--- a/man/ChangeLog	Thu Mar 14 03:54:28 2002 +0000
+++ b/man/ChangeLog	Thu Mar 14 11:50:17 2002 +0000
@@ -1,3 +1,12 @@
+2002-03-14  Stephen J. Turnbull  <stephen@xemacs.org>
+
+	* emodules.texi (Using DEFUN): Add @ref{Lisp Primitives}.
+
+	* lispref/mule.texi (Unicode Support): 
+	* xemacs/custom.texi (Behaviors):
+	* lispref/customize.texi (Enabling Behavior):
+	New nodes.
+
 2002-03-12  Ben Wing  <ben@xemacs.org>
 
 	* The Great Mule Merge: placeholder.

--- a/man/emodules.texi	Thu Mar 14 03:54:28 2002 +0000
+++ b/man/emodules.texi	Thu Mar 14 11:50:17 2002 +0000
@@ -855,11 +855,12 @@
 @cindex functions, Lisp
 @cindex functions, defining
 
-  Although the full syntax of a function declaration is discussed in the
-@value{emacs} internals manual in greater depth, what follows is a brief
-description of how to define and implement a new Lisp primitive in a
-module.  This is done using the @code{DEFUN} macro.  Here is a small
-example:
+  The full syntax of a function declaration is discussed in the
+@value{emacs} internals manual in greater depth.  @ref{Writing Lisp
+Primitives,,,internals,@value{emacs} Internals Manual}.  What follows is
+a brief description of how to define and implement a new Lisp primitive
+in a module.  This is done using the @code{DEFUN} macro.  Here is a
+small example:
 
 @example
 @cartouche

--- a/man/lispref/customize.texi	Thu Mar 14 03:54:28 2002 +0000
+++ b/man/lispref/customize.texi	Thu Mar 14 11:50:17 2002 +0000
@@ -16,6 +16,7 @@
 * Group Definitions::
 * Variable Definitions::
 * Customization Types::
+* Enabling Behavior::
 @end menu
 
 @node Common Keywords
@@ -287,7 +288,7 @@
 customization buffer.  The @code{saved-value} property is actually a
 list whose car is an expression which evaluates to the value.
 
-@node Customization Types
+@node Customization Types, Enabling Behavior, Variable Definitions, Customization
 @section Customization Types
 
   When you define a user option with @code{defcustom}, you must specify
@@ -748,3 +749,106 @@
 @code{checkbox} associated with this item.
 @end ignore
 @end table
+
+
+@node Enabling Behavior, , Customization Types, Customization
+@subsection Enabling Behavior
+@cindex behavior
+
+@c #### Does this belong here?
+
+Some functionality requires a fair amount of effort to enable globally
+in a session.  For example, someone who discovers filladapt and really
+likes it must toggle it separately in each buffer.  On the other hand,
+after trying it for a while she might like to disable it everywhere,
+having decided it doesn't work very well for her.  Such a functionality
+is called a @dfn{behavior}.
+
+@code{define-behavior} allows the programmer to register functions to
+enable or disable a package globally in a session.  The user sees a
+consistent interface through the @code{enable-behavior} and
+@code{disable-behavior} functions.  These functions were introduced in
+XEmacs 21.5.6.
+
+@defvar behavior-hash-table
+
+Internal table of registered behaviors.
+@end defvar
+
+@defvar behavior-history
+
+History of entered behaviors.
+@end defvar
+
+@defun define-behavior name doc-string [cl-keys ...]
+
+Define a behavior named @var{name}.
+
+@var{doc-string} must be specified.  It is a description of what the
+behavior does when it's enabled and how to further control it (typically
+through custom variables).  Accepted keywords are
+
+@table @code
+@item :title
+A "pretty" version of the name, for use in menus.  If omitted
+a prettified name will be generated.
+
+@item :require
+A single symbol or a list of such symbols, which need to be
+present at enable time, or will be loaded using @code{require}.
+
+@item :enable
+A function of no variables, which turns the behavior on.
+
+@item :disable
+A function of no variables, which turns the behavior off.
+@end table
+
+Behaviors are assumed to be global, and to take effect immediately; if
+the underlying package is per-buffer, the enabler may have to scan all
+existing buffers and frob them.  When a behavior is disabled, it should
+completely go away @strong{everywhere}, as if it were never invoked at
+all.
+
+The @code{:disable} keyword can be missing.  This is bad practice.  In
+such a case, attempting to disable the behavior will signal an error
+unless you use the @code{force} option.
+@end defun
+
+@defun read-behavior prompt [require-match [initial-contents [history [default]]]]
+
+Return a behavior symbol from the minibuffer, prompting with string
+@var{prompt}.
+
+The optional arguments @var{require-match}, @var{initial-contents},
+@var{history}, and @var{default} are passed to @code{completing-read},
+and have semantics derived from that function.  @ref{Minibuffer
+Completion}.  The default value of @var{history} is
+@code{behavior-history}.
+@end defun
+
+@defun behavior-enabled-p name
+
+Return non-nil if the behavior registered under @var{name} is enabled.
+
+Unimplemented in 21.5.6.
+@end defun
+
+@defun enable-behavior behavior [force]
+Enable the behavior registered under the symbol @var{behavior}.
+
+The optional argument @var{force} is unimplemented in 21.5.6.
+
+Called interactively, prompt the user for @var{behavior}, and take
+@var{force} from the prefix argument.
+@end defun
+
+@defun disable-behavior (behavior &optional force)
+Disable the behavior registered under the symbol @var{behavior}.
+
+The optional argument @var{force} is unimplemented in 21.5.6.
+
+Called interactively, prompt the user for @var{behavior}, and take
+@var{force} from the prefix argument.
+@end defun
+

