xemacs-beta: man/internals/internals.texi comparison

comparison man/internals/internals.texi @ 398:74fd4e045ea6 r21-2-29

Import from CVS: tag r21-2-29

author	cvs
date	Mon, 13 Aug 2007 11:13:30 +0200
parents	aabb7f5b1c81
children	a86b2b5e0111

comparison

equal deleted inserted replaced

-:f4aeb21a5bad
+:74fd4e045ea6
 @setfilename ../../info/internals.info
 @settitle XEmacs Internals Manual
 @c %**end of header
 @ifinfo
+@dircategory XEmacs Editor
+@direntry
+* Internals: (internals).       XEmacs Internals Manual.
+@end direntry
 Copyright @copyright{} 1992 - 1996 Ben Wing.
 Copyright @copyright{} 1996, 1997 Sun Microsystems.
 Copyright @copyright{} 1994 - 1998 Free Software Foundation.
 Copyright @copyright{} 1994, 1995 Board of Trustees, University of Illinois.
 @setchapternewpage odd
 @finalout
 @titlepage
 @title XEmacs Internals Manual
-@subtitle Version 1.2, October 1998
+@subtitle Version 1.3, August 1999
 @author Ben Wing
 @author Martin Buchholz
 @author Hrvoje Niksic
+@author Matthias Neubauer
+@author Olivier Galibert
 @page
 @vskip 0pt plus 1fill
 @noindent
 Copyright @copyright{} 1992 - 1996 Ben Wing. @*
 Copyright @copyright{} 1996, 1997 Sun Microsystems, Inc. @*
 Copyright @copyright{} 1994 - 1998 Free Software Foundation. @*
 Copyright @copyright{} 1994, 1995 Board of Trustees, University of Illinois.
 @sp 2
-Version 1.2 @*
+Version 1.3 @*
-October 1998.@*
+August 1999.@*
 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.
 * The XEmacs Object System (Abstractly Speaking)::
 * How Lisp Objects Are Represented in C::
 * Rules When Writing New C Code::
 * A Summary of the Various XEmacs Modules::
 * Allocation of Objects in XEmacs Lisp::
+* Dumping::
 * Events and the Event Loop::
 * Evaluation; Stack Frames; Bindings::
 * Symbols and Variables::
 * Buffers and Textual Representation::
 * MULE Character Sets and Encodings::
 * The Lisp Reader and Compiler::
 * Lstreams::
 * Consoles; Devices; Frames; Windows::
 * The Redisplay Mechanism::
 * Extents::
-* Faces and Glyphs::
+* Faces::
+* Glyphs::
 * Specifiers::
 * Menus::
 * Subprocesses::
 * Interface to X Windows::
-* Index::                   Index including concepts, functions, variables,
+* Index::
-and other terms.
+@detailmenu --- The Detailed Node Listing ---
---- The Detailed Node Listing ---
-Here are other nodes that are inferiors of those already listed,
-mentioned here so you can get to them in one step:
 A History of Emacs
 * Through Version 18::          Unification prevails.
 * Lucid Emacs::                 One version 19 Emacs.
 * GNU Emacs 19::                The other version 19 Emacs.
+* GNU Emacs 20::                The other version 20 Emacs.
 * XEmacs::                      The continuation of Lucid Emacs.
 Rules When Writing New C Code
 * General Coding Rules::
 * Writing Lisp Primitives::
 * Adding Global Lisp Variables::
+* Coding for Mule::
 * Techniques for XEmacs Developers::
+Coding for Mule
+* Character-Related Data Types::
+* Working With Character and Byte Positions::
+* Conversion to and from External Data::
+* General Guidelines for Writing Mule-Aware Code::
+* An Example of Mule-Aware Code::
 A Summary of the Various XEmacs Modules
 * Low-Level Modules::
 * Basic Lisp Modules::
 Allocation of Objects in XEmacs Lisp
 * Introduction to Allocation::
 * Garbage Collection::
 * GCPROing::
+* Garbage Collection - Step by Step::
 * Integers and Characters::
 * Allocation from Frob Blocks::
 * lrecords::
 * Low-level allocation::
 * Pure Space::
 * Symbol::
 * Marker::
 * String::
 * Compiled Function::
+Garbage Collection - Step by Step
+* Invocation::
+* garbage_collect_1::
+* mark_object::
+* gc_sweep::
+* sweep_lcrecords_1::
+* compact_string_chars::
+* sweep_strings::
+* sweep_bit_vectors_1::
+Dumping
+* Overview::
+* Data descriptions::
+* Dumping phase::
+* Reloading phase::
+Dumping phase
+* Object inventory::
+* Address allocation::
+* The header::
+* Data dumping::
+* Pointers dumping::
 Events and the Event Loop
 * Introduction to Events::
 * Main Loop::
 * Specifics of the Event Gathering Mechanism::
 MULE Character Sets and Encodings
 * Character Sets::
 * Encodings::
 * Internal Mule Encodings::
+* CCL::
 Encodings
 * Japanese EUC (Extended Unix Code)::
 * JIS7::
 Internal Mule Encodings
 * Internal String Encoding::
 * Internal Character Encoding::
-The Lisp Reader and Compiler
 Lstreams
+* Creating an Lstream::         Creating an lstream object.
+* Lstream Types::               Different sorts of things that are streamed.
+* Lstream Functions::           Functions for working with lstreams.
+* Lstream Methods::             Creating new lstream types.
 Consoles; Devices; Frames; Windows
 * Introduction to Consoles; Devices; Frames; Windows::
 * Point::
 * Window Hierarchy::
+* The Window Object::
 The Redisplay Mechanism
 * Critical Redisplay Sections::
 * Line Start Cache::
+* Redisplay Piece by Piece::
 Extents
 * Introduction to Extents::     Extents are ranges over text, with properties.
 * Extent Ordering::             How extents are ordered internally.
 * Format of the Extent Info::   The extent information in a buffer or string.
 * Zero-Length Extents::         A weird special case.
-* Mathematics of Extent Ordering::      A rigorous foundation.
+* Mathematics of Extent Ordering::  A rigorous foundation.
 * Extent Fragments::            Cached information useful for redisplay.
-Faces and Glyphs
+@end detailmenu
-Specifiers
-Menus
-Subprocesses
-Interface to X Windows
 @end menu
 @node A History of Emacs, XEmacs From the Outside, Top, Top
 @chapter A History of Emacs
 @cindex history of Emacs
 * GNU Emacs 19::                The other version 19 Emacs.
 * GNU Emacs 20::                The other version 20 Emacs.
 * XEmacs::                      The continuation of Lucid Emacs.
 @end menu
-@node Through Version 18
+@node Through Version 18, Lucid Emacs, A History of Emacs, A History of Emacs
 @section Through Version 18
 @cindex Gosling, James
 @cindex Great Usenet Renaming
 Although the history of the early versions of GNU Emacs is unclear,
 version 18.58 released ?????.
 @item
 version 18.59 released October 31, 1992.
 @end itemize
-@node Lucid Emacs
+@node Lucid Emacs, GNU Emacs 19, Through Version 18, A History of Emacs
 @section Lucid Emacs
 @cindex Lucid Emacs
 @cindex Lucid Inc.
 @cindex Energize
 @cindex Epoch
 version 20.3 (the first stable version of XEmacs 20.x) released November 30,
 1997.
 version 20.4 released February 28, 1998.
 @end itemize
-@node GNU Emacs 19
+@node GNU Emacs 19, GNU Emacs 20, Lucid Emacs, A History of Emacs
 @section GNU Emacs 19
 @cindex GNU Emacs 19
 @cindex FSF Emacs
 About a year after the initial release of Lucid Emacs, the FSF
 worse.  Lucid soon began incorporating features from GNU Emacs 19 into
 Lucid Emacs; the work was mostly done by Richard Mlynarik, who had been
 working on and using GNU Emacs for a long time (back as far as version
 16 or 17).
-@node GNU Emacs 20
+@node GNU Emacs 20, XEmacs, GNU Emacs 19, A History of Emacs
 @section GNU Emacs 20
 @cindex GNU Emacs 20
 @cindex FSF Emacs
 On February 2, 1997 work began on GNU Emacs to integrate Mule.  The first
 version 20.2 released September 20, 1997.
 @item
 version 20.3 released August 19, 1998.
 @end itemize
-@node XEmacs
+@node XEmacs,  , GNU Emacs 20, A History of Emacs
 @section XEmacs
 @cindex XEmacs
 @cindex Sun Microsystems
 @cindex University of Illinois
 windows, frames, events) that are useful for implementing an editor.
 Some of these objects (in particular windows and frames) have
 displayable representations, and XEmacs provides a function
 @code{redisplay()} that ensures that the display of all such objects
 matches their internal state.  Most of the time, a standard Lisp
-environment is in a @dfn{read-eval-print} loop -- i.e. ``read some Lisp
+environment is in a @dfn{read-eval-print} loop---i.e. ``read some Lisp
 code, execute it, and print the results''.  XEmacs has a similar loop:
 @itemize @bullet
 @item
 read an event
 handler for some or all classes of errors. (If no handler is registered,
 a default handler, generally installed by the top-level event loop, is
 executed; this prints out the error and continues.) Routines can also
 specify cleanup code (called an @dfn{unwind-protect}) that will be
 called when control exits from a block of code, no matter how that exit
-occurs -- i.e. even if a function deeply nested below it causes a
+occurs---i.e. even if a function deeply nested below it causes a
 non-local exit back to the top level.
 Note that this facility has appeared in some recent vintages of C, in
 particular Visual C++ and other PC compilers written for the Microsoft
 Win32 API.
 that if you declare a local variable in a particular function, and then
 call another function, that subfunction can ``see'' the local variable
 you declared.  This is actually considered a bug in Emacs Lisp and in
 all other early dialects of Lisp, and was corrected in Common Lisp. (In
 Common Lisp, you can still declare dynamically scoped variables if you
-want to -- they are sometimes useful -- but variables by default are
+want to---they are sometimes useful---but variables by default are
 @dfn{lexically scoped} as in C.)
 @end enumerate
 For those familiar with Lisp, Emacs Lisp is modelled after MacLisp, an
 early dialect of Lisp developed at MIT (no relation to the Macintosh
 most other data structures in Lisp.
 @item char
 An object representing a single character of text; chars behave like
 integers in many ways but are logically considered text rather than
 numbers and have a different read syntax. (the read syntax for a char
-contains the char itself or some textual encoding of it -- for example,
+contains the char itself or some textual encoding of it---for example,
 a Japanese Kanji character might be encoded as @samp{^[$(B#&^[(B} using the
-ISO-2022 encoding standard -- rather than the numerical representation
+ISO-2022 encoding standard---rather than the numerical representation
 of the char; this way, if the mapping between chars and integers
 changes, which is quite possible for Kanji characters and other extended
 characters, the same character will still be created.  Note that some
 primitives confuse chars and integers.  The worst culprit is @code{eq},
 which makes a special exception and considers a char to be @code{eq} to
 @example
 1.983e-4
 @end example
-converts to a float whose value is 1983.23e-4, or .0001983.
+converts to a float whose value is 1.983e-4, or .0001983.
 @example
 ?b
 @end example
 The tag describes the type of the Lisp object.  For integers and chars,
 the lower 28 bits contain the value of the integer or char; for all
 others, the lower 28 bits contain a pointer.  The mark bit is used
 during garbage-collection, and is always 0 when garbage collection is
 not happening. (The way that garbage collection works, basically, is that it
-loops over all places where Lisp objects could exist -- this includes
+loops over all places where Lisp objects could exist---this includes
 all global variables in C that contain Lisp objects [including
 @code{Vobarray}, the C equivalent of @code{obarray}; through this, all
-Lisp variables will get marked], plus various other places -- and
+Lisp variables will get marked], plus various other places---and
 recursively scans through the Lisp objects, marking each object it finds
 by setting the mark bit.  Then it goes through the lists of all objects
 allocated, freeing the ones that are not marked and turning off the mark
 bit of the ones that are marked.)
 machines/compilers do this, and on the ones that don't, a more
 complicated definition is selected by defining
 @code{EXPLICIT_SIGN_EXTEND}.
 Note that when @code{ERROR_CHECK_TYPECHECK} is defined, the extractor
-macros become more complicated -- they check the tag bits and/or the
+macros become more complicated---they check the tag bits and/or the
 type field in the first four bytes of a record type to ensure that the
 object is really of the correct type.  This is great for catching places
-where an incorrect type is being dereferenced -- this typically results
+where an incorrect type is being dereferenced---this typically results
 in a pointer being dereferenced as the wrong type of structure, with
 unpredictable (and sometimes not easily traceable) results.
 There are similar @code{XSET@var{TYPE}()} macros that construct a Lisp
 object.  These macros are of the form @code{XSET@var{TYPE}
 * Adding Global Lisp Variables::
 * Coding for Mule::
 * Techniques for XEmacs Developers::
 @end menu
-@node General Coding Rules
+@node General Coding Rules, Writing Lisp Primitives, Rules When Writing New C Code, Rules When Writing New C Code
 @section General Coding Rules
 The C code is actually written in a dialect of C called @dfn{Clean C},
 meaning that it can be compiled, mostly warning-free, with either a C or
 C++ compiler.  Coding in Clean C has several advantages over plain C.
 the same directory as the C sources) and @file{lisp.h}.  @file{config.h}
 must always be included before any other header files (including
 system header files) to ensure that certain tricks played by various
 @file{s/} and @file{m/} files work out correctly.
