[chise/xemacs-chise.git.1] / etc / NEWS

-*- mode:outline -*-
\f
* Introduction
==============

This file presents some general information about XEmacs.  It is
primarily about the changes in recent XEmacs versions and its release
history.

Use `C-c C-f' to move to the next equal level of outline, and
`C-c C-b' to move to previous equal level.  `C-h m' will give more
info about the Outline mode.  Many commands are also available through
the menubar.

Users who would like to know which capabilities have been introduced
in each release should look at the appropriate section of this file.
Starting with version 20.0, XEmacs includes ChangeLogs, which can be
consulted for a more detailed list of changes.

Users interested in some of the details of how XEmacs differs from GNU
Emacs should read the section "What's Different?" near the end of this
file.

    N.B.  The term "GNU Emacs" refers to any release of Emacs Version
    19 from the Free Software Foundation's GNU Project.  (We do not
    say just "Emacs" as Richard M. Stallman ["RMS"] prefers, because
    it is clearly a more generic term.)  The term "XEmacs" refers to
    this program or to its predecessors "Era" and "Lucid Emacs".  The
    predecessor of all these program is called "Emacs 18".  When no
    particular version is implied, "Emacs" will be used.

\f
* Changes in XEmacs 21.2
========================

** XEmacs now locates the early package hierarchy at ~/.xemacs/packages.

This has changed from simply ~/.xemacs.

** `delete-key-deletes-forward' now defaults to t.

`delete-key-deletes-forward' is the variable that regulates the
behaviour of the delete key on the systems that offer both a backspace
and a delete key.  If set to nil, the key labeled "Delete" will delete
backward.  If set to non-nil, the "Delete" key will delete forward,
except on keyboards where a "Backspace" key is not provided.

Unless our implementation has bugs, the only reason why you would want 
to set `delete-key-deletes-forward' to nil is if you want to use the
Delete key to delete backwards, despite the presence (according to
Xlib) of a BackSpace key on the keyboard.
 
** Interactive searching and matching case improvements.

Case sensitiveness in searching operations is normally controlled by
the variable `case-fold-search' (if non-nil, case is ignored while
searching).  This mechanism has now been slightly improved for
interactive searches: if the search string (or regexp) contains
uppercase characters, the searching is forced to be case-sensitive,
`case-fold-search'.   

The new behavior affects all functions performing interactive
searches, like `zap-to-char', `list-matching-lines', `tags-search'
etc.  The incremental search facility has always behaved that way.

** You can now create "indirect buffers", like in GNU Emacs.  An
indirect buffer shares its text with another buffer ("base buffer"),
but has its own major mode, local variables, extents, and narrowing.
An indirect buffer has a name of its own, distinct from those of the
base buffer and all other buffers.  An indirect buffer cannot itself
be visiting a file (though its base buffer can be).  The base buffer
cannot itself be indirect.

Use (make-indirect-buffer BASE-BUFFER NAME) to make an indirect buffer
named NAME whose base is BASE-BUFFER.  If BASE-BUFFER is itself an
indirect buffer, its base buffer is used as the base for the new
buffer.

You can make an indirect buffer current, or switch to it in a window,
just as you would a non-indirect buffer.

The function `buffer-base-buffer' returns a buffer's base buffer or
nil, if given an ordinary (non-indirect) buffer.  The function
`buffer-indirect-children' returns a list of the indirect children of
a base buffer.

** User names following the tilde character can now be completed at
file name prompts; e.g. `C-x C-f ~hni<TAB>' will complete to
`~hniksic/'.  To make this operation faster, a cache of user names is
maintained internally.

The new primitives available for this purpose are functions named
`user-name-completion' and `user-name-all-completions'.

** Native widgets can be displayed in buffers.

The glyph system has been extended to allow the display of glyphs that
are implemented as native window-system widgets. Thus you can embed
buttons, scrollbars, combo boxes, edit fields and progress gauges in a
buffer. As a side effect subwindow support now works once again.

This support is currently only available under MS-Windows.

** X-Face support is now available under MS-Windows

If an X-Face libary built under MS-Windows is available then XEmacs
will use this at build time.

** The font-menu is now available under MS-Windows

** MS-Windows support for selection is much more robust

Generally selection should now do what you would expect under
MS-Windows: the middle mouse button will paste your current selection
or the clipboard; conversions from different types of selection to the
clipboard can be made; the kill-ring and friends will be updated as
per X.

