/[emacs]/emacs/etc/NEWS
ViewVC logotype

Diff of /emacs/etc/NEWS

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 1.613 by eliz, Tue Feb 26 16:51:30 2002 UTC revision 1.613.2.1 by fx, Fri Jul 5 22:18:13 2002 UTC
# Line 1  Line 1 
1  GNU Emacs NEWS -- history of user-visible changes.  2001-03-15  GNU Emacs NEWS -- history of user-visible changes.  2002-0705
2  Copyright (C) 1999, 2000, 2001, 2002 Free Software Foundation, Inc.  Copyright (C) 2002 Free Software Foundation, Inc.
3  See the end for copying conditions.  See the end for copying conditions.
4    
5  Please send Emacs bug reports to bug-gnu-emacs@gnu.org.  Please send Emacs bug reports to bug-gnu-emacs@gnu.org.
# Line 12  When you add a new item, please add it w Line 12  When you add a new item, please add it w
12  so we will look at it  so we will look at it
13    
14    
15  * Installation Changes in Emacs 21.3  * Changes in Emacs 22.1
16    
17  ** Emacs now supports new configure options `--program-prefix',  ** The Emacs character set is now a superset of Unicode (it has about
18  `--program-suffix' and `--program-transform-name' that affect the names of  four times the code space, which should be plenty).
 installed programs.  
19    
20  ---  The internal encoding used for buffers and strings is now
21  ** Leim is now part of the Emacs distribution.  Unicode-based and called `utf-8-emacs'.  utf-8-emacs is backwards
22  You no longer need to download a separate tarball in order to build  compatible with the UTF-8 encoding of Unicode.  The `emacs-mule'
23  Emacs with Leim.  coding system can still read and write data in the old internal
24    encoding.
25    
26  ---  There are still charsets which contain disjoint sets of characters
27  ** Support for FreeBSD/Alpha has been added.  where this is necessary or useful, especially for various Far Eastern
28    sets which are problematic with Unicode.
29    
30  ---  Since the internal encoding is also used by default for byte-compiled
31  ** Support for BSD/OS 5.0 was added.  files -- i.e. the normal coding system for byte-compiled Lisp files is
32    now utf-8-Emacs -- Lisp containing non-ASCII characters which is
33    compiled by Emacs 22 can't be read by earlier versions of Emacs.
34    Files compiled by Emacs 20 or 21 are loaded correctly as emacs-mule
35    (whether or not they contain multibyte characters), which makes
36    loading them somewhat slower than Emacs 22-compiled files.  Thus it
37    may be worth recompiling existing .elc files which don't need to be
38    shared with older Emacsen.
39    
40    ** There are assorted new coding systems/aliases -- see
41  * Changes in Emacs 21.3  M-x list-coding-systems.
   
 +++  
 ** Emacs now supports ICCCM Extended Segments in X selections.  
   
 Some versions of X, notably XFree86, use Extended Segments to encode  
 in X selections characters that belong to character sets which are not  
 part of the list of standard charsets supported by the ICCCM spec.  
 Examples of such non-standard character sets include ISO 8859-14, ISO  
 8859-15, KOI8-R, and BIG5.  The new coding system  
 `compound-text-with-extensions' supports these extensions, and is now  
 used by default for encoding and decoding X selections.  If you don't  
 want this support, set `selection-coding-system' to `compound-text'.  
   
 +++  
 ** The parameters of automatic hscrolling can now be customized.  
 The variable `automatic-hscroll-margin' determines how many columns  
 away from the window edge point is allowed to get before automatic  
 hscrolling will horizontally scroll the window.  The default value is 5.  
   
 The variable `automatic-hscroll-step' determines how many columns  
 automatic hscrolling will scroll the window when point gets too close  
 to the window edge.  If its value is zero, the default, Emacs scrolls  
 the window so as to center point.  If its value is an integer, it says  
 how many columns to scroll.  If the value is a floating-point number,  
 it gives the fraction of the window's width to scroll the window.  
   
 ** New display feature: focus follows mouse.  If you set the variable  
 x-autoselect-window to non-nil value, moving the mouse to a different  
 Emacs window will select that window.  The default is nil, so that  
 this feature is not enabled.  
   
 ** The new command `describe-text-at' pops up a buffer with description  
 of text properties, overlays, and widgets at point, and lets you get  
 more information about them, by clicking on mouse-sensitive areas or  
 moving there and pressing RET.  
   
 ** Emacs normally highlights mouse sensitive text whenever the mouse  
 is over the text.  By setting the new variable `mouse-highlight', you  
 can optionally enable mouse highlighting only after you move the  
 mouse, so that highlighting disappears when you press a key.  You can  
 also disable mouse highlighting.  
   
 ** font-lock: in modes like C and Lisp where the fontification assumes that  
 an open-paren in column 0 is always outside of any string or comment,  
 font-lock now highlights any such open-paren-in-column-zero in bold-red  
 if it is inside a string or a comment, to indicate that it can cause  
 trouble with fontification and/or indentation.  
   
 +++  
 ** There's a new face `minibuffer-prompt'.  
 Emacs adds this face to the list of text properties stored in the  
 variable `minibuffer-prompt-properties', which is used to display the  
 prompt string.  
   
 +++  
 ** The new face `mode-line-inactive' is used to display the mode line  
 of non-selected windows.  The `mode-line' face is now used to display  
 the mode line of the currently selected window.  
   
 The new variable `mode-line-in-non-selected-windows' controls whether  
 the `mode-line-inactive' face is used.  
   
 ** A menu item "Show/Hide" was added to the top-level menu "Options".  
 This menu allows you to turn various display features on and off (like  
 tool bar and the menu bar itself).  You can also move the vertical  
 scroll bar to either side here or turn it off completely.  There is also  
 a menu-item to toggle displaying of current date and time, current line  
 and column number in the mode-line.  
   
 ** Speedbar has moved from the "Tools" top level menu to "Show/Hide".  
   
 ** Emacs can now indicate in the mode-line the presence of new e-mails in  
 directory in addition to file.  See the documentation of the user option  
 `display-time-mail-directory'.  
   
 +++  
 ** The new option `Info-scroll-prefer-subnodes' causes Info to behave  
 like the stand-alone Info reader (from the GNU Texinfo package) as far  
 as motion between nodes and their subnodes is concerned.  If it is t  
 (the default), Emacs behaves as before when you type SPC in a menu: it  
 visits the subnode pointed to by the first menu entry.  If this option  
 is nil, SPC scrolls to the end of the current node, and only then goes  
 to the first menu item, like the stand-alone reader does.  
   
 This change was already in Emacs 21.1, but wasn't advertised in the  
 NEWS.  
   
 ---  
 ** LDAP support now defaults to ldapsearch from OpenLDAP version 2.  
   
 +++  
 ** You can now disable pc-selection-mode after enabling it.  
 M-x pc-selection-mode behaves like a proper minor mode, and with no  
 argument it toggles the mode.  
   
 Turning off PC-Selection mode restores the global key bindings  
 that were replaced by turning on the mode.  
   
 ** Changes in support of colors on character terminals  
   
 +++  
 *** The new command-line option --color=MODE lets you specify a standard  
 mode for a tty color support.  It is meant to be used on character  
 terminals whose capabilities are not set correctly in the terminal  
 database, or with terminal emulators which support colors, but don't  
 set the TERM environment variable to a name of a color-capable  
 terminal.  "emacs --color" uses the same color commands as GNU `ls'  
 when invoked with "ls --color", so if your terminal can support colors  
 in "ls --color", it will support "emacs --color" as well.  See the  
 user manual for the possible values of the MODE parameter.  
   
 ---  
 *** Emacs now supports several character terminals which provide more  
 than 8 colors.  For example, for `xterm', 16-color, 88-color, and  
 256-color modes are supported.  Emacs automatically notes at startup  
 the extended number of colors, and defines the appropriate entries for  
 all of these colors.  
   
 ---  
 *** There's a new support for colors on `rxvt' terminal emulator.  
   
 +++  
 ** Emacs can now be invoked in full-screen mode on a windowed display.  
   
 When Emacs is invoked on a window system, the new command-line options  
 `--fullwidth', `--fullheight', and `--fullscreen' produce a frame  
 whose width, height, or both width and height take up the entire  
 screen size.  (For now, this works only on GNU and Unix systems, and  
 not with every window manager.)  
   
 ** Info-index finally offers completion.  
   
 ** shell-mode now supports programmable completion using `pcomplete'.  
   
 ** Controlling the left and right fringe widths.  
   
 The left and right fringe widths can now be controlled by setting the  
 `left-fringe' and `right-fringe' frame parameters to an integer value  
 specifying the width in pixels.  Setting the width to 0 effectively  
 removes the corresponding fringe.  
   
 The actual fringe widths may deviate from the specified widths, since  
 the combined fringe widths must match an integral number of columns.  
 The extra width is distributed evenly between the left and right fringe.  
 For force a specific fringe width, specify the width as a negative  
 integer (if both widths are negative, only the left fringe gets the  
 specified width).  
   
 Setting the width to nil (the default), restores the default fringe  
 width which is the minimum number of pixels necessary to display any  
 of the currently defined fringe bitmaps.  The width of the built-in  
 fringe bitmaps is 8 pixels.  
   
 ** Changes in C-h bindings:  
   
 C-h e displays the *Messages* buffer.  
   
 C-h followed by a control character is used for displaying files  
     that do not change:  
   
 C-h C-f displays the FAQ.  
 C-h C-e displays the PROBLEMS file.  
   
 The info-search bindings on C-h C-f, C-h C-k and C-h C-i  
 have been moved to C-h F, C-h K and C-h S.  
   
 C-h c, C-h k, C-h w, and C-h f now handle remapped interactive commands.  
   
 - C-h c and C-h k report the actual command (after possible remapping)  
   run by the key sequence.  
   
 - C-h w and C-h f on a command which has been remapped now report the  
   command it is remapped to, and the keys which can be used to run  
   that command.  
   
 For example, if C-k is bound to kill-line, and kill-line is remapped  
 to new-kill-line, these commands now report:  
   
 - C-h c and C-h k C-k reports:  
   C-k runs the command new-kill-line  
   
 - C-h w and C-h f kill-line reports:  
   kill-line is remapped to new-kill-line which is on C-k, <deleteline>  
   
 - C-h w and C-h f new-kill-line reports:  
   new-kill-line is on C-k  
   
 ** C-w in incremental search now grabs either a character or a word,  
 making the decision in a heuristic way.  This new job is done by the  
 command `isearch-yank-word-or-char'.  To restore the old behavior,  
 bind C-w to `isearch-yank-word' in `isearch-mode-map'.  
   
 ** In GUD mode when talking to GDB, C-x C-a C-j "jumps" the program  
 counter to the specified source line (the one where point is).  
   
 ** GUD mode improvements for jdb:  
   
 *** Search for source files using jdb classpath and class  
     information. Fast startup since there is no need to scan all  
     source files up front. There is also no need to create and maintain  
     lists of source directories to scan. Look at `gud-jdb-use-classpath'  
     and `gud-jdb-classpath' customization variables documentation.  
   
 *** Supports the standard breakpoint (gud-break, gud-clear)  
     set/clear operations from java source files under the classpath, stack  
     traversal (gud-up, gud-down), and run until current stack finish  
     (gud-finish).  
   
 *** Supports new jdb (Java 1.2 and later) in addition to oldjdb  
     (Java 1.1 jdb).  
   
 *** The previous method of searching for source files has been  
     preserved in case someone still wants/needs to use it.  
     Set gud-jdb-use-classpath to nil.  
   
   Added Customization Variables  
   
 *** gud-jdb-command-name.  What command line to use to invoke jdb.  
   
 *** gud-jdb-use-classpath. Allows selection of java source file searching  
     method: set to t for new method, nil to scan gud-jdb-directories for  
     java sources (previous method).  
   
 *** gud-jdb-directories. List of directories to scan and search for java  
     classes using the original gud-jdb method (if gud-jdb-use-classpath  
     is nil).  
   
   Minor Improvements  
   
 *** Do not allow debugger output history variable to grow without bounds.  
   
 ** hide-ifdef-mode now uses overlays rather than selective-display  
 to hide its text.  This should be mostly transparent but slightly  
 changes the behavior of motion commands line C-e and C-p.  
   
 ** In Dired's ! command (dired-do-shell-command), `*' and `?' now  
 control substitution of the file names only when they are surrounded  
 by whitespace.  This means you can now use them as shell wildcards  
 too.  If you want to use just plain `*' as a wildcard, type `*""'; the  
 doublequotes make no difference in the shell, but they prevent  
 special treatment in `dired-do-shell-command'.  
   
 ** Unquoted `$' in file names do not signal an error any more when  
 the corresponding environment variable does not exist.  
 Instead, the `$ENVVAR' text is left as is, so that `$$' quoting  
 is only rarely needed.  
   
 ** jit-lock can now be delayed with `jit-lock-defer-time'.  
   
 If this variable is non-nil, its value should be the amount of Emacs  
 idle time in seconds to wait before starting fontification.  For  
 example, if you set `jit-lock-defer-time' to 0.25, fontification will  
 only happen after 0.25s of idle time.  
   
 +++  
 ** Marking commands extend the region when invoked multiple times.  If  
 you hit M-C-SPC (mark-sexp), M-@ (mark-word), M-h (mark-paragraph), or  
 C-M-h (mark-defun) repeatedly, the marked region will now be extended  
 each time, so you can mark the next two sexps with M-C-SPC M-C-SPC,  
 for example.  This feature also works for mark-end-of-sentence, if you  
 bind that to a key.  
   
 ** In the *Occur* buffer, `o' switches to it in another window, and  
 C-o displays the current line's occurrence in another window without  
 switching to it.  
   
 ** When you specify a frame size with --geometry, the size applies to  
 all frames you create.  A position specified with --geometry only  
 affects the initial frame.  
   
 +++  
 ** M-h (mark-paragraph) now accepts a prefix arg.  
 With positive arg, M-h marks the current and the following paragraphs;  
 if the arg is negative, it marks the current and the preceding  
 paragraphs.  
   
 ** In Dired, the w command now copies the current line's file name  
 into the kill ring.  
   
 ** The variables dired-free-space-program and dired-free-space-args  
 have been renamed to directory-free-space-program and  
 directory-free-space-args, and they now apply whenever Emacs puts a  
 directory listing into a buffer.  
   
 ** mouse-wheels can now scroll a specific fraction of the window  
 (rather than a fixed number of lines) and the scrolling is `progressive'.  
   
 ** The keyboard-coding-system is now automatically set based on  
 your current locale settings.  If it turns out that your terminal  
 does not support the encoding implied by your locale (for example,  
 it inserts non-ASCII chars if you hit M-i), you will need to add  
   
         (set-keyboard-coding-system nil)  
   
 to your .emacs to revert to the old behavior.  
   
 ** A new coding system `euc-tw' has been added for traditional Chinese  
 in CNS encoding; it accepts both Big 5 and CNS as input; on saving,  
 Big 5 is then converted to CNS.  
   
 +++  
 ** Emacs now reads the standard abbrevs file ~/.abbrev_defs  
 automatically at startup, if it exists.  When Emacs offers to save  
 modified buffers, it saves the abbrevs too if they have changed.  It  
 can do this either silently or asking for confirmation first,  
 according to the value of `save-abbrevs'.  
   
 ** Display of hollow cursors now obeys the buffer-local value (if any)  
 of `cursor-in-non-selected-windows' in the buffer that the cursor  
 appears in.  
   
