File: gforth.ds

package info (click to toggle)
gforth 0.4.9.19990617-1
links: PTS
area: main
in suites: potato
size: 5,040 kB
ctags: 868
sloc: ansic: 3,794; sh: 1,928; lisp: 1,335; makefile: 649; sed: 129
file content (11461 lines) | stat: -rw-r--r-- 419,680 bytes
\input texinfo   @c -*-texinfo-*-
@comment The source is gforth.ds, from which gforth.texi is generated

@comment TODO: nac29jan99 - a list of things to add in the next edit:
@comment 1. x-ref all ambiguous or implementation-defined features?
@comment 2. Describe the use of Auser Avariable AConstant A, etc.
@comment 3. words in miscellaneous section need a home.
@comment 4. search for TODO for other minor and major works required.
@comment 5. [rats] change all @var to @i in Forth source so that info
@comment    file looks decent.
@c          Not an improvement IMO - anton
@c          and anyway, this should be taken up
@c          with Karl Berry (the texinfo guy) - anton
@comment .. would be useful to have a word that identified all deferred words
@comment should semantics stuff in intro be moved to another section


@comment %**start of header (This is for running Texinfo on a region.)
@setfilename gforth.info
@settitle Gforth Manual
@dircategory GNU programming tools
@direntry
* Gforth: (gforth).             A fast interpreter for the Forth language.
@end direntry
@comment @setchapternewpage odd
@comment TODO this gets left in by HTML converter
@macro progstyle {}
Programming style note:
@end macro
@comment %**end of header (This is for running Texinfo on a region.)


@comment ----------------------------------------------------------
@comment macros for beautifying glossary entries
@comment if these are used, need to strip them out for HTML converter
@comment else they get repeated verbatim in HTML output.
@comment .. not working yet.

@macro GLOSS-START {}
@iftex
@ninerm
@end iftex
@end macro

@macro GLOSS-END {}
@iftex
@rm
@end iftex
@end macro

@comment ----------------------------------------------------------


@include version.texi

@ifinfo
This file documents Gforth @value{VERSION}

Copyright @copyright{} 1995-1999 Free Software Foundation, Inc.

     Permission is granted to make and distribute verbatim copies of
     this manual provided the copyright notice and this permission notice
     are preserved on all copies.
     
@ignore
     Permission is granted to process this file through TeX and print the
     results, provided the printed document carries a copying permission
     notice identical to this one except for the removal of this paragraph
     (this paragraph not being relevant to the printed manual).
     
@end ignore
     Permission is granted to copy and distribute modified versions of this
     manual under the conditions for verbatim copying, provided also that the
     sections entitled "Distribution" and "General Public License" are
     included exactly as in the original, and provided that the entire
     resulting derived work is distributed under the terms of a permission
     notice identical to this one.
     
     Permission is granted to copy and distribute translations of this manual
     into another language, under the above conditions for modified versions,
     except that the sections entitled "Distribution" and "General Public
     License" may be included in a translation approved by the author instead
     of in the original English.
@end ifinfo

@finalout
@titlepage
@sp 10
@center @titlefont{Gforth Manual}
@sp 2
@center for version @value{VERSION}
@sp 2
@center Neal Crook
@center Anton Ertl
@center Bernd Paysan
@center Jens Wilke
@sp 3
@center This manual is permanently under construction and was last updated on 04-May-1999

@comment  The following two commands start the copyright page.
@page
@vskip 0pt plus 1filll
Copyright @copyright{} 1995--1999 Free Software Foundation, Inc.

@comment !! Published by ... or You can get a copy of this manual ...

     Permission is granted to make and distribute verbatim copies of
     this manual provided the copyright notice and this permission notice
     are preserved on all copies.
     
     Permission is granted to copy and distribute modified versions of this
     manual under the conditions for verbatim copying, provided also that the
     sections entitled "Distribution" and "General Public License" are
     included exactly as in the original, and provided that the entire
     resulting derived work is distributed under the terms of a permission
     notice identical to this one.
     
     Permission is granted to copy and distribute translations of this manual
     into another language, under the above conditions for modified versions,
     except that the sections entitled "Distribution" and "General Public
     License" may be included in a translation approved by the author instead
     of in the original English.
@end titlepage


@node Top, License, (dir), (dir)
@ifinfo
Gforth is a free implementation of ANS Forth available on many
personal machines. This manual corresponds to version @value{VERSION}.
@end ifinfo

@menu
* License::                     The GPL
* Goals::                       About the Gforth Project
* Gforth Environment::          Starting (and exiting) Gforth
* Introduction::                An introduction to ANS Forth
* Words::                       Forth words available in Gforth
* Error messages::              How to interpret them
* Tools::                       Programming tools
* ANS conformance::             Implementation-defined options etc.
* Model::                       The abstract machine of Gforth
* Integrating Gforth::          Forth as scripting language for applications
* Emacs and Gforth::            The Gforth Mode
* Image Files::                 @code{.fi} files contain compiled code
* Engine::                      The inner interpreter and the primitives
* Binding to System Library::   
* Cross Compiler::              The Cross Compiler
* Bugs::                        How to report them
* Origin::                      Authors and ancestors of Gforth
* Forth-related information::   Books and places to look on the WWW
* Word Index::                  An item for each Forth word
* Concept Index::               A menu covering many topics

@detailmenu --- The Detailed Node Listing ---

Goals of Gforth

* Gforth Extensions Sinful?::

Gforth Environment

* Invoking Gforth::             Getting in
* Leaving Gforth::              Getting out
* Command-line editing::        
* Upper and lower case::
* Environment variables::       ..that affect how Gforth starts up
* Gforth Files::                What gets installed and where

An Introduction to ANS Forth

* Introducing the Text Interpreter::
* Stacks and Postfix notation::
* Your first definition::
* How does that work?::
* Forth is written in Forth::
* Review - elements of a Forth system::
* Where to go next::
* Exercises::

Forth Words

* Notation::                    
* Comments::
* Boolean Flags::
* Arithmetic::                  
* Stack Manipulation::          
* Memory::                      
* Control Structures::          
* Defining Words::              
* The Text Interpreter::
* Tokens for Words::            
* Word Lists::                   
* Environmental Queries::
* Files::                       
* Blocks::                      
* Other I/O::                   
* Programming Tools::           
* Assembler and Code Words::    
* Threading Words::             
* Locals::                      
* Structures::                  
* Object-oriented Forth::       
* Passing Commands to the OS::
* Miscellaneous Words::

Arithmetic

* Single precision::            
* Bitwise operations::          
* Double precision::            Double-cell integer arithmetic
* Numeric comparison::
* Mixed precision::             Operations with single and double-cell integers
* Floating Point::              

Stack Manipulation

* Data stack::                  
* Floating point stack::        
* Return stack::                
* Locals stack::                
* Stack pointer manipulation::  

Memory

* Memory model::                
* Dictionary allocation::       
* Heap Allocation::             
* Memory Access::               
* Address arithmetic::          
* Memory Blocks::               

Control Structures

* Selection::                   IF.. ELSE.. ENDIF
* Simple Loops::                BEGIN..
* Counted Loops::               DO
* Arbitrary control structures::
* Calls and returns::
* Exception Handling::          

Defining Words

* Simple Defining Words::       Variables, values and constants
* Colon Definitions::
* User-defined Defining Words::
* Supplying names::
* Interpretation and Compilation Semantics::

The Text Interpreter

* Input Sources::
* Number Conversion::
* Interpret/Compile states::
* Literals::
* Interpreter Directives::

Word Lists

* Why use word lists?::
* Word list examples::

Files

* Forth source files::
* General files::         
* Search Paths::                 
* Forth Search Paths::    
* General Search Paths::        

Other I/O

* Simple numeric output::       Predefined formats
* Formatted numeric output::    Formatted (pictured) output
* String Formats::              How Forth stores strings in memory
* Displaying characters and strings:: Other stuff
* Input::                       Input

Programming Tools

* Debugging::                   Simple and quick.
* Assertions::                  Making your programs self-checking.
* Singlestep Debugger::		Executing your program word by word.

Locals

* Gforth locals::               
* ANS Forth locals::            

Gforth locals

* Where are locals visible by name?::  
* How long do locals live?::    
* Programming Style::           
* Implementation::              

Structures

* Why explicit structure support?::  
* Structure Usage::             
* Structure Naming Convention::  
* Structure Implementation::    
* Structure Glossary::          

Object-oriented Forth

* Why object-oriented programming?::
* Object-Oriented Terminology::
* Objects::
* OOF::
* Mini-OOF::
* Comparison with other object models::  

The @file{objects.fs} model

* Properties of the Objects model::  
* Basic Objects Usage::         
* The Objects base class::            
* Creating objects::            
* Object-Oriented Programming Style::  
* Class Binding::               
* Method conveniences::         
* Classes and Scoping::         
* Object Interfaces::           
* Objects Implementation::      
* Objects Glossary::            

The @file{oof.fs} model

* Properties of the OOF model::
* Basic OOF Usage::
* The OOF base class::
* Class Declaration::
* Class Implementation::

The @file{mini-oof.fs} model

* Basic Mini-OOF Usage::
* Mini-OOF Example::
* Mini-OOF Implementation::

Tools

* ANS Report::                  Report the words used, sorted by wordset.

ANS conformance

* 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::  

The Core Words

* core-idef::                   Implementation Defined Options                   
* core-ambcond::                Ambiguous Conditions                
* core-other::                  Other System Documentation                  

The optional Block word set

* block-idef::                  Implementation Defined Options
* block-ambcond::               Ambiguous Conditions               
* block-other::                 Other System Documentation                 

The optional Double Number word set

* double-ambcond::              Ambiguous Conditions              

The optional Exception word set

* exception-idef::              Implementation Defined Options              

The optional Facility word set

* facility-idef::               Implementation Defined Options               
* facility-ambcond::            Ambiguous Conditions            

The optional File-Access word set

* file-idef::                   Implementation Defined Options
* file-ambcond::                Ambiguous Conditions                

The optional Floating-Point word set

* floating-idef::               Implementation Defined Options
* floating-ambcond::            Ambiguous Conditions            

The optional Locals word set

* locals-idef::                 Implementation Defined Options                 
* locals-ambcond::              Ambiguous Conditions              

The optional Memory-Allocation word set

* memory-idef::                 Implementation Defined Options                 

The optional Programming-Tools word set

* programming-idef::            Implementation Defined Options            
* programming-ambcond::         Ambiguous Conditions         

The optional Search-Order word set

* search-idef::                 Implementation Defined Options                 
* search-ambcond::              Ambiguous Conditions              

Image Files

* Image Licensing Issues::      Distribution terms for images.
* Image File Background::       Why have image files?
* Non-Relocatable Image Files::   don't always work.
* Data-Relocatable Image Files::  are better.
* Fully Relocatable Image Files:: better yet.
* Stack and Dictionary Sizes::  Setting the default sizes for an image.
* Running Image Files::         @code{gforth -i @i{file}} or @i{file}.
* Modifying the Startup Sequence::  and turnkey applications.

Fully Relocatable Image Files

* gforthmi::                    The normal way
* cross.fs::                    The hard way

Engine

* Portability::                 
* Threading::                   
* Primitives::                  
* Performance::                 

Threading

* Scheduling::                  
* Direct or Indirect Threaded?::  
* DOES>::                       

Primitives

* Automatic Generation::        
* TOS Optimization::            
* Produced code::               

Cross Compiler

* Using the Cross Compiler::
* How the Cross Compiler Works::

Other Forth-related information

* Internet resources::
* Books::
* The Forth Interest Group::
* Conferences::

@end detailmenu
@end menu

@node License, Goals, Top, Top
@unnumbered GNU GENERAL PUBLIC LICENSE
@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
@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
@unnumbered Preface
@cindex Preface
This manual documents Gforth. Some introductory material is provided for
readers who are unfamiliar with Forth or who are migrating to Gforth
from other Forth compilers. However, this manual is primarily a
reference manual.
@end iftex

@comment TODO much more blurb here.

@c ******************************************************************
@node Goals, Gforth Environment, License, Top
@comment node-name,     next,           previous, up
@chapter Goals of Gforth
@cindex goals of the Gforth project
The goal of the Gforth Project is to develop a standard model for
ANS Forth. This can be split into several subgoals:

@itemize @bullet
@item
Gforth should conform to the ANS Forth Standard.
@item
It should be a model, i.e. it should define all the
implementation-dependent things.
@item
It should become standard, i.e. widely accepted and used. This goal
is the most difficult one.
@end itemize

To achieve these goals Gforth should be
@itemize @bullet
@item
Similar to previous models (fig-Forth, F83)
@item
Powerful. It should provide for all the things that are considered
necessary today and even some that are not yet considered necessary.
@item
Efficient. It should not get the reputation of being exceptionally
slow.
@item
Free.
@item
Available on many machines/easy to port.
@end itemize

Have we achieved these goals? Gforth conforms to the ANS Forth
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
change. It certainly has not yet become a de facto standard, but it
appears to be quite popular. It has some similarities to and some
differences from previous models. It has some powerful features, but not
yet everything that we envisioned. We certainly have achieved our
execution speed goals (@pxref{Performance}).  It is free and available
on many machines.

@menu
* Gforth Extensions Sinful?::
@end menu

@node Gforth Extensions Sinful?, , Goals, Goals
@comment node-name,     next,           previous, up
@section Is it a Sin to use Gforth Extensions?
@cindex Gforth extensions

If you've been paying attention, you will have realised that there is an
ANS (American National Standard) for Forth. As you read through the rest
of this manual, you will see documentation for @i{Standard} words, and
documentation for some appealing Gforth @i{extensions}. You might ask
yourself the question: @i{``Given that there is a standard, would I be
committing a sin to use (non-Standard) Gforth extensions?''}

The answer to that question is somewhat pragmatic and somewhat
philosophical. Consider these points:

@itemize @bullet
@item
A number of the Gforth extensions can be implemented in ANS Forth using
files provided in the @file{compat/} directory. These are mentioned in
the text in passing.
@item
Forth has a rich historical precedent for programmers taking advantage
of implementation-dependent features of their tools (for example,
relying on a knowledge of the dictionary structure). Sometimes these
techniques are necessary to extract every last bit of performance from
the hardware, sometimes they are just a programming shorthand.
@item
The best way to break the rules is to know what the rules are. To learn
the rules, there is no substitute for studying the text of the Standard
itself. In particular, Appendix A of the Standard (@var{Rationale})
provides a valuable insight into the thought processes of the technical
committee.
@item
The best reason to break a rule is because you have to; because it's
more productive to do that, because it makes your code run fast enough
or because you can see no Standard way to achieve what you want to
achieve.
@end itemize

The tool @file{ans-report.fs} (@pxref{ANS Report}) makes it easy to
analyse your program and determine what non-Standard definitions it
relies upon.


@c ******************************************************************
@node Gforth Environment, Introduction, Goals, Top
@chapter Gforth Environment
@cindex Gforth environment

Note: ultimately, the gforth man page will be auto-generated from the
material in this chapter.

@menu
* Invoking Gforth::             Getting in
* Leaving Gforth::              Getting out
* Command-line editing::        
* Upper and lower case::
* Environment variables::       ..that affect how Gforth starts up
* Gforth Files::                What gets installed and where
@end menu

@xref{Image Files} for related information about the creation of images.

@comment ----------------------------------------------
@node Invoking Gforth, Leaving Gforth, ,Gforth Environment
@section Invoking Gforth
@cindex invoking Gforth
@cindex running Gforth
@cindex command-line options
@cindex options on the command line
@cindex flags on the command line

Gforth is made up of two parts; an executable ``engine'' (named
@file{gforth} or @file{gforth-fast}) and an image file. To start it, you
will usually just say @code{gforth} -- this automatically loads the
default image file @file{gforth.fi}. In many other cases the default
Gforth image will be invoked like this:
@example
gforth [file | -e forth-code] ...
@end example
@noindent
This interprets the contents of the files and the Forth code in the order they
are given.

In addition to the @file{gforth} engine, there is also an engine called
@file{gforth-fast}, which is faster, but gives less informative error
messages (@pxref{Error messages}).

In general, the command line looks like this:

@example
gforth[-fast] [engine options] [image options]
@end example

The engine options must come before the rest of the command
line. They are:

@table @code
@cindex -i, command-line option
@cindex --image-file, command-line option
@item --image-file @i{file}
@itemx -i @i{file}
Loads the Forth image @i{file} instead of the default
@file{gforth.fi} (@pxref{Image Files}).

@cindex --path, command-line option
@cindex -p, command-line option
@item --path @i{path}
@itemx -p @i{path}
Uses @i{path} for searching the image file and Forth source code files
instead of the default in the environment variable @code{GFORTHPATH} or
the path specified at installation time (e.g.,
@file{/usr/local/share/gforth/0.2.0:.}). A path is given as a list of
directories, separated by @samp{:} (on Unix) or @samp{;} (on other OSs).

@cindex --dictionary-size, command-line option
@cindex -m, command-line option
@cindex @i{size} parameters for command-line options
@cindex size of the dictionary and the stacks
@item --dictionary-size @i{size}
@itemx -m @i{size}
Allocate @i{size} space for the Forth dictionary space instead of
using the default specified in the image (typically 256K). The
@i{size} specification for this and subsequent options consists of
an integer and a unit (e.g.,
@code{4M}). The unit can be one of @code{b} (bytes), @code{e} (element
size, in this case Cells), @code{k} (kilobytes), @code{M} (Megabytes),
@code{G} (Gigabytes), and @code{T} (Terabytes). If no unit is specified,
@code{e} is used.

@cindex --data-stack-size, command-line option
@cindex -d, command-line option
@item --data-stack-size @i{size}
@itemx -d @i{size}
Allocate @i{size} space for the data stack instead of using the
default specified in the image (typically 16K).

@cindex --return-stack-size, command-line option
@cindex -r, command-line option
@item --return-stack-size @i{size}
@itemx -r @i{size}
Allocate @i{size} space for the return stack instead of using the
default specified in the image (typically 15K).

@cindex --fp-stack-size, command-line option
@cindex -f, command-line option
@item --fp-stack-size @i{size}
@itemx -f @i{size}
Allocate @i{size} space for the floating point stack instead of
using the default specified in the image (typically 15.5K). In this case
the unit specifier @code{e} refers to floating point numbers.

@cindex --locals-stack-size, command-line option
@cindex -l, command-line option
@item --locals-stack-size @i{size}
@itemx -l @i{size}
Allocate @i{size} space for the locals stack instead of using the
default specified in the image (typically 14.5K).

@cindex -h, command-line option
@cindex --help, command-line option
@item --help
@itemx -h
Print a message about the command-line options

@cindex -v, command-line option
@cindex --version, command-line option
@item --version
@itemx -v
Print version and exit

@cindex --debug, command-line option
@item --debug
Print some information useful for debugging on startup.

@cindex --offset-image, command-line option
@item --offset-image
Start the dictionary at a slightly different position than would be used
otherwise (useful for creating data-relocatable images,
@pxref{Data-Relocatable Image Files}).

@cindex --no-offset-im, command-line option
@item --no-offset-im
Start the dictionary at the normal position.

@cindex --clear-dictionary, command-line option
@item --clear-dictionary
Initialize all bytes in the dictionary to 0 before loading the image
(@pxref{Data-Relocatable Image Files}).

@cindex --die-on-signal, command-line-option
@item --die-on-signal
Normally Gforth handles most signals (e.g., the user interrupt SIGINT,
or the segmentation violation SIGSEGV) by translating it into a Forth
@code{THROW}. With this option, Gforth exits if it receives such a
signal. This option is useful when the engine and/or the image might be
severely broken (such that it causes another signal before recovering
from the first); this option avoids endless loops in such cases.
@end table

@cindex loading files at startup
@cindex executing code on startup
@cindex batch processing with Gforth
As explained above, the image-specific command-line arguments for the
default image @file{gforth.fi} consist of a sequence of filenames and
@code{-e @var{forth-code}} options that are interpreted in the sequence
in which they are given. The @code{-e @var{forth-code}} or
@code{--evaluate @var{forth-code}} option evaluates the Forth
code. This option takes only one argument; if you want to evaluate more
Forth words, you have to quote them or use @code{-e} several times. To exit
after processing the command line (instead of entering interactive mode)
append @code{-e bye} to the command line.

@cindex versions, invoking other versions of Gforth
If you have several versions of Gforth installed, @code{gforth} will
invoke the version that was installed last. @code{gforth-@i{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:
On startup the system first executes the system initialization file
(unless the option @code{--no-init-file} is given; note that the system
resulting from using this option may not be ANS Forth conformant). Then
the user initialization file @file{.gforth.fs} is executed, unless the
option @code{--no-rc} is given; this file is first searched in @file{.},
then in @file{~}, then in the normal path (see above).



@comment ----------------------------------------------
@node Leaving Gforth, Command-line editing, Invoking Gforth, Gforth Environment
@section Leaving Gforth
@cindex Gforth - leaving
@cindex leaving Gforth

You can leave Gforth by typing @code{bye} or @kbd{Ctrl-d} (at the start
of a line) or (if you invoked Gforth with the @code{--die-on-signal}
option) @kbd{Ctrl-c}. When you leave Gforth, all of your definitions and
data are discarded. @xref{Image Files} for ways of saving the state of
the system before leaving Gforth.

doc-bye

@comment ----------------------------------------------
@node Command-line editing, Upper and lower case,Leaving Gforth,Gforth Environment
@section Command-line editing
@cindex command-line editing

Gforth maintains a history file that records every line that you type to
the text interpreter. This file is preserved between sessions, and is
used to provide a command-line recall facility; if you type ctrl-P
repeatedly you can recall successively older commands from this (or
previous) session(s). The full list of command-line editing facilities is:

@comment use @table? - anton
@itemize @bullet
@item
@kbd{Ctrl-p} (``previous'') (or up-arrow) to recall successively older
commands from the history buffer.
@item
@kbd{Ctrl-n} (``next'') (or down-arrow) to recall successively newer commands
from the history buffer.
@item
@kbd{Ctrl-f} (or right-arrow) to move the cursor right, non-destructively.
@item
@kbd{Ctrl-b} (or left-arrow) to move the cursor left, non-destructively.
@item
@kbd{Ctrl-h} (backspace) to delete the character to the left of the cursor,
closing up the line.
@item
@kbd{Ctrl-k} to delete (``kill'') from the cursor to the end of the line.
@item
@kbd{Ctrl-a} to move the cursor to the start of the line.
@item
@kbd{Ctrl-e} to move the cursor to the end of the line.
@item
@key{RET} (@kbd{Ctrl-m}) or @key{LFD} (@kbd{Ctrl-j}) to submit the current
line.
@item
@key{TAB} to step through all possible full-word completions of the word
currently being typed.
@item
@kbd{Ctrl-d} at the start of the line to terminate Gforth (gracefully,
using @code{bye}).
@end itemize

When editing, displayable characters are inserted to the left of the
cursor position; the line is always in ``insert'' (as opposed to
``overstrike'') mode.

@cindex history file
@cindex @file{.gforth-history}
On Unix systems, the history file is @file{~/.gforth-history} by
default@footnote{i.e. it is stored in the user's home directory.}. You
can find out the name and location of your history file using:

@example 
history-file type \ Unix-class systems

history-file type \ Other systems
history-dir  type
@end example

If you enter long definitions by hand, you can use a text editor to
paste them out of the history file into a Forth source file for reuse at
a later time.

Gforth never trims the size of the history file, so you should do this
periodically, if necessary.

@comment this is all defined in history.fs



@comment ----------------------------------------------
@node Upper and lower case, Environment variables,Command-line editing,Gforth Environment
@section Upper and lower case
@cindex case-sensitivity
@cindex upper and lower case

Gforth is case-insensitive, so you can enter definitions and invoke
Standard words using upper, lower or mixed case (however,
@pxref{core-idef, Implementation-defined options, Implementation-defined
options}).

ANS Forth only @i{requires} implementations to recognise Standard words
when they are typed entirely in upper case. Therefore, a Standard
program must use upper case for all Standard words. You can use whatever
case you like for words that you define, but in a standard program you
have to use the words in the same case that you defined them.

Gforth supports case sensitivity through @code{table}s (case-sensitive
wordlists, @pxref{Word Lists}).

Two people have asked how to convert Gforth to case sensitivity; while
we think this is a bad idea, you can change all wordlists into tables
like this:

@example
' table-find forth-wordlist wordlist-map @ !
@end example

Note that you now have to type the predefined words in the same case
that we defined them, which are varying.  You may want to convert them
to your favourite case before doing this operation (I won't explain how,
because if you are even contemplating to do this, you'd better have
enough knowledge of Forth systems to know this already).

@comment ----------------------------------------------
@node Environment variables, Gforth Files, Upper and lower case,Gforth Environment
@section Environment variables
@cindex environment variables

Gforth uses these environment variables:

@itemize @bullet
@item
@cindex GFORTHHIST - environment variable
GFORTHHIST - (Unix systems only) specifies the directory in which to
open/create the history file, @file{.gforth-history}. Default:
@code{$HOME}.

@item
@cindex GFORTHPATH - environment variable
GFORTHPATH - specifies the path used when searching for the gforth image file and
for Forth source-code files.

@item
@cindex GFORTH - environment variable
GFORTH - used by @file{gforthmi} @xref{gforthmi}.

@item
@cindex GFORTHD - environment variable
GFORTHD - used by @file{gforthmi} @xref{gforthmi}.

@item
@cindex TMP, TEMP - environment variable
TMP, TEMP - (non-Unix systems only) used as a potential location for the
history file.
@end itemize

@comment also POSIXELY_CORRECT LINES COLUMNS HOME but no interest in
@comment mentioning these.

All the Gforth environment variables default to sensible values if they
are not set.


@comment ----------------------------------------------
@node Gforth Files, ,Environment variables,Gforth Environment
@section Gforth files
@cindex Gforth files

When you Gforth on a Unix system in the default places, it installs
files in these locations:

@itemize @bullet
@item
@file{/usr/local/bin/gforth}
@item
@file{/usr/local/bin/gforthmi}
@item
@file{/usr/local/man/man1/gforth.1} - man page.
@item
@file{/usr/local/info} - the Info version of this manual.
@item
@file{/usr/local/lib/gforth/<version>/...} - Gforth @file{.fi} files.
@item
@file{/usr/local/share/gforth/<version>/TAGS} - Emacs TAGS file.
@item
@file{/usr/local/share/gforth/<version>/...} - Gforth source files.
@item
@file{.../emacs/site-lisp/gforth.el} - Emacs gforth mode.
@end itemize

You can select different places for installation by using
@code{configure} options (listed with @code{configure --help}).

@c ******************************************************************
@node Introduction, Words, Gforth Environment, Top
@comment node-name,     next,           previous, up
@chapter An Introduction to ANS Forth
@cindex Forth - an introduction

The primary purpose of this manual is to document Gforth. However, since
Forth is not a widely-known language and there is a lack of up-to-date
teaching material, it seems worthwhile to provide some introductory
material. @xref{Forth-related information} for other sources of Forth-related
information.

The examples in this section should work on any ANS Forth; the
output shown was produced using Gforth. Each example attempts to
reproduce the exact output that Gforth produces. If you try out the
examples (and you should), what you should type is shown @kbd{like this}
and Gforth's response is shown @code{like this}. The single exception is
that, where the example shows @key{RET} it means that you should
press the ``carriage return'' key. Unfortunately, some output formats for
this manual cannot show the difference between @kbd{this} and
@code{this} which will make trying out the examples harder (but not
impossible).

Forth is an unusual language. It provides an interactive development
environment which includes both an interpreter and compiler. Forth
programming style encourages you to break a problem down into many
@cindex factoring
small fragments (@dfn{factoring}), and then to develop and test each
fragment interactively. Forth advocates assert that breaking the
edit-compile-test cycle used by conventional programming languages can
lead to great productivity improvements.

@menu
* Introducing the Text Interpreter::
* Stacks and Postfix notation::
* Your first definition::
* How does that work?::
* Forth is written in Forth::
* Review - elements of a Forth system::
* Where to go next::
* Exercises::
@end menu

@comment ----------------------------------------------
@node Introducing the Text Interpreter, Stacks and Postfix notation, Introduction, Introduction
@section Introducing the Text Interpreter
@cindex text interpreter
@cindex outer interpreter

@c IMO this is too detailed and the pace is too slow for
@c an introduction.  If you know German, take a look at
@c http://www.complang.tuwien.ac.at/anton/lvas/skriptum-stack.html 
@c to see how I do it - anton 

When you invoke the Forth image, you will see a startup banner printed
and nothing else (if you have Gforth installed on your system, try
invoking it now, by typing @kbd{gforth@key{RET}}). Forth is now running
its command line interpreter, which is called the @dfn{Text Interpreter}
(also known as the @dfn{Outer Interpreter}).  (You will learn a lot
about the text interpreter as you read through this chapter, but
@pxref{The Text Interpreter} for more detail).

Although it's not obvious, Forth is actually waiting for your
input. Type a number and press the @key{RET} key:

@example
@kbd{45@key{RET}}  ok
@end example

Rather than give you a prompt to invite you to input something, the text
interpreter prints a status message @i{after} it has processed a line
of input. The status message in this case (``@code{ ok}'' followed by
carriage-return) indicates that the text interpreter was able to process
all of your input successfully. Now type something illegal:

@example
@kbd{qwer341@key{RET}}
:1: Undefined word
qwer341
^^^^^^^
$400D2BA8 Bounce
$400DBDA8 no.extensions
@end example

The exact text, other than the ``Undefined word'' may differ slightly on
your system, but the effect is the same; when the text interpreter
detects an error, it discards any remaining text on a line, resets
certain internal state and prints an error message. @xref{Error
messages} for a detailed description of error messages.

The text interpreter waits for you to press carriage-return, and then
processes your input line. Starting at the beginning of the line, it
breaks the line into groups of characters separated by spaces. For each
group of characters in turn, it makes two attempts to do something:

@itemize @bullet
@item
It tries to treat it as a command. It does this by searching a @dfn{name
dictionary}. If the group of characters matches an entry in the name
dictionary, the name dictionary provides the text interpreter with
information that allows the text interpreter perform some actions. In
Forth jargon, we say that the group
@cindex word
@cindex definition
@cindex execution token
@cindex xt
of characters names a @dfn{word}, that the dictionary search returns an
@dfn{execution token (xt)} corresponding to the @dfn{definition} of the
word, and that the text interpreter executes the xt. Often, the terms
@dfn{word} and @dfn{definition} are used interchangeably.
@item
If the text interpreter fails to find a match in the name dictionary, it
tries to treat the group of characters as a number in the current number
base (when you start up Forth, the current number base is base 10). If
the group of characters legitimately represents a number, the text
interpreter pushes the number onto a stack (we'll learn more about that
in the next section).
@end itemize

If the text interpreter is unable to do either of these things with any
group of characters, it discards the group of characters and the rest of
the line, then prints an error message. If the text interpreter reaches
the end of the line without error, it prints the status message ``@code{ ok}''
followed by carriage-return.

This is the simplest command we can give to the text interpreter:

@example
@key{RET}  ok
@end example

The text interpreter did everything we asked it to do (nothing) without
an error, so it said that everything is ``@code{ ok}''. Try a slightly longer
command:

@example
@kbd{12 dup fred dup@key{RET}}
:1: Undefined word
12 dup fred dup
       ^^^^
$400D2BA8 Bounce
$400DBDA8 no.extensions
@end example

When you press the carriage-return key, the text interpreter starts to
work its way along the line:

@itemize @bullet
@item
When it gets to the space after the @code{2}, it takes the group of
characters @code{12} and looks them up in the name
dictionary@footnote{We can't tell if it found them or not, but assume
for now that it did not}. There is no match for this group of characters
in the name dictionary, so it tries to treat them as a number. It is
able to do this successfully, so it puts the number, 12, ``on the stack''
(whatever that means).
@item
The text interpreter resumes scanning the line and gets the next group
of characters, @code{dup}. It looks it up in the name dictionary and
(you'll have to take my word for this) finds it, and executes the word
@code{dup} (whatever that means).
@item
Once again, the text interpreter resumes scanning the line and gets the
group of characters @code{fred}. It looks them up in the name
dictionary, but can't find them. It tries to treat them as a number, but
they don't represent any legal number.
@end itemize

At this point, the text interpreter gives up and prints an error
message. The error message shows exactly how far the text interpreter
got in processing the line. In particular, it shows that the text
interpreter made no attempt to do anything with the final character
group, @code{dup}, even though we have good reason to believe that the
text interpreter would have no problem looking that word up and
executing it a second time.


@comment ----------------------------------------------
@node Stacks and Postfix notation, Your first definition, Introducing the Text Interpreter, Introduction
@section Stacks, postfix notation and parameter passing
@cindex text interpreter
@cindex outer interpreter

In procedural programming languages (like C and Pascal), the
building-block of programs is the @dfn{function} or @dfn{procedure}. These
functions or procedures are called with @dfn{explicit parameters}. For
example, in C we might write:

@example
total = total + new_volume(length,height,depth);
@end example

@noindent
where new_volume is a function-call to another piece of code, and total,
length, height and depth are all variables. length, height and depth are
parameters to the function-call.

In Forth, the equivalent of the function or procedure is the
@dfn{definition} and parameters are implicitly passed between
definitions using a shared stack that is visible to the
programmer. Although Forth does support variables, the existence of the
stack means that they are used far less often than in most other
programming languages. When the text interpreter encounters a number, it
will place (@dfn{push}) it on the stack. There are several stacks (the
actual number is implementation-dependent ...) and the particular stack
used for any operation is implied unambiguously by the operation being
performed. The stack used for all integer operations is called the @dfn{data
stack} and, since this is the stack used most commonly, references to
``the data stack'' are often abbreviated to ``the stack''.

The stacks have a last-in, first-out (LIFO) organisation. If you type:

@example
@kbd{1 2 3@key{RET}}  ok
@end example

Then this instructs the text interpreter to placed three numbers on the
(data) stack. An analogy for the behaviour of the stack is to take a
pack of playing cards and deal out the ace (1), 2 and 3 into a pile on
the table. The 3 was the last card onto the pile (``last-in'') and if
you take a card off the pile then, unless you're prepared to fiddle a
bit, the card that you take off will be the 3 (``first-out''). The
number that will be first-out of the stack is called the @dfn{top of
stack}, which
@cindex TOS definition
is often abbreviated to @dfn{TOS}.

To understand how parameters are passed in Forth, consider the
behaviour of the definition @code{+} (pronounced ``plus''). You will not
be surprised to learn that this definition performs addition. More
precisely, it adds two number together and produces a result. Where does
it get the two numbers from? It takes the top two numbers off the
stack. Where does it place the result? On the stack. You can act-out the
behaviour of @code{+} with your playing cards like this:

@itemize @bullet
@item
Pick up two cards from the stack on the table
@item
Stare at them intently and ask yourself ``what @i{is} the sum of these two
numbers''
@item
Decide that the answer is 5
@item
Shuffle the two cards back into the pack and find a 5
@item
Put a 5 on the remaining ace that's on the table.
@end itemize

If you don't have a pack of cards handy but you do have Forth running,
you can use the definition @code{.s} to show the current state of the stack,
without affecting the stack. Type:

@example
@kbd{clearstack 1 2 3@key{RET}} ok
@kbd{.s@key{RET}} <3> 1 2 3  ok
@end example

The text interpreter looks up the word @code{clearstack} and executes
it; it tidies up the stack and removes any entries that may have been
left on it by earlier examples. The text interpreter pushes each of the
three numbers in turn onto the stack. Finally, the text interpreter
looks up the word @code{.s} and executes it. The effect of executing
@code{.s} is to print the ``<3>'' (the total number of items on the stack)
followed by a list of all the items on the stack; the item on the far
right-hand side is the TOS.

You can now type:

@example
@kbd{+ .s@key{RET}} <2> 1 5  ok
@end example

@noindent
which is correct; there are now 2 items on the stack and the result of
the addition is 5.

If you're playing with cards, try doing a second addition: pick up the
two cards, work out that their sum is 6, shuffle them into the pack,
look for a 6 and place that on the table. You now have just one item on
the stack. What happens if you try to do a third addition? Pick up the
first card, pick up the second card -- ah! There is no second card. This
is called a @dfn{stack underflow} and consitutes an error. If you try to
do the same thing with Forth it will report an error (probably a Stack
Underflow or an Invalid Memory Address error).

The opposite situation to a stack underflow is a @dfn{stack overflow},
which simply accepts that there is a finite amount of storage space
reserved for the stack. To stretch the playing card analogy, if you had
enough packs of cards and you piled the cards up on the table, you would
eventually be unable to add another card; you'd hit the ceiling. Gforth
allows you to set the maximum size of the stacks. In general, the only
time that you will get a stack overflow is because a definition has a
bug in it and is generating data on the stack uncontrollably.

There's one final use for the playing card analogy. If you model your
stack using a pack of playing cards, the maximum number of items on
your stack will be 52 (I assume you didn't use the Joker). The maximum
@i{value} of any item on the stack is 13 (the King). In fact, the only
possible numbers are positive integer numbers 1 through 13; you can't
have (for example) 0 or 27 or 3.52 or -2. If you change the way you
think about some of the cards, you can accommodate different
numbers. For example, you could think of the Jack as representing 0,
the Queen as representing -1 and the King as representing -2. Your
*range* remains unchanged (you can still only represent a total of 13
numbers) but the numbers that you can represent are -2 through 10.

In that analogy, the limit was the amount of information that a single
stack entry could hold, and Forth has a similar limit. In Forth, the
size of a stack entry is called a @dfn{cell}. The actual size of a cell is
implementation dependent and affects the maximum value that a stack
entry can hold. A Standard Forth provides a cell size of at least
16-bits, and most desktop systems use a cell size of 32-bits.

Forth does not do any type checking for you, so you are free to
manipulate and combine stack items in any way you wish. A convenient way
of treating stack items is as 2's complement signed integers, and that
is what Standard words like @code{+} do. Therefore you can type:

@example
@kbd{-5 12 + .s@key{RET}} <1> 7  ok
@end example

If you use numbers and definitions like @code{+} in order to turn Forth
into a great big pocket calculator, you will realise that it's rather
different from a normal calculator. Rather than typing 2 + 3 = you had
to type 2 3 + (ignore the fact that you had to use @code{.s} to see the
result). The terminology used to describe this difference is to say that
your calculator uses @dfn{Infix Notation} (parameters and operators are
mixed) whilst Forth uses @dfn{Postfix Notation} (parameters and
operators are separate), also called @dfn{Reverse Polish Notation}.