+When including header files, always use angle brackets, not double
+quotes, except when the file to be included is in the same directory as
+the including file.  If either file is a generated file, then that is
+not likely to be the case.  In order to understand why we have this
+rule, imagine what happens when you do a build in the source directory
+using @samp{./configure} and another build in another directory using
+@samp{../work/configure}.  There will be two different @file{config.h}
+files.  Which one will be used if you @samp{#include "config.h"}?
 @strong{All global and static variables that are to be modifiable must
 be declared uninitialized.}  This means that you may not use the
 ``declare with initializer'' form for these variables, such as @code{int
 some_variable = 0;}.  The reason for this has to do with some kludges
 done during the dumping process: If possible, the initialized data
 segment is re-mapped so that it becomes part of the (unmodifiable) code
 segment in the dumped executable.  This allows this memory to be shared
 among multiple running XEmacs processes.  XEmacs is careful to place as
 much constant data as possible into initialized variables (in
-particular, into what's called the @dfn{pure space} -- see below) during
+particular, into what's called the @dfn{pure space}---see below) during
 the @file{temacs} phase.
 @cindex copy-on-write
 @strong{Please note:} This kludge only works on a few systems nowadays,
 and is rapidly becoming irrelevant because most modern operating systems
 The C source code makes heavy use of C preprocessor macros.  One popular
 macro style is:
 @example
-#define FOO(var, value) do @{		\
+#define FOO(var, value) do @{           \
-Lisp_Object FOO_value = (value);	\
+Lisp_Object FOO_value = (value);      \
-... /* compute using FOO_value */	\
+... /* compute using FOO_value */     \
-(var) = bar;				\
+(var) = bar;                          \
 @} while (0)
 @end example
 The @code{do @{...@} while (0)} is a standard trick to allow FOO to have
 statement semantics, so that it can safely be used within an @code{if}
 case of @code{GET_EXTERNAL_LIST_LENGTH}, validating the properness of
 the list.  The macros @code{EXTERNAL_LIST_LOOP_DELETE_IF} and
 @code{LIST_LOOP_DELETE_IF} delete elements from a lisp list satisfying some
 predicate.
-@node Writing Lisp Primitives
+@node Writing Lisp Primitives, Adding Global Lisp Variables, General Coding Rules, Rules When Writing New C Code
 @section Writing Lisp Primitives
 Lisp primitives are Lisp functions implemented in C.  The details of
 interfacing the C function so that Lisp can call it are handled by a few
 C macros.  The only way to really understand how to write new C code is
 @file{eval.c} is a very good file to look through for examples;
 @file{lisp.h} contains the definitions for important macros and
 functions.
-@node Adding Global Lisp Variables
+@node Adding Global Lisp Variables, Coding for Mule, Writing Lisp Primitives, Rules When Writing New C Code
 @section Adding Global Lisp Variables
 Global variables whose names begin with @samp{Q} are constants whose
 value is a symbol of a particular name.  The name of the variable should
 be derived from the name of the symbol using the same rules as for Lisp
 garbage-collection mechanism won't know that the object in this variable
 is in use, and will happily collect it and reuse its storage for another
 Lisp object, and you will be the one who's unhappy when you can't figure
 out how your variable got overwritten.
-@node Coding for Mule
+@node Coding for Mule, Techniques for XEmacs Developers, Adding Global Lisp Variables, Rules When Writing New C Code
 @section Coding for Mule
 @cindex Coding for Mule
 Although Mule support is not compiled by default in XEmacs, many people
 are using it, and we consider it crucial that new code works correctly
 * Conversion to and from External Data::
 * General Guidelines for Writing Mule-Aware Code::
 * An Example of Mule-Aware Code::
 @end menu
-@node Character-Related Data Types
+@node Character-Related Data Types, Working With Character and Byte Positions, Coding for Mule, Coding for Mule
 @subsection Character-Related Data Types
 First, let's review the basic character-related datatypes used by
 XEmacs.  Note that the separate @code{typedef}s are not mandatory in the
 current implementation (all of them boil down to @code{unsigned char} or
 which are equivalent to @code{unsigned char}.  Obviously, an
 @code{Extcount} is the distance between two @code{Extbyte}s.  Extbytes
 and Extcounts are not all that frequent in XEmacs code.
 @end table
-@node Working With Character and Byte Positions
+@node Working With Character and Byte Positions, Conversion to and from External Data, Character-Related Data Types, Coding for Mule
 @subsection Working With Character and Byte Positions
 Now that we have defined the basic character-related types, we can look
 at the macros and functions designed for work with them and for
 conversion between them.  Most of these macros are defined in
 @example
 Bufbyte *charptr_n_addr (Bufbyte *p, Charcount cc);
 @end example
 @end table
-@node Conversion to and from External Data
+@node Conversion to and from External Data, General Guidelines for Writing Mule-Aware Code, Working With Character and Byte Positions, Coding for Mule
 @subsection Conversion to and from External Data
 When an external function, such as a C library function, returns a
 @code{char} pointer, you should almost never treat it as @code{Bufbyte}.
 This is because these returned strings may contain 8bit characters which
 These macros convert external text of a specific format to its internal
 representation, with the external format being incoded into the name of
 the macro.
 @end table
-@node General Guidelines for Writing Mule-Aware Code
+@node General Guidelines for Writing Mule-Aware Code, An Example of Mule-Aware Code, Conversion to and from External Data, Coding for Mule
 @subsection General Guidelines for Writing Mule-Aware Code
 This section contains some general guidance on how to write Mule-aware
 code, as well as some pitfalls you should avoid.
 they receive.  This increases efficiency because that way external data
 needs to be decoded only once, when it is read.  After that, it is
 passed around in internal format.
 @end table
-@node An Example of Mule-Aware Code
+@node An Example of Mule-Aware Code,  , General Guidelines for Writing Mule-Aware Code, Coding for Mule
 @subsection An Example of Mule-Aware Code
 As an example of Mule-aware code, we shall will analyze the
 @code{string} function, which conses up a Lisp string from the character
 arguments it receives.  Here is the definition, pasted from
 over the XEmacs code.  For starters, I recommend
 @code{Fnormalize_menu_item_name} in @file{menubar.c}.  After you have
 understood this section of the manual and studied the examples, you can
 proceed writing new Mule-aware code.
-@node Techniques for XEmacs Developers
+@node Techniques for XEmacs Developers,  , Coding for Mule, Rules When Writing New C Code
 @section Techniques for XEmacs Developers
 To make a quantified XEmacs, do: @code{make quantmacs}.
 You simply can't dump Quantified and Purified images.  Run the image
 Unfortunately, Emacs Lisp is slow, and is going to stay slow.  Function
 calls in elisp are especially expensive.  Iterating over a long list is
 going to be 30 times faster implemented in C than in Elisp.
-To get started debugging XEmacs, take a look at the @file{gdbinit} and
+To get started debugging XEmacs, take a look at the @file{.gdbinit} and
-@file{dbxrc} files in the @file{src} directory.
+@file{.dbxrc} files in the @file{src} directory.
 @xref{Q2.1.15 - How to Debug an XEmacs problem with a debugger,,,
 xemacs-faq, XEmacs FAQ}.
 After making source code changes, run @code{make check} to ensure that
 you haven't introduced any regressions.  If you're feeling ambitious,
 * Modules for Interfacing with the Operating System::
 * Modules for Interfacing with X Windows::
 * Modules for Internationalization::
 @end menu
-@node Low-Level Modules
+@node Low-Level Modules, Basic Lisp Modules, A Summary of the Various XEmacs Modules, A Summary of the Various XEmacs Modules
 @section Low-Level Modules
 @example
 config.h
 @end example
 This is not currently used.
-@node Basic Lisp Modules
+@node Basic Lisp Modules, Modules for Standard Editing Operations, Low-Level Modules, A Summary of the Various XEmacs Modules
 @section Basic Lisp Modules
 @example
 emacsfns.h
 lisp-disunion.h
 declarations (i.e. a simple declaration like @code{struct foo;} where
 the structure itself is defined elsewhere) should be placed into the
 typedefs section as necessary.
 @file{lrecord.h} contains the basic structures and macros that implement
-all record-type Lisp objects -- i.e. all objects whose type is a field
+all record-type Lisp objects---i.e. all objects whose type is a field
 in their C structure, which includes all objects except the few most
 basic ones.
 @file{lisp.h} contains prototypes for most of the exported functions in
 the various modules.  Lisp primitives defined using @code{DEFUN} that
 not dependent on any particular object type, and interfaces to
 particular types of objects using a standardized interface of
 type-specific methods.  This scheme is a fundamental principle of
 object-oriented programming and is heavily used throughout XEmacs.  The
 great advantage of this is that it allows for a clean separation of
-functionality into different modules -- new classes of Lisp objects, new
+functionality into different modules---new classes of Lisp objects, new
 event interfaces, new device types, new stream interfaces, etc. can be
 added transparently without affecting code anywhere else in XEmacs.
 Because the different subsystems are divided into general and specific
 code, adding a new subtype within a subsystem will in general not
 require changes to the generic subsystem code or affect any of the other
 @end example
 @file{symbols.c} implements the handling of symbols, obarrays, and
 retrieving the values of symbols.  Much of the code is devoted to
 handling the special @dfn{symbol-value-magic} objects that define
-special types of variables -- this includes buffer-local variables,
+special types of variables---this includes buffer-local variables,
 variable aliases, variables that forward into C variables, etc.  This
 module is initialized extremely early (right after @file{alloc.c}),
 because it is here that the basic symbols @code{t} and @code{nil} are
 created, and those symbols are used everywhere throughout XEmacs.
 structures.  Note that the byte-code @emph{compiler} is written in Lisp.
-@node Modules for Standard Editing Operations
+@node Modules for Standard Editing Operations, Editor-Level Control Flow Modules, Basic Lisp Modules, A Summary of the Various XEmacs Modules
 @section Modules for Standard Editing Operations
 @example
 buffer.c
 buffer.h
 This module implements the undo mechanism for tracking buffer changes.
 Most of this could be implemented in Lisp.
-@node Editor-Level Control Flow Modules
+@node Editor-Level Control Flow Modules, Modules for the Basic Displayable Lisp Objects, Modules for Standard Editing Operations, A Summary of the Various XEmacs Modules
 @section Editor-Level Control Flow Modules
 @example
 event-Xt.c
 event-stream.c
 @example
 keyboard.c
 @end example
 @file{keyboard.c} contains functions that implement the actual editor
-command loop -- i.e. the event loop that cyclically retrieves and
+command loop---i.e. the event loop that cyclically retrieves and
 dispatches events.  This code is also rather tricky, just like
 @file{event-stream.c}.
 bootstrapping implementations early in temacs, before the echo-area Lisp
 code is loaded).
-@node Modules for the Basic Displayable Lisp Objects
+@node Modules for the Basic Displayable Lisp Objects, Modules for other Display-Related Lisp Objects, Editor-Level Control Flow Modules, A Summary of the Various XEmacs Modules
 @section Modules for the Basic Displayable Lisp Objects
 @example
 device-ns.h
 device-stream.c
 is part of the redisplay mechanism or the code for particular object
 types such as scrollbars.
-@node Modules for other Display-Related Lisp Objects
+@node Modules for other Display-Related Lisp Objects, Modules for the Redisplay Mechanism, Modules for the Basic Displayable Lisp Objects, A Summary of the Various XEmacs Modules
 @section Modules for other Display-Related Lisp Objects
 @example
 faces.c
 faces.h
 @example
 font-lock.c
 @end example
-This file provides C support for syntax highlighting -- i.e.
+This file provides C support for syntax highlighting---i.e.
 highlighting different syntactic constructs of a source file in
 different colors, for easy reading.  The C support is provided so that
 this is fast.
 These modules decode GIF-format image files, for use with glyphs.
-@node Modules for the Redisplay Mechanism
+@node Modules for the Redisplay Mechanism, Modules for Interfacing with the File System, Modules for other Display-Related Lisp Objects, A Summary of the Various XEmacs Modules
 @section Modules for the Redisplay Mechanism
 @example
 redisplay-output.c
 redisplay-tty.c
 These files provide some miscellaneous TTY-output functions and should
 probably be merged into @file{redisplay-tty.c}.
-@node Modules for Interfacing with the File System
+@node Modules for Interfacing with the File System, Modules for Other Aspects of the Lisp Interpreter and Object System, Modules for the Redisplay Mechanism, A Summary of the Various XEmacs Modules
 @section Modules for Interfacing with the File System
 @example
 lstream.c
 lstream.h
 for expanding symbolic links, on systems that don't implement it or have
 a broken implementation.
-@node Modules for Other Aspects of the Lisp Interpreter and Object System
+@node Modules for Other Aspects of the Lisp Interpreter and Object System, Modules for Interfacing with the Operating System, Modules for Interfacing with the File System, A Summary of the Various XEmacs Modules
 @section Modules for Other Aspects of the Lisp Interpreter and Object System
 @example
 elhash.c
 elhash.h
 @cindex mark method
 Opaque objects can also have an arbitrary @dfn{mark method} associated
 with them, in case the block of memory contains other Lisp objects that
 need to be marked for garbage-collection purposes. (If you need other
 object methods, such as a finalize method, you should just go ahead and
-create a new Lisp object type -- it's not hard.)
+create a new Lisp object type---it's not hard.)
 @example
 abbrev.c
 various security applications on the Internet.