 ** The default values of `tooltip-delay' and `tooltip-hide-delay'  
 were changed.  
   
 ** On terminals whose erase-char is ^H (Backspace), Emacs  
 now uses normal-erase-is-backspace-mode.  
   
 ** Etags changes.  
   
 *** In Prolog, etags creates tags for rules in addition to predicates.  
   
 *** In Perl, packages are tags.  Subroutine tags are named from their  
 package.  You can jump to sub tags as you did before, by the sub name, or  
 additionally by looking for package::sub.  
   
 *** New language PHP: tags are functions, classes and defines.  If  
 the --members option is specified to etags, tags are vars also.  
   
 +++  
 ** The command line option --no-windows has been changed to  
 --no-window-system.  The old one still works, but is deprecated.  
   
 ** `buffer-menu' and `list-buffers' now list buffers whose names begin  
 with a space, if they visit files.  
   
 ** You can now customize fill-nobreak-predicate to control where  
 filling can break lines.  We provide two sample predicates,  
 fill-single-word-nobreak-p and fill-french-nobreak-p.  
   
 ** New user option `add-log-always-start-new-record'.  
 When this option is enabled, M-x add-change-log-entry will always  
 start a new record regardless of when the last record is.  
   
 ** New user option `sgml-xml'.  
 When this option is enabled, SGML tags are inserted in XML style,  
 i.e., there is always a closing tag.  
 When not customized, it becomes buffer-local when it can be inferred  
 from the file name or buffer contents.  
   
 ** New user option `isearch-resume-enabled'.  
 This option can be disabled, to avoid the normal behaviour of isearch  
 which puts calls to `isearch-resume' in the command history.  
   
 ** When the *scratch* buffer is recreated, its mode is set from  
 initial-major-mode, which normally is lisp-interaction-mode,  
 instead of using default-major-mode.  
   
 ** Byte compiler warning and error messages have been brought more  
 in line with the output of other GNU tools.  
   
 ** Lisp-mode now uses font-lock-doc-face for the docstrings.  
   
 ** perl-mode has a new variable `perl-indent-continued-arguments'.  
   
 ** `special-display-buffer-names' and `special-display-regexps' now  
 understand two new boolean pseudo-frame-parameters `same-frame' and  
 `same-window'.  
   
 ** When pure storage overflows while dumping, Emacs now prints how  
 much pure storage it will approximately need.  
   
 ** M-x setenv now expands environment variables of the form `$foo' and  
 `${foo}' in the specified new value of the environment variable.  To  
 include a `$' in the value, use `$$'.  
   
 +++  
 ** File-name completion can now ignore directories.  
 If an element of the list in `completion-ignored-extensions' ends in a  
 slash `/', it indicates a subdirectory that should be ignored when  
 completing file names.  Elements of `completion-ignored-extensions'  
 which do not end in a slash are never considered when a completion  
 candidate is a directory.  
   
 ** New user option `inhibit-startup-buffer-menu'.  
 When loading many files, for instance with `emacs *', Emacs normally  
 displays a buffer menu.  This option turns the buffer menu off.  
   
 ** Rmail now displays 5-digit message ids in its summary buffer.  
   
 ** When using M-x revert-buffer in a compilation buffer to rerun a  
 compilation, it is now made sure that the compilation buffer is reused  
 in case it has been renamed.  
   
 ** On MS Windows, the "system caret" now follows the cursor.  
 This enables Emacs to work better with programs that need to track  
 the cursor, for example screen magnifiers and text to speech programs.  
   
 ** Tooltips now work on MS Windows.  
 See the Emacs 21.1 NEWS entry for tooltips for details.  
   
 ** Pointing devices with more than 3 buttons are now supported on MS Windows.  
 The new variable `w32-pass-extra-mouse-buttons-to-system' controls  
 whether Emacs should handle the extra buttons itself (the default), or  
 pass them to Windows to be handled with system-wide functions.  
   
 ** Under X11, it is possible to swap Alt and Meta (and Super and Hyper).  
 The new variables `x-alt-keysym', `x-hyper-keysym', `x-meta-keysym',  
 and `x-super-keysym' can be used to choose which keysyms Emacs should  
 use for the modifiers.  For example, the following two lines swap  
 Meta and Alt:  
     (setq x-alt-keysym 'meta)  
     (setq x-meta-keysym 'alt)  
   
 ---  
 ** A French translation of the `Emacs Survival Guide' is available.  
   
 ---  
 ** A French translation of the Emacs Tutorial is available.  
   
 ** New modes and packages  
   
 +++  
 *** Calc is now part of the Emacs distribution.  
   
 Calc is an advanced desk calculator and mathematical tool written in  
 Emacs Lisp.  Its documentation is in a separate manual; within Emacs,  
 type "C-h i m calc RET" to read that manual.  A reference card is  
 available in `etc/calccard.tex' and `etc/calccard.ps'.  
   
 +++  
 *** The Emacs Lisp Reference Manual is now part of the distribution.  
   
 The ELisp reference manual in Info format is built as part of the  
 Emacs build procedure and installed together with the Emacs User  
 Manual.  A menu item was added to the menu bar that makes it easy  
 accessible (Help->More Manuals->Emacs Lisp Reference).  
   
 *** The Introduction to Programming in Emacs Lisp manual is now part of  
 the distribution.  
   
 This manual is now part of the standard distribution and is installed,  
 together with the Emacs User Manual, into the Info directory.  A menu  
 item was added to the menu bar that makes it easy accessible  
 (Help->More Manuals->Introduction to Emacs Lisp).  
   
 *** The ruler-mode.el library provides a minor mode for displaying an  
 "active" ruler in the header line.  You can use the mouse to visually  
 change the `fill-column', `window-margins' and `tab-stop-list'  
 settings.  
   
 *** The reveal.el package provides the minor modes `reveal-mode' and  
 `global-reveal-mode' which will make text visible on the fly as you  
 move your cursor into hidden region of the buffer.  
 It should work with any package that uses overlays to hide parts  
 of a buffer, such as outline-minor-mode, hs-minor-mode, hide-ifdef-mode, ...  
   
 *** master-mode.el implements a minor mode for scrolling a slave  
 buffer without leaving your current buffer, the master buffer.  
   
 It can be used by sql.el, for example: the SQL buffer is the master  
 and its SQLi buffer is the slave.  This allows you to scroll the SQLi  
 buffer containing the output from the SQL buffer containing the  
 commands.  
   
 This is how to use sql.el and master.el together: the variable  
 sql-buffer contains the slave buffer.  It is a local variable in the  
 SQL buffer.  
   
 (add-hook 'sql-mode-hook  
    (function (lambda ()  
                (master-mode t)  
                (master-set-slave sql-buffer))))  
 (add-hook 'sql-set-sqli-hook  
    (function (lambda ()  
                (master-set-slave sql-buffer))))  
   
   
 * Lisp Changes in Emacs 21.3  
   
 ** Interactive commands can be remapped through keymaps.  
   
 This is an alternative to using defadvice or substitute-key-definition  
 to modify the behaviour of a key binding using the normal keymap  
 binding and lookup functionality.  
   
 When a key sequence is bound to a command, and that command is  
 remapped to another command, that command is run instead of the  
 original command.  
   
 Example:  
 Suppose that minor mode my-mode has defined the commands  
 my-kill-line and my-kill-word, and it wants C-k (and any other key  
 bound to kill-line) to run the command my-kill-line instead of  
 kill-line, and likewise it wants to run my-kill-word instead of  
 kill-word.  
   
 Instead of rebinding C-k and the other keys in the minor mode map,  
 command remapping allows you to directly map kill-line into  
 my-kill-line and kill-word into my-kill-word through the minor mode  
 map using define-key:  
   
    (define-key my-mode-map [remap kill-line] 'my-kill-line)  
    (define-key my-mode-map [remap kill-word] 'my-kill-word)  
   
 Now, when my-mode is enabled, and the user enters C-k or M-d,  
 the commands my-kill-line and my-kill-word are run.  
   
 Notice that only one level of remapping is supported.  In the above  
 example, this means that if my-kill-line is remapped to other-kill,  
 then C-k still runs my-kill-line.  
   
 The following changes have been made to provide command remapping:  
   
 - Command remappings are defined using `define-key' with a prefix-key  
   `remap', i.e. `(define-key MAP [remap CMD] DEF)' remaps command CMD  
   to definition DEF in keymap MAP.  The definition is not limited to  
   another command; it can be anything accepted for a normal binding.  
   
 - The new function `remap-command' returns the binding for a remapped  
   command in the current keymaps, or nil if it isn't remapped.  
   
 - key-binding now remaps interactive commands unless the optional  
   third argument NO-REMAP is non-nil.  
   
 - where-is-internal now returns nil for a remapped command (e.g.  
   kill-line if my-mode is enabled), and the actual key binding for  
   the command it is remapped to (e.g. C-k for my-kill-line).  
   It also has a new optional fifth argument, NO-REMAP, which inhibits  
   remapping if non-nil (e.g. it returns C-k for kill-line and  
   <kill-line> for my-kill-line).  
   
 - The new variable `this-original-command' contains the original  
   command before remapping.  It is equal to `this-command' when the  
   command was not remapped.  
   
 ** Atomic change groups.  
   
 To perform some changes in the current buffer "atomically" so that  
 they either all succeed or are all undone, use `atomic-change-group'  
 around the code that makes changes.  For instance:  
   
   (atomic-change-group  
     (insert foo)  
     (delete-region x y))  
   
 If an error (or other nonlocal exit) occurs inside the body of  
 `atomic-change-group', it unmakes all the changes in that buffer that  
 were during the execution of the body.  The change group has no effect  
 on any other buffers--any such changes remain.  
   
 If you need something more sophisticated, you can directly call the  
 lower-level functions that `atomic-change-group' uses.  Here is how.  
   
 To set up a change group for one buffer, call `prepare-change-group'.  
 Specify the buffer as argument; it defaults to the current buffer.  
 This function returns a "handle" for the change group.  You must save  
 the handle to activate the change group and then finish it.  
   
 Before you change the buffer again, you must activate the change  
 group.  Pass the handle to `activate-change-group' afterward to  
 do this.  
   
 After you make the changes, you must finish the change group.  You can  
 either accept the changes or cancel them all.  Call  
 `accept-change-group' to accept the changes in the group as final;  
 call `cancel-change-group' to undo them all.  
   
 You should use `unwind-protect' to make sure the group is always  
 finished.  The call to `activate-change-group' should be inside the  
 `unwind-protect', in case the user types C-g just after it runs.  
 (This is one reason why `prepare-change-group' and  
 `activate-change-group' are separate functions.)  Once you finish the  
 group, don't use the handle again--don't try to finish the same group  
 twice.  
   
 To make a multibuffer change group, call `prepare-change-group' once  
 for each buffer you want to cover, then use `nconc' to combine the  
 returned values, like this:  
   
   (nconc (prepare-change-group buffer-1)  
          (prepare-change-group buffer-2))  
   
 You can then activate the multibuffer change group with a single call  
 to `activate-change-group', and finish it with a single call to  
 `accept-change-group' or `cancel-change-group'.  
   
 Nested use of several change groups for the same buffer works as you  
 would expect.  Non-nested use of change groups for the same buffer  
 will lead to undesirable results, so don't let it happen; the first  
 change group you start for any given buffer should be the last one  
 finished.  
   
 ** New function substring-no-properties.  
   
 +++  
 *** The meanings of scroll-up-aggressively and scroll-down-aggressively  
 have been interchanged, so that the former now controls scrolling up,  
 and the latter now controls scrolling down.  
   
 +++  
 ** New function window-body-height.  
   
 This is like window-height but does not count the mode line  
 or the header line.  
   
 +++  
 ** The function `atan' now accepts an optional second argument.  
   
 When called with 2 arguments, as in `(atan Y X)', `atan' returns the  
 angle in radians between the vector [X, Y] and the X axis.  (This is  
 equivalent to the standard C library function `atan2'.)  
   
 +++  
 ** You can now make a window as short as one line.  
   
 A window that is just one line tall does not display either a mode  
 line or a header line, even if the variables `mode-line-format' and  
 `header-line-format' call for them.  A window that is two lines tall  
 cannot display both a mode line and a header line at once; if the  
 variables call for both, only the mode line actually appears.  
   
 +++  
 ** The new frame parameter `tty-color-mode' specifies the mode to use  
 for color support on character terminal frames.  Its value can be a  
 number of colors to support, or a symbol.  See the Emacs Lisp  
 Reference manual for more detailed documentation.  
   
 ** Already true in Emacs 21.1, but not emphasized clearly enough:  
   
 Multibyte buffers can now faithfully record all 256 character codes  
 from 0 to 255.  As a result, most of the past reasons to use unibyte  
 buffers no longer exist.  We only know of three reasons to use them  
 now:  
   
 1. If you prefer to use unibyte text all of the time.  
   
 2. For reading files into temporary buffers, when you want to avoid  
 the time it takes to convert the format.  
   
 3. For binary files where format conversion would be pointless and  
 wasteful.  
   
 ** If text has a `keymap' property, that keymap takes precedence  
 over minor mode keymaps.  
   
 ** A hex escape in a string forces the string to be multibyte.  
 An octal escape makes it unibyte.  
   
 ** The position after an invisible, intangible character  
 is considered an unacceptable value for point;  
 intangibility processing effectively treats the following character  
 as part of the intangible region even if it is not itself intangible.  
   
 Thus, point can go before an invisible, intangible region, but not  
 after it.  This prevents C-f and C-b from appearing to stand still on  
 the screen.  
   
 +++  
 ** define-abbrev now accepts an optional argument SYSTEM-FLAG.  If  
 non-nil, this marks the abbrev as a "system" abbrev, which means that  
 it won't be stored in the user's abbrevs file if he saves the abbrevs.  
 Major modes that predefine some abbrevs should always specify this  
 flag.  
42    
43  ** Support for Mocklisp has been removed.  ** New charset implementation with many new charsets.
44    See M-x list-character-sets.  New charsets can be defined conveniently
45    as tables of unicodes.
46    
47  ** The function insert-string is now obsolete.  The dimension of a charset is now 0, 1, 2, or 3, and the size of each
48    dimension is no longer limited to 94 or 96.
49    
50  ** The precedence of file-name-handlers has been changed.  Generic characters no longer exist.  
 Instead of blindly choosing the first handler that matches,  
 find-file-name-handler now gives precedence to a file-name handler  
 that matches near the end of the file name.  More specifically, the  
 handler whose (match-beginning 0) is the largest is chosen.  
 In case of ties, the old "first matched" rule applies.  
51    
52  ** Dense keymaps now handle inheritance correctly.  A dynamic charset priority list is used to infer the charset of
53  Previously a dense keymap would hide all of the simple-char key  unicodes for display &c.
 bindings of the parent keymap.  
54    
55  ** jit-lock obeys a new text-property `jit-lock-defer-multiline'.  ** The following facilities are obsolete:
 If a piece of text with that property gets contextually refontified  
 (see jit-lock-defer-contextually), then all of that text will  
 be refontified.  This is useful when the syntax of a textual element  
 depends on text several lines further down (and when font-lock-multiline  
 is not appropriate to solve that problem).  For example in Perl:  
56    
57          s{  Minor modes: unify-8859-on-encoding-mode, unify-8859-on-decoding-mode
                 foo  
         }{  
                 bar  
         }e  
   
 Adding/removing the last `e' changes the `bar' from being a piece of  
 text to being a piece of code, so you'd put a jit-lock-defer-multiline  
 property over the second half of the command to force (deferred)  
 refontification of `bar' whenever the `e' is added/removed.  
   
 ** describe-vector now takes a second argument `describer' which is  
 called to print the entries' values.  It defaults to `princ'.  
   
 ** defcustom and other custom declarations now use a default group  
 (the last group defined in the same file) when no :group was given.  
   