Whilst postfix notation might look confusing to begin with, it has
several important advantages:

@itemize @bullet
@item
it is unambiguous
@item
it is more concise
@item
it fits naturally with a stack-based system
@end itemize

To examine these claims in more detail, consider these sums:

@example
6 + 5 * 4 =
4 * 5 + 6 =
@end example

If you're just learning maths or your maths is very rusty, you will
probably come up with the answer 44 for the first and 26 for the
second. If you are a bit of a whizz at maths you will remember the
@i{convention} that multiplication takes precendence over addition, and
you'd come up with the answer 26 both times. To explain the answer 26
to someone who got the answer 44, you'd probably rewrite the first sum
like this:

@example
6 + (5 * 4) =
@end example

If what you really wanted was to perform the addition before the
multiplication, you would have to use parentheses to force it.

If you did the first two sums on a pocket calculator you would probably
get the right answers, unless you were very cautious and entered them using
these keystroke sequences:

6 + 5 = * 4 =
4 * 5 = + 6 =

Postfix notation is unambiguous because the order that the operators
are applied is always explicit; that also means that parentheses are
never required. The operators are @i{active} (the act of quoting the
operator makes the operation occur) which removes the need for ``=''.

The sum 6 + 5 * 4 can be written (in postfix notation) in two
equivalent ways:

@example
6 5 4 * +      or:
5 4 * 6 +
@end example

An important thing that you should notice about this notation is that
the @i{order} of the numbers does not change; if you want to subtract
2 from 10 you type @code{10 2 -}.

The reason that Forth uses postfix notation is very simple to explain: it
makes the implementation extremely simple, and it follows naturally from
using the stack as a mechanism for passing parameters. Another way of
thinking about this is to realise that all Forth definitions are
@i{active}; they execute as they are encountered by the text
interpreter. The result of this is that the syntax of Forth is trivially
simple.



@comment ----------------------------------------------
@node Your first definition, How does that work?, Stacks and Postfix notation, Introduction
@section Your first Forth definition
@cindex first definition

Until now, the examples we've seen have been trivial; we've just been
using Forth as a bigger-than-pocket calculator. Also, each calculation
we've shown has been a ``one-off'' -- to repeat it we'd need to type it in
again@footnote{That's not quite true. If you press the up-arrow key on
your keyboard you should be able to scroll back to any earlier command,
edit it and re-enter it.} In this section we'll see how to add new
words to Forth's vocabulary.

The easiest way to create a new word is to use a @dfn{colon
definition}. We'll define a few and try them out before worrying too
much about how they work. Try typing in these examples; be careful to
copy the spaces accurately:

@example
: add-two 2 + . ;
: greet ." Hello and welcome" ;
: demo 5 add-two ;
@end example

@noindent
Now try them out:

@example
@kbd{greet@key{RET}} Hello and welcome  ok
@kbd{greet greet@key{RET}} Hello and welcomeHello and welcome  ok
@kbd{4 add-two@key{RET}} 6  ok
@kbd{demo@key{RET}} 7  ok
@kbd{9 greet demo add-two@key{RET}} Hello and welcome7 11  ok
@end example

The first new thing that we've introduced here is the pair of words
@code{:} and @code{;}. These are used to start and terminate a new
definition, respectively. The first word after the @code{:} is the name
for the new definition.

As you can see from the examples, a definition is built up of words that
have already been defined; Forth makes no distinction between
definitions that existed when you started the system up, and those that
you define yourself.

The examples also introduce the words @code{.} (dot), @code{."}
(dot-quote) and @code{dup} (dewp). Dot takes the value from the top of
the stack and displays it. It's like @code{.s} except that it only
displays the top item of the stack and it is destructive; after it has
executed, the number is no longer on the stack. There is always one
space printed after the number, and no spaces before it. Dot-quote
defines a string (a sequence of characters) that will be printed when
the word is executed. The string can contain any printable characters
except @code{"}. A @code{"} has a special function; it is not a Forth
word but it acts as a delimiter (the way that delimiters work is
described in the next section). Finally, @code{dup} duplicates the value
at the top of the stack. Try typing @code{5 dup .s} to see what it does.

We already know that the text interpreter searches through the
dictionary to locate names. If you've followed the examples earlier, you
will already have a definition called @code{add-two}. Lets try modifying
it by typing in a new definition:

@example
@kbd{: add-two dup . ." + 2 =" 2 + . ;@key{RET}} redefined add-two  ok
@end example

Forth recognised that we were defining a word that already exists, and
printed a message to warn us of that fact. Let's try out the new
definition:

@example
@kbd{9 add-two@key{RET}} 9 + 2 =11  ok
@end example

@noindent
All that we've actually done here, though, is to create a new
definition, with a particular name. The fact that there was already a
definition with the same name did not make any difference to the way
that the new definition was created (except that Forth printed a warning
message). The old definition of add-two still exists (try @code{demo}
again to see that this is true). Any new definition will use the new
definition of @code{add-two}, but old definitions continue to use the
version that already existed at the time that they were @code{compiled}.

Before you go on to the next section, try defining and redefining some
words of your own.

@comment ----------------------------------------------
@node How does that work?, Forth is written in Forth, Your first definition, Introduction
@section How does that work?
@cindex parsing words

@c That's pretty deep (IMO way too deep) for an introduction. - anton

@c Is it a good idea to talk about the interpretation semantics of a
@c number? We don't have an xt to go along with it. - anton

@c Now that I have eliminated execution semantics, I wonder if it would not
@c be better to keep them (or add run-time semantics), to make it easier to
@c explain what compilation semantics usually does. - anton

Now we're going to take another look at the definition of @code{add-two}
from the previous section. From our knowledge of the way that the text
interpreter works, we would have expected this result when we tried to
define @code{add-two}:

@example
@kbd{: add-two 2 + . " ;@key{RET}}
  ^^^^^^^
Error: Undefined word
@end example

The reason that this didn't happen is bound up in the way that @code{:}
works. The word @code{:} does two special things. The first special
thing that it does prevents the text interpreter from ever seeing the
characters @code{add-two}. The text interpreter uses a variable called
@cindex modifying >IN
@code{>IN} (pronounced ''to-in'') to keep track of where it is in the
input line. When it encounters the word @code{:} it behaves in exactly
the same way as it does for any other word; it looks it up in the name
dictionary, finds its xt and executes it. When @code{:} executes, it
looks at the input buffer, finds the word @code{add-two} and advances the
value of @code{>IN} to point past it. It then does some other stuff
associated with creating the new definition (including creating an entry
for @code{add-two} in the name dictionary). When the execution of @code{:}
completes, control returns to the text interpreter, which is oblivious
to the fact that it has been tricked into ignoring part of the input
line.

@cindex parsing words
Words like @code{:} -- words that advance the value of @code{>IN} and so
prevent the text interpreter from acting on the whole of the input line
-- are called @dfn{parsing words}.

@cindex @code{state} - effect on the text interpreter
@cindex text interpreter - effect of state
The second special thing that @code{:} does is change the value of a
variable called @code{state}, which affects the way that the text
interpreter behaves. When Gforth starts up, @code{state} has the value
0, and the text interpreter is said to be @dfn{interpreting}. During a
colon definition (started with @code{:}), @code{state} is set to -1 and
the text interpreter is said to be @dfn{compiling}. The word @code{;}
ends the definition -- one of the things that it does is to change the
value of @code{state} back to 0.

We have already seen how the text interpreter behaves when it is
interpreting; it looks for each character sequence in the dictionary,
finds its xt and executes it, or it converts it to a number and pushes
it onto the stack, or it fails to do either and generates an error.

When the text interpreter is compiling, its behaviour is slightly
different; it still looks for each character sequence in the dictionary
and finds it, or converts it to a number, or fails to do either and
generates an error.  But instead of the execution token of a word it
finds and executes the compilation token.  For most words executing the
compilation token results in laying down (@dfn{compiling}) the execution
token, i.e., some magic to make that xt or number get executed or pushed
at a later time; at the time that @code{add-two} is
@dfn{executed}. Therefore, when you execute @code{add-two} its
@dfn{run-time effect} is exactly the same as if you had typed @code{2 +
.} outside of a definition, and pressed carriage-return.

In Forth, every word or number can be described in terms of two
properties:

@itemize @bullet
@item
Its @dfn{interpretation semantics}, represented by the execution token.
@item
Its @dfn{compilation semantics}, represented by the compilation token.
@end itemize

The value of @code{state} determines whether the text interpreter will
use the compilation or interpretation semantics of a word or number that
it encounters.

@itemize @bullet
@item
@cindex interpretation semantics
When the text interpreter encounters a word or number in @dfn{interpret}
state, it performs the @dfn{interpretation semantics} of the word or
number.
@item
@cindex compilation semantics
When the text interpreter encounters a word or number in @dfn{compile}
state, it performs the @dfn{compilation semantics} of the word or
number.
@end itemize

@noindent
Numbers are always treated in a fixed way:

@itemize @bullet
@item
When the number is @dfn{interpreted}, its behaviour is to push the number onto the stack.
@item
When the number is @dfn{compiled}, a piece of code is appended to the
current definition that pushes the number when it runs. (In other words,
the compilation semantics of a number are to postpone its interpretation
semantics until the run-time of the definition that it is being compiled
into.)
@end itemize

The behaviour of a word is not so regular, but most have @i{default
compilation semantics} which means that they behave like this:

@itemize @bullet
@item
The @dfn{interpretation semantics} of the word are to do something useful.
@item
The @dfn{compilation semantics} of the word are to append its
@dfn{interpretation semantics} to the current definition (so that its
run-time behaviour is to do something useful).
@end itemize

@cindex immediate words
The actual behaviour of any particular word depends upon the way in
which it was defined. When the text interpreter finds the word in the
name dictionary, it not only retrieves the xt for the word, it also
retrieves some flags: the @dfn{compile-only} flag and the @dfn{immediate
flag}. The compile-only flag indicates that the word has no
interpretation semantics (the run-time behaviour for the default
compilation semantics is not affected by this flag, however); any
attempt to interpret a word that has the compile-only flag set will
generate an error (for example, @code{IF} has no interpretation
semantics). The immediate flag changes the compilation semantics of the
word; if it is set, the compilation semantics are equal to the
interpretation semantics (again ignoring the compile-only flag).  it. In
other words, these so-called @dfn{immediate} words behave like this:

@itemize @bullet
@item
The @dfn{interpretation semantics} of the word are to do something useful.
@item
The @dfn{compilation semantics} of the word are to do something useful
(and actually the same thing); i.e., it is executed during compilation.
@end itemize

This example shows the difference between an immediate and a
non-immediate word:

@example
: show-state state @@ . ;
: show-state-now show-state ; immediate
: word1 show-state ;
: word2 show-state-now ;
@end example

The word @code{immediate} after the definition of @code{show-state-now}
makes that word an immediate word. These definitions introduce a new
word: @code{@@} (pronounced ``fetch''). This word fetches the value of a
variable, and leaves it on the stack. Therefore, the behaviour of
@code{show-state} is to print a number that represents the current value
of @code{state}.

When you execute @code{word1}, it prints the number 0, indicating that
the system is interpreting. When the text interpreter compiled the
definition of @code{word1}, it encountered @code{show-state} whose
compilation semantics are to append its interpretation semantics to the
current definition. When you execute @code{word1}, it performs the
interpretation semantics of @code{show-state}.  At the time that @code{word1}
(and therefore @code{show-state}) are executed, the system is
interpreting.

When you pressed @key{RET} after entering the definition of @code{word2},
you should have seen the number -1 printed, followed by ``@code{
ok}''. When the text interpreter compiled the definition of
@code{word2}, it encountered @code{show-state-now}, an immediate word,
whose compilation semantics are therefore to perform its interpretation
semantics. It is executed straight away (even before the text
interpreter has moved on to process another group of characters; the
@code{;} in this example). The effect of executing it are to display the
value of @code{state} @i{at the time that the definition of}
@code{word2} @i{is being defined}. Printing -1 demonstrates that the
system is compiling at this time. If you execute @code{word2} it does
nothing at all.

@cindex @code{."}, how it works
Before leaving the subject of immediate words, consider the behaviour of
@code{."} in the definition of @code{greet}, in the previous
section. This word is both a parsing word and an immediate word. Notice
that there is a space between @code{."} and the start of the text
@code{Hello and welcome}, but that there is no space between the last
letter of @code{welcome} and the @code{"} character. The reason for this
is that @code{."} is a Forth word; it must have a space after it so that
the text interpreter can identify it. The @code{"} is not a Forth word;
it is a @dfn{delimiter}. The examples earlier show that, when the string
is displayed, there is neither a space before the @code{H} nor after the
@code{e}. Since @code{."} is an immediate word, it executes at the time
that @code{greet} is defined. When it executes, its behaviour is to
search forward in the input line looking for the delimiter. When it
finds the delimiter, it updates @code{>IN} to point past the
delimiter. It also compiles some magic code into the definition of
@code{greet}; the xt of a run-time routine that prints a text string. It
compiles the string @code{Hello and welcome} into memory so that it is
available to be printed later. When the text interpreter gains control,
the next word it finds in the input stream is @code{;} and so it
terminates the definition of @code{greet}.


@comment ----------------------------------------------
@node Forth is written in Forth, Review - elements of a Forth system, How does that work?, Introduction
@section Forth is written in Forth
@cindex structure of Forth programs

When you start up a Forth compiler, a large number of definitions
already exist. In Forth, you develop a new application using bottom-up
programming techniques to create new definitions that are defined in
terms of existing definitions. As you create each definition you can
test and debug it interactively.

If you have tried out the examples in this section, you will probably
have typed them in by hand; when you leave Gforth, your definitions will
be lost. You can avoid this by using a text editor to enter Forth source
code into a file, and then loading code from the file using
@code{include} (@xref{Forth source files}). A Forth source file is
processed by the text interpreter, just as though you had typed it in by
hand@footnote{Actually, there are some subtle differences -- see
@ref{The Text Interpreter}.}.

Gforth also supports the traditional Forth alternative to using text
files for program entry (@xref{Blocks}).

In common with many, if not most, Forth compilers, most of Gforth is
actually written in Forth. All of the @file{.fs} files in the
installation directory@footnote{For example,
@file{/usr/local/share/gforth...}} are Forth source files, which you can
study to see examples of Forth programming.

Gforth maintains a history file that records every line that you type to
the text interpreter. This file is preserved between sessions, and is
used to provide a command-line recall facility. If you enter long
definitions by hand, you can use a text editor to paste them out of the
history file into a Forth source file for reuse at a later time
(@pxref{Command-line editing} for more information).


@comment ----------------------------------------------
@node Review - elements of a Forth system, Where to go next, Forth is written in Forth, Introduction
@section Review - elements of a Forth system
@cindex elements of a Forth system

To summarise this chapter:

@itemize @bullet
@item
Forth programs use @dfn{factoring} to break a problem down into small
fragments called @dfn{words} or @dfn{definitions}.
@item
Forth program development is an interactive process.
@item
The main command loop that accepts input, and controls both
interpretation and compilation, is called the @dfn{text interpreter}
(also known as the @dfn{outer interpreter}).
@item
Forth has a very simple syntax, consisting of words and numbers
separated by spaces or carriage-return characters. Any additional syntax
is imposed by @dfn{parsing words}.
@item
Forth uses a stack to pass parameters between words. As a result, it
uses postfix notation.
@item
To use a word that has previously been defined, the text interpreter
searches for the word in the @dfn{name dictionary}.
@item
Words have @dfn{interpretation semantics} and @dfn{compilation semantics}.
@item
The text interpreter uses the value of @code{state} to select between
the use of the @dfn{interpretation semantics} and the  @dfn{compilation
semantics} of a word that it encounters.
@item
The relationship between the @dfn{interpretation semantics} and
@dfn{compilation semantics} for a word
depend upon the way in which the word was defined (for example, whether
it is an @dfn{immediate} word).
@item
Forth definitions can be implemented in Forth (called @dfn{high-level
definitions}) or in some other way (usually a lower-level language and
as a result often called @dfn{low-level definitions}, @dfn{code
definitions} or @dfn{primitives}).
@item
Many Forth systems are implemented mainly in Forth.
@end itemize


@comment ----------------------------------------------
@node Where to go next,Exercises,Review - elements of a Forth system, Introduction
@section Where To Go Next
@cindex where to go next

Amazing as it may seem, if you have read (and understood) this far, you
know almost all the fundamentals about the inner workings of a Forth
system. You certainly know enough to be able to read and understand the
rest of this manual and the ANS Forth document, to learn more about the
facilities that Forth in general and Gforth in particular provide. Even
scarier, you know almost enough to implement your own Forth system.
However, that's not a good idea just yet... better to try writing some
programs in Gforth.

Forth has such a rich vocabulary that it can be hard to know where to
start in learning it. This section suggests a few sets of words that are
enough to write small but useful programs. Use the word index in this
document to learn more about each word, then try it out and try to write
small definitions using it. Start by experimenting with these words:

@itemize @bullet
@item
Arithmetic: @code{+ - * / /MOD */ ABS INVERT}
@item
Comparison: @code{MIN MAX =}
@item
Logic: @code{AND OR XOR NOT}
@item
Stack manipulation: @code{DUP DROP SWAP OVER}
@item
Loops and decisions: @code{IF ELSE ENDIF ?DO I LOOP}
@item
Input/Output: @code{. ." EMIT CR KEY}
@item
Defining words: @code{: ; CREATE}
@item
Memory allocation words: @code{ALLOT ,}
@item
Tools: @code{SEE WORDS .S MARKER}
@end itemize

When you have mastered those, go on to:

@itemize @bullet
@item
More defining words: @code{VARIABLE CONSTANT VALUE TO CREATE DOES>}
@item
Memory access: @code{@@ !}
@end itemize

When you have mastered these, there's nothing for it but to read through
the whole of this manual and find out what you've missed.

@comment ----------------------------------------------
@node Exercises, ,Where to go next, Introduction
@section Exercises
@cindex exercises

TODO: provide a set of programming excercises linked into the stuff done
already and into other sections of the manual. Provide solutions to all
the exercises in a .fs file in the distribution.

@c Get some inspiration from Starting Forth and Kelly&Spies.

@c excercises:
@c 1. take inches and convert to feet and inches.
@c 2. take temperature and convert from fahrenheight to celcius;
@c    may need to care about symmetric vs floored??
@c 3. take input line and do character substitution
@c    to encipher or decipher
@c 4. as above but work on a file for in and out
@c 5. take input line and convert to pig-latin 
@c
@c thing of sets of things to exercise then come up with
@c problems that need those things.


@c ******************************************************************
@node Words, Error messages, Introduction, Top
@chapter Forth Words
@cindex words

@menu
* Notation::                    
* Comments::
* Boolean Flags::
* Arithmetic::                  
* Stack Manipulation::          
* Memory::                      
* Control Structures::          
* Defining Words::              
* The Text Interpreter::
* Tokens for Words::            
* Word Lists::                   
* Environmental Queries::
* Files::                       
* Blocks::                      
* Other I/O::                   
* Programming Tools::           
* Assembler and Code Words::    
* Threading Words::             
* Locals::                      
* Structures::                  
* Object-oriented Forth::       
* Passing Commands to the OS::
* Miscellaneous Words::
@end menu

@node Notation, Comments, Words, Words
@section Notation
@cindex notation of glossary entries
@cindex format of glossary entries
@cindex glossary notation format
@cindex word glossary entry format

The Forth words are described in this section in the glossary notation
that has become a de-facto standard for Forth texts, i.e.,

@format
@i{word}     @i{Stack effect}   @i{wordset}   @i{pronunciation}
@end format
@i{Description}

@table @var
@item word
The name of the word.

@item Stack effect
@cindex stack effect
The stack effect is written in the notation @code{@i{before} --
@i{after}}, where @i{before} and @i{after} describe the top 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,
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
notation. Also, return stack effects are not shown in @i{stack
effect}, but in @i{Description}. The name of a stack item describes
the type and/or the function of the item. See below for a discussion of
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
@i{ -- }. 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.

@cindex pronounciation of words
@item pronunciation
How the word is pronounced.

@cindex wordset
@item wordset
The ANS Forth standard is divided into several word sets. A standard
system need not support all of them. Therefore, in theory, the fewer
word sets your program uses the more portable it will be. However, we
suspect that most ANS Forth systems on personal machines will feature
all word sets. Words that are not defined in ANS Forth have
@code{gforth} or @code{gforth-internal} as word set. @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 word set field.

@item Description
A description of the behaviour of the word.
@end table

@cindex types of stack items
@cindex stack item types
The type of a stack item is specified by the character(s) the name
starts with:

@table @code
@item f
@cindex @code{f}, stack item type
Boolean flags, i.e. @code{false} or @code{true}.
@item c
@cindex @code{c}, stack item type
Char
@item w
@cindex @code{w}, stack item type
Cell, can contain an integer or an address
@item n
@cindex @code{n}, stack item type
signed integer
@item u
@cindex @code{u}, stack item type
unsigned integer
@item d
@cindex @code{d}, stack item type
double sized signed integer
@item ud
@cindex @code{ud}, stack item type
double sized unsigned integer
@item r
@cindex @code{r}, stack item type
Float (on the FP stack)
@item a-
@cindex @code{a_}, stack item type
Cell-aligned address
@item c-
@cindex @code{c_}, stack item type
Char-aligned address (note that a Char may have two bytes in Windows NT)
@item f-
@cindex @code{f_}, stack item type
Float-aligned address
@item df-
@cindex @code{df_}, stack item type
Address aligned for IEEE double precision float
@item sf-
@cindex @code{sf_}, stack item type
Address aligned for IEEE single precision float
@item xt
@cindex @code{xt}, stack item type
Execution token, same size as Cell
@item wid
@cindex @code{wid}, stack item type
Word list ID, same size as Cell
@item f83name
@cindex @code{f83name}, stack item type
Pointer to a name structure
@item "
@cindex @code{"}, stack item type
string in the input stream (not on the stack). The terminating character
is a blank by default. If it is not a blank, it is shown in @code{<>}
quotes.
@end table

@node Comments, Boolean Flags, Notation, Words
@section Comments
@cindex comments

Forth supports two styles of comment; the traditional @i{in-line} comment,
@code{(} and its modern cousin, the @i{comment to end of line}; @code{\}.

doc-(
doc-\
doc-\G

@node Boolean Flags, Arithmetic, Comments, Words
@section Boolean Flags
@cindex Boolean flags

A Boolean flag is cell-sized. A cell with all bits clear represents the
flag @code{false} and a flag with all bits set represents the flag
@code{true}. Words that check a flag (for example, @code{IF}) will treat
a cell that has @i{any} bit set as @code{true}.

doc-true
doc-false
doc-on
doc-off

@node Arithmetic, Stack Manipulation, Boolean Flags, Words
@section Arithmetic
@cindex arithmetic words

@cindex division with potentially negative operands
Forth arithmetic is not checked, i.e., you will not hear about integer
overflow on addition or multiplication, you may hear about division by
zero if you are lucky. The operator is written after the operands, but
the operands are still in the original order. I.e., the infix @code{2-1}
corresponds to @code{2 1 -}. Forth offers a variety of division
operators. If you perform division with potentially negative operands,
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
former, @pxref{Mixed precision}).
@comment TODO discuss the different division forms and the std approach

@menu
* Single precision::            
* Bitwise operations::          
* Double precision::            Double-cell integer arithmetic
* Numeric comparison::
* Mixed precision::             Operations with single and double-cell integers
* Floating Point::              
@end menu

@node Single precision, Bitwise operations, Arithmetic, Arithmetic
@subsection Single precision
@cindex single precision arithmetic words

By default, numbers in Forth are single-precision integers that are 1
cell in size. They can be signed or unsigned, depending upon how you
treat them. @xref{Number Conversion} for the rules used by the text
interpreter for recognising single-precision integers.

doc-+
doc-1+
doc--
doc-1-
doc-*
doc-/
doc-mod
doc-/mod
doc-negate
doc-abs
doc-min
doc-max
doc-d>s
doc-floored

@node Bitwise operations, Double precision, Single precision, Arithmetic
@subsection Bitwise operations
@cindex bitwise operation words

doc-and
doc-or
doc-xor
doc-invert
doc-lshift
doc-rshift
doc-2*
doc-d2*
doc-2/
doc-d2/

@node Double precision, Numeric comparison, Bitwise operations, Arithmetic
@subsection Double precision
@cindex double precision arithmetic words

@xref{Number Conversion} for the rules used by the text interpreter for
recognising double-precision integers.

A double precision number is represented by a cell pair, with the most
significant cell at the TOS. It is trivial to convert an unsigned
single to an (unsigned) double; simply push a @code{0} onto the
TOS. Since numbers are represented by Gforth using 2's complement
arithmetic, converting a signed single to a (signed) double requires
sign-extension across the most significant cell. This can be achieved
using @code{s>d}. The moral of the story is that you cannot convert a
number without knowing whether it represents an unsigned or a
signed number.

doc-s>d
doc-d+
doc-d-
doc-dnegate
doc-dabs
doc-dmin
doc-dmax

@node Numeric comparison, Mixed precision, Double precision, Arithmetic
@subsection Numeric comparison
@cindex numeric comparison words

doc-<
doc-<=
doc-<>
doc-=
doc->
doc->=

doc-0<
doc-0<=
doc-0<>
doc-0=
doc-0>
doc-0>=

doc-u<
doc-u<=
@c TODO why u<> and u= ... they are the same as <> and =
@c commented them out because they are unnecessary
@c doc-u<>
@c doc-u=
doc-u>
doc-u>=

doc-within

doc-d<
doc-d<=
doc-d<>
doc-d=
doc-d>
doc-d>=

doc-d0<
doc-d0<=
doc-d0<>
doc-d0=
doc-d0>
doc-d0>=

doc-du<
doc-du<=
@c doc-du<>
@c doc-du=
doc-du>
doc-du>=

@node Mixed precision, Floating Point, Numeric comparison, Arithmetic
@subsection Mixed precision
@cindex mixed precision arithmetic words

doc-m+
doc-*/
doc-*/mod
doc-m*
doc-um*
doc-m*/
doc-um/mod
doc-fm/mod
doc-sm/rem

@node Floating Point,  , Mixed precision, Arithmetic
@subsection Floating Point
@cindex floating point arithmetic words

@xref{Number Conversion} for the rules used by the text interpreter for
recognising floating-point numbers.

Gforth has a separate floating point
stack, but the documentation uses the unified notation.

@cindex floating-point arithmetic, pitfalls
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}
(@url{http://www.validgh.com/goldberg/paper.ps}).

doc-d>f
doc-f>d
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-f2*
doc-f2/
doc-1/f
doc-precision
doc-set-precision

@cindex angles in trigonometric operations
@cindex trigonometric operations
Angles in floating point operations are given in radians (a full circle
has 2 pi radians).

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
doc-pi

@cindex equality of floats
@cindex floating-point comparisons
One particular problem with floating-point arithmetic is that comparison
for equality often fails when you would expect it to succeed.  For this
reason approximate equality is often preferred (but you still have to
know what you are doing).  The comparison words are:

doc-f~rel
doc-f~abs
doc-f=
doc-f~
doc-f<>

doc-f<
doc-f<=
doc-f>
doc-f>=

doc-f0<
doc-f0<=
doc-f0<>
doc-f0=
doc-f0>
doc-f0>=


@node Stack Manipulation, Memory, Arithmetic, Words
@section Stack Manipulation
@cindex stack manipulation words

@cindex floating-point stack in the standard
Gforth maintains a number of separate stacks:

@cindex data stack
@cindex parameter stack
@itemize @bullet
@item
A data stack (also known as the @dfn{parameter stack}) -- for
characters, cells, addresses, and double cells.

@cindex floating-point stack
@item
A floating point stack -- for floating point numbers.

@cindex return stack
@item
A return stack -- for storing the return addresses of colon
definitions and other (non-FP) data.

@cindex locals stack
@item
A locals stack for storing local variables.
@end itemize

@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
@cindex data stack manipulation words
@cindex stack manipulations words, data stack

doc-drop
doc-nip
doc-dup
doc-over
doc-tuck
doc-swap
doc-pick
doc-rot
doc--rot
doc-?dup
doc-roll
doc-2drop
doc-2nip
doc-2dup
doc-2over
doc-2tuck
doc-2swap
doc-2rot

@node Floating point stack, Return stack, Data stack, Stack Manipulation
@subsection Floating point stack
@cindex floating-point stack manipulation words
@cindex stack manipulation words, floating-point stack

Whilst every sane Forth has a separate floating-point stack, it is not
strictly required; an ANS Forth system could theoretically keep
floating-point numbers on the data stack. As an 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 they work also for a
unified stack model, but we do not recommend trying it. Instead, just
say that your program has an environmental dependency on a separate
floating-point stack.

doc-floating-stack

doc-fdrop
doc-fnip
doc-fdup
doc-fover
doc-ftuck
doc-fswap
doc-fpick
doc-frot

@node Return stack, Locals stack, Floating point stack, Stack Manipulation
@subsection Return stack
@cindex return stack manipulation words
@cindex stack manipulation words, return stack

@cindex return stack and locals
@cindex locals and return stack
A Forth system is allowed to keep local variables on the
return stack. This is reasonable, as local variables usually eliminate
the need to use the return stack explicitly. So, if you want to produce
a standard compliant program and you are using local variables in a
word, forget about return stack manipulations in that word (refer to the
standard document for the exact rules).

doc->r
doc-r>
doc-r@
doc-rdrop
doc-2>r
doc-2r>
doc-2r@
doc-2rdrop

@node Locals stack, Stack pointer manipulation, Return stack, Stack Manipulation
@subsection Locals stack

@comment TODO

@node Stack pointer manipulation,  , Locals stack, Stack Manipulation
@subsection Stack pointer manipulation
@cindex stack pointer manipulation words

doc-sp0
doc-sp@
doc-sp!
doc-fp0
doc-fp@
doc-fp!
doc-rp0
doc-rp@
doc-rp!
doc-lp0
doc-lp@
doc-lp!

@node Memory, Control Structures, Stack Manipulation, Words
@section Memory
@cindex memory words

@menu
* Memory model::                
* Dictionary allocation::       
* Heap Allocation::             
* Memory Access::               
* Address arithmetic::          
* Memory Blocks::               
@end menu

@node Memory model, Dictionary allocation, Memory, Memory
@subsection ANS Forth and Gforth memory models

@c The ANS Forth description is a mess (e.g., is the heap part of
@c the dictionary?), so let's not stick to closely with it.

ANS Forth considers a Forth system as consisting of several memories, of
which only @dfn{data space} is managed and accessible with the memory
words.  Memory not necessarily in data space includes the stacks, the
code (called code space) and the headers (called name space). In Gforth
everything is in data space, but the code for the primitives is usually
read-only.

Data space is divided into a number of areas: The (data space portion of
the) dictionary@footnote{Sometimes, the term @dfn{dictionary} is used to
refer to the search data structure embodied in word lists and headers,
because it is used for looking up names, just as you would in a
conventional dictionary.}, the heap, and a number of system-allocated
buffers.

In ANS Forth data space is also divided into contiguous regions.  You
can only use address arithmetic within a contiguous region, not between
them.  Usually each allocation gives you one contiguous region, but the
dictionary allocation words have additional rules (@pxref{Dictionary
allocation}).

Gforth provides one big address space, and address arithmetic can be
performed between any addresses. However, in the dictionary headers or
code are interleaved with data, so almost the only contiguous data space
regions there are those described by ANS Forth as contiguous; but you
can be sure that the dictionary is allocated towards increasing
addresses even between contiguous regions.  The memory order of
allocations in the heap is platform-dependent (and possibly different
from one run to the next).

@subsubsection ANS Forth dictionary details

@c !! I have deleted some of the stuff this section refers to - anton

This section is just informative, you can skip it if you are in a hurry.

When you create a colon definition, the text interpreter compiles the
code for the definition into the code space and compiles the name
of the definition into the header space, together with other
information about the definition (such as its execution token).

When you create a variable, the execution of @code{variable} will
compile some code, assign one cell in data space, and compile the name
of the variable into the header space.

@cindex memory regions - relationship between them
ANS Forth does not specify the relationship between the three memory
regions, and specifies that a Standard program must not access code or
data space directly -- it may only access data space directly. In
addition, the Standard defines what relationships you may and may not
rely on when allocating regions in data space. These constraints are
simply a reflection of the many diverse techniques that are used to
implement Forth systems; understanding and following the requirements of
the Standard allows you to write portable programs -- programs that run
in the same way on any of these diverse systems. Another way of looking
at this is to say that ANS Forth was designed to permit compliant Forth
systems to be implemented in many diverse ways.

@cindex memory regions - how they are assigned
Here are some examples of ways in which name, code and data spaces
might be assigned in different Forth implementations:

@itemize @bullet
@item
For a Forth system that runs from RAM under a general-purpose operating
system, it can be convenient to interleave name, code and data spaces in
a single contiguous memory region. This organisation can be
memory-efficient (for example, because the relationship between the name
dictionary entry and the associated code space entry can be
implicit, rather than requiring an explicit memory pointer to reference
from the header space and the code space). This is the
organisation used by Gforth, as this example@footnote{The addresses
in the example have been truncated to fit it onto the page, and the
addresses and data shown will not match the output from your system} shows:
@example
hex
variable fred 123456 fred !
variable jim abcd jim !
: foo + / - ;
' fred 10 - 50 dump 
..80: 5C 46 0E 40  84 66 72 65 - 64 20 20 20  20 20 20 20  \F.@.fred       
..90: D0 9B 04 08  00 00 00 00 - 56 34 12 00  80 46 0E 40  ........V4...F.@
..A0: 83 6A 69 6D  20 20 20 20 - D0 9B 04 08  00 00 00 00  .jim    ........
..B0: CD AB 00 00  9C 46 0E 40 - 83 66 6F 6F  20 20 20 20  .....F.@.foo    
..C0: 80 9B 04 08  00 00 00 00 - E4 2E 05 08  0C 2F 05 08  ............./..
@end example

@item
For a high-performance system running on a modern RISC processor with a
modified Harvard architecture (one that has a unified main memory but
separate instruction and data caches), it is desirable to separate
processor instructions from processor data. This encourages a high cache
density and therefore a high cache hit rate. The Forth code space
is not necessarily made up entirely of processor instructions; its
nature is dependent upon the Forth implementation. 

@item
A Forth compiler that runs on a segmented 8086 processor could be
designed to interleave the name, code and data spaces within a single
64Kbyte segment. A more common implementation choice is to use a
separate 64Kbyte segment for each region, which provides more memory
overall but provides an address map in which only the data space is
accessible.

@item
Microprocessors exist that run Forth (or many of the primitives required
to implement the Forth virtual machine efficiently) directly. On these
processors, the relationship between name, code and data spaces may be
imposed as a side-effect of the architecture of the processor.

@item
A Forth compiler that executes from ROM on an embedded system needs its
data space separated from the name and code spaces so that the data
space can be mapped to a RAM area.

@item 
A Forth compiler that runs on an embedded system may have a requirement
for a small memory footprint. On such a system it can be useful to
separate the header space from the data and code spaces; once the
application has been compiled, the header space is no longer
required@footnote{more strictly speaking, most applications can be
designed so that this is the case}. The header space can be deleted
entirely, or could be stored in memory on a remote @i{host} system for
debug and development purposes. In the latter case, the compiler running
on the @i{target} system could implement a protocol across a
communication link that would allow it to interrogate the header space.
@end itemize


@node Dictionary allocation, Heap Allocation, Memory model, Memory
@subsection Dictionary allocation
@cindex reserving data space
@cindex data space - reserving some

Dictionary allocation is a stack-oriented allocation scheme, i.e., if
you want to deallocate X, you also deallocate everything
allocated after X.

The allocations using the words below are contiguous and grow the region
towards increasing addresses.  Other words that allocate dictionary
memory of any kind (i.e., defining words including @code{:noname}) end
the contiguous region and start a new one.

In ANS Forth only @code{create}d words are guaranteed to produce an
address that is the start of the following contiguous region.  In
particular, the cell allocated by @code{variable} is not guaranteed to
be contiguous with following @code{allot}ed memory.

You can deallocate memory by using @code{allot} with a negative argument
(with some restrictions, see @code{allot}). For larger deallocations use
@code{marker}.


doc-here
doc-unused
doc-allot
doc-c,
doc-f,
doc-,
doc-2,
@cindex user space
doc-udp
doc-uallot

Memory accesses have to be aligned (@pxref{Address arithmetic}). So of
course you should allocate memory in an aligned way, too. I.e., before
allocating allocating a cell, @code{here} must be cell-aligned, etc.
The words below align @code{here} if it is not already.  Basically it is
only already aligned for a type, if the last allocation was a multiple
of the size of this type and if @code{here} was aligned for this type
before.

After freshly @code{create}ing a word, @code{here} is @code{align}ed in
ANS Forth (@code{maxalign}ed in Gforth).

doc-align
doc-falign
doc-sfalign
doc-dfalign
doc-maxalign
doc-cfalign


@node Heap Allocation, Memory Access, Dictionary allocation, Memory
@subsection Heap allocation
@cindex heap allocation
@cindex dynamic allocation of memory
@cindex memory-allocation word set

Heap allocation supports deallocation of allocated memory in any
order. Dictionary allocation is not affected by it (i.e., it does not
end a contiguous region). In Gforth, these words are implemented using
the standard C library calls malloc(), free() and resize().

doc-allocate
doc-free
doc-resize


@node Memory Access, Address arithmetic, Heap Allocation, Memory
@subsection Memory Access
@cindex memory access words

doc-@
doc-!
doc-+!
doc-c@
doc-c!
doc-2@
doc-2!
doc-f@
doc-f!
doc-sf@
doc-sf!
doc-df@
doc-df!

@node Address arithmetic, Memory Blocks, Memory Access, Memory
@subsection Address arithmetic
@cindex address arithmetic words

Address arithmetic is the foundation on which data structures like
arrays, records (@pxref{Structures}) and objects (@pxref{Object-oriented
Forth}) are built.

ANS Forth does not specify the sizes of the data types. Instead, it
offers a number of words for computing sizes and doing address
arithmetic. Address arithmetic is performed in terms of address units
(aus); on most systems the address unit is one byte. Note that a
character may have more than one au, so @code{chars} is no noop (on
systems where it is a noop, it compiles to nothing).

@cindex alignment of addresses for types
ANS Forth also defines words for aligning addresses for specific
types. Many computers require that accesses to specific data types
must only occur at specific addresses; e.g., that cells may only be
accessed at addresses divisible by 4. Even if a machine allows unaligned
accesses, it can usually perform aligned accesses faster. 

For the performance-conscious: alignment operations are usually only
necessary during the definition of a data structure, not during the
(more frequent) accesses to it.

ANS Forth defines no words for character-aligning addresses. This is not
an oversight, but reflects the fact that addresses that are not
char-aligned have no use in the standard and therefore will not be
created.

@cindex @code{CREATE} and alignment
ANS Forth guarantees that addresses returned by @code{CREATE}d words
are cell-aligned; in addition, Gforth guarantees that these addresses
are aligned for all purposes.