-@node Modules for Interfacing with the Operating System
+@node Modules for Interfacing with the Operating System, Modules for Interfacing with X Windows, Modules for Other Aspects of the Lisp Interpreter and Object System, A Summary of the Various XEmacs Modules
 @section Modules for Interfacing with the Operating System
 @example
 callproc.c
 process.c
 These modules are used for MS-DOS support, which does not work in
 XEmacs.
-@node Modules for Interfacing with X Windows
+@node Modules for Interfacing with X Windows, Modules for Internationalization, Modules for Interfacing with the Operating System, A Summary of the Various XEmacs Modules
 @section Modules for Interfacing with X Windows
 @example
 Emacs.ad.h
 @end example
 Don't touch this code; something is liable to break if you do.
-@node Modules for Internationalization
+@node Modules for Internationalization,  , Modules for Interfacing with X Windows, A Summary of the Various XEmacs Modules
 @section Modules for Internationalization
 @example
 mule-canna.c
 mule-ccl.c
 Asian-language support, and is not currently used.
-@node Allocation of Objects in XEmacs Lisp, Events and the Event Loop, A Summary of the Various XEmacs Modules, Top
+@node Allocation of Objects in XEmacs Lisp, Dumping, A Summary of the Various XEmacs Modules, Top
 @chapter Allocation of Objects in XEmacs Lisp
 @menu
 * Introduction to Allocation::
 * Garbage Collection::
 * GCPROing::
+* Garbage Collection - Step by Step::
 * Integers and Characters::
 * Allocation from Frob Blocks::
 * lrecords::
 * Low-level allocation::
 * Pure Space::
 * Marker::
 * String::
 * Compiled Function::
 @end menu
-@node Introduction to Allocation
+@node Introduction to Allocation, Garbage Collection, Allocation of Objects in XEmacs Lisp, Allocation of Objects in XEmacs Lisp
 @section Introduction to Allocation
 Emacs Lisp, like all Lisps, has garbage collection.  This means that
 the programmer never has to explicitly free (destroy) an object; it
 happens automatically when the object becomes inaccessible.  Most
 format; this includes conses, strings, vectors, and sometimes symbols.
 With the exception of vectors, objects in this category are allocated in
 @dfn{frob blocks}, i.e. large blocks of memory that are subdivided into
 individual objects.  This saves a lot on malloc overhead, since there
 are typically quite a lot of these objects around, and the objects are
-small.  (A cons, for example, occupies 8 bytes on 32-bit machines -- 4
+small.  (A cons, for example, occupies 8 bytes on 32-bit machines---4
 bytes for each of the two objects it contains.) Vectors are individually
 @code{malloc()}ed since they are of variable size.  (It would be
 possible, and desirable, to allocate vectors of certain small sizes out
 of frob blocks, but it isn't currently done.) Strings are handled
 specially: Each string is allocated in two parts, a fixed size structure
 abstraction, the FSF scheme is not nearly as clean or as easy to
 extend. (FSF calls items of type (c) @code{Lisp_Misc} and items of type
 (d) @code{Lisp_Vectorlike}, with separate tags for each, although
 @code{Lisp_Vectorlike} is also used for vectors.)
-@node Garbage Collection
+@node Garbage Collection, GCPROing, Introduction to Allocation, Allocation of Objects in XEmacs Lisp
 @section Garbage Collection
 @cindex garbage collection
 @cindex mark and sweep
 Garbage collection is simple in theory but tricky to implement.
 Finally, note that garbage collection can be invoked explicitly
 by calling @code{garbage-collect} but is also called automatically
 by @code{eval}, once a certain amount of memory has been allocated
 since the last garbage collection (according to @code{gc-cons-threshold}).
-@node GCPROing
+@node GCPROing, Garbage Collection - Step by Step, Garbage Collection, Allocation of Objects in XEmacs Lisp
 @section @code{GCPRO}ing
 @code{GCPRO}ing is one of the ugliest and trickiest parts of Emacs
 internals.  The basic idea is that whenever garbage collection
 occurs, all in-use objects must be reachable somehow or
 variable @samp{gcprolist} pointing to the head of the list and the nth
 local @code{gcpro} variable pointing to the first @code{gcpro} variable
 in the next enclosing stack frame.  Each @code{GCPRO}ed thing is an
 lvalue, and the @code{struct gcpro} local variable contains a pointer to
 this lvalue.  This is why things will mess up badly if you don't pair up
-the @code{GCPRO}s and @code{UNGCPRO}s -- you will end up with
+the @code{GCPRO}s and @code{UNGCPRO}s---you will end up with
 @code{gcprolist}s containing pointers to @code{struct gcpro}s or local
 @code{Lisp_Object} variables in no-longer-active stack frames.
 @item
 It is actually possible for a single @code{struct gcpro} to
 anything that looks like a reference to an object as a reference.  This
 will result in a few objects not getting collected when they should, but
 it obviates the need for @code{GCPRO}ing, and allows garbage collection
 to happen at any point at all, such as during object allocation.
-@node Integers and Characters
+@node Garbage Collection - Step by Step, Integers and Characters, GCPROing, Allocation of Objects in XEmacs Lisp
+@section Garbage Collection - Step by Step
+@cindex garbage collection step by step
+@menu
+* Invocation::
+* garbage_collect_1::
+* mark_object::
+* gc_sweep::
+* sweep_lcrecords_1::
+* compact_string_chars::
+* sweep_strings::
+* sweep_bit_vectors_1::
+@end menu
+@node Invocation, garbage_collect_1, Garbage Collection - Step by Step, Garbage Collection - Step by Step
+@subsection Invocation
+@cindex garbage collection, invocation
+The first thing that anyone should know about garbage collection is:
+when and how the garbage collector is invoked. One might think that this
+could happen every time new memory is allocated, e.g. new objects are
+created, but this is @emph{not} the case. Instead, we have the following
+situation:
+The entry point of any process of garbage collection is an invocation
+of the function @code{garbage_collect_1} in file @code{alloc.c}. The
+invocation can occur @emph{explicitly} by calling the function
+@code{Fgarbage_collect} (in addition this function provides information
+about the freed memory), or can occur @emph{implicitly} in four different
+situations:
+@enumerate
+@item
+In function @code{main_1} in file @code{emacs.c}. This function is called
+at each startup of xemacs. The garbage collection is invoked after all
+initial creations are completed, but only if a special internal error
+checking-constant @code{ERROR_CHECK_GC} is defined.
+@item
+In function @code{disksave_object_finalization} in file
+@code{alloc.c}. The only purpose of this function is to clear the
+objects from memory which need not be stored with xemacs when we dump out
+an executable. This is only done by @code{Fdump_emacs} or by
+@code{Fdump_emacs_data} respectively (both in @code{emacs.c}). The
+actual clearing is accomplished by making these objects unreachable and
+starting a garbage collection. The function is only used while building
+xemacs.
+@item
+In function @code{Feval / eval} in file @code{eval.c}. Each time the
+well known and often used function eval is called to evaluate a form,
+one of the first things that could happen, is a potential call of
+@code{garbage_collect_1}. There exist three global variables,
+@code{consing_since_gc} (counts the created cons-cells since the last
+garbage collection), @code{gc_cons_threshold} (a specified threshold
+after which a garbage collection occurs) and @code{always_gc}. If
+@code{always_gc} is set or if the threshold is exceeded, the garbage
+collection will start.
+@item
+In function @code{Ffuncall / funcall} in file @code{eval.c}. This
+function evaluates calls of elisp functions and works according to
+@code{Feval}.
+@end enumerate
+The upshot is that garbage collection can basically occur everywhere
+@code{Feval}, respectively @code{Ffuncall}, is used - either directly or
+through another function. Since calls to these two functions are
+hidden in various other functions, many calls to
+@code{garabge_collect_1} are not obviously foreseeable, and therefore
+unexpected. Instances where they are used that are worth remembering are
+various elisp commands, as for example @code{or},
+@code{and}, @code{if}, @code{cond}, @code{while}, @code{setq}, etc.,
+miscellaneous @code{gui_item_...} functions, everything related to
+@code{eval} (@code{Feval_buffer}, @code{call0}, ...) and inside
+@code{Fsignal}. The latter is used to handle signals, as for example the
+ones raised by every @code{QUITE}-macro triggered after pressing Ctrl-g.
+@node garbage_collect_1, mark_object, Invocation, Garbage Collection - Step by Step
+@subsection @code{garbage_collect_1}
+@cindex @code{garbage_collect_1}
+We can now describe exactly what happens after the invocation takes
+place.
+@enumerate
+@item
+There are several cases in which the garbage collector is left immediately:
+when we are already garbage collecting (@code{gc_in_progress}), when
+the garbage collection is somehow forbidden
+(@code{gc_currently_forbidden}), when we are currently displaying something
+(@code{in_display}) or when we are preparing for the armageddon of the
+whole system (@code{preparing_for_armageddon}).
+@item
+Next the correct frame in which to put
+all the output occurring during garbage collecting is determined. In
+order to be able to restore the old display's state after displaying the
+message, some data about the current cursor position has to be
+saved. The variables @code{pre_gc_curser} and @code{cursor_changed} take
+care of that.
+@item
+The state of @code{gc_currently_forbidden} must be restored after
+the garbage collection, no matter what happens during the process. We
+accomplish this by @code{record_unwind_protect}ing the suitable function
+@code{restore_gc_inhibit} together with the current value of
+@code{gc_currently_forbidden}.
+@item
+If we are concurrently running an interactive xemacs session, the next step
+is simply to show the garbage collector's cursor/message.
+@item
+The following steps are the intrinsic steps of the garbage collector,
+therefore @code{gc_in_progress} is set.
+@item
+For debugging purposes, it is possible to copy the current C stack
+frame. However, this seems to be a currently unused feature.
+@item
+Before actually starting to go over all live objects, references to
+objects that are no longer used are pruned. We only have to do this for events
+(@code{clear_event_resource}) and for specifiers
+(@code{cleanup_specifiers}).
+@item
+Now the mark phase begins and marks all accessible elements. In order to
+start from
+all slots that serve as roots of accessibility, the function
+@code{mark_object} is called for each root individually to go out from
+there to mark all reachable objects. All roots that are traversed are
+shown in their processed order:
+@itemize @bullet
+@item
+all constant symbols and static variables that are registered via
+@code{staticpro}@ in the array @code{staticvec}.
+@xref{Adding Global Lisp Variables}.
+@item
+all Lisp objects that are created in C functions and that must be
+protected from freeing them. They are registered in the global
+list @code{gcprolist}.
+@xref{GCPROing}.
+@item
+all local variables (i.e. their name fields @code{symbol} and old
+values @code{old_values}) that are bound during the evaluation by the Lisp
+engine. They are stored in @code{specbinding} structs pushed on a stack
+called @code{specpdl}.
+@xref{Dynamic Binding; The specbinding Stack; Unwind-Protects}.
+@item
+all catch blocks that the Lisp engine encounters during the evaluation
+cause the creation of structs @code{catchtag} inserted in the list
+@code{catchlist}. Their tag (@code{tag}) and value (@code{val} fields
+are freshly created objects and therefore have to be marked.
+@xref{Catch and Throw}.
+@item
+every function application pushes new structs @code{backtrace}
+on the call stack of the Lisp engine (@code{backtrace_list}). The unique
+parts that have to be marked are the fields for each function
+(@code{function}) and all their arguments (@code{args}).
+@xref{Evaluation}.
+@item
+all objects that are used by the redisplay engine that must not be freed
+are marked by a special function called @code{mark_redisplay} (in
+@code{redisplay.c}).
+@item
+all objects created for profiling purposes are allocated by C functions
+instead of using the lisp allocation mechanisms. In order to receive the
+right ones during the sweep phase, they also have to be marked
+manually. That is done by the function @code{mark_profiling_info}
+@end itemize
+@item
+Hash tables in XEmacs belong to a kind of special objects that
+make use of a concept often called 'weak pointers'.
+To make a long story short, these kind of pointers are not followed
+during the estimation of the live objects during garbage collection.
+Any object referenced only by weak pointers is collected
+anyway, and the reference to it is cleared. In hash tables there are
+different usage patterns of them, manifesting in different types of hash
+tables, namely 'non-weak', 'weak', 'key-weak' and 'value-weak'
+(internally also 'key-car-weak' and 'value-car-weak') hash tables, each
+clearing entries depending on different conditions. More information can
+be found in the documentation to the function @code{make-hash-table}.
+Because there are complicated dependency rules about when and what to
+mark while processing weak hash tables, the standard @code{marker}
+method is only active if it is marking non-weak hash tables. As soon as
+a weak component is in the table, the hash table entries are ignored
+while marking. Instead their marking is done each separately by the
+function @code{finish_marking_weak_hash_tables}. This function iterates
+over each hash table entry @code{hentries} for each weak hash table in
+@code{Vall_weak_hash_tables}. Depending on the type of a table, the
+appropriate action is performed.
+If a table is acting as @code{HASH_TABLE_KEY_WEAK}, and a key already marked,
+everything reachable from the @code{value} component is marked. If it is
+acting as a @code{HASH_TABLE_VALUE_WEAK} and the value component is
+already marked, the marking starts beginning only from the
+@code{key} component.
+If it is a @code{HASH_TABLE_KEY_CAR_WEAK} and the car
+of the key entry is already marked, we mark both the @code{key} and
+@code{value} components.
+Finally, if the table is of the type @code{HASH_TABLE_VALUE_CAR_WEAK}
+and the car of the value components is already marked, again both the
+@code{key} and the @code{value} components get marked.