 ** emacsserver now runs pre-command-hook and post-command-hook when  
 it receives a request from emacsclient.  
   
 ** The variable `recursive-load-depth-limit' has been deleted.  
 Emacs now signals an error if the same file is loaded with more  
 than 3 levels of nesting.  
   
 ** The default values of paragraph-start and indent-line-function have  
 been changed to reflect those used in Text mode rather than those used  
 in Indented-Text mode.  
   
 ** If a major mode function has a non-nil `no-clone-indirect'  
 property, `clone-indirect-buffer' signals an error if you use  
 it in that buffer.  
   
 ** If you set `query-replace-skip-read-only' non-nil,  
 `query-replace' and related functions simply ignore  
 a match if part of it has a read-only property.  
   
 ** In `replace-match', the replacement text no longer inherits  
 properties from surrounding text.  
   
 ** New function `buffer-local-value'.  
   
 - Function: buffer-local-value variable buffer  
   
 This function returns the buffer-local binding of VARIABLE (a symbol)  
 in buffer BUFFER.  If VARIABLE does not have a buffer-local binding in  
 buffer BUFFER, it returns the default value of VARIABLE instead.  
   
 ** The default value of `paragraph-start' and `indent-line-function' has  
 been changed to reflect the one used in Text mode rather than the one  
 used in Indented Text mode.  
   
 ** New function `text-clone-create'.  Text clones are chunks of text  
 that are kept identical by transparently propagating changes from one  
 clone to the other.  
   
 ** font-lock can manage arbitrary text-properties beside `face'.  
 *** the FACENAME returned in font-lock-keywords can be a list  
 of the form (face FACE PROP1 VAL1 PROP@ VAL2 ...) so you can set  
 other properties than `face'.  
 *** font-lock-extra-managed-props can be set to make sure those extra  
 properties are automatically cleaned up by font-lock.  
   
 ** The new function `run-mode-hooks' and the new macro `delay-mode-hooks'  
 are used by define-derived-mode to make sure the mode hook for the  
 parent mode is run at the end of the child mode.  
   
 ** `provide' and `featurep' now accept an optional second argument  
 to test/provide subfeatures.  Also `provide' now checks `after-load-alist'  
 and run any code associated with the provided feature.  
   
 ** The variable `compilation-parse-errors-filename-function' can  
 be used to transform filenames found in compilation output.  
   
 +++  
 ** Functions `file-name-sans-extension' and `file-name-extension' now  
 ignore the leading dots in file names, so that file names such as  
 `.emacs' are treated as extensionless.  
   
 ** Functions `user-uid' and `user-real-uid' now return floats if the  
 user UID doesn't fit in a Lisp integer.  Function `user-full-name'  
 accepts a float as UID parameter.  
   
 ** `define-key-after' now accepts keys longer than 1.  
   
 ** `define-derived-mode' now accepts nil as the parent.  
   
 ** The local variable `no-byte-compile' in elisp files is now obeyed.  
   
 ** New functions `keymap-prompt' and `current-active-maps'.  
   
 ** New function `describe-buffer-bindings'.  
   
 ** New vars `exec-suffixes' and `load-suffixes' used when  
 searching for an executable resp. an elisp file.  
   
 ** Variable aliases have been implemented  
   
 - Macro: defvaralias ALIAS-VAR BASE-VAR  
   
 This defines the symbol ALIAS-VAR as a variable alias for symbol  
 BASE-VAR.  This means that retrieving the value of ALIAS-VAR returns  
 the value of BASE-VAR, and changing the value of ALIAS-VAR changes the  
 value of BASE-VAR.  
   
 - Function: indirect-variable VARIABLE  
   
 This function returns the variable at the end of the chain of aliases  
 of VARIABLE.  If VARIABLE is not a symbol, or if VARIABLE is not  
 defined as an alias, the function returns VARIABLE.  
   
 It might be noteworthy that variables aliases work for all kinds of  
 variables, including buffer-local and frame-local variables.  
   
 ** Functions from `post-gc-hook' are run at the end of garbage  
 collection.  The hook is run with GC inhibited, so use it with care.  
   
 ** If the second argument to `copy-file' is the name of a directory,  
 the file is copied to that directory instead of signaling an error.  
   
 ** The variables most-positive-fixnum and most-negative-fixnum  
 have been moved from the CL package to the core.  
   
 ** On MS Windows, locale-coding-system is used to interact with the OS.  
 The Windows specific variable w32-system-coding-system, which was  
 formerly used for that purpose is now an alias for locale-coding-system.  
   
 ** New packages:  
   
 *** The new package syntax.el provides an efficient way to find the  
 current syntactic context (as returned by parse-partial-sexp).  
   
 *** The TCL package tcl-mode.el was replaced by tcl.el.  
 This was actually done in Emacs-21.1, and was not documented.  
   
 *** The new package Ibuffer provides a powerful, completely  
 customizable replacement for buff-menu.el.  
58    
59    
60  * Installation Changes in Emacs 21.1  * Lisp changes in Emacs 22.1
   
 See the INSTALL file for information on installing extra libraries and  
 fonts to take advantage of the new graphical features and extra  
 charsets in this release.  
   
 ** Support for GNU/Linux on IA64 machines has been added.  
   
 ** Support for LynxOS has been added.  
   
 ** There are new configure options associated with the support for  
 images and toolkit scrollbars.  Use the --help option in `configure'  
 to list them.  
   
 ** You can build a 64-bit Emacs for SPARC/Solaris systems which  
 support 64-bit executables and also on Irix 6.5.  This increases the  
 maximum buffer size.  See etc/MACHINES for instructions.  Changes to  
 build on other 64-bit systems should be straightforward modulo any  
 necessary changes to unexec.  
   
 ** There is a new configure option `--disable-largefile' to omit  
 Unix-98-style support for large files if that is available.  
   
 ** There is a new configure option `--without-xim' that instructs  
 Emacs to not use X Input Methods (XIM), if these are available.  
   
 ** `movemail' defaults to supporting POP.  You can turn this off using  
 the --without-pop configure option, should that be necessary.  
   
 ** This version can be built for the Macintosh, but does not implement  
 all of the new display features described below.  The port currently  
 lacks unexec, asynchronous processes, and networking support.  See the  
 "Emacs and the Mac OS" appendix in the Emacs manual, for the  
 description of aspects specific to the Mac.  
   
 ** Note that the MS-Windows port does not yet implement various of the  
 new display features described below.  
   
   
 * Changes in Emacs 21.1  
   
 ** Emacs has a new redisplay engine.  
   
 The new redisplay handles characters of variable width and height.  
 Italic text can be used without redisplay problems.  Fonts containing  
 oversized characters, i.e. characters larger than the logical height  
 of a font can be used.  Images of various formats can be displayed in  
 the text.  
   
 ** Emacs has a new face implementation.  
   
 The new faces no longer fundamentally use X font names to specify the  
 font.  Instead, each face has several independent attributes--family,  
 height, width, weight and slant--that it may or may not specify.  
 These attributes can be merged from various faces, and then together  
 specify a font.  
   
 Faces are supported on terminals that can display color or fonts.  
 These terminal capabilities are auto-detected.  Details can be found  
 under Lisp changes, below.  
   
 ** Emacs can display faces on TTY frames.  
   
 Emacs automatically detects terminals that are able to display colors.  
 Faces with a weight greater than normal are displayed extra-bright, if  
 the terminal supports it.  Faces with a weight less than normal and  
 italic faces are displayed dimmed, if the terminal supports it.  
 Underlined faces are displayed underlined if possible.  Other face  
 attributes such as `overline', `strike-through', and `box' are ignored  
 on terminals.  
   
 The command-line options `-fg COLOR', `-bg COLOR', and `-rv' are now  
 supported on character terminals.  
   
 Emacs automatically remaps all X-style color specifications to one of  
 the colors supported by the terminal.  This means you could have the  
 same color customizations that work both on a windowed display and on  
 a TTY or when Emacs is invoked with the -nw option.  
   
 ** New default font is Courier 12pt under X.  
   
 ** Sound support  
   
 Emacs supports playing sound files on GNU/Linux and FreeBSD (Voxware  
 driver and native BSD driver, a.k.a. Luigi's driver).  Currently  
 supported file formats are RIFF-WAVE (*.wav) and Sun Audio (*.au).  
 You must configure Emacs with the option `--with-sound=yes' to enable  
 sound support.  
   
 ** Emacs now resizes mini-windows if appropriate.  
   
 If a message is longer than one line, or minibuffer contents are  
 longer than one line, Emacs can resize the minibuffer window unless it  
 is on a frame of its own.  You can control resizing and the maximum  
 minibuffer window size by setting the following variables:  
   
 - User option: max-mini-window-height  
   
 Maximum height for resizing mini-windows.  If a float, it specifies a  
 fraction of the mini-window frame's height.  If an integer, it  
 specifies a number of lines.  
   
 Default is 0.25.  
   
 - User option: resize-mini-windows  
   
 How to resize mini-windows.  If nil, don't resize.  If t, always  
 resize to fit the size of the text.  If `grow-only', let mini-windows  
 grow only, until they become empty, at which point they are shrunk  
 again.  
   
 Default is `grow-only'.  
   
 ** LessTif support.  
   
 Emacs now runs with the LessTif toolkit (see  
 <http://www.lesstif.org>).  You will need version 0.92.26, or later.  
   
 ** LessTif/Motif file selection dialog.  
   
 When Emacs is configured to use LessTif or Motif, reading a file name  
 from a menu will pop up a file selection dialog if `use-dialog-box' is  
 non-nil.  
   
 ** File selection dialog on MS-Windows is supported.  
   
 When a file is visited by clicking File->Open, the MS-Windows version  
 now pops up a standard file selection dialog where you can select a  
 file to visit.  File->Save As also pops up that dialog.  
   
 ** Toolkit scroll bars.  
   
 Emacs now uses toolkit scroll bars if available.  When configured for  
 LessTif/Motif, it will use that toolkit's scroll bar.  Otherwise, when  
 configured for Lucid and Athena widgets, it will use the Xaw3d scroll  
 bar if Xaw3d is available.  You can turn off the use of toolkit scroll  
 bars by specifying `--with-toolkit-scroll-bars=no' when configuring  
 Emacs.  
   
 When you encounter problems with the Xaw3d scroll bar, watch out how  
 Xaw3d is compiled on your system.  If the Makefile generated from  
 Xaw3d's Imakefile contains a `-DNARROWPROTO' compiler option, and your  
 Emacs system configuration file `s/your-system.h' does not contain a  
 define for NARROWPROTO, you might consider adding it.  Take  
 `s/freebsd.h' as an example.  
   
 Alternatively, if you don't have access to the Xaw3d source code, take  
 a look at your system's imake configuration file, for example in the  
 directory `/usr/X11R6/lib/X11/config' (paths are different on  
 different systems).  You will find files `*.cf' there.  If your  
 system's cf-file contains a line like `#define NeedWidePrototypes NO',  
 add a `#define NARROWPROTO' to your Emacs system configuration file.  
   
 The reason for this is that one Xaw3d function uses `double' or  
 `float' function parameters depending on the setting of NARROWPROTO.  
 This is not a problem when Imakefiles are used because each system's  
 imake configuration file contains the necessary information.  Since  
 Emacs doesn't use imake, this has do be done manually.  
   
 ** Tool bar support.  
   
 Emacs supports a tool bar at the top of a frame under X.  For details  
 of how to define a tool bar, see the page describing Lisp-level  
 changes.  Tool-bar global minor mode controls whether or not it is  
 displayed and is on by default.  The appearance of the bar is improved  
 if Emacs has been built with XPM image support.  Otherwise monochrome  
 icons will be used.  
   
 To make the tool bar more useful, we need contributions of extra icons  
 for specific modes (with copyright assignments).  
   
 ** Tooltips.  
   
 Tooltips are small X windows displaying a help string at the current  
 mouse position.  The Lisp package `tooltip' implements them.  You can  
 turn them off via the user option `tooltip-mode'.  
   
 Tooltips also provides support for GUD debugging.  If activated,  
 variable values can be displayed in tooltips by pointing at them with  
 the mouse in source buffers.  You can customize various aspects of the  
 tooltip display in the group `tooltip'.  
   
 ** Automatic Hscrolling  
   
 Horizontal scrolling now happens automatically if  
 `automatic-hscrolling' is set (the default).  This setting can be  
 customized.  
   
 If a window is scrolled horizontally with set-window-hscroll, or  
 scroll-left/scroll-right (C-x <, C-x >), this serves as a lower bound  
 for automatic horizontal scrolling.  Automatic scrolling will scroll  
 the text more to the left if necessary, but won't scroll the text more  
 to the right than the column set with set-window-hscroll etc.  
   
 ** When using a windowing terminal, each Emacs window now has a cursor  
 of its own.  By default, when a window is selected, the cursor is  
 solid; otherwise, it is hollow.  The user-option  
 `cursor-in-non-selected-windows' controls how to display the  
 cursor in non-selected windows.  If nil, no cursor is shown, if  
 non-nil a hollow box cursor is shown.  
   
 ** Fringes to the left and right of windows are used to display  
 truncation marks, continuation marks, overlay arrows and alike.  The  
 foreground, background, and stipple of these areas can be changed by  
 customizing face `fringe'.  
   
 ** The mode line under X is now drawn with shadows by default.  
 You can change its appearance by modifying the face `mode-line'.  
 In particular, setting the `:box' attribute to nil turns off the 3D  
 appearance of the mode line.  (The 3D appearance makes the mode line  
 occupy more space, and thus might cause the first or the last line of  
 the window to be partially obscured.)  
   
 The variable `mode-line-inverse-video', which was used in older  
 versions of emacs to make the mode-line stand out, is now deprecated.  
 However, setting it to nil will cause the `mode-line' face to be  
 ignored, and mode-lines to be drawn using the default text face.  
   
 ** Mouse-sensitive mode line.  
   
 Different parts of the mode line have been made mouse-sensitive on all  
 systems which support the mouse.  Moving the mouse to a  
 mouse-sensitive part in the mode line changes the appearance of the  
 mouse pointer to an arrow, and help about available mouse actions is  
 displayed either in the echo area, or in the tooltip window if you  
 have enabled one.  
   
 Currently, the following actions have been defined:  
   
 - Mouse-1 on the buffer name in the mode line goes to the next buffer.  
   
 - Mouse-3 on the buffer-name goes to the previous buffer.  
   
 - Mouse-2 on the read-only or modified status in the mode line (`%' or  
 `*') toggles the status.  
   
 - Mouse-3 on the mode name displays a minor-mode menu.  
   
 ** Hourglass pointer  
   
 Emacs can optionally display an hourglass pointer under X.  You can  
 turn the display on or off by customizing group `cursor'.  
   
 ** Blinking cursor  
   
 M-x blink-cursor-mode toggles a blinking cursor under X and on  
 terminals having terminal capabilities `vi', `vs', and `ve'.  Blinking  
 and related parameters like frequency and delay can be customized in  
 the group `cursor'.  
   
 ** New font-lock support mode `jit-lock-mode'.  
   
 This support mode is roughly equivalent to `lazy-lock' but is  
 generally faster.  It supports stealth and deferred fontification.  
 See the documentation of the function `jit-lock-mode' for more  
 details.  
   
 Font-lock uses jit-lock-mode as default support mode, so you don't  
 have to do anything to activate it.  
   
 ** The default binding of the Delete key has changed.  
   
 The new user-option `normal-erase-is-backspace' can be set to  
 determine the effect of the Delete and Backspace function keys.  
   
