other component in your system may make trouble).
   
   @item Character-aligned address requirements:
   installation-dependent. Currently a character is represented by a C
   @code{unsigned char}; in the future we might switch to @code{wchar_t}
   (Comments on that requested).
   
   @item character-set extensions and matching of names:
   Any character except the ASCII NUL charcter can be used in a
   name. Matching is case-insensitive. The matching is performed using the
   C function @code{strncasecmp}, whose function is probably influenced by
   the locale. E.g., the @code{C} locale does not know about accents and
   umlauts, so they are matched case-sensitively in that locale. For
   portability reasons it is best to write programs such that they work in
   the @code{C} locale. Then one can use libraries written by a Polish
   programmer (who might use words containing ISO Latin-2 encoded
   characters) and by a French programmer (ISO Latin-1) in the same program
   (of course, @code{WORDS} will produce funny results for some of the
   words (which ones, depends on the font you are using)). Also, the locale
   you prefer may not be available in other operating systems. Hopefully,
   Unicode will solve these problems one day.
   
   @item conditions under which control characters match a space delimiter:
   If @code{WORD} is called with the space character as a delimiter, all
   white-space characters (as identified by the C macro @code{isspace()})
   are delimiters. @code{PARSE}, on the other hand, treats space like other
   delimiters. @code{PARSE-WORD} treats space like @code{WORD}, but behaves
   like @code{PARSE} otherwise. @code{(NAME)}, which is used by the outer
   interpreter (aka text interpreter) by default, treats all white-space
   characters as delimiters.
   
   @item format of the control flow stack:
   The data stack is used as control flow stack. The size of a control flow
   stack item in cells is given by the constant @code{cs-item-size}. At the
   time of this writing, an item consists of a (pointer to a) locals list
   (third), an address in the code (second), and a tag for identifying the
   item (TOS). The following tags are used: @code{defstart},
   @code{live-orig}, @code{dead-orig}, @code{dest}, @code{do-dest},
   @code{scopestart}.
   
   @item conversion of digits > 35
   The characters @code{[\]^_'} are the digits with the decimal value
   36@minus{}41. There is no way to input many of the larger digits.
   
   @item display after input terminates in @code{ACCEPT} and @code{EXPECT}:
   The cursor is moved to the end of the entered string. If the input is
   terminated using the @kbd{Return} key, a space is typed.
   
   @item exception abort sequence of @code{ABORT"}:
   The error string is stored into the variable @code{"error} and a
   @code{-2 throw} is performed.
   
   @item input line terminator:
   For interactive input, @kbd{C-m} and @kbd{C-j} terminate lines. One of
   these characters is typically produced when you type the @kbd{Enter} or
   @kbd{Return} key.
   
   @item maximum size of a counted string:
   @code{s" /counted-string" environment? drop .}. Currently 255 characters
   on all ports, but this may change.
   
   @item maximum size of a parsed string:
   Given by the constant @code{/line}. Currently 255 characters.
   
   @item maximum size of a definition name, in characters:
   31
   
   @item maximum string length for @code{ENVIRONMENT?}, in characters:
   31
   
   @item method of selecting the user input device:
   The user input device is the standard input. There is currently no way to
   change it from within Gforth. However, the input can typically be
   redirected in the command line that starts Gforth.
   
   @item method of selecting the user output device:
   The user output device is the standard output. It cannot be redirected
   from within Gforth, but typically from the command line that starts
   Gforth. Gforth uses buffered output, so output on a terminal does not
   become visible before the next newline or buffer overflow. Output on
   non-terminals is invisible until the buffer overflows.
   
   @item methods of dictionary compilation:
   What are we expected to document here?
   
   @item number of bits in one address unit:
   @code{s" address-units-bits" environment? drop .}. 8 in all current
   ports.
   
   @item number representation and arithmetic:
   Processor-dependent. Binary two's complement on all current ports.
   
   @item ranges for integer types:
   Installation-dependent. Make environmental queries for @code{MAX-N},
   @code{MAX-U}, @code{MAX-D} and @code{MAX-UD}. The lower bounds for
   unsigned (and positive) types is 0. The lower bound for signed types on
   two's complement and one's complement machines machines can be computed
   by adding 1 to the upper bound.
   
   @item read-only data space regions:
   The whole Forth data space is writable.
   
   @item size of buffer at @code{WORD}:
   @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
   shared with the pictured numeric output string. If overwriting
   @code{PAD} is acceptable, it is as large as the remaining dictionary
   space, although only as much can be sensibly used as fits in a counted
   string.
   
   @item size of one cell in address units:
   @code{1 cells .}.
   
   @item size of one character in address units:
   @code{1 chars .}. 1 on all current ports.
   
   @item size of the keyboard terminal buffer:
   Varies. You can determine the size at a specific time using @code{lp@
   tib - .}. It is shared with the locals stack and TIBs of files that
   include the current file. You can change the amount of space for TIBs
   and locals stack at Gforth startup with the command line option
   @code{-l}.
   
   @item size of the pictured numeric output buffer:
   @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
   shared with @code{WORD}.
   
   @item size of the scratch area returned by @code{PAD}:
   The remainder of dictionary space. You can even use the unused part of
   the data stack space. The current size can be computed with @code{sp@
   pad - .}.
   
   @item system case-sensitivity characteristics:
   Dictionary searches are case insensitive. However, as explained above
   under @i{character-set extensions}, the matching for non-ASCII
   characters is determined by the locale you are using. In the default
   @code{C} locale all non-ASCII characters are matched case-sensitively.
   
   @item system prompt:
   @code{ ok} in interpret state, @code{ compiled} in compile state.
   
   @item division rounding:
   installation dependent. @code{s" floored" environment? drop .}. We leave
   the choice to @code{gcc} (what to use for @code{/}) and to you (whether to use
   @code{fm/mod}, @code{sm/rem} or simply @code{/}).
   
   @item values of @code{STATE} when true:
   -1.
   
   @item values returned after arithmetic overflow:
   On two's complement machines, arithmetic is performed modulo
   2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
   arithmetic (with appropriate mapping for signed types). Division by zero
   typically results in a @code{-55 throw} (floatingpoint unidentified
   fault), although a @code{-10 throw} (divide by zero) would be more
   appropriate.
   
   @item whether the current definition can be found after @t{DOES>}:
   No.
   
   @end table
   
   @c ---------------------------------------------------------------------
   @node core-ambcond, core-other, core-idef, The Core Words
   @subsection Ambiguous conditions
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item a name is neither a word nor a number:
   @code{-13 throw} (Undefined word)
   
   @item a definition name exceeds the maximum length allowed:
   @code{-19 throw} (Word name too long)
   
   @item addressing a region not inside the various data spaces of the forth system:
   The stacks, code space and name space are accessible. Machine code space is
   typically readable. Accessing other addresses gives results dependent on
   the operating system. On decent systems: @code{-9 throw} (Invalid memory
   address).
   
   @item argument type incompatible with parameter:
   This is usually not caught. Some words perform checks, e.g., the control
   flow words, and issue a @code{ABORT"} or @code{-12 THROW} (Argument type
   mismatch).
   
   @item attempting to obtain the execution token of a word with undefined execution semantics:
   You get an execution token representing the compilation semantics
   instead.
   