+Again, there are lists with comparable properties called weak
+lists. There exist different peculiarities of their types called
+@code{simple}, @code{assoc}, @code{key-assoc} and
+@code{value-assoc}. You can find further details about them in the
+description to the function @code{make-weak-list}. The scheme of their
+marking is similar: all weak lists are listed in @code{Qall_weak_lists},
+therefore we iterate over them. The marking is advanced until we hit an
+already marked pair. Then we know that during a former run all
+the rest has been marked completely. Again, depending on the special
+type of the weak list, our jobs differ. If it is a @code{WEAK_LIST_SIMPLE}
+and the elem is marked, we mark the @code{cons} part. If it is a
+@code{WEAK_LIST_ASSOC} and not a pair or a pair with both marked car and
+cdr, we mark the @code{cons} and the @code{elem}. If it is a
+@code{WEAK_LIST_KEY_ASSOC} and not a pair or a pair with a marked car of
+the elem, we mark the @code{cons} and the @code{elem}. Finally, if it is
+a @code{WEAK_LIST_VALUE_ASSOC} and not a pair or a pair with a marked
+cdr of the elem, we mark both the @code{cons} and the @code{elem}.
+Since, by marking objects in reach from weak hash tables and weak lists,
+other objects could get marked, this perhaps implies further marking of
+other weak objects, both finishing functions are redone as long as
+yet unmarked objects get freshly marked.
+@item
+After completing the special marking for the weak hash tables and for the weak
+lists, all entries that point to objects that are going to be swept in
+the further process are useless, and therefore have to be removed from
+the table or the list.
+The function @code{prune_weak_hash_tables} does the job for weak hash
+tables. Totally unmarked hash tables are removed from the list
+@code{Vall_weak_hash_tables}. The other ones are treated more carefully
+by scanning over all entries and removing one as soon as one of
+the components @code{key} and @code{value} is unmarked.
+The same idea applies to the weak lists. It is accomplished by
+@code{prune_weak_lists}: An unmarked list is pruned from
+@code{Vall_weak_lists} immediately. A marked list is treated more
+carefully by going over it and removing just the unmarked pairs.
+@item
+The function @code{prune_specifiers} checks all listed specifiers held
+in @code{Vall_speficiers} and removes the ones from the lists that are
+unmarked.
+@item
+All syntax tables are stored in a list called
+@code{Vall_syntax_tables}. The function @code{prune_syntax_tables} walks
+through it and unlinks the tables that are unmarked.
+@item
+Next, we will attack the complete sweeping - the function
+@code{gc_sweep} which holds the predominance.
+@item
+First, all the variables with respect to garbage collection are
+reset. @code{consing_since_gc} - the counter of the created cells since
+the last garbage collection - is set back to 0, and
+@code{gc_in_progress} is not @code{true} anymore.
+@item
+In case the session is interactive, the displayed cursor and message are
+removed again.
+@item
+The state of @code{gc_inhibit} is restored to the former value by
+unwinding the stack.
+@item
+A small memory reserve is always held back that can be reached by
+@code{breathing_space}. If nothing more is left, we create a new reserve
+and exit.
+@end enumerate
+@node mark_object, gc_sweep, garbage_collect_1, Garbage Collection - Step by Step
+@subsection @code{mark_object}
+@cindex @code{mark_object}
+The first thing that is checked while marking an object is whether the
+object is a real Lisp object @code{Lisp_Type_Record} or just an integer
+or a character. Integers and characters are the only two types that are
+stored directly - without another level of indirection, and therefore they
+don't have to be marked and collected.
+@xref{How Lisp Objects Are Represented in C}.
+The second case is the one we have to handle. It is the one when we are
+dealing with a pointer to a Lisp object. But, there exist also three
+possibilities, that prevent us from doing anything while marking: The
+object is read only which prevents it from being garbage collected,
+i.e. marked (@code{C_READONLY_RECORD_HEADER}). The object in question is
+already marked, and need not be marked for the second time (checked by
+@code{MARKED_RECORD_HEADER_P}). If it is a special, unmarkable object
+(@code{UNMARKABLE_RECORD_HEADER_P}, apparently, these are objects that
+sit in some const space, and can therefore not be marked, see
+@code{this_one_is_unmarkable} in @code{alloc.c}).
+Now, the actual marking is feasible. We do so by once using the macro
+@code{MARK_RECORD_HEADER} to mark the object itself (actually the
+special flag in the lrecord header), and calling its special marker
+"method" @code{marker} if available. The marker method marks every
+other object that is in reach from our current object. Note, that these
+marker methods should not call @code{mark_object} recursively, but
+instead should return the next object from where further marking has to
+be performed.
+In case another object was returned, as mentioned before, we reiterate
+the whole @code{mark_object} process beginning with this next object.
+@node gc_sweep, sweep_lcrecords_1, mark_object, Garbage Collection - Step by Step
+@subsection @code{gc_sweep}
+@cindex @code{gc_sweep}
+The job of this function is to free all unmarked records from memory. As
+we know, there are different types of objects implemented and managed, and
+consequently different ways to free them from memory.
+@xref{Introduction to Allocation}.
+We start with all objects stored through @code{lcrecords}. All
+bulkier objects are allocated and handled using that scheme of
+@code{lcrecords}. Each object is @code{malloc}ed separately
+instead of placing it in one of the contiguous frob blocks. All types
+that are currently stored
+using @code{lcrecords}'s  @code{alloc_lcrecord} and
+@code{make_lcrecord_list} are the types: vectors, buffers,
+char-table, char-table-entry, console, weak-list, database, device,
+ldap, hash-table, command-builder, extent-auxiliary, extent-info, face,
+coding-system, frame, image-instance, glyph, popup-data, gui-item,
+keymap, charset, color_instance, font_instance, opaque, opaque-list,
+process, range-table, specifier, symbol-value-buffer-local,
+symbol-value-lisp-magic, symbol-value-varalias, toolbar-button,
+tooltalk-message, tooltalk-pattern, window, and window-configuration. We
+take care of them in the fist place
+in order to be able to handle and to finalize items stored in them more
+easily. The function @code{sweep_lcrecords_1} as described below is
+doing the whole job for us.
+For a description about the internals: @xref{lrecords}.
+Our next candidates are the other objects that behave quite differently
+than everything else: the strings. They consists of two parts, a
+fixed-size portion (@code{struct Lisp_string}) holding the string's
+length, its property list and a pointer to the second part, and the
+actual string data, which is stored in string-chars blocks comparable to
+frob blocks. In this block, the data is not only freed, but also a
+compression of holes is made, i.e. all strings are relocated together.
+@xref{String}. This compacting phase is performed by the function
+@code{compact_string_chars}, the actual sweeping by the function
+@code{sweep_strings} is described below.
+After that, the other types are swept step by step using functions
+@code{sweep_conses}, @code{sweep_bit_vectors_1},
+@code{sweep_compiled_functions}, @code{sweep_floats},
+@code{sweep_symbols}, @code{sweep_extents}, @code{sweep_markers} and
+@code{sweep_extents}.  They are the fixed-size types cons, floats,
+compiled-functions, symbol, marker, extent, and event stored in
+so-called "frob blocks", and therefore we can basically do the same on
+every type objects, using the same macros, especially defined only to
+handle everything with respect to fixed-size blocks. The only fixed-size
+type that is not handled here are the fixed-size portion of strings,
+because we took special care of them earlier.
+The only big exceptions are bit vectors stored differently and
+therefore treated differently by the function @code{sweep_bit_vectors_1}
+described later.
+At first, we need some brief information about how
+these fixed-size types are managed in general, in order to understand
+how the sweeping is done. They have all a fixed size, and are therefore
+stored in big blocks of memory - allocated at once - that can hold a
+certain amount of objects of one type. The macro
+@code{DECLARE_FIXED_TYPE_ALLOC} creates the suitable structures for
+every type. More precisely, we have the block struct
+(holding a pointer to the previous block @code{prev} and the
+objects in @code{block[]}), a pointer to current block
+(@code{current_..._block)}) and its last index
+(@code{current_..._block_index}), and a pointer to the free list that
+will be created. Also a macro @code{FIXED_TYPE_FROM_BLOCK} plus some
+related macros exists that are used to obtain a new object, either from
+the free list @code{ALLOCATE_FIXED_TYPE_1} if there is an unused object
+of that type stored or by allocating a completely new block using
+@code{ALLOCATE_FIXED_TYPE_FROM_BLOCK}.
+The rest works as follows: all of them define a
+macro @code{UNMARK_...} that is used to unmark the object. They define a
+macro @code{ADDITIONAL_FREE_...} that defines additional work that has
+to be done when converting an object from in use to not in use (so far,
+only markers use it in order to unchain them). Then, they all call
+the macro @code{SWEEP_FIXED_TYPE_BLOCK} instantiated with their type name
+and their struct name.
+This call in particular does the following: we go over all blocks
+starting with the current moving towards the oldest.
+For each block, we look at every object in it. If the object already
+freed (checked with @code{FREE_STRUCT_P} using the first pointer of the
+object), or if it is
+set to read only (@code{C_READONLY_RECORD_HEADER_P}, nothing must be
+done. If it is unmarked (checked with @code{MARKED_RECORD_HEADER_P}), it
+is put in the free list and set free (using the macro
+@code{FREE_FIXED_TYPE}, otherwise it stays in the block, but is unmarked
+(by @code{UNMARK_...}). While going through one block, we note if the
+whole block is empty. If so, the whole block is freed (using
+@code{xfree}) and the free list state is set to the state it had before
+handling this block.
+@node sweep_lcrecords_1, compact_string_chars, gc_sweep, Garbage Collection - Step by Step
+@subsection @code{sweep_lcrecords_1}
+@cindex @code{sweep_lcrecords_1}
+After nullifying the complete lcrecord statistics, we go over all
+lcrecords two separate times. They are all chained together in a list with
+a head called @code{all_lcrecords}.
+The first loop calls for each object its @code{finalizer} method, but only
+in the case that it is not read only
+(@code{C_READONLY_RECORD_HEADER_P)}, it is not already marked
+(@code{MARKED_RECORD_HEADER_P}), it is not already in a free list (list of
+freed objects, field @code{free}) and finally it owns a finalizer
+method.
+The second loop actually frees the appropriate objects again by iterating
+through the whole list. In case an object is read only or marked, it
+has to persist, otherwise it is manually freed by calling
+@code{xfree}. During this loop, the lcrecord statistics are kept up to
+date by calling @code{tick_lcrecord_stats} with the right arguments,
+@node compact_string_chars, sweep_strings, sweep_lcrecords_1, Garbage Collection - Step by Step
+@subsection @code{compact_string_chars}
+@cindex @code{compact_string_chars}
+The purpose of this function is to compact all the data parts of the
+strings that are held in so-called @code{string_chars_block}, i.e. the
+strings that do not exceed a certain maximal length.
+The procedure with which this is done is as follows. We are keeping two
+positions in the @code{string_chars_block}s using two pointer/integer
+pairs, namely @code{from_sb}/@code{from_pos} and
+@code{to_sb}/@code{to_pos}. They stand for the actual positions, from
+where to where, to copy the actually handled string.
+While going over all chained @code{string_char_block}s and their held
+strings, staring at @code{first_string_chars_block}, both pointers
+are advanced and eventually a string is copied from @code{from_sb} to
+@code{to_sb}, depending on the status of the pointed at strings.
+More precisely, we can distinguish between the following actions.
+@itemize @bullet
+@item
+The string at @code{from_sb}'s position could be marked as free, which
+is indicated by an invalid pointer to the pointer that should point back
+to the fixed size string object, and which is checked by
+@code{FREE_STRUCT_P}. In this case, the @code{from_sb}/@code{from_pos}
+is advanced to the next string, and nothing has to be copied.
+@item
+Also, if a string object itself is unmarked, nothing has to be
+copied. We likewise advance the @code{from_sb}/@code{from_pos}
+pair as described above.
+@item
+In all other cases, we have a marked string at hand. The string data
+must be moved from the from-position to the to-position. In case
+there is not enough space in the actual @code{to_sb}-block, we advance
+this pointer to the beginning of the next block before copying. In case the
+from and to positions are different, we perform the
+actual copying using the library function @code{memmove}.
+@end itemize
+After compacting, the pointer to the current
+@code{string_chars_block}, sitting in @code{current_string_chars_block},
+is reset on the last block to which we moved a string,
+i.e. @code{to_block}, and all remaining blocks (we know that they just
+carry garbage) are explicitly @code{xfree}d.
+@node sweep_strings, sweep_bit_vectors_1, compact_string_chars, Garbage Collection - Step by Step
+@subsection @code{sweep_strings}
+@cindex @code{sweep_strings}
+The sweeping for the fixed sized string objects is essentially exactly
+the same as it is for all other fixed size types. As before, the freeing
+into the suitable free list is done by using the macro
+@code{SWEEP_FIXED_SIZE_BLOCK} after defining the right macros
+@code{UNMARK_string} and @code{ADDITIONAL_FREE_string}. These two
+definitions are a little bit special compared to the ones used
+for the other fixed size types.
+@code{UNMARK_string} is defined the same way except some additional code
+used for updating the bookkeeping information.
+For strings, @code{ADDITIONAL_FREE_string} has to do something in
+addition: in case, the string was not allocated in a
+@code{string_chars_block} because it exceeded the maximal length, and
+therefore it was @code{malloc}ed separately, we know also @code{xfree}
+it explicitly.