 On window systems, the default value of this option is chosen  
 according to the keyboard used.  If the keyboard has both a Backspace  
 key and a Delete key, and both are mapped to their usual meanings, the  
 option's default value is set to t, so that Backspace can be used to  
 delete backward, and Delete can be used to delete forward.  On  
 keyboards which either have only one key (usually labeled DEL), or two  
 keys DEL and BS which produce the same effect, the option's value is  
 set to nil, and these keys delete backward.  
   
 If not running under a window system, setting this option accomplishes  
 a similar effect by mapping C-h, which is usually generated by the  
 Backspace key, to DEL, and by mapping DEL to C-d via  
 `keyboard-translate'.  The former functionality of C-h is available on  
 the F1 key.  You should probably not use this setting on a text-only  
 terminal if you don't have both Backspace, Delete and F1 keys.  
   
 Programmatically, you can call function normal-erase-is-backspace-mode  
 to toggle the behavior of the Delete and Backspace keys.  
   
 ** The default for user-option `next-line-add-newlines' has been  
 changed to nil, i.e. C-n will no longer add newlines at the end of a  
 buffer by default.  
   
 ** The <home> and <end> keys now move to the beginning or end of the  
 current line, respectively.  C-<home> and C-<end> move to the  
 beginning and end of the buffer.  
   
 ** Emacs now checks for recursive loads of Lisp files.  If the  
 recursion depth exceeds `recursive-load-depth-limit', an error is  
 signaled.  
   
 ** When an error is signaled during the loading of the user's init  
 file, Emacs now pops up the *Messages* buffer.  
   
 ** Emacs now refuses to load compiled Lisp files which weren't  
 compiled with Emacs.  Set `load-dangerous-libraries' to t to change  
 this behavior.  
   
 The reason for this change is an incompatible change in XEmacs's byte  
 compiler.  Files compiled with XEmacs can contain byte codes that let  
 Emacs dump core.  
   
 ** Toggle buttons and radio buttons in menus.  
   
 When compiled with LessTif (or Motif) support, Emacs uses toolkit  
 widgets for radio and toggle buttons in menus.  When configured for  
 Lucid, Emacs draws radio buttons and toggle buttons similar to Motif.  
   
 ** The menu bar configuration has changed.  The new configuration is  
 more CUA-compliant.  The most significant change is that Options is  
 now a separate menu-bar item, with Mule and Customize as its submenus.  
   
 ** Item Save Options on the Options menu allows saving options set  
 using that menu.  
   
 ** Highlighting of trailing whitespace.  
   
 When `show-trailing-whitespace' is non-nil, Emacs displays trailing  
 whitespace in the face `trailing-whitespace'.  Trailing whitespace is  
 defined as spaces or tabs at the end of a line.  To avoid busy  
 highlighting when entering new text, trailing whitespace is not  
 displayed if point is at the end of the line containing the  
 whitespace.  
   
 ** C-x 5 1 runs the new command delete-other-frames which deletes  
 all frames except the selected one.  
   
 ** The new user-option `confirm-kill-emacs' can be customized to  
 let Emacs ask for confirmation before exiting.  
   
 ** The header line in an Info buffer is now displayed as an emacs  
 header-line (which is like a mode-line, but at the top of the window),  
 so that it remains visible even when the buffer has been scrolled.  
 This behavior may be disabled by customizing the option  
 `Info-use-header-line'.  
   
 ** Polish, Czech, German, and French translations of Emacs' reference card  
 have been added.  They are named `pl-refcard.tex', `cs-refcard.tex',  
 `de-refcard.tex' and `fr-refcard.tex'.  Postscript files are included.  
   
 ** An `Emacs Survival Guide', etc/survival.tex, is available.  
   
 ** A reference card for Dired has been added.  Its name is  
 `dired-ref.tex'.  A French translation is available in  
 `fr-drdref.tex'.  
   
 ** C-down-mouse-3 is bound differently.  Now if the menu bar is not  
 displayed it pops up a menu containing the items which would be on the  
 menu bar.  If the menu bar is displayed, it pops up the major mode  
 menu or the Edit menu if there is no major mode menu.  
   
 ** Variable `load-path' is no longer customizable through Customize.  
   
 You can no longer use `M-x customize-variable' to customize `load-path'  
 because it now contains a version-dependent component.  You can still  
 use `add-to-list' and `setq' to customize this variable in your  
 `~/.emacs' init file or to modify it from any Lisp program in general.  
   
 ** C-u C-x = provides detailed information about the character at  
 point in a pop-up window.  
   
 ** Emacs can now support 'wheeled' mice (such as the MS IntelliMouse)  
 under XFree86.  To enable this, use the `mouse-wheel-mode' command, or  
 customize the variable `mouse-wheel-mode'.  
   
 The variables `mouse-wheel-follow-mouse' and `mouse-wheel-scroll-amount'  
 determine where and by how much buffers are scrolled.  
   
 ** Emacs' auto-save list files are now by default stored in a  
 sub-directory `.emacs.d/auto-save-list/' of the user's home directory.  
 (On MS-DOS, this subdirectory's name is `_emacs.d/auto-save.list/'.)  
 You can customize `auto-save-list-file-prefix' to change this location.  
   
 ** The function `getenv' is now callable interactively.  
   
 ** The new user-option `even-window-heights' can be set to nil  
 to prevent `display-buffer' from evening out window heights.  
   
 ** The new command M-x delete-trailing-whitespace RET will delete the  
 trailing whitespace within the current restriction.  You can also add  
 this function to `write-file-hooks' or `local-write-file-hooks'.  
   
 ** When visiting a file with M-x find-file-literally, no newlines will  
 be added to the end of the buffer even if `require-final-newline' is  
 non-nil.  
   
 ** The new user-option `find-file-suppress-same-file-warnings' can be  
 set to suppress warnings ``X and Y are the same file'' when visiting a  
 file that is already visited under a different name.  
   
 ** The new user-option `electric-help-shrink-window' can be set to  
 nil to prevent adjusting the help window size to the buffer size.  
   
 ** New command M-x describe-character-set reads a character set name  
 and displays information about that.  
   
 ** The new variable `auto-mode-interpreter-regexp' contains a regular  
 expression matching interpreters, for file mode determination.  
   
 This regular expression is matched against the first line of a file to  
 determine the file's mode in `set-auto-mode' when Emacs can't deduce a  
 mode from the file's name.  If it matches, the file is assumed to be  
 interpreted by the interpreter matched by the second group of the  
 regular expression.  The mode is then determined as the mode  
 associated with that interpreter in `interpreter-mode-alist'.  
   
 ** New function executable-make-buffer-file-executable-if-script-p is  
 suitable as an after-save-hook as an alternative to `executable-chmod'.  
   
 ** The most preferred coding-system is now used to save a buffer if  
 buffer-file-coding-system is `undecided' and it is safe for the buffer  
 contents.  (The most preferred is set by set-language-environment or  
 by M-x prefer-coding-system.)  Thus if you visit an ASCII file and  
 insert a non-ASCII character from your current language environment,  
 the file will be saved silently with the appropriate coding.  
 Previously you would be prompted for a safe coding system.  
   
 ** The many obsolete language `setup-...-environment' commands have  
 been removed -- use `set-language-environment'.  
   
 ** The new Custom option `keyboard-coding-system' specifies a coding  
 system for keyboard input.  
   
 ** New variable `inhibit-iso-escape-detection' determines if Emacs'  
 coding system detection algorithm should pay attention to ISO2022's  
 escape sequences.  If this variable is non-nil, the algorithm ignores  
 such escape sequences.  The default value is nil, and it is  
 recommended not to change it except for the special case that you  
 always want to read any escape code verbatim.  If you just want to  
 read a specific file without decoding escape codes, use C-x RET c  
 (`universal-coding-system-argument').  For instance, C-x RET c latin-1  
 RET C-x C-f filename RET.  
   
 ** Variable `default-korean-keyboard' is initialized properly from the  
 environment variable `HANGUL_KEYBOARD_TYPE'.  
   
 ** New command M-x list-charset-chars reads a character set name and  
 displays all characters in that character set.  
   
 ** M-x set-terminal-coding-system (C-x RET t) now allows CCL-based  
 coding systems such as cpXXX and cyrillic-koi8.  
   
 ** Emacs now attempts to determine the initial language environment  
 and preferred and locale coding systems systematically from the  
 LC_ALL, LC_CTYPE, and LANG environment variables during startup.  
   
 ** New language environments `Polish', `Latin-8' and `Latin-9'.  
 Latin-8 and Latin-9 correspond respectively to the ISO character sets  
 8859-14 (Celtic) and 8859-15 (updated Latin-1, with the Euro sign).  
 GNU Intlfonts doesn't support these yet but recent X releases have  
 8859-15.  See etc/INSTALL for information on obtaining extra fonts.  
 There are new Leim input methods for Latin-8 and Latin-9 prefix (only)  
 and Polish `slash'.  
   
 ** New language environments `Dutch' and `Spanish'.  
 These new environments mainly select appropriate translations  
 of the tutorial.  
   
 ** In Ethiopic language environment, special key bindings for  
 function keys are changed as follows.  This is to conform to "Emacs  
 Lisp Coding Convention".  
   
     new  command                            old-binding  
     ---  -------                            -----------  
     f3   ethio-fidel-to-sera-buffer         f5  
     S-f3 ethio-fidel-to-sera-region         f5  
     C-f3 ethio-fidel-to-sera-mail-or-marker f5  
   
     f4   ethio-sera-to-fidel-buffer         unchanged  
     S-f4 ethio-sera-to-fidel-region         unchanged  
     C-f4 ethio-sera-to-fidel-mail-or-marker unchanged  
   
     S-f5 ethio-toggle-punctuation           f3  
     S-f6 ethio-modify-vowel                 f6  
     S-f7 ethio-replace-space                f7  
     S-f8 ethio-input-special-character      f8  
     S-f9 ethio-replace-space                unchanged  
     C-f9 ethio-toggle-space                 f2  
   
 ** There are new Leim input methods.  
 New input methods "turkish-postfix", "turkish-alt-postfix",  
 "greek-mizuochi", "TeX", and "greek-babel" are now part of the Leim  
 package.  
   
 ** The rule of input method "slovak" is slightly changed.  Now the  
 rules for translating "q" and "Q" to "`" (backquote) are deleted, thus  
 typing them inserts "q" and "Q" respectively.  Rules for translating  
 "=q", "+q", "=Q", and "+Q" to "`" are also deleted.  Now, to input  
 "`", you must type "=q".  
   
 ** When your terminal can't display characters from some of the ISO  
 8859 character sets but can display Latin-1, you can display  
 more-or-less mnemonic sequences of ASCII/Latin-1 characters instead of  
 empty boxes (under a window system) or question marks (not under a  
 window system).  Customize the option `latin1-display' to turn this  
 on.  
   
 ** M-; now calls comment-dwim which tries to do something clever based  
 on the context.  M-x kill-comment is now an alias to comment-kill,  
 defined in newcomment.el.  You can choose different styles of region  
 commenting with the variable `comment-style'.  
   
 ** New user options `display-time-mail-face' and  
 `display-time-use-mail-icon' control the appearance of mode-line mail  
 indicator used by the display-time package.  On a suitable display the  
 indicator can be an icon and is mouse-sensitive.  
   
 ** On window-systems, additional space can be put between text lines  
 on the display using several methods  
   
 - By setting frame parameter `line-spacing' to PIXELS.  PIXELS must be  
 a positive integer, and specifies that PIXELS number of pixels should  
 be put below text lines on the affected frame or frames.  
   
 - By setting X resource `lineSpacing', class `LineSpacing'.  This is  
 equivalent to specifying the frame parameter.  
   
 - By specifying `--line-spacing=N' or `-lsp N' on the command line.  
   
 - By setting buffer-local variable `line-spacing'.  The meaning is  
 the same, but applies to the a particular buffer only.  
   
 ** The new command `clone-indirect-buffer' can be used to create  
 an indirect buffer that is a twin copy of the current buffer.  The  
 command `clone-indirect-buffer-other-window', bound to C-x 4 c,  
 does the same but displays the indirect buffer in another window.  
   
 ** New user options `backup-directory-alist' and  
 `make-backup-file-name-function' control the placement of backups,  
 typically in a single directory or in an invisible sub-directory.  
   
 ** New commands iso-iso2sgml and iso-sgml2iso convert between Latin-1  
 characters and the corresponding SGML (HTML) entities.  
   
 ** New X resources recognized  
   
 *** The X resource `synchronous', class `Synchronous', specifies  
 whether Emacs should run in synchronous mode.  Synchronous mode  
 is useful for debugging X problems.  
   
 Example:  
   
   emacs.synchronous: true  
   
 *** The X resource `visualClass, class `VisualClass', specifies the  
 visual Emacs should use.  The resource's value should be a string of  
 the form `CLASS-DEPTH', where CLASS is the name of the visual class,  
 and DEPTH is the requested color depth as a decimal number.  Valid  
 visual class names are  
   
   TrueColor  
   PseudoColor  
   DirectColor  
   StaticColor  
   GrayScale  
   StaticGray  
   
 Visual class names specified as X resource are case-insensitive, i.e.  
 `pseudocolor', `Pseudocolor' and `PseudoColor' all have the same  
 meaning.  
   
 The program `xdpyinfo' can be used to list the visual classes  
 supported on your display, and which depths they have.  If  
 `visualClass' is not specified, Emacs uses the display's default  
 visual.  
   
 Example:  
   
   emacs.visualClass: TrueColor-8  
   
 *** The X resource `privateColormap', class `PrivateColormap',  
 specifies that Emacs should use a private colormap if it is using the  
 default visual, and that visual is of class PseudoColor.  Recognized  
 resource values are `true' or `on'.  
   
 Example:  
   
   emacs.privateColormap: true  
   
 ** Faces and frame parameters.  
   
 There are four new faces `scroll-bar', `border', `cursor' and `mouse'.  
 Setting the frame parameters `scroll-bar-foreground' and  
 `scroll-bar-background' sets foreground and background color of face  
 `scroll-bar' and vice versa.  Setting frame parameter `border-color'  
 sets the background color of face `border' and vice versa.  Likewise  
 for frame parameters `cursor-color' and face `cursor', and frame  
 parameter `mouse-color' and face `mouse'.  
   
 Changing frame parameter `font' sets font-related attributes of the  
 `default' face and vice versa.  Setting frame parameters  
 `foreground-color' or `background-color' sets the colors of the  
 `default' face and vice versa.  
   
 ** New face `menu'.  
   
 The face `menu' can be used to change colors and font of Emacs' menus.  
   
 ** New frame parameter `screen-gamma' for gamma correction.  
   
 The new frame parameter `screen-gamma' specifies gamma-correction for  
 colors.  Its value may be nil, the default, in which case no gamma  
 correction occurs, or a number > 0, usually a float, that specifies  
 the screen gamma of a frame's display.  
   
 PC monitors usually have a screen gamma of 2.2.  smaller values result  
 in darker colors.  You might want to try a screen gamma of 1.5 for LCD  
 color displays.  The viewing gamma Emacs uses is 0.4545. (1/2.2).  
   
 The X resource name of this parameter is `screenGamma', class  
 `ScreenGamma'.  
   
 ** Tabs and variable-width text.  
   
 Tabs are now displayed with stretch properties; the width of a tab is  
 defined as a multiple of the normal character width of a frame, and is  
 independent of the fonts used in the text where the tab appears.  
 Thus, tabs can be used to line up text in different fonts.  
   
 ** Enhancements of the Lucid menu bar  
   
 *** The Lucid menu bar now supports the resource "margin".  
   
         emacs.pane.menubar.margin: 5  
   
 The default margin is 4 which makes the menu bar appear like the  
 LessTif/Motif one.  
   