   @item dividing by zero:
   typically results in a @code{-55 throw} (floating point unidentified
   fault), although a @code{-10 throw} (divide by zero) would be more
   appropriate.
   
   @item insufficient data stack or return stack space:
   Not checked. This typically results in mysterious illegal memory
   accesses, producing @code{-9 throw} (Invalid memory address) or
   @code{-23 throw} (Address alignment exception).
   
   @item insufficient space for loop control parameters:
   like other return stack overflows.
   
   @item insufficient space in the dictionary:
   Not checked. Similar results as stack overflows. However, typically the
   error appears at a different place when one inserts or removes code.
   
   @item interpreting a word with undefined interpretation semantics:
   For some words, we defined interpretation semantics. For the others:
   @code{-14 throw} (Interpreting a compile-only word). Note that this is
   checked only by the outer (aka text) interpreter; if the word is
   @code{execute}d in some other way, it will typically perform it's
   compilation semantics even in interpret state. (We could change @code{'}
   and relatives not to give the xt of such words, but we think that would
   be too restrictive).
   
   @item modifying the contents of the input buffer or a string literal:
   These are located in writable memory and can be modified.
   
   @item overflow of the pictured numeric output string:
   Not checked.
   
   @item parsed string overflow:
   @code{PARSE} cannot overflow. @code{WORD} does not check for overflow.
   
   @item producing a result out of range:
   On two's complement machines, arithmetic is performed modulo
   2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
   arithmetic (with appropriate mapping for signed types). Division by zero
   typically results in a @code{-55 throw} (floatingpoint unidentified
   fault), although a @code{-10 throw} (divide by zero) would be more
   appropriate. @code{convert} and @code{>number} currently overflow
   silently.
   
   @item reading from an empty data or return stack:
   The data stack is checked by the outer (aka text) interpreter after
   every word executed. If it has underflowed, a @code{-4 throw} (Stack
   underflow) is performed. Apart from that, the stacks are not checked and
   underflows can result in similar behaviour as overflows (of adjacent
   stacks).
   
   @item unexepected end of the input buffer, resulting in an attempt to use a zero-length string as a name:
   @code{Create} and its descendants perform a @code{-16 throw} (Attempt to
   use zero-length string as a name). Words like @code{'} probably will not
   find what they search. Note that it is possible to create zero-length
   names with @code{nextname} (should it not?).
   
   @item @code{>IN} greater than input buffer:
   The next invocation of a parsing word returns a string wih length 0.
   
   @item @code{RECURSE} appears after @code{DOES>}:
   Compiles a recursive call to the defining word not to the defined word.
   
   @item argument input source different than current input source for @code{RESTORE-INPUT}:
   @code{-12 THROW}. Note that, once an input file is closed (e.g., because
   the end of the file was reached), its source-id may be
   reused. Therefore, restoring an input source specification referencing a
   closed file may lead to unpredictable results instead of a @code{-12
   THROW}.
   
   In the future, Gforth may be able to retore input source specifications
   from other than the current input soruce.
   
   @item data space containing definitions gets de-allocated:
   Deallocation with @code{allot} is not checked. This typically resuls in
   memory access faults or execution of illegal instructions.
   
   @item data space read/write with incorrect alignment:
   Processor-dependent. Typically results in a @code{-23 throw} (Address
   alignment exception). Under Linux on a 486 or later processor with
   alignment turned on, incorrect alignment results in a @code{-9 throw}
   (Invalid memory address). There are reportedly some processors with
   alignment restrictions that do not report them.
   
   @item data space pointer not properly aligned, @code{,}, @code{C,}:
   Like other alignment errors.
   
   @item less than u+2 stack items (@code{PICK} and @code{ROLL}):
   Not checked. May cause an illegal memory access.
   
   @item loop control parameters not available:
   Not checked. The counted loop words simply assume that the top of return
   stack items are loop control parameters and behave accordingly.
   
   @item most recent definition does not have a name (@code{IMMEDIATE}):
   @code{abort" last word was headerless"}.
   
   @item name not defined by @code{VALUE} used by @code{TO}:
   @code{-32 throw} (Invalid name argument)
   
   @item name not found (@code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]}):
   @code{-13 throw} (Undefined word)
   
   @item parameters are not of the same type (@code{DO}, @code{?DO}, @code{WITHIN}):
   Gforth behaves as if they were of the same type. I.e., you can predict
   the behaviour by interpreting all parameters as, e.g., signed.
   
   @item @code{POSTPONE} or @code{[COMPILE]} applied to @code{TO}:
   Assume @code{: X POSTPONE TO ; IMMEDIATE}. @code{X} is equivalent to
   @code{TO}.
   
   @item String longer than a counted string returned by @code{WORD}:
   Not checked. The string will be ok, but the count will, of course,
   contain only the least significant bits of the length.
   
   @item u greater than or equal to the number of bits in a cell (@code{LSHIFT}, @code{RSHIFT}):
   Processor-dependent. Typical behaviours are returning 0 and using only
   the low bits of the shift count.
   
   @item word not defined via @code{CREATE}:
   @code{>BODY} produces the PFA of the word no matter how it was defined.
   
   @code{DOES>} changes the execution semantics of the last defined word no
   matter how it was defined. E.g., @code{CONSTANT DOES>} is equivalent to
   @code{CREATE , DOES>}.
   
   @item words improperly used outside @code{<#} and @code{#>}:
   Not checked. As usual, you can expect memory faults.
   
   @end table
   
   
   @c ---------------------------------------------------------------------
   @node core-other,  , core-ambcond, The Core Words
   @subsection Other system documentation
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item nonstandard words using @code{PAD}:
   None.
   
   @item operator's terminal facilities available:
   After processing the command line, Gforth goes into interactive mode,
   and you can give commands to Gforth interactively. The actual facilities
   available depend on how you invoke Gforth.
   
   @item program data space available:
   @code{sp@ here - .} gives the space remaining for dictionary and data
   stack together.
   
   @item return stack space available:
   By default 16 KBytes. The default can be overridden with the @code{-r}
   switch (@pxref{Invocation}) when Gforth starts up.
   
   @item stack space available:
   @code{sp@ here - .} gives the space remaining for dictionary and data
   stack together.
   
   @item system dictionary space required, in address units:
   Type @code{here forthstart - .} after startup. At the time of this
   writing, this gives 70108 (bytes) on a 32-bit system.
   @end table
   
   
   @c =====================================================================
   @node The optional Block word set, The optional Double Number word set, The Core Words, ANS conformance
   @section The optional Block word set
   @c =====================================================================
   
   @menu
   * block-idef::                  Implementation Defined Options                  
   * block-ambcond::               Ambiguous Conditions               
   * block-other::                 Other System Documentation                 
   @end menu
   
   
   @c ---------------------------------------------------------------------
   @node block-idef, block-ambcond, The optional Block word set, The optional Block word set
   @subsection Implementation Defined Options
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item the format for display by @code{LIST}:
   First the screen number is displayed, then 16 lines of 64 characters,
   each line preceded by the line number.
   