+@node sweep_bit_vectors_1,  , sweep_strings, Garbage Collection - Step by Step
+@subsection @code{sweep_bit_vectors_1}
+@cindex @code{sweep_bit_vectors_1}
+Bit vectors are also one of the rare types that are @code{malloc}ed
+individually. Consequently, while sweeping, all further needless
+bit vectors must be freed by hand. This is done, as one might imagine,
+the expected way: since they are all registered in a list called
+@code{all_bit_vectors}, all elements of that list are traversed,
+all unmarked bit vectors are unlinked by calling @code{xfree} and all of
+them become unmarked.
+In addition, the bookkeeping information used for garbage
+collector's output purposes is updated.
+@node Integers and Characters, Allocation from Frob Blocks, Garbage Collection - Step by Step, Allocation of Objects in XEmacs Lisp
 @section Integers and Characters
 Integer and character Lisp objects are created from integers using the
 macros @code{XSETINT()} and @code{XSETCHAR()} or the equivalent
 functions @code{make_int()} and @code{make_char()}. (These are actually
 @code{XSETINT()} and the like will truncate values given to them that
 are too big; i.e. you won't get the value you expected but the tag bits
 will at least be correct.
-@node Allocation from Frob Blocks
+@node Allocation from Frob Blocks, lrecords, Integers and Characters, Allocation of Objects in XEmacs Lisp
 @section Allocation from Frob Blocks
 The uninitialized memory required by a @code{Lisp_Object} of a particular type
 is allocated using
 @code{ALLOCATE_FIXED_TYPE()}.  This only occurs inside of the
 @code{ALLOCATE_FIXED_TYPE_FROM_BLOCK()}, which looks at the end of the
 last frob block for space, and creates a new frob block if there is
 none. (There are actually two versions of these macros, one of which is
 more defensive but less efficient and is used for error-checking.)
-@node lrecords
+@node lrecords, Low-level allocation, Allocation from Frob Blocks, Allocation of Objects in XEmacs Lisp
 @section lrecords
 [see @file{lrecord.h}]
 All lrecords have at the beginning of their structure a @code{struct
 (i.e. declared with a @code{_SEQUENCE_IMPLEMENTATION} macro.)  This should
 simply return the object's size in bytes, exactly as you might expect.
 For an example, see the methods for window configurations and opaques.
 @end enumerate
-@node Low-level allocation
+@node Low-level allocation, Pure Space, lrecords, Allocation of Objects in XEmacs Lisp
 @section Low-level allocation
 Memory that you want to allocate directly should be allocated using
 @code{xmalloc()} rather than @code{malloc()}.  This implements
 error-checking on the return value, and once upon a time did some more
 routines.  These routines also call @code{INCREMENT_CONS_COUNTER()} at the
 appropriate times; this keeps statistics on how much memory is
 allocated, so that garbage-collection can be invoked when the
 threshold is reached.
-@node Pure Space
+@node Pure Space, Cons, Low-level allocation, Allocation of Objects in XEmacs Lisp
 @section Pure Space
 Not yet documented.
-@node Cons
+@node Cons, Vector, Pure Space, Allocation of Objects in XEmacs Lisp
 @section Cons
 Conses are allocated in standard frob blocks.  The only thing to
 note is that conses can be explicitly freed using @code{free_cons()}
 and associated functions @code{free_list()} and @code{free_alist()}.  This
 generating extra objects and thereby triggering GC sooner.
 However, you have to be @emph{extremely} careful when doing this.
 If you mess this up, you will get BADLY BURNED, and it has happened
 before.
-@node Vector
+@node Vector, Bit Vector, Cons, Allocation of Objects in XEmacs Lisp
 @section Vector
 As mentioned above, each vector is @code{malloc()}ed individually, and
 all are threaded through the variable @code{all_vectors}.  Vectors are
 marked strangely during garbage collection, by kludging the size field.
 Note that the @code{struct Lisp_Vector} is declared with its
 @code{contents} field being a @emph{stretchy} array of one element.  It
 is actually @code{malloc()}ed with the right size, however, and access
 to any element through the @code{contents} array works fine.
-@node Bit Vector
+@node Bit Vector, Symbol, Vector, Allocation of Objects in XEmacs Lisp
 @section Bit Vector
 Bit vectors work exactly like vectors, except for more complicated
 code to access an individual bit, and except for the fact that bit
 vectors are lrecords while vectors are not. (The only difference here is
 that there's an lrecord implementation pointer at the beginning and the
 tag field in bit vector Lisp words is ``lrecord'' rather than
 ``vector''.)
-@node Symbol
+@node Symbol, Marker, Bit Vector, Allocation of Objects in XEmacs Lisp
 @section Symbol
 Symbols are also allocated in frob blocks.  Note that the code
 exists for symbols to be either lrecords (category (c) above)
 or simple types (category (b) above), and are lrecords by
 chained through their @code{next} field.
 Remember that @code{intern} looks up a symbol in an obarray, creating
 one if necessary.
-@node Marker
+@node Marker, String, Symbol, Allocation of Objects in XEmacs Lisp
 @section Marker
 Markers are allocated in frob blocks, as usual.  They are kept
 in a buffer unordered, but in a doubly-linked list so that they
 can easily be removed. (Formerly this was a singly-linked list,
 but in some cases garbage collection took an extraordinarily
 long time due to the O(N^2) time required to remove lots of
 markers from a buffer.) Markers are removed from a buffer in
 the finalize stage, in @code{ADDITIONAL_FREE_marker()}.
-@node String
+@node String, Compiled Function, Marker, Allocation of Objects in XEmacs Lisp
 @section String
 As mentioned above, strings are a special case.  A string is logically
 two parts, a fixed-size object (containing the length, property list,
 and a pointer to the actual data), and the actual data in the string.
 Note that there is one situation not handled: a string that is too big
 to fit into a string-chars block.  Such strings, called @dfn{big
 strings}, are all @code{malloc()}ed as their own block. (#### Although it
 would make more sense for the threshold for big strings to be somewhat
 lower, e.g. 1/2 or 1/4 the size of a string-chars block.  It seems that
-this was indeed the case formerly -- indeed, the threshold was set at
+this was indeed the case formerly---indeed, the threshold was set at
-1/8 -- but Mly forgot about this when rewriting things for 19.8.)
+1/8---but Mly forgot about this when rewriting things for 19.8.)
 Note also that the string data in string-chars blocks is padded as
 necessary so that proper alignment constraints on the @code{struct
 Lisp_String} back pointers are maintained.
 string data (which would normally be obtained from the now-non-existent
 @code{struct Lisp_String}) at the beginning of the dead string data gap.
 The string compactor recognizes this special 0xFFFFFFFF marker and
 handles it correctly.
-@node Compiled Function
+@node Compiled Function,  , String, Allocation of Objects in XEmacs Lisp
 @section Compiled Function
 Not yet documented.
-@node Events and the Event Loop, Evaluation; Stack Frames; Bindings, Allocation of Objects in XEmacs Lisp, Top
+@node Dumping, Events and the Event Loop, Allocation of Objects in XEmacs Lisp, Top
+@chapter Dumping
+@section What is dumping and its justification
+The C code of XEmacs is just a Lisp engine with a lot of built-in
+primitives useful for writing an editor.  The editor itself is written
+mostly in Lisp, and represents around 100K lines of code.  Loading and
+executing the initialization of all this code takes a bit a time (five
+to ten times the usual startup time of current xemacs) and requires
+having all the lisp source files around.  Having to reload them each
+time the editor is started would not be acceptable.
+The traditional solution to this problem is called dumping: the build
+process first creates the lisp engine under the name @file{temacs}, then
+runs it until it has finished loading and initializing all the lisp
+code, and eventually creates a new executable called @file{xemacs}
+including both the object code in @file{temacs} and all the contents of
+the memory after the initialization.
+This solution, while working, has a huge problem: the creation of the
+new executable from the actual contents of memory is an extremely
+system-specific process, quite error-prone, and which interferes with a
+lot of system libraries (like malloc).  It is even getting worse
+nowadays with libraries using constructors which are automatically
+called when the program is started (even before main()) which tend to
+crash when they are called multiple times, once before dumping and once
+after (IRIX 6.x libz.so pulls in some C++ image libraries thru
+dependencies which have this problem).  Writing the dumper is also one
+of the most difficult parts of porting XEmacs to a new operating system.
+Basically, `dumping' is an operation that is just not officially
+supported on many operating systems.
+The aim of the portable dumper is to solve the same problem as the
+system-specific dumper, that is to be able to reload quickly, using only
+a small number of files, the fully initialized lisp part of the editor,
+without any system-specific hacks.
+@menu
+* Overview::
+* Data descriptions::
+* Dumping phase::
+* Reloading phase::
+* Remaining issues::
+@end menu
+@node Overview, Data descriptions, Dumping, Dumping
+@section Overview
+The portable dumping system has to:
+@enumerate
+@item
+At dump time, write all initialized, non-quickly-rebuildable data to a
+file [Note: currently named @file{xemacs.dmp}, but the name will
+change], along with all informations needed for the reloading.
+@item
+When starting xemacs, reload the dump file, relocate it to its new
+starting address if needed, and reinitialize all pointers to this
+data.  Also, rebuild all the quickly rebuildable data.
+@end enumerate
+@node Data descriptions, Dumping phase, Overview, Dumping
+@section Data descriptions
+The more complex task of the dumper is to be able to write lisp objects
+(lrecords) and C structs to disk and reload them at a different address,
+updating all the pointers they include in the process.  This is done by
+using external data descriptions that give information about the layout
+of the structures in memory.
+The specification of these descriptions is in lrecord.h.  A description
+of an lrecord is an array of struct lrecord_description.  Each of these
+structs include a type, an offset in the structure and some optional
+parameters depending on the type.  For instance, here is the string
+description:
+@example
+static const struct lrecord_description string_description[] = @{
+@{ XD_BYTECOUNT,         offsetof (Lisp_String, size) @},
+@{ XD_OPAQUE_DATA_PTR,   offsetof (Lisp_String, data), XD_INDIRECT(0, 1) @},
+@{ XD_LISP_OBJECT,       offsetof (Lisp_String, plist) @},
+@{ XD_END @}
+@};
+@end example
+The first line indicates a member of type Bytecount, which is used by
+the next, indirect directive.  The second means "there is a pointer to
+some opaque data in the field @code{data}".  The length of said data is
+given by the expression @code{XD_INDIRECT(0, 1)}, which means "the value
+in the 0th line of the description (welcome to C) plus one".  The third
+line means "there is a Lisp_Object member @code{plist} in the Lisp_String
+structure".  @code{XD_END} then ends the description.
+This gives us all the information we need to move around what is pointed
+to by a structure (C or lrecord) and, by transitivity, everything that
+it points to.  The only missing information for dumping is the size of
+the structure.  For lrecords, this is part of the
+lrecord_implementation, so we don't need to duplicate it.  For C
+structures we use a struct struct_description, which includes a size
+field and a pointer to an associated array of lrecord_description.
+@node Dumping phase, Reloading phase, Data descriptions, Dumping
+@section Dumping phase
+Dumping is done by calling the function pdump() (in alloc.c) which is
+invoked from Fdump_emacs (in emacs.c).  This function performs a number
+of tasks.
+@menu
+* Object inventory::
+* Address allocation::
+* The header::
+* Data dumping::
+* Pointers dumping::
+@end menu
+@node Object inventory, Address allocation, Dumping phase, Dumping phase
+@subsection Object inventory
+The first task is to build the list of the objects to dump.  This
+includes:
+@itemize @bullet
+@item lisp objects
+@item C structures
+@end itemize
+We end up with one @code{pdump_entry_list_elmt} per object group (arrays
+of C structs are kept together) which includes a pointer to the first
+object of the group, the per-object size and the count of objects in the
+group, along with some other information which is initialized later.
+These entries are linked together in @code{pdump_entry_list} structures
+and can be enumerated thru either:
+@enumerate
+@item
+the @code{pdump_object_table}, an array of @code{pdump_entry_list}, one
+per lrecord type, indexed by type number.
+@item
+the @code{pdump_opaque_data_list}, used for the opaque data which does
+not include pointers, and hence does not need descriptions.
+@item
+the @code{pdump_struct_table}, which is a vector of
+@code{struct_description}/@code{pdump_entry_list} pairs, used for
+non-opaque C structures.
+@end enumerate
+This uses a marking strategy similar to the garbage collector.  Some
+differences though:
+@enumerate
+@item
+We do not use the mark bit (which does not exist for C structures
+anyway), we use a big hash table instead.
+@item
+We do not use the mark function of lrecords but instead rely on the
+external descriptions.  This happens essentially because we need to
+follow pointers to C structures and opaque data in addition to
+Lisp_Object members.
+@end enumerate
+This is done by @code{pdump_register_object}, which handles Lisp_Object
+variables, and pdump_register_struct which handles C structures, which
+both delegate the description management to pdump_register_sub.
+The hash table doubles as a map object to pdump_entry_list_elmt (i.e.
+allows us to look up a pdump_entry_list_elmt with the object it points
+to).  Entries are added with @code{pdump_add_entry()} and looked up with
+@code{pdump_get_entry()}.  There is no need for entry removal.  The hash
+value is computed quite basically from the object pointer by
+@code{pdump_make_hash()}.