 *** Arrows that indicate sub-menus are now drawn with shadows, as in  
 LessTif and Motif.  
   
 ** A block cursor can be drawn as wide as the glyph under it under X.  
   
 As an example: if a block cursor is over a tab character, it will be  
 drawn as wide as that tab on the display.  To do this, set  
 `x-stretch-cursor' to a non-nil value.  
   
 ** Empty display lines at the end of a buffer may be marked with a  
 bitmap (this is similar to the tilde displayed by vi and Less).  
   
 This behavior is activated by setting the buffer-local variable  
 `indicate-empty-lines' to a non-nil value.  The default value of this  
 variable is found in `default-indicate-empty-lines'.  
   
 ** There is a new "aggressive" scrolling method.  
   
 When scrolling up because point is above the window start, if the  
 value of the buffer-local variable `scroll-up-aggressively' is a  
 number, Emacs chooses a new window start so that point ends up that  
 fraction of the window's height from the top of the window.  
   
 When scrolling down because point is below the window end, if the  
 value of the buffer-local variable `scroll-down-aggressively' is a  
 number, Emacs chooses a new window start so that point ends up that  
 fraction of the window's height from the bottom of the window.  
   
 ** You can now easily create new *Info* buffers using either  
 M-x clone-buffer, C-u m <entry> RET or C-u g <entry> RET.  
 M-x clone-buffer can also be used on *Help* and several other special  
 buffers.  
   
 ** The command `Info-search' now uses a search history.  
   
 ** Listing buffers with M-x list-buffers (C-x C-b) now shows  
 abbreviated file names.  Abbreviations can be customized by changing  
 `directory-abbrev-alist'.  
   
 ** A new variable, backup-by-copying-when-privileged-mismatch, gives  
 the highest file uid for which backup-by-copying-when-mismatch will be  
 forced on.  The assumption is that uids less than or equal to this  
 value are special uids (root, bin, daemon, etc.--not real system  
 users) and that files owned by these users should not change ownership,  
 even if your system policy allows users other than root to edit them.  
   
 The default is 200; set the variable to nil to disable the feature.  
   
 ** The rectangle commands now avoid inserting undesirable spaces,  
 notably at the end of lines.  
   
 All these functions have been rewritten to avoid inserting unwanted  
 spaces, and an optional prefix now allows them to behave the old way.  
   
 ** The function `replace-rectangle' is an alias for `string-rectangle'.  
   
 ** The new command M-x string-insert-rectangle is like `string-rectangle',  
 but inserts text instead of replacing it.  
   
 ** The new command M-x query-replace-regexp-eval acts like  
 query-replace-regexp, but takes a Lisp expression which is evaluated  
 after each match to get the replacement text.  
   
 ** M-x query-replace recognizes a new command `e' (or `E') that lets  
 you edit the replacement string.  
   
 ** The new command mail-abbrev-complete-alias, bound to `M-TAB'  
 (if you load the library `mailabbrev'), lets you complete mail aliases  
 in the text, analogous to lisp-complete-symbol.  
   
 ** The variable `echo-keystrokes' may now have a floating point value.  
   
 ** If your init file is compiled (.emacs.elc), `user-init-file' is set  
 to the source name (.emacs.el), if that exists, after loading it.  
   
 ** The help string specified for a menu-item whose definition contains  
 the property `:help HELP' is now displayed under X, on MS-Windows, and  
 MS-DOS, either in the echo area or with tooltips.  Many standard menus  
 displayed by Emacs now have help strings.  
   
 --  
 ** New user option `read-mail-command' specifies a command to use to  
 read mail from the menu etc.  
   
 ** The environment variable `EMACSLOCKDIR' is no longer used on MS-Windows.  
 This environment variable was used when creating lock files.  Emacs on  
 MS-Windows does not use this variable anymore.  This change was made  
 before Emacs 21.1, but wasn't documented until now.  
   
 ** Highlighting of mouse-sensitive regions is now supported in the  
 MS-DOS version of Emacs.  
   
 ** The new command `msdos-set-mouse-buttons' forces the MS-DOS version  
 of Emacs to behave as if the mouse had a specified number of buttons.  
 This comes handy with mice that don't report their number of buttons  
 correctly.  One example is the wheeled mice, which report 3 buttons,  
 but clicks on the middle button are not passed to the MS-DOS version  
 of Emacs.  
   
 ** Customize changes  
   
 *** Customize now supports comments about customized items.  Use the  
 `State' menu to add comments, or give a prefix argument to  
 M-x customize-set-variable or M-x customize-set-value.  Note that  
 customization comments will cause the customizations to fail in  
 earlier versions of Emacs.  
   
 *** The new option `custom-buffer-done-function' says whether to kill  
 Custom buffers when you've done with them or just bury them (the  
 default).  
   
 *** If Emacs was invoked with the `-q' or `--no-init-file' options, it  
 does not allow you to save customizations in your `~/.emacs' init  
 file.  This is because saving customizations from such a session would  
 wipe out all the other customizationss you might have on your init  
 file.  
   
 ** If Emacs was invoked with the `-q' or `--no-init-file' options, it  
 does not save disabled and enabled commands for future sessions, to  
 avoid overwriting existing customizations of this kind that are  
 already in your init file.  
   
 ** New features in evaluation commands  
   
 *** The commands to evaluate Lisp expressions, such as C-M-x in Lisp  
 modes, C-j in Lisp Interaction mode, and M-:, now bind the variables  
 print-level, print-length, and debug-on-error based on the new  
 customizable variables eval-expression-print-level,  
 eval-expression-print-length, and eval-expression-debug-on-error.  
   
 The default values for the first two of these variables are 12 and 4  
 respectively, which means that `eval-expression' now prints at most  
 the first 12 members of a list and at most 4 nesting levels deep (if  
 the list is longer or deeper than that, an ellipsis `...'  is  
 printed).  
   
 <RET> or <mouse-2> on the printed text toggles between an abbreviated  
 printed representation and an unabbreviated one.  
   
 The default value of eval-expression-debug-on-error is t, so any error  
 during evaluation produces a backtrace.  
   
 *** The function `eval-defun' (C-M-x) now loads Edebug and instruments  
 code when called with a prefix argument.  
   
 ** CC mode changes.  
   
 Note: This release contains changes that might not be compatible with  
 current user setups (although it's believed that these  
 incompatibilities will only show in very uncommon circumstances).  
 However, since the impact is uncertain, these changes may be rolled  
 back depending on user feedback.  Therefore there's no forward  
 compatibility guarantee wrt the new features introduced in this  
 release.  
   
 *** The hardcoded switch to "java" style in Java mode is gone.  
 CC Mode used to automatically set the style to "java" when Java mode  
 is entered.  This has now been removed since it caused too much  
 confusion.  
   
 However, to keep backward compatibility to a certain extent, the  
 default value for c-default-style now specifies the "java" style for  
 java-mode, but "gnu" for all other modes (as before).  So you won't  
 notice the change if you haven't touched that variable.  
   
 *** New cleanups, space-before-funcall and compact-empty-funcall.  
 Two new cleanups have been added to c-cleanup-list:  
   
 space-before-funcall causes a space to be inserted before the opening  
 parenthesis of a function call, which gives the style "foo (bar)".  
   
 compact-empty-funcall causes any space before a function call opening  
 parenthesis to be removed if there are no arguments to the function.  
 It's typically useful together with space-before-funcall to get the  
 style "foo (bar)" and "foo()".  
   
 *** Some keywords now automatically trigger reindentation.  
 Keywords like "else", "while", "catch" and "finally" have been made  
 "electric" to make them reindent automatically when they continue an  
 earlier statement.  An example:  
   
 for (i = 0; i < 17; i++)  
   if (a[i])  
     res += a[i]->offset;  
 else  
   
 Here, the "else" should be indented like the preceding "if", since it  
 continues that statement. CC Mode will automatically reindent it after  
 the "else" has been typed in full, since it's not until then it's  
 possible to decide whether it's a new statement or a continuation of  
 the preceding "if".  
   
 CC Mode uses Abbrev mode to achieve this, which is therefore turned on  
 by default.  
   
 *** M-a and M-e now moves by sentence in multiline strings.  
 Previously these two keys only moved by sentence in comments, which  
 meant that sentence movement didn't work in strings containing  
 documentation or other natural language text.  
   
 The reason it's only activated in multiline strings (i.e. strings that  
 contain a newline, even when escaped by a '\') is to avoid stopping in  
 the short strings that often reside inside statements.  Multiline  
 strings almost always contain text in a natural language, as opposed  
 to other strings that typically contain format specifications,  
 commands, etc.  Also, it's not that bothersome that M-a and M-e misses  
 sentences in single line strings, since they're short anyway.  
   
 *** Support for autodoc comments in Pike mode.  
 Autodoc comments for Pike are used to extract documentation from the  
 source, like Javadoc in Java.  Pike mode now recognize this markup in  
 comment prefixes and paragraph starts.  
   
 *** The comment prefix regexps on c-comment-prefix may be mode specific.  
 When c-comment-prefix is an association list, it specifies the comment  
 line prefix on a per-mode basis, like c-default-style does.  This  
 change came about to support the special autodoc comment prefix in  
 Pike mode only.  
   
 *** Better handling of syntactic errors.  
 The recovery after unbalanced parens earlier in the buffer has been  
 improved; CC Mode now reports them by dinging and giving a message  
 stating the offending line, but still recovers and indent the  
 following lines in a sane way (most of the time).  An "else" with no  
 matching "if" is handled similarly.  If an error is discovered while  
 indenting a region, the whole region is still indented and the error  
 is reported afterwards.  
   
 *** Lineup functions may now return absolute columns.  
 A lineup function can give an absolute column to indent the line to by  
 returning a vector with the desired column as the first element.  
   
 *** More robust and warning-free byte compilation.  
 Although this is strictly not a user visible change (well, depending  
 on the view of a user), it's still worth mentioning that CC Mode now  
 can be compiled in the standard ways without causing trouble.  Some  
 code have also been moved between the subpackages to enhance the  
 modularity somewhat.  Thanks to Martin Buchholz for doing the  
 groundwork.  
   
 *** c-style-variables-are-local-p now defaults to t.  
 This is an incompatible change that has been made to make the behavior  
 of the style system wrt global variable settings less confusing for  
 non-advanced users.  If you know what this variable does you might  
 want to set it to nil in your .emacs, otherwise you probably don't  
 have to bother.  
   
 Defaulting c-style-variables-are-local-p to t avoids the confusing  
 situation that occurs when a user sets some style variables globally  
 and edits both a Java and a non-Java file in the same Emacs session.  
 If the style variables aren't buffer local in this case, loading of  
 the second file will cause the default style (either "gnu" or "java"  
 by default) to override the global settings made by the user.  
   
 *** New initialization procedure for the style system.  
 When the initial style for a buffer is determined by CC Mode (from the  
 variable c-default-style), the global values of style variables now  
 take precedence over the values specified by the chosen style.  This  
 is different than the old behavior: previously, the style-specific  
 settings would override the global settings.  This change makes it  
 possible to do simple configuration in the intuitive way with  
 Customize or with setq lines in one's .emacs file.  
   
 By default, the global value of every style variable is the new  
 special symbol set-from-style, which causes the value to be taken from  
 the style system.  This means that in effect, only an explicit setting  
 of a style variable will cause the "overriding" behavior described  
 above.  
   
 Also note that global settings override style-specific settings *only*  
 when the initial style of a buffer is chosen by a CC Mode major mode  
 function.  When a style is chosen in other ways --- for example, by a  
 call like (c-set-style "gnu") in a hook, or via M-x c-set-style ---  
 then the style-specific values take precedence over any global style  
 values.  In Lisp terms, global values override style-specific values  
 only when the new second argument to c-set-style is non-nil; see the  
 function documentation for more info.  
   
 The purpose of these changes is to make it easier for users,  
 especially novice users, to do simple customizations with Customize or  
 with setq in their .emacs files.  On the other hand, the new system is  
 intended to be compatible with advanced users' customizations as well,  
 such as those that choose styles in hooks or whatnot.  This new system  
 is believed to be almost entirely compatible with current  
 configurations, in spite of the changed precedence between style and  
 global variable settings when a buffer's default style is set.  
   
 (Thanks to Eric Eide for clarifying this explanation a bit.)  
   
 **** c-offsets-alist is now a customizable variable.  
 This became possible as a result of the new initialization behavior.  
   
 This variable is treated slightly differently from the other style  
 variables; instead of using the symbol set-from-style, it will be  
 completed with the syntactic symbols it doesn't already contain when  
 the style is first initialized.  This means it now defaults to the  
 empty list to make all syntactic elements get their values from the  
 style system.  
   
 **** Compatibility variable to restore the old behavior.  
 In case your configuration doesn't work with this change, you can set  
 c-old-style-variable-behavior to non-nil to get the old behavior back  
 as far as possible.  
   
 *** Improvements to line breaking and text filling.  
 CC Mode now handles this more intelligently and seamlessly wrt the  
 surrounding code, especially inside comments.  For details see the new  
 chapter about this in the manual.  
   
 **** New variable to recognize comment line prefix decorations.  
 The variable c-comment-prefix-regexp has been added to properly  
 recognize the line prefix in both block and line comments.  It's  
 primarily used to initialize the various paragraph recognition and  
 adaptive filling variables that the text handling functions uses.  
   
 **** New variable c-block-comment-prefix.  
 This is a generalization of the now obsolete variable  
 c-comment-continuation-stars to handle arbitrary strings.  
   
 **** CC Mode now uses adaptive fill mode.  
 This to make it adapt better to the paragraph style inside comments.  
   
 It's also possible to use other adaptive filling packages inside CC  
 Mode, notably Kyle E. Jones' Filladapt mode (http://wonderworks.com/).  
 A new convenience function c-setup-filladapt sets up Filladapt for use  
 inside CC Mode.  
   
 Note though that the 2.12 version of Filladapt lacks a feature that  
 causes it to work suboptimally when c-comment-prefix-regexp can match  
 the empty string (which it commonly does).  A patch for that is  
 available from the CC Mode web site (http://www.python.org/emacs/  
 cc-mode/).  
   
 **** The variables `c-hanging-comment-starter-p' and  
 `c-hanging-comment-ender-p', which controlled how comment starters and  
 enders were filled, are not used anymore.  The new version of the  
 function `c-fill-paragraph' keeps the comment starters and enders as  
 they were before the filling.  
   
 **** It's now possible to selectively turn off auto filling.  
 The variable c-ignore-auto-fill is used to ignore auto fill mode in  
 specific contexts, e.g. in preprocessor directives and in string  
 literals.  
   
 **** New context sensitive line break function c-context-line-break.  
 It works like newline-and-indent in normal code, and adapts the line  
 prefix according to the comment style when used inside comments.  If  
 you're normally using newline-and-indent, you might want to switch to  
 this function.  
   
 *** Fixes to IDL mode.  
 It now does a better job in recognizing only the constructs relevant  
 to IDL.  E.g. it no longer matches "class" as the beginning of a  
 struct block, but it does match the CORBA 2.3 "valuetype" keyword.  
 Thanks to Eric Eide.  
   
 *** Improvements to the Whitesmith style.  
 It now keeps the style consistently on all levels and both when  
 opening braces hangs and when they don't.  
   
 **** New lineup function c-lineup-whitesmith-in-block.  
   
 *** New lineup functions c-lineup-template-args and c-indent-multi-line-block.  
 See their docstrings for details.  c-lineup-template-args does a  
 better job of tracking the brackets used as parens in C++ templates,  
 and is used by default to line up continued template arguments.  
   