   @item the length of a line affected by @code{\}:
   64 characters.
   @end table
   
   
   @c ---------------------------------------------------------------------
   @node block-ambcond, block-other, block-idef, The optional Block word set
   @subsection Ambiguous conditions
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item correct block read was not possible:
   Typically results in a @code{throw} of some OS-derived value (between
   -512 and -2048). If the blocks file was just not long enough, blanks are
   supplied for the missing portion.
   
   @item I/O exception in block transfer:
   Typically results in a @code{throw} of some OS-derived value (between
   -512 and -2048).
   
   @item invalid block number:
   @code{-35 throw} (Invalid block number)
   
   @item a program directly alters the contents of @code{BLK}:
   The input stream is switched to that other block, at the same
   position. If the storing to @code{BLK} happens when interpreting
   non-block input, the system will get quite confused when the block ends.
   
   @item no current block buffer for @code{UPDATE}:
   @code{UPDATE} has no effect.
   
   @end table
   
   
   @c ---------------------------------------------------------------------
   @node block-other,  , block-ambcond, The optional Block word set
   @subsection Other system documentation
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item any restrictions a multiprogramming system places on the use of buffer addresses:
   No restrictions (yet).
   
   @item the number of blocks available for source and data:
   depends on your disk space.
   
   @end table
   
   
   @c =====================================================================
   @node The optional Double Number word set, The optional Exception word set, The optional Block word set, ANS conformance
   @section The optional Double Number word set
   @c =====================================================================
   
   @menu
   * double-ambcond::              Ambiguous Conditions              
   @end menu
   
   
   @c ---------------------------------------------------------------------
   @node double-ambcond,  , The optional Double Number word set, The optional Double Number word set
   @subsection Ambiguous conditions
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item @var{d} outside of range of @var{n} in @code{D>S}:
   The least significant cell of @var{d} is produced.
   
   @end table
   
   
   @c =====================================================================
   @node The optional Exception word set, The optional Facility word set, The optional Double Number word set, ANS conformance
   @section The optional Exception word set
   @c =====================================================================
   
   @menu
   * exception-idef::              Implementation Defined Options              
   @end menu
   
   
   @c ---------------------------------------------------------------------
   @node exception-idef,  , The optional Exception word set, The optional Exception word set
   @subsection Implementation Defined Options
   @c ---------------------------------------------------------------------
   
   @table @i
   @item @code{THROW}-codes used in the system:
   The codes -256@minus{}-511 are used for reporting signals (see
   @file{errore.fs}). The codes -512@minus{}-2047 are used for OS errors
   (for file and memory allocation operations). The mapping from OS error
   numbers to throw code is -512@minus{}@var{errno}. One side effect of
   this mapping is that undefined OS errors produce a message with a
   strange number; e.g., @code{-1000 THROW} results in @code{Unknown error
   488} on my system.
   @end table
   
   @c =====================================================================
   @node The optional Facility word set, The optional File-Access word set, The optional Exception word set, ANS conformance
   @section The optional Facility word set
   @c =====================================================================
   
   @menu
   * facility-idef::               Implementation Defined Options               
   * facility-ambcond::            Ambiguous Conditions            
   @end menu
   
   
   @c ---------------------------------------------------------------------
   @node facility-idef, facility-ambcond, The optional Facility word set, The optional Facility word set
   @subsection Implementation Defined Options
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item encoding of keyboard events (@code{EKEY}):
   Not yet implemeted.
   
   @item duration of a system clock tick
   System dependent. With respect to @code{MS}, the time is specified in
   microseconds. How well the OS and the hardware implement this, is
   another question.
   
   @item repeatability to be expected from the execution of @code{MS}:
   System dependent. On Unix, a lot depends on load. If the system is
   lightly loaded, and the delay is short enough that Gforth does not get
   swapped out, the performance should be acceptable. Under MS-DOS and
   other single-tasking systems, it should be good.
   
   @end table
   
   
   @c ---------------------------------------------------------------------
   @node facility-ambcond,  , facility-idef, The optional Facility word set
   @subsection Ambiguous conditions
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item @code{AT-XY} can't be performed on user output device:
   Largely terminal dependant. No range checks are done on the arguments.
   No errors are reported. You may see some garbage appearing, you may see
   simply nothing happen.
   
   @end table
   
   
   @c =====================================================================
   @node The optional File-Access word set, The optional Floating-Point word set, The optional Facility word set, ANS conformance
   @section The optional File-Access word set
   @c =====================================================================
   
   @menu
   * file-idef::                   Implementation Defined Options                   
   * file-ambcond::                Ambiguous Conditions                
   @end menu
   
   
   @c ---------------------------------------------------------------------
   @node file-idef, file-ambcond, The optional File-Access word set, The optional File-Access word set
   @subsection Implementation Defined Options
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item File access methods used:
   @code{R/O}, @code{R/W} and @code{BIN} work as you would
   expect. @code{W/O} translates into the C file opening mode @code{w} (or
   @code{wb}): The file is cleared, if it exists, and created, if it does
   not (both with @code{open-file} and @code{create-file}).  Under Unix
   @code{create-file} creates a file with 666 permissions modified by your
   umask.
   
   @item file exceptions:
   The file words do not raise exceptions (except, perhaps, memory access
   faults when you pass illegal addresses or file-ids).
   
   @item file line terminator:
   System-dependent. Gforth uses C's newline character as line
   terminator. What the actual character code(s) of this are is
   system-dependent.
   
   @item file name format
   System dependent. Gforth just uses the file name format of your OS.
   
   @item information returned by @code{FILE-STATUS}:
   @code{FILE-STATUS} returns the most powerful file access mode allowed
   for the file: Either @code{R/O}, @code{W/O} or @code{R/W}. If the file
   cannot be accessed, @code{R/O BIN} is returned. @code{BIN} is applicable
   along with the retured mode.
   
   @item input file state after an exception when including source:
   All files that are left via the exception are closed.
   
   @item @var{ior} values and meaning:
   The @var{ior}s returned by the file and memory allocation words are
   intended as throw codes. They typically are in the range
   -512@minus{}-2047 of OS errors.  The mapping from OS error numbers to
   @var{ior}s is -512@minus{}@var{errno}.
   
   @item maximum depth of file input nesting:
   limited by the amount of return stack, locals/TIB stack, and the number
   of open files available. This should not give you troubles.
   
   @item maximum size of input line:
   @code{/line}. Currently 255.
   
   @item methods of mapping block ranges to files:
   Currently, the block words automatically access the file
   @file{blocks.fb} in the currend working directory. More sophisticated
   methods could be implemented if there is demand (and a volunteer).
   
   @item number of string buffers provided by @code{S"}:
   1
   
   @item size of string buffer used by @code{S"}:
   @code{/line}. currently 255.
   
   @end table
   
   @c ---------------------------------------------------------------------
   @node file-ambcond,  , file-idef, The optional File-Access word set
   @subsection Ambiguous conditions
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item attempting to position a file outside it's boundaries:
   @code{REPOSITION-FILE} is performed as usual: Afterwards,
   @code{FILE-POSITION} returns the value given to @code{REPOSITION-FILE}.
   
   @item attempting to read from file positions not yet written:
   End-of-file, i.e., zero characters are read and no error is reported.
   
   @item @var{file-id} is invalid (@code{INCLUDE-FILE}):
   An appropriate exception may be thrown, but a memory fault or other
   problem is more probable.
   