Note that the ANS Forth word @code{char} has nothing to do with address
arithmetic.

doc-chars
doc-char+
doc-cells
doc-cell+
doc-cell
doc-aligned
doc-floats
doc-float+
doc-float
doc-faligned
doc-sfloats
doc-sfloat+
doc-sfaligned
doc-dfloats
doc-dfloat+
doc-dfaligned
doc-maxaligned
doc-cfaligned
doc-address-unit-bits

@node Memory Blocks,  , Address arithmetic, Memory
@subsection Memory Blocks
@cindex memory block words
@cindex character strings - moving and copying

Memory blocks often represent character strings; @xref{String Formats}
for ways of storing character strings in memory. @xref{Displaying
characters and strings} for other string-processing words.

Some of these words work on address units. Others work on character
units (increments of @code{CHAR}), and expect a @code{CHAR}-aligned
address. Choose the correct operation depending upon your data type.

When copying characters between overlapping memory regions, choose
carefully between @code{cmove} and @code{cmove>}.

You can only use any of these words @i{portably} to access data space.

@comment TODO - think the naming of the arguments is wrong for move
@comment well, really it seems to be the Standard that's wrong; it
@comment describes MOVE as a word that requires a CELL-aligned source
@comment and destination address but a xtranfer count that need not
@comment be a multiple of CELL.
doc-move
doc-erase
doc-cmove
doc-cmove>
doc-fill
doc-blank
doc-compare
doc-search
doc--trailing
doc-/string

@comment TODO examples


@node Control Structures, Defining Words, Memory, Words
@section Control Structures
@cindex control structures

Control structures in Forth cannot be used interpretively, only in a
colon definition@footnote{To be precise, they have no interpretation
semantics (@pxref{Interpretation and Compilation Semantics}).}. We do
not like this limitation, but have not seen a satisfying way around it
yet, although many schemes have been proposed.

@menu
* Selection::                   IF ... ELSE ... ENDIF
* Simple Loops::                BEGIN ...
* Counted Loops::               DO
* Arbitrary control structures::
* Calls and returns::
* Exception Handling::          
@end menu

@node Selection, Simple Loops, Control Structures, Control Structures
@subsection Selection
@cindex selection control structures
@cindex control structures for selection

@c what's the purpose of all these @i? Maybe we should define a macro
@c so we can produce logical markup.  - anton

@cindex @code{IF} control structure
@example
@i{flag}
IF
  @i{code}
ENDIF
@end example
@noindent

@var{code} is executed if @var{flag} is non-zero (that's truth as far as
@code{IF} etc. are concerned).

@example
@i{flag}
IF
  @i{code1}
ELSE
  @i{code2}
ENDIF
@end example

If @var{flag} is true, perform @var{code1}, otherwise @var{code2}.

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
recommend using @code{ENDIF}, because it is less confusing for people
who also know other languages (and is not prone to reinforcing negative
prejudices against Forth in these people). Adding @code{ENDIF} to a
system that only supplies @code{THEN} is simple:
@example
: ENDIF   POSTPONE THEN ; immediate
@end example

[According to @cite{Webster's New Encyclopedic Dictionary}, @dfn{then
(adv.)}  has the following meanings:
@quotation
... 2b: following next after in order ... 3d: as a necessary consequence
(if you were there, then you saw them).
@end quotation
Forth's @code{THEN} has the meaning 2b, whereas @code{THEN} in Pascal
and many other programming languages has the meaning 3d.]

Gforth also provides the words @code{?DUP-IF} and @code{?DUP-0=-IF}, so
you can avoid using @code{?dup}. Using these alternatives is also more
efficient than using @code{?dup}. Definitions in ANS Forth
for @code{ENDIF}, @code{?DUP-IF} and @code{?DUP-0=-IF} are provided in
@file{compat/control.fs}.

@cindex @code{CASE} control structure
@example
@i{n}
CASE
  @i{n1} OF @i{code1} ENDOF
  @i{n2} OF @i{code2} ENDOF
  @dots{}
ENDCASE
@end example

Executes the first @i{codei}, where the @i{ni} is equal to
@i{n}. A default case can be added by simply writing the code after
the last @code{ENDOF}. It may use @i{n}, which is on top of the stack,
but must not consume it.

@node Simple Loops, Counted Loops, Selection, Control Structures
@subsection Simple Loops
@cindex simple loops
@cindex loops without count 

@cindex @code{WHILE} loop
@example
BEGIN
  @i{code1}
  @i{flag}
WHILE
  @i{code2}
REPEAT
@end example

@i{code1} is executed and @i{flag} is computed. If it is true,
@i{code2} is executed and the loop is restarted; If @i{flag} is
false, execution continues after the @code{REPEAT}.

@cindex @code{UNTIL} loop
@example
BEGIN
  @i{code}
  @i{flag}
UNTIL
@end example

@i{code} is executed. The loop is restarted if @code{flag} is false.

@cindex endless loop
@cindex loops, endless
@example
BEGIN
  @i{code}
AGAIN
@end example

This is an endless loop.

@node Counted Loops, Arbitrary control structures, Simple Loops, Control Structures
@subsection Counted Loops
@cindex counted loops
@cindex loops, counted
@cindex @code{DO} loops

The basic counted loop is:
@example
@i{limit} @i{start}
?DO
  @i{body}
LOOP
@end example

This performs one iteration for every integer, starting from @i{start}
and up to, but excluding @i{limit}. The counter, or @i{index}, can be
accessed with @code{i}. For example, the loop:
@example
10 0 ?DO
  i .
LOOP
@end example
@noindent
prints @code{0 1 2 3 4 5 6 7 8 9}

The index of the innermost loop can be accessed with @code{i}, the index
of the next loop with @code{j}, and the index of the third loop with
@code{k}.

doc-i
doc-j
doc-k

The loop control data are kept on the return stack, so there are some
restrictions on mixing return stack accesses and counted loop words. In
particuler, if you put values on the return stack outside the loop, you
cannot read them inside the loop@footnote{well, not in a way that is
portable.}. If you put values on the return stack within a loop, you
have to remove them before the end of the loop and before accessing the
index of the loop.

There are several variations on the counted loop:

@itemize @bullet
@item
@code{LEAVE} leaves the innermost counted loop immediately; execution
continues after the associated @code{LOOP} or @code{NEXT}. For example:

@example
10 0 ?DO  i DUP . 3 = IF LEAVE THEN LOOP
@end example
prints @code{0 1 2 3}


@item
@code{UNLOOP} prepares for an abnormal loop exit, e.g., via
@code{EXIT}. @code{UNLOOP} removes the loop control parameters from the
return stack so @code{EXIT} can get to its return address. For example:

@example
: demo 10 0 ?DO i DUP . 3 = IF UNLOOP EXIT THEN LOOP ." Done" ;
@end example
prints @code{0 1 2 3}


@item
If @i{start} is greater than @i{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 @i{start} is greater than
@i{limit}; @code{+DO} is for signed loop parameters, @code{U+DO} for
unsigned loop parameters.

@item
@code{?DO} can 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.

@item
@code{LOOP} can be replaced with @code{@i{n} +LOOP}; this updates the
index by @i{n} instead of by 1. The loop is terminated when the border
between @i{limit-1} and @i{limit} is crossed. E.g.:

@example
4 0 +DO  i .  2 +LOOP
@end example
@noindent
prints @code{0 2}

@example
4 1 +DO  i .  2 +LOOP
@end example
@noindent
prints @code{1 3}


@cindex negative increment for counted loops
@cindex counted loops with negative increment
The behaviour of @code{@i{n} +LOOP} is peculiar when @i{n} is negative:

@example
-1 0 ?DO  i .  -1 +LOOP
@end example
@noindent
prints @code{0 -1}

@example
0 0 ?DO  i .  -1 +LOOP
@end example
prints nothing.

Therefore we recommend avoiding @code{@i{n} +LOOP} with negative
@i{n}. One alternative is @code{@i{u} -LOOP}, which reduces the
index by @i{u} each iteration. The loop is terminated when the border
between @i{limit+1} and @i{limit} is crossed. Gforth also provides
@code{-DO} and @code{U-DO} for down-counting loops. E.g.:

@example
-2 0 -DO  i .  1 -LOOP
@end example
@noindent
prints @code{0 -1}

@example
-1 0 -DO  i .  1 -LOOP
@end example
@noindent
prints @code{0}

@example
0 0 -DO  i .  1 -LOOP
@end example
@noindent
prints nothing.

@end itemize

Unfortunately, @code{+DO}, @code{U+DO}, @code{-DO}, @code{U-DO} and
@code{-LOOP} are not defined in ANS Forth. However, an implementation
for these words that uses only standard words is provided in
@file{compat/loops.fs}.


@cindex @code{FOR} loops
Another counted loop is:
@example
@i{n}
FOR
  @i{body}
NEXT
@end example
This is the preferred loop of native code compiler writers who are too
lazy to optimize @code{?DO} loops properly. This loop structure is not
defined in ANS Forth. In Gforth, this loop iterates @i{n+1} times;
@code{i} produces values starting with @i{n} and ending with 0. Other
Forth systems may behave differently, even if 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
@cindex control structures, user-defined

@cindex control-flow stack
ANS Forth permits and supports using control structures in a non-nested
way. Information about incomplete control structures is stored on the
control-flow stack. This stack may be implemented on the Forth data
stack, and this is what we have done in Gforth.

@cindex @code{orig}, control-flow stack item
@cindex @code{dest}, control-flow stack item
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
building any control structure possible (except control structures that
need storage, like calls, coroutines, and backtracking).

doc-if
doc-ahead
doc-then
doc-begin
doc-until
doc-again
doc-cs-pick
doc-cs-roll

The Standard words @code{CS-PICK} and @code{CS-ROLL} allow you to
manipulate the control-flow stack in a portable way. Without them, you
would need to know how many stack items are occupied by a control-flow
entry (many systems use one cell. In Gforth they currently take three,
but this may change in the future).

Some standard control structure words are built from these words:

doc-else
doc-while
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:

doc-?do
doc-+do
doc-u+do
doc--do
doc-u-do
doc-do
doc-for
doc-loop
doc-+loop
doc--loop
doc-next
doc-leave
doc-?leave
doc-unloop
doc-done

The standard does not allow using @code{CS-PICK} and @code{CS-ROLL} on
@i{do-sys}. Gforth allows it, but it's your job to ensure that for
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
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{DONE}).

Another group of control structure words are:

doc-case
doc-endcase
doc-of
doc-endof

@i{case-sys} and @i{of-sys} cannot be processed using @code{CS-PICK} and
@code{CS-ROLL}.

@subsubsection Programming Style

In order to ensure readability we recommend that you do not create
arbitrary control structures directly, but define new control structure
words for the control structure you want and use these words in your
program. For example, instead of writing:

@example
BEGIN
  ...
IF [ 1 CS-ROLL ]
  ...
AGAIN THEN
@end example

@noindent
we recommend defining control structure words, e.g.,

@example
: WHILE ( DEST -- ORIG DEST )
 POSTPONE IF
 1 CS-ROLL ; immediate

: REPEAT ( orig dest -- )
 POSTPONE AGAIN
 POSTPONE THEN ; immediate
@end example

@noindent
and then using these to create the control structure:

@example
BEGIN
  ...
WHILE
  ...
REPEAT
@end example

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
necessary to define them.

@node Calls and returns, Exception Handling, Arbitrary control structures, Control Structures
@subsection Calls and returns
@cindex calling a definition
@cindex returning from a definition

@cindex recursive definitions
A definition can be called simply be writing the name of the definition
to be called. Normally a definition is invisible during its own
definition. If you want to write a directly recursive definition, you
can use @code{recursive} to make the current definition visible, or
@code{recurse} to call the current definition directly.

doc-recursive
doc-recurse

@comment TODO add example of the two recursion methods
@quotation
@progstyle
I prefer using @code{recursive} to @code{recurse}, because calling the
definition by name is more descriptive (if the name is well-chosen) than
the somewhat cryptic @code{recurse}.  E.g., in a quicksort
implementation, it is much better to read (and think) ``now sort the
partitions'' than to read ``now do a recursive call''.
@end quotation

For mutual recursion, use @code{Defer}red words, like this:

@example
Defer foo

: bar ( ... -- ... )
 ... foo ... ;

:noname ( ... -- ... )
 ... bar ... ;
IS foo
@end example

Deferred words are discussed in more detail in @ref{Simple
Defining Words}.

The current definition returns control to the calling definition when
the end of the definition is reached or @code{EXIT} is encountered.

doc-exit
doc-;s

@node Exception Handling,  , Calls and returns, Control Structures
@subsection Exception Handling
@cindex exceptions

If your program detects a fatal error condition, the simplest action
that it can take is to @code{quit}. This resets the return stack and
restarts the text interpreter, but does not print any error message.

The next stage in severity is to execute @code{abort}, which has the
same effect as @code{quit}, with the addition that it resets the data
stack.

A slightly more sophisticated approach is use use @code{abort"}, which
compiles a string to be used as an error message and does a conditional
@code{abort} at run-time. For example:

@example
@kbd{: checker abort" That flag was true" ." A false flag" ;@key{RET}}  ok
@kbd{0 checker@key{RET}} A false flag ok
@kbd{1 checker@key{RET}}
:1: That flag was true
1 checker
  ^^^^^^^
$400D1648 throw 
$400E4660
@end example

These simple techniques allow a program to react to a fatal error
condition, but they are not exactly user-friendly. The ANS Forth
Exception word set provides the pair of words @code{throw} and
@code{catch}, which can be used to provide sophisticated error-handling.

@code{catch} has a similar behaviour to @code{execute}, in that it takes
an @i{xt} as a parameter and starts execution of the xt. However,
before passing control to the xt, @code{catch} pushes an
@dfn{exception frame} onto the @dfn{exception stack}. This exception
frame is used to restore the system to a known state if a detected error
occurs during the execution of the xt. A typical way to use @code{catch}
would be:

@example
... ['] foo catch IF ...
@end example

@c TOS is undefined. - anton
Whilst @code{foo} executes, it can call other words to any level of
nesting, as usual.  If @code{foo} (and all the words that it calls)
execute successfully, control will ultimately pass to the word following
the @code{catch}, and there will be a 0 at TOS.  However, if any word
detects an error, it can terminate the execution of @code{foo} by
pushing a non-zero error code onto the stack and then performing a
@code{throw}. The execution of @code{throw} will pass control to the
word following the @code{catch}, but this time the TOS will hold the
error code. Therefore, the @code{IF} in the example can be used to
determine whether @code{foo} executed successfully.

This simple example shows how you can use @code{throw} and @code{catch}
to ``take over'' exception handling from the system:
@example
: my-div ['] / catch if ." DIVIDE ERROR" else ." OK.. " . then ;
@end example

The next example is more sophisticated and shows a multi-level
@code{throw} and @code{catch}. To understand this example, start at the
definition of @code{top-level} and work backwards:

@example
: lowest-level ( -- c )
    key dup 27 = if
	1 throw \ ESCAPE key pressed
    else
	." lowest-level successfull" CR
    then
;

: lower-level ( -- c )
    lowest-level
    \ at this level consider a CTRL-U to be a fatal error
    dup 21 = if \ CTRL-U
	2 throw
    else
	." lower-level successfull" CR
    then
;

: low-level ( -- c )
    ['] lower-level catch
    ?dup if
	\ error occurred - do we recognise it?
	dup 1 = if
	    \ ESCAPE key pressed.. pretend it was an E
	    [char] E
	else throw \ propogate the error upwards
	then
    then
    ." low-level successfull" CR
;

: top-level ( -- )
    CR ['] low-level catch \ CATCH is used like EXECUTE
    ?dup if \ error occurred..
	." Error " . ." occurred - contact your supplier"
    else
	." The '" emit ." ' key was pressed" CR
    then
;
@end example

The ANS Forth document assigns @code{throw} codes thus:

@itemize @bullet
@item
codes in the range -1 -- -255 are reserved to be assigned by the
Standard. Assignments for codes in the range -1 -- -58 are currently
documented in the Standard. In particular, @code{-1 throw} is equivalent
to @code{abort} and @code{-2 throw} is equivalent to @code{abort"}.
@item
codes in the range -256 -- -4095 are reserved to be assigned by the system.
@item
all other codes may be assigned by programs.
@end itemize

Gforth provides the word @code{exception} as a mechanism for assigning
system throw codes to applications. This allows multiple applications to
co-exist in memory without any clash of @code{throw} codes. A definition
of @code{exception} in ANS Forth is provided in
@file{compat/exception.fs}.

doc-quit
doc-abort
doc-abort"

doc-catch
doc-throw
doc---exception-exception


@c -------------------------------------------------------------
@node Defining Words, The Text Interpreter, Control Structures, Words
@section Defining Words
@cindex defining words

@menu
* Simple Defining Words::       Variables, values and constants
* Colon Definitions::
* User-defined Defining Words::
* Supplying names::
* Interpretation and Compilation Semantics::
@end menu

@node Simple Defining Words, Colon Definitions, Defining Words, Defining Words
@subsection Simple Defining Words
@cindex simple defining words
@cindex defining words, simple

@c split this section?

Defining words are used to create new entries in the dictionary. The
simplest defining word is @code{CREATE}. @code{CREATE} is used like
this:

@example
CREATE new-word1
@end example

@code{CREATE} is a parsing word that generates a dictionary entry for
@code{new-word1}. When @code{new-word1} is executed, all that it does is
leave an address on the stack. The address represents the value of
the data space pointer (@code{HERE}) at the time that @code{new-word1}
was defined. Therefore, @code{CREATE} is a way of associating a name
with the address of a region of memory.

doc-create

By extending this example to reserve some memory in data space, we end
up with a @i{variable}. Here are two different ways to do it:

@example
CREATE new-word2 1 cells allot  \ reserve 1 cell - initial value undefined
CREATE new-word3 4 ,            \ reserve 1 cell and initialise it (to 4)
@end example

The variable can be examined and modified using @code{@@} (``fetch'') and
@code{!} (``store'') like this:

@example
new-word2 @@ .      \ get address, fetch from it and display
1234 new-word2 !   \ new value, get address, store to it
@end example

As a final refinement, the whole code sequence can be wrapped up in a
defining word (pre-empting the subject of the next section), making it
easier to create new variables:

@example
: myvariable ( "name" -- a-addr ) CREATE 0 , ;

myvariable foo
myvariable joe

45 3 * foo !   \ set foo to 135
1234 joe !     \ set joe to 1234
3 joe +!       \ increment joe by 3.. to 1237
@end example

Not surprisingly, there is no need to define @code{myvariable}, since
Forth already has a definition @code{Variable}. It behaves in exactly
the same way as @code{myvariable}. Forth also provides @code{2Variable}
and @code{fvariable} for double and floating-point variables,
respectively.

doc-variable
doc-2variable
doc-fvariable

@cindex arrays
A similar mechanism can be used to create arrays. For example, an
80-character text input buffer:

@example
CREATE text-buf 80 chars allot

text-buf 0 chars c@@ \ the 1st character (offset 0)
text-buf 3 chars c@@ \ the 4th character (offset 3)
@end example

You can build arbitrarily complex data structures by allocating
appropriate areas of memory. @xref{Structures} for further discussions
of this, and to learn about some Gforth tools that make it easier.

@cindex user variables
@cindex user space
The defining word @code{User} behaves in the same way as @code{Variable}.
The difference is that it reserves space in @i{user (data) space} rather
than normal data space. In a Forth system that has a multi-tasker, each
task has its own set of user variables.

doc-user

@comment TODO is that stuff about user variables strictly correct? Is it
@comment just terminal tasks that have user variables?
@comment should document tasker.fs (with some examples) elsewhere
@comment in this manual, then expand on user space and user variables.

After @code{CREATE} and @code{Variable}s, the next defining word to
consider is @code{Constant}. @code{Constant} allows you to declare a
fixed value and refer to it by name. For example:

@example
12 Constant INCHES-PER-FOOT
3E+08 fconstant SPEED-O-LIGHT
@end example

A @code{Variable} can be both read and written, so its run-time
behaviour is to supply an address through which its current value can be
manipulated. In contrast, the value of a @code{Constant} cannot be
changed once it has been declared@footnote{Well, often it can be -- but
not in a Standard, portable way. It's safer to use a @code{Value} (read
on).} so it's not necessary to supply the address -- it is more
efficient to return the value of the constant directly. That's exactly
what happens; the run-time effect of a constant is to put its value on
the top of the stack (@ref{User-defined Defining Words} describes one
way of implementing @code{Constant}).

Gforth also provides @code{2Constant} and @code{fconstant} for defining
double and floating-point constants, respectively.

doc-constant
doc-2constant
doc-fconstant

@c that's too deep, and it's not necessarily true for all ANS Forths. - anton
Constants in Forth behave differently from their equivalents in other
programming languages. In other languages, a constant (such as an EQU in
assembler or a #define in C) only exists at compile-time; in the
executable program the constant has been translated into an absolute
number and, unless you are using a symbolic debugger, it's impossible to
know what abstract thing that number represents. In Forth a constant has
an entry in the header space and remains there after the code that
uses it has been defined. In fact, it must remain in the dictionary
since it has run-time duties to perform. For example:

@example
12 Constant INCHES-PER-FOOT
: FEET-TO-INCHES ( n1 -- n2 ) INCHES-PER-FOOT * ;
@end example

@cindex in-lining of constants
When @code{FEET-TO-INCHES} is executed, it will in turn execute the xt
associated with the constant @code{INCHES-PER-FOOT}. If you use
@code{see} to decompile the definition of @code{FEET-TO-INCHES}, you can
see that it makes a call to @code{INCHES-PER-FOOT}. Some Forth compilers
attempt to optimise constants by in-lining them where they are used. You
can force Gforth to in-line a constant like this:

@example
: FEET-TO-INCHES ( n1 -- n2 ) [ INCHES-PER-FOOT ] LITERAL * ;
@end example

If you use @code{see} to decompile @i{this} version of
@code{FEET-TO-INCHES}, you can see that @code{INCHES-PER-FOOT} is no
longer present. @xref{Interpret/Compile states} and @ref{Literals} on
how this works.

In-lining constants in this way might improve execution time
fractionally, and can ensure that a constant is now only referenced at
compile-time. However, the definition of the constant still remains in
the dictionary. Some Forth compilers provide a mechanism for controlling
a second dictionary for holding transient words such that this second
dictionary can be deleted later in order to recover memory
space. However, there is no standard way of doing this.

One aspect of constants and variables that can sometimes be confusing is
that they have different stack effects; one returns its value whilst the
other returns the address of its value. The defining word @code{Value}
provides an alternative to @code{Variable}, and has the same stack
effect as a constant. A @code{Value} needs an additional word, @code{TO}
to allow its value to be changed. Here are some examples:

@example
12 Value APPLES \ a Value is initialised when it is declared.. like a
                \ constant but unlike a variable
34 TO APPLES    \ Change the value of APPLES. TO is a parsing word
APPLES          \ puts 34 on the top of the stack.
@end example

doc-value
doc-to

The defining word @code{Defer} allows you to define a word by name
without defining its behaviour; the definition of its behaviour is
deferred. Here are two situation where this can be useful:

@itemize @bullet
@item
Where you want to allow the behaviour of a word to be altered later, and
for all precompiled references to the word to change when its behaviour
is changed.
@item
For mutual recursion; @xref{Calls and returns}.
@end itemize

In the following example, @code{foo} always invokes the version of
@code{greet} that prints ``@code{Good morning}'' whilst @code{bar}
always invokes the version that prints ``@code{Hello}''. There is no way
of getting @code{foo} to use the later version without re-ordering the
source code and recompilng it.

@example
: greet ." Good morning" ;
: foo ... greet ... ;
: greet ." Hello" ;
: bar ... greet ... ;
@end example

This problem can be solved by defining @code{greet} as a @code{Defer}red
word. The behaviour of a @code{Defer}red word can be defined and
redefined at any time by using @code{IS} to associate the xt of a
previously-defined word with it. The previous example becomes:

@example
Defer greet
: foo ... greet ... ;
: bar ... greet ... ;
: greet1 ." Good morning" ;
: greet2 ." Hello" ;
' greet2 <IS> greet  \ make greet behave like greet2
@end example

One thing to note is that @code{<IS>} consumes it's name when it is
executed.  If you want to specify the name at compile time, use
@code{[IS]}:

@example
: set-greet ( xt -- )
  [IS] greet ;

' greet1 set-greet
@end example

A deferred word can only inherit default semantics from the xt (because
that is all that an xt can represent -- @pxref{Tokens for Words} for
more discussion of this). However, the semantics of the deferred word
itself can be modified at the time that it is defined. For example:

@example
: bar .... ; compile-only
Defer fred immediate
Defer jim

' bar <IS> jim  \ jim has default semantics
' bar <IS> fred \ fred is immediate
@end example

doc-defer
doc-<is>
doc-[is]
@comment TODO document these: what's defers [is]
doc-what's
doc-defers

Definitions in ANS Forth for @code{defer}, @code{<is>} and
@code{[is]} are provided in @file{compat/defer.fs}.

The defining word @code{Alias} allows you to define a word by name that
has the same behaviour as some other word. Here are two situation where
this can be useful:

@itemize @bullet
@item
When you want access to a word's definition from a different word list
(for an example of this, see the definition of the @code{Root} word list
in the Gforth source).
@item
When you want to create a synonym; a definition that can be known by
either of two names (for example, @code{THEN} and @code{ENDIF} are
aliases).
@end itemize

The word whose behaviour the alias is to inherit is represented by an
xt. Therefore, the alias only inherits default semantics from its
ancestor. The semantics of the alias itself can be modified at the time
that it is defined. For example:

@example
: foo ... ; immediate

' foo Alias bar \ bar is not an immediate word
' foo Alias fooby immediate \ fooby is an immediate word
@end example

@c "combined words" is an undefined term
Words that are aliases have the same xt, different headers in the
dictionary, and consequently different name tokens (@pxref{Tokens for
Words}) and possibly different immediate flags.  An alias can only have
default or immediate compilation semantics; you can define aliases for
combined words with @code{interpret/compile:}.

@c distribute this to the appropriate paragraphs? - anton
doc-alias

@node Colon Definitions, User-defined Defining Words, Simple Defining Words, Defining Words
@subsection Colon Definitions
@cindex colon definitions

@example
: name ( ... -- ... )
    word1 word2 word3 ;
@end example

@noindent
Creates a word called @code{name} that, upon execution, executes
@code{word1 word2 word3}. @code{name} is a @dfn{(colon) definition}.

The explanation above is somewhat superficial. @xref{Your first
definition} for simple examples of colon definitions, then
@xref{Interpretation and Compilation Semantics} for an in-depth
discussion of some of the issues involved.

doc-:
doc-;

@node User-defined Defining Words, Supplying names, Colon Definitions, Defining Words
@subsection User-defined Defining Words
@cindex user-defined defining words
@cindex defining words, user-defined

You can create a new defining word by wrapping defining-time code around
an existing defining word and putting the sequence in a colon
definition. For example, suppose that you have a word @code{stats} that
gathers statistics about colon definitions given the @i{xt} of the
definition, and you want every colon definition in your application to
make a call to @code{stats}. You can define and use a new version of
@code{:} like this:

@example
: stats ( xt -- ) DUP ." (Gathering statistics for " . ." )"
  ... ;  \ other code

: my: : lastxt postpone literal ['] stats compile, ;

my: foo + - ;
@end example

When @code{foo} is defined using @code{my:} these steps occur:

@itemize @bullet
@item
@code{my:} is executed.
@item
The @code{:} within the definition (the one between @code{my:} and
@code{lastxt}) is executed, and does just what it always does; it parses
the input stream for a name, builds a dictionary header for the name
@code{foo} and switches @code{state} from interpret to compile.
@item
The word @code{lastxt} is executed. It puts the @i{xt} for the word that is
being defined -- @code{foo} -- onto the stack.
@item
The code that was produced by @code{postpone literal} is executed; this
causes the value on the stack to be compiled as a literal in the code
area of @code{foo}.
@item
The code @code{['] stats} compiles a literal into the definition of
@code{my:}. When @code{compile,} is executed, that literal -- the
execution token for @code{stats} -- is layed down in the code area of
@code{foo} , following the literal@footnote{Strictly speaking, the
mechanism that @code{compile,} uses to convert an @i{xt} into something
in the code area is implementation-dependent. A threaded implementation
might spit out the execution token directly whilst another
implementation might spit out a native code sequence.}.
@item
At this point, the execution of @code{my:} is complete, and control
returns to the text interpreter. The text interpreter is in compile
state, so subsequent text @code{+ -} is compiled into the definition of
@code{foo} and the @code{;} terminates the definition as always.
@end itemize

You can use @code{see} to decompile a word that was defined using
@code{my:} and see how it is different from a normal @code{:}
definition. For example:

@example
: bar + - ;  \ like foo but using : rather than my:
see bar
: bar
  + - ;
see foo
: foo
  107645672 stats + - ;

\ use ' stats . to show that 107645672 is the xt for stats
@end example


@c a deferred word is not neccessary for these examples. - anton
Rather than edit your application's source code to change every @code{:}
to a @code{my:}, use a deferred word:

@example
: real: : ;     \ retain access to the original
defer :         \ redefine as a deferred word
' my: IS :      \ use special version of :
\
\ load application here
\
' real: IS :    \ go back to the original
@end example

You can use techniques like this to make new defining words in terms of
@i{any} existing defining word.


@cindex defining defining words
@cindex @code{CREATE} ... @code{DOES>}
If you want the words defined with your defining words to behave
differently from words defined with standard defining words, you can
write your defining word like this:

@example
: def-word ( "name" -- )
    CREATE @i{code1}
DOES> ( ... -- ... )
    @i{code2} ;

def-word name
@end example

@cindex child words
This fragment defines a @dfn{defining word} @code{def-word} and then
executes it.  When @code{def-word} executes, it @code{CREATE}s a new
word, @code{name}, and executes the code @i{code1}. The code @i{code2}
is not executed at this time. The word @code{name} is sometimes called a
@dfn{child} of @code{def-word}.

When you execute @code{name}, the address of the body of @code{name} is
put on the data stack and @i{code2} is executed (the address of the body
of @code{name} is the address @code{HERE} returns immediately after the
@code{CREATE}).

@cindex atavism in child words
You can use @code{def-word} to define a set of child words that behave
differently, though atavistically; they all have a common run-time
behaviour determined by @i{code2}. Typically, the @i{code1} sequence
builds a data area in the body of the child word. The structure of the
data is common to all children of @code{def-word}, but the data values
are specific -- and private -- to each child word. When a child word is
executed, the address of its private data area is passed as a parameter
on TOS to be used and manipulated@footnote{It is legitimate both to read
and write to this data area.} by @i{code2}.

The two fragments of code that make up the defining words act (are
executed) at two completely separate times:

@itemize @bullet
@item
At @i{define time}, the defining word executes @i{code1} to generate a
child word
@item
At @i{child execution time}, when a child word is invoked, @i{code2}
is executed, using parameters (data) that are private and specific to
the child word.
@end itemize

@c NAC I think this is a really bad example, because it diminishes
@c rather than emphasising the fact that some important stuff happens
@c at define time, and other important stuff happens at child-invocation
@c time, and that those two times are potentially very different.

@c Well, IMO CREATE-DOES> is usually presented with much ado, making
@c people think that it's hard to understand, and making those people who
@c understand it easily think that it's hyped.  I prefer presenting it in a
@c diminished way and only emphasize the special issues later. - anton

In other words, if you make the following definitions:
@example
: def-word1 ( "name" -- )
    CREATE @i{code1} ;

: action1 ( ... -- ... )
    @i{code2} ;

def-word1 name1
@end example

Using @code{name1 action1} is equivalent to using @code{name}.

The classic example is that you can define @code{CONSTANT} in this way:

@example
: CONSTANT ( w "name" -- )
    CREATE ,
DOES> ( -- w )
    @@ ;
@end example

@comment There is a beautiful description of how this works and what
@comment it does in the Forthwrite 100th edition.. as well as an elegant
@comment commentary on the Counting Fruits problem.

When you create a constant with @code{5 CONSTANT five}, a set of
define-time actions take place; first a new word @code{five} is created,
then the value 5 is laid down in the body of @code{five} with
@code{,}. When @code{five} is invoked, the address of the body is put on
the stack, and @code{@@} retrieves the value 5. The word @code{five} has
no code of its own; it simply contains a data field and a pointer to the
code that follows @code{DOES>} in its defining word. That makes words
created in this way very compact.

The final example in this section is intended to remind you that space
reserved in @code{CREATE}d words is @i{data} space and therefore can be
both read and written by a Standard program@footnote{Exercise: use this
example as a starting point for your own implementation of @code{Value}
and @code{TO} -- if you get stuck, investigate the behaviour of @code{'} and
@code{[']}.}:

@example
: foo ( "name" -- )
    CREATE -1 ,
DOES> ( -- )
    @@ . ;

foo first-word
foo second-word

123 ' first-word >BODY !
@end example

If @code{first-word} had been a @code{CREATE}d word, we could simply
have executed it to get the address of its data field. However, since it
was defined to have @code{DOES>} actions, its execution semantics are to
perform those @code{DOES>} actions. To get the address of its data field
it's necessary to use @code{'} to get its xt, then @code{>BODY} to
translate the xt into the address of the data field.  When you execute
@code{first-word}, it will display @code{123}. When you execute
@code{second-word} it will display @code{-1}.

@cindex stack effect of @code{DOES>}-parts
@cindex @code{DOES>}-parts, stack effect
In the examples above the stack comment after the @code{DOES>} specifies
the stack effect of the defined words, not the stack effect of the
following code (the following code expects the address of the body on
the top of stack, which is not reflected in the stack comment). This is
the convention that I use and recommend (it clashes a bit with using
locals declarations for stack effect specification, though).

@subsubsection Applications of @code{CREATE..DOES>}
@cindex @code{CREATE} ... @code{DOES>}, applications

You may wonder how to use this feature. Here are some usage patterns:

@cindex factoring similar colon definitions
When you see a sequence of code occurring several times, and you can
identify a meaning, you will factor it out as a colon definition. When
you see similar colon definitions, you can factor them using
@code{CREATE..DOES>}. E.g., an assembler usually defines several words
that look very similar:
@example
: ori, ( reg-target reg-source n -- )
    0 asm-reg-reg-imm ;
: andi, ( reg-target reg-source n -- )
    1 asm-reg-reg-imm ;
@end example

@noindent
This could be factored with:
@example
: reg-reg-imm ( op-code -- )
    CREATE ,
DOES> ( reg-target reg-source n -- )
    @@ asm-reg-reg-imm ;

0 reg-reg-imm ori,
1 reg-reg-imm andi,
@end example

@cindex currying
Another view of @code{CREATE..DOES>} is to consider it as a crude way to
supply a part of the parameters for a word (known as @dfn{currying} in
the functional language community). E.g., @code{+} needs two
parameters. Creating versions of @code{+} with one parameter fixed can
be done like this:
@example
: curry+ ( n1 -- )
    CREATE ,
DOES> ( n2 -- n1+n2 )
    @@ + ;

 3 curry+ 3+
-2 curry+ 2-
@end example

@subsubsection The gory details of @code{CREATE..DOES>}
@cindex @code{CREATE} ... @code{DOES>}, details

doc-does>

@cindex @code{DOES>} in a separate definition
This means that you need not use @code{CREATE} and @code{DOES>} in the
same definition; you can put the @code{DOES>}-part in a separate
definition. This allows us to, e.g., select among different @code{DOES>}-parts:
@example
: does1 
DOES> ( ... -- ... )
    ... ;

: does2
DOES> ( ... -- ... )
    ... ;

: def-word ( ... -- ... )
    create ...
    IF
       does1
    ELSE
       does2
    ENDIF ;
@end example

In this example, the selection of whether to use @code{does1} or
@code{does2} is made at compile-time; at the time that the child word is
@code{CREATE}d.

@cindex @code{DOES>} in interpretation state
In a standard program you can apply a @code{DOES>}-part only if the last
word was defined with @code{CREATE}. In Gforth, the @code{DOES>}-part
will override the behaviour of the last word defined in any case. In a
standard program, you can use @code{DOES>} only in a colon
definition. In Gforth, you can also use it in interpretation state, in a
kind of one-shot mode; for example:
@example
CREATE name ( ... -- ... )
  @i{initialization}
DOES>
  @i{code} ;
@end example

@noindent
is equivalent to the standard:
@example
:noname
DOES>
    @i{code} ;
CREATE name EXECUTE ( ... -- ... )
    @i{initialization}
@end example

You can get the address of the body of a word with:

doc->body

@node Supplying names, Interpretation and Compilation Semantics, User-defined Defining Words, Defining Words
@subsection Supplying the name of a defined word
@cindex names for defined words
@cindex defining words, name parameter

@cindex defining words, name given in a string
By default, a defining word takes the name for the defined word from the
input stream. Sometimes you want to supply the name from a string. You
can do this with:

doc-nextname

For example:

@example
s" foo" nextname create
@end example
@noindent
is equivalent to:
@example
create foo
@end example

@cindex defining words without name
Sometimes you want to define an @dfn{anonymous word}; a word without a
name. You can do this with:

doc-:noname

This leaves the execution token for the word on the stack after the
closing @code{;}. Here's an example in which a deferred word is
initialised with an @code{xt} from an anonymous colon definition:
@example
Defer deferred
:noname ( ... -- ... )
  ... ;
IS deferred
@end example

@noindent
Gforth provides an alternative way of doing this, using two separate
words:

doc-noname
@cindex execution token of last defined word
doc-lastxt

@noindent
The previous example can be rewritten using @code{noname} and
@code{lastxt}:

@example
Defer deferred
noname : ( ... -- ... )
  ... ;
lastxt IS deferred
@end example

@noindent
@code{noname} and @code{nextname} work with any defining word, not just
@code{:}.

@code{lastxt} also works when the last word was not defined as
@code{noname}. It also has the useful property that is is valid as soon
as the header for a definition has been build. Thus:

@example
lastxt . : foo [ lastxt . ] ; ' foo .
@end example

@noindent
prints 3 numbers; the last two are the same.


@node Interpretation and Compilation Semantics,  , Supplying names, Defining Words
@subsection Interpretation and Compilation Semantics
@cindex semantics, interpretation and compilation

@cindex interpretation semantics
The @dfn{interpretation semantics} of a word are what the text
interpreter does when it encounters the word in interpret state. It also
appears in some other contexts, e.g., the execution token returned by
@code{' @i{word}} identifies the interpretation semantics of
@i{word} (in other words, @code{' @i{word} execute} is equivalent to
interpret-state text interpretation of @code{@i{word}}).

@cindex compilation semantics
The @dfn{compilation semantics} of a word are what the text interpreter
does when it encounters the word in compile state. It also appears in
other contexts, e.g, @code{POSTPONE @i{word}} compiles@footnote{In
standard terminology, ``appends to the current definition''.} the
compilation semantics of @i{word}.