The only thing selection doesn't do is set the clipboard automatically
as this would break the MS-Windows model. If you want this behaviour
then set `selection-sets-clipboard' to t

\f
* Lisp and internal changes in XEmacs 21.2
==========================================

** Much effort has been invested to make XEmacs Lisp faster:

*** Many basic lisp operations are now faster.
This is especially the case when running a Mule-enabled XEmacs.

A general overhaul of the lisp engine should produce a speedup of 1.4
in a Latin-1 XEmacs, and 2.1 in a Mule XEmacs.  These numbers were
obtained running `(byte-compile "simple.el")', which should be a
pretty typical test of "pure" Lisp.

*** Lisp hash tables have been re-implemented.  The Common Lisp style
hash table interface has been made standard, and moved from cl.el into
fast C code (See the section on hash tables in the XEmacs Lisp
Reference).  A speedup factor of 3 can be expected with code that
makes intensive use of hash tables.

*** The garbage collector has been tuned, leading to a speedup of
1.16.

*** The family of functions that iterate over lists, like `memq', and
`rassq', have been made a little faster (typically 1.3).

*** Lisp function calls are faster, by approximately a factor of two.
However, defining inline functions (via defsubst) still makes sense
for tight loops.

*** Finally, a few functions have had dramatic performance
improvements.  For example, `(last long-list)' is now 30 times faster.

Of course, your mileage will vary.

Many operations do not see any improvement.  Surprisingly, running
(font-lock-fontify-buffer) does not use the Lisp engine much at all.
Speeding up your favorite slow operation is an excellent project to
improve XEmacs.  Don't forget to profile!

** XEmacs finally has an automated test suite!
Although this is not yet very sophisticated, it is already responsible
for several important bug fixes in XEmacs.  To try it out, simply use
the makefile target `make check' after building XEmacs.

** Hash tables have been reimplemented.
As was pointed out above, the standard interface to hash tables is now
the Common Lisp interface, as described in Common Lisp, the Language
(CLtL2, by Steele).  The older interface (functions with names
containing the phrase `hashtable') will continue to work, but the
preferred interface now has names containing the phrase `hash-table'.

Here's the executive overview: create hash tables using
make-hash-table, and use gethash, puthash, remhash, maphash and
clrhash to manipulate entries in the hash table.  See the (updated)
Lisp Reference Manual for details.

** Lisp code handles circular lists much more robustly.
Many basic lisp functions used to loop forever when given a circular
list, expecting you to C-g (quit) out of the loop.  Now this is more
likely to trigger a `circular-list' error.  Printing a circular list
now results in something like this:

    (let ((x (cons 'foo 'foo)))
      (setcdr x x)
      x)
      => (foo ... <circular list>)

An extra bonus is that checking for circularities is not just
friendlier, but actually faster than checking for C-g.

** The new form `ignore-file-errors', similar to `ignore-errors' may
be used as a short-hand for condition-case when you wish to ignore
file-related error.  For example:

    (ignore-file-errors (delete-file "foo"))

** The arguments to `locate-file' are now much more "lispy".  As
before, the usage is:

    (locate-file FILENAME PATH-LIST &optional SUFFIXES MODE)

Except that SUFFIXES are now a list of strings instead of a single,
colon-separated string.  MODE is now a symbol or a list of symbols
(symbols `exists', `executable', `writable', and `readable' are
supported) instead of an integer code.  See the documentation for
details.

Of course, the old form is still accepted for backward compatibility.

** `translate-region' has been improved in several ways.  Its TABLE
argument used to be a 256-character string.  In addition to this, it
can now also be a vector or a char-table (which is useful for Mule.)
If TABLE a vector or a generic char-table, you can map characters to
strings instead of to other characters.  For instance:

    (let ((table (make-char-table 'generic)))
      (put-char-table ?a "the letter a" table)
      (put-char-table ?b "" table)
      (put-char-table ?c ?\n table)
      (translate-region (point-min) (point-max) table))

** The `keywordp' function now returns non-nil only on symbols
interned in the global obarray.  For example:

    (keywordp (intern ":foo" [0]))
      => nil
    (keywordp (intern ":foo"))       ; The same as (keywordp :foo)
      => t

This behaviour is compatible with other code which treats symbols
beginning with colon as keywords only if they are interned in the
global obarray.  `keywordp' used to wrongly return t in both cases
above.