   @item I/O exception reading or closing @var{file-id} (@code{include-file}, @code{included}):
   The @var{ior} produced by the operation, that discovered the problem, is
   thrown.
   
   @item named file cannot be opened (@code{included}):
   The @var{ior} produced by @code{open-file} is thrown.
   
   @item requesting an unmapped block number:
   There are no unmapped legal block numbers. On some operating systems,
   writing a block with a large number may overflow the file system and
   have an error message as consequence.
   
   @item using @code{source-id} when @code{blk} is non-zero:
   @code{source-id} performs its function. Typically it will give the id of
   the source which loaded the block. (Better ideas?)
   
   @end table
   
   
   @c =====================================================================
   @node  The optional Floating-Point word set, The optional Locals word set, The optional File-Access word set, ANS conformance
   @section The optional Floating-Point word set
   @c =====================================================================
   
   @menu
   * floating-idef::               Implementation Defined Options
   * floating-ambcond::            Ambiguous Conditions            
   @end menu
   
   
   @c ---------------------------------------------------------------------
   @node floating-idef, floating-ambcond, The optional Floating-Point word set, The optional Floating-Point word set
   @subsection Implementation Defined Options
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item format and range of floating point numbers:
   System-dependent; the @code{double} type of C.
   
   @item results of @code{REPRESENT} when @var{float} is out of range:
   System dependent; @code{REPRESENT} is implemented using the C library
   function @code{ecvt()} and inherits its behaviour in this respect.
   
   @item rounding or truncation of floating-point numbers:
   System dependent; the rounding behaviour is inherited from the hosting C
   compiler. IEEE-FP-based (i.e., most) systems by default round to
   nearest, and break ties by rounding to even (i.e., such that the last
   bit of the mantissa is 0).
   
   @item size of floating-point stack:
   @code{s" FLOATING-STACK" environment? drop .}. Can be changed at startup
   with the command-line option @code{-f}.
   
   @item width of floating-point stack:
   @code{1 floats}.
   
   @end table
   
   
   @c ---------------------------------------------------------------------
   @node floating-ambcond,  , floating-idef, The optional Floating-Point word set
   @subsection Ambiguous conditions
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item @code{df@@} or @code{df!} used with an address that is not double-float  aligned:
   System-dependent. Typically results in an alignment fault like other
   alignment violations.
   
   @item @code{f@@} or @code{f!} used with an address that is not float  aligned:
   System-dependent. Typically results in an alignment fault like other
   alignment violations.
   
   @item Floating-point result out of range:
   System-dependent. Can result in a @code{-55 THROW} (Floating-point
   unidentified fault), or can produce a special value representing, e.g.,
   Infinity.
   
   @item @code{sf@@} or @code{sf!} used with an address that is not single-float  aligned:
   System-dependent. Typically results in an alignment fault like other
   alignment violations.
   
   @item BASE is not decimal (@code{REPRESENT}, @code{F.}, @code{FE.}, @code{FS.}):
   The floating-point number is converted into decimal nonetheless.
   
   @item Both arguments are equal to zero (@code{FATAN2}):
   System-dependent. @code{FATAN2} is implemented using the C library
   function @code{atan2()}.
   
   @item Using ftan on an argument @var{r1} where cos(@var{r1}) is zero:
   System-dependent. Anyway, typically the cos of @var{r1} will not be zero
   because of small errors and the tan will be a very large (or very small)
   but finite number.
   
   @item @var{d} cannot be presented precisely as a float in @code{D>F}:
   The result is rounded to the nearest float.
   
   @item dividing by zero:
   @code{-55 throw} (Floating-point unidentified fault)
   
   @item exponent too big for conversion (@code{DF!}, @code{DF@@}, @code{SF!}, @code{SF@@}):
   System dependent. On IEEE-FP based systems the number is converted into
   an infinity.
   
   @item @var{float}<1 (@code{facosh}):
   @code{-55 throw} (Floating-point unidentified fault)
   
   @item @var{float}=<-1 (@code{flnp1}):
   @code{-55 throw} (Floating-point unidentified fault). On IEEE-FP systems
   negative infinity is typically produced for @var{float}=-1.
   
   @item @var{float}=<0 (@code{fln}, @code{flog}):
   @code{-55 throw} (Floating-point unidentified fault). On IEEE-FP systems
   negative infinity is typically produced for @var{float}=0.
   
   @item @var{float}<0 (@code{fasinh}, @code{fsqrt}):
   @code{-55 throw} (Floating-point unidentified fault). @code{fasinh}
   produces values for these inputs on my Linux box (Bug in the C library?)
   
   @item |@var{float}|>1 (@code{facos}, @code{fasin}, @code{fatanh}):
   @code{-55 throw} (Floating-point unidentified fault).
   
   @item integer part of float cannot be represented by @var{d} in @code{f>d}:
   @code{-55 throw} (Floating-point unidentified fault).
   
   @item string larger than pictured numeric output area (@code{f.}, @code{fe.}, @code{fs.}):
   This does not happen.
   @end table
   
   
   
   @c =====================================================================
   @node  The optional Locals word set, The optional Memory-Allocation word set, The optional Floating-Point word set, ANS conformance
   @section The optional Locals word set
   @c =====================================================================
   
   @menu
   * locals-idef::                 Implementation Defined Options                 
   * locals-ambcond::              Ambiguous Conditions              
   @end menu
   
   
   @c ---------------------------------------------------------------------
   @node locals-idef, locals-ambcond, The optional Locals word set, The optional Locals word set
   @subsection Implementation Defined Options
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item maximum number of locals in a definition:
   @code{s" #locals" environment? drop .}. Currently 15. This is a lower
   bound, e.g., on a 32-bit machine there can be 41 locals of up to 8
   characters. The number of locals in a definition is bounded by the size
   of locals-buffer, which contains the names of the locals.
   
   @end table
   
   
   @c ---------------------------------------------------------------------
   @node locals-ambcond,  , locals-idef, The optional Locals word set
   @subsection Ambiguous conditions
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item executing a named local in interpretation state:
   @code{-14 throw} (Interpreting a compile-only word).
   
   @item @var{name} not defined by @code{VALUE} or @code{(LOCAL)} (@code{TO}):
   @code{-32 throw} (Invalid name argument)
   
   @end table
   
   
   @c =====================================================================
   @node  The optional Memory-Allocation word set, The optional Programming-Tools word set, The optional Locals word set, ANS conformance
   @section The optional Memory-Allocation word set
   @c =====================================================================
   
   @menu
   * memory-idef::                 Implementation Defined Options                 
   @end menu
   
   
   @c ---------------------------------------------------------------------
   @node memory-idef,  , The optional Memory-Allocation word set, The optional Memory-Allocation word set
   @subsection Implementation Defined Options
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item values and meaning of @var{ior}:
   The @var{ior}s returned by the file and memory allocation words are
   intended as throw codes. They typically are in the range
   -512@minus{}-2047 of OS errors.  The mapping from OS error numbers to
   @var{ior}s is -512@minus{}@var{errno}.
   