 *** c-lineup-comment now preserves alignment with a comment on the  
 previous line.  It used to instead preserve comments that started in  
 the column specified by comment-column.  
   
 *** c-lineup-C-comments handles "free form" text comments.  
 In comments with a long delimiter line at the start, the indentation  
 is kept unchanged for lines that start with an empty comment line  
 prefix.  This is intended for the type of large block comments that  
 contain documentation with its own formatting.  In these you normally  
 don't want CC Mode to change the indentation.  
   
 *** The `c' syntactic symbol is now relative to the comment start  
 instead of the previous line, to make integers usable as lineup  
 arguments.  
   
 *** All lineup functions have gotten docstrings.  
   
 *** More preprocessor directive movement functions.  
 c-down-conditional does the reverse of c-up-conditional.  
 c-up-conditional-with-else and c-down-conditional-with-else are  
 variants of these that also stops at "#else" lines (suggested by Don  
 Provan).  
   
 *** Minor improvements to many movement functions in tricky situations.  
   
 ** Dired changes  
   
 *** New variable `dired-recursive-deletes' determines if the delete  
 command will delete non-empty directories recursively.  The default  
 is, delete only empty directories.  
   
 *** New variable `dired-recursive-copies' determines if the copy  
 command will copy directories recursively.  The default is, do not  
 copy directories recursively.  
   
 *** In command `dired-do-shell-command' (usually bound to `!') a `?'  
 in the shell command has a special meaning similar to `*', but with  
 the difference that the command will be run on each file individually.  
   
 *** The new command `dired-find-alternate-file' (usually bound to `a')  
 replaces the Dired buffer with the buffer for an alternate file or  
 directory.  
   
 *** The new command `dired-show-file-type' (usually bound to `y') shows  
 a message in the echo area describing what type of file the point is on.  
 This command invokes the external program `file' do its work, and so  
 will only work on systems with that program, and will be only as  
 accurate or inaccurate as it is.  
   
 *** Dired now properly handles undo changes of adding/removing `-R'  
 from ls switches.  
   
 *** Dired commands that prompt for a destination file now allow the use  
 of the `M-n' command in the minibuffer to insert the source filename,  
 which the user can then edit.  This only works if there is a single  
 source file, not when operating on multiple marked files.  
   
 ** Gnus changes.  
   
 The Gnus NEWS entries are short, but they reflect sweeping changes in  
 four areas: Article display treatment, MIME treatment,  
 internationalization and mail-fetching.  
   
 *** The mail-fetching functions have changed.  See the manual for the  
 many details.  In particular, all procmail fetching variables are gone.  
   
 If you used procmail like in  
   
 (setq nnmail-use-procmail t)  
 (setq nnmail-spool-file 'procmail)  
 (setq nnmail-procmail-directory "~/mail/incoming/")  
 (setq nnmail-procmail-suffix "\\.in")  
   
 this now has changed to  
   
 (setq mail-sources  
       '((directory :path "~/mail/incoming/"  
                    :suffix ".in")))  
   
 More information is available in the info doc at Select Methods ->  
 Getting Mail -> Mail Sources  
   
 *** Gnus is now a MIME-capable reader.  This affects many parts of  
 Gnus, and adds a slew of new commands.  See the manual for details.  
 Separate MIME packages like RMIME, mime-compose etc., will probably no  
 longer work; remove them and use the native facilities.  
   
 The FLIM/SEMI package still works with Emacs 21, but if you want to  
 use the native facilities, you must remove any mailcap.el[c] that was  
 installed by FLIM/SEMI version 1.13 or earlier.  
   
 *** Gnus has also been multilingualized.  This also affects too many  
 parts of Gnus to summarize here, and adds many new variables.  There  
 are built-in facilities equivalent to those of gnus-mule.el, which is  
 now just a compatibility layer.  
   
 *** gnus-mule.el is now just a compatibility layer over the built-in  
 Gnus facilities.  
   
 *** gnus-auto-select-first can now be a function to be  
 called to position point.  
   
 *** The user can now decide which extra headers should be included in  
 summary buffers and NOV files.  
   
 *** `gnus-article-display-hook' has been removed.  Instead, a number  
 of variables starting with `gnus-treat-' have been added.  
   
 *** The Gnus posting styles have been redone again and now work in a  
 subtly different manner.  
   
 *** New web-based backends have been added: nnslashdot, nnwarchive  
 and nnultimate.  nnweb has been revamped, again, to keep up with  
 ever-changing layouts.  
   
 *** Gnus can now read IMAP mail via nnimap.  
   
 *** There is image support of various kinds and some sound support.  
   
 ** Changes in Texinfo mode.  
   
 *** A couple of new key bindings have been added for inserting Texinfo  
 macros  
   
   Key binding   Macro  
   -------------------------  
   C-c C-c C-s   @strong  
   C-c C-c C-e   @emph  
   C-c C-c u     @uref  
   C-c C-c q     @quotation  
   C-c C-c m     @email  
   C-c C-o       @<block> ... @end <block>  
   M-RET         @item  
   
 *** The " key now inserts either " or `` or '' depending on context.  
   
 ** Changes in Outline mode.  
   
 There is now support for Imenu to index headings.  A new command  
 `outline-headers-as-kill' copies the visible headings in the region to  
 the kill ring, e.g. to produce a table of contents.  
   
 ** Changes to Emacs Server  
   
 *** The new option `server-kill-new-buffers' specifies what to do  
 with buffers when done with them.  If non-nil, the default, buffers  
 are killed, unless they were already present before visiting them with  
 Emacs Server.  If nil, `server-temp-file-regexp' specifies which  
 buffers to kill, as before.  
   
 Please note that only buffers are killed that still have a client,  
 i.e. buffers visited with `emacsclient --no-wait' are never killed in  
 this way.  
   
 ** Both emacsclient and Emacs itself now accept command line options  
 of the form +LINE:COLUMN in addition to +LINE.  
   
 ** Changes to Show Paren mode.  
   
 *** Overlays used by Show Paren mode now use a priority property.  
 The new user option show-paren-priority specifies the priority to  
 use.  Default is 1000.  
   
 ** New command M-x check-parens can be used to find unbalanced paren  
 groups and strings in buffers in Lisp mode (or other modes).  
   
 ** Changes to hideshow.el  
   
 *** Generalized block selection and traversal  
   
 A block is now recognized by its start and end regexps (both strings),  
 and an integer specifying which sub-expression in the start regexp  
 serves as the place where a `forward-sexp'-like function can operate.  
 See the documentation of variable `hs-special-modes-alist'.  
   
 *** During incremental search, if Hideshow minor mode is active,  
 hidden blocks are temporarily shown.  The variable `hs-headline' can  
 be used in the mode line format to show the line at the beginning of  
 the open block.  
   
 *** User option `hs-hide-all-non-comment-function' specifies a  
 function to be called at each top-level block beginning, instead of  
 the normal block-hiding function.  
   
 *** The command `hs-show-region' has been removed.  
   
 *** The key bindings have changed to fit the Emacs conventions,  
 roughly imitating those of Outline minor mode.  Notably, the prefix  
 for all bindings is now `C-c @'.  For details, see the documentation  
 for `hs-minor-mode'.  
   
 *** The variable `hs-show-hidden-short-form' has been removed, and  
 hideshow.el now always behaves as if this variable were set to t.  
   
 ** Changes to Change Log mode and Add-Log functions  
   
 *** If you invoke `add-change-log-entry' from a backup file, it makes  
 an entry appropriate for the file's parent.  This is useful for making  
 log entries by comparing a version with deleted functions.  
   
 **** New command M-x change-log-merge merges another log into the  
 current buffer.  
   
 *** New command M-x change-log-redate fixes any old-style date entries  
 in a log file.  
   
 *** Change Log mode now adds a file's version number to change log  
 entries if user-option `change-log-version-info-enabled' is non-nil.  
 Unless the file is under version control the search for a file's  
 version number is performed based on regular expressions from  
 `change-log-version-number-regexp-list' which can be customized.  
 Version numbers are only found in the first 10 percent of a file.  
   
 *** Change Log mode now defines its own faces for font-lock highlighting.  
   
 ** Changes to cmuscheme  
   
 *** The user-option `scheme-program-name' has been renamed  
 `cmuscheme-program-name' due to conflicts with xscheme.el.  
   
 ** Changes in Font Lock  
   
 *** The new function `font-lock-remove-keywords' can be used to remove  
 font-lock keywords from the current buffer or from a specific major mode.  
   
 *** Multi-line patterns are now supported.  Modes using this, should  
 set font-lock-multiline to t in their font-lock-defaults.  
   
 *** `font-lock-syntactic-face-function' allows major-modes to choose  
 the face used for each string/comment.  
   
 *** A new standard face `font-lock-doc-face'.  
 Meant for Lisp docstrings, Javadoc comments and other "documentation in code".  
   
 ** Changes to Shell mode  
   
 *** The `shell' command now accepts an optional argument to specify the buffer  
 to use, which defaults to "*shell*".  When used interactively, a  
 non-default buffer may be specified by giving the `shell' command a  
 prefix argument (causing it to prompt for the buffer name).  
   
 ** Comint (subshell) changes  
   
 These changes generally affect all modes derived from comint mode, which  
 include shell-mode, gdb-mode, scheme-interaction-mode, etc.  
   
 *** Comint now by default interprets some carriage-control characters.  
 Comint now removes CRs from CR LF sequences, and treats single CRs and  
 BSs in the output in a way similar to a terminal (by deleting to the  
 beginning of the line, or deleting the previous character,  
 respectively).  This is achieved by adding `comint-carriage-motion' to  
 the `comint-output-filter-functions' hook by default.  
   
 *** By default, comint no longer uses the variable `comint-prompt-regexp'  
 to distinguish prompts from user-input.  Instead, it notices which  
 parts of the text were output by the process, and which entered by the  
 user, and attaches `field' properties to allow emacs commands to use  
 this information.  Common movement commands, notably beginning-of-line,  
 respect field boundaries in a fairly natural manner.  To disable this  
 feature, and use the old behavior, customize the user option  
 `comint-use-prompt-regexp-instead-of-fields'.  
   
 *** Comint now includes new features to send commands to running processes  
 and redirect the output to a designated buffer or buffers.  
   
 *** The command M-x comint-redirect-send-command reads a command and  
 buffer name from the mini-buffer.  The command is sent to the current  
 buffer's process, and its output is inserted into the specified buffer.  
   
 The command M-x comint-redirect-send-command-to-process acts like  
 M-x comint-redirect-send-command but additionally reads the name of  
 the buffer whose process should be used from the mini-buffer.  
   
 *** Packages based on comint now highlight user input and program prompts,  
 and support choosing previous input with mouse-2.  To control these features,  
 see the user-options `comint-highlight-input' and `comint-highlight-prompt'.  
   
 *** The new command `comint-write-output' (usually bound to `C-c C-s')  
 saves the output from the most recent command to a file.  With a prefix  
 argument, it appends to the file.  
   
 *** The command `comint-kill-output' has been renamed `comint-delete-output'  
 (usually bound to `C-c C-o'); the old name is aliased to it for  
 compatibility.  
   
 *** The new function `comint-add-to-input-history' adds commands to the input  
 ring (history).  
   
 *** The new variable `comint-input-history-ignore' is a regexp for  
 identifying history lines that should be ignored, like tcsh time-stamp  
 strings, starting with a `#'.  The default value of this variable is "^#".  
   
 ** Changes to Rmail mode  
   
 *** The new user-option rmail-user-mail-address-regexp can be  
 set to fine tune the identification of the correspondent when  
 receiving new mail.  If it matches the address of the sender, the  
 recipient is taken as correspondent of a mail.  If nil, the default,  
 `user-login-name' and `user-mail-address' are used to exclude yourself  
 as correspondent.  
   
 Usually you don't have to set this variable, except if you collect  
 mails sent by you under different user names.  Then it should be a  
 regexp matching your mail addresses.  
   
 *** The new user-option rmail-confirm-expunge controls whether and how  
 to ask for confirmation before expunging deleted messages from an  
 Rmail file.  You can choose between no confirmation, confirmation  
 with y-or-n-p, or confirmation with yes-or-no-p.  Default is to ask  
 for confirmation with yes-or-no-p.  
   
 *** RET is now bound in the Rmail summary to rmail-summary-goto-msg,  
 like `j'.  
   
 *** There is a new user option `rmail-digest-end-regexps' that  
 specifies the regular expressions to detect the line that ends a  
 digest message.  
   
 *** The new user option `rmail-automatic-folder-directives' specifies  
 in which folder to put messages automatically.  
   
 *** The new function `rmail-redecode-body' allows to fix a message  
 with non-ASCII characters if Emacs happens to decode it incorrectly  
 due to missing or malformed "charset=" header.  
   
 ** The new user-option `mail-envelope-from' can be used to specify  
 an envelope-from address different from user-mail-address.  
   
 ** The variable mail-specify-envelope-from controls whether to  
 use the -f option when sending mail.  
   
 ** The Rmail command `o' (`rmail-output-to-rmail-file') now writes the  
 current message in the internal `emacs-mule' encoding, rather than in  
 the encoding taken from the variable `buffer-file-coding-system'.  
 This allows to save messages whose characters cannot be safely encoded  
 by the buffer's coding system, and makes sure the message will be  
 displayed correctly when you later visit the target Rmail file.  
   
 If you want your Rmail files be encoded in a specific coding system  
 other than `emacs-mule', you can customize the variable  
 `rmail-file-coding-system' to set its value to that coding system.  
   
 ** Changes to TeX mode  
   
 *** The default mode has been changed from `plain-tex-mode' to  
 `latex-mode'.  
   
 *** latex-mode now has a simple indentation algorithm.  
   
 *** M-f and M-p jump around \begin...\end pairs.  
   
 *** Added support for outline-minor-mode.  
   
 ** Changes to RefTeX mode  
   
 *** RefTeX has new support for index generation.  Index entries can be  
     created with `C-c <', with completion available on index keys.  
     Pressing `C-c /' indexes the word at the cursor with a default  
     macro.  `C-c >' compiles all index entries into an alphabetically  
     sorted *Index* buffer which looks like the final index.  Entries  
     can be edited from that buffer.  
   
 *** Label and citation key selection now allow to select several  
     items and reference them together (use `m' to mark items, `a' or  
     `A' to use all marked entries).  
   
 *** reftex.el has been split into a number of smaller files to reduce  
     memory use when only a part of RefTeX is being used.  
   
 *** a new command `reftex-view-crossref-from-bibtex' (bound to `C-c &'  
     in BibTeX-mode) can be called in a BibTeX database buffer in order  
     to show locations in LaTeX documents where a particular entry has  
     been cited.  
   
 ** Emacs Lisp mode now allows multiple levels of outline headings.  
 The level of a heading is determined from the number of leading  
 semicolons in a heading line.  Toplevel forms starting with a `('  
 in column 1 are always made leaves.  
   
 ** The M-x time-stamp command (most commonly used on write-file-hooks)  
 has the following new features:  
   
 *** The patterns for finding the time stamp and for updating a pattern  
 may match text spanning multiple lines.  For example, some people like  
 to have the filename and date on separate lines.  The new variable  
 time-stamp-inserts-lines controls the matching for multi-line patterns.  
   
 *** More than one time stamp can be updated in the same file.  This  
 feature is useful if you need separate time stamps in a program source  
 file to both include in formatted documentation and insert in the  
 compiled binary.  The same time-stamp will be written at each matching  
 pattern.  The variable time-stamp-count enables this new feature; it  
 defaults to 1.  
   
 ** Partial Completion mode now completes environment variables in  
 file names.  
   
 ** Ispell changes  
   
 *** The command `ispell' now spell-checks a region if  
 transient-mark-mode is on, and the mark is active.  Otherwise it  
 spell-checks the current buffer.  
   