@cindex execution semantics
The standard also talks about @dfn{execution semantics}. They are used
only for defining the interpretation and compilation semantics of many
words. By default, the interpretation semantics of a word are to
@code{execute} its execution semantics, and the compilation semantics of
a word are to @code{compile,} its execution semantics.@footnote{In
standard terminology: The default interpretation semantics are its
execution semantics; the default compilation semantics are to append its
execution semantics to the execution semantics of the current
definition.}

@comment TODO expand, make it co-operate with new sections on text interpreter.

@cindex immediate words
@cindex compile-only words
You can change the semantics of the most-recently defined word:

doc-immediate
doc-compile-only
doc-restrict

Note that ticking (@code{'}) a compile-only word gives an error
(``Interpreting a compile-only word'').

Gforth also allows you to define words with arbitrary combinations of
interpretation and compilation semantics.

doc-interpret/compile:

This feature was introduced for implementing @code{TO} and @code{S"}. I
recommend that you do not define such words, as cute as they may be:
they make it hard to get at both parts of the word in some contexts.
E.g., assume you want to get an execution token for the compilation
part. Instead, define two words, one that embodies the interpretation
part, and one that embodies the compilation part.  Once you have done
that, you can define a combined word with @code{interpret/compile:} for
the convenience of your users.

You might try to use this feature to provide an optimizing
implementation of the default compilation semantics of a word. For
example, by defining:
@example
:noname
   foo bar ;
:noname
   POSTPONE foo POSTPONE bar ;
interpret/compile: opti-foobar
@end example

@noindent
as an optimizing version of:

@example
: foobar
    foo bar ;
@end example

Unfortunately, this does not work correctly with @code{[compile]},
because @code{[compile]} assumes that the compilation semantics of all
@code{interpret/compile:} words are non-default. I.e., @code{[compile]
opti-foobar} would compile compilation semantics, whereas
@code{[compile] foobar} would compile interpretation semantics.

@cindex state-smart words (are a bad idea)
Some people try to use @dfn{state-smart} words to emulate the feature provided
by @code{interpret/compile:} (words are state-smart if they check
@code{STATE} during execution). E.g., they would try to code
@code{foobar} like this:

@example
: foobar
  STATE @@
  IF ( compilation state )
    POSTPONE foo POSTPONE bar
  ELSE
    foo bar
  ENDIF ; immediate
@end example

Although this works if @code{foobar} is only processed by the text
interpreter, it does not work in other contexts (like @code{'} or
@code{POSTPONE}). E.g., @code{' foobar} will produce an execution token
for a state-smart word, not for the interpretation semantics of the
original @code{foobar}; when you execute this execution token (directly
with @code{EXECUTE} or indirectly through @code{COMPILE,}) in compile
state, the result will not be what you expected (i.e., it will not
perform @code{foo bar}). State-smart words are a bad idea. Simply don't
write them@footnote{For a more detailed discussion of this topic, see
@cite{@code{State}-smartness -- Why it is Evil and How to Exorcise it} by Anton
Ertl; presented at EuroForth '98 and available from
@url{http://www.complang.tuwien.ac.at/papers/ertl98.ps.gz}}!

@cindex defining words with arbitrary semantics combinations
It is also possible to write defining words that define words with
arbitrary combinations of interpretation and compilation semantics. In
general, they look like this:

@example
: def-word
    create-interpret/compile
    @i{code1}
interpretation>
    @i{code2}
<interpretation
compilation>
    @i{code3}
<compilation ;
@end example

For a @i{word} defined with @code{def-word}, the interpretation
semantics are to push the address of the body of @i{word} and perform
@i{code2}, and the compilation semantics are to push the address of
the body of @i{word} and perform @i{code3}. E.g., @code{constant}
can also be defined like this (except that the defined constants don't
behave correctly when @code{[compile]}d):

@example
: constant ( n "name" -- )
    create-interpret/compile
    ,
interpretation> ( -- n )
    @@
<interpretation
compilation> ( compilation. -- ; run-time. -- n )
    @@ postpone literal
<compilation ;
@end example

doc-create-interpret/compile
doc-interpretation>
doc-<interpretation
doc-compilation>
doc-<compilation

Words defined with @code{interpret/compile:} and
@code{create-interpret/compile} have an extended header structure that
differs from other words; however, unless you try to access them with
plain address arithmetic, you should not notice this. Words for
accessing the header structure usually know how to deal with this; e.g.,
@code{'} @i{word} @code{>body} also gives you the body of a word created
with @code{create-interpret/compile}.

doc-postpone
@comment TODO -- expand glossary text for POSTPONE

@c ----------------------------------------------------------
@node The Text Interpreter, Tokens for Words, Defining Words, Words
@section  The Text Interpreter
@cindex interpreter - outer
@cindex text interpreter
@cindex outer interpreter

@c Should we really describe all these ugly details?  IMO the text
@c interpreter should be much cleaner, but that may not be possible within
@c ANS Forth. - anton

The text interpreter@footnote{This is an expanded version of the
material in @ref{Introducing the Text Interpreter}.} is an endless loop
that processes input from the current input device. It is also called
the outer interpreter, in contrast to the inner interpreter
(@pxref{Engine}) which executes the compiled Forth code on interpretive
implementations.

@cindex interpret state
@cindex compile state
The text interpreter operates in one of two states: @dfn{interpret
state} and @dfn{compile state}. The current state is defined by the
aptly-named variable, @code{state}.

This section starts by describing how the text interpreter behaves when
it is in interpret state, processing input from the user input device --
the keyboard. This is the mode that a Forth system is in after it starts
up.

@cindex input buffer
@cindex terminal input buffer
The text interpreter works from an area of memory called the @dfn{input
buffer}@footnote{When the text interpreter is processing input from the
keyboard, this area of memory is called the @dfn{terminal input buffer}
(TIB) and is addressed by the (obsolescent) words @code{TIB} and
@code{#TIB}.}, which stores your keyboard input when you press the
@key{RET} key. Starting at the beginning of the input buffer, it skips
leading spaces (called @dfn{delimiters}) then parses a string (a
sequence of non-space characters) until it reaches either a space
character or the end of the buffer. Having parsed a string, it makes two
attempts to process it:

@cindex dictionary
@itemize @bullet
@item
It looks for the string in a @dfn{dictionary} of definitions. If the
string is found, the string names a @dfn{definition} (also known as a
@dfn{word}) and the dictionary search returns information that allows
the text interpreter to perform the word's @dfn{interpretation
semantics}. In most cases, this simply means that the word will be
executed.
@item
If the string is not found in the dictionary, the text interpreter
attempts to treat it as a number, using the rules described in
@ref{Number Conversion}. If the string represents a legal number in the
current radix, the number is pushed onto a parameter stack (the data
stack for integers, the floating-point stack for floating-point
numbers).
@end itemize

If both attempts fail, or if the word is found in the dictionary but has
no interpretation semantics@footnote{This happens if the word was
defined as @code{COMPILE-ONLY}.} the text interpreter discards the
remainder of the input buffer, issues an error message and waits for
more input. If one of the attempts succeeds, the text interpreter
repeats the parsing process until the whole of the input buffer has been
processed, at which point it prints the status message ``@code{ ok}''
and waits for more input.

@cindex parse area
The text interpreter keeps track of its position in the input buffer by
updating a variable called @code{>IN} (pronounced ``to-in''). The value
of @code{>IN} starts out as 0, indicating an offset of 0 from the start
of the input buffer. The region from offset @code{>IN @@} to the end of
the input buffer is called the @dfn{parse area}@footnote{In other words,
the text interpreter processes the contents of the input buffer by
parsing strings from the parse area until the parse area is empty.}.
This example shows how @code{>IN} changes as the text interpreter parses
the input buffer:

@example
: remaining >IN @@ SOURCE 2 PICK - -ROT + SWAP
  CR ." ->" TYPE ." <-" ; IMMEDIATE 

1 2 3 remaining + remaining . 

: foo 1 2 3 remaining SWAP remaining ;
@end example

@noindent
The result is:

@example
->+ remaining .<-
->.<-5  ok

->SWAP remaining ;-<
->;<-  ok
@end example

@cindex parsing words
The value of @code{>IN} can also be modified by a word in the input
buffer that is executed by the text interpreter.  This means that a word
can ``trick'' the text interpreter into either skipping a section of the
input buffer@footnote{This is how parsing words work.} or into parsing a
section twice. For example:

@example
: lat ." <<lat>>" ;
: flat ." <<flat>>" >IN DUP @@ 3 - SWAP ! ;
@end example

@noindent
When @code{flat} is executed, this output is produced@footnote{Exercise
for the reader: what would happen if the @code{3} were replaced with
@code{4}?}:

@example
<<flat>><<lat>>
@end example

@noindent
Two important notes about the behaviour of the text interpreter:

@itemize @bullet
@item
It processes each input string to completion before parsing additional
characters from the input buffer.
@item
It treats the input buffer as a read-only region (and so must your code).
@end itemize

@noindent
When the text interpreter is in compile state, its behaviour changes in
these ways:

@itemize @bullet
@item
If a parsed string is found in the dictionary, the text interpreter will
perform the word's @dfn{compilation semantics}. In most cases, this
simply means that the execution semantics of the word will be appended
to the current definition.
@item
When a number is encountered, it is compiled into the current definition
(as a literal) rather than being pushed onto a parameter stack.
@item
If an error occurs, @code{state} is modified to put the text interpreter
back into interpret state.
@item
Each time a line is entered from the keyboard, Gforth prints
``@code{ compiled}'' rather than `` @code{ok}''.
@end itemize

@cindex text interpreter - input sources
When the text interpreter is using an input device other than the
keyboard, its behaviour changes in these ways:

@itemize @bullet
@item
When the parse area is empty, the text interpreter attempts to refill
the input buffer from the input source. When the input source is
exhausted, the input source is set back to the user input device.
@item
It doesn't print out ``@code{ ok}'' or ``@code{ compiled}'' messages each
time the parse area is emptied.
@item
If an error occurs, the input source is set back to the user input
device.
@end itemize

@ref{Input Sources} describes this in more detail.

doc->in
doc-source

doc-tib
doc-#tib

@menu
* Input Sources::
* Number Conversion::
* Interpret/Compile states::
* Literals::
* Interpreter Directives::
@end menu

@node Input Sources, Number Conversion, The Text Interpreter, The Text Interpreter
@subsection Input Sources
@cindex input sources
@cindex text interpreter - input sources

By default, the text interpreter accepts input from the user input
device (the keyboard) when Forth starts up. The text interpreter can
process input from any of these sources:

@itemize @bullet
@item
The user input device -- the keyboard.
@item
A file, using the words described in @ref{Forth source files}.
@item
A block, using the words described in @ref{Blocks}.
@item
A text string, using @code{evaluate}.
@end itemize

A program can identify the current input device from the values of
@code{source-id} and @code{blk}.

doc-source-id
doc-blk

doc-save-input
doc-restore-input

doc-evaluate


@node Number Conversion, Interpret/Compile states, Input Sources, The Text Interpreter
@subsection Number Conversion
@cindex number conversion
@cindex double-cell numbers, input format
@cindex input format for double-cell numbers
@cindex single-cell numbers, input format
@cindex input format for single-cell numbers
@cindex floating-point numbers, input format
@cindex input format for floating-point numbers

This section describes the rules that the text interpreter uses when it
tries to convert a string into a number.

Let <digit> represent any character that is a legal digit in the current
number base@footnote{For example, 0-9 when the number base is decimal or
0-9, A-F when the number base is hexadecimal.}.

Let <decimal digit> represent any character in the range 0-9.

Let @{@i{a b}@} represent the @i{optional} presence of any of the characters
in the braces (@i{a} or @i{b} or neither).

Let * represent any number of instances of the previous character
(including none).

Let any other character represent itself.

@noindent
Now, the conversion rules are:

@itemize @bullet
@item
A string of the form <digit><digit>* is treated as a single-precision
(cell-sized) positive integer. Examples are 0 123 6784532 32343212343456 42
@item
A string of the form -<digit><digit>* is treated as a single-precision
(cell-sized) negative integer, and is represented using 2's-complement
arithmetic. Examples are -45 -5681 -0
@item
A string of the form <digit><digit>*.<digit>* is treated as a double-precision
(double-cell-sized) positive integer. Examples are 3465. 3.465 34.65
(all three of these represent the same number).
@item
A string of the form -<digit><digit>*.<digit>* is treated as a
double-precision (double-cell-sized) negative integer, and is
represented using 2's-complement arithmetic. Examples are -3465. -3.465
-34.65 (all three of these represent the same number).
@item
A string of the form @{+ -@}<decimal digit>@{.@}<decimal digit>*@{e
E@}@{+ -@}<decimal digit><decimal digit>* is treated as a floating-point
number. Examples are 1e 1e0 1.e 1.e0 +1e+0 (which all represent the same
number) +12.E-4
@end itemize

By default, the number base used for integer number conversion is given
by the contents of the variable @code{base}.  Note that a lot of
confusion can result from unexpected values of @code{base}.  If you
change @code{base} anywhere, make sure to save the old value and restore
it afterwards.  In general I recommend keeping @code{base} decimal, and
using the prefixes described below for the popular non-decimal bases.

doc-dpl
doc-base
doc-hex
doc-decimal

@cindex '-prefix for character strings
@cindex &-prefix for decimal numbers
@cindex %-prefix for binary numbers
@cindex $-prefix for hexadecimal numbers
Gforth allows you to override the value of @code{base} by using a
prefix@footnote{Some Forth implementations provide a similar scheme by
implementing @code{$} etc. as parsing words that process the subsequent
number in the input stream and push it onto the stack. For example, see
@cite{Number Conversion and Literals}, by Wil Baden; Forth Dimensions
20(3) pages 26--27. In such implementations, unlike in Gforth, a space
is required between the prefix and the number.} before the first digit
of an (integer) number. Four prefixes are supported:

@itemize @bullet
@item
@code{&} -- decimal
@item
@code{%} -- binary
@item
@code{$} -- hexadecimal
@item
@code{'} -- base @code{max-char+1}
@end itemize

Here are some examples, with the equivalent decimal number shown after
in braces:

-$41 (-65), %1001101 (205), %1001.0001 (145 - a double-precision number),
'AB (16706; ascii A is 65, ascii B is 66, number is 65*256 + 66),
'ab (24930; ascii a is 97, ascii B is 98, number is 97*256 + 98),
&905 (905), $abc (2478), $ABC (2478).

@cindex number conversion - traps for the unwary
@noindent
Number conversion has a number of traps for the unwary:

@itemize @bullet
@item
You cannot determine the current number base using the code sequence
@code{base @@ .} -- the number base is always 10 in the current number
base. Instead, use something like @code{base @@ dec.}
@item
If the number base is set to a value greater than 14 (for example,
hexadecimal), the number 123E4 is ambiguous; the conversion rules allow
it to be intepreted as either a single-precision integer or a
floating-point number (Gforth treats it as an integer). The ambiguity
can be resolved by explicitly stating the sign of the mantissa and/or
exponent: 123E+4 or +123E4 -- if the number base is decimal, no
ambiguity arises; either representation will be treated as a
floating-point number.
@item
There is a word @code{bin} but it does @i{not} set the number base!
It is used to specify file types.
@item
ANS Forth requires the @code{.} of a double-precision number to
be the final character in the string. Allowing the @code{.} to be
anywhere after the first digit is a Gforth extension.
@item
The number conversion process does not check for overflow.
@item
In Gforth, number conversion to floating-point numbers always use base
10, irrespective of the value of @code{base}. In ANS Forth,
conversion to floating-point numbers whilst the value of
@code{base} is not 10 is an ambiguous condition.
@end itemize

@ref{Input} describes words that you can use to read numbers into your
programs.

@node Interpret/Compile states, Literals, Number Conversion, The Text Interpreter
@subsection Interpret/Compile states
@cindex Interpret/Compile states

A standard program is not permitted to change @code{state}
explicitly. However, it can change @code{state} implicitly, using the
words @code{[} and @code{]}. When @code{[} is executed it switches
@code{state} to interpret state, and therefore the text interpreter
starts interpreting. When @code{]} is executed it switches @code{state}
to compile state and therefore the text interpreter starts
compiling. The most common usage for these words is to compile literals,
as shown in @ref{Literals}. However, they give you the freedom to switch
modes at will.

@c This is a bad example: It's non-standard, and it's not necessary.
@c However, I can't think of a good example for switching into compile
@c state when there is no current word (@code{state}-smart words are not a
@c good reason).  So maybe we should use an example for switching into
@c interpret @code{state} in a colon def. - anton

Here is an example of building a jump-table of execution
tokens:

@example
: AA ." this is A" ;
: BB ." this is B" ;
: CC ." this is C" ;

create table ' aa COMPILE, ' bb COMPILE, ' cc COMPILE,
: go ( n -- ) \ n is offset into table.. 0 for 1st entry
  cells table + @ execute ;
@end example

@noindent
Now @code{0 go} will display ``@code{this is A}''. The table can be
built far more neatly@footnote{The source code is neater.. what is
compiled in memory in each case is identical.} like this:

@example
create table ] aa bb cc [
@end example

The problem with this code is that it is not portable; it will only work
on systems where code space and data space co-incide. The reason is that
both tables @i{compile} execution tokens -- into code space. The
Standard only allows data space to be assigned for a @code{CREATE}d
word. In addition, the Standard only allows @code{@@} to access data
space, whilst this example is using it to access code space. The only
portable, Standard way to build this table is to build it in data space,
like this:

@example
create table ' aa , ' bb , ' cc ,
@end example

@noindent
A similar technique can be used to build a table of constants:

@example
create primes 1 , 3 , 5 , 7 , 11 ,
@end example

doc-state
doc-[
doc-]

@node Literals, Interpreter Directives, Interpret/Compile states, The Text Interpreter
@subsection Literals
@cindex Literals

Often, you want to use a number within a colon definition. When you do
this, the text interpreter automatically compiles the number as a
@i{literal}. A literal is a number whose run-time effect is to be pushed
onto the stack.  If you had to do some maths to generate the number, you
might write it like this:

@example
: HOUR-TO-SEC ( n1 -- n2 )
  60 *      \ to minutes
  60 * ;    \ to seconds
@end example

It is very clear what this definition is doing, but it's inefficient
since it is performing 2 multiples at run-time. An alternative would be
to write:

@example
: HOUR-TO-SEC ( n1 -- n2 )
  3600 * ;  \ to seconds
@end example

Which does the same thing, and has the advantage of using a single
multiply. Ideally, we'd like the efficiency of the second with the
readability of the first.

@code{Literal} allows us to achieve that. It takes a number from the
stack and lays it down in the current definition just as though the
number had been typed directly into the definition. Our first attempt
might look like this:

@example
60          \ mins per hour
60 *        \ seconds per minute
: HOUR-TO-SEC ( n1 -- n2 )
  Literal * ;  \ to seconds
@end example

But this produces the error message @code{unstructured}. What happened?
The stack notation for @code{:} is (@i{ -- colon-sys}) and the size of
@i{colon-sys} is implementation-defined. In other words, once we start a
colon definition we can't portably access anything that was on the stack
before the definition began@footnote{@cite{Two Problems in ANS Forth},
by Thomas Worthington; Forth Dimensions 20(2) pages 32--34 describes
some situations where you might want to access stack items above
colon-sys, and provides a solution to the problem.}. The correct way of
solving this problem in this instance is to use @code{[ ]} like this:

@example
: HOUR-TO-SEC ( n1 -- n2 )
  [ 60          \ minutes per hour
    60 * ]      \ seconds per minute
  LITERAL * ;   \ to seconds
@end example

doc-literal
doc-]L
doc-2literal
doc-fliteral

@node Interpreter Directives, , Literals, The Text Interpreter
@subsection Interpreter Directives
@cindex interpreter directives

These words are usually used in interpret state; typically to control
which parts of a source file are processed by the text
interpreter. There are only a few ANS Forth Standard words, but Gforth
supplements these with a rich set of immediate control structure words
to compensate for the fact that the non-immediate versions can only be
used in compile state (@pxref{Control Structures}). Typical usages:

@example
FALSE Constant ASSEMBLER
.
.
ASSEMBLER [IF]
: ASSEMBLER-FEATURE
  ...
;
[ENDIF]
.
.
: SEE
  ... \ general-purpose SEE code
  [ ASSEMBLER [IF] ]
  ... \ assembler-specific SEE code
  [ [ENDIF] ]
;
@end example

doc-[IF]
doc-[ELSE]
doc-[THEN]
doc-[ENDIF]

doc-[IFDEF]
doc-[IFUNDEF]

doc-[?DO]
doc-[DO]
doc-[FOR]
doc-[LOOP]
doc-[+LOOP]
doc-[NEXT]

doc-[BEGIN]
doc-[UNTIL]
doc-[AGAIN]
doc-[WHILE]
doc-[REPEAT]



@c -------------------------------------------------------------
@node Tokens for Words, Word Lists, The Text Interpreter, Words
@section Tokens for Words
@cindex tokens for words

This section describes the creation and use of tokens that represent
words.

Named words have information stored in their header space entries to
indicate any non-default semantics (@pxref{Interpretation and
Compilation Semantics}). The semantics can be modified, using
@code{immediate} and/or @code{compile-only}, at the time that the words
are defined. Unnamed words have (by definition) no header space
entry, and therefore must have default semantics.

Named words have interpretation and compilation semantics. Unnamed words
just have execution semantics.

@cindex xt
@cindex execution token
The execution semantics of an unnamed word are represented by an
@dfn{execution token} (@i{xt}). As explained in @ref{Supplying names},
the execution token of the last word defined can be produced with
@code{lastxt}.

The interpretation semantics of a named word are also represented by an
execution token. You can produce the execution token using @code{'} or
@code{[']}. A simple example shows the difference between the two:

@example
: greet ( -- )   ." Hello" ;
: foo ( -- xt )  ['] greet execute ; \ ['] parses greet at compile-time
: bar ( -- )     ' execute ; \  '  parses at run-time

\ the next four lines all do the same thing
foo
bar greet
greet
' greet EXECUTE
@end example

An execution token occupies one cell.
@cindex code field address
@cindex CFA
In Gforth, the abstract data type @i{execution token} is implemented
as a code field address (CFA).
@comment TODO note that the standard does not say what it represents..
@comment and you cannot necessarily compile it in all Forths (eg native
@comment compilers?).

For literals, use @code{'} in interpreted code and @code{[']} in
compiled code. Gforth's @code{'} and @code{[']} behave somewhat
unusually by complaining about compile-only words. To get the execution
token for a compile-only word @i{name}, use @code{COMP' @i{name} DROP}
or @code{[COMP'] @i{name} DROP}.

@cindex compilation token
The compilation semantics of a named word are represented by a
@dfn{compilation token} consisting of two cells: @i{w xt}. The top cell
@i{xt} is an execution token. The compilation semantics represented by
the compilation token can be performed with @code{execute}, which
consumes the whole compilation token, with an additional stack effect
determined by the represented compilation semantics.

At present, the @i{w} part of a compilation token is an execution token,
and the @i{xt} part represents either @code{execute} or
@code{compile,}@footnote{Depending upon the compilation semantics of the
word. If the word has default compilation semantics, the @i{xt} will
represent @code{compile,}. Otherwise (e.g., for immediate words), the
@i{xt} will represent @code{execute}.}. However, don't rely on that
knowledge, unless necessary; future versions of Gforth may introduce
unusual compilation tokens (e.g., a compilation token that represents
the compilation semantics of a literal).

You can compile the compilation semantics with @code{postpone,}. I.e.,
@code{COMP' @i{word} postpone,} is equivalent to @code{postpone
@i{word}}.

@cindex name token
@cindex name field address
@cindex NFA
Named words are also represented by the @dfn{name token}, (@i{nt}). In
Gforth, the abstract data type @emph{name token} is implemented as a
name field address (NFA).

doc-execute
doc-compile,
doc-[']
doc-'
doc-[comp']
doc-comp'
doc-postpone,

doc-find-name
doc-name>int
doc-name?int
doc-name>comp
doc-name>string

@c -------------------------------------------------------------
@node Word Lists, Environmental Queries, Tokens for Words, Words
@section Word Lists
@cindex word lists
@cindex header space

A wordlist is a list of named words; you can add new words and look up
words by name (and you can remove words in a restricted way with
markers).  Every named (and @code{reveal}ed) word is in one wordlist.

@cindex search order stack
The text interpreter searches the wordlists present in the search order
(a stack of wordlists), from the top to the bottom.  Within each
wordlist, the search starts conceptually at the newest word; i.e., if
two words in a wordlist have the same name, the newer word is found.

@cindex compilation word list
New words are added to the @dfn{compilation wordlist} (aka current
wordlist).

@cindex wid
A word list is identified by a cell-sized word list identifier (@i{wid})
in much the same way as a file is identified by a file handle. The
numerical value of the wid has no (portable) meaning, and might change
from session to session.

The ANS Forth ``Search order'' word set is intended to provide a set of
low-level tools that allow various different schemes to be
implemented. Gforth provides @code{vocabulary}, a traditional Forth
word.  @file{compat/vocabulary.fs} provides an implementation in ANS
Standard Forth.

@comment TODO: locals section refers to here, saying that every word list (aka
@comment vocabulary) has its own methods for searching etc. Need to document that.

@comment the thisone- prefix is used to pick out the true definition of a
@comment word from the source files, rather than some alias.
doc-forth-wordlist
doc-definitions
doc-get-current
doc-set-current
doc-get-order
doc---thisone-set-order
doc-wordlist
doc-table
doc-push-order
doc-previous
doc-also
doc---thisone-forth
doc-only
doc---thisone-order

doc-find
doc-search-wordlist

doc-words
doc-vlist

doc-mappedwordlist
doc-root
doc-vocabulary
doc-seal
doc-vocs
doc-current
doc-context

@menu
* Why use word lists?::
* Word list examples::
@end menu

@node Why use word lists?, Word list examples, Word Lists, Word Lists
@subsection Why use word lists?
@cindex word lists - why use them?

Here are some reasons for using multiple word lists:

@itemize @bullet
@item
To improve compilation speed by reducing the number of header space
entries that must be searched. This is achieved by creating a new
word list that contains all of the definitions that are used in the
definition of a Forth system but which would not usually be used by
programs running on that system. That word list would be on the search
list when the Forth system was compiled but would be removed from the
search list for normal operation. This can be a useful technique for
low-performance systems (for example, 8-bit processors in embedded
systems) but is unlikely to be necessary in high-performance desktop
systems.
@item
To prevent a set of words from being used outside the context in which
they are valid. Two classic examples of this are an integrated editor
(all of the edit commands are defined in a separate word list; the
search order is set to the editor word list when the editor is invoked;
the old search order is restored when the editor is terminated) and an
integrated assembler (the op-codes for the machine are defined in a
separate word list which is used when a @code{CODE} word is defined).
@item
To prevent a name-space clash between multiple definitions with the same
name. For example, when building a cross-compiler you might have a word
@code{IF} that generates conditional code for your target system. By
placing this definition in a different word list you can control whether
the host system's @code{IF} or the target system's @code{IF} get used in
any particular context by controlling the order of the word lists on the
search order stack.
@end itemize

@node Word list examples, ,Why use word lists?, Word Lists
@subsection Word list examples
@cindex word lists - examples

Here is an example of creating and using a new wordlist using ANS
Forth Standard words:

@example
wordlist constant my-new-words-wordlist
: my-new-words get-order nip my-new-words-wordlist swap set-order ;

\ add it to the search order
also my-new-words

\ alternatively, add it to the search order and make it
\ the compilation word list
also my-new-words definitions
\ type "order" to see the problem
@end example

The problem with this example is that @code{order} has no way to
associate the name @code{my-new-words} with the wid of the word list (in
Gforth, @code{order} and @code{vocs} will display @code{???}  for a wid
that has no associated name). There is no Standard way of associating a
name with a wid.

In Gforth, this example can be re-coded using @code{vocabulary}, which
associates a name with a wid:

@example
vocabulary my-new-words

\ add it to the search order
my-new-words

\ alternatively, add it to the search order and make it
\ the compilation word list
my-new-words definitions
\ type "order" to see that the problem is solved
@end example

@c -------------------------------------------------------------
@node Environmental Queries, Files, Word Lists, Words
@section Environmental Queries
@cindex environmental queries

ANS Forth introduced the idea of ``environmental queries'' as a way
for a program running on a system to determine certain characteristics of the system.
The Standard specifies a number of strings that might be recognised by a system.

The Standard requires that the header space used for environmental queries
be distinct from the header space used for definitions.

Typically, environmental queries are supported by creating a set of
definitions in a word list that is @i{only} used during environmental
queries; that is what Gforth does. There is no Standard way of adding
definitions to the set of recognised environmental queries, but any
implementation that supports the loading of optional word sets must have
some mechanism for doing this (after loading the word set, the
associated environmental query string must return @code{true}). In
Gforth, the word list used to honour environmental queries can be
manipulated just like any other word list.

doc-environment?
doc-environment-wordlist

doc-gforth
doc-os-class

Note that, whilst the documentation for (e.g.) @code{gforth} shows it
returning two items on the stack, querying it using @code{environment?}
will return an additional item; the @code{true} flag that shows that the
string was recognised.

@comment TODO Document the standard strings or note where they are documented herein

Here are some examples of using environmental queries:

@example
s" address-unit-bits" environment? 0=
[IF]
     cr .( environmental attribute address-units-bits unknown... ) cr
[THEN]

s" block" environment? [IF] DROP include block.fs [THEN]

s" gforth" environment? [IF] 2DROP include compat/vocabulary.fs [THEN]

s" gforth" environment? [IF] .( Gforth version ) TYPE
                        [ELSE] .( Not Gforth..) [THEN]
@end example


Here is an example of adding a definition to the environment word list:

@example
get-current environment-wordlist set-current
true constant block
true constant block-ext
set-current
@end example

You can see what definitions are in the environment word list like this:

@example
get-order 1+ environment-wordlist swap set-order words previous
@end example


@c -------------------------------------------------------------
@node Files, Blocks, Environmental Queries, Words
@section Files
@cindex files
@cindex I/O - file-handling

Gforth provides facilities for accessing files that are stored in the
host operating system's file-system. Files that are processed by Gforth
can be divided into two categories:

@itemize @bullet
@item
Files that are processed by the Text Interpreter (@dfn{Forth source files}).
@item
Files that are processed by some other program (@dfn{general files}).
@end itemize

@menu
* Forth source files::
* General files::         
* Search Paths::                 
* Forth Search Paths::    
* General Search Paths::        
@end menu


@c -------------------------------------------------------------
@node Forth source files, General files, Files, Files
@subsection Forth source files
@cindex including files
@cindex Forth source files

The simplest way to interpret the contents of a file is to use one of
these two formats:

@example
include mysource.fs
s" mysource.fs" included
@end example

Sometimes you want to include a file only if it is not included already
(by, say, another source file). In that case, you can use one of these
fomats:

@example
require mysource.fs
needs mysource.fs
s" mysource.fs" required
@end example

@cindex stack effect of included files
@cindex including files, stack effect
I recommend that you write your source files such that interpreting them
does not change the stack. This allows using these files with
@code{required} and friends without complications. For example:

@example
1 require foo.fs drop
@end example

doc-include-file
doc-included
doc-included?
doc-include
doc-required
doc-require
doc-needs
doc-init-included-files

A definition in ANS Forth for @code{required} is provided in
@file{compat/required.fs}.

@c -------------------------------------------------------------
@node General files, Search Paths, Forth source files, Files
@subsection General files
@cindex general files
@cindex file-handling

Files are opened/created by name and type. The following types are
recognised:

doc-r/o
doc-r/w
doc-w/o
doc-bin

When a file is opened/created, it returns a file identifier,
@i{wfileid} that is used for all other file commands. All file
commands also return a status value, @i{wior}, that is 0 for a
successful operation and an implementation-defined non-zero value in the
case of an error.

doc-open-file
doc-create-file

doc-close-file
doc-delete-file
doc-rename-file
doc-read-file
doc-read-line
doc-write-file
doc-write-line
doc-emit-file
doc-flush-file

doc-file-status
doc-file-position
doc-reposition-file
doc-file-size
doc-resize-file

@c ---------------------------------------------------------
@node Search Paths, Forth Search Paths, General files, Files
@subsection Search Paths
@cindex path for @code{included}
@cindex file search path
@cindex @code{include} search path
@cindex search path for files

If you specify an absolute filename (i.e., a filename starting with
@file{/} or @file{~}, or with @file{:} in the second position (as in
@samp{C:...})) for @code{included} and friends, that file is included
just as you would expect.

For relative filenames, Gforth uses a search path similar to Forth's
search order (@pxref{Word Lists}). It tries to find the given filename
in the directories present in the path, and includes the first one it
finds. There are separate search paths for Forth source files and
general files.

If the search path contains the directory @file{.} (as it should), this
refers to the directory that the present file was @code{included}
from. This allows files to include other files relative to their own
position (irrespective of the current working directory or the absolute
position).  This feature is essential for libraries consisting of
several files, where a file may include other files from the library.
It corresponds to @code{#include "..."} in C. If the current input
source is not a file, @file{.} refers to the directory of the innermost
file being included, or, if there is no file being included, to the
current working directory.

Use @file{~+} to refer to the current working directory (as in the
@code{bash}).

If the filename starts with @file{./}, the search path is not searched
(just as with absolute filenames), and the @file{.} has the same meaning
as described above.

@c ---------------------------------------------------------
@node Forth Search Paths, General Search Paths, Search Paths, Files
@subsubsection Forth Search Paths
@cindex search path control - Forth

The search path is initialized when you start Gforth (@pxref{Invoking
Gforth}). You can display it and change it using these words:

doc-.fpath
doc-fpath+
doc-fpath=
doc-open-fpath-file

Here is an example of using @code{fpath} and @code{require}:

@example
fpath= /usr/lib/forth/|./
require timer.fs
@end example

@c ---------------------------------------------------------
@node General Search Paths,  , Forth Search Paths, Files
@subsubsection General Search Paths
@cindex search path control - for user applications

Your application may need to search files in several directories, like
@code{included} does. To facilitate this, Gforth allows you to define
and use your own search paths, by providing generic equivalents of the
Forth search path words:

doc-.path
doc-path+
doc-path=
doc-open-path-file

Here's an example of creating a search path:

@example
\ Make a buffer for the path:
create mypath   100 chars ,     \ maximum length (is checked)
                0 ,             \ real len
                100 chars allot \ space for path
@end example

@c -------------------------------------------------------------
@node Blocks, Other I/O, Files, Words
@section Blocks
@cindex I/O - blocks
@cindex blocks

When you run Gforth on a modern desk-top computer, it runs under the
control of an operating system which provides certain services.  One of
these services is @var{file services}, which allows Forth source code
and data to be stored in files and read into Gforth (@pxref{Files}).

Traditionally, Forth has been an important programming language on
systems where it has interfaced directly to the underlying hardware with
no intervening operating system. Forth provides a mechanism, called
@dfn{blocks}, for accessing mass storage on such systems.

A block is a 1024-byte data area, which can be used to hold data or
Forth source code. No structure is imposed on the contents of the
block. A block is identified by its number; blocks are numbered
contiguously from 1 to an implementation-defined maximum.

A typical system that used blocks but no operating system might use a
single floppy-disk drive for mass storage, with the disks formatted to
provide 256-byte sectors. Blocks would be implemented by assigning the
first four sectors of the disk to block 1, the second four sectors to
block 2 and so on, up to the limit of the capacity of the disk. The disk
would not contain any file system information, just the set of blocks.

@cindex blocks file
On systems that do provide file services, blocks are typically
implemented by storing a sequence of blocks within a single @dfn{blocks
file}.  The size of the blocks file will be an exact multiple of 1024
bytes, corresponding to the number of blocks it contains. This is the
mechanism that Gforth uses.

@cindex @file{blocks.fb}
Only 1 blocks file can be open at a time. If you use block words without
having specified a blocks file, Gforth defaults to the blocks file
@file{blocks.fb}. Gforth uses the Forth search path when attempting to
locate a blocks file (@pxref{Forth Search Paths}).

@cindex block buffers
When you read and write blocks under program control, Gforth uses a
number of @dfn{block buffers} as intermediate storage. These buffers are
not used when you use @code{load} to interpret the contents of a block.

The behaviour of the block buffers is directly analagous to that of a
cache. Each block buffer has three states:

@itemize @bullet
@item
Unassigned
@item
Assigned-clean
@item
Assigned-dirty
@end itemize

Initially, all block buffers are @i{unassigned}. In order to access a
block, the block (specified by its block number) must be assigned to a
block buffer.

The assignment of a block to a block buffer is performed by @code{block}
or @code{buffer}. Use @code{block} when you wish to modify the existing
contents of a block. Use @code{buffer} when you don't care about the
existing contents of the block@footnote{The ANS Forth definition of
@code{buffer} is intended not to cause disk I/O; if the data associated
with the particular block is already stored in a block buffer due to an
earlier @code{block} command, @code{buffer} will return that block
buffer and the existing contents of the block will be
available. Otherwise, @code{buffer} will simply assign a new, empty
block buffer for the block.}.

Once a block has been assigned to a block buffer, the block buffer state
becomes @i{assigned-clean}. Data can now be manipulated within the
block buffer.

When the contents of a block buffer is changed it is necessary,
@i{before calling} @code{block} @i{or} @code{buffer} @i{again}, to
either abandon the changes (by doing nothing) or commit the changes,
using @code{update}. Using @code{update} does not change the blocks
file; it simply changes a block buffer's state to @i{assigned-dirty}.

The word @code{flush} causes all @i{assigned-dirty} blocks to be
written back to the blocks file on disk. Leaving Gforth using @code{bye}
also causes a @code{flush} to be performed.

In Gforth, @code{block} and @code{buffer} use a @i{direct-mapped}
algorithm to assign a block buffer to a block. That means that any
particular block can only be assigned to one specific block buffer,
called (for the particular operation) the @i{victim buffer}. If the
victim buffer is @i{unassigned} or @i{assigned-clean} it can be
allocated to the new block immediately. If it is @i{assigned-dirty}
its current contents must be written out to disk before it can be
allocated to the new block.

Although no structure is imposed on the contents of a block, it is
traditional to display the contents as 16 lines each of 64 characters.  A
block provides a single, continuous stream of input (for example, it
acts as a single parse area) -- there are no end-of-line characters
within a block, and no end-of-file character at the end of a
block. There are two consequences of this:

@itemize @bullet
@item
The last character of one line wraps straight into the first character
of the following line
@item
The word @code{\} -- comment to end of line -- requires special
treatment; in the context of a block it causes all characters until the
end of the current 64-character ``line'' to be ignored.
@end itemize

In Gforth, when you use @code{block} with a non-existent block number,
the current block file will be extended to the appropriate size and the
block buffer will be initialised with spaces.