+The roots for the marking are:
+@enumerate
+@item
+the @code{staticpro}'ed variables (there is a special @code{staticpro_nodump()}
+call for protected variables we do not want to dump).
+@item
+the @code{pdump_wire}'d variables (@code{staticpro} is equivalent to
+@code{staticpro_nodump()} + @code{pdump_wire()}).
+@item
+the @code{dumpstruct}'ed variables, which points to C structures.
+@end enumerate
+This does not include the GCPRO'ed variables, the specbinds, the
+catchtags, the backlist, the redisplay or the profiling info, since we
+do not want to rebuild the actual chain of lisp calls which end up to
+the dump-emacs call, only the global variables.
+Weak lists and weak hash tables are dumped as if they were their
+non-weak equivalent (without changing their type, of course).  This has
+not yet been a problem.
+@node Address allocation, The header, Object inventory, Dumping phase
+@subsection Address allocation
+The next step is to allocate the offsets of each of the objects in the
+final dump file.  This is done by @code{pdump_allocate_offset()} which
+is called indirectly by @code{pdump_scan_by_alignment()}.
+The strategy to deal with alignment problems uses these facts:
+@enumerate
+@item
+real world alignment requirements are powers of two.
+@item
+the C compiler is required to adjust the size of a struct so that you
+can have an array of them next to each other.  This means you can have a
+upper bound of the alignment requirements of a given structure by
+looking at which power of two its size is a multiple.
+@item
+the non-variant part of variable size lrecords has an alignment
+requirement of 4.
+@end enumerate
+Hence, for each lrecord type, C struct type or opaque data block the
+alignment requirement is computed as a power of two, with a minimum of
+2^2 for lrecords.  @code{pdump_scan_by_alignment()} then scans all the
+@code{pdump_entry_list_elmt}'s, the ones with the highest requirements
+first.  This ensures the best packing.
+The maximum alignment requirement we take into account is 2^8.
+@code{pdump_allocate_offset()} only has to do a linear allocation,
+starting at offset 256 (this leaves room for the header and keep the
+alignments happy).
+@node The header, Data dumping, Address allocation, Dumping phase
+@subsection The header
+The next step creates the file and writes a header with a signature and
+some random informations in it (number of staticpro, number of assigned
+lrecord types, etc...).  The reloc_address field, which indicates at
+which address the file should be loaded if we want to avoid post-reload
+relocation, is set to 0.  It then seeks to offset 256 (base offset for
+the objects).
+@node Data dumping, Pointers dumping, The header, Dumping phase
+@subsection Data dumping
+The data is dumped in the same order as the addresses were allocated by
+@code{pdump_dump_data()}, called from @code{pdump_scan_by_alignment()}.
+This function copies the data to a temporary buffer, relocates all
+pointers in the object to the addresses allocated in step Address
+Allocation, and writes it to the file.  Using the same order means that,
+if we are careful with lrecords whose size is not a multiple of 4, we
+are ensured that the object is always written at the offset in the file
+allocated in step Address Allocation.
+@node Pointers dumping,  , Data dumping, Dumping phase
+@subsection Pointers dumping
+A bunch of tables needed to reassign properly the global pointers are
+then written.  They are:
+@enumerate
+@item the staticpro array
+@item the dumpstruct array
+@item the lrecord_implementation_table array
+@item a vector of all the offsets to the objects in the file that include a
+description (for faster relocation at reload time)
+@item the pdump_wired and pdump_wired_list arrays
+@end enumerate
+For each of the arrays we write both the pointer to the variables and
+the relocated offset of the object they point to.  Since these variables
+are global, the pointers are still valid when restarting the program and
+are used to regenerate the global pointers.
+The @code{pdump_wired_list} array is a special case.  The variables it
+points to are the head of weak linked lists of lisp objects of the same
+type.  Not all objects of this list are dumped so the relocated pointer
+we associate with them points to the first dumped object of the list, or
+Qnil if none is available.  This is also the reason why they are not
+used as roots for the purpose of object enumeration.
+This is the end of the dumping part.
+@node Reloading phase, Remaining issues, Dumping phase, Dumping
+@section Reloading phase
+@subsection File loading
+The file is mmap'ed in memory (which ensures a PAGESIZE alignment, at
+least 4096), or if mmap is unavailable or fails, a 256-bytes aligned
+malloc is done and the file is loaded.
+Some variables are reinitialized from the values found in the header.
+The difference between the actual loading address and the reloc_address
+is computed and will be used for all the relocations.
+@subsection Putting back the staticvec
+The staticvec array is memcpy'd from the file and the variables it
+points to are reset to the relocated objects addresses.
+@subsection Putting back the dumpstructed variables
+The variables pointed to by dumpstruct in the dump phase are reset to
+the right relocated object addresses.
+@subsection lrecord_implementations_table
+The lrecord_implementations_table is reset to its dump time state and
+the right lrecord_type_index values are put in.
+@subsection Object relocation
+All the objects are relocated using their description and their offset
+by @code{pdump_reloc_one}.  This step is unnecessary if the
+reloc_address is equal to the file loading address.
+@subsection Putting back the pdump_wire and pdump_wire_list variables
+Same as Putting back the dumpstructed variables.
+@subsection Reorganize the hash tables
+Since some of the hash values in the lisp hash tables are
+address-dependent, their layout is now wrong.  So we go through each of
+them and have them resorted by calling @code{pdump_reorganize_hash_table}.
+@node Remaining issues,  , Reloading phase, Dumping
+@section Remaining issues
+The build process will have to start a post-dump xemacs, ask it the
+loading address (which will, hopefully, be always the same between
+different xemacs invocations) and relocate the file to the new address.
+This way the object relocation phase will not have to be done, which
+means no writes in the objects and that, because of the use of mmap, the
+dumped data will be shared between all the xemacs running on the
+computer.
+Some executable signature will be necessary to ensure that a given dump
+file is really associated with a given executable, or random crashes
+will occur.  Maybe a random number set at compile or configure time thru
+a define.  This will also allow for having differently-compiled xemacsen
+on the same system (mule and no-mule comes to mind).
+The DOC file contents should probably end up in the dump file.
+@node Events and the Event Loop, Evaluation; Stack Frames; Bindings, Dumping, Top
 @chapter Events and the Event Loop
 @menu
 * Introduction to Events::
 * Main Loop::
 * Other Event Loop Functions::
 * Converting Events::
 * Dispatching Events; The Command Builder::
 @end menu
-@node Introduction to Events
+@node Introduction to Events, Main Loop, Events and the Event Loop, Events and the Event Loop
 @section Introduction to Events
 An event is an object that encapsulates information about an
 interesting occurrence in the operating system.  Events are
 generated either by user action, direct (e.g. typing on the
 XEmacs has its own types of events (called @dfn{Emacs events}),
 which provides an abstract layer on top of the system-dependent
 nature of the most basic events that are received.  Part of the
 complex nature of the XEmacs event collection process involves
 converting from the operating-system events into the proper
-Emacs events -- there may not be a one-to-one correspondence.
+Emacs events---there may not be a one-to-one correspondence.
 Emacs events are documented in @file{events.h}; I'll discuss them
 later.
-@node Main Loop
+@node Main Loop, Specifics of the Event Gathering Mechanism, Introduction to Events, Events and the Event Loop
 @section Main Loop
 The @dfn{command loop} is the top-level loop that the editor is always
 running.  It loops endlessly, calling @code{next-event} to retrieve an
 event and @code{dispatch-event} to execute it. @code{dispatch-event} does
 one console), and the engine that looks up keystrokes and
 constructs full key sequences is called the @dfn{command builder}.
 This is documented elsewhere.
 The guts of the command loop are in @code{command_loop_1()}.  This
-function doesn't catch errors, though -- that's the job of
+function doesn't catch errors, though---that's the job of
 @code{command_loop_2()}, which is a condition-case (i.e. error-trapping)
 wrapper around @code{command_loop_1()}.  @code{command_loop_1()} never
 returns, but may get thrown out of.
 When an error occurs, @code{cmd_error()} is called, which usually
 wrapper similar to @code{command_loop_2()}.  Note also that
 @code{initial_command_loop()} sets up a catch for @code{top-level} when
 invoking @code{top_level_1()}, just like when it invokes
 @code{command_loop_2()}.
-@node Specifics of the Event Gathering Mechanism
+@node Specifics of the Event Gathering Mechanism, Specifics About the Emacs Event, Main Loop, Events and the Event Loop
 @section Specifics of the Event Gathering Mechanism
 Here is an approximate diagram of the collection processes
 at work in XEmacs, under TTY's (TTY's are simpler than X
 so we'll look at this first):
 which repeatedly calls `next-event'
 and then dispatches the event
 using `dispatch-event'
 @end example
-@node Specifics About the Emacs Event
+@node Specifics About the Emacs Event, The Event Stream Callback Routines, Specifics of the Event Gathering Mechanism, Events and the Event Loop
 @section Specifics About the Emacs Event
-@node The Event Stream Callback Routines
+@node The Event Stream Callback Routines, Other Event Loop Functions, Specifics About the Emacs Event, Events and the Event Loop
 @section The Event Stream Callback Routines
-@node Other Event Loop Functions
+@node Other Event Loop Functions, Converting Events, The Event Stream Callback Routines, Events and the Event Loop
 @section Other Event Loop Functions
 @code{detect_input_pending()} and @code{input-pending-p} look for
 input by calling @code{event_stream->event_pending_p} and looking in
 @code{[V]unread-command-event} and the @code{command_event_queue} (they
 @code{read-char} calls @code{next-command-event} and uses
 @code{event_to_character()} to return the character equivalent.  With
 the right kind of input method support, it is possible for (read-char)
 to return a Kanji character.
-@node Converting Events
+@node Converting Events, Dispatching Events; The Command Builder, Other Event Loop Functions, Events and the Event Loop
 @section Converting Events
 @code{character_to_event()}, @code{event_to_character()},
 @code{event-to-character}, and @code{character-to-event} convert between
 characters and keypress events corresponding to the characters.  If the
 event was not a keypress, @code{event_to_character()} returns -1 and
 @code{event-to-character} returns @code{nil}.  These functions convert
 between character representation and the split-up event representation
 (keysym plus mod keys).
-@node Dispatching Events; The Command Builder
+@node Dispatching Events; The Command Builder,  , Converting Events, Events and the Event Loop
 @section Dispatching Events; The Command Builder
 Not yet documented.
 @node Evaluation; Stack Frames; Bindings, Symbols and Variables, Events and the Event Loop, Top
 * Dynamic Binding; The specbinding Stack; Unwind-Protects::
 * Simple Special Forms::
 * Catch and Throw::
 @end menu
-@node Evaluation
+@node Evaluation, Dynamic Binding; The specbinding Stack; Unwind-Protects, Evaluation; Stack Frames; Bindings, Evaluation; Stack Frames; Bindings
 @section Evaluation
 @code{Feval()} evaluates the form (a Lisp object) that is passed to
 it.  Note that evaluation is only non-trivial for two types of objects:
 symbols and conses.  A symbol is evaluated simply by calling
 @code{call3()} call a function, passing it the argument(s) given (the
 arguments are given as separate C arguments rather than being passed as
 an array).  @code{apply1()} uses @code{Fapply()} while the others use
 @code{Ffuncall()} to do the real work.
-@node Dynamic Binding; The specbinding Stack; Unwind-Protects
+@node Dynamic Binding; The specbinding Stack; Unwind-Protects, Simple Special Forms, Evaluation, Evaluation; Stack Frames; Bindings
 @section Dynamic Binding; The specbinding Stack; Unwind-Protects
 @example
 struct specbinding
 @{
 a local-variable binding (@code{func} is 0, @code{symbol} is not
 @code{nil}, and @code{old_value} holds the old value, which is stored as
 the symbol's value).
 @end enumerate
-@node Simple Special Forms
+@node Simple Special Forms, Catch and Throw, Dynamic Binding; The specbinding Stack; Unwind-Protects, Evaluation; Stack Frames; Bindings
 @section Simple Special Forms
 @code{or}, @code{and}, @code{if}, @code{cond}, @code{progn},
 @code{prog1}, @code{prog2}, @code{setq}, @code{quote}, @code{function},
 @code{let*}, @code{let}, @code{while}
 Note that, with the exeption of @code{Fprogn}, these functions are
 typically called in real life only in interpreted code, since the byte
 compiler knows how to convert calls to these functions directly into
 byte code.
-@node Catch and Throw
+@node Catch and Throw,  , Simple Special Forms, Evaluation; Stack Frames; Bindings
 @section Catch and Throw
 @example
 struct catchtag
 @{
 * Introduction to Symbols::
 * Obarrays::
 * Symbol Values::
 @end menu
-@node Introduction to Symbols
+@node Introduction to Symbols, Obarrays, Symbols and Variables, Symbols and Variables
 @section Introduction to Symbols
 A symbol is basically just an object with four fields: a name (a
 string), a value (some Lisp object), a function (some Lisp object), and
 a property list (usually a list of alternating keyword/value pairs).
 there can be a distinct function and variable with the same name.  The
 property list is used as a more general mechanism of associating
 additional values with particular names, and once again the namespace is
 independent of the function and variable namespaces.
-@node Obarrays
+@node Obarrays, Symbol Values, Introduction to Symbols, Symbols and Variables
 @section Obarrays
 The identity of symbols with their names is accomplished through a
 structure called an obarray, which is just a poorly-implemented hash
 table mapping from strings to symbols whose name is that string. (I say
 a new one, and @code{unintern} to remove a symbol from an obarray.  This
 returns the removed symbol. (Remember: You can't put the symbol back
 into any obarray.) Finally, @code{mapatoms} maps over all of the symbols
 in an obarray.