   @end table
   
   @c =====================================================================
   @node  The optional Programming-Tools word set, The optional Search-Order word set, The optional Memory-Allocation word set, ANS conformance
   @section The optional Programming-Tools word set
   @c =====================================================================
   
   @menu
   * programming-idef::            Implementation Defined Options            
   * programming-ambcond::         Ambiguous Conditions         
   @end menu
   
   
   @c ---------------------------------------------------------------------
   @node programming-idef, programming-ambcond, The optional Programming-Tools word set, The optional Programming-Tools word set
   @subsection Implementation Defined Options
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item ending sequence for input following @code{;code} and @code{code}:
   Not implemented (yet).
   
   @item manner of processing input following @code{;code} and @code{code}:
   Not implemented (yet).
   
   @item search order capability for @code{EDITOR} and @code{ASSEMBLER}:
   Not implemented (yet). If they were implemented, they would use the
   search order wordset.
   
   @item source and format of display by @code{SEE}:
   The source for @code{see} is the intermediate code used by the inner
   interpreter.  The current @code{see} tries to output Forth source code
   as well as possible.
   
   @end table
   
   @c ---------------------------------------------------------------------
   @node programming-ambcond,  , programming-idef, The optional Programming-Tools word set
   @subsection Ambiguous conditions
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item deleting the compilation wordlist (@code{FORGET}):
   Not implemented (yet).
   
   @item fewer than @var{u}+1 items on the control flow stack (@code{CS-PICK}, @code{CS-ROLL}):
   This typically results in an @code{abort"} with a descriptive error
   message (may change into a @code{-22 throw} (Control structure mismatch)
   in the future). You may also get a memory access error. If you are
   unlucky, this ambiguous condition is not caught.
   
   @item @var{name} can't be found (@code{forget}):
   Not implemented (yet).
   
   @item @var{name} not defined via @code{CREATE}:
   @code{;code} is not implemented (yet). If it were, it would behave like
   @code{DOES>} in this respect, i.e., change the execution semantics of
   the last defined word no matter how it was defined.
   
   @item @code{POSTPONE} applied to @code{[IF]}:
   After defining @code{: X POSTPONE [IF] ; IMMEDIATE}. @code{X} is
   equivalent to @code{[IF]}.
   
   @item reaching the end of the input source before matching @code{[ELSE]} or @code{[THEN]}:
   Continue in the same state of conditional compilation in the next outer
   input source. Currently there is no warning to the user about this.
   
   @item removing a needed definition (@code{FORGET}):
   Not implemented (yet).
   
   @end table
   
   
   @c =====================================================================
   @node  The optional Search-Order word set,  , The optional Programming-Tools word set, ANS conformance
   @section The optional Search-Order word set
   @c =====================================================================
   
   @menu
   * search-idef::                 Implementation Defined Options                 
   * search-ambcond::              Ambiguous Conditions              
   @end menu
   
   
   @c ---------------------------------------------------------------------
   @node search-idef, search-ambcond, The optional Search-Order word set, The optional Search-Order word set
   @subsection Implementation Defined Options
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item maximum number of word lists in search order:
   @code{s" wordlists" environment? drop .}. Currently 16.
   
   @item minimum search order:
   @code{root root}.
   
   @end table
   
   @c ---------------------------------------------------------------------
   @node search-ambcond,  , search-idef, The optional Search-Order word set
   @subsection Ambiguous conditions
   @c ---------------------------------------------------------------------
   
   @table @i
   
   @item changing the compilation wordlist (during compilation):
   The definition is put into the wordlist that is the compilation wordlist
   when @code{REVEAL} is executed (by @code{;}, @code{DOES>},
   @code{RECURSIVE}, etc.).
   
   @item search order empty (@code{previous}):
   @code{abort" Vocstack empty"}.
   
   @item too many word lists in search order (@code{also}):
   @code{abort" Vocstack full"}.
   
   @end table
   
   
   @node Model, Emacs and Gforth, ANS conformance, Top
   @chapter Model
   
   @node Emacs and Gforth, Internals, Model, Top
   @chapter Emacs and Gforth
   
   Gforth comes with @file{gforth.el}, an improved version of
   @file{forth.el} by Goran Rydqvist (icluded in the TILE package). The
   improvements are a better (but still not perfect) handling of
   indentation. I have also added comment paragraph filling (@kbd{M-q}),
   commenting (@kbd{C-x \}) and uncommenting (@kbd{C-u C-x \}) regions and
   removing debugging tracers (@kbd{C-x ~}, @pxref{Debugging}). I left the
   stuff I do not use alone, even though some of it only makes sense for
   TILE. To get a description of these features, enter Forth mode and type
   @kbd{C-h m}.
   
   In addition, Gforth supports Emacs quite well: The source code locations
   given in error messages, debugging output (from @code{~~}) and failed
   assertion messages are in the right format for Emacs' compilation mode
   (@pxref{Compilation, , Running Compilations under Emacs, emacs, Emacs
   Manual}) so the source location corresponding to an error or other
   message is only a few keystrokes away (@kbd{C-x `} for the next error,
   @kbd{C-c C-c} for the error under the cursor).
   
   Also, if you @code{include} @file{etags.fs}, a new @file{TAGS} file
   (@pxref{Tags, , Tags Tables, emacs, Emacs Manual}) will be produced that
   contains the definitions of all words defined afterwards. You can then
   find the source for a word using @kbd{M-.}. Note that emacs can use
   several tags files at the same time (e.g., one for the Gforth sources
   and one for your program, @pxref{Select Tags Table,,Selecting a Tags
   Table,emacs, Emacs Manual}). The TAGS file for the preloaded words is
   @file{$(datadir)/gforth/$(VERSION)/TAGS} (e.g.,
   @file{/usr/local/share/gforth/0.2/TAGS}).
   
   To get all these benefits, add the following lines to your @file{.emacs}
   file:
   
   @example
   (autoload 'forth-mode "gforth.el")
   (setq auto-mode-alist (cons '("\\.fs\\'" . forth-mode) auto-mode-alist))
   @end example
   
   @node Internals, Bugs, Emacs and Gforth, Top
 @chapter Internals  @chapter Internals
   
 Reading this section is not necessary for programming with gforth. It  Reading this section is not necessary for programming with Gforth. It
 should be helpful for finding your way in the gforth sources.  should be helpful for finding your way in the Gforth sources.
   
   The ideas in this section have also been published in the papers
   @cite{ANS fig/GNU/??? Forth} (in German) by Bernd Paysan, presented at
   the Forth-Tagung '93 and @cite{A Portable Forth Engine} by M. Anton
   Ertl, presented at EuroForth '93; the latter is available at
   @*@file{http://www.complang.tuwien.ac.at/papers/ertl93.ps.Z}.
   
   @menu
   * Portability::                 
   * Threading::                   
   * Primitives::                  
   * System Architecture::         
   * Performance::                 
   @end menu
   
   @node Portability, Threading, Internals, Internals
 @section Portability  @section Portability
   
 One of the main goals of the effort is availability across a wide range  One of the main goals of the effort is availability across a wide range
Line 1394  Double-Word Integers, gcc.info, GNU C Ma Line 3302  Double-Word Integers, gcc.info, GNU C Ma
 double numbers. GNU C is available for free on all important (and many  double numbers. GNU C is available for free on all important (and many
 unimportant) UNIX machines, VMS, 80386s running MS-DOS, the Amiga, and  unimportant) UNIX machines, VMS, 80386s running MS-DOS, the Amiga, and
 the Atari ST, so a Forth written in GNU C can run on all these  the Atari ST, so a Forth written in GNU C can run on all these
 machines@footnote{Due to Apple's look-and-feel lawsuit it is not  machines.
 available on the Mac (@pxref{Boycott, , Protect Your Freedom--Fight  
 ``Look And Feel'', gcc.info, GNU C Manual}).}.  
   