Gforth doesn't encourage the use of blocks; the mechanism is only
provided for backward compatibility -- ANS Forth requires blocks to be
available when files are.

Common techniques that are used when working with blocks include:

@itemize @bullet
@item
A screen editor that allows you to edit blocks without leaving the Forth
environment.
@item
Shadow screens; where every code block has an associated block
containing comments (for example: code in odd block numbers, comments in
even block numbers). Typically, the block editor provides a convenient
mechanism to toggle between code and comments.
@item
Load blocks; a single block (typically block 1) contains a number of
@code{thru} commands which @code{load} the whole of the application.
@end itemize

See Frank Sergeant's Pygmy Forth to see just how well blocks can be
integrated into a Forth programming environment.

@comment TODO what about errors on open-blocks?
doc-open-blocks
doc-use
doc-get-block-fid
doc-block-position

doc-scr
doc-list

doc---block-block
doc-buffer

doc-update
doc-updated?
doc-save-buffers
doc-empty-buffers
doc-empty-buffer
doc-flush

doc-load
doc-thru
doc-+load
doc-+thru
xdoc--gforth--->
doc-block-included

@c -------------------------------------------------------------
@node Other I/O, Programming Tools, Blocks, Words
@section Other I/O
@cindex I/O - keyboard and display

@menu
* Simple numeric output::       Predefined formats
* Formatted numeric output::    Formatted (pictured) output
* String Formats::              How Forth stores strings in memory
* Displaying characters and strings:: Other stuff
* Input::                       Input
@end menu

@node Simple numeric output, Formatted numeric output, Other I/O, Other I/O
@subsection Simple numeric output
@cindex numeric output - simple/free-format

The simplest output functions are those that display numbers from the
data or floating-point stacks. Floating-point output is always displayed
using base 10. Numbers displayed from the data stack use the value stored
in @code{base}.

doc-.
doc-dec.
doc-hex.
doc-u.
doc-.r
doc-u.r
doc-d.
doc-ud.
doc-d.r
doc-ud.r
doc-f.
doc-fe.
doc-fs.

Examples of printing the number 1234.5678E23 in the different floating-point output
formats are shown below:

@example
f. 123456779999999000000000000.
fe. 123.456779999999E24
fs. 1.23456779999999E26
@end example


@node Formatted numeric output, String Formats, Simple numeric output, Other I/O
@subsection Formatted numeric output
@cindex formatted numeric output
@cindex pictured numeric output
@cindex numeric output - formatted

Forth traditionally uses a technique called @dfn{pictured numeric
output} for formatted printing of integers.  In this technique, digits
are extracted from the number (using the current output radix defined by
@code{base}), converted to ASCII codes and appended to a string that is
built in a scratch-pad area of memory (@pxref{core-idef,
Implementation-defined options, Implementation-defined
options}). Arbitrary characters can be appended to the string during the
extraction process. The completed string is specified by an address
and length and can be manipulated (@code{TYPE}ed, copied, modified)
under program control.

All of the words described in the previous section for simple numeric
output are implemented in Gforth using pictured numeric output.

Three important things to remember about Pictured Numeric Output:

@itemize @bullet
@item
It always operates on double-precision numbers; to display a
single-precision number, convert it first (@pxref{Double precision} for
ways of doing this).
@item
It always treats the double-precision number as though it were
unsigned. The examples below show ways of printing signed numbers.
@item
The string is built up from right to left; least significant digit first.
@end itemize

doc-<#
doc-#
doc-#s
doc-hold
doc-sign
doc-#>

doc-represent

Here are some examples of using pictured numeric output:

@example
: my-u. ( u -- )
  \ Simplest use of pns.. behaves like Standard u. 
  0              \ convert to unsigned double
  <#             \ start conversion
  #s             \ convert all digits
  #>             \ complete conversion
  TYPE SPACE ;   \ display, with trailing space

: cents-only ( u -- )
  0              \ convert to unsigned double
  <#             \ start conversion
  # #            \ convert two least-significant digits
  #>             \ complete conversion, discard other digits
  TYPE SPACE ;   \ display, with trailing space

: dollars-and-cents ( u -- )
  0              \ convert to unsigned double
  <#             \ start conversion
  # #            \ convert two least-significant digits
  [char] . hold  \ insert decimal point
  #s             \ convert remaining digits
  [char] $ hold  \ append currency symbol
  #>             \ complete conversion
  TYPE SPACE ;   \ display, with trailing space

: my-. ( n -- )
  \ handling negatives.. behaves like Standard .
  s>d            \ convert to signed double
  swap over dabs \ leave sign byte followed by unsigned double
  <#             \ start conversion
  #s             \ convert all digits
  rot sign       \ get at sign byte, append "-" if needed
  #>             \ complete conversion
  TYPE SPACE ;   \ display, with trailing space

: account. ( n -- )
  \ accountants don't like minus signs, they use braces
  \ for negative numbers
  s>d            \ convert to signed double
  swap over dabs \ leave sign byte followed by unsigned double
  <#             \ start conversion
  2 pick         \ get copy of sign byte
  0< IF [char] ) hold THEN \ right-most character of output
  #s             \ convert all digits
  rot            \ get at sign byte
  0< IF [char] ( hold THEN
  #>             \ complete conversion
  TYPE SPACE ;   \ display, with trailing space
@end example

Here are some examples of using these words:

@example
1 my-u. 1
hex -1 my-u. decimal FFFFFFFF
1 cents-only 01
1234 cents-only 34
2 dollars-and-cents $0.02
1234 dollars-and-cents $12.34
123 my-. 123
-123 my. -123
123 account. 123
-456 account. (456)
@end example


@node String Formats, Displaying characters and strings, Formatted numeric output, Other I/O
@subsection String Formats
@cindex strings - see character strings
@cindex character strings - formats
@cindex I/O - see character strings

Forth commonly uses two different methods for representing character
strings:

@itemize @bullet
@item
@cindex address of counted string
As a @dfn{counted string}, represented by a @i{c-addr}. The char
addressed by @i{c-addr} contains a character-count, @i{n}, of the
string and the string occupies the subsequent @i{n} char addresses in
memory.
@item
As cell pair on the stack; @i{c-addr u}, where @i{u} is the length
of the string in characters, and @i{c-addr} is the address of the
first byte of the string.
@end itemize

ANS Forth encourages the use of the second format when representing
strings on the stack, whilst conceeding that the counted string format
remains useful as a way of storing strings in memory.

doc-count

@xref{Memory Blocks} for words that move, copy and search
for strings. @xref{Displaying characters and strings,} for words that
display characters and strings.


@node Displaying characters and strings, Input, String Formats, Other I/O
@subsection Displaying characters and strings
@cindex characters - compiling and displaying
@cindex character strings - compiling and displaying

This section starts with a glossary of Forth words and ends with a set
of examples.

doc-bl
doc-space
doc-spaces
doc-emit
doc-toupper
doc-."
doc-.(
doc-type
doc-cr
@cindex cursor control
doc-at-xy
doc-page
doc-s"
doc-c"
doc-char
doc-[char]
doc-sliteral

As an example, consider the following text, stored in a file @file{test.fs}:

@example
.( text-1)
: my-word
  ." text-2" cr
  .( text-3)
;

." text-4"

: my-char
  [char] ALPHABET emit
  char emit
;
@end example

When you load this code into Gforth, the following output is generated:

@example
@kbd{include test.fs @key{RET}} text-1text-3text-4 ok
@end example

@itemize @bullet
@item
Messages @code{text-1} and @code{text-3} are displayed because @code{.(} 
is an immediate word; it behaves in the same way whether it is used inside
or outside a colon definition.
@item
Message @code{text-4} is displayed because of Gforth's added interpretation
semantics for @code{."}.
@item
Message @code{text-2} is @i{not} displayed, because the text interpreter
performs the compilation semantics for @code{."} within the definition of
@code{my-word}.
@end itemize

Here are some examples of executing @code{my-word} and @code{my-char}:

@example
@kbd{my-word @key{RET}} text-2
 ok
@kbd{my-char fred @key{RET}} Af ok
@kbd{my-char jim @key{RET}} Aj ok
@end example

@itemize @bullet
@item
Message @code{text-2} is displayed because of the run-time behaviour of
@code{."}.
@item
@code{[char]} compiles the ``A'' from ``ALPHABET'' and puts its display code
on the stack at run-time. @code{emit} always displays the character
when @code{my-char} is executed.
@item
@code{char} parses a string at run-time and the second @code{emit} displays
the first character of the string.
@item
If you type @code{see my-char} you can see that @code{[char]} discarded
the text ``LPHABET'' and only compiled the display code for ``A'' into the
definition of @code{my-char}.
@end itemize



@node Input, , Displaying characters and strings, Other I/O
@subsection Input
@cindex input
@cindex I/O - see input
@cindex parsing a string

@xref{String Formats} for ways of storing character strings in memory.

@comment TODO examples for >number >float accept key key? pad parse word refill
@comment then index them

doc-key
doc-key?
doc->number
doc->float
doc-accept
doc-pad
doc-parse
doc-word
doc-sword
doc-refill
@comment obsolescent words..
doc-convert
doc-query
doc-expect
doc-span


@c -------------------------------------------------------------
@node Programming Tools, Assembler and Code Words, Other I/O, Words
@section Programming Tools
@cindex programming tools

@menu
* Debugging::                   Simple and quick.
* Assertions::                  Making your programs self-checking.
* Singlestep Debugger::		Executing your program word by word.
@end menu

@node Debugging, Assertions, Programming Tools, Programming Tools
@subsection Debugging
@cindex debugging

Languages with a slow edit/compile/link/test development loop tend to
require sophisticated tracing/stepping debuggers to facilate
productive debugging.

A much better (faster) way in fast-compiling 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 simple debugging aids provided in @file{debugs.fs}
are meant to support this style of debugging. In addition, there are
words for non-destructively inspecting the stack and memory:

doc-.s
doc-f.s

There is a word @code{.r} but it does @i{not} display the return
stack! It is used for formatted numeric output.

doc-depth
doc-fdepth
doc-clearstack
doc-?
doc-dump

The word @code{~~} prints debugging information (by default the source
location and the stack contents). It is easy to insert. If you use Emacs
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).

The default actions of @code{~~} 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

doc-see
doc-marker

Here's an example of using @code{marker} at the start of a source file
that you are debugging; it ensures that you only ever have one copy of
the file's definitions compiled at any time:

@example
[IFDEF] my-code
    my-code
[ENDIF]

marker my-code
init-included-files

\ .. definitions start here
\ .
\ .
\ end
@end example



@node Assertions, Singlestep Debugger, Debugging, Programming Tools
@subsection Assertions
@cindex assertions

It is a good idea to make your programs self-checking, especially if you
make an assumption that may become invalid during maintenance (for
example, that a certain field of a data structure is never zero). Gforth
supports @dfn{assertions} for this purpose. They are used like this:

@example
assert( @i{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-)

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.

The value of @code{assert-level} is evaluated at compile-time, not at
run-time. Therefore 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 different pieces of code at different
@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).

Definitions in ANS Forth for these assertion words are provided
in @file{compat/assert.fs}.


@node Singlestep Debugger, , Assertions, Programming Tools
@subsection Singlestep Debugger
@cindex singlestep Debugger
@cindex debugging Singlestep
@cindex @code{dbg}
@cindex @code{BREAK:}
@cindex @code{BREAK"}

When you create a new word there's often the need to check whether it
behaves correctly or not. You can do this by typing @code{dbg
badword}. A debug session might look like this:

@example
: badword 0 DO i . LOOP ;  ok
2 dbg badword 
: badword  
Scanning code...

Nesting debugger ready!

400D4738  8049BC4 0              -> [ 2 ] 00002 00000 
400D4740  8049F68 DO             -> [ 0 ] 
400D4744  804A0C8 i              -> [ 1 ] 00000 
400D4748 400C5E60 .              -> 0 [ 0 ] 
400D474C  8049D0C LOOP           -> [ 0 ] 
400D4744  804A0C8 i              -> [ 1 ] 00001 
400D4748 400C5E60 .              -> 1 [ 0 ] 
400D474C  8049D0C LOOP           -> [ 0 ] 
400D4758  804B384 ;              ->  ok
@end example

Each line displayed is one step. You always have to hit return to
execute the next word that is displayed. If you don't want to execute
the next word in a whole, you have to type @kbd{n} for @code{nest}. Here is
an overview what keys are available:

@table @i

@item @key{RET}
Next; Execute the next word.

@item n
Nest; Single step through next word.

@item u
Unnest; Stop debugging and execute rest of word. If we got to this word
with nest, continue debugging with the calling word.

@item d
Done; Stop debugging and execute rest.

@item s
Stop; Abort immediately.

@end table

Debugging large application with this mechanism is very difficult, because
you have to nest very deeply into the program before the interesting part
begins. This takes a lot of time. 

To do it more directly put a @code{BREAK:} command into your source code.
When program execution reaches @code{BREAK:} the single step debugger is
invoked and you have all the features described above.

If you have more than one part to debug it is useful to know where the
program has stopped at the moment. You can do this by the 
@code{BREAK" string"} command. This behaves like @code{BREAK:} except that
string is typed out when the ``breakpoint'' is reached.

doc-dbg
doc-BREAK:
doc-BREAK"


@c -------------------------------------------------------------
@node Assembler and Code Words, Threading Words, Programming Tools, Words
@section Assembler and Code Words
@cindex assembler
@cindex code words

Gforth provides some words for defining primitives (words written in
machine code), and for defining 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,}@footnote{This isn't portable,
because these words emit stuff in @i{data} space; it works because
Gforth has unified code/data spaces. Assembler isn't likely to be
portable anyway.}.

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.

The typical usage of these @code{code} words can be shown most easily by
analogy to the equivalent high-level defining words:

@example
: foo                                        code foo
   <high-level Forth words>                        <assembler>
;                                            end-code

: bar                                        : bar
   <high-level Forth words>                     <high-level Forth words>
   CREATE                                       CREATE
      <high-level Forth words>                     <high-level Forth words>
   DOES>                                        ;code
      <high-level Forth words>                     <assembler>
;                                            end-code
@end example

@code{flush-icache} is always present. The other words are rarely used
and reside in @code{code.fs}, which is usually not loaded. You can load
it with @code{require code.fs}.

@cindex registers of the inner interpreter
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.

@cindex code words, portable
Another option for implementing normal and defining words efficiently
is to add the desired functionality to the source of Gforth. For normal
words you just have to edit @file{primitives} (@pxref{Automatic
Generation}). Defining words (equivalent to @code{;CODE} words, for fast
defined words) may require changes in @file{engine.c}, @file{kernel.fs},
@file{prims2x.fs}, and possibly @file{cross.fs}.


@c -------------------------------------------------------------
@node Threading Words, Locals, Assembler and Code Words, Words
@section Threading Words
@cindex threading words

@cindex code address
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 incomplete. It is also pretty low-level;
some day it will hopefully be made unnecessary by an internals wordset
that abstracts implementation details away completely.

doc-threading-method
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:

You can recognize words defined by a @code{CREATE}...@code{DOES>} word
with @code{>does-code}. If the word was defined in that way, the value
returned is non-zero and identifies the @code{DOES>} used by the
defining word.
@comment TODO should that be ``identifies the xt of the DOES> ??''

@c -------------------------------------------------------------
@node Locals, Structures, Threading Words, Words
@section Locals
@cindex locals

Local variables can make Forth programming more enjoyable and Forth
programs easier to read. Unfortunately, the locals of ANS Forth are
laden with restrictions. Therefore, we provide not only the ANS Forth
locals wordset, but also our own, more powerful locals wordset (we
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
@*@url{http://www.complang.tuwien.ac.at/papers/ertl94l.ps.gz}.

@menu
* Gforth locals::               
* ANS Forth locals::            
@end menu

@node Gforth locals, ANS Forth locals, Locals, Locals
@subsection Gforth locals
@cindex Gforth locals
@cindex locals, Gforth style

Locals can be defined with

@example
@{ local1 local2 ... -- comment @}
@end example
or
@example
@{ local1 local2 ... @}
@end example

E.g.,
@example
: max @{ n1 n2 -- n3 @}
 n1 n2 > if
   n1
 else
   n2
 endif ;
@end example

The similarity of locals definitions with stack comments is intended. A
locals definition often replaces the stack comment of a word. The order
of the locals corresponds to the order in a stack comment and everything
after the @code{--} is really a comment.

This similarity has one disadvantage: It is too easy to confuse locals
declarations with stack comments, causing bugs and making them hard to
find. However, this problem can be avoided by appropriate coding
conventions: Do not use both notations in the same program. If you do,
they should be distinguished using additional means, e.g. by position.

@cindex types of locals
@cindex locals types
The name of the local may be preceded by a type specifier, e.g.,
@code{F:} for a floating point value:

@example
: CX* @{ F: Ar F: Ai F: Br F: Bi -- Cr Ci @}
\ complex multiplication
 Ar Br f* Ai Bi f* f-
 Ar Bi f* Ai Br f* f+ ;
@end example

@cindex flavours of locals
@cindex locals flavours
@cindex value-flavoured locals
@cindex variable-flavoured locals
Gforth currently supports cells (@code{W:}, @code{W^}), doubles
(@code{D:}, @code{D^}), floats (@code{F:}, @code{F^}) and characters
(@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{TO}. A variable-flavoured local (defined with @code{W^} etc.)
produces its address (which becomes invalid when the variable's scope is
left). E.g., the standard word @code{emit} can be defined in terms of
@code{type} like this:

@example
: emit @{ C^ char* -- @}
    char* 1 type ;
@end example

@cindex default type of locals
@cindex locals, default type
A local without type specifier is a @code{W:} local. Both flavours of
locals are initialized with values from the data or FP stack.

Currently there is no way to define locals with user-defined data
structures, but we are working on it.

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?
@cindex locals visibility
@cindex visibility of locals
@cindex scope of locals

Basically, the answer is that locals are visible where you would expect
it in block-structured languages, and sometimes a little longer. If you
want to restrict the scope of a local, enclose its definition in
@code{SCOPE}...@code{ENDSCOPE}.

doc-scope
doc-endscope

These words behave like control structure words, so you can use them
with @code{CS-PICK} and @code{CS-ROLL} to restrict the scope in
arbitrary ways.

If you want a more exact answer to the visibility question, here's the
basic principle: A local is visible in all places that can only be
reached through the definition of the local@footnote{In compiler
construction terminology, all places dominated by the definition of the
local.}. In other words, it is not visible in places that can be reached
without going through the definition of the local. E.g., locals defined
in @code{IF}...@code{ENDIF} are visible until the @code{ENDIF}, locals
defined in @code{BEGIN}...@code{UNTIL} are visible after the
@code{UNTIL} (until, e.g., a subsequent @code{ENDSCOPE}).

The reasoning behind this solution is: We want to have the locals
visible as long as it is meaningful. The user can always make the
visibility shorter by using explicit scoping. In a place that can
only be reached through the definition of a local, the meaning of a
local name is clear. In other places it is not: How is the local
initialized at the control flow path that does not contain the
definition? Which local is meant, if the same name is defined twice in
two independent control flow paths?

This should be enough detail for nearly all users, so you can skip the
rest of this section. If you really must know all the gory details and
options, read on.

In order to implement this rule, the compiler has to know which places
are unreachable. It knows this automatically after @code{AHEAD},
@code{AGAIN}, @code{EXIT} and @code{LEAVE}; in other cases (e.g., after
most @code{THROW}s), you can use the word @code{UNREACHABLE} to tell the
compiler that the control flow never reaches that place. If
@code{UNREACHABLE} is not used where it could, the only consequence is
that the visibility of some locals is more limited than the rule above
says. If @code{UNREACHABLE} is used where it should not (i.e., if you
lie to the compiler), buggy code will be produced.

doc-unreachable

Another problem with this rule is that at @code{BEGIN}, the compiler
does not know which locals will be visible on the incoming
back-edge. All problems discussed in the following are due to this
ignorance of the compiler (we discuss the problems using @code{BEGIN}
loops as examples; the discussion also applies to @code{?DO} and other
loops). Perhaps the most insidious example is:
@example
AHEAD
BEGIN
  x
[ 1 CS-ROLL ] THEN
  @{ x @}
  ...
UNTIL
@end example

This should be legal according to the visibility rule. The use of
@code{x} can only be reached through the definition; but that appears
textually below the use.

From this example it is clear that the visibility rules cannot be fully
implemented without major headaches. Our implementation treats common
cases as advertised and the exceptions are treated in a safe way: The
compiler makes a reasonable guess about the locals visible after a
@code{BEGIN}; if it is too pessimistic, the
user will get a spurious error about the local not being defined; if the
compiler is too optimistic, it will notice this later and issue a
warning. In the case above the compiler would complain about @code{x}
being undefined at its use. You can see from the obscure examples in
this section that it takes quite unusual control structures to get the
compiler into trouble, and even then it will often do fine.

If the @code{BEGIN} is reachable from above, the most optimistic guess
is that all locals visible before the @code{BEGIN} will also be
visible after the @code{BEGIN}. This guess is valid for all loops that
are entered only through the @code{BEGIN}, in particular, for normal
@code{BEGIN}...@code{WHILE}...@code{REPEAT} and
@code{BEGIN}...@code{UNTIL} loops and it is implemented in our
compiler. When the branch to the @code{BEGIN} is finally generated by
@code{AGAIN} or @code{UNTIL}, the compiler checks the guess and
warns the user if it was too optimistic:
@example
IF
  @{ x @}
BEGIN
  \ x ? 
[ 1 cs-roll ] THEN
  ...
UNTIL
@end example

Here, @code{x} lives only until the @code{BEGIN}, but the compiler
optimistically assumes that it lives until the @code{THEN}. It notices
this difference when it compiles the @code{UNTIL} and issues a
warning. The user can avoid the warning, and make sure that @code{x}
is not used in the wrong area by using explicit scoping:
@example
IF
  SCOPE
  @{ x @}
  ENDSCOPE
BEGIN
[ 1 cs-roll ] THEN
  ...
UNTIL
@end example

Since the guess is optimistic, there will be no spurious error messages
about undefined locals.

If the @code{BEGIN} is not reachable from above (e.g., after
@code{AHEAD} or @code{EXIT}), the compiler cannot even make an
optimistic guess, as the locals visible after the @code{BEGIN} may be
defined later. Therefore, the compiler assumes that no locals are
visible after the @code{BEGIN}. However, the user can use
@code{ASSUME-LIVE} to make the compiler assume that the same locals are
visible at the BEGIN as at the point where the top control-flow stack
item was created.

doc-assume-live

E.g.,
@example
@{ x @}
AHEAD
ASSUME-LIVE
BEGIN
  x
[ 1 CS-ROLL ] THEN
  ...
UNTIL
@end example

Other cases where the locals are defined before the @code{BEGIN} can be
handled by inserting an appropriate @code{CS-ROLL} before the
@code{ASSUME-LIVE} (and changing the control-flow stack manipulation
behind the @code{ASSUME-LIVE}).

Cases where locals are defined after the @code{BEGIN} (but should be
visible immediately after the @code{BEGIN}) can only be handled by
rearranging the loop. E.g., the ``most insidious'' example above can be
arranged into:
@example
BEGIN
  @{ x @}
  ... 0=
WHILE
  x
REPEAT
@end example

@node How long do locals live?, Programming Style, Where are locals visible by name?, Gforth locals
@subsubsection How long do locals live?
@cindex locals lifetime
@cindex lifetime of locals

The right answer for the lifetime question would be: A local lives at
least as long as it can be accessed. For a value-flavoured local this
means: until the end of its visibility. However, a variable-flavoured
local could be accessed through its address far beyond its visibility
scope. Ultimately, this would mean that such locals would have to be
garbage collected. Since this entails un-Forth-like implementation
complexities, I adopted the same cowardly solution as some other
languages (e.g., C): The local lives only as long as it is visible;
afterwards its address is invalid (and programs that access it
afterwards are erroneous).

@node Programming Style, Implementation, How long do locals live?, Gforth locals
@subsubsection Programming Style
@cindex locals programming style
@cindex programming style, locals

The freedom to define locals anywhere has the potential to change
programming styles dramatically. In particular, the need to use the
return stack for intermediate storage vanishes. Moreover, all stack
manipulations (except @code{PICK}s and @code{ROLL}s with run-time
determined arguments) can be eliminated: If the stack items are in the
wrong order, just write a locals definition for all of them; then
write the items in the order you want.

This seems a little far-fetched and eliminating stack manipulations is
unlikely to become a conscious programming objective. Still, the number
of stack manipulations will be reduced dramatically if local variables
are used liberally (e.g., compare @code{max} in @ref{Gforth locals} with
a traditional implementation of @code{max}).

This shows one potential benefit of locals: making Forth programs more
readable. Of course, this benefit will only be realized if the
programmers continue to honour the principle of factoring instead of
using the added latitude to make the words longer.

@cindex single-assignment style for locals
Using @code{TO} can and should be avoided.  Without @code{TO},
every value-flavoured local has only a single assignment and many
advantages of functional languages apply to Forth. I.e., programs are
easier to analyse, to optimize and to read: It is clear from the
definition what the local stands for, it does not turn into something
different later.

E.g., a definition using @code{TO} might look like this:
@example
: strcmp @{ addr1 u1 addr2 u2 -- n @}
 u1 u2 min 0
 ?do
   addr1 c@@ addr2 c@@ -
   ?dup-if
     unloop exit
   then
   addr1 char+ TO addr1
   addr2 char+ TO addr2
 loop
 u1 u2 - ;
@end example
Here, @code{TO} is used to update @code{addr1} and @code{addr2} at
every loop iteration. @code{strcmp} is a typical example of the
readability problems of using @code{TO}. When you start reading
@code{strcmp}, you think that @code{addr1} refers to the start of the
string. Only near the end of the loop you realize that it is something
else.

This can be avoided by defining two locals at the start of the loop that
are initialized with the right value for the current iteration.
@example
: strcmp @{ addr1 u1 addr2 u2 -- n @}
 addr1 addr2
 u1 u2 min 0 
 ?do @{ s1 s2 @}
   s1 c@@ s2 c@@ -
   ?dup-if
     unloop exit
   then
   s1 char+ s2 char+
 loop
 2drop
 u1 u2 - ;
@end example
Here it is clear from the start that @code{s1} has a different value
in every loop iteration.

@node Implementation,  , Programming Style, Gforth locals
@subsubsection Implementation
@cindex locals implementation
@cindex implementation of locals

@cindex locals stack
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
also eliminates the problems and restrictions of using the return stack
as locals stack. Like the other stacks, the locals stack grows toward
lower addresses. A few primitives allow an efficient implementation:

doc-@local#
doc-f@local#
doc-laddr#
doc-lp+!#
doc-lp!
doc->l
doc-f>l

In addition to these primitives, some specializations of these
primitives for commonly occurring inline arguments are provided for
efficiency reasons, e.g., @code{@@local0} as specialization of
@code{@@local#} for the inline argument 0. The following compiling words
compile the right specialized version, or the general version, as
appropriate:

doc-compile-@local
doc-compile-f@local
doc-compile-lp+!

Combinations of conditional branches and @code{lp+!#} like
@code{?branch-lp+!#} (the locals pointer is only changed if the branch
is taken) are provided for efficiency and correctness in loops.

A special area in the dictionary space is reserved for keeping the
local variable names. @code{@{} switches the dictionary pointer to this
area and @code{@}} switches it back and generates the locals
initializing code. @code{W:} etc.@ are normal defining words. This
special area is cleared at the start of every colon definition.

@cindex word list for defining locals
A special feature of Gforth's dictionary is used to implement the
definition of locals without type specifiers: every word list (aka
vocabulary) has its own methods for searching
etc. (@pxref{Word Lists}). For the present purpose we defined a word list
with a special search method: When it is searched for a word, it
actually creates that word using @code{W:}. @code{@{} changes the search
order to first search the word list containing @code{@}}, @code{W:} etc.,
and then the word list for defining locals without type specifiers.

The lifetime rules support a stack discipline within a colon
definition: The lifetime of a local is either nested with other locals
lifetimes or it does not overlap them.

At @code{BEGIN}, @code{IF}, and @code{AHEAD} no code for locals stack
pointer manipulation is generated. Between control structure words
locals definitions can push locals onto the locals stack. @code{AGAIN}
is the simplest of the other three control flow words. It has to
restore the locals stack depth of the corresponding @code{BEGIN}
before branching. The code looks like this:
@format
@code{lp+!#} current-locals-size @minus{} dest-locals-size
@code{branch} <begin>
@end format

@code{UNTIL} is a little more complicated: If it branches back, it
must adjust the stack just like @code{AGAIN}. But if it falls through,
the locals stack must not be changed. The compiler generates the
following code:
@format
@code{?branch-lp+!#} <begin> current-locals-size @minus{} dest-locals-size
@end format
The locals stack pointer is only adjusted if the branch is taken.

@code{THEN} can produce somewhat inefficient code:
@format
@code{lp+!#} current-locals-size @minus{} orig-locals-size
<orig target>:
@code{lp+!#} orig-locals-size @minus{} new-locals-size
@end format
The second @code{lp+!#} adjusts the locals stack pointer from the
level at the @i{orig} point to the level after the @code{THEN}. The
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
adjustment from the current level to the right level after the
@code{THEN}.

@cindex locals information on the control-flow stack
@cindex control-flow stack items, locals information
In a conventional Forth implementation a dest control-flow stack entry
is just the target address and an orig entry is just the address to be
patched. Our locals implementation adds a word list to every orig or dest
item. It is the list of locals visible (or assumed visible) at the point
described by the entry. Our implementation also adds a tag to identify
the kind of entry, in particular to differentiate between live and dead
(reachable and unreachable) orig entries.

A few unusual operations have to be performed on locals word lists:

doc-common-list
doc-sub-list?
doc-list-size

Several features of our locals word list implementation make these
operations easy to implement: The locals word lists are organised as
linked lists; the tails of these lists are shared, if the lists
contain some of the same locals; and the address of a name is greater
than the address of the names behind it in the list.

Another important implementation detail is the variable
@code{dead-code}. It is used by @code{BEGIN} and @code{THEN} to
determine if they can be reached directly or only through the branch
that they resolve. @code{dead-code} is set by @code{UNREACHABLE},
@code{AHEAD}, @code{EXIT} etc., and cleared at the start of a colon
definition, by @code{BEGIN} and usually by @code{THEN}.

Counted loops are similar to other loops in most respects, but
@code{LEAVE} requires special attention: It performs basically the same
service as @code{AHEAD}, but it does not create a control-flow stack
entry. Therefore the information has to be stored elsewhere;
traditionally, the information was stored in the target fields of the
branches created by the @code{LEAVE}s, by organizing these fields into a
linked list. Unfortunately, this clever trick does not provide enough
space for storing our extended control flow information. Therefore, we
introduce another stack, the leave stack. It contains the control-flow
stack entries for all unresolved @code{LEAVE}s.

Local names are kept until the end of the colon definition, even if
they are no longer visible in any control-flow path. In a few cases
this may lead to increased space needs for the locals name area, but
usually less than reclaiming this space would cost in code size.


@node ANS Forth locals,  , Gforth locals, Locals
@subsection ANS Forth locals
@cindex locals, ANS Forth style

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
possible syntaxes is a subset of the syntax we used in the Gforth locals
wordset, i.e.:

@example
@{ local1 local2 ... -- comment @}
@end example
@noindent
or
@example
@{ local1 local2 ... @}
@end example

The order of the locals corresponds to the order in a stack comment. The
restrictions are:

@itemize @bullet
@item
Locals can only be cell-sized values (no type specifiers are allowed).
@item
Locals can be defined only outside control structures.
@item
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
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
stack easier.
@item
The whole definition must be in one line.
@end itemize

Locals defined in this way behave like @code{VALUE}s (@xref{Simple
Defining Words}). I.e., they are initialized from the stack. Using their
name produces their value. Their value can be changed using @code{TO}.

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
another ANS Forth system, use @file{compat/anslocal.fs} to implement the
syntax on the other system.

Note that a syntax shown in the standard, section A.13 looks
similar, but is quite different in having the order of locals
reversed. Beware!

The ANS Forth locals wordset itself consists of a word:

doc-(local)

The ANS Forth locals extension wordset defines a syntax using @code{locals|}, but it is so
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
problem with this syntax is that the locals are defined in an order
reversed with respect to the standard stack comment notation, making
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
locals wordset.


@c ----------------------------------------------------------
@node Structures, Object-oriented Forth, Locals, Words
@section  Structures
@cindex structures
@cindex records

This section presents the structure package that comes with Gforth. A
version of the package implemented in ANS Forth is available in
@file{compat/struct.fs}. This package was inspired by a posting on
comp.lang.forth in 1989 (unfortunately I don't remember, by whom;
possibly John Hayes). A version of this section has been published in
???. Marcel Hendrix provided helpful comments.

@menu
* Why explicit structure support?::  
* Structure Usage::             
* Structure Naming Convention::  
* Structure Implementation::    
* Structure Glossary::          
@end menu

@node Why explicit structure support?, Structure Usage, Structures, Structures
@subsection Why explicit structure support?

@cindex address arithmetic for structures
@cindex structures using address arithmetic
If we want to use a structure containing several fields, we could simply
reserve memory for it, and access the fields using address arithmetic
(@pxref{Address arithmetic}). As an example, consider a structure with
the following fields

@table @code
@item a
is a float
@item b
is a cell
@item c
is a float
@end table

Given the (float-aligned) base address of the structure we get the
address of the field

@table @code
@item a
without doing anything further.
@item b
with @code{float+}
@item c
with @code{float+ cell+ faligned}
@end table

It is easy to see that this can become quite tiring. 

Moreover, it is not very readable, because seeing a
@code{cell+} tells us neither which kind of structure is
accessed nor what field is accessed; we have to somehow infer the kind
of structure, and then look up in the documentation, which field of
that structure corresponds to that offset.

Finally, this kind of address arithmetic also causes maintenance
troubles: If you add or delete a field somewhere in the middle of the
structure, you have to find and change all computations for the fields
afterwards.

So, instead of using @code{cell+} and friends directly, how
about storing the offsets in constants:

@example
0 constant a-offset
0 float+ constant b-offset
0 float+ cell+ faligned c-offset
@end example

Now we can get the address of field @code{x} with @code{x-offset
+}. This is much better in all respects. Of course, you still
have to change all later offset definitions if you add a field. You can
fix this by declaring the offsets in the following way:

@example
0 constant a-offset
a-offset float+ constant b-offset
b-offset cell+ faligned constant c-offset
@end example

Since we always use the offsets with @code{+}, we could use a defining
word @code{cfield} that includes the @code{+} in the action of the
defined word:

@example
: cfield ( n "name" -- )
    create ,
does> ( name execution: addr1 -- addr2 )
    @@ + ;

0 cfield a
0 a float+ cfield b
0 b cell+ faligned cfield c
@end example

Instead of @code{x-offset +}, we now simply write @code{x}.

The structure field words now can be used quite nicely. However,
their definition is still a bit cumbersome: We have to repeat the
name, the information about size and alignment is distributed before
and after the field definitions etc.  The structure package presented
here addresses these problems.

@node Structure Usage, Structure Naming Convention, Why explicit structure support?, Structures
@subsection Structure Usage
@cindex structure usage

@cindex @code{field} usage
@cindex @code{struct} usage
@cindex @code{end-struct} usage
You can define a structure for a (data-less) linked list with:
@example
struct
    cell% field list-next
end-struct list%
@end example

With the address of the list node on the stack, you can compute the
address of the field that contains the address of the next node with
@code{list-next}. E.g., you can determine the length of a list
with:

@example
: list-length ( list -- n )
\ "list" is a pointer to the first element of a linked list
\ "n" is the length of the list
    0 BEGIN ( list1 n1 )
        over
    WHILE ( list1 n1 )
        1+ swap list-next @@ swap
    REPEAT
    nip ;
@end example

You can reserve memory for a list node in the dictionary with
@code{list% %allot}, which leaves the address of the list node on the
stack. For the equivalent allocation on the heap you can use @code{list%
%alloc} (or, for an @code{allocate}-like stack effect (i.e., with ior),
use @code{list% %allocate}). You can get the the size of a list
node with @code{list% %size} and its alignment with @code{list%
%alignment}.

Note that in ANS Forth the body of a @code{create}d word is
@code{aligned} but not necessarily @code{faligned};
therefore, if you do a:
@example
create @emph{name} foo% %allot
@end example

@noindent
then the memory alloted for @code{foo%} is
guaranteed to start at the body of @code{@emph{name}} only if
@code{foo%} contains only character, cell and double fields.

@cindex strcutures containing structures
You can include a structure @code{foo%} as a field of
another structure, like this:
@example
struct
...
    foo% field ...
...
end-struct ...
@end example

@cindex structure extension
@cindex extended records
Instead of starting with an empty structure, you can extend an
existing structure. E.g., a plain linked list without data, as defined
above, is hardly useful; You can extend it to a linked list of integers,
like this:@footnote{This feature is also known as @emph{extended
records}. It is the main innovation in the Oberon language; in other
words, adding this feature to Modula-2 led Wirth to create a new
language, write a new compiler etc.  Adding this feature to Forth just
required a few lines of code.}

@example
list%
    cell% field intlist-int
end-struct intlist%
@end example

@code{intlist%} is a structure with two fields:
@code{list-next} and @code{intlist-int}.

@cindex structures containing arrays
You can specify an array type containing @emph{n} elements of
type @code{foo%} like this:

@example
foo% @emph{n} *
@end example

You can use this array type in any place where you can use a normal
type, e.g., when defining a @code{field}, or with
@code{%allot}.

@cindex first field optimization
The first field is at the base address of a structure and the word
for this field (e.g., @code{list-next}) actually does not change
the address on the stack. You may be tempted to leave it away in the
interest of run-time and space efficiency. This is not necessary,
because the structure package optimizes this case and compiling such
words does not generate any code. So, in the interest of readability
and maintainability you should include the word for the field when
accessing the field.

@node Structure Naming Convention, Structure Implementation, Structure Usage, Structures
@subsection Structure Naming Convention
@cindex structure naming convention

The field names that come to (my) mind are often quite generic, and,
if used, would cause frequent name clashes. E.g., many structures
probably contain a @code{counter} field. The structure names
that come to (my) mind are often also the logical choice for the names
of words that create such a structure.

Therefore, I have adopted the following naming conventions: 

@itemize @bullet
@cindex field naming convention
@item
The names of fields are of the form
@code{@emph{struct}-@emph{field}}, where
@code{@emph{struct}} is the basic name of the structure, and
@code{@emph{field}} is the basic name of the field. You can
think of field words as converting the (address of the)
structure into the (address of the) field.

@cindex structure naming convention
@item
The names of structures are of the form
@code{@emph{struct}%}, where
@code{@emph{struct}} is the basic name of the structure.
@end itemize

This naming convention does not work that well for fields of extended
structures; e.g., the integer list structure has a field
@code{intlist-int}, but has @code{list-next}, not
@code{intlist-next}.