--- a/man/lispref/mule.texi	Thu Mar 14 03:54:28 2002 +0000
+++ b/man/lispref/mule.texi	Thu Mar 14 11:50:17 2002 +0000
@@ -1,6 +1,6 @@
 @c -*-texinfo-*-
 @c This is part of the XEmacs Lisp Reference Manual.
-@c Copyright (C) 1996 Ben Wing.
+@c Copyright (C) 1996 Ben Wing, 2001-2002 Free Software Foundation.
 @c See the file lispref.texi for copying conditions.
 @setfilename ../../info/internationalization.info
 @node MULE, Tips, Internationalization, top
@@ -23,6 +23,7 @@
 * Coding Systems::      Ways of representing a string of chars using integers.
 * CCL::                 A special language for writing fast converters.
 * Category Tables::     Subdividing charsets into groups.
+* Unicode Support::     The universal coded character set.
 @end menu
 
 @node Internationalization Terminology, Charsets, , MULE
@@ -2009,7 +2010,7 @@
 
   This section is not yet written.
 
-@node Category Tables, , CCL, MULE
+@node Category Tables, Unicode Support, CCL, MULE
 @section Category Tables
 
   A category table is a type of char table used for keeping track of
@@ -2069,3 +2070,114 @@
 Valid values are @code{nil} or a bit vector of size 95.
 @end defun
 