 Writing in a portable language has the reputation of producing code that  Writing in a portable language has the reputation of producing code that
 is slower than assembly. For our Forth engine we repeatedly looked at  is slower than assembly. For our Forth engine we repeatedly looked at
Line 1414  machines some compiler versions produce Line 3320  machines some compiler versions produce
 explicit register declarations are used. So by default  explicit register declarations are used. So by default
 @code{-DFORCE_REG} is not used.  @code{-DFORCE_REG} is not used.
   
   @node Threading, Primitives, Portability, Internals
 @section Threading  @section Threading
   
 GNU C's labels as values extension (available since @code{gcc-2.0},  GNU C's labels as values extension (available since @code{gcc-2.0},
Line 1445  goto *ca; Line 3352  goto *ca;
 Of course we have packaged the whole thing neatly in macros called  Of course we have packaged the whole thing neatly in macros called
 @code{NEXT} and @code{NEXT1} (the part of NEXT after fetching the cfa).  @code{NEXT} and @code{NEXT1} (the part of NEXT after fetching the cfa).
   
   @menu
   * Scheduling::                  
   * Direct or Indirect Threaded?::  
   * DOES>::                       
   @end menu
   
   @node Scheduling, Direct or Indirect Threaded?, Threading, Threading
 @subsection Scheduling  @subsection Scheduling
   
 There is a little complication: Pipelined and superscalar processors,  There is a little complication: Pipelined and superscalar processors,
Line 1463  NEXT; Line 3377  NEXT;
 the NEXT comes strictly after the other code, i.e., there is nearly no  the NEXT comes strictly after the other code, i.e., there is nearly no
 scheduling. After a little thought the problem becomes clear: The  scheduling. After a little thought the problem becomes clear: The
 compiler cannot know that sp and ip point to different addresses (and  compiler cannot know that sp and ip point to different addresses (and
 the version of @code{gcc} we used would not know it even if it could),  the version of @code{gcc} we used would not know it even if it was
 so it could not move the load of the cfa above the store to the  possible), so it could not move the load of the cfa above the store to
 TOS. Indeed the pointers could be the same, if code on or very near the  the TOS. Indeed the pointers could be the same, if code on or very near
 top of stack were executed. In the interest of speed we chose to forbid  the top of stack were executed. In the interest of speed we chose to
 this probably unused ``feature'' and helped the compiler in scheduling:  forbid this probably unused ``feature'' and helped the compiler in
 NEXT is divided into the loading part (@code{NEXT_P1}) and the goto part  scheduling: NEXT is divided into the loading part (@code{NEXT_P1}) and
 (@code{NEXT_P2}). @code{+} now looks like:  the goto part (@code{NEXT_P2}). @code{+} now looks like:
 @example  @example
 n=sp[0]+sp[1];  n=sp[0]+sp[1];
 sp++;  sp++;
Line 1477  NEXT_P1; Line 3391  NEXT_P1;
 sp[0]=n;  sp[0]=n;
 NEXT_P2;  NEXT_P2;
 @end example  @end example
 This can be scheduled optimally by the compiler (see \sect{TOS}).  This can be scheduled optimally by the compiler.
   
 This division can be turned off with the switch @code{-DCISC_NEXT}. This  This division can be turned off with the switch @code{-DCISC_NEXT}. This
 switch is on by default on machines that do not profit from scheduling  switch is on by default on machines that do not profit from scheduling
 (e.g., the 80386), in order to preserve registers.  (e.g., the 80386), in order to preserve registers.
   
   @node Direct or Indirect Threaded?, DOES>, Scheduling, Threading
 @subsection Direct or Indirect Threaded?  @subsection Direct or Indirect Threaded?
   
 Both! After packaging the nasty details in macro definitions we  Both! After packaging the nasty details in macro definitions we
Line 1499  are inherently machine-dependent, but th Line 3414  are inherently machine-dependent, but th
 lines. I.e., even porting direct threading to a new machine is a small  lines. I.e., even porting direct threading to a new machine is a small
 effort.  effort.
   
   @node DOES>,  , Direct or Indirect Threaded?, Threading
 @subsection DOES>  @subsection DOES>
 One of the most complex parts of a Forth engine is @code{dodoes}, i.e.,  One of the most complex parts of a Forth engine is @code{dodoes}, i.e.,
 the chunk of code executed by every word defined by a  the chunk of code executed by every word defined by a
Line 1516  again. We use this approach for the indi Line 3432  again. We use this approach for the indi
 a cell unused in most words is a bit wasteful, but on the machines we  a cell unused in most words is a bit wasteful, but on the machines we
 are targetting this is hardly a problem. The other reason for having a  are targetting this is hardly a problem. The other reason for having a
 code field size of two cells is to avoid having different image files  code field size of two cells is to avoid having different image files
 for direct and indirect threaded systems (@pxref{image-format}).  for direct and indirect threaded systems (@pxref{System Architecture}).
   
 The other approach is that the code field points or jumps to the cell  The other approach is that the code field points or jumps to the cell
 after @code{DOES}. In this variant there is a jump to @code{dodoes} at  after @code{DOES}. In this variant there is a jump to @code{dodoes} at
Line 1530  used up by the jump to the code address Line 3446  used up by the jump to the code address
 this approach for direct threading. We did not want to add another  this approach for direct threading. We did not want to add another
 cell to the code field.  cell to the code field.
   
   @node Primitives, System Architecture, Threading, Internals
 @section Primitives  @section Primitives
   
   @menu
   * Automatic Generation::        
   * TOS Optimization::            
   * Produced code::               
   @end menu
   
   @node Automatic Generation, TOS Optimization, Primitives, Primitives
 @subsection Automatic Generation  @subsection Automatic Generation
   
 Since the primitives are implemented in a portable language, there is no  Since the primitives are implemented in a portable language, there is no
Line 1573  looks like this: Line 3497  looks like this:
 @example  @example
 I_plus: /* + ( n1 n2 -- n ) */  /* label, stack effect */  I_plus: /* + ( n1 n2 -- n ) */  /* label, stack effect */
 /*  */                          /* documentation */  /*  */                          /* documentation */
 {  @{
 DEF_CA                          /* definition of variable ca (indirect threading) */  DEF_CA                          /* definition of variable ca (indirect threading) */
 Cell n1;                        /* definitions of variables */  Cell n1;                        /* definitions of variables */
 Cell n2;  Cell n2;
Line 1582  n1 = (Cell) sp[1];              /* input Line 3506  n1 = (Cell) sp[1];              /* input
 n2 = (Cell) TOS;  n2 = (Cell) TOS;
 sp += 1;                        /* stack adjustment */  sp += 1;                        /* stack adjustment */
 NAME("+")                       /* debugging output (with -DDEBUG) */  NAME("+")                       /* debugging output (with -DDEBUG) */
 {  @{
 n = n1+n2;                      /* C code taken from the source */  n = n1+n2;                      /* C code taken from the source */
 }  @}
 NEXT_P1;                        /* NEXT part 1 */  NEXT_P1;                        /* NEXT part 1 */
 TOS = (Cell)n;                  /* output */  TOS = (Cell)n;                  /* output */
 NEXT_P2;                        /* NEXT part 2 */  NEXT_P2;                        /* NEXT part 2 */
 }  @}
 @end example  @end example
   
 This looks long and inefficient, but the GNU C compiler optimizes quite  This looks long and inefficient, but the GNU C compiler optimizes quite
Line 1610  where the programmer has to take the act Line 3534  where the programmer has to take the act
 account, most notably @code{?dup}, but also words that do not (always)  account, most notably @code{?dup}, but also words that do not (always)
 fall through to NEXT.  fall through to NEXT.
   