@node Structure Implementation, Structure Glossary, Structure Naming Convention, Structures
@subsection Structure Implementation
@cindex structure implementation
@cindex implementation of structures

The central idea in the implementation is to pass the data about the
structure being built on the stack, not in some global
variable. Everything else falls into place naturally once this design
decision is made.

The type description on the stack is of the form @emph{align
size}. Keeping the size on the top-of-stack makes dealing with arrays
very simple.

@code{field} is a defining word that uses @code{Create}
and @code{DOES>}. The body of the field contains the offset
of the field, and the normal @code{DOES>} action is simply:

@example
@ +
@end example

@noindent
i.e., add the offset to the address, giving the stack effect
@i{addr1 -- addr2} for a field.

@cindex first field optimization, implementation
This simple structure is slightly complicated by the optimization
for fields with offset 0, which requires a different
@code{DOES>}-part (because we cannot rely on there being
something on the stack if such a field is invoked during
compilation). Therefore, we put the different @code{DOES>}-parts
in separate words, and decide which one to invoke based on the
offset. For a zero offset, the field is basically a noop; it is
immediate, and therefore no code is generated when it is compiled.

@node Structure Glossary,  , Structure Implementation, Structures
@subsection Structure Glossary
@cindex structure glossary

doc-%align
doc-%alignment
doc-%alloc
doc-%allocate
doc-%allot
doc-cell%
doc-char%
doc-dfloat%
doc-double%
doc-end-struct
doc-field
doc-float%
doc-naligned
doc-sfloat%
doc-%size
doc-struct

@c -------------------------------------------------------------
@node Object-oriented Forth, Passing Commands to the OS, Structures, Words
@section Object-oriented Forth

Gforth comes with three packages for object-oriented programming:
@file{objects.fs}, @file{oof.fs}, and @file{mini-oof.fs}; none of them
is preloaded, so you have to @code{include} them before use. The most
important differences between these packages (and others) are discussed
in @ref{Comparison with other object models}. All packages are written
in ANS Forth and can be used with any other ANS Forth.

@menu
* Why object-oriented programming?::
* Object-Oriented Terminology::
* Objects::
* OOF::
* Mini-OOF::
* Comparison with other object models::  
@end menu


@node Why object-oriented programming?, Object-Oriented Terminology, , Object-oriented Forth
@subsubsection Why object-oriented programming?
@cindex object-oriented programming motivation
@cindex motivation for object-oriented programming

Often we have to deal with several data structures (@emph{objects}),
that have to be treated similarly in some respects, but differently in
others. Graphical objects are the textbook example: circles, triangles,
dinosaurs, icons, and others, and we may want to add more during program
development. We want to apply some operations to any graphical object,
e.g., @code{draw} for displaying it on the screen. However, @code{draw}
has to do something different for every kind of object.
@comment TODO add some other operations eg perimeter, area
@comment and tie in to concrete examples later..

We could implement @code{draw} as a big @code{CASE}
control structure that executes the appropriate code depending on the
kind of object to be drawn. This would be not be very elegant, and,
moreover, we would have to change @code{draw} every time we add
a new kind of graphical object (say, a spaceship).

What we would rather do is: When defining spaceships, we would tell
the system: ``Here's how you @code{draw} a spaceship; you figure
out the rest''.

This is the problem that all systems solve that (rightfully) call
themselves object-oriented; the object-oriented packages presented here
solve this problem (and not much else).
@comment TODO ?list properties of oo systems.. oo vs o-based?

@node Object-Oriented Terminology, Objects, Why object-oriented programming?, Object-oriented Forth
@subsubsection Object-Oriented Terminology
@cindex object-oriented terminology
@cindex terminology for object-oriented programming

This section is mainly for reference, so you don't have to understand
all of it right away.  The terminology is mainly Smalltalk-inspired.  In
short:

@table @emph
@cindex class
@item class
a data structure definition with some extras.

@cindex object
@item object
an instance of the data structure described by the class definition.

@cindex instance variables
@item instance variables
fields of the data structure.

@cindex selector
@cindex method selector
@cindex virtual function
@item selector
(or @emph{method selector}) a word (e.g.,
@code{draw}) that performs an operation on a variety of data
structures (classes). A selector describes @emph{what} operation to
perform. In C++ terminology: a (pure) virtual function.

@cindex method
@item method
the concrete definition that performs the operation
described by the selector for a specific class. A method specifies
@emph{how} the operation is performed for a specific class.

@cindex selector invocation
@cindex message send
@cindex invoking a selector
@item selector invocation
a call of a selector. One argument of the call (the TOS (top-of-stack))
is used for determining which method is used. In Smalltalk terminology:
a message (consisting of the selector and the other arguments) is sent
to the object.

@cindex receiving object
@item receiving object
the object used for determining the method executed by a selector
invocation. In the @file{objects.fs} model, it is the object that is on
the TOS when the selector is invoked. (@emph{Receiving} comes from
the Smalltalk @emph{message} terminology.)

@cindex child class
@cindex parent class
@cindex inheritance
@item child class
a class that has (@emph{inherits}) all properties (instance variables,
selectors, methods) from a @emph{parent class}. In Smalltalk
terminology: The subclass inherits from the superclass. In C++
terminology: The derived class inherits from the base class.

@end table

@c If you wonder about the message sending terminology, it comes from
@c a time when each object had it's own task and objects communicated via
@c message passing; eventually the Smalltalk developers realized that
@c they can do most things through simple (indirect) calls. They kept the
@c terminology.


@node Objects, OOF, Object-Oriented Terminology, Object-oriented Forth
@subsection The @file{objects.fs} model
@cindex objects
@cindex object-oriented programming

@cindex @file{objects.fs}
@cindex @file{oof.fs}

This section describes the @file{objects.fs} package. This material also has been published in @cite{Yet Another Forth Objects Package} by Anton Ertl and appeared in Forth Dimensions 19(2), pages 37--43 (@url{http://www.complang.tuwien.ac.at/forth/objects/objects.html}).
@c McKewan's and Zsoter's packages

This section assumes that you have read @ref{Structures}.

The techniques on which this model is based have been used to implement
the parser generator, Gray, and have also been used in Gforth for
implementing the various flavours of word lists (hashed or not,
case-sensitive or not, special-purpose word lists for locals etc.).


@menu
* Properties of the Objects model::  
* Basic Objects Usage::         
* The Objects base class::            
* Creating objects::            
* Object-Oriented Programming Style::  
* Class Binding::               
* Method conveniences::         
* Classes and Scoping::         
* Object Interfaces::           
* Objects Implementation::      
* Objects Glossary::            
@end menu

Marcel Hendrix provided helpful comments on this section. Andras Zsoter
and Bernd Paysan helped me with the related works section.

@node Properties of the Objects model, Basic Objects Usage, Objects, Objects
@subsubsection Properties of the @file{objects.fs} model
@cindex @file{objects.fs} properties

@itemize @bullet
@item
It is straightforward to pass objects on the stack. Passing
selectors on the stack is a little less convenient, but possible.

@item
Objects are just data structures in memory, and are referenced by their
address. You can create words for objects with normal defining words
like @code{constant}. Likewise, there is no difference between instance
variables that contain objects and those that contain other data.

@item
Late binding is efficient and easy to use.

@item
It avoids parsing, and thus avoids problems with state-smartness
and reduced extensibility; for convenience there are a few parsing
words, but they have non-parsing counterparts. There are also a few
defining words that parse. This is hard to avoid, because all standard
defining words parse (except @code{:noname}); however, such
words are not as bad as many other parsing words, because they are not
state-smart.

@item
It does not try to incorporate everything. It does a few things and does
them well (IMO). In particular, this model was not designed to support
information hiding (although it has features that may help); you can use
a separate package for achieving this.

@item
It is layered; you don't have to learn and use all features to use this
model. Only a few features are necessary (@xref{Basic Objects Usage},
@xref{The Objects base class}, @xref{Creating objects}.), the others
are optional and independent of each other.

@item
An implementation in ANS Forth is available.

@end itemize


@node Basic Objects Usage, The Objects base class, Properties of the Objects model, Objects
@subsubsection Basic @file{objects.fs} Usage
@cindex basic objects usage
@cindex objects, basic usage

You can define a class for graphical objects like this:

@cindex @code{class} usage
@cindex @code{end-class} usage
@cindex @code{selector} usage
@example
object class \ "object" is the parent class
  selector draw ( x y graphical -- )
end-class graphical
@end example

This code defines a class @code{graphical} with an
operation @code{draw}.  We can perform the operation
@code{draw} on any @code{graphical} object, e.g.:

@example
100 100 t-rex draw
@end example

@noindent
where @code{t-rex} is a word (say, a constant) that produces a
graphical object.

@comment TODO add a 2nd operation eg perimeter.. and use for
@comment a concrete example

@cindex abstract class
How do we create a graphical object? With the present definitions,
we cannot create a useful graphical object. The class
@code{graphical} describes graphical objects in general, but not
any concrete graphical object type (C++ users would call it an
@emph{abstract class}); e.g., there is no method for the selector
@code{draw} in the class @code{graphical}.

For concrete graphical objects, we define child classes of the
class @code{graphical}, e.g.:

@cindex @code{overrides} usage
@cindex @code{field} usage in class definition
@example
graphical class \ "graphical" is the parent class
  cell% field circle-radius

:noname ( x y circle -- )
  circle-radius @@ draw-circle ;
overrides draw

:noname ( n-radius circle -- )
  circle-radius ! ;
overrides construct

end-class circle
@end example

Here we define a class @code{circle} as a child of @code{graphical},
with field @code{circle-radius} (which behaves just like a field
(@pxref{Structures}); it defines (using @code{overrides}) new methods
for the selectors @code{draw} and @code{construct} (@code{construct} is
defined in @code{object}, the parent class of @code{graphical}).

Now we can create a circle on the heap (i.e.,
@code{allocate}d memory) with:

@cindex @code{heap-new} usage
@example
50 circle heap-new constant my-circle
@end example

@noindent
@code{heap-new} invokes @code{construct}, thus
initializing the field @code{circle-radius} with 50. We can draw
this new circle at (100,100) with:

@example
100 100 my-circle draw
@end example

@cindex selector invocation, restrictions
@cindex class definition, restrictions
Note: You can only invoke a selector if the object on the TOS
(the receiving object) belongs to the class where the selector was
defined or one of its descendents; e.g., you can invoke
@code{draw} only for objects belonging to @code{graphical}
or its descendents (e.g., @code{circle}).  Immediately before
@code{end-class}, the search order has to be the same as
immediately after @code{class}.

@node The Objects base class, Creating objects, Basic Objects Usage, Objects
@subsubsection The @file{object.fs} base class
@cindex @code{object} class

When you define a class, you have to specify a parent class.  So how do
you start defining classes? There is one class available from the start:
@code{object}. It is ancestor for all classes and so is the
only class that has no parent. It has two selectors: @code{construct}
and @code{print}.

@node Creating objects, Object-Oriented Programming Style, The Objects base class, Objects
@subsubsection Creating objects
@cindex creating objects
@cindex object creation
@cindex object allocation options

@cindex @code{heap-new} discussion
@cindex @code{dict-new} discussion
@cindex @code{construct} discussion
You can create and initialize an object of a class on the heap with
@code{heap-new} ( ... class -- object ) and in the dictionary
(allocation with @code{allot}) with @code{dict-new} (
... class -- object ). Both words invoke @code{construct}, which
consumes the stack items indicated by "..." above.

@cindex @code{init-object} discussion
@cindex @code{class-inst-size} discussion
If you want to allocate memory for an object yourself, you can get its
alignment and size with @code{class-inst-size 2@@} ( class --
align size ). Once you have memory for an object, you can initialize
it with @code{init-object} ( ... class object -- );
@code{construct} does only a part of the necessary work.

@node Object-Oriented Programming Style, Class Binding, Creating objects, Objects
@subsubsection Object-Oriented Programming Style
@cindex object-oriented programming style

This section is not exhaustive.

@cindex stack effects of selectors
@cindex selectors and stack effects
In general, it is a good idea to ensure that all methods for the
same selector have the same stack effect: when you invoke a selector,
you often have no idea which method will be invoked, so, unless all
methods have the same stack effect, you will not know the stack effect
of the selector invocation.

One exception to this rule is methods for the selector
@code{construct}. We know which method is invoked, because we
specify the class to be constructed at the same place. Actually, I
defined @code{construct} as a selector only to give the users a
convenient way to specify initialization. The way it is used, a
mechanism different from selector invocation would be more natural
(but probably would take more code and more space to explain).

@node Class Binding, Method conveniences, Object-Oriented Programming Style, Objects
@subsubsection Class Binding
@cindex class binding
@cindex early binding

@cindex late binding
Normal selector invocations determine the method at run-time depending
on the class of the receiving object. This run-time selection is called
@i{late binding}.

Sometimes it's preferable to invoke a different method. For example,
you might want to use the simple method for @code{print}ing
@code{object}s instead of the possibly long-winded @code{print} method
of the receiver class. You can achieve this by replacing the invocation
of @code{print} with:

@cindex @code{[bind]} usage
@example
[bind] object print
@end example

@noindent
in compiled code or:

@cindex @code{bind} usage
@example
bind object print
@end example

@cindex class binding, alternative to
@noindent
in interpreted code. Alternatively, you can define the method with a
name (e.g., @code{print-object}), and then invoke it through the
name. Class binding is just a (often more convenient) way to achieve
the same effect; it avoids name clutter and allows you to invoke
methods directly without naming them first.

@cindex superclass binding
@cindex parent class binding
A frequent use of class binding is this: When we define a method
for a selector, we often want the method to do what the selector does
in the parent class, and a little more. There is a special word for
this purpose: @code{[parent]}; @code{[parent]
@emph{selector}} is equivalent to @code{[bind] @emph{parent
selector}}, where @code{@emph{parent}} is the parent
class of the current class. E.g., a method definition might look like:

@cindex @code{[parent]} usage
@example
:noname
  dup [parent] foo \ do parent's foo on the receiving object
  ... \ do some more
; overrides foo
@end example

@cindex class binding as optimization
In @cite{Object-oriented programming in ANS Forth} (Forth Dimensions,
March 1997), Andrew McKewan presents class binding as an optimization
technique. I recommend not using it for this purpose unless you are in
an emergency. Late binding is pretty fast with this model anyway, so the
benefit of using class binding is small; the cost of using class binding
where it is not appropriate is reduced maintainability.

While we are at programming style questions: You should bind
selectors only to ancestor classes of the receiving object. E.g., say,
you know that the receiving object is of class @code{foo} or its
descendents; then you should bind only to @code{foo} and its
ancestors.

@node Method conveniences, Classes and Scoping, Class Binding, Objects
@subsubsection Method conveniences
@cindex method conveniences

In a method you usually access the receiving object pretty often.  If
you define the method as a plain colon definition (e.g., with
@code{:noname}), you may have to do a lot of stack
gymnastics. To avoid this, you can define the method with @code{m:
... ;m}. E.g., you could define the method for
@code{draw}ing a @code{circle} with

@cindex @code{this} usage
@cindex @code{m:} usage
@cindex @code{;m} usage
@example
m: ( x y circle -- )
  ( x y ) this circle-radius @@ draw-circle ;m
@end example

@cindex @code{exit} in @code{m: ... ;m}
@cindex @code{exitm} discussion
@cindex @code{catch} in @code{m: ... ;m}
When this method is executed, the receiver object is removed from the
stack; you can access it with @code{this} (admittedly, in this
example the use of @code{m: ... ;m} offers no advantage). Note
that I specify the stack effect for the whole method (i.e. including
the receiver object), not just for the code between @code{m:}
and @code{;m}. You cannot use @code{exit} in
@code{m:...;m}; instead, use
@code{exitm}.@footnote{Moreover, for any word that calls
@code{catch} and was defined before loading
@code{objects.fs}, you have to redefine it like I redefined
@code{catch}: @code{: catch this >r catch r> to-this ;}}

@cindex @code{inst-var} usage
You will frequently use sequences of the form @code{this
@emph{field}} (in the example above: @code{this
circle-radius}). If you use the field only in this way, you can
define it with @code{inst-var} and eliminate the
@code{this} before the field name. E.g., the @code{circle}
class above could also be defined with:

@example
graphical class
  cell% inst-var radius

m: ( x y circle -- )
  radius @@ draw-circle ;m
overrides draw

m: ( n-radius circle -- )
  radius ! ;m
overrides construct

end-class circle
@end example

@code{radius} can only be used in @code{circle} and its
descendent classes and inside @code{m:...;m}.

@cindex @code{inst-value} usage
You can also define fields with @code{inst-value}, which is
to @code{inst-var} what @code{value} is to
@code{variable}.  You can change the value of such a field with
@code{[to-inst]}.  E.g., we could also define the class
@code{circle} like this:

@example
graphical class
  inst-value radius

m: ( x y circle -- )
  radius draw-circle ;m
overrides draw

m: ( n-radius circle -- )
  [to-inst] radius ;m
overrides construct

end-class circle
@end example


@node Classes and Scoping, Object Interfaces, Method conveniences, Objects
@subsubsection Classes and Scoping
@cindex classes and scoping
@cindex scoping and classes

Inheritance is frequent, unlike structure extension. This exacerbates
the problem with the field name convention (@pxref{Structure Naming
Convention}): One always has to remember in which class the field was
originally defined; changing a part of the class structure would require
changes for renaming in otherwise unaffected code.

@cindex @code{inst-var} visibility
@cindex @code{inst-value} visibility
To solve this problem, I added a scoping mechanism (which was not in my
original charter): A field defined with @code{inst-var} (or
@code{inst-value}) is visible only in the class where it is defined and in
the descendent classes of this class.  Using such fields only makes
sense in @code{m:}-defined methods in these classes anyway.

This scoping mechanism allows us to use the unadorned field name,
because name clashes with unrelated words become much less likely.

@cindex @code{protected} discussion
@cindex @code{private} discussion
Once we have this mechanism, we can also use it for controlling the
visibility of other words: All words defined after
@code{protected} are visible only in the current class and its
descendents. @code{public} restores the compilation
(i.e. @code{current}) word list that was in effect before. If you
have several @code{protected}s without an intervening
@code{public} or @code{set-current}, @code{public}
will restore the compilation word list in effect before the first of
these @code{protected}s.

@node Object Interfaces, Objects Implementation, Classes and Scoping, Objects
@subsubsection Object Interfaces
@cindex object interfaces
@cindex interfaces for objects

In this model you can only call selectors defined in the class of the
receiving objects or in one of its ancestors. If you call a selector
with a receiving object that is not in one of these classes, the
result is undefined; if you are lucky, the program crashes
immediately.

@cindex selectors common to hardly-related classes
Now consider the case when you want to have a selector (or several)
available in two classes: You would have to add the selector to a
common ancestor class, in the worst case to @code{object}. You
may not want to do this, e.g., because someone else is responsible for
this ancestor class.

The solution for this problem is interfaces. An interface is a
collection of selectors. If a class implements an interface, the
selectors become available to the class and its descendents. A class
can implement an unlimited number of interfaces. For the problem
discussed above, we would define an interface for the selector(s), and
both classes would implement the interface.

As an example, consider an interface @code{storage} for
writing objects to disk and getting them back, and a class
@code{foo} that implements it. The code would look like this:

@cindex @code{interface} usage
@cindex @code{end-interface} usage
@cindex @code{implementation} usage
@example
interface
  selector write ( file object -- )
  selector read1 ( file object -- )
end-interface storage

bar class
  storage implementation

... overrides write
... overrides read
...
end-class foo
@end example

@noindent
(I would add a word @code{read} @i{( file -- object )} that uses
@code{read1} internally, but that's beyond the point illustrated
here.)

Note that you cannot use @code{protected} in an interface; and
of course you cannot define fields.

In the Neon model, all selectors are available for all classes;
therefore it does not need interfaces. The price you pay in this model
is slower late binding, and therefore, added complexity to avoid late
binding.

@node Objects Implementation, Objects Glossary, Object Interfaces, Objects
@subsubsection @file{objects.fs} Implementation
@cindex @file{objects.fs} implementation

@cindex @code{object-map} discussion
An object is a piece of memory, like one of the data structures
described with @code{struct...end-struct}. It has a field
@code{object-map} that points to the method map for the object's
class.

@cindex method map
@cindex virtual function table
The @emph{method map}@footnote{This is Self terminology; in C++
terminology: virtual function table.} is an array that contains the
execution tokens (@i{xt}s) of the methods for the object's class. Each
selector contains an offset into a method map.

@cindex @code{selector} implementation, class
@code{selector} is a defining word that uses
@code{CREATE} and @code{DOES>}. The body of the
selector contains the offset; the @code{does>} action for a
class selector is, basically:

@example
( object addr ) @@ over object-map @@ + @@ execute
@end example

Since @code{object-map} is the first field of the object, it
does not generate any code. As you can see, calling a selector has a
small, constant cost.

@cindex @code{current-interface} discussion
@cindex class implementation and representation
A class is basically a @code{struct} combined with a method
map. During the class definition the alignment and size of the class
are passed on the stack, just as with @code{struct}s, so
@code{field} can also be used for defining class
fields. However, passing more items on the stack would be
inconvenient, so @code{class} builds a data structure in memory,
which is accessed through the variable
@code{current-interface}. After its definition is complete, the
class is represented on the stack by a pointer (e.g., as parameter for
a child class definition).

A new class starts off with the alignment and size of its parent,
and a copy of the parent's method map. Defining new fields extends the
size and alignment; likewise, defining new selectors extends the
method map. @code{overrides} just stores a new @i{xt} in the method
map at the offset given by the selector.

@cindex class binding, implementation
Class binding just gets the @i{xt} at the offset given by the selector
from the class's method map and @code{compile,}s (in the case of
@code{[bind]}) it.

@cindex @code{this} implementation
@cindex @code{catch} and @code{this}
@cindex @code{this} and @code{catch}
I implemented @code{this} as a @code{value}. At the
start of an @code{m:...;m} method the old @code{this} is
stored to the return stack and restored at the end; and the object on
the TOS is stored @code{TO this}. This technique has one
disadvantage: If the user does not leave the method via
@code{;m}, but via @code{throw} or @code{exit},
@code{this} is not restored (and @code{exit} may
crash). To deal with the @code{throw} problem, I have redefined
@code{catch} to save and restore @code{this}; the same
should be done with any word that can catch an exception. As for
@code{exit}, I simply forbid it (as a replacement, there is
@code{exitm}).

@cindex @code{inst-var} implementation
@code{inst-var} is just the same as @code{field}, with
a different @code{DOES>} action:
@example
@@ this +
@end example
Similar for @code{inst-value}.

@cindex class scoping implementation
Each class also has a word list that contains the words defined with
@code{inst-var} and @code{inst-value}, and its protected
words. It also has a pointer to its parent. @code{class} pushes
the word lists of the class and all its ancestors onto the search order stack,
and @code{end-class} drops them.

@cindex interface implementation
An interface is like a class without fields, parent and protected
words; i.e., it just has a method map. If a class implements an
interface, its method map contains a pointer to the method map of the
interface. The positive offsets in the map are reserved for class
methods, therefore interface map pointers have negative
offsets. Interfaces have offsets that are unique throughout the
system, unlike class selectors, whose offsets are only unique for the
classes where the selector is available (invokable).

This structure means that interface selectors have to perform one
indirection more than class selectors to find their method. Their body
contains the interface map pointer offset in the class method map, and
the method offset in the interface method map. The
@code{does>} action for an interface selector is, basically:

@example
( object selector-body )
2dup selector-interface @@ ( object selector-body object interface-offset )
swap object-map @@ + @@ ( object selector-body map )
swap selector-offset @@ + @@ execute
@end example

where @code{object-map} and @code{selector-offset} are
first fields and generate no code.

As a concrete example, consider the following code:

@example
interface
  selector if1sel1
  selector if1sel2
end-interface if1

object class
  if1 implementation
  selector cl1sel1
  cell% inst-var cl1iv1

' m1 overrides construct
' m2 overrides if1sel1
' m3 overrides if1sel2
' m4 overrides cl1sel2
end-class cl1

create obj1 object dict-new drop
create obj2 cl1    dict-new drop
@end example

The data structure created by this code (including the data structure
for @code{object}) is shown in the <a
href="objects-implementation.eps">figure</a>, assuming a cell size of 4.
@comment TODO add this diagram..

@node Objects Glossary,  , Objects Implementation, Objects
@subsubsection @file{objects.fs} Glossary
@cindex @file{objects.fs} Glossary

doc---objects-bind
doc---objects-<bind>
doc---objects-bind'
doc---objects-[bind]
doc---objects-class
doc---objects-class->map
doc---objects-class-inst-size
doc---objects-class-override!
doc---objects-construct
doc---objects-current'
doc---objects-[current]
doc---objects-current-interface
doc---objects-dict-new
doc---objects-drop-order
doc---objects-end-class
doc---objects-end-class-noname
doc---objects-end-interface
doc---objects-end-interface-noname
doc---objects-exitm
doc---objects-heap-new
doc---objects-implementation
doc---objects-init-object
doc---objects-inst-value
doc---objects-inst-var
doc---objects-interface
doc---objects-;m
doc---objects-m:
doc---objects-method
doc---objects-object
doc---objects-overrides
doc---objects-[parent]
doc---objects-print
doc---objects-protected
doc---objects-public
doc---objects-push-order
doc---objects-selector
doc---objects-this
doc---objects-<to-inst>
doc---objects-[to-inst]
doc---objects-to-this
doc---objects-xt-new

@c -------------------------------------------------------------
@node OOF, Mini-OOF, Objects, Object-oriented Forth
@subsection The @file{oof.fs} model
@cindex oof
@cindex object-oriented programming

@cindex @file{objects.fs}
@cindex @file{oof.fs}

This section describes the @file{oof.fs} package.

The package described in this section has been used in bigFORTH since 1991, and
used for two large applications: a chromatographic system used to
create new medicaments, and a graphic user interface library (MINOS).

You can find a description (in German) of @file{oof.fs} in @cite{Object
oriented bigFORTH} by Bernd Paysan, published in @cite{Vierte Dimension}
10(2), 1994.

@menu
* Properties of the OOF model::
* Basic OOF Usage::
* The OOF base class::
* Class Declaration::
* Class Implementation::
@end menu

@node Properties of the OOF model, Basic OOF Usage, OOF, OOF
@subsubsection Properties of the @file{oof.fs} model
@cindex @file{oof.fs} properties

@itemize @bullet
@item
This model combines object oriented programming with information
hiding. It helps you writing large application, where scoping is
necessary, because it provides class-oriented scoping.

@item
Named objects, object pointers, and object arrays can be created,
selector invocation uses the ``object selector'' syntax. Selector invocation
to objects and/or selectors on the stack is a bit less convenient, but
possible.

@item
Selector invocation and instance variable usage of the active object is
straightforward, since both make use of the active object.

@item
Late binding is efficient and easy to use.

@item
State-smart objects parse selectors. However, extensibility is provided
using a (parsing) selector @code{postpone} and a selector @code{'}.

@item
An implementation in ANS Forth is available.

@end itemize


@node Basic OOF Usage, The OOF base class, Properties of the OOF model, OOF
@subsubsection Basic @file{oof.fs} Usage
@cindex @file{oof.fs} usage

This section uses the same example as for @code{objects} (@pxref{Basic Objects Usage}).

You can define a class for graphical objects like this:

@cindex @code{class} usage
@cindex @code{class;} usage
@cindex @code{method} usage
@example
object class graphical \ "object" is the parent class
  method draw ( x y graphical -- )
class;
@end example

This code defines a class @code{graphical} with an
operation @code{draw}.  We can perform the operation
@code{draw} on any @code{graphical} object, e.g.:

@example
100 100 t-rex draw
@end example

@noindent
where @code{t-rex} is an object or object pointer, created with e.g.
@code{graphical : t-rex}.

@cindex abstract class
How do we create a graphical object? With the present definitions,
we cannot create a useful graphical object. The class
@code{graphical} describes graphical objects in general, but not
any concrete graphical object type (C++ users would call it an
@emph{abstract class}); e.g., there is no method for the selector
@code{draw} in the class @code{graphical}.

For concrete graphical objects, we define child classes of the
class @code{graphical}, e.g.:

@example
graphical class circle \ "graphical" is the parent class
  cell var circle-radius
how:
  : draw ( x y -- )
    circle-radius @@ draw-circle ;

  : init ( n-radius -- (
    circle-radius ! ;
class;
@end example

Here we define a class @code{circle} as a child of @code{graphical},
with a field @code{circle-radius}; it defines new methods for the
selectors @code{draw} and @code{init} (@code{init} is defined in
@code{object}, the parent class of @code{graphical}).

Now we can create a circle in the dictionary with:

@example
50 circle : my-circle
@end example

@noindent
@code{:} invokes @code{init}, thus initializing the field
@code{circle-radius} with 50. We can draw this new circle at (100,100)
with:

@example
100 100 my-circle draw
@end example

@cindex selector invocation, restrictions
@cindex class definition, restrictions
Note: You can only invoke a selector if the receiving object belongs to
the class where the selector was defined or one of its descendents;
e.g., you can invoke @code{draw} only for objects belonging to
@code{graphical} or its descendents (e.g., @code{circle}). The scoping
mechanism will check if you try to invoke a selector that is not
defined in this class hierarchy, so you'll get an error at compilation
time.


@node The OOF base class, Class Declaration, Basic OOF Usage, OOF
@subsubsection The @file{oof.fs} base class
@cindex @file{oof.fs} base class

When you define a class, you have to specify a parent class.  So how do
you start defining classes? There is one class available from the start:
@code{object}. You have to use it as ancestor for all classes. It is the
only class that has no parent. Classes are also objects, except that
they don't have instance variables; class manipulation such as
inheritance or changing definitions of a class is handled through
selectors of the class @code{object}.

@code{object} provides a number of selectors:

@itemize @bullet
@item
@code{class} for subclassing, @code{definitions} to add definitions
later on, and @code{class?} to get type informations (is the class a
subclass of the class passed on the stack?).
doc---object-class
doc---object-definitions
doc---object-class?

@item
@code{init} and @code{dispose} as constructor and destructor of the
object. @code{init} is invocated after the object's memory is allocated,
while @code{dispose} also handles deallocation. Thus if you redefine
@code{dispose}, you have to call the parent's dispose with @code{super
dispose}, too.
doc---object-init
doc---object-dispose

@item
@code{new}, @code{new[]}, @code{:}, @code{ptr}, @code{asptr}, and
@code{[]} to create named and unnamed objects and object arrays or
object pointers.
doc---object-new
doc---object-new[]
doc---object-:
doc---object-ptr
doc---object-asptr
doc---object-[]

@item
@code{::} and @code{super} for explicit scoping. You should use explicit
scoping only for super classes or classes with the same set of instance
variables. Explicitly-scoped selectors use early binding.
doc---object-::
doc---object-super

@item
@code{self} to get the address of the object
doc---object-self

@item
@code{bind}, @code{bound}, @code{link}, and @code{is} to assign object
pointers and instance defers.
doc---object-bind
doc---object-bound
doc---object-link
doc---object-is

@item
@code{'} to obtain selector tokens, @code{send} to invocate selectors
form the stack, and @code{postpone} to generate selector invocation code.
doc---object-'
doc---object-postpone

@item
@code{with} and @code{endwith} to select the active object from the
stack, and enable its scope. Using @code{with} and @code{endwith}
also allows you to create code using selector @code{postpone} without being
trapped by the state-smart objects.
doc---object-with
doc---object-endwith

@end itemize

@node Class Declaration, Class Implementation, The OOF base class, OOF
@subsubsection Class Declaration
@cindex class declaration

@itemize @bullet
@item
Instance variables
doc---oof-var

@item
Object pointers
doc---oof-ptr
doc---oof-asptr

@item
Instance defers
doc---oof-defer

@item
Method selectors
doc---oof-early
doc---oof-method

@item
Class-wide variables
doc---oof-static

@item
End declaration
doc---oof-how:
doc---oof-class;

@end itemize

@c -------------------------------------------------------------
@node Class Implementation,  , Class Declaration, OOF
@subsubsection Class Implementation
@cindex class implementation

@c -------------------------------------------------------------
@node Mini-OOF, Comparison with other object models, OOF, Object-oriented Forth
@subsection The @file{mini-oof.fs} model
@cindex mini-oof

Gforth's third object oriented Forth package is a 12-liner. It uses a
mixture of the @file{object.fs} and the @file{oof.fs} syntax,
and reduces to the bare minimum of features. This is based on a posting
of Bernd Paysan in comp.arch.

@menu
* Basic Mini-OOF Usage::
* Mini-OOF Example::
* Mini-OOF Implementation::
@end menu

@c -------------------------------------------------------------
@node Basic Mini-OOF Usage, Mini-OOF Example, , Mini-OOF
@subsubsection Basic @file{mini-oof.fs} Usage
@cindex mini-oof usage

There is a base class (@code{class}, which allocates one cell for the
object pointer) plus seven other words: to define a method, a variable,
a class; to end a class, to resolve binding, to allocate an object and
to compile a class method.
@comment TODO better description of the last one

doc-object
doc-method
doc-var
doc-class
doc-end-class
doc-defines
doc-new
doc-::


@c -------------------------------------------------------------
@node Mini-OOF Example, Mini-OOF Implementation, Basic Mini-OOF Usage, Mini-OOF
@subsubsection Mini-OOF Example
@cindex mini-oof example

A short example shows how to use this package. This example, in slightly
extended form, is supplied as @file{moof-exm.fs}
@comment TODO could flesh this out with some comments from the Forthwrite article

@example
object class
  method init
  method draw
end-class graphical
@end example

This code defines a class @code{graphical} with an
operation @code{draw}.  We can perform the operation
@code{draw} on any @code{graphical} object, e.g.:

@example
100 100 t-rex draw
@end example

where @code{t-rex} is an object or object pointer, created with e.g.
@code{graphical new Constant t-rex}.

For concrete graphical objects, we define child classes of the
class @code{graphical}, e.g.:

@example
graphical class
  cell var circle-radius
end-class circle \ "graphical" is the parent class

:noname ( x y -- )
  circle-radius @@ draw-circle ; circle defines draw
:noname ( r -- )
  circle-radius ! ; circle defines init
@end example

There is no implicit init method, so we have to define one. The creation
code of the object now has to call init explicitely.

@example
circle new Constant my-circle
50 my-circle init
@end example

It is also possible to add a function to create named objects with
automatic call of @code{init}, given that all objects have @code{init}
on the same place:

@example
: new: ( .. o "name" -- )
    new dup Constant init ;
80 circle new: large-circle
@end example

We can draw this new circle at (100,100) with:

@example
100 100 my-circle draw
@end example

@node Mini-OOF Implementation, , Mini-OOF Example, Mini-OOF
@subsubsection @file{mini-oof.fs} Implementation

Object-oriented systems with late binding typically use a
``vtable''-approach: the first variable in each object is a pointer to a
table, which contains the methods as function pointers. The vtable
may also contain other information.

So first, let's declare methods:

@example
: method ( m v -- m' v ) Create  over , swap cell+ swap
  DOES> ( ... o -- ... ) @ over @ + @ execute ;
@end example

During method declaration, the number of methods and instance
variables is on the stack (in address units). @code{method} creates
one method and increments the method number. To execute a method, it
takes the object, fetches the vtable pointer, adds the offset, and
executes the @i{xt} stored there. Each method takes the object it is
invoked from as top of stack parameter. The method itself should
consume that object.

Now, we also have to declare instance variables

@example
: var ( m v size -- m v' ) Create  over , +
  DOES> ( o -- addr ) @ + ;
@end example

As before, a word is created with the current offset. Instance
variables can have different sizes (cells, floats, doubles, chars), so
all we do is take the size and add it to the offset. If your machine
has alignment restrictions, put the proper @code{aligned} or
@code{faligned} before the variable, to adjust the variable
offset. That's why it is on the top of stack.

We need a starting point (the base object) and some syntactic sugar:

@example
Create object  1 cells , 2 cells ,
: class ( class -- class methods vars ) dup 2@ ;
@end example

For inheritance, the vtable of the parent object has to be
copied when a new, derived class is declared. This gives all the
methods of the parent class, which can be overridden, though.

@example
: end-class  ( class methods vars -- )
  Create  here >r , dup , 2 cells ?DO ['] noop , 1 cells +LOOP
  cell+ dup cell+ r> rot @ 2 cells /string move ;
@end example

The first line creates the vtable, initialized with
@code{noop}s. The second line is the inheritance mechanism, it
copies the xts from the parent vtable.

We still have no way to define new methods, let's do that now:

@example
: defines ( xt class -- ) ' >body @ + ! ;
@end example

To allocate a new object, we need a word, too:

@example
: new ( class -- o )  here over @ allot swap over ! ;
@end example

Sometimes derived classes want to access the method of the
parent object. There are two ways to achieve this with Mini-OOF:
first, you could use named words, and second, you could look up the
vtable of the parent object.

@example
: :: ( class "name" -- ) ' >body @ + @ compile, ;
@end example


Nothing can be more confusing than a good example, so here is
one. First let's declare a text object (called
@code{button}), that stores text and position:

@example
object class
  cell var text
  cell var len
  cell var x
  cell var y
  method init
  method draw
end-class button
@end example

@noindent
Now, implement the two methods, @code{draw} and @code{init}:

@example
:noname ( o -- )
 >r r@ x @ r@ y @ at-xy  r@ text @ r> len @ type ;
 button defines draw
:noname ( addr u o -- )
 >r 0 r@ x ! 0 r@ y ! r@ len ! r> text ! ;
 button defines init
@end example

@noindent
To demonstrate inheritance, we define a class @code{bold-button}, with no
new data and no new methods:

@example
button class
end-class bold-button

: bold   27 emit ." [1m" ;
: normal 27 emit ." [0m" ;
@end example

@noindent
The class @code{bold-button} has a different draw method to
@code{button}, but the new method is defined in terms of the draw method
for @code{button}:

@example
:noname bold [ button :: draw ] normal ; bold-button defines draw
@end example

@noindent
Finally, create two objects and apply methods:

@example
button new Constant foo
s" thin foo" foo init
page
foo draw
bold-button new Constant bar
s" fat bar" bar init
1 bar y !
bar draw
@end example


@node Comparison with other object models, , Mini-OOF, Object-oriented Forth
@subsubsection Comparison with other object models
@cindex comparison of object models
@cindex object models, comparison

Many object-oriented Forth extensions have been proposed (@cite{A survey
of object-oriented Forths} (SIGPLAN Notices, April 1996) by Bradford
J. Rodriguez and W. F. S. Poehlman lists 17). This section discusses the
relation of the object models described here to two well-known and two
closely-related (by the use of method maps) models.