** The first argument to `intern-soft' may now also be a symbol, like
with `unintern'.  If given a symbol, `intern-soft' will look for that
exact symbol rather than for any string.  This is useful when you want
to check whether a specific symbol is interned in an obarray, e.g.:

    (intern "foo")
    (intern-soft "foo")
      => foo
    (intern-soft (make-symbol "foo"))
      => nil

** Functions for decoding base64 encoding are now available; see
`base64-encode-region', `base64-encode-string', `base64-decode-region'
and `base64-decode-string'.

\f
* Changes in XEmacs 21.0
========================

** XEmacs has been unbundled into constituent installable packages.
See the Info documentation under "Packages" for more information.
See the file `etc/PACKAGES' in the distribution for a partial list of
packages available at the time of the 21.0 release.

** XEmacs is now supported under Microsoft Windows 95/98 and Windows
NT operating systems.  For starters, look at the XEmacs on Windows FAQ
at <URL:http://jagor.srce.hr/~hniksic/xemacs-on-windows-faq.txt>.  To
discuss Windows-specific issues, subscribe to the mailing list at
<xemacs-nt-request@xemacs.org>.

** XEmacs will now use `XEmacs' as its application class if it finds
any `XEmacs' resources in the resource database.  Otherwise, it will
continue to use the `Emacs' class.

** The options menu has been ported to Custom.
This means that each entry in the options menu acts as if you had customized
the corresponding variable by hand.  ### WARNING: there is currently no
upgrading function to help you port your old options settings to the new
format. Consequently, if you want to modify the options for XEmacs 21, you
will have to set them all again through the menu, and remove the code loading
.xemacs-options from your .emacs.

** When the Zmacs region is active, `M-x query-replace' and the other
replace commands now operate on the region contents only.

** XEmacs now is able to choose X visuals and use private colormaps.
The '-visual <visualStr>' command line option or the '.EmacsVisual'
Xresource controls which visual XEmacs will use, and
'-privateColormap' or '.privateColormap' will force XEmacs to create a
private colormap for use.  The syntax for the visual string is
"<visual><bitdepth>" where <visual> is one of 'StaticColor',
'TrueColor', 'GrayScale', 'PseudoColor' or 'DirectColor' and
<bitdepth> is the appropriate number of bits per pixel.  If an invalid
or non-supported combination is entered, XEmacs attempts to find a happy
medium.  The X creation mechanism will then determine if it needs to
create a colormap for use, or the presence of the private flags will
force it to create it.

** The `imenu' package has been ported to XEmacs and is available as a
package.

** `echo-keystrokes' can now be a floating-point number, so that you
can set it to intervals shorter than one second.

   (setq echo-keystrokes 0.1)

** The new command `center-to-window-line' works like `recenter'
(bound to `C-l'), only it does not redisplay the whole display area.

** The M-. command will now first search through exact tags matches,
and then through inexact matches, as one would expect.

** The new variable `user-full-name' can be used to customize one's
name when using the Emacs mail and news reading facilities.

Normally, `user-full-name' is a function that returns the full name of
a user or UID, as specified by the system -- for instance,
(user-full-name "root") returns something like "Super-User".  However,
when the function is called without arguments, it will return the
value of the `user-full-name' variable.  The `user-full-name' variable
is initialized using the environment variable NAME and (failing that)
the user's system name.

The behavior of the `user-full-name' function with an argument
specified is unchanged.

** The new command `M-x customize-changed-options' lets you customize
all the options whose default values have changed in recent Emacs
versions.  You specify a previous Emacs version number as argument,
and the command creates a customization buffer showing all the
customizable options whose default values were changed since that
version.

If you don't specify a particular version number argument, then the
customization buffer shows all the customizable options for which
Emacs versions of changes are recorded.

** The new command `add-log-convert' can be used to convert the
old-style (pre-20.3) ChangeLog buffers to new style, for
consistency.  A reminder: if you wish to revert to old-style
ChangeLogs instead, customize the value of `add-log-time-format'
variable.

** The new command `zap-up-to-char' is now available.  It is similar
to `zap-to-char', except that it does not delete the searched-for
character.  It is not bound to a key by default.

** You can now store a number into a register with `C-u NUMBER C-x r n'
REG, increment it by INC with `C-u INC C-x r + REG' (to increment by
one, omit C-u INC), and insert it in the buffer with `C-x r g REG'.
This is useful for writing keyboard macros.

** The M-: command, when given a prefix argument, will now insert its
result to the current buffer.

** The `C-h c' command, when given a prefix argument, will now insert
the message into the current buffer.

** Horizontally split windows may now be dragged using the mouse.
Because of this, the dividers between vertical windows are always
visible.  To turn it off, set `vertical-divider-always-visible-p' to
nil.