-@node Symbol Values
+@node Symbol Values,  , Obarrays, Symbols and Variables
 @section Symbol Values
 The value field of a symbol normally contains a Lisp object.  However,
 a symbol can be @dfn{unbound}, meaning that it logically has no value.
 This is internally indicated by storing a special Lisp object, called
 * Markers and Extents::         Tagging locations within a buffer.
 * Bufbytes and Emchars::        Representation of individual characters.
 * The Buffer Object::           The Lisp object corresponding to a buffer.
 @end menu
-@node Introduction to Buffers
+@node Introduction to Buffers, The Text in a Buffer, Buffers and Textual Representation, Buffers and Textual Representation
 @section Introduction to Buffers
 A buffer is logically just a Lisp object that holds some text.
 In this, it is like a string, but a buffer is optimized for
 frequent insertion and deletion, while a string is not.  Furthermore:
 and @dfn{buffer of the selected window}, and the distinction between
 @dfn{point} of the current buffer and @dfn{window-point} of the selected
 window. (This latter distinction is explained in detail in the section
 on windows.)
-@node The Text in a Buffer
+@node The Text in a Buffer, Buffer Lists, Introduction to Buffers, Buffers and Textual Representation
 @section The Text in a Buffer
 The text in a buffer consists of a sequence of zero or more
 characters.  A @dfn{character} is an integer that logically represents
 a letter, number, space, or other unit of text.  Most of the characters
 Bufbytes underscores the fact that we are working with a string of bytes
 in the internal Emacs buffer representation rather than in one of a
 number of possible alternative representations (e.g. EUC-encoded text,
 etc.).
-@node Buffer Lists
+@node Buffer Lists, Markers and Extents, The Text in a Buffer, Buffers and Textual Representation
 @section Buffer Lists
 Recall earlier that buffers are @dfn{permanent} objects, i.e.  that
 they remain around until explicitly deleted.  This entails that there is
 a list of all the buffers in existence.  This list is actually an
 respectively.  You can also force a new buffer to be created using
 @code{generate-new-buffer}, which takes a name and (if necessary) makes
 a unique name from this by appending a number, and then creates the
 buffer.  This is basically like the symbol operation @code{gensym}.
-@node Markers and Extents
+@node Markers and Extents, Bufbytes and Emchars, Buffer Lists, Buffers and Textual Representation
 @section Markers and Extents
 Among the things associated with a buffer are things that are
 logically attached to certain buffer positions.  This can be used to
 keep track of a buffer position when text is inserted and deleted, so
 is no way to determine what markers are in a buffer if you are just
 given the buffer.  Extents remain in a buffer until they are detached
 (which could happen as a result of text being deleted) or the buffer is
 deleted, and primitives do exist to enumerate the extents in a buffer.
-@node Bufbytes and Emchars
+@node Bufbytes and Emchars, The Buffer Object, Markers and Extents, Buffers and Textual Representation
 @section Bufbytes and Emchars
 Not yet documented.
-@node The Buffer Object
+@node The Buffer Object,  , Bufbytes and Emchars, Buffers and Textual Representation
 @section The Buffer Object
 Buffers contain fields not directly accessible by the Lisp programmer.
 We describe them here, naming them by the names used in the C code.
 Many are accessible indirectly in Lisp programs via Lisp primitives.
 * Encodings::
 * Internal Mule Encodings::
 * CCL::
 @end menu
-@node Character Sets
+@node Character Sets, Encodings, MULE Character Sets and Encodings, MULE Character Sets and Encodings
 @section Character Sets
 A character set (or @dfn{charset}) is an ordered set of characters.  A
 particular character in a charset is indexed using one or more
 @dfn{position codes}, which are non-negative integers.  The number of
 160 - 255       Latin-1                 32 - 127
 @end example
 This is a bit ad-hoc but gets the job done.
-@node Encodings
+@node Encodings, Internal Mule Encodings, Character Sets, MULE Character Sets and Encodings
 @section Encodings
 An @dfn{encoding} is a way of numerically representing characters from
 one or more character sets.  If an encoding only encompasses one
 character set, then the position codes for the characters in that
 @menu
 * Japanese EUC (Extended Unix Code)::
 * JIS7::
 @end menu
-@node Japanese EUC (Extended Unix Code)
+@node Japanese EUC (Extended Unix Code), JIS7, Encodings, Encodings
 @subsection Japanese EUC (Extended Unix Code)
 This encompasses the character sets Printing-ASCII, Japanese-JISX0201,
 and Japanese-JISX0208-Kana (half-width katakana, the right half of
 JISX0201).  It uses 8-bit bytes.
 Japanese-JISX0208        PC1 + 0x80 | PC2 + 0x80
 Japanese-JISX0212        PC1 + 0x80 | PC2 + 0x80
 @end example
-@node JIS7
+@node JIS7,  , Japanese EUC (Extended Unix Code), Encodings
 @subsection JIS7
 This encompasses the character sets Printing-ASCII,
 Japanese-JISX0201-Roman (the left half of JISX0201; this character set
 is very similar to Printing-ASCII and is a 94-character charset),
 0x1B 0x28 0x42    ESC ( B            invoke Printing-ASCII
 @end example
 Initially, Printing-ASCII is invoked.
-@node Internal Mule Encodings
+@node Internal Mule Encodings, CCL, Encodings, MULE Character Sets and Encodings
 @section Internal Mule Encodings
 In XEmacs/Mule, each character set is assigned a unique number, called a
 @dfn{leading byte}.  This is used in the encodings of a character.
 Leading bytes are in the range 0x80 - 0xFF (except for ASCII, which has
 @menu
 * Internal String Encoding::
 * Internal Character Encoding::
 @end menu
-@node Internal String Encoding
+@node Internal String Encoding, Internal Character Encoding, Internal Mule Encodings, Internal Mule Encodings
 @subsection Internal String Encoding
 ASCII characters are encoded using their position code directly.  Other
 characters are encoded using their leading byte followed by their
 position code(s) with the high bit set.  Characters in private character
 None of the standard non-modal encodings meet all of these
 conditions.  For example, EUC satisfies only (2) and (3), while
 Shift-JIS and Big5 (not yet described) satisfy only (2). (All
 non-modal encodings must satisfy (2), in order to be unambiguous.)
-@node Internal Character Encoding
+@node Internal Character Encoding,  , Internal String Encoding, Internal Mule Encodings
 @subsection Internal Character Encoding
 One 19-bit word represents a single character.  The word is
 separated into three fields:
 @end example
 Note that character codes 0 - 255 are the same as the ``binary encoding''
 described above.
-@node CCL
+@node CCL,  , Internal Mule Encodings, MULE Character Sets and Encodings
 @section CCL
 @example
 CCL PROGRAM SYNTAX:
 CCL_PROGRAM := (CCL_MAIN_BLOCK
 * Lstream Types::               Different sorts of things that are streamed.
 * Lstream Functions::           Functions for working with lstreams.
 * Lstream Methods::             Creating new lstream types.
 @end menu
-@node Creating an Lstream
+@node Creating an Lstream, Lstream Types, Lstreams, Lstreams
 @section Creating an Lstream
 Lstreams come in different types, depending on what is being interfaced
 to.  Although the primitive for creating new lstreams is
 @code{Lstream_new()}, generally you do not call this directly.  Instead,
 Open for reading, but ``read'' never returns partial MULE characters.
 @item "wc"
 Open for writing, but never writes partial MULE characters.
 @end table
-@node Lstream Types
+@node Lstream Types, Lstream Functions, Creating an Lstream, Lstreams
 @section Lstream Types
 @table @asis
 @item stdio
 @item decoding
 @item encoding
 @end table
-@node Lstream Functions
+@node Lstream Functions, Lstream Methods, Lstream Types, Lstreams
 @section Lstream Functions
-@deftypefun {Lstream *} Lstream_new (Lstream_implementation *@var{imp}, CONST char *@var{mode})
+@deftypefun {Lstream *} Lstream_new (Lstream_implementation *@var{imp}, const char *@var{mode})
 Allocate and return a new Lstream.  This function is not really meant to
 be called directly; rather, each stream type should provide its own
 stream creation function, which creates the stream and does any other
 necessary creation stuff (e.g. opening a file).
 @end deftypefun
 @end deftypefn
 @deftypefn Macro void Lstream_ungetc (Lstream *@var{stream}, int @var{c})
 Push one byte back onto the input queue.  This will be the next byte
 read from the stream.  Any number of bytes can be pushed back and will
-be read in the reverse order they were pushed back -- most recent
+be read in the reverse order they were pushed back---most recent
-first. (This is necessary for consistency -- if there are a number of
+first. (This is necessary for consistency---if there are a number of
 bytes that have been unread and I read and unread a byte, it needs to be
 the first to be read again.) This is a macro and so it is very
 efficient.  The @var{c} argument is only evaluated once but the @var{stream}
 argument is evaluated more than once.
 @end deftypefn
 @deftypefunx int Lstream_fgetc (Lstream *@var{stream})
 @deftypefunx void Lstream_fungetc (Lstream *@var{stream}, int @var{c})
 Function equivalents of the above macros.
 @end deftypefun
-@deftypefun int Lstream_read (Lstream *@var{stream}, void *@var{data}, int @var{size})
+@deftypefun ssize_t Lstream_read (Lstream *@var{stream}, void *@var{data}, size_t @var{size})
 Read @var{size} bytes of @var{data} from the stream.  Return the number
 of bytes read.  0 means EOF. -1 means an error occurred and no bytes
 were read.
 @end deftypefun
-@deftypefun int Lstream_write (Lstream *@var{stream}, void *@var{data}, int @var{size})
+@deftypefun ssize_t Lstream_write (Lstream *@var{stream}, void *@var{data}, size_t @var{size})
 Write @var{size} bytes of @var{data} to the stream.  Return the number
 of bytes written.  -1 means an error occurred and no bytes were written.
 @end deftypefun
-@deftypefun void Lstream_unread (Lstream *@var{stream}, void *@var{data}, int @var{size})
+@deftypefun void Lstream_unread (Lstream *@var{stream}, void *@var{data}, size_t @var{size})
 Push back @var{size} bytes of @var{data} onto the input queue.  The next
 call to @code{Lstream_read()} with the same size will read the same
 bytes back.  Note that this will be the case even if there is other
 pending unread data.
 @end deftypefun
 @end deftypefun
 @deftypefun void Lstream_reopen (Lstream *@var{stream})
 Reopen a closed stream.  This enables I/O on it again.  This is not
 meant to be called except from a wrapper routine that reinitializes
-variables and such -- the close routine may well have freed some
+variables and such---the close routine may well have freed some
 necessary storage structures, for example.
 @end deftypefun
 @deftypefun void Lstream_rewind (Lstream *@var{stream})
 Rewind the stream to the beginning.
 @end deftypefun
-@node Lstream Methods
+@node Lstream Methods,  , Lstream Functions, Lstreams
 @section Lstream Methods
-@deftypefn {Lstream Method} int reader (Lstream *@var{stream}, unsigned char *@var{data}, int @var{size})
+@deftypefn {Lstream Method} ssize_t reader (Lstream *@var{stream}, unsigned char *@var{data}, size_t @var{size})
 Read some data from the stream's end and store it into @var{data}, which
 can hold @var{size} bytes.  Return the number of bytes read.  A return
 value of 0 means no bytes can be read at this time.  This may be because
 of an EOF, or because there is a granularity greater than one byte that
 the stream imposes on the returned data, and @var{size} is less than
 calls @code{Lstream_read()} with a very small size.
 This function can be @code{NULL} if the stream is output-only.
 @end deftypefn
-@deftypefn {Lstream Method} int writer (Lstream *@var{stream}, CONST unsigned char *@var{data}, int @var{size})
+@deftypefn {Lstream Method} ssize_t writer (Lstream *@var{stream}, const unsigned char *@var{data}, size_t @var{size})
 Send some data to the stream's end.  Data to be sent is in @var{data}
 and is @var{size} bytes.  Return the number of bytes sent.  This
 function can send and return fewer bytes than is passed in; in that
 case, the function will just be called again until there is no data left
 or 0 is returned.  A return value of 0 means that no more data can be
 @deftypefn {Lstream Method} int rewinder (Lstream *@var{stream})
 Rewind the stream.  If this is @code{NULL}, the stream is not seekable.
 @end deftypefn
 @deftypefn {Lstream Method} int seekable_p (Lstream *@var{stream})
-Indicate whether this stream is seekable -- i.e. it can be rewound.
+Indicate whether this stream is seekable---i.e. it can be rewound.
 This method is ignored if the stream does not have a rewind method.  If
 this method is not present, the result is determined by whether a rewind
 method is present.
 @end deftypefn
 * Point::
 * Window Hierarchy::
 * The Window Object::
 @end menu
-@node Introduction to Consoles; Devices; Frames; Windows
+@node Introduction to Consoles; Devices; Frames; Windows, Point, Consoles; Devices; Frames; Windows, Consoles; Devices; Frames; Windows
 @section Introduction to Consoles; Devices; Frames; Windows
 A window-system window that you see on the screen is called a
 @dfn{frame} in Emacs terminology.  Each frame is subdivided into one or
 more non-overlapping panes, called (confusingly) @dfn{windows}.  Each
 window, but every frame remembers the last window in it that was
 selected, and changing the selected frame causes the remembered window
 within it to become the selected window.  Similar relationships apply
 for consoles to devices and devices to frames.