@cindex Neon model
The most popular model currently seems to be the Neon model (see
@cite{Object-oriented programming in ANS Forth} (Forth Dimensions, March
1997) by Andrew McKewan) but this model has a number of limitations
@footnote{A longer version of this critique can be
found in @cite{On Standardizing Object-Oriented Forth Extensions} (Forth
Dimensions, May 1997) by Anton Ertl.}:

@itemize @bullet
@item
It uses a @code{@emph{selector
object}} syntax, which makes it unnatural to pass objects on the
stack.

@item
It requires that the selector parses the input stream (at
compile time); this leads to reduced extensibility and to bugs that are+
hard to find.

@item
It allows using every selector to every object;
this eliminates the need for classes, but makes it harder to create
efficient implementations. 
@end itemize

@cindex Pountain's object-oriented model
Another well-known publication is @cite{Object-Oriented Forth} (Academic
Press, London, 1987) by Dick Pountain. However, it is not really about
object-oriented programming, because it hardly deals with late
binding. Instead, it focuses on features like information hiding and
overloading that are characteristic of modular languages like Ada (83).

@cindex Zsoter's object-oriented model
In @cite{Does late binding have to be slow?} (Forth Dimensions 18(1) 1996, pages 31-35)
Andras Zsoter describes a model that makes heavy use of an active object
(like @code{this} in @file{objects.fs}): The active object is not only
used for accessing all fields, but also specifies the receiving object
of every selector invocation; you have to change the active object
explicitly with @code{@{ ... @}}, whereas in @file{objects.fs} it
changes more or less implicitly at @code{m: ... ;m}. Such a change at
the method entry point is unnecessary with the Zsoter's model, because
the receiving object is the active object already. On the other hand, the explicit
change is absolutely necessary in that model, because otherwise no one
could ever change the active object. An ANS Forth implementation of this
model is available at @url{http://www.forth.org/fig/oopf.html}.

@cindex @file{oof.fs}, differences to other models
The @file{oof.fs} model combines information hiding and overloading
resolution (by keeping names in various word lists) with object-oriented
programming. It sets the active object implicitly on method entry, but
also allows explicit changing (with @code{>o...o>} or with
@code{with...endwith}). It uses parsing and state-smart objects and
classes for resolving overloading and for early binding: the object or
class parses the selector and determines the method from this. If the
selector is not parsed by an object or class, it performs a call to the
selector for the active object (late binding), like Zsoter's model.
Fields are always accessed through the active object. The big
disadvantage of this model is the parsing and the state-smartness, which
reduces extensibility and increases the opportunities for subtle bugs;
essentially, you are only safe if you never tick or @code{postpone} an
object or class (Bernd disagrees, but I (Anton) am not convinced).

@cindex @file{mini-oof.fs}, differences to other models
The @file{mini-oof.fs} model is quite similar to a very stripped-down version of
the @file{objects.fs} model, but syntactically it is a mixture of the @file{objects.fs} and
@file{oof.fs} models.

@c -------------------------------------------------------------
@node Passing Commands to the OS, Miscellaneous Words, Object-oriented Forth, Words
@section Passing Commands to the Operating System
@cindex operating system - passing commands
@cindex shell commands

Gforth allows you to pass an arbitrary string to the host operating
system shell (if such a thing exists) for execution.

doc-sh
doc-system
doc-$?
doc-getenv

@c -------------------------------------------------------------
@node Miscellaneous Words,  , Passing Commands to the OS, Words
@section Miscellaneous Words
@cindex miscellaneous words

@comment TODO find homes for these

These section lists the ANS Forth words that are not documented
elsewhere in this manual. Ultimately, they all need proper homes.

doc-ms
doc-time&date

doc-[compile]

The following ANS Forth words are not currently supported by Gforth 
(@pxref{ANS conformance}):

@code{EDITOR} 
@code{EKEY} 
@code{EKEY>CHAR} 
@code{EKEY?} 
@code{EMIT?} 
@code{FORGET} 

@c ******************************************************************
@node Error messages, Tools, Words, Top
@chapter Error messages
@cindex error messages
@cindex backtrace

A typical Gforth error message looks like this:

@example
in file included from :-1
in file included from ./yyy.fs:1
./xxx.fs:4: Invalid memory address
bar
^^^
$400E664C @@
$400E6664 foo
@end example

The message identifying the error is @code{Invalid memory address}.  The
error happened when text-interpreting line 4 of the file
@file{./xxx.fs}. This line is given (it contains @code{bar}), and the
word on the line where the error happened, is pointed out (with
@code{^^^}).

The file containing the error was included in line 1 of @file{./yyy.fs},
and @file{yyy.fs} was included from a non-file (in this case, by giving
@file{yyy.fs} as command-line parameter to Gforth).

At the end of the error message you find a return stack dump that can be
interpreted as a backtrace (possibly empty). On top you find the top of
the return stack when the @code{throw} happened, and at the bottom you
find the return stack entry just above the return stack of the topmost
text interpreter.

To the right of most return stack entries you see a guess for the word
that pushed that return stack entry as its return address. This gives a
backtrace. In our case we see that @code{bar} called @code{foo}, and
@code{foo} called @code{@@} (and @code{@@} had an @emph{Invalid memory
address} exception).

Note that the backtrace is not perfect: We don't know which return stack
entries are return addresses (so we may get false positives); and in
some cases (e.g., for @code{abort"}) we cannot determine from the return
address the word that pushed the return address, so for some return
addresses you see no names in the return stack dump.

@cindex @code{catch} and backtraces
The return stack dump represents the return stack at the time when a
specific @code{throw} was executed.  In programs that make use of
@code{catch}, it is not necessarily clear which @code{throw} should be
used for the return stack dump (e.g., consider one @code{throw} that
indicates an error, which is caught, and during recovery another error
happens; which @code{throw} should be used for the stack dump).  Gforth
presents the return stack dump for the first @code{throw} after the last
executed (not returned-to) @code{catch}; this works well in the usual
case.

@cindex @code{gforth-fast} and backtraces
@cindex @code{gforth-fast}, difference from @code{gforth}
@cindex backtraces with @code{gforth-fast}
@cindex return stack dump with @code{gforth-fast}
@code{gforth} is able to do a return stack dump for throws generated
from primitives (e.g., invalid memory address, stack empty etc.);
@code{gforth-fast} is only able to do a return stack dump from a
directly called @code{throw} (including @code{abort} etc.).  This is the
only difference (apart from a speed factor of between 1.15 (K6-2) and
1.6 (21164A)) between @code{gforth} and @code{gforth-fast}.  Given an
exception caused by a primitive in @code{gforth-fast}, you will
typically see no return stack dump at all; however, if the exception is
caught by @code{catch} (e.g., for restoring some state), and then
@code{throw}n again, the return stack dump will be for the first such
@code{throw}.

@c ******************************************************************
@node Tools, ANS conformance, Error messages, Top
@chapter Tools

@menu
* ANS Report::                  Report the words used, sorted by wordset.
@end menu

See also @ref{Emacs and Gforth}.

@node ANS Report,  , Tools, Tools
@section @file{ans-report.fs}: Report the words used, sorted by wordset
@cindex @file{ans-report.fs}
@cindex report the words used in your program
@cindex words used in your program

If you want to label a Forth program as ANS Forth Program, you must
document which wordsets the program uses; for extension wordsets, it is
helpful to list the words the program requires from these wordsets
(because Forth systems are allowed to provide only some words of them).

The @file{ans-report.fs} tool makes it easy for you to determine which
words from which wordset and which non-ANS words your application
uses. You simply have to include @file{ans-report.fs} before loading the
program you want to check. After loading your program, you can get the
report with @code{print-ans-report}. A typical use is to run this as
batch job like this:
@example
gforth ans-report.fs myprog.fs -e "print-ans-report bye"
@end example

The output looks like this (for @file{compat/control.fs}):
@example
The program uses the following words
from CORE :
: POSTPONE THEN ; immediate ?dup IF 0= 
from BLOCK-EXT :
\ 
from FILE :
( 
@end example

@subsection Caveats

Note that @file{ans-report.fs} just checks which words are used, not whether
they are used in an ANS Forth conforming way!

Some words are defined in several wordsets in the
standard. @file{ans-report.fs} reports them for only one of the
wordsets, and not necessarily the one you expect. It depends on usage
which wordset is the right one to specify. E.g., if you only use the
compilation semantics of @code{S"}, it is a Core word; if you also use
its interpretation semantics, it is a File word.

@c ******************************************************************
@node ANS conformance, Model, Tools, Top
@chapter ANS conformance
@cindex ANS conformance of Gforth

To the best of our knowledge, Gforth is an

ANS Forth System
@itemize @bullet
@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

@cindex system documentation
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 =====================================================================
@cindex core words, system documentation
@cindex system documentation, core words

@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 ---------------------------------------------------------------------
@cindex core words, implementation-defined options
@cindex implementation-defined options, core words


@table @i
@item (Cell) aligned addresses:
@cindex cell-aligned addresses
@cindex 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:
@cindex @code{EMIT} and non-graphic characters
@cindex non-graphic characters and @code{EMIT}
The character is output using the C library function (actually, macro)
@code{putc}.

@item character editing of @code{ACCEPT} and @code{EXPECT}:
@cindex character editing of @code{ACCEPT} and @code{EXPECT}
@cindex editing in @code{ACCEPT} and @code{EXPECT}
@cindex @code{ACCEPT}, editing
@cindex @code{EXPECT}, editing
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). @xref{Command-line editing}.

@item character set:
@cindex 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:
@cindex 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:
@cindex character-set extensions and matching of names
@cindex case-sensitivity for name lookup
@cindex name lookup, case-sensitivity
@cindex locale and case-sensitivity
Any character except the ASCII NUL character can be used in a
name. Matching is case-insensitive (except in @code{TABLE}s). 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:
@cindex space delimiters
@cindex control characters as delimiters
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:
@cindex control-flow stack, format
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
@cindex 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}:
@cindex @code{EXPECT}, display after end of input
@cindex @code{ACCEPT}, display after end of input
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"}:
@cindex exception abort sequence of @code{ABORT"}
@cindex @code{ABORT"}, exception abort sequence
The error string is stored into the variable @code{"error} and a
@code{-2 throw} is performed.

@item input line terminator:
@cindex input line terminator
@cindex line terminator on input
@cindex newline character on input
For interactive input, @kbd{C-m} (CR) and @kbd{C-j} (LF) 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:
@cindex maximum size of a counted string
@cindex counted string, maximum size
@code{s" /counted-string" environment? drop .}. Currently 255 characters
on all ports, but this may change.

@item maximum size of a parsed string:
@cindex maximum size of a parsed string
@cindex parsed string, maximum size
Given by the constant @code{/line}. Currently 255 characters.

@item maximum size of a definition name, in characters:
@cindex maximum size of a definition name, in characters
@cindex name, maximum length
31

@item maximum string length for @code{ENVIRONMENT?}, in characters:
@cindex maximum string length for @code{ENVIRONMENT?}, in characters
@cindex @code{ENVIRONMENT?} string length, maximum
31

@item method of selecting the user input device:
@cindex user input device, method of selecting
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:
@cindex user output device, method of selecting
@code{EMIT} and @code{TYPE} output to the file-id stored in the value
@code{outfile-id} (@code{stdout} by default). Gforth uses unbuffered
output when the user output device is a terminal, otherwise the output
is buffered.

@item methods of dictionary compilation:
What are we expected to document here?

@item number of bits in one address unit:
@cindex number of bits in one address unit
@cindex address unit, size in bits
@code{s" address-units-bits" environment? drop .}. 8 in all current
ports.

@item number representation and arithmetic:
@cindex number representation and arithmetic
Processor-dependent. Binary two's complement on all current ports.

@item ranges for integer types:
@cindex ranges for integer types
@cindex integer types, ranges
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:
@cindex read-only data space regions
@cindex data-space, read-only regions
The whole Forth data space is writable.

@item size of buffer at @code{WORD}:
@cindex size of buffer at @code{WORD}
@cindex @code{WORD} buffer size
@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:
@cindex cell size
@code{1 cells .}.

@item size of one character in address units:
@cindex char size
@code{1 chars .}. 1 on all current ports.

@item size of the keyboard terminal buffer:
@cindex size of the keyboard terminal buffer
@cindex terminal buffer, size
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:
@cindex size of the pictured numeric output buffer
@cindex pictured numeric output buffer, size
@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}:
@cindex size of the scratch area returned by @code{PAD}
@cindex @code{PAD} size
The remainder of dictionary space. @code{unused pad here - - .}.

@item system case-sensitivity characteristics:
@cindex case-sensitivity characteristics
Dictionary searches are case-insensitive (except in
@code{TABLE}s). 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:
@cindex system prompt
@cindex prompt
@code{ ok} in interpret state, @code{ compiled} in compile state.

@item division rounding:
@cindex 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:
@cindex @code{STATE} values
-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} (Floating-point 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>}:
@cindex @t{DOES>}, visibility of current definition
No.

@end table

@c ---------------------------------------------------------------------
@node core-ambcond, core-other, core-idef, The Core Words
@subsection Ambiguous conditions
@c ---------------------------------------------------------------------
@cindex core words, ambiguous conditions
@cindex ambiguous conditions, core words

@table @i

@item a name is neither a word nor a number:
@cindex name not found
@cindex undefined word
@code{-13 throw} (Undefined word). Actually, @code{-13 bounce}, which
preserves the data and FP stack, so you don't lose more work than
necessary.

@item a definition name exceeds the maximum length allowed:
@cindex word name too long
@code{-19 throw} (Word name too long)

@item addressing a region not inside the various data spaces of the forth system:
@cindex Invalid memory address
The stacks, code space and header 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:
@cindex argument type mismatch
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:
@cindex Interpreting a compile-only word, for @code{'} etc.
@cindex execution token of words with undefined execution semantics
@code{-14 throw} (Interpreting a compile-only word). In some cases, you
get an execution token for @code{compile-only-error} (which performs a
@code{-14 throw} when executed).

@item dividing by zero:
@cindex dividing by zero
@cindex floating point unidentified fault, integer division
On better platforms, this produces a @code{-10 throw} (Division by
zero); on other systems, this typically results in a @code{-55 throw}
(Floating-point unidentified fault).

@item insufficient data stack or return stack space:
@cindex insufficient data stack or return stack space
@cindex stack overflow
@cindex address alignment exception, stack overflow
@cindex Invalid memory address, stack overflow
Depending on the operating system, the installation, and the invocation
of Gforth, this is either checked by the memory management hardware, or
it is not checked. If it is checked, you typically get a @code{-3 throw}
(Stack overflow), @code{-5 throw} (Return stack overflow), or @code{-9
throw} (Invalid memory address) (depending on the platform and how you
achieved the overflow) as soon as the overflow happens. If it is not
checked, overflows typically result in mysterious illegal memory
accesses, producing @code{-9 throw} (Invalid memory address) or
@code{-23 throw} (Address alignment exception); they might also destroy
the internal data structure of @code{ALLOCATE} and friends, resulting in
various errors in these words.

@item insufficient space for loop control parameters:
@cindex insufficient space for loop control parameters
like other return stack overflows.

@item insufficient space in the dictionary:
@cindex insufficient space in the dictionary
@cindex dictionary overflow
If you try to allot (either directly with @code{allot}, or indirectly
with @code{,}, @code{create} etc.) more memory than available in the
dictionary, you get a @code{-8 throw} (Dictionary overflow). If you try
to access memory beyond the end of the dictionary, the results are
similar to stack overflows.

@item interpreting a word with undefined interpretation semantics:
@cindex interpreting a word with undefined interpretation semantics
@cindex Interpreting a compile-only word
For some words, we have defined interpretation semantics. For the
others: @code{-14 throw} (Interpreting a compile-only word).

@item modifying the contents of the input buffer or a string literal:
@cindex 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:
@cindex overflow of the pictured numeric output string
@cindex pictured numeric output string, overflow
@code{-17 throw} (Pictured numeric ouput string overflow).

@item parsed string overflow:
@cindex parsed string overflow
@code{PARSE} cannot overflow. @code{WORD} does not check for overflow.

@item producing a result out of range:
@cindex 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{-10 throw} (divide by zero) or @code{-55
throw} (floating point unidentified fault). @code{convert} and
@code{>number} currently overflow silently.

@item reading from an empty data or return stack:
@cindex stack empty
@cindex stack underflow
@cindex return stack underflow
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, stacks may be checked or not,
depending on operating system, installation, and invocation. If they are
caught by a check, they typically result in @code{-4 throw} (Stack
underflow), @code{-6 throw} (Return stack underflow) or @code{-9 throw}
(Invalid memory address), depending on the platform and which stack
underflows and by how much. Note that even if the system uses checking
(through the MMU), your program may have to underflow by a significant
number of stack items to trigger the reaction (the reason for this is
that the MMU, and therefore the checking, works with a page-size
granularity).  If there is no checking, the symptoms resulting from an
underflow are similar to those from an overflow.  Unbalanced return
stack errors result in a variaty of symptoms, including @code{-9 throw}
(Invalid memory address) and Illegal Instruction (typically @code{-260
throw}).

@item unexpected end of the input buffer, resulting in an attempt to use a zero-length string as a name:
@cindex unexpected end of the input buffer
@cindex zero-length string as a name
@cindex Attempt to use 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:
@cindex @code{>IN} greater than input buffer
The next invocation of a parsing word returns a string with length 0.

@item @code{RECURSE} appears after @code{DOES>}:
@cindex @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}:
@cindex argument input source different than current input source for @code{RESTORE-INPUT}
@cindex argument type mismatch, @code{RESTORE-INPUT}
@cindex @code{RESTORE-INPUT}, Argument type mismatch
@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 restore input source specifications
from other than the current input source.

@item data space containing definitions gets de-allocated:
@cindex data space containing definitions gets de-allocated
Deallocation with @code{allot} is not checked. This typically results in
memory access faults or execution of illegal instructions.

@item data space read/write with incorrect alignment:
@cindex data space read/write with incorrect alignment
@cindex alignment faults
@cindex address alignment exception
Processor-dependent. Typically results in a @code{-23 throw} (Address
alignment exception). Under Linux-Intel 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 violations.

@item data space pointer not properly aligned, @code{,}, @code{C,}:
@cindex 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}):
Like other stack underflows.

@item loop control parameters not available:
@cindex 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}):
@cindex most recent definition does not have a name (@code{IMMEDIATE})
@cindex last word was headerless
@code{abort" last word was headerless"}.

@item name not defined by @code{VALUE} used by @code{TO}:
@cindex name not defined by @code{VALUE} used by @code{TO}
@cindex @code{TO} on non-@code{VALUE}s
@cindex Invalid name argument, @code{TO}
@code{-32 throw} (Invalid name argument) (unless name is a local or was
defined by @code{CONSTANT}; in the latter case it just changes the constant).

@item name not found (@code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]}):
@cindex name not found (@code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]})
@cindex undefined word, @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}):
@cindex 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}:
@cindex @code{POSTPONE} or @code{[COMPILE]} applied to @code{TO}
Assume @code{: X POSTPONE TO ; IMMEDIATE}. @code{X} performs the
compilation semantics of @code{TO}.

@item String longer than a counted string returned by @code{WORD}:
@cindex string longer than a counted string returned by @code{WORD}
@cindex @code{WORD}, string overflow
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}):
@cindex @code{LSHIFT}, large shift counts
@cindex @code{RSHIFT}, large shift counts
Processor-dependent. Typical behaviours are returning 0 and using only
the low bits of the shift count.

@item word not defined via @code{CREATE}:
@cindex @code{>BODY} of non-@code{CREATE}d words
@code{>BODY} produces the PFA of the word no matter how it was defined.

@cindex @code{DOES>} of non-@code{CREATE}d words
@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 ---------------------------------------------------------------------
@cindex other system documentation, core words
@cindex core words, other system documentation

@table @i
@item nonstandard words using @code{PAD}:
@cindex @code{PAD} use by nonstandard words
None.

@item operator's terminal facilities available:
@cindex 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:
@cindex program data space available
@cindex data space available
@code{UNUSED .} gives the remaining dictionary space. The total
dictionary space can be specified with the @code{-m} switch
(@pxref{Invoking Gforth}) when Gforth starts up.

@item return stack space available:
@cindex return stack space available
You can compute the total return stack space in cells with
@code{s" RETURN-STACK-CELLS" environment? drop .}. You can specify it at
startup time with the @code{-r} switch (@pxref{Invoking Gforth}).

@item stack space available:
@cindex stack space available
You can compute the total data stack space in cells with
@code{s" STACK-CELLS" environment? drop .}. You can specify it at
startup time with the @code{-d} switch (@pxref{Invoking Gforth}).

@item system dictionary space required, in address units:
@cindex system dictionary space required, in address units
Type @code{here forthstart - .} after startup. At the time of this
writing, this gives 80080 (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 =====================================================================
@cindex system documentation, block words
@cindex block words, system documentation

@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 ---------------------------------------------------------------------
@cindex implementation-defined options, block words
@cindex block words, implementation-defined options

@table @i
@item the format for display by @code{LIST}:
@cindex @code{LIST} display format
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{\}:
@cindex length of a line affected by @code{\}
@cindex @code{\}, line length in blocks
64 characters.
@end table


@c ---------------------------------------------------------------------
@node block-ambcond, block-other, block-idef, The optional Block word set
@subsection Ambiguous conditions
@c ---------------------------------------------------------------------
@cindex block words, ambiguous conditions
@cindex ambiguous conditions, block words

@table @i
@item correct block read was not possible:
@cindex block read 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:
@cindex I/O exception in block transfer
@cindex block transfer, I/O exception
Typically results in a @code{throw} of some OS-derived value (between
-512 and -2048).

@item invalid block number:
@cindex invalid block number
@cindex block number invalid
@code{-35 throw} (Invalid block number)

@item a program directly alters the contents of @code{BLK}:
@cindex @code{BLK}, altering @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}:
@cindex @code{UPDATE}, no current block buffer
@code{UPDATE} has no effect.

@end table

@c ---------------------------------------------------------------------
@node block-other,  , block-ambcond, The optional Block word set
@subsection Other system documentation
@c ---------------------------------------------------------------------
@cindex other system documentation, block words
@cindex block words, other system documentation

@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 =====================================================================
@cindex system documentation, double words
@cindex double words, system documentation

@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 ---------------------------------------------------------------------
@cindex double words, ambiguous conditions
@cindex ambiguous conditions, double words

@table @i
@item @i{d} outside of range of @i{n} in @code{D>S}:
@cindex @code{D>S}, @i{d} out of range of @i{n} 
The least significant cell of @i{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 =====================================================================
@cindex system documentation, exception words
@cindex exception words, system documentation

@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 ---------------------------------------------------------------------
@cindex implementation-defined options, exception words
@cindex exception words, implementation-defined options

@table @i
@item @code{THROW}-codes used in the system:
@cindex @code{THROW}-codes used in the system
The codes -256@minus{}-511 are used for reporting signals. The mapping
from OS signal numbers to throw codes is -256@minus{}@i{signal}. The
codes -512@minus{}-2047 are used for OS errors (for file and memory
allocation operations). The mapping from OS error numbers to throw codes
is -512@minus{}@code{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 =====================================================================
@cindex system documentation, facility words
@cindex facility words, system documentation

@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 ---------------------------------------------------------------------
@cindex implementation-defined options, facility words
@cindex facility words, implementation-defined options

@table @i
@item encoding of keyboard events (@code{EKEY}):
@cindex keyboard events, encoding in @code{EKEY}
@cindex @code{EKEY}, encoding of keyboard events
Not yet implemented.

@item duration of a system clock tick:
@cindex duration of a system clock tick
@cindex clock tick duration
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}:
@cindex repeatability to be expected from the execution of @code{MS}
@cindex @code{MS}, repeatability to be expected
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 ---------------------------------------------------------------------
@cindex facility words, ambiguous conditions
@cindex ambiguous conditions, facility words

@table @i
@item @code{AT-XY} can't be performed on user output device:
@cindex @code{AT-XY} can't be performed on user output device
Largely terminal dependent. 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 =====================================================================
@cindex system documentation, file words
@cindex file words, system documentation

@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 ---------------------------------------------------------------------
@cindex implementation-defined options, file words
@cindex file words, implementation-defined options

@table @i
@item file access methods used:
@cindex 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 (with both @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:
@cindex 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:
@cindex 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:
@cindex file name format
System dependent. Gforth just uses the file name format of your OS.

@item information returned by @code{FILE-STATUS}:
@cindex @code{FILE-STATUS}, returned information
@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 returned mode.

@item input file state after an exception when including source:
@cindex exception when including source
All files that are left via the exception are closed.

@item @i{ior} values and meaning:
@cindex @i{ior} values and meaning
The @i{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
@i{ior}s is -512@minus{}@i{errno}.

@item maximum depth of file input nesting:
@cindex maximum depth of file input nesting
@cindex file input nesting, maximum depth
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:
@cindex maximum size of input line
@cindex input line size, maximum
@code{/line}. Currently 255.

@item methods of mapping block ranges to files:
@cindex mapping block ranges to files
@cindex files containing blocks
@cindex blocks in files
By default, blocks are accessed in the file @file{blocks.fb} in the
current working directory. The file can be switched with @code{USE}.

@item number of string buffers provided by @code{S"}:
@cindex @code{S"}, number of string buffers
1

@item size of string buffer used by @code{S"}:
@cindex @code{S"}, size of string buffer
@code{/line}. currently 255.

@end table

@c ---------------------------------------------------------------------
@node file-ambcond,  , file-idef, The optional File-Access word set
@subsection Ambiguous conditions
@c ---------------------------------------------------------------------
@cindex file words, ambiguous conditions
@cindex ambiguous conditions, file words

@table @i
@item attempting to position a file outside its boundaries:
@cindex @code{REPOSITION-FILE}, outside the file'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:
@cindex reading from file positions not yet written
End-of-file, i.e., zero characters are read and no error is reported.

@item @i{file-id} is invalid (@code{INCLUDE-FILE}):
@cindex @code{INCLUDE-FILE}, @i{file-id} is invalid 
An appropriate exception may be thrown, but a memory fault or other
problem is more probable.

@item I/O exception reading or closing @i{file-id} (@code{INCLUDE-FILE}, @code{INCLUDED}):
@cindex @code{INCLUDE-FILE}, I/O exception reading or closing @i{file-id}
@cindex @code{INCLUDED}, I/O exception reading or closing @i{file-id}
The @i{ior} produced by the operation, that discovered the problem, is
thrown.

@item named file cannot be opened (@code{INCLUDED}):
@cindex @code{INCLUDED}, named file cannot be opened
The @i{ior} produced by @code{open-file} is thrown.

@item requesting an unmapped block number:
@cindex unmapped block numbers
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:
@cindex @code{SOURCE-ID}, behaviour 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 =====================================================================
@cindex system documentation, floating-point words
@cindex floating-point words, system documentation

@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 ---------------------------------------------------------------------
@cindex implementation-defined options, floating-point words
@cindex floating-point words, implementation-defined options

@table @i
@item format and range of floating point numbers:
@cindex format and range of floating point numbers
@cindex floating point numbers, format and range
System-dependent; the @code{double} type of C.

@item results of @code{REPRESENT} when @i{float} is out of range:
@cindex  @code{REPRESENT}, results when @i{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:
@cindex rounding of floating-point numbers
@cindex truncation of floating-point numbers
@cindex floating-point numbers, rounding or truncation
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:
@cindex floating-point stack size
@code{s" FLOATING-STACK" environment? drop .} gives the total size of
the floating-point stack (in floats). You can specify this on startup
with the command-line option @code{-f} (@pxref{Invoking Gforth}).

@item width of floating-point stack:
@cindex floating-point stack width 
@code{1 floats}.

@end table


@c ---------------------------------------------------------------------
@node floating-ambcond,  , floating-idef, The optional Floating-Point word set
@subsection Ambiguous conditions
@c ---------------------------------------------------------------------
@cindex floating-point words, ambiguous conditions
@cindex ambiguous conditions, floating-point words

@table @i
@item @code{df@@} or @code{df!} used with an address that is not double-float  aligned:
@cindex @code{df@@} or @code{df!} used with an address that is not double-float  aligned
System-dependent. Typically results in a @code{-23 THROW} like other
alignment violations.

@item @code{f@@} or @code{f!} used with an address that is not float  aligned:
@cindex @code{f@@} used with an address that is not float aligned
@cindex @code{f!} used with an address that is not float aligned
System-dependent. Typically results in a @code{-23 THROW} like other
alignment violations.

@item floating-point result out of range:
@cindex 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:
@cindex @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 @code{base} is not decimal (@code{REPRESENT}, @code{F.}, @code{FE.}, @code{FS.}):
@cindex @code{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}):
@cindex @code{FATAN2}, both arguments are equal to zero
System-dependent. @code{FATAN2} is implemented using the C library
function @code{atan2()}.

@item Using @code{FTAN} on an argument @i{r1} where cos(@i{r1}) is zero:
@cindex @code{FTAN} on an argument @i{r1} where cos(@i{r1}) is zero
System-dependent. Anyway, typically the cos of @i{r1} will not be zero
because of small errors and the tan will be a very large (or very small)
but finite number.

@item @i{d} cannot be presented precisely as a float in @code{D>F}:
@cindex @code{D>F}, @i{d} cannot be presented precisely as a float
The result is rounded to the nearest float.

@item dividing by zero:
@cindex dividing by zero, floating-point
@cindex floating-point dividing by zero
@cindex floating-point unidentified fault, FP divide-by-zero
@code{-55 throw} (Floating-point unidentified fault)

@item exponent too big for conversion (@code{DF!}, @code{DF@@}, @code{SF!}, @code{SF@@}):
@cindex 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 @i{float}<1 (@code{FACOSH}):
@cindex @code{FACOSH}, @i{float}<1
@cindex floating-point unidentified fault, @code{FACOSH}
@code{-55 throw} (Floating-point unidentified fault)

@item @i{float}=<-1 (@code{FLNP1}):
@cindex @code{FLNP1}, @i{float}=<-1
@cindex floating-point unidentified fault, @code{FLNP1}
@code{-55 throw} (Floating-point unidentified fault). On IEEE-FP systems
negative infinity is typically produced for @i{float}=-1.

@item @i{float}=<0 (@code{FLN}, @code{FLOG}):
@cindex @code{FLN}, @i{float}=<0
@cindex @code{FLOG}, @i{float}=<0
@cindex floating-point unidentified fault, @code{FLN} or @code{FLOG}
@code{-55 throw} (Floating-point unidentified fault). On IEEE-FP systems
negative infinity is typically produced for @i{float}=0.

@item @i{float}<0 (@code{FASINH}, @code{FSQRT}):
@cindex @code{FASINH}, @i{float}<0
@cindex @code{FSQRT}, @i{float}<0
@cindex floating-point unidentified fault, @code{FASINH} or @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 |@i{float}|>1 (@code{FACOS}, @code{FASIN}, @code{FATANH}):
@cindex @code{FACOS}, |@i{float}|>1
@cindex @code{FASIN}, |@i{float}|>1
@cindex @code{FATANH}, |@i{float}|>1
@cindex floating-point unidentified fault, @code{FACOS}, @code{FASIN} or @code{FATANH}
@code{-55 throw} (Floating-point unidentified fault).

@item integer part of float cannot be represented by @i{d} in @code{F>D}:
@cindex @code{F>D}, integer part of float cannot be represented by @i{d}
@cindex floating-point unidentified fault, @code{F>D}
@code{-55 throw} (Floating-point unidentified fault).

@item string larger than pictured numeric output area (@code{f.}, @code{fe.}, @code{fs.}):
@cindex 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 =====================================================================
@cindex system documentation, locals words
@cindex locals words, system documentation

@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 ---------------------------------------------------------------------
@cindex implementation-defined options, locals words
@cindex locals words, implementation-defined options

@table @i
@item maximum number of locals in a definition:
@cindex maximum number of locals in a definition
@cindex locals, maximum number 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 ---------------------------------------------------------------------
@cindex locals words, ambiguous conditions
@cindex ambiguous conditions, locals words

@table @i
@item executing a named local in interpretation state:
@cindex local in interpretation state
@cindex Interpreting a compile-only word, for a local
Locals have no interpretation semantics. If you try to perform the
interpretation semantics, you will get a @code{-14 throw} somewhere
(Interpreting a compile-only word). If you perform the compilation
semantics, the locals access will be compiled (irrespective of state).

@item @i{name} not defined by @code{VALUE} or @code{(LOCAL)} (@code{TO}):
@cindex name not defined by @code{VALUE} or @code{(LOCAL)} used by @code{TO}
@cindex @code{TO} on non-@code{VALUE}s and non-locals
@cindex Invalid name argument, @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 =====================================================================
@cindex system documentation, memory-allocation words
@cindex memory-allocation words, system documentation

@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 ---------------------------------------------------------------------
@cindex implementation-defined options, memory-allocation words
@cindex memory-allocation words, implementation-defined options

@table @i
@item values and meaning of @i{ior}:
@cindex  @i{ior} values and meaning
The @i{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
@i{ior}s is -512@minus{}@i{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 =====================================================================
@cindex system documentation, programming-tools words
@cindex programming-tools words, system documentation

@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 ---------------------------------------------------------------------
@cindex implementation-defined options, programming-tools words
@cindex programming-tools words, implementation-defined options

@table @i
@item ending sequence for input following @code{;CODE} and @code{CODE}:
@cindex @code{;CODE} ending sequence
@cindex @code{CODE} ending sequence
@code{END-CODE}

@item manner of processing input following @code{;CODE} and @code{CODE}:
@cindex @code{;CODE}, processing input
@cindex @code{CODE}, processing input
The @code{ASSEMBLER} vocabulary is pushed on the search order stack, and
the input is processed by the text interpreter, (starting) in interpret
state.

@item search order capability for @code{EDITOR} and @code{ASSEMBLER}:
@cindex @code{ASSEMBLER}, search order capability
The ANS Forth search order word set.

@item source and format of display by @code{SEE}:
@cindex @code{SEE}, source and format of output
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 ---------------------------------------------------------------------
@cindex programming-tools words, ambiguous conditions
@cindex ambiguous conditions, programming-tools words

@table @i

@item deleting the compilation word list (@code{FORGET}):
@cindex @code{FORGET}, deleting the compilation word list
Not implemented (yet).

@item fewer than @i{u}+1 items on the control-flow stack (@code{CS-PICK}, @code{CS-ROLL}):
@cindex @code{CS-PICK}, fewer than @i{u}+1 items on the control flow-stack
@cindex @code{CS-ROLL}, fewer than @i{u}+1 items on the control flow-stack
@cindex control-flow stack underflow
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 @i{name} can't be found (@code{FORGET}):
@cindex @code{FORGET}, @i{name} can't be found
Not implemented (yet).

@item @i{name} not defined via @code{CREATE}:
@cindex @code{;CODE}, @i{name} not defined via @code{CREATE}
@code{;CODE} behaves like @code{DOES>} in this respect, i.e., it changes
the execution semantics of the last defined word no matter how it was
defined.

@item @code{POSTPONE} applied to @code{[IF]}:
@cindex @code{POSTPONE} applied to @code{[IF]}
@cindex @code{[IF]} and @code{POSTPONE}
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]}:
@cindex @code{[IF]}, 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}):
@cindex @code{FORGET}, removing a needed definition
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 =====================================================================
@cindex system documentation, search-order words
@cindex search-order words, system documentation

@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 ---------------------------------------------------------------------
@cindex implementation-defined options, search-order words
@cindex search-order words, implementation-defined options

@table @i
@item maximum number of word lists in search order:
@cindex maximum number of word lists in search order
@cindex search order, maximum depth
@code{s" wordlists" environment? drop .}. Currently 16.

@item minimum search order:
@cindex minimum search order
@cindex search order, minimum
@code{root root}.

@end table

@c ---------------------------------------------------------------------
@node search-ambcond,  , search-idef, The optional Search-Order word set
@subsection Ambiguous conditions
@c ---------------------------------------------------------------------
@cindex search-order words, ambiguous conditions
@cindex ambiguous conditions, search-order words

@table @i
@item changing the compilation word list (during compilation):
@cindex changing the compilation word list (during compilation)
@cindex compilation word list, change before definition ends
The word is entered into the word list that was the compilation word list
at the start of the definition. Any changes to the name field (e.g.,
@code{immediate}) or the code field (e.g., when executing @code{DOES>})
are applied to the latest defined word (as reported by @code{last} or
@code{lastxt}), if possible, irrespective of the compilation word list.

@item search order empty (@code{previous}):
@cindex @code{previous}, search order empty
@cindex vocstack empty, @code{previous}
@code{abort" Vocstack empty"}.

@item too many word lists in search order (@code{also}):
@cindex @code{also}, too many word lists in search order
@cindex vocstack full, @code{also}
@code{abort" Vocstack full"}.

@end table

@c ***************************************************************
@node Model, Integrating Gforth, ANS conformance, Top
@chapter Model

This chapter has yet to be written. It will contain information, on
which internal structures you can rely.

@c ***************************************************************
@node Integrating Gforth, Emacs and Gforth, Model, Top
@chapter Integrating Gforth into C programs

This is not yet implemented.

Several people like to use Forth as scripting language for applications
that are otherwise written in C, C++, or some other language.

The Forth system ATLAST provides facilities for embedding it into
applications; unfortunately it has several disadvantages: most
importantly, it is not based on ANS Forth, and it is apparently dead
(i.e., not developed further and not supported). The facilities
provided by Gforth in this area are inspired by ATLAST's facilities, so
making the switch should not be hard.

We also tried to design the interface such that it can easily be
implemented by other Forth systems, so that we may one day arrive at a
standardized interface. Such a standard interface would allow you to
replace the Forth system without having to rewrite C code.

You embed the Gforth interpreter by linking with the library
@code{libgforth.a} (give the compiler the option @code{-lgforth}).  All
global symbols in this library that belong to the interface, have the
prefix @code{forth_}. (Global symbols that are used internally have the
prefix @code{gforth_}).

You can include the declarations of Forth types and the functions and
variables of the interface with @code{#include <forth.h>}.

Types.

Variables.

Data and FP Stack pointer. Area sizes.

functions.

forth_init(imagefile)
forth_evaluate(string) exceptions?
forth_goto(address) (or forth_execute(xt)?)
forth_continue() (a corountining mechanism)

Adding primitives.

No checking.

Signals?

Accessing the Stacks

@c ******************************************************************
@node Emacs and Gforth, Image Files, Integrating Gforth, Top
@chapter Emacs and Gforth
@cindex Emacs and Gforth

@cindex @file{gforth.el}
@cindex @file{forth.el}
@cindex Rydqvist, Goran
@cindex comment editing commands
@cindex @code{\}, editing with Emacs
@cindex debug tracer editing commands
@cindex @code{~~}, removal with Emacs
@cindex Forth mode in Emacs
Gforth comes with @file{gforth.el}, an improved version of
@file{forth.el} by Goran Rydqvist (included in the TILE package). The
improvements are:

@itemize @bullet
@item
A better (but still not perfect) handling of indentation.
@item
Comment paragraph filling (@kbd{M-q})
@item
Commenting (@kbd{C-x \}) and uncommenting (@kbd{C-u C-x \}) of regions
@item
Removal of debugging tracers (@kbd{C-x ~}, @pxref{Debugging}).
@end itemize

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}.