** XEmacs/Mule (internationalization) changes.

*** Mule support now works on TTY's.  Use `set-terminal-coding-system'
and `set-keyboard-coding-system' to specify the coding system of your
display and keyboard.

*** Egg/SJ3 input method is now officially supported.  Quail and
Egg/Skk have been available through the generalized Leim since 20.3.

*** Localized Japanese menubars are available if XEmacs is built with
XFONTSET and either the X11 libraries are built with X_LOCALE defined
or the native C libraries support Japanese localization.  This has
been available since 20.3, only it hasn't been announced before.

** Jamie Zawinski's `gdb-highlight' extension is now distributed with
the `debug' package.  gdb-highlight makes most objects printed in a
gdb buffer be mouse-sensitive: as text shows up in the buffer, it is
parsed, and objects which are recognized have context-sensitive
commands attached to them.  To use it, add the following to `.emacs':

  (add-hook 'gdb-mode-hook (lambda () (require 'gdb-highlight)))

** The package popper.el is now included in the edit-utils package.
It has been greatly enhanced with respect to the one once included
with the ilisp package and should work well under XEmacs 21.0.

** Gnuserv changes

*** Like the old 'gnudoit' program. Gnuclient -batch now can read from stdin.

*** Gnuclient -batch no longer breaks off the output at the first LF.

** C mode changes

*** Multiline macros are now handled, both as they affect indentation,
and as recognized syntax.  New syntactic symbol cpp-macro-cont is
assigned to second and subsequent lines of a multiline macro
definition.

*** A new style "user" which captures all non-hook-ified
(i.e. top-level) .emacs file variable setings and customizations.
Style "cc-mode" is an alias for "user" and is deprecated.  "gnu" style
is still the default however.

*** "java" style now conforms to Sun's JDK coding style.

*** There are new commands c-beginning-of-defun, c-end-of-defun which
are alternatives which you could bind to C-M-a and C-M-e if you prefer
them.  They do not have key bindings by default.

*** New and improved implementations of M-a (c-beginning-of-statement)
and M-e (c-end-of-statement).

*** C++ namespace blocks are supported, with new syntactic symbols
namespace-open, namespace-close, and innamespace.

*** File local variable settings of c-file-style and c-file-offsets
makes the style variables local to that buffer only.

*** New indentation functions c-lineup-close-paren,
c-indent-one-line-block, c-lineup-dont-change.

*** Improvements (hopefully!) to the way CC Mode is loaded.  You
should now be able to do a (require 'cc-mode) to get the entire
package loaded properly for customization in your .emacs file.  A new
variable c-initialize-on-load controls this and is t by default.

** In Text mode, now only blank lines separate paragraphs.
This makes it possible to get the full benefit of Adaptive Fill mode
in Text mode, and other modes derived from it (such as Mail mode).
TAB in Text mode now runs the command indent-relative; this makes a
practical difference only when you use indented paragraphs.

As a result, the old Indented Text mode is now identical to Text mode,
and is an alias for it.

If you want spaces at the beginning of a line to start a paragraph,
use the new mode, Paragraph Indent Text mode.

** Changes to Gnus, the XEmacs newsreader.

*** New functionality for using Gnus as an offline newsreader has been
added.  A plethora of new commands and modes have been added.  See the
Gnus manual for the full story.

*** The nndraft backend has returned, but works differently than
before.  All Message buffers are now also articles in the nndraft
group, which is created automatically.

*** `gnus-alter-header-function' can now be used to alter header
values.

*** `gnus-summary-goto-article' now accept Message-ID's.

*** A new Message command for deleting text in the body of a message
outside the region: `C-c C-v'.

*** You can now post to component group in nnvirtual groups with
`C-u C-c C-c'.

*** `nntp-rlogin-program' -- new variable to ease customization.

*** `C-u C-c C-c' in `gnus-article-edit-mode' will now inhibit
re-highlighting of the article buffer.

*** New element in `gnus-boring-article-headers' -- `long-to'.

*** `M-i' symbolic prefix command.  See the section "Symbolic
Prefixes" in the Gnus manual for details.

*** `L' and `I' in the summary buffer now take the symbolic prefix
`a' to add the score rule to the "all.SCORE" file.

*** `gnus-simplify-subject-functions' variable to allow greater
control over simplification.

*** `A T' -- new command for fetching the current thread.

*** `/ T' -- new command for including the current thread in the
limit.

*** `M-RET' is a new Message command for breaking cited text.

*** \\1-expressions are now valid in `nnmail-split-methods'.

*** The `custom-face-lookup' function has been removed.
If you used this function in your initialization files, you must
rewrite them to use `face-spec-set' instead.