+
+@c Added 2002-03-13 sjt
+@node Unicode Support, , Category Tables, MULE
+@section Unicode Support
+@cindex unicode
+@cindex utf-8
+@cindex utf-16
+@cindex ucs-2
+@cindex ucs-4
+@cindex bmp
+@cindex basic multilingual plance
+
+Unicode support was added by Ben Wing to XEmacs 21.5.6.
+
+@defun set-language-unicode-precedence-list list
+Set the language-specific precedence list used for Unicode decoding.
+This is a list of charsets, which are consulted in order for a translation
+matching a given Unicode character.  If no matches are found, the charsets
+in the default precedence list (see
+@code{set-default-unicode-precedence-list}) are consulted, and then all
+remaining charsets, in some arbitrary order.
+
+The language-specific precedence list is meant to be set as part of the
+language environment initialization; the default precedence list is meant
+to be set by the user.
+@end defun
+
+@defun language-unicode-precedence-list
+Return the language-specific precedence list used for Unicode decoding.
+See @code{set-language-unicode-precedence-list} for more information.
+@end defun
+
+@defun set-default-unicode-precedence-list list
+Set the default precedence list used for Unicode decoding.
+This is meant to be set by the user.  See
+`set-language-unicode-precedence-list' for more information.
+@end defun
+
+@defun default-unicode-precedence-list
+Return the default precedence list used for Unicode decoding.
+See @code{set-language-unicode-precedence-list} for more information.
+@end defun
+
+@defun set-unicode-conversion character code
+Add conversion information between Unicode codepoints and characters.
+@var{character} is one of the following:
+
+@c #### fix this markup
+-- A character (in which case @var{code} must be a non-negative integer)
+-- A vector of characters (in which case @var{code} must be a vector of
+   non-negative integers of the same length)
+
+Values of @var{code} above 2^20 - 1 are allowed for the purpose of specifying
+private characters, but will cause errors when converted to UTF-16 or UTF-32.
+UCS-4 and UTF-8 can handle values to 2^31 - 1, but XEmacs Lisp integers top
+out at 2^30 - 1.
+@end defun
+
+@defun character-to-unicode character
+Convert @var{character} to Unicode codepoint.
+When there is no international support (i.e. MULE is not defined),
+this function simply does @code{char-to-int}.
+@end defun
+
+@defun unicode-to-character code [charsets]
+Convert Unicode codepoint @var{code} to character.
+@var{code} should be a non-negative integer.
+If @var{charsets} is given, it should be a list of charsets, and only those
+charsets will be consulted, in the given order, for a translation.
+Otherwise, the default ordering of all charsets will be given (see
+@code{set-unicode-charset-precedence}).
+
+When there is no international support (i.e. MULE is not defined),
+this function simply does @code{int-to-char} and ignores the
+@var{charsets} argument.
+@end defun
+
+@defun parse-unicode-translation-table filename charset start end offset flags
+Parse Unicode translation data in @var{filename} for MULE @var{charset}.
+Data is text, in the form of one translation per line -- charset
+codepoint followed by Unicode codepoint.  Numbers are decimal or hex
+\(preceded by 0x).  Comments are marked with a #.  Charset codepoints
+for two-dimensional charsets should have the first octet stored in the
+high 8 bits of the hex number and the second in the low 8 bits.
+
+If @var{start} and @var{end} are given, only charset codepoints within
+the given range will be processed.  If @var{offset} is given, that value
+will be added to all charset codepoints in the file to obtain the
+internal charset codepoint.  @var{start} and @var{end} apply to the
+codepoints in the file, before @var{offset} is applied.
+
+(Note that, as usual, we assume that octets are in the range 32 to
+127 or 33 to 126.  If you have a table in kuten form, with octets in
+the range 1 to 94, you will have to use an offset of 5140,
+i.e. 0x2020.)
+
+@var{flags}, if specified, control further how the tables are interpreted
+and are used to special-case certain known table weirdnesses in the
+Unicode tables:
+
+@table @code
+@item ignore-first-column'
+Exactly as it sounds.  The JIS X 0208 tables have 3 columns of data instead
+of 2; the first is the Shift-JIS codepoint.
+
+@item big5
+The charset codepoint is a Big Five codepoint; convert it to the
+proper hacked-up codepoint in `chinese-big5-1' or `chinese-big5-2'.
+@end table
+@end defun
+

--- a/man/xemacs/custom.texi	Thu Mar 14 03:54:28 2002 +0000
+++ b/man/xemacs/custom.texi	Thu Mar 14 11:50:17 2002 +0000
@@ -17,6 +17,9 @@
 @menu
 * Minor Modes::     Each minor mode is one feature you can turn on
                      independently of any others.
+* Behaviors::       Like minor modes, behaviors are independent of other
+                     features, but behaviors are usually enabled globally,
+                     while minor modes are per-buffer and often temporary.
 * Variables::       Many Emacs commands examine Emacs variables
                      to decide what to do; by setting variables,
                      you can control their functioning.
@@ -75,6 +78,42 @@
 as you type them.  For example, @samp{amd} might expand to @samp{abbrev
 mode}.  @xref{Abbrevs}, for full information.
 
+@c Updated for 21.5.6 2002/03/13 sjt
+@node Behaviors
+@section Behaviors
+@cindex behavior
+
+Some functionality requires a fair amount of effort to enable globally
+in a session.  For example, someone who discovers filladapt and really
+likes it must toggle it separately in each buffer.  On the other hand,
+after trying it for a while she might like to disable it everywhere,
+having decided it doesn't work very well for her.  Such a functionality
+is called a @dfn{behavior}.
+
+The package developer will register behaviors with XEmacs.  Then the
+user invokes the @code{enable-behavior} and @code{disable-behavior}
+functions to enable or disable a given behavior.  The behavior registry
+was introduced in XEmacs 21.5.6.
+
+@defun enable-behavior behavior [force]
+Called interactively, prompt the user, read a behavior symbol name with
+completion for @var{behavior}, and take @var{force} from the prefix
+argument.  Then enable the behavior registered under the symbol
+@var{behavior}.
+
+The optional argument @var{force} is unimplemented in 21.5.6.
+@end defun
+
+@defun disable-behavior behavior [force]
+Called interactively, prompt the user, read a behavior symbol name with
+completion for @var{behavior}, and take @var{force} from the prefix
+argument.  Then disable the behavior registered under the symbol
+@var{behavior}.
+
+The optional argument @var{force} is unimplemented in 21.5.6.
+@end defun
+
+
 @node Variables
 @section Variables
 @cindex variable