 *** Support for synchronous subprocesses - DOS/Windoze - has been  
 added.  
   
 *** An "alignment error" bug was fixed when a manual spelling  
 correction is made and re-checked.  
   
 *** An Italian, Portuguese, and Slovak dictionary definition has been added.  
   
 *** Region skipping performance has been vastly improved in some  
 cases.  
   
 *** Spell checking HTML buffers has been improved and isn't so strict  
 on syntax errors.  
   
 *** The buffer-local words are now always placed on a new line at the  
 end of the buffer.  
   
 *** Spell checking now works in the MS-DOS version of Emacs.  
   
 ** Makefile mode changes  
   
 *** The mode now uses the abbrev table `makefile-mode-abbrev-table'.  
   
 *** Conditionals and include statements are now highlighted when  
 Fontlock mode is active.  
   
 ** Isearch changes  
   
 *** Isearch now puts a call to `isearch-resume' in the command history,  
 so that searches can be resumed.  
   
 *** In Isearch mode, C-M-s and C-M-r are now bound like C-s and C-r,  
 respectively, i.e. you can repeat a regexp isearch with the same keys  
 that started the search.  
   
 *** In Isearch mode, mouse-2 in the echo area now yanks the current  
 selection into the search string rather than giving an error.  
   
 *** There is a new lazy highlighting feature in incremental search.  
   
 Lazy highlighting is switched on/off by customizing variable  
 `isearch-lazy-highlight'.  When active, all matches for the current  
 search string are highlighted.  The current match is highlighted as  
 before using face `isearch' or `region'.  All other matches are  
 highlighted using face `isearch-lazy-highlight-face' which defaults to  
 `secondary-selection'.  
   
 The extra highlighting makes it easier to anticipate where the cursor  
 will end up each time you press C-s or C-r to repeat a pending search.  
 Highlighting of these additional matches happens in a deferred fashion  
 using "idle timers," so the cycles needed do not rob isearch of its  
 usual snappy response.  
   
 If `isearch-lazy-highlight-cleanup' is set to t, highlights for  
 matches are automatically cleared when you end the search.  If it is  
 set to nil, you can remove the highlights manually with `M-x  
 isearch-lazy-highlight-cleanup'.  
   
 ** VC Changes  
   
 VC has been overhauled internally.  It is now modular, making it  
 easier to plug-in arbitrary version control backends.  (See Lisp  
 Changes for details on the new structure.)  As a result, the mechanism  
 to enable and disable support for particular version systems has  
 changed: everything is now controlled by the new variable  
 `vc-handled-backends'.  Its value is a list of symbols that identify  
 version systems; the default is '(RCS CVS SCCS).  When finding a file,  
 each of the backends in that list is tried in order to see whether the  
 file is registered in that backend.  
   
 When registering a new file, VC first tries each of the listed  
 backends to see if any of them considers itself "responsible" for the  
 directory of the file (e.g. because a corresponding subdirectory for  
 master files exists).  If none of the backends is responsible, then  
 the first backend in the list that could register the file is chosen.  
 As a consequence, the variable `vc-default-back-end' is now obsolete.  
   
 The old variable `vc-master-templates' is also obsolete, although VC  
 still supports it for backward compatibility.  To define templates for  
 RCS or SCCS, you should rather use the new variables  
 vc-{rcs,sccs}-master-templates.  (There is no such feature under CVS  
 where it doesn't make sense.)  
   
 The variables `vc-ignore-vc-files' and `vc-handle-cvs' are also  
 obsolete now, you must set `vc-handled-backends' to nil or exclude  
 `CVS' from the list, respectively, to achieve their effect now.  
   
 *** General Changes  
   
 The variable `vc-checkout-carefully' is obsolete: the corresponding  
 checks are always done now.  
   
 VC Dired buffers are now kept up-to-date during all version control  
 operations.  
   
 `vc-diff' output is now displayed in `diff-mode'.  
 `vc-print-log' uses `log-view-mode'.  
 `vc-log-mode' (used for *VC-Log*) has been replaced by `log-edit-mode'.  
   
 The command C-x v m (vc-merge) now accepts an empty argument as the  
 first revision number.  This means that any recent changes on the  
 current branch should be picked up from the repository and merged into  
 the working file (``merge news'').  
   
 The commands C-x v s (vc-create-snapshot) and C-x v r  
 (vc-retrieve-snapshot) now ask for a directory name from which to work  
 downwards.  
   
 *** Multiple Backends  
   
 VC now lets you register files in more than one backend.  This is  
 useful, for example, if you are working with a slow remote CVS  
 repository.  You can then use RCS for local editing, and occasionally  
 commit your changes back to CVS, or pick up changes from CVS into your  
 local RCS archives.  
   
 To make this work, the ``more local'' backend (RCS in our example)  
 should come first in `vc-handled-backends', and the ``more remote''  
 backend (CVS) should come later.  (The default value of  
 `vc-handled-backends' already has it that way.)  
   
 You can then commit changes to another backend (say, RCS), by typing  
 C-u C-x v v RCS RET (i.e. vc-next-action now accepts a backend name as  
 a revision number).  VC registers the file in the more local backend  
 if that hasn't already happened, and commits to a branch based on the  
 current revision number from the more remote backend.  
   
 If a file is registered in multiple backends, you can switch to  
 another one using C-x v b (vc-switch-backend).  This does not change  
 any files, it only changes VC's perspective on the file.  Use this to  
 pick up changes from CVS while working under RCS locally.  
   
 After you are done with your local RCS editing, you can commit your  
 changes back to CVS using C-u C-x v v CVS RET.  In this case, the  
 local RCS archive is removed after the commit, and the log entry  
 buffer is initialized to contain the entire RCS change log of the file.  
   
 *** Changes for CVS  
   
 There is a new user option, `vc-cvs-stay-local'.  If it is `t' (the  
 default), then VC avoids network queries for files registered in  
 remote repositories.  The state of such files is then only determined  
 by heuristics and past information.  `vc-cvs-stay-local' can also be a  
 regexp to match against repository hostnames; only files from hosts  
 that match it are treated locally.  If the variable is nil, then VC  
 queries the repository just as often as it does for local files.  
   
 If `vc-cvs-stay-local' is on, then VC also makes local backups of  
 repository versions.  This means that ordinary diffs (C-x v =) and  
 revert operations (C-x v u) can be done completely locally, without  
 any repository interactions at all.  The name of a local version  
 backup of FILE is FILE.~REV.~, where REV is the repository version  
 number.  This format is similar to that used by C-x v ~  
 (vc-version-other-window), except for the trailing dot.  As a matter  
 of fact, the two features can each use the files created by the other,  
 the only difference being that files with a trailing `.' are deleted  
 automatically after commit.  (This feature doesn't work on MS-DOS,  
 since DOS disallows more than a single dot in the trunk of a file  
 name.)  
   
 If `vc-cvs-stay-local' is on, and there have been changes in the  
 repository, VC notifies you about it when you actually try to commit.  
 If you want to check for updates from the repository without trying to  
 commit, you can either use C-x v m RET to perform an update on the  
 current file, or you can use C-x v r RET to get an update for an  
 entire directory tree.  
   
 The new user option `vc-cvs-use-edit' indicates whether VC should call  
 "cvs edit" to make files writeable; it defaults to `t'.  (This option  
 is only meaningful if the CVSREAD variable is set, or if files are  
 "watched" by other developers.)  
   
 The commands C-x v s (vc-create-snapshot) and C-x v r  
 (vc-retrieve-snapshot) are now also implemented for CVS.  If you give  
 an empty snapshot name to the latter, that performs a `cvs update',  
 starting at the given directory.  
   
 *** Lisp Changes in VC  
   
 VC has been restructured internally to make it modular.  You can now  
 add support for arbitrary version control backends by writing a  
 library that provides a certain set of backend-specific functions, and  
 then telling VC to use that library.  For example, to add support for  
 a version system named SYS, you write a library named vc-sys.el, which  
 provides a number of functions vc-sys-... (see commentary at the top  
 of vc.el for a detailed list of them).  To make VC use that library,  
 you need to put it somewhere into Emacs' load path and add the symbol  
 `SYS' to the list `vc-handled-backends'.  
   
 ** The customizable EDT emulation package now supports the EDT  
 SUBS command and EDT scroll margins.  It also works with more  
 terminal/keyboard configurations and it now works under XEmacs.  
 See etc/edt-user.doc for more information.  
   
 ** New modes and packages  
   
 *** The new global minor mode `minibuffer-electric-default-mode'  
 automatically hides the `(default ...)' part of minibuffer prompts when  
 the default is not applicable.  
   
 *** Artist is an Emacs lisp package that allows you to draw lines,  
 rectangles and ellipses by using your mouse and/or keyboard.  The  
 shapes are made up with the ascii characters |, -, / and \.  
   
 Features are:  
   
 - Intersecting: When a `|' intersects with a `-', a `+' is  
   drawn, like this:   |         \ /  
                     --+--        X  
                       |         / \  
   
 - Rubber-banding: When drawing lines you can interactively see the  
   result while holding the mouse button down and moving the mouse.  If  
   your machine is not fast enough (a 386 is a bit too slow, but a  
   pentium is well enough), you can turn this feature off.  You will  
   then see 1's and 2's which mark the 1st and 2nd endpoint of the line  
   you are drawing.  
   
 - Arrows: After having drawn a (straight) line or a (straight)  
   poly-line, you can set arrows on the line-ends by typing < or >.  
   
 - Flood-filling: You can fill any area with a certain character by  
   flood-filling.  
   
 - Cut copy and paste: You can cut, copy and paste rectangular  
   regions.  Artist also interfaces with the rect package (this can be  
   turned off if it causes you any trouble) so anything you cut in  
   artist can be yanked with C-x r y and vice versa.  
   
 - Drawing with keys: Everything you can do with the mouse, you can  
   also do without the mouse.  
   
 - Aspect-ratio: You can set the variable artist-aspect-ratio to  
   reflect the height-width ratio for the font you are using. Squares  
   and circles are then drawn square/round.  Note, that once your  
   ascii-file is shown with font with a different height-width ratio,  
   the squares won't be square and the circles won't be round.  
   
 - Drawing operations: The following drawing operations are implemented:  
   
     lines               straight-lines  
     rectangles          squares  
     poly-lines          straight poly-lines  
     ellipses            circles  
     text (see-thru)     text (overwrite)  
     spray-can           setting size for spraying  
     vaporize line       vaporize lines  
     erase characters    erase rectangles  
   
   Straight lines are lines that go horizontally, vertically or  
   diagonally.  Plain lines go in any direction.  The operations in  
   the right column are accessed by holding down the shift key while  
   drawing.  
   
   It is possible to vaporize (erase) entire lines and connected lines  
   (rectangles for example) as long as the lines being vaporized are  
   straight and connected at their endpoints.  Vaporizing is inspired  
   by the drawrect package by Jari Aalto <jari.aalto@poboxes.com>.  
   
 - Picture mode compatibility: Artist is picture mode compatible (this  
   can be turned off).  
   
 *** The new package Eshell is an operating system command shell  
 implemented entirely in Emacs Lisp.  Use `M-x eshell' to invoke it.  
 It functions similarly to bash and zsh, and allows running of Lisp  
 functions and external commands using the same syntax.  It supports  
 history lists, aliases, extended globbing, smart scrolling, etc.  It  
 will work on any platform Emacs has been ported to.  And since most of  
 the basic commands -- ls, rm, mv, cp, ln, du, cat, etc. -- have been  
 rewritten in Lisp, it offers an operating-system independent shell,  
 all within the scope of your Emacs process.  
   
 *** The new package timeclock.el is a mode is for keeping track of time  
 intervals.  You can use it for whatever purpose you like, but the  
 typical scenario is to keep track of how much time you spend working  
 on certain projects.  
   
 *** The new package hi-lock.el provides commands to highlight matches  
 of interactively entered regexps.  For example,  
   
   M-x highlight-regexp RET clearly RET RET  
   
 will highlight all occurrences of `clearly' using a yellow background  
 face.  New occurrences of `clearly' will be highlighted as they are  
 typed.  `M-x unhighlight-regexp RET' will remove the highlighting.  
 Any existing face can be used for highlighting and a set of  
 appropriate faces is provided.  The regexps can be written into the  
 current buffer in a form that will be recognized the next time the  
 corresponding file is read.  There are commands to highlight matches  
 to phrases and to highlight entire lines containing a match.  
   
 *** The new package zone.el plays games with Emacs' display when  
 Emacs is idle.  
   
 *** The new package tildify.el allows to add hard spaces or other text  
 fragments in accordance with the current major mode.  
   
 *** The new package xml.el provides a simple but generic XML  
 parser. It doesn't parse the DTDs however.  
   
 *** The comment operations are now provided by the newcomment.el  
 package which allows different styles of comment-region and should  
 be more robust while offering the same functionality.  
 `comment-region' now doesn't always comment a-line-at-a-time, but only  
 comments the region, breaking the line at point if necessary.  
   
 *** The Ebrowse package implements a C++ class browser and tags  
 facilities tailored for use with C++.  It is documented in a  
 separate Texinfo file.  
   
 *** The PCL-CVS package available by either running M-x cvs-examine or  
 by visiting a CVS administrative directory (with a prefix argument)  
 provides an alternative interface to VC-dired for CVS.  It comes with  
 `log-view-mode' to view RCS and SCCS logs and `log-edit-mode' used to  
 enter check-in log messages.  
   
 *** The new package called `woman' allows to browse Unix man pages  
 without invoking external programs.  
   
 The command `M-x woman' formats manual pages entirely in Emacs Lisp  
 and then displays them, like `M-x manual-entry' does.  Unlike  
 `manual-entry', `woman' does not invoke any external programs, so it  
 is useful on systems such as MS-DOS/MS-Windows where the `man' and  
 Groff or `troff' commands are not readily available.  
   
 The command `M-x woman-find-file' asks for the file name of a man  
 page, then formats and displays it like `M-x woman' does.  
   
 *** The new command M-x re-builder offers a convenient interface for  
 authoring regular expressions with immediate visual feedback.  
   
 The buffer from which the command was called becomes the target for  
 the regexp editor popping up in a separate window.  Matching text in  
 the target buffer is immediately color marked during the editing.  
 Each sub-expression of the regexp will show up in a different face so  
 even complex regexps can be edited and verified on target data in a  
 single step.  
   
 On displays not supporting faces the matches instead blink like  
 matching parens to make them stand out.  On such a setup you will  
 probably also want to use the sub-expression mode when the regexp  
 contains such to get feedback about their respective limits.  
   
 *** glasses-mode is a minor mode that makes  
 unreadableIdentifiersLikeThis readable.  It works as glasses, without  
 actually modifying content of a buffer.  
   
 *** The package ebnf2ps translates an EBNF to a syntactic chart in  
 PostScript.  
   
 Currently accepts ad-hoc EBNF, ISO EBNF and Bison/Yacc.  
   
 The ad-hoc default EBNF syntax has the following elements:  
   
     ;           comment (until end of line)  
     A           non-terminal  
     "C"         terminal  
     ?C?         special  
     $A          default non-terminal  
     $"C"        default terminal  
     $?C?        default special  
     A = B.      production (A is the header and B the body)  
     C D         sequence (C occurs before D)  
     C | D       alternative (C or D occurs)  
     A - B       exception (A excluding B, B without any non-terminal)  
     n * A       repetition (A repeats n (integer) times)  
     (C)         group (expression C is grouped together)  
     [C]         optional (C may or not occurs)  
     C+          one or more occurrences of C  
     {C}+        one or more occurrences of C  
     {C}*        zero or more occurrences of C  
     {C}         zero or more occurrences of C  
     C / D       equivalent to: C {D C}*  
     {C || D}+   equivalent to: C {D C}*  
     {C || D}*   equivalent to: [C {D C}*]  
     {C || D}    equivalent to: [C {D C}*]  
   
 Please, see ebnf2ps documentation for EBNF syntax and how to use it.  
   