@cindex source location of error or debugging output in Emacs
@cindex error output, finding the source location in Emacs
@cindex debugging output, finding the source location in Emacs
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).

@cindex @file{TAGS} file
@cindex @file{etags.fs}
@cindex viewing the source of a word in Emacs
Also, if you @code{include} @file{etags.fs}, a new @file{TAGS} file will
be produced (@pxref{Tags, , Tags Tables, emacs, Emacs Manual}) 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/0.2.0/TAGS}).

@cindex @file{.emacs}
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

@c ******************************************************************
@node Image Files, Engine, Emacs and Gforth, Top
@chapter Image Files
@cindex image file
@cindex @file{.fi} files
@cindex precompiled Forth code
@cindex dictionary in persistent form
@cindex persistent form of dictionary

An image file is a file containing an image of the Forth dictionary,
i.e., compiled Forth code and data residing in the dictionary.  By
convention, we use the extension @code{.fi} for image files.

@menu
* Image Licensing Issues::      Distribution terms for images.
* Image File Background::       Why have image files?
* Non-Relocatable Image Files::   don't always work.
* Data-Relocatable Image Files::  are better.
* Fully Relocatable Image Files:: better yet.
* Stack and Dictionary Sizes::  Setting the default sizes for an image.
* Running Image Files::         @code{gforth -i @i{file}} or @i{file}.
* Modifying the Startup Sequence::  and turnkey applications.
@end menu

@node Image Licensing Issues, Image File Background, Image Files, Image Files
@section Image Licensing Issues
@cindex license for images
@cindex image license

An image created with @code{gforthmi} (@pxref{gforthmi}) or
@code{savesystem} (@pxref{Non-Relocatable Image Files}) includes the
original image; i.e., according to copyright law it is a derived work of
the original image.

Since Gforth is distributed under the GNU GPL, the newly created image
falls under the GNU GPL, too. In particular, this means that if you
distribute the image, you have to make all of the sources for the image
available, including those you wrote.  For details see @ref{License, ,
GNU General Public License (Section 3)}.

If you create an image with @code{cross} (@pxref{cross.fs}), the image
contains only code compiled from the sources you gave it; if none of
these sources is under the GPL, the terms discussed above do not apply
to the image. However, if your image needs an engine (a gforth binary)
that is under the GPL, you should make sure that you distribute both in
a way that is at most a @emph{mere aggregation}, if you don't want the
terms of the GPL to apply to the image.

@node Image File Background, Non-Relocatable Image Files, Image Licensing Issues, Image Files
@section Image File Background
@cindex image file background

Our Forth system consists not only of primitives, but also of
definitions written in Forth. Since the Forth compiler itself belongs to
those definitions, it is not possible to start the system with the
primitives and the Forth source alone. Therefore we provide the Forth
code as an image file in nearly executable form. When Gforth starts up,
a C routine loads the image file into memory, optionally relocates the
addresses, then sets up the memory (stacks etc.) according to
information in the image file, and (finally) starts executing Forth
code.

The image file variants represent different compromises between the
goals of making it easy to generate image files and making them
portable.

@cindex relocation at run-time
Win32Forth 3.4 and Mitch Bradley's @code{cforth} use relocation at
run-time. This avoids many of the complications discussed below (image
files are data relocatable without further ado), but costs performance
(one addition per memory access).

@cindex relocation at load-time
By contrast, the Gforth loader performs relocation at image load time. The
loader also has to replace tokens that represent primitive calls with the
appropriate code-field addresses (or code addresses in the case of
direct threading).

There are three kinds of image files, with different degrees of
relocatability: non-relocatable, data-relocatable, and fully relocatable
image files.

@cindex image file loader
@cindex relocating loader
@cindex loader for image files
These image file variants have several restrictions in common; they are
caused by the design of the image file loader:

@itemize @bullet
@item
There is only one segment; in particular, this means, that an image file
cannot represent @code{ALLOCATE}d memory chunks (and pointers to
them). The contents of the stacks are not represented, either.

@item
The only kinds of relocation supported are: adding the same offset to
all cells that represent data addresses; and replacing special tokens
with code addresses or with pieces of machine code.

If any complex computations involving addresses are performed, the
results cannot be represented in the image file. Several applications that
use such computations come to mind:
@itemize @minus
@item
Hashing addresses (or data structures which contain addresses) for table
lookup. If you use Gforth's @code{table}s or @code{wordlist}s for this
purpose, you will have no problem, because the hash tables are
recomputed automatically when the system is started. If you use your own
hash tables, you will have to do something similar.

@item
There's a cute implementation of doubly-linked lists that uses
@code{XOR}ed addresses. You could represent such lists as singly-linked
in the image file, and restore the doubly-linked representation on
startup.@footnote{In my opinion, though, you should think thrice before
using a doubly-linked list (whatever implementation).}

@item
The code addresses of run-time routines like @code{docol:} cannot be
represented in the image file (because their tokens would be replaced by
machine code in direct threaded implementations). As a workaround,
compute these addresses at run-time with @code{>code-address} from the
executions tokens of appropriate words (see the definitions of
@code{docol:} and friends in @file{kernel.fs}).

@item
On many architectures addresses are represented in machine code in some
shifted or mangled form. You cannot put @code{CODE} words that contain
absolute addresses in this form in a relocatable image file. Workarounds
are representing the address in some relative form (e.g., relative to
the CFA, which is present in some register), or loading the address from
a place where it is stored in a non-mangled form.
@end itemize
@end itemize

@node  Non-Relocatable Image Files, Data-Relocatable Image Files, Image File Background, Image Files
@section Non-Relocatable Image Files
@cindex non-relocatable image files
@cindex image file, non-relocatable

These files are simple memory dumps of the dictionary. They are specific
to the executable (i.e., @file{gforth} file) they were created
with. What's worse, they are specific to the place on which the
dictionary resided when the image was created. Now, there is no
guarantee that the dictionary will reside at the same place the next
time you start Gforth, so there's no guarantee that a non-relocatable
image will work the next time (Gforth will complain instead of crashing,
though).

You can create a non-relocatable image file with

doc-savesystem

@node Data-Relocatable Image Files, Fully Relocatable Image Files, Non-Relocatable Image Files, Image Files
@section Data-Relocatable Image Files
@cindex data-relocatable image files
@cindex image file, data-relocatable

These files contain relocatable data addresses, but fixed code addresses
(instead of tokens). They are specific to the executable (i.e.,
@file{gforth} file) they were created with. For direct threading on some
architectures (e.g., the i386), data-relocatable images do not work. You
get a data-relocatable image, if you use @file{gforthmi} with a
Gforth binary that is not doubly indirect threaded (@pxref{Fully
Relocatable Image Files}).

@node Fully Relocatable Image Files, Stack and Dictionary Sizes, Data-Relocatable Image Files, Image Files
@section Fully Relocatable Image Files
@cindex fully relocatable image files
@cindex image file, fully relocatable

@cindex @file{kern*.fi}, relocatability
@cindex @file{gforth.fi}, relocatability
These image files have relocatable data addresses, and tokens for code
addresses. They can be used with different binaries (e.g., with and
without debugging) on the same machine, and even across machines with
the same data formats (byte order, cell size, floating point
format). However, they are usually specific to the version of Gforth
they were created with. The files @file{gforth.fi} and @file{kernl*.fi}
are fully relocatable.

There are two ways to create a fully relocatable image file:

@menu
* gforthmi::                    The normal way
* cross.fs::                    The hard way
@end menu

@node gforthmi, cross.fs, Fully Relocatable Image Files, Fully Relocatable Image Files
@subsection @file{gforthmi}
@cindex @file{comp-i.fs}
@cindex @file{gforthmi}

You will usually use @file{gforthmi}. If you want to create an
image @i{file} that contains everything you would load by invoking
Gforth with @code{gforth @i{options}}, you simply say:
@example
gforthmi @i{file} @i{options}
@end example

E.g., if you want to create an image @file{asm.fi} that has the file
@file{asm.fs} loaded in addition to the usual stuff, you could do it
like this:

@example
gforthmi asm.fi asm.fs
@end example

@file{gforthmi} is implemented as a sh script and works like this: It
produces two non-relocatable images for different addresses and then
compares them. Its output reflects this: first you see the output (if
any) of the two Gforth invocations that produce the nonrelocatable image
files, then you see the output of the comparing program: It displays the
offset used for data addresses and the offset used for code addresses;
moreover, for each cell that cannot be represented correctly in the
image files, it displays a line like the following one:

@example
     78DC         BFFFFA50         BFFFFA40
@end example

This means that at offset $78dc from @code{forthstart}, one input image
contains $bffffa50, and the other contains $bffffa40. Since these cells
cannot be represented correctly in the output image, you should examine
these places in the dictionary and verify that these cells are dead
(i.e., not read before they are written).

If you type @file{gforthmi} with no arguments, it prints some usage
instructions.

@cindex @code{savesystem} during @file{gforthmi}
@cindex @code{bye} during @file{gforthmi}
@cindex doubly indirect threaded code
@cindex environment variable @code{GFORTHD}
@cindex @code{GFORTHD} environment variable
@cindex @code{gforth-ditc}
There are a few wrinkles: After processing the passed @i{options}, the
words @code{savesystem} and @code{bye} must be visible. A special doubly
indirect threaded version of the @file{gforth} executable is used for
creating the nonrelocatable images; you can pass the exact filename of
this executable through the environment variable @code{GFORTHD}
(default: @file{gforth-ditc}); if you pass a version that is not doubly
indirect threaded, you will not get a fully relocatable image, but a
data-relocatable image (because there is no code address offset). The
normal @file{gforth} executable is used for creating the relocatable
image; you can pass the exact filename of this executable through the
environment variable @code{GFORTH}.

@node cross.fs,  , gforthmi, Fully Relocatable Image Files
@subsection @file{cross.fs}
@cindex @file{cross.fs}
@cindex cross-compiler
@cindex metacompiler

You can also use @code{cross}, a batch compiler that accepts a Forth-like
programming language. This @code{cross} language has to be documented
yet.

@cindex target compiler
@code{cross} also allows you to create image files for machines with
different data sizes and data formats than the one used for generating
the image file. You can also use it to create an application image that
does not contain a Forth compiler. These features are bought with
restrictions and inconveniences in programming. E.g., addresses have to
be stored in memory with special words (@code{A!}, @code{A,}, etc.) in
order to make the code relocatable.


@node Stack and Dictionary Sizes, Running Image Files, Fully Relocatable Image Files, Image Files
@section Stack and Dictionary Sizes
@cindex image file, stack and dictionary sizes
@cindex dictionary size default
@cindex stack size default

If you invoke Gforth with a command line flag for the size
(@pxref{Invoking Gforth}), the size you specify is stored in the
dictionary. If you save the dictionary with @code{savesystem} or create
an image with @file{gforthmi}, this size will become the default
for the resulting image file. E.g., the following will create a
fully relocatable version of @file{gforth.fi} with a 1MB dictionary:

@example
gforthmi gforth.fi -m 1M
@end example

In other words, if you want to set the default size for the dictionary
and the stacks of an image, just invoke @file{gforthmi} with the
appropriate options when creating the image.

@cindex stack size, cache-friendly
Note: For cache-friendly behaviour (i.e., good performance), you should
make the sizes of the stacks modulo, say, 2K, somewhat different. E.g.,
the default stack sizes are: data: 16k (mod 2k=0); fp: 15.5k (mod
2k=1.5k); return: 15k(mod 2k=1k); locals: 14.5k (mod 2k=0.5k).

@node Running Image Files, Modifying the Startup Sequence, Stack and Dictionary Sizes, Image Files
@section Running Image Files
@cindex running image files
@cindex invoking image files
@cindex image file invocation

@cindex -i, invoke image file
@cindex --image file, invoke image file
You can invoke Gforth with an image file @i{image} instead of the
default @file{gforth.fi} with the @code{-i} flag (@pxref{Invoking Gforth}):
@example
gforth -i @i{image}
@end example

@cindex executable image file
@cindex image file, executable
If your operating system supports starting scripts with a line of the
form @code{#! ...}, you just have to type the image file name to start
Gforth with this image file (note that the file extension @code{.fi} is
just a convention). I.e., to run Gforth with the image file @i{image},
you can just type @i{image} instead of @code{gforth -i @i{image}}.
This works because every @code{.fi} file starts with a line of this
format:

@example
#! /usr/local/bin/gforth-0.4.0 -i
@end example

The file and pathname for the Gforth engine specified on this line is
the specific Gforth executable that it was built against; i.e. the value
of the environment variable @code{GFORTH} at the time that
@file{gforthmi} was executed.

You can make use of the same shell capability to make a Forth source
file into an executable. For example, if you place this text in a file:

@example
#! /usr/local/bin/gforth

." Hello, world" CR
bye
@end example

@noindent
and then make the file executable (chmod +x in Unix), you can run it
directly from the command line. The sequence @code{#!} is used in two
ways; firstly, it is recognised as a ``magic sequence'' by the operating
system@footnote{The Unix kernel actually recognises two types of files:
executable files and files of data, where the data is processed by an
interpreter that is specified on the ``interpreter line'' -- the first
line of the file, starting with the sequence #!. There may be a small
limit (e.g., 32) on the number of characters that may be specified on
the interpreter line.} secondly it is treated as a comment character by
Gforth. Because of the second usage, a space is required between
@code{#!} and the path to the executable.

The disadvantage of this latter technique, compared with using
@file{gforthmi}, is that it is slower; the Forth source code is compiled
on-the-fly, each time the program is invoked.

@comment TODO describe the #! magic with reference to the Power Tools book.

doc-#!

@node Modifying the Startup Sequence,  , Running Image Files, Image Files
@section Modifying the Startup Sequence
@cindex startup sequence for image file
@cindex image file initialization sequence
@cindex initialization sequence of image file

You can add your own initialization to the startup sequence through the
deferred word @code{'cold}. @code{'cold} is invoked just before the
image-specific command line processing (by default, loading files and
evaluating (@code{-e}) strings) starts.

A sequence for adding your initialization usually looks like this:

@example
:noname
    Defers 'cold \ do other initialization stuff (e.g., rehashing wordlists)
    ... \ your stuff
; IS 'cold
@end example

@cindex turnkey image files
@cindex image file, turnkey applications
You can make a turnkey image by letting @code{'cold} execute a word
(your turnkey application) that never returns; instead, it exits Gforth
via @code{bye} or @code{throw}.

@cindex command-line arguments, access
@cindex arguments on the command line, access
You can access the (image-specific) command-line arguments through the
variables @code{argc} and @code{argv}. @code{arg} provides convenient
access to @code{argv}.

If @code{'cold} exits normally, Gforth processes the command-line
arguments as files to be loaded and strings to be evaluated.  Therefore,
@code{'cold} should remove the arguments it has used in this case.

doc-'cold
doc-argc
doc-argv
doc-arg


@c ******************************************************************
@node Engine, Binding to System Library, Image Files, Top
@chapter Engine
@cindex engine
@cindex virtual machine

Reading this chapter is not necessary for programming with Gforth. It
may 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
@*@url{http://www.complang.tuwien.ac.at/papers/ertl93.ps.Z}.

@menu
* Portability::                 
* Threading::                   
* Primitives::                  
* Performance::                 
@end menu

@node Portability, Threading, Engine, Engine
@section Portability
@cindex engine portability

An important goal of the Gforth Project is availability across a wide
range of personal machines. fig-Forth, and, to a lesser extent, F83,
achieved this goal by manually coding the engine in assembly language
for several then-popular processors. This approach is very
labor-intensive and the results are short-lived due to progress in
computer architecture.

@cindex C, using C for the engine
Others have avoided this problem by coding in C, e.g., Mitch Bradley
(cforth), Mikael Patel (TILE) and Dirk Zoller (pfe). This approach is
particularly popular for UNIX-based Forths due to the large variety of
architectures of UNIX machines. Unfortunately an implementation in C
does not mix well with the goals of efficiency and with using
traditional techniques: Indirect or direct threading cannot be expressed
in C, and switch threading, the fastest technique available in C, is
significantly slower. Another problem with C is that it is very
cumbersome to express double integer arithmetic.

@cindex GNU C for the engine
@cindex long long
Fortunately, there is a portable language that does not have these
limitations: GNU C, the version of C processed by the GNU C compiler
(@pxref{C Extensions, , Extensions to the C Language Family, gcc.info,
GNU C Manual}). Its labels as values feature (@pxref{Labels as Values, ,
Labels as Values, gcc.info, GNU C Manual}) makes direct and indirect
threading possible, its @code{long long} type (@pxref{Long Long, ,
Double-Word Integers, gcc.info, GNU C Manual}) corresponds to Forth's
double numbers@footnote{Unfortunately, long longs are not implemented
properly on all machines (e.g., on alpha-osf1, long longs are only 64
bits, the same size as longs (and pointers), but they should be twice as
long according to @pxref{Long Long, , Double-Word Integers, gcc.info, GNU
C Manual}). So, we had to implement doubles in C after all. Still, on
most machines we can use long longs and achieve better performance than
with the emulation package.}. GNU C is available for free on all
important (and many 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 machines.

Writing in a portable language has the reputation of producing code that
is slower than assembly. For our Forth engine we repeatedly looked at
the code produced by the compiler and eliminated most compiler-induced
inefficiencies by appropriate changes in the source code.

@cindex explicit register declarations
@cindex --enable-force-reg, configuration flag
@cindex -DFORCE_REG
However, register allocation cannot be portably influenced by the
programmer, leading to some inefficiencies on register-starved
machines. We use explicit register declarations (@pxref{Explicit Reg
Vars, , Variables in Specified Registers, gcc.info, GNU C Manual}) to
improve the speed on some machines. They are turned on by using the
configuration flag @code{--enable-force-reg} (@code{gcc} switch
@code{-DFORCE_REG}). Unfortunately, this feature not only depends on the
machine, but also on the compiler version: On some machines some
compiler versions produce incorrect code when certain explicit register
declarations are used. So by default @code{-DFORCE_REG} is not used.

@node Threading, Primitives, Portability, Engine
@section Threading
@cindex inner interpreter implementation
@cindex threaded code implementation

@cindex labels as values
GNU C's labels as values extension (available since @code{gcc-2.0},
@pxref{Labels as Values, , Labels as Values, gcc.info, GNU C Manual})
makes it possible to take the address of @i{label} by writing
@code{&&@i{label}}.  This address can then be used in a statement like
@code{goto *@i{address}}. I.e., @code{goto *&&x} is the same as
@code{goto x}.

@cindex @code{NEXT}, indirect threaded
@cindex indirect threaded inner interpreter
@cindex inner interpreter, indirect threaded
With this feature an indirect threaded @code{NEXT} looks like:
@example
cfa = *ip++;
ca = *cfa;
goto *ca;
@end example
@cindex instruction pointer
For those unfamiliar with the names: @code{ip} is the Forth instruction
pointer; the @code{cfa} (code-field address) corresponds to ANS Forths
execution token and points to the code field of the next word to be
executed; The @code{ca} (code address) fetched from there points to some
executable code, e.g., a primitive or the colon definition handler
@code{docol}.

@cindex @code{NEXT}, direct threaded
@cindex direct threaded inner interpreter
@cindex inner interpreter, direct threaded
Direct threading is even simpler:
@example
ca = *ip++;
goto *ca;
@end example

Of course we have packaged the whole thing neatly in macros called
@code{NEXT} and @code{NEXT1} (the part of @code{NEXT} after fetching the cfa).

@menu
* Scheduling::                  
* Direct or Indirect Threaded?::  
* DOES>::                       
@end menu

@node Scheduling, Direct or Indirect Threaded?, Threading, Threading
@subsection Scheduling
@cindex inner interpreter optimization

There is a little complication: Pipelined and superscalar processors,
i.e., RISC and some modern CISC machines can process independent
instructions while waiting for the results of an instruction. The
compiler usually reorders (schedules) the instructions in a way that
achieves good usage of these delay slots. However, on our first tries
the compiler did not do well on scheduling primitives. E.g., for
@code{+} implemented as
@example
n=sp[0]+sp[1];
sp++;
sp[0]=n;
NEXT;
@end example
the @code{NEXT} comes strictly after the other code, i.e., there is nearly no
scheduling. After a little thought the problem becomes clear: The
compiler cannot know that @code{sp} and @code{ip} point to different
addresses (and the version of @code{gcc} we used would not know it even
if it was possible), so it could not move the load of the cfa above the
store to the TOS. Indeed the pointers could be the same, if code on or
very near the top of stack were executed. In the interest of speed we
chose to forbid this probably unused ``feature'' and helped the compiler
in scheduling: @code{NEXT} is divided into the loading part (@code{NEXT_P1})
and the goto part (@code{NEXT_P2}). @code{+} now looks like:
@example
n=sp[0]+sp[1];
sp++;
NEXT_P1;
sp[0]=n;
NEXT_P2;
@end example
This can be scheduled optimally by the compiler.

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
(e.g., the 80386), in order to preserve registers.

@node Direct or Indirect Threaded?, DOES>, Scheduling, Threading
@subsection Direct or Indirect Threaded?
@cindex threading, direct or indirect?

@cindex -DDIRECT_THREADED
Both! After packaging the nasty details in macro definitions we
realized that we could switch between direct and indirect threading by
simply setting a compilation flag (@code{-DDIRECT_THREADED}) and
defining a few machine-specific macros for the direct-threading case.
On the Forth level we also offer access words that hide the
differences between the threading methods (@pxref{Threading Words}).

Indirect threading is implemented completely machine-independently.
Direct threading needs routines for creating jumps to the executable
code (e.g. to @code{docol} or @code{dodoes}). These routines are inherently
machine-dependent, but they do not amount to many source lines. Therefore,
even porting direct threading to a new machine requires little effort.

@cindex --enable-indirect-threaded, configuration flag
@cindex --enable-direct-threaded, configuration flag
The default threading method is machine-dependent. You can enforce a
specific threading method when building Gforth with the configuration
flag @code{--enable-direct-threaded} or
@code{--enable-indirect-threaded}. Note that direct threading is not
supported on all machines.

@node DOES>,  , Direct or Indirect Threaded?, Threading
@subsection DOES>
@cindex @code{DOES>} implementation

@cindex @code{dodoes} routine
@cindex @code{DOES>}-code
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
@code{CREATE}...@code{DOES>} pair. The main problem here is: How to find
the Forth code to be executed, i.e. the code after the
@code{DOES>} (the @code{DOES>}-code)? There are two solutions:

In fig-Forth the code field points directly to the @code{dodoes} and the
@code{DOES>}code address is stored in the cell after the code address (i.e. at
@code{@i{CFA} cell+}). It may seem that this solution is illegal in
the Forth-79 and all later standards, because in fig-Forth this address
lies in the body (which is illegal in these standards). However, by
making the code field larger for all words this solution becomes legal
again. We use this approach for the indirect threaded version and for
direct threading on some machines. Leaving a cell unused in most words
is a bit wasteful, but on the machines we are targeting this is hardly a
problem. The other reason for having a code field size of two cells is
to avoid having different image files for direct and indirect threaded
systems (direct threaded systems require two-cell code fields on many
machines).

@cindex @code{DOES>}-handler
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
this address (the @code{DOES>}-handler). @code{dodoes} can then get the
@code{DOES>}-code address by computing the code address, i.e., the address of
the jump to dodoes, and add the length of that jump field. A variant of
this is to have a call to @code{dodoes} after the @code{DOES>}; then the
return address (which can be found in the return register on RISCs) is
the @code{DOES>}-code address. Since the two cells available in the code field
are used up by the jump to the code address in direct threading on many
architectures, we use this approach for direct threading on these
architectures. We did not want to add another cell to the code field.

@node Primitives, Performance, Threading, Engine
@section Primitives
@cindex primitives, implementation
@cindex virtual machine instructions, implementation

@menu
* Automatic Generation::        
* TOS Optimization::            
* Produced code::               
@end menu

@node Automatic Generation, TOS Optimization, Primitives, Primitives
@subsection Automatic Generation
@cindex primitives, automatic generation

@cindex @file{prims2x.fs}
Since the primitives are implemented in a portable language, there is no
longer any need to minimize the number of primitives. On the contrary,
having many primitives has an advantage: speed. In order to reduce the
number of errors in primitives and to make programming them easier, we
provide a tool, the primitive generator (@file{prims2x.fs}), that
automatically generates most (and sometimes all) of the C code for a
primitive from the stack effect notation.  The source for a primitive
has the following form:

@cindex primitive source format
@format
@i{Forth-name}	@i{stack-effect}	@i{category}	[@i{pronounc.}]
[@code{""}@i{glossary entry}@code{""}]
@i{C code}
[@code{:}
@i{Forth code}]
@end format

The items in brackets are optional. The category and glossary fields
are there for generating the documentation, the Forth code is there
for manual implementations on machines without GNU C. E.g., the source
for the primitive @code{+} is:
@example
+    n1 n2 -- n    core    plus
n = n1+n2;
@end example

This looks like a specification, but in fact @code{n = n1+n2} is C
code. Our primitive generation tool extracts a lot of information from
the stack effect notations@footnote{We use a one-stack notation, even
though we have separate data and floating-point stacks; The separate
notation can be generated easily from the unified notation.}: The number
of items popped from and pushed on the stack, their type, and by what
name they are referred to in the C code. It then generates a C code
prelude and postlude for each primitive. The final C code for @code{+}
looks like this:

@example
I_plus:	/* + ( n1 n2 -- n ) */  /* label, stack effect */
/*  */                          /* documentation */
@{
DEF_CA                          /* definition of variable ca (indirect threading) */
Cell n1;                        /* definitions of variables */
Cell n2;
Cell n;
n1 = (Cell) sp[1];              /* input */
n2 = (Cell) TOS;
sp += 1;                        /* stack adjustment */
NAME("+")                       /* debugging output (with -DDEBUG) */
@{
n = n1+n2;                      /* C code taken from the source */
@}
NEXT_P1;                        /* NEXT part 1 */
TOS = (Cell)n;                  /* output */
NEXT_P2;                        /* NEXT part 2 */
@}
@end example

This looks long and inefficient, but the GNU C compiler optimizes quite
well and produces optimal code for @code{+} on, e.g., the R3000 and the
HP RISC machines: Defining the @code{n}s does not produce any code, and
using them as intermediate storage also adds no cost.

There are also other optimizations that are not illustrated by this
example: assignments between simple variables are usually for free (copy
propagation). If one of the stack items is not used by the primitive
(e.g.  in @code{drop}), the compiler eliminates the load from the stack
(dead code elimination). On the other hand, there are some things that
the compiler does not do, therefore they are performed by
@file{prims2x.fs}: The compiler does not optimize code away that stores
a stack item to the place where it just came from (e.g., @code{over}).

While programming a primitive is usually easy, there are a few cases
where the programmer has to take the actions of the generator into
account, most notably @code{?dup}, but also words that do not (always)
fall through to @code{NEXT}.

@node TOS Optimization, Produced code, Automatic Generation, Primitives
@subsection TOS Optimization
@cindex TOS optimization for primitives
@cindex primitives, keeping the TOS in a register

An important optimization for stack machine emulators, e.g., Forth
engines, is keeping  one or more of the top stack items in
registers.  If a word has the stack effect @i{in1}...@i{inx} @code{--}
@i{out1}...@i{outy}, keeping the top @i{n} items in registers
@itemize @bullet
@item
is better than keeping @i{n-1} items, if @i{x>=n} and @i{y>=n},
due to fewer loads from and stores to the stack.
@item is slower than keeping @i{n-1} items, if @i{x<>y} and @i{x<n} and
@i{y<n}, due to additional moves between registers.
@end itemize

@cindex -DUSE_TOS
@cindex -DUSE_NO_TOS
In particular, keeping one item in a register is never a disadvantage,
if there are enough registers. Keeping two items in registers is a
disadvantage for frequent words like @code{?branch}, constants,
variables, literals and @code{i}. Therefore our generator only produces
code that keeps zero or one items in registers. The generated C code
covers both cases; the selection between these alternatives is made at
C-compile time using the switch @code{-DUSE_TOS}. @code{TOS} in the C
code for @code{+} is just a simple variable name in the one-item case,
otherwise it is a macro that expands into @code{sp[0]}. Note that the
GNU C compiler tries to keep simple variables like @code{TOS} in
registers, and it usually succeeds, if there are enough registers.

@cindex -DUSE_FTOS
@cindex -DUSE_NO_FTOS
The primitive generator performs the TOS optimization for the
floating-point stack, too (@code{-DUSE_FTOS}). For floating-point
operations the benefit of this optimization is even larger:
floating-point operations take quite long on most processors, but can be
performed in parallel with other operations as long as their results are
not used. If the FP-TOS is kept in a register, this works. If
it is kept on the stack, i.e., in memory, the store into memory has to
wait for the result of the floating-point operation, lengthening the
execution time of the primitive considerably.

The TOS optimization makes the automatic generation of primitives a
bit more complicated. Just replacing all occurrences of @code{sp[0]} by
@code{TOS} is not sufficient. There are some special cases to
consider:
@itemize @bullet
@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,
if the TOS optimization is turned on.
@item Primitives with stack effects of the form @code{--}
@i{out1}...@i{outy} must store the TOS to the stack at the start.
Likewise, primitives with the stack effect @i{in1}...@i{inx} @code{--}
must load the TOS from the stack at the end. But for the null stack
effect @code{--} no stores or loads should be generated.
@end itemize

@node Produced code,  , TOS Optimization, Primitives
@subsection Produced code
@cindex primitives, assembly code listing

@cindex @file{engine.s}
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
look at the resulting file @file{engine.s}.

@node  Performance,  , Primitives, Engine
@section Performance
@cindex performance of some Forth interpreters
@cindex engine performance
@cindex benchmarking Forth systems
@cindex Gforth 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}.

@cindex Win32Forth performance
@cindex NT Forth performance
@cindex eforth performance
@cindex ThisForth performance
@cindex PFE performance
@cindex TILE performance
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-0.9.14 (compiled with @code{gcc-2.6.3} with the default
configuration for Linux: @code{-O2 -fomit-frame-pointer -DUSE_REGS
-DUNROLL_NEXT}), ThisForth Beta (compiled with @code{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  1.58  3.18  8.58
bubble    1.00  1.31  1.41   1.48  0.88  1.50        3.88
matmul    1.00  1.47  1.35   1.46  0.74  1.58        4.09
fib       1.00  1.52  1.34   1.22  0.86  1.74  2.99  4.30
@end example

You may be quite surprised by the good performance of Gforth when
compared with systems written in assembly language. One important reason
for the disappointing performance of these other 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 @code{NEXT} (@pxref{Image File Background}).

Only Eforth with the peephole optimizer has a performance that is
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 of the latter systems to
standard C, which makes efficient threading impossible (however, the
measured implementation of PFE uses a GNU C extension: @pxref{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.

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.

Note that there have been several releases of Win32Forth since the
release presented here, so the results presented above may have little
predictive value for the performance of Win32Forth today.

@cindex @file{Benchres}
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 direct
threaded version used here. The paper available at
@*@url{http://www.complang.tuwien.ac.at/papers/ertl&maierhofer95.ps.gz};
it also contains numbers for some native code systems. You can find a
newer version of these measurements at
@url{http://www.complang.tuwien.ac.at/forth/performance.html}. You can
find numbers for Gforth on various machines in @file{Benchres}.

@c ******************************************************************
@node Binding to System Library, Cross Compiler, Engine, Top
@chapter Binding to System Library

@node Cross Compiler, Bugs, Binding to System Library, Top
@chapter Cross Compiler

Cross Compiler

@menu
* Using the Cross Compiler::
* How the Cross Compiler Works::
@end menu

@node Using the Cross Compiler, How the Cross Compiler Works, Cross Compiler, Cross Compiler
@section Using the Cross Compiler

@node How the Cross Compiler Works, , Using the Cross Compiler, Cross Compiler
@section How the Cross Compiler Works

@node Bugs, Origin, Cross Compiler, Top
@appendix Bugs
@cindex bug reporting

Known bugs are described in the file @file{BUGS} in the Gforth distribution.

If you find a bug, please send a bug report to
@email{bug-gforth@@gnu.org}. A bug report should include this
information:

@itemize @bullet
@item
The Gforth version used (it is announced at the start of an
interactive Gforth session).
@item
The machine and operating system (on Unix
systems @code{uname -a} will report this information).
@item
The installation options (send the file @file{config.status}).
@item
A complete list of changes (if any) you (or your installer) have made to the
Gforth sources.
@item
A program (or a sequence of keyboard commands) that reproduces the bug.
@item
A description of what you think constitutes the buggy behaviour.
@end itemize

For a thorough guide on reporting bugs read @ref{Bug Reporting, , How
to Report Bugs, gcc.info, GNU C Manual}.


@node Origin, Forth-related information, Bugs, Top
@appendix Authors and Ancestors of Gforth

@section Authors and Contributors
@cindex authors of Gforth
@cindex contributors to Gforth

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, Marcel Hendrix, John
Wavrik, Barrie Stott, Marc de Groot, and Jorge Acerada. Since the
release of Gforth-0.2.1 there were also helpful comments from many
others; thank you all, sorry for not listing you here (but digging
through my mailbox to extract your names is on my to-do list). Since the
release of Gforth-0.4.0 Neal Crook worked on the manual.

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 did not meet
physically for the first 4 years of development.

@section Pedigree
@cindex pedigree of Gforth

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.

Henry 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 developed 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 Forth-related information, Word Index, Origin, Top
@appendix Other Forth-related information
@cindex Forth-related information

@menu
* Internet resources::
* Books::
* The Forth Interest Group::
* Conferences::
@end menu


@node Internet resources, Books, Forth-related information, Forth-related information
@section Internet resources
@cindex internet resources

@cindex comp.lang.forth
@cindex frequently asked questions
There is an active newsgroup (comp.lang.forth) discussing Forth and
Forth-related issues. A frequently-asked-questions (FAQ) list
is posted to the newsgroup regulary, and archived at these sites:

@itemize @bullet
@item
@url{ftp://rtfm.mit.edu/pub/usenet-by-group/comp.lang.forth/}
@item
@url{ftp://ftp.forth.org/pub/Forth/FAQ/}
@end itemize

The FAQ list should be considered mandatory reading before posting to
the newsgroup.

Here are some other web sites holding Forth-related material:

@itemize @bullet
@item
@url{http://www.taygeta.com/forth.html} -- Skip Carter's Forth pages.
@item
@url{http://www.jwdt.com/~paysan/gforth.html} -- the Gforth home page.
@item
@url{http://www.minerva.com/uathena.htm} -- home of ANS Forth Standard.
@item
@url{http://dec.bournemouth.ac.uk/forth/index.html} -- the Forth
Research page, including links to the Journal of Forth Application and
Research (JFAR) and a searchable Forth bibliography.
@end itemize


@node Books, The Forth Interest Group, Internet resources, Forth-related information
@section Books
@cindex books on Forth

As the Standard is relatively new, there are not many books out yet. It
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
deviations of the book. However, books based on the Forth-83 standard
should be ok, because ANS Forth is primarily an extension of Forth-83.

@cindex standard document for ANS Forth
@cindex ANS Forth document
The definite reference if you want to write ANS Forth programs is, of
course, the ANS Forth document. It is available in printed form from the
National Standards Institute Sales Department (Tel.: USA (212) 642-4900;
Fax.: USA (212) 302-1286) as document @cite{X3.215-1994} for about
$200. You 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
(@url{http://www.taygeta.com/forth/dpans.html}; this is my favourite
format); this HTML version also includes the answers to Requests for
Interpretation (RFIs). Some pointers to these versions can be found
through @*@url{http://www.complang.tuwien.ac.at/projects/forth.html}.

@cindex introductory book on Forth
@cindex book on Forth, introductory
@cindex Woehr, Jack: @cite{Forth: The New Model}
@cindex @cite{Forth: The new model} (book)
@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
cover the whole standard. It also contains interesting background
information (Jack Woehr was in the ANS Forth Technical Committee). It is
not appropriate for complete newbies, but programmers experienced in
other languages should find it ok.

@cindex Conklin, Edward K., and Elizabeth Rather: @cite{Forth Programmer's Handbook}
@cindex Rather, Elizabeth and Edward K. Conklin: @cite{Forth Programmer's Handbook}
@cindex @cite{Forth Programmer's Handbook} (book)
@cite{Forth Programmer's Handbook} by Edward K. Conklin, Elizabeth
D. Rather and the technical staff of Forth, Inc. (Forth, Inc., 1997;
ISBN 0-9662156-0-5) contains little introductory material. The majority
of the book is similar to @ref{Words}, but the book covers most of the
standard words and some non-standard words (whereas this manual is
quite incomplete). In addition, the book contains a chapter on
programming style. The major drawback of this book is that it usually
does not identify what is standard and what is specific to the Forth
system described in the book (probably one of Forth, Inc.'s systems).
Fortunately, many of the non-standard programming practices described in
the book work in Gforth, too.  Still, this drawback makes the book
hardly more useful than a pre-ANS book.

@node The Forth Interest Group, Conferences, Books, Forth-related information
@section The Forth Interest Group
@cindex Forth interest group (FIG)

The Forth Interest Group (FIG) is a world-wide, non-profit,
member-supported organisation. It publishes a regular magazine,
@var{FORTH Dimensions}, and offers other benefits of membership. You can
contact the FIG through their office email address:
@email{office@@forth.org} or by visiting their web site at
@url{http://www.forth.org/}. This web site also includes links to FIG
chapters in other countries and American cities
(@url{http://www.forth.org/chapters.html}).

@node Conferences, , The Forth Interest Group, Forth-related information
@section Conferences
@cindex Conferences

There are several regular conferences related to Forth. They are all
well-publicised in @var{FORTH Dimensions} and on the comp.lang.forth
news group:

@itemize @bullet
@item
FORML -- the Forth modification laboratory convenes every year near
Monterey, California.
@item
The Rochester Forth Conference -- an annual conference traditionally
held in Rochester, New York.
@item
EuroForth -- this European conference takes place annually.
@end itemize


@node Word Index, Concept Index, Forth-related information, Top
@unnumbered Word Index

This index is a list of Forth words that have ``glossary'' entries
within this manual. Each word is listed with its stack effect and
wordset.

@printindex fn

@node Concept Index,  , Word Index, Top
@unnumbered Concept and Word Index

Not all entries listed in this index are present verbatim in the
text. This index also duplicates, in abbreviated form, all of the words
listed in the Word Index (only the names are listed for the words here).

@printindex cp

@contents
@bye