-@node Point
+@node Point, Window Hierarchy, Introduction to Consoles; Devices; Frames; Windows, Consoles; Devices; Frames; Windows
 @section Point
 Recall that every buffer has a current insertion position, called
 @dfn{point}.  Now, two or more windows may be displaying the same buffer,
 and the text cursor in the two windows (i.e. @code{point}) can be in
 want to retrieve the correct value of @code{point} for a window,
 you must special-case on the selected window and retrieve the
 buffer's point instead.  This is related to why @code{save-window-excursion}
 does not save the selected window's value of @code{point}.
-@node Window Hierarchy
+@node Window Hierarchy, The Window Object, Point, Consoles; Devices; Frames; Windows
 @section Window Hierarchy
 @cindex window hierarchy
 @cindex hierarchy of windows
 If a frame contains multiple windows (panes), they are always created
 @dfn{one above the other}.
 @item
 Leaf windows also have markers in their @code{start} (the
 first buffer position displayed in the window) and @code{pointm}
-(the window's stashed value of @code{point} -- see above) fields,
+(the window's stashed value of @code{point}---see above) fields,
 while combination windows have nil in these fields.
 @item
 The list of children for a window is threaded through the
 @code{next} and @code{prev} fields of each child window.
 does nothing except set a special @code{dead} bit to 1 and clear out the
 @code{next}, @code{prev}, @code{hchild}, and @code{vchild} fields, for
 GC purposes.
 @item
-Most frames actually have two top-level windows -- one for the
+Most frames actually have two top-level windows---one for the
 minibuffer and one (the @dfn{root}) for everything else.  The modeline
 (if present) separates these two.  The @code{next} field of the root
 points to the minibuffer, and the @code{prev} field of the minibuffer
 points to the root.  The other @code{next} and @code{prev} fields are
 @code{nil}, and the frame points to both of these windows.
 frames have no root window, and the @code{next} of the minibuffer window
 is @code{nil} but the @code{prev} points to itself. (#### This is an
 artifact that should be fixed.)
 @end enumerate
-@node The Window Object
+@node The Window Object,  , Window Hierarchy, Consoles; Devices; Frames; Windows
 @section The Window Object
 Windows have the following accessible fields:
 @table @code
 @end enumerate
 @menu
 * Critical Redisplay Sections::
 * Line Start Cache::
+* Redisplay Piece by Piece::
 @end menu
-@node Critical Redisplay Sections
+@node Critical Redisplay Sections, Line Start Cache, The Redisplay Mechanism, The Redisplay Mechanism
 @section Critical Redisplay Sections
 @cindex critical redisplay sections
 Within this section, we are defenseless and assume that the
 following cannot happen:
 we simply return. #### We should abort instead.
 #### If a frame-size change does occur we should probably
 actually be preempting redisplay.
-@node Line Start Cache
+@node Line Start Cache, Redisplay Piece by Piece, Critical Redisplay Sections, The Redisplay Mechanism
 @section Line Start Cache
 @cindex line start cache
 The traditional scrolling code in Emacs breaks in a variable height
 world.  It depends on the key assumption that the number of lines that
 information basically for free.  In those cases where a user is simply
 scrolling around viewing a buffer there is a high probability that this
 is sufficient to always provide the needed information.  The second
 thing we can do is be smart about invalidating the cache.
-TODO -- Be smart about invalidating the cache.  Potential places:
+TODO---Be smart about invalidating the cache.  Potential places:
 @itemize @bullet
 @item
 Insertions at end-of-line which don't cause line-wraps do not alter the
 starting positions of any display lines.  These types of buffer
 @end itemize
 In case you're wondering, the Second Golden Rule of Redisplay is not
 applicable.
-@node Extents, Faces and Glyphs, The Redisplay Mechanism, Top
+@node Redisplay Piece by Piece,  , Line Start Cache, The Redisplay Mechanism
+@section Redisplay Piece by Piece
+@cindex Redisplay Piece by Piece
+As you can begin to see redisplay is complex and also not well
+documented. Chuck no longer works on XEmacs so this section is my take
+on the workings of redisplay.
+Redisplay happens in three phases:
+@enumerate
+@item
+Determine desired display in area that needs redisplay.
+Implemented by @code{redisplay.c}
+@item
+Compare desired display with current display
+Implemented by @code{redisplay-output.c}
+@item
+Output changes Implemented by @code{redisplay-output.c},
+@code{redisplay-x.c}, @code{redisplay-msw.c} and @code{redisplay-tty.c}
+@end enumerate
+Steps 1 and 2 are device-independant and relatively complex.  Step 3 is
+mostly device-dependent.
+Determining the desired display
+Display attributes are stored in @code{display_line} structures. Each
+@code{display_line} consists of a set of @code{display_block}'s and each
+@code{display_block} contains a number of @code{rune}'s. Generally
+dynarr's of @code{display_line}'s are held by each window representing
+the current display and the desired display.
+The @code{display_line} structures are tighly tied to buffers which
+presents a problem for redisplay as this connection is bogus for the
+modeline. Hence the @code{display_line} generation routines are
+duplicated for generating the modeline. This means that the modeline
+display code has many bugs that the standard redisplay code does not.
+The guts of @code{display_line} generation are in
+@code{create_text_block}, which creates a single display line for the
+desired locale. This incrementally parses the characters on the current
+line and generates redisplay structures for each.
+Gutter redisplay is different. Because the data to display is stored in
+a string we cannot use @code{create_text_block}. Instead we use
+@code{create_text_string_block} which performs the same function as
+@code{create_text_block} but for strings. Many of the complexities of
+@code{create_text_block} to do with cursor handling and selective
+display have been removed.
+@node Extents, Faces, The Redisplay Mechanism, Top
 @chapter Extents
 @menu
 * Introduction to Extents::     Extents are ranges over text, with properties.
 * Extent Ordering::             How extents are ordered internally.
 * Format of the Extent Info::   The extent information in a buffer or string.
 * Zero-Length Extents::         A weird special case.
-* Mathematics of Extent Ordering::      A rigorous foundation.
+* Mathematics of Extent Ordering::  A rigorous foundation.
 * Extent Fragments::            Cached information useful for redisplay.
 @end menu
-@node Introduction to Extents
+@node Introduction to Extents, Extent Ordering, Extents, Extents
 @section Introduction to Extents
 Extents are regions over a buffer, with a start and an end position
 denoting the region of the buffer included in the extent.  In
 addition, either end can be closed or open, meaning that the endpoint
 automatically go inside or out of extents as necessary with no
 further work needing to be done.  It didn't work out that way,
 however, and just ended up complexifying and buggifying all the
 rest of the code.)
-@node Extent Ordering
+@node Extent Ordering, Format of the Extent Info, Introduction to Extents, Extents
 @section Extent Ordering
 Extents are compared using memory indices.  There are two orderings
 for extents and both orders are kept current at all times.  The normal
 or @dfn{display} order is as follows:
 The display order and the e-order are complementary orders: any
 theorem about the display order also applies to the e-order if you swap
 all occurrences of ``display order'' and ``e-order'', ``less than'' and
 ``greater than'', and ``extent start'' and ``extent end''.
-@node Format of the Extent Info
+@node Format of the Extent Info, Zero-Length Extents, Extent Ordering, Extents
 @section Format of the Extent Info
 An extent-info structure consists of a list of the buffer or string's
 extents and a @dfn{stack of extents} that lists all of the extents over
 a particular position.  The stack-of-extents info is used for
-optimization purposes -- it basically caches some info that might
+optimization purposes---it basically caches some info that might
 be expensive to compute.  Certain otherwise hard computations are easy
 given the stack of extents over a particular position, and if the
 stack of extents over a nearby position is known (because it was
 calculated at some prior point in time), it's easy to move the stack
 of extents to the proper position.
 between two extents.  Note also that callers of these functions should
 not be aware of the fact that the extent list is implemented as an
 array, except for the fact that positions are integers (this should be
 generalized to handle integers and linked list equally well).
-@node Zero-Length Extents
+@node Zero-Length Extents, Mathematics of Extent Ordering, Format of the Extent Info, Extents
 @section Zero-Length Extents
 Extents can be zero-length, and will end up that way if their endpoints
 are explicitly set that way or if their detachable property is nil
 and all the text in the extent is deleted. (The exception is open-open
 Note that closed-open, non-detachable zero-length extents behave
 exactly like markers and that open-closed, non-detachable zero-length
 extents behave like the ``point-type'' marker in Mule.
-@node Mathematics of Extent Ordering
+@node Mathematics of Extent Ordering, Extent Fragments, Zero-Length Extents, Extents
 @section Mathematics of Extent Ordering
 @cindex extent mathematics
 @cindex mathematics of extents
 @cindex extent ordering
 Proof: If @math{F2} does not include @math{I} then its start index is
 greater than @math{I} and thus it is greater than any extent in
 @math{S}, including @math{F}.  Otherwise, @math{F2} includes @math{I}
 and thus is in @math{S}, and thus @math{F2 >= F}.
-@node Extent Fragments
+@node Extent Fragments,  , Mathematics of Extent Ordering, Extents
 @section Extent Fragments
 @cindex extent fragment
 Imagine that the buffer is divided up into contiguous, non-overlapping
 @dfn{runs} of text such that no extent starts or ends within a run
 (extents that abut the run don't count).
 An extent fragment is a structure that holds data about the run that
 contains a particular buffer position (if the buffer position is at the
-junction of two runs, the run after the position is used) -- the
+junction of two runs, the run after the position is used)---the
 beginning and end of the run, a list of all of the extents in that run,
 the @dfn{merged face} that results from merging all of the faces
 corresponding to those extents, the begin and end glyphs at the
 beginning of the run, etc.  This is the information that redisplay needs
 in order to display this run.
 Extent fragments have to be very quick to update to a new buffer
 position when moving linearly through the buffer.  They rely on the
 stack-of-extents code, which does the heavy-duty algorithmic work of
 determining which extents overly a particular position.
-@node Faces and Glyphs, Specifiers, Extents, Top
+@node Faces, Glyphs, Extents, Top
-@chapter Faces and Glyphs
+@chapter Faces
 Not yet documented.
-@node Specifiers, Menus, Faces and Glyphs, Top
+@node Glyphs, Specifiers, Faces, Top
+@chapter Glyphs
+Glyphs are graphical elements that can be displayed in XEmacs buffers or
+gutters. We use the term graphical element here in the broadest possible
+sense since glyphs can be as mundane as text to as arcane as a native
+tab widget.
+In XEmacs, glyphs represent the uninstantiated state of graphical
+elements, i.e. they hold all the information necessary to produce an
+image on-screen but the image does not exist at this stage.
+Glyphs are lazily instantiated by calling one of the glyph
+functions. This usually occurs within redisplay when
+@code{Fglyph_height} is called. Instantiation causes an image-instance
+to be created and cached. This cache is on a device basis for all glyphs
+except glyph-widgets, and on a window basis for glyph widgets.  The
+caching is done by @code{image_instantiate} and is necessary because it
+is generally possible to display an image-instance in multiple
+domains. For instance if we create a Pixmap, we can actually display
+this on multiple windows - even though we only need a single Pixmap
+instance to do this. If caching wasn't done then it would be necessary
+to create image-instances for every displayable occurrance of a glyph -
+and every usage - and this would be extremely memory and cpu intensive.
+Widget-glyphs (a.k.a native widgets) are not cached in this way. This is
+because widget-glyph image-instances on screen are toolkit windows, and
+thus cannot be reused in multiple XEmacs domains. Thus widget-glyphs are
+cached on a window basis.
+Any action on a glyph first consults the cache before actually
+instantiating a widget.
+@section Widget-Glyphs in the MS-Windows Environment
+To Do
+@section Widget-Glyphs in the X Environment
+Widget-glyphs under X make heavy use of lwlib for manipulating the
+native toolkit objects. This is primarily so that different toolkits can
+be supported for widget-glyphs, just as they are supported for features
+such as menubars etc.
+Lwlib is extremely poorly documented and quite hairy so here is my
+understanding of what goes on.
+Lwlib maintains a set of widget_instances which mirror the hierarchical
+state of Xt widgets. I think this is so that widgets can be updated and
+manipulated generically by the lwlib library. For instance
+update_one_widget_instance can cope with multiple types of widget and
+multiple types of toolkit. Each element in the widget hierarchy is updated
+from its corresponding widget_instance by walking the widget_instance
+tree recursively.
+This has desirable properties such as lw_modify_all_widgets which is
+called from glyphs-x.c and updates all the properties of a widget
+without having to know what the widget is or what toolkit it is from.
+Unfortunately this also has hairy properrties such as making the lwlib
+code quite complex. And of course lwlib has to know at some level what
+the widget is and how to set its properties.
+@node Specifiers, Menus, Glyphs, Top
 @chapter Specifiers
 Not yet documented.
 @node Menus, Subprocesses, Specifiers, Top
 @item tty_name
 The name of the terminal that the subprocess is using,
 or @code{nil} if it is using pipes.
 @end table
-@node Interface to X Windows, Index, Subprocesses, Top
+@node Interface to X Windows, Index , Subprocesses, Top
 @chapter Interface to X Windows
 Not yet documented.
 @include index.texi

Mercurial > hg > xemacs-beta

comparison man/internals/internals.texi @ 398:74fd4e045ea6 r21-2-29