   @node TOS Optimization, Produced code, Automatic Generation, Primitives
 @subsection TOS Optimization  @subsection TOS Optimization
   
 An important optimization for stack machine emulators, e.g., Forth  An important optimization for stack machine emulators, e.g., Forth
 engines, is keeping  one or more of the top stack items in  engines, is keeping  one or more of the top stack items in
 registers.  If a word has the stack effect {@var{in1}...@var{inx} @code{--}  registers.  If a word has the stack effect @var{in1}...@var{inx} @code{--}
 @var{out1}...@var{outy}}, keeping the top @var{n} items in registers  @var{out1}...@var{outy}, keeping the top @var{n} items in registers
 @itemize  @itemize
 @item  @item
 is better than keeping @var{n-1} items, if @var{x>=n} and @var{y>=n},  is better than keeping @var{n-1} items, if @var{x>=n} and @var{y>=n},
Line 1654  consider: Line 3579  consider:
 @item In the case of @code{dup ( w -- w w )} the generator must not  @item In the case of @code{dup ( w -- w w )} the generator must not
 eliminate the store to the original location of the item on the stack,  eliminate the store to the original location of the item on the stack,
 if the TOS optimization is turned on.  if the TOS optimization is turned on.
 @item Primitives with stack effects of the form {@code{--}  @item Primitives with stack effects of the form @code{--}
 @var{out1}...@var{outy}} must store the TOS to the stack at the start.  @var{out1}...@var{outy} must store the TOS to the stack at the start.
 Likewise, primitives with the stack effect {@var{in1}...@var{inx} @code{--}}  Likewise, primitives with the stack effect @var{in1}...@var{inx} @code{--}
 must load the TOS from the stack at the end. But for the null stack  must load the TOS from the stack at the end. But for the null stack
 effect @code{--} no stores or loads should be generated.  effect @code{--} no stores or loads should be generated.
 @end itemize  @end itemize
   
   @node Produced code,  , TOS Optimization, Primitives
 @subsection Produced code  @subsection Produced code
   
 To see what assembly code is produced for the primitives on your machine  To see what assembly code is produced for the primitives on your machine
 with your compiler and your flag settings, type @code{make engine.s} and  with your compiler and your flag settings, type @code{make engine.s} and
 look at the resulting file @file{engine.c}.  look at the resulting file @file{engine.s}.
   
   @node System Architecture, Performance, Primitives, Internals
 @section System Architecture  @section System Architecture
   
 Our Forth system consists not only of primitives, but also of  Our Forth system consists not only of primitives, but also of
Line 1689  possible, because we do not want to dist Line 3616  possible, because we do not want to dist
 same image file, and to make it easy for the users to use their image  same image file, and to make it easy for the users to use their image
 files on many machines. We currently need to create a different image  files on many machines. We currently need to create a different image
 file for machines with different cell sizes and different byte order  file for machines with different cell sizes and different byte order
 (little- or big-endian)@footnote{We consider adding information to the  (little- or big-endian)@footnote{We are considering adding information to the
 image file that enables the loader to change the byte order.}.  image file that enables the loader to change the byte order.}.
   
 Forth code that is going to end up in a portable image file has to  Forth code that is going to end up in a portable image file has to
 comply to some restrictions: addresses have to be stored in memory  comply to some restrictions: addresses have to be stored in memory with
 with special words (@code{A!}, @code{A,}, etc.) in order to make the  special words (@code{A!}, @code{A,}, etc.) in order to make the code
 code relocatable. Cells, floats, etc., have to be stored at the  relocatable. Cells, floats, etc., have to be stored at the natural
 natural alignment boundaries@footnote{E.g., store floats (8 bytes) at  alignment boundaries@footnote{E.g., store floats (8 bytes) at an address
 an address dividable by~8. This happens automatically in our system  dividable by~8. This happens automatically in our system when you use
 when you use the ANSI alignment words.}, in order to avoid alignment  the ANS Forth alignment words.}, in order to avoid alignment faults on
 faults on machines with stricter alignment. The image file is produced  machines with stricter alignment. The image file is produced by a
 by a metacompiler (@file{cross.fs}).  metacompiler (@file{cross.fs}).
   
 So, unlike the image file of Mitch Bradleys @code{cforth}, our image  So, unlike the image file of Mitch Bradleys @code{cforth}, our image
 file is not directly executable, but has to undergo some manipulations  file is not directly executable, but has to undergo some manipulations
Line 1709  at run-time. The loader also has to repl Line 3636  at run-time. The loader also has to repl
 primitive calls with the appropriate code-field addresses (or code  primitive calls with the appropriate code-field addresses (or code
 addresses in the case of direct threading).  addresses in the case of direct threading).
   
   @node  Performance,  , System Architecture, Internals
   @section Performance
   
   On RISCs the Gforth engine is very close to optimal; i.e., it is usually
   impossible to write a significantly faster engine.
   
   On register-starved machines like the 386 architecture processors
   improvements are possible, because @code{gcc} does not utilize the
   registers as well as a human, even with explicit register declarations;
   e.g., Bernd Beuster wrote a Forth system fragment in assembly language
   and hand-tuned it for the 486; this system is 1.19 times faster on the
   Sieve benchmark on a 486DX2/66 than Gforth compiled with
   @code{gcc-2.6.3} with @code{-DFORCE_REG}.
   
   However, this potential advantage of assembly language implementations
   is not necessarily realized in complete Forth systems: We compared
   Gforth (direct threaded, compiled with @code{gcc-2.6.3} and
   @code{-DFORCE_REG}) with Win32Forth 1.2093, LMI's NT Forth (Beta, May
   1994) and Eforth (with and without peephole (aka pinhole) optimization
   of the threaded code); all these systems were written in assembly
   language. We also compared Gforth with three systems written in C:
   PFE-0.9.11 (compiled with @code{gcc-2.6.3} with the default
   configuration for Linux: @code{-O2 -fomit-frame-pointer -DUSE_REGS}),
   ThisForth Beta (compiled with gcc-2.6.3 -O3 -fomit-frame-pointer;
   ThisForth employs peephole optimization of the threaded code) and TILE
   (compiled with @code{make opt}). We benchmarked Gforth, PFE, ThisForth
   and TILE on a 486DX2/66 under Linux. Kenneth O'Heskin kindly provided
   the results for Win32Forth and NT Forth on a 486DX2/66 with similar
   memory performance under Windows NT. Marcel Hendrix ported Eforth to
   Linux, then extended it to run the benchmarks, added the peephole
   optimizer, ran the benchmarks and reported the results.
    