*** Cancelling now uses the current select method.  Symbolic prefix
`a' forces normal posting method.

*** New command to translate M******** sm*rtq**t*s into proper text
-- `W d'.

*** For easier debugging of nntp, you can set `nntp-record-commands'
to a non-nil value.

*** nntp now uses ~/.authinfo, a .netrc-like file, for controlling
where and how to send AUTHINFO to NNTP servers.

*** A command for editing group parameters from the summary buffer
has been added.

*** A history of where mails have been split is available.

*** A new article date command has been added -- `article-date-iso8601'.

*** Subjects can be simplified when threading by setting
`gnus-score-thread-simplify'.

*** A new function for citing in Message has been added --
`message-cite-original-without-signature'.

*** `article-strip-all-blank-lines' -- new article command.

*** A new Message command to kill to the end of the article has
been added.

*** A minimum adaptive score can be specified by using the
`gnus-adaptive-word-minimum' variable.

*** The "lapsed date" article header can be kept continually
updated by the `gnus-start-date-timer' command.

*** Web listserv archives can be read with the nnlistserv backend.

*** Old dejanews archives can now be read by nnweb.

*** Byte-compilation of user-specs now works under XEmacs.

** The `dir' files are no longer essential for functioning of the Info
subsystem.  If the `dir' file does not exist in an Info directory, the
relevant information will be generated on-the-fly.

This behavior can be customized, look for `Info-auto-generate-directory'
and `Info-save-auto-generated-dir' in the `info' customization group.

\f
* Lisp and internal changes in XEmacs 21.0
==========================================

** It is now possible to build XEmacs with support for 31-bit Lisp
integers (normally, Lisp integers are only 28 bits wide on 32-bit
machines.)  Configure with --use-minimal-tagbits to test.  With this
change, the maximum buffer size on 32-bit machines is increased from
128M to 1G.  This setting will be made default in a future XEmacs
version.

** Specifier changes.

*** When instantiating a specifier, the window locale now has a higher
precedence than the buffer locale.  This is because the window locale
is more specific than the buffer locale.

*** The new macro `let-specifier' can be used to temporarily add
specifications to specifiers.  See the documentation for details.

*** The new specifiers `vertical-scrollbar-visible-p' and
`horizontal-scrollbar-visible-p' may be used to control scrollbar
visibility.  Previously, the only way to remove a scrollbar was to set
its size to 0.  This method is still supported for backward
compatibility.

*** The new specifiers `scrollbar-on-left-p' and `scrollbar-on-top-p'
may be used to control the position of the vertical and horizontal
toolbar.  Previously, their position could be changed only through the
use of X resources.

*** The new draggable vertical dividers between windows may be turned
off using the `vertical-divider-always-visible-p' specifier.  When
this is set to nil, the vertical dividers between windows are shown
only when needed, and they are not draggable.

Other properties of the vertical dividers may be controlled using
`vertical-divider-shadow-thickness', `vertical-divider-line-width' and
`vertical-divider-spacing' specifiers, which see.

** Frame focus management changes.

*** When the variable focus-follows-mouse is non-nil, `select-frame'
no longer permanently selects a different frame.  The frame selection
is temporary and is reverted when the current command terminates, much
like the buffer selected by `set-buffer'.  This is the same as in FSF
Emacs.

*** The new function `focus-frame' sets the window system focus to
FRAME (and selects it), regardless of the value of
`focus-follows-mouse'.  Doing this is not well behaved, so be
absolutely sure that you want this.

The code that uses `select-frame' only to get the window manager focus
should be changed to use `set-frame-focus' instead, so that they keep
working when `focus-follows-mouse' is non-nil.

*** The special forms `save-selected-frame' and `with-selected-frame'
can now be used to temporarily change selected frame.

*** The behavior of `other-frame' command (`C-x 5 o') is unaffected by
these changes.

** The function `select-window' now has an optional second argument
NORECORD which if non-nil inhibits the recording of a buffer change.

** The function `vertical-motion' now correctly handles the second,
optional WINDOW argument.  A new third argument PIXELS, if non-nil,
indicates that the returned motion should be in pixels.

** The new function `vertical-motion-pixels' is similar to
vertical-motion but takes as input a vertical motion in pixels.