   We used four small benchmarks: the ubiquitous Sieve; bubble-sorting and
   matrix multiplication come from the Stanford integer benchmarks and have
   been translated into Forth by Martin Fraeman; we used the versions
   included in the TILE Forth package, but with bigger data set sizes; and
   a recursive Fibonacci number computation for benchmarking calling
   performance. The following table shows the time taken for the benchmarks
   scaled by the time taken by Gforth (in other words, it shows the speedup
   factor that Gforth achieved over the other systems).
   
   @example
   relative      Win32-    NT       eforth       This-
     time  Gforth Forth Forth eforth  +opt   PFE Forth  TILE
   sieve     1.00  1.39  1.14   1.39  0.85  1.78  3.18  8.58
   bubble    1.00  1.31  1.41   1.48  0.88  1.67        3.88
   matmul    1.00  1.47  1.35   1.46  1.16  2.36        4.09
   fib       1.00  1.52  1.34   1.22  1.13  1.93  2.99  4.30
   @end example
   
   You may find the good performance of Gforth compared with the systems
   written in assembly language quite surprising. One important reason for
   the disappointing performance of these systems is probably that they are
   not written optimally for the 486 (e.g., they use the @code{lods}
   instruction). In addition, Win32Forth uses a comfortable, but costly
   method for relocating the Forth image: like @code{cforth}, it computes
   the actual addresses at run time, resulting in two address computations
   per NEXT (@pxref{System Architecture}).
   
   Only Eforth with the peephole optimizer performs comparable to
   Gforth. The speedups achieved with peephole optimization of threaded
   code are quite remarkable. Adding a peephole optimizer to Gforth should
   cause similar speedups.
   
   The speedup of Gforth over PFE, ThisForth and TILE can be easily
   explained with the self-imposed restriction to standard C, which makes
   efficient threading impossible (however, the measured implementation of
   PFE uses a GNU C extension: @ref{Global Reg Vars, , Defining Global
   Register Variables, gcc.info, GNU C Manual}).  Moreover, current C
   compilers have a hard time optimizing other aspects of the ThisForth
   and the TILE source.
   
   Note that the performance of Gforth on 386 architecture processors
   varies widely with the version of @code{gcc} used. E.g., @code{gcc-2.5.8}
   failed to allocate any of the virtual machine registers into real
   machine registers by itself and would not work correctly with explicit
   register declarations, giving a 1.3 times slower engine (on a 486DX2/66
   running the Sieve) than the one measured above.
   
   In @cite{Translating Forth to Efficient C} by M. Anton Ertl and Martin
   Maierhofer (presented at EuroForth '95), an indirect threaded version of
   Gforth is compared with Win32Forth, NT Forth, PFE, and ThisForth; that
   version of Gforth is 2\%@minus{}8\% slower on a 486 than the version
   used here. The paper available at
   @*@file{http://www.complang.tuwien.ac.at/papers/ertl&maierhofer95.ps.gz};
   it also contains numbers for some native code systems. You can find
   numbers for Gforth on various machines in @file{Benchres}.
   
   @node Bugs, Origin, Internals, Top
   @chapter Bugs
   
   Known bugs are described in the file BUGS in the Gforth distribution.
   
   If you find a bug, please send a bug report to
   @code{gforth-bugs@@mips.complang.tuwien.ac.at}. A bug report should
   describe the Gforth version used (it is announced at the start of an
   interactive Gforth session), the machine and operating system (on Unix
   systems you can use @code{uname -a} to produce this information), the
   installation options (send the @code{config.status} file), and a
   complete list of changes you (or your installer) have made to the Gforth
   sources (if any); it should contain a program (or a sequence of keyboard
   commands) that reproduces the bug and a description of what you think
   constitutes the buggy behaviour.
   
   For a thorough guide on reporting bugs read @ref{Bug Reporting, , How
   to Report Bugs, gcc.info, GNU C Manual}.
   
   
   @node Origin, Word Index, Bugs, Top
   @chapter Authors and Ancestors of Gforth
   
   @section Authors and Contributors
   
   The Gforth project was started in mid-1992 by Bernd Paysan and Anton
   Ertl. The third major author was Jens Wilke.  Lennart Benschop (who was
   one of Gforth's first users, in mid-1993) and Stuart Ramsden inspired us
   with their continuous feedback. Lennart Benshop contributed
   @file{glosgen.fs}, while Stuart Ramsden has been working on automatic
   support for calling C libraries. Helpful comments also came from Paul
   Kleinrubatscher, Christian Pirker, Dirk Zoller and Marcel Hendrix.
   
   Gforth also owes a lot to the authors of the tools we used (GCC, CVS,
   and autoconf, among others), and to the creators of the Internet: Gforth
   was developed across the Internet, and its authors have not met
   physically yet.
   
   @section Pedigree
   
   Gforth descends from BigForth (1993) and fig-Forth. Gforth and PFE (by
   Dirk Zoller) will cross-fertilize each other. Of course, a significant
   part of the design of Gforth was prescribed by ANS Forth.
   
   Bernd Paysan wrote BigForth, a descendent from TurboForth, an unreleased
   32 bit native code version of VolksForth for the Atari ST, written
   mostly by Dietrich Weineck.
   
   VolksForth descends from F83. It was written by Klaus Schleisiek, Bernd
   Pennemann, Georg Rehfeld and Dietrich Weineck for the C64 (called
   UltraForth there) in the mid-80s and ported to the Atari ST in 1986.
   
   Hennry Laxen and Mike Perry wrote F83 as a model implementation of the
   Forth-83 standard. !! Pedigree? When?
   
   A team led by Bill Ragsdale implemented fig-Forth on many processors in
   1979. Robert Selzer and Bill Ragsdale developed the original
   implementation of fig-Forth for the 6502 based on microForth.
   
   The principal architect of microForth was Dean Sanderson. microForth was
   FORTH, Inc.'s first off-the-shelf product. It was developped in 1976 for
   the 1802, and subsequently implemented on the 8080, the 6800 and the
   Z80.
   
   All earlier Forth systems were custom-made, usually by Charles Moore,
   who discovered (as he puts it) Forth during the late 60s. The first full
   Forth existed in 1971.
   
   A part of the information in this section comes from @cite{The Evolution
   of Forth} by Elizabeth D. Rather, Donald R. Colburn and Charles
   H. Moore, presented at the HOPL-II conference and preprinted in SIGPLAN
   Notices 28(3), 1993.  You can find more historical and genealogical
   information about Forth there.
   
   @node Word Index, Node Index, Origin, Top
   @chapter Word Index
   
   This index is as incomplete as the manual. Each word is listed with
   stack effect and wordset.
   
   @printindex fn
   
   @node Node Index,  , Word Index, Top
   @chapter Node Index
   
   This index is even less complete than the manual.
   
 @contents  @contents
 @bye  @bye
   

Removed from v.1.3  
changed lines
  Added in v.1.31


FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>