** The new functions window-text-area-pixel-{width,height,edges} can
be used to obtain information about the text-displaying area of a
window.

** The new functions `shrink-window-pixels' and `enlarge-window-pixels'
can be used to adjust the size of a window by a pixel amount.

** The new function `window-displayed-text-pixel-height' can be used
to determine the height of the text actually displayed in a window.

** The arithmetic comparison functions <, >, =, /= now accept a
variable number of arguments.

This means that if you want to test whether A < B < C, you can write
it as (< A B C) instead of (and (< A B) (< B C)).  Likewise,
(apply #'> LIST) now tests if LIST is monotonously increasing -- and
so on.

** The XEmacs hashtables now have a consistent read/print syntax.
This means that a hashtable will be readably printed in a
structure-like form:

   #s(hashtable size 2 data (key1 value1 key2 value2))

When XEmacs reads this form, it will create a new hashtable according
to description.  This allows you to easily dump hashtables to files
using `prin1', and read them back in using `read'.

If `print-readably' is non-nil, a more relaxed syntax is used; for
instance:

   #<hashtable size 2/13 data (key1 value1 key2 value2) 0x874d>

** It is now possible to build XEmacs with LDAP support.
You will need to install a LDAP library first.  The following have
been tested:
  - LDAP 3.3 from the University of Michigan
    (get it from <URL:http://www.umich.edu/~dirsvcs/ldap/>)
  - OpenLDAP 1.0.3 from the OpenLDAP Foundation
    (get it from <URL:http://www.openldap.org/>)
  - LDAP SDK 1.0 from Netscape Corp.
    (get it from <URL:http://developer.netscape.com/>)

** When profiling is in effect, a call-count of all recorded functions
is now calculated.  This information is stored in
`call-count-profile-table', and is utilized by `profile-results' as
well as the new command `profile-call-count-results'.

** It is now an error to change the value of a symbol whose name
starts with a colon, if it is interned in the standard obarray.

However, setting such a symbol to its proper value, which is that
symbol itself, is not an error.  This is for the sake of programs that
support pre-19.12 XEmacs and pre-20 GNU Emacs by explicitly setting
these variables to themselves.

** The `concat' function no longer accepts integer arguments.

** The new function `string' concatenates all its argument characters
and returns the resulting string.  This is consistent with other
functions, like `list', `vector', etc.

** The function `temp-directory' is now available to return the
directory to store temporary files.  On Unix this will be obtained
from TMPDIR, defaulting to `/tmp'.

** The function load-average now accepts an optional argument
USE-FLOATS.  If it is non-nil, the load average values are returned as
floating point numbers, rather than as integers to be divided by 100.

** The `make-event' function now supports the TYPE and PLIST
arguments, which can be used to create various events from Lisp.  See
the documentation for details.

** `function-interactive' is a new function that returns the
interactive specification of a funcallable object.

** The new `lmessage' function allows printing of a formatted message
with a particular label.

   (lmessage 'progress "Processing... %d" counter)

This function is more convenient than `display-message' because it
automatically applies `format' to its arguments.

** The new `lwarn' function, analogous to `lmessage', allows printing
a formatted warning, with a non-default CLASS or LABEL.

** The new function `split-path' can now be used to explode the
components of a colon-separated search path into a list.

   (split-path "foo:bar")
     => ("foo" "bar")

** Specifiers and symbols whose value is a specifier are now allowed
as modeline specifications.

** defcustom now accepts the keyword `:version'.  Use this to specify
in which version of Emacs a certain variable's default value changed.
For example,

   (defcustom foo-max 34 "*Maximum number of foo's allowed."
     :type 'integer
     :group 'foo
     :version "21.0")

This information is used to control the customize-changed-options
command.

** The line number tracking in modeline is now efficient, even for
very large buffers.  This is achieved by caching the line numbers of
recent buffer positions, and reusing them.  This cache is used only in
the buffers where `line-number-mode' is in effect.

** When the new GNU Malloc aka Doug Lea Malloc is available, it will
be used.  This should result in better performance on Linux systems
with libc6.

** The code XEmacs uses to assemble its various paths into the
directory hierarchy has been rewritten to support the package system.
Look under "Startup Paths" in the Info documentation for more
information.

*** site-lisp is now longer part of the load-path by default.
Its use is deprecated, but you can specify --with-site-lisp=yes at the
configure command line to get it back.

*** `Info-default-directory-list' is now obsolete.  If you want to
change the path which XEmacs uses to search for info files, set
`Info-directory-list' instead.