 *** The package align.el will align columns within a region, using M-x  
 align.  Its mode-specific rules, based on regular expressions,  
 determine where the columns should be split.  In C and C++, for  
 example, it will align variable names in declaration lists, or the  
 equal signs of assignments.  
   
 *** `paragraph-indent-minor-mode' is a new minor mode supporting  
 paragraphs in the same style as `paragraph-indent-text-mode'.  
   
 *** bs.el is a new package for buffer selection similar to  
 list-buffers or electric-buffer-list.  Use M-x bs-show to display a  
 buffer menu with this package.  See the Custom group `bs'.  
   
 *** find-lisp.el is a package emulating the Unix find command in Lisp.  
   
 *** calculator.el is a small calculator package that is intended to  
 replace desktop calculators such as xcalc and calc.exe.  Actually, it  
 is not too small - it has more features than most desktop calculators,  
 and can be customized easily to get many more functions.  It should  
 not be confused with "calc" which is a much bigger mathematical tool  
 which answers different needs.  
   
 *** The minor modes cwarn-mode and global-cwarn-mode highlights  
 suspicious C and C++ constructions.  Currently, assignments inside  
 expressions, semicolon following `if', `for' and `while' (except, of  
 course, after a `do .. while' statement), and C++ functions with  
 reference parameters are recognized.  The modes require font-lock mode  
 to be enabled.  
   
 *** smerge-mode.el provides `smerge-mode', a simple minor-mode for files  
 containing diff3-style conflict markers, such as generated by RCS.  
   
 *** 5x5.el is a simple puzzle game.  
   
 *** hl-line.el provides `hl-line-mode', a minor mode to highlight the  
 current line in the current buffer.  It also provides  
 `global-hl-line-mode' to provide the same behaviour in all buffers.  
   
 *** ansi-color.el translates ANSI terminal escapes into text-properties.  
   
 Please note: if `ansi-color-for-comint-mode' and  
 `global-font-lock-mode' are non-nil, loading ansi-color.el will  
 disable font-lock and add `ansi-color-apply' to  
 `comint-preoutput-filter-functions' for all shell-mode buffers.  This  
 displays the output of "ls --color=yes" using the correct foreground  
 and background colors.  
   
 *** delphi.el provides a major mode for editing the Delphi (Object  
 Pascal) language.  
   
 *** quickurl.el provides a simple method of inserting a URL based on  
 the text at point.  
   
 *** sql.el provides an interface to SQL data bases.  
   
 *** fortune.el uses the fortune program to create mail/news signatures.  
   
 *** whitespace.el is a package for warning about and cleaning bogus  
 whitespace in a file.  
   
 *** PostScript mode (ps-mode) is a new major mode for editing PostScript  
 files. It offers: interaction with a PostScript interpreter, including  
 (very basic) error handling; fontification, easily customizable for  
 interpreter messages; auto-indentation; insertion of EPSF templates and  
 often used code snippets; viewing of BoundingBox; commenting out /  
 uncommenting regions; conversion of 8bit characters to PostScript octal  
 codes. All functionality is accessible through a menu.  
   
 *** delim-col helps to prettify columns in a text region or rectangle.  
61    
62  Here is an example of columns:  New functions: characterp, max-char, map-charset-chars,
63    define-charset-alias, primary-charset, set-primary-charset,
64    unify-charset, clear-charset-maps, charset-priority-list,
65    set-charset-priority, define-coding-system,
66    define-coding-system-alias, coding-system-aliases
67    
68  horse   apple   bus  Changed functions: copy-sequence, decode-char, encode-char,
69  dog     pineapple       car     EXTRA  set-fontset-font, new-fontset, modify-syntax-entry, define-charset,
70  porcupine       strawberry      airplane  modify-category-entry
71    
72  Doing the following settings:  Obsoleted: char-bytes, chars-in-region, set-coding-priority,
73    char-valid-p
    (setq delimit-columns-str-before "[ ")  
    (setq delimit-columns-str-after " ]")  
    (setq delimit-columns-str-separator ", ")  
    (setq delimit-columns-separator "\t")  
   
   
 Selecting the lines above and typing:  
   
    M-x delimit-columns-region  
   
 It results:  
   
 [ horse    , apple     , bus     ,       ]  
 [ dog      , pineapple , car     , EXTRA ]  
 [ porcupine, strawberry, airplane,       ]  
   
 delim-col has the following options:  
   
    delimit-columns-str-before           Specify a string to be inserted  
                                         before all columns.  
   
    delimit-columns-str-separator        Specify a string to be inserted  
                                         between each column.  
   
    delimit-columns-str-after            Specify a string to be inserted  
                                         after all columns.  
   
    delimit-columns-separator            Specify a regexp which separates  
                                         each column.  
   
 delim-col has the following commands:  
   
    delimit-columns-region       Prettify all columns in a text region.  
    delimit-columns-rectangle    Prettify all columns in a text rectangle.  
   
 *** Recentf mode maintains a menu for visiting files that were  
 operated on recently.  User option recentf-menu-filter specifies a  
 menu filter function to change the menu appearance. For example, the  
 recent file list can be displayed:  
   
 - organized by major modes, directories or user defined rules.  
 - sorted by file paths, file names, ascending or descending.  
 - showing paths relative to the current default-directory  
   
 The `recentf-filter-changer' menu filter function allows to  
 dynamically change the menu appearance.  
   
 *** elide-head.el provides a mechanism for eliding boilerplate header  
 text.  
   
 *** footnote.el provides `footnote-mode', a minor mode supporting use  
 of footnotes.  It is intended for use with Message mode, but isn't  
 specific to Message mode.  
   
 *** diff-mode.el provides `diff-mode', a major mode for  
 viewing/editing context diffs (patches).  It is selected for files  
 with extension `.diff', `.diffs', `.patch' and `.rej'.  
   
 *** EUDC, the Emacs Unified Directory Client, provides a common user  
 interface to access directory servers using different directory  
 protocols.  It has a separate manual.  
   
 *** autoconf.el provides a major mode for editing configure.in files  
 for Autoconf, selected automatically.  
   
 *** windmove.el provides moving between windows.  
   
 *** crm.el provides a facility to read multiple strings from the  
 minibuffer with completion.  
   
 *** todo-mode.el provides management of TODO lists and integration  
 with the diary features.  
   
 *** autoarg.el provides a feature reported from Twenex Emacs whereby  
 numeric keys supply prefix args rather than self inserting.  
   
 *** The function `turn-off-auto-fill' unconditionally turns off Auto  
 Fill mode.  
   
 *** pcomplete.el is a library that provides programmable completion  
 facilities for Emacs, similar to what zsh and tcsh offer.  The main  
 difference is that completion functions are written in Lisp, meaning  
 they can be profiled, debugged, etc.  
   
 *** antlr-mode is a new major mode for editing ANTLR grammar files.  
 It is automatically turned on for files whose names have the extension  
 `.g'.  
   
 ** Changes in sort.el  
   
 The function sort-numeric-fields interprets numbers starting with `0'  
 as octal and numbers starting with `0x' or `0X' as hexadecimal.  The  
 new user-option sort-numeric-base can be used to specify a default  
 numeric base.  
   
 ** Changes to Ange-ftp  
   
 *** Ange-ftp allows you to specify of a port number in remote file  
 names cleanly.  It is appended to the host name, separated by a hash  
 sign, e.g. `/foo@bar.org#666:mumble'.  (This syntax comes from EFS.)  
   
 *** If the new user-option `ange-ftp-try-passive-mode' is set, passive  
 ftp mode will be used if the ftp client supports that.  
   
 *** Ange-ftp handles the output of the w32-style clients which  
 output ^M at the end of lines.  
   
 ** The recommended way of using Iswitchb is via the new global minor  
 mode `iswitchb-mode'.  
   
 ** Just loading the msb package doesn't switch on Msb mode anymore.  
 If you have `(require 'msb)' in your .emacs, please replace it with  
 `(msb-mode 1)'.  
   
 ** Flyspell mode has various new options.  See the `flyspell' Custom  
 group.  
   
 ** The user option `backward-delete-char-untabify-method' controls the  
 behavior of `backward-delete-char-untabify'.  The following values  
 are recognized:  
   
 `untabify' -- turn a tab to many spaces, then delete one space;  
 `hungry'   -- delete all whitespace, both tabs and spaces;  
 `all'      -- delete all whitespace, including tabs, spaces and newlines;  
 nil        -- just delete one character.  
   
 Default value is `untabify'.  
   
 [This change was made in Emacs 20.3 but not mentioned then.]  
   
 ** In Cperl mode `cperl-invalid-face' should now be a normal face  
 symbol, not double-quoted.  
   
 ** Some packages are declared obsolete, to be removed in a future  
 version.  They are:  auto-show, c-mode, hilit19, hscroll, ooutline,  
 profile, rnews, rnewspost, and sc.  Their implementations have been  
 moved to lisp/obsolete.  
   
 ** auto-compression mode is no longer enabled just by loading jka-compr.el.  
 To control it, set `auto-compression-mode' via Custom or use the  
 `auto-compression-mode' command.  
   
 ** `browse-url-gnome-moz' is a new option for  
 `browse-url-browser-function', invoking Mozilla in GNOME, and  
 `browse-url-kde' can be chosen for invoking the KDE browser.  
   
 ** The user-option `browse-url-new-window-p' has been renamed to  
 `browse-url-new-window-flag'.  
   
 ** The functions `keep-lines', `flush-lines' and `how-many' now  
 operate on the active region in Transient Mark mode.  
   
 ** `gnus-user-agent' is a new possibility for `mail-user-agent'.  It  
 is like `message-user-agent', but with all the Gnus paraphernalia.  
   
 ** The Strokes package has been updated.  If your Emacs has XPM  
 support, you can use it for pictographic editing.  In Strokes mode,  
 use C-mouse-2 to compose a complex stoke and insert it into the  
 buffer.  You can encode or decode a strokes buffer with new commands  
 M-x strokes-encode-buffer and M-x strokes-decode-buffer.  There is a  
 new command M-x strokes-list-strokes.  
   
 ** Hexl contains a new command `hexl-insert-hex-string' which inserts  
 a string of hexadecimal numbers read from the mini-buffer.  
   
 ** Hexl mode allows to insert non-ASCII characters.  
   
 The non-ASCII characters are encoded using the same encoding as the  
 file you are visiting in Hexl mode.  
   
 ** Shell script mode changes.  
   
 Shell script mode (sh-script) can now indent scripts for shells  
 derived from sh and rc.  The indentation style is customizable, and  
 sh-script can attempt to "learn" the current buffer's style.  
   
 ** Etags changes.  
   
 *** In DOS, etags looks for file.cgz if it cannot find file.c.  
   
 *** New option --ignore-case-regex is an alternative to --regex.  It is now  
 possible to bind a regexp to a language, by prepending the regexp with  
 {lang}, where lang is one of the languages that `etags --help' prints out.  
 This feature is useful especially for regex files, where each line contains  
 a regular expression.  The manual contains details.  
   
 *** In C and derived languages, etags creates tags for function  
 declarations when given the --declarations option.  
   
 *** In C++, tags are created for "operator".  The tags have the form  
 "operator+", without spaces between the keyword and the operator.  
   
 *** You shouldn't generally need any more the -C or -c++ option: etags  
 automatically switches to C++ parsing when it meets the `class' or  
 `template' keywords.  
   
 *** Etags now is able to delve at arbitrary deeps into nested structures in  
 C-like languages.  Previously, it was limited to one or two brace levels.  
   
 *** New language Ada: tags are functions, procedures, packages, tasks, and  
 types.  
   
 *** In Fortran, `procedure' is not tagged.  
   
 *** In Java, tags are created for "interface".  
   
 *** In Lisp, "(defstruct (foo", "(defun (operator" and similar constructs  
 are now tagged.  
   
 *** In makefiles, tags the targets.  
   
 *** In Perl, the --globals option tags global variables.  my and local  
 variables are tagged.  
   
 *** New language Python: def and class at the beginning of a line are tags.  
   
 *** .ss files are Scheme files, .pdb is Postscript with C syntax, .psw is  
 for PSWrap.  
   
 ** Changes in etags.el  
   
 *** The new user-option tags-case-fold-search can be used to make  
 tags operations case-sensitive or case-insensitive.  The default  
 is to use the same setting as case-fold-search.  
   
 *** You can display additional output with M-x tags-apropos by setting  
 the new variable tags-apropos-additional-actions.  
   
 If non-nil, the variable's value should be a list of triples (TITLE  
 FUNCTION TO-SEARCH).  For each triple, M-x tags-apropos processes  
 TO-SEARCH and lists tags from it.  TO-SEARCH should be an alist,  
 obarray, or symbol.  If it is a symbol, the symbol's value is used.  
   
 TITLE is a string to use to label the list of tags from TO-SEARCH.  
   
 FUNCTION is a function to call when an entry is selected in the Tags  
 List buffer.  It is called with one argument, the selected symbol.  
   
 A useful example value for this variable might be something like:  
   
   '(("Emacs Lisp" Info-goto-emacs-command-node obarray)  
     ("Common Lisp" common-lisp-hyperspec common-lisp-hyperspec-obarray)  
     ("SCWM" scwm-documentation scwm-obarray))  
   
 *** The face tags-tag-face can be used to customize the appearance  
 of tags in the output of M-x tags-apropos.  
   
 *** Setting tags-apropos-verbose to a non-nil value displays the  
 names of tags files in the *Tags List* buffer.  
   
 *** You can now search for tags that are part of the filename itself.  
 If you have tagged the files topfile.c subdir/subfile.c  
 /tmp/tempfile.c, you can now search for tags "topfile.c", "subfile.c",  
 "dir/sub", "tempfile", "tempfile.c".  If the tag matches the file name,  
 point will go to the beginning of the file.  
   
 *** Compressed files are now transparently supported if  
 auto-compression-mode is active.  You can tag (with Etags) and search  
 (with find-tag) both compressed and uncompressed files.  
   
 *** Tags commands like M-x tags-search no longer change point  
 in buffers where no match is found.  In buffers where a match is  
 found, the original value of point is pushed on the marker ring.  
   
 ** Fortran mode has a new command `fortran-strip-sequence-nos' to  
 remove text past column 72.  The syntax class of `\' in Fortran is now  
 appropriate for C-style escape sequences in strings.  
   
 ** SGML mode's default `sgml-validate-command' is now `nsgmls'.  
   
 ** A new command `view-emacs-problems' (C-h P) displays the PROBLEMS file.  
   
 ** The Dabbrev package has a new user-option `dabbrev-ignored-regexps'  
 containing a list of regular expressions.  Buffers matching a regular  
 expression from that list, are not checked.  
   
 ** Emacs can now figure out modification times of remote files.  
 When you do C-x C-f /user@host:/path/file RET and edit the file,  
 and someone else modifies the file, you will be prompted to revert  
 the buffer, just like for the local files.  
   
 ** The buffer menu (C-x C-b) no longer lists the *Buffer List* buffer.  
   
 ** When invoked with a prefix argument, the command `list-abbrevs' now  
 displays local abbrevs, only.  
   
 ** Refill minor mode provides preliminary support for keeping  
 paragraphs filled as you modify them.  
   
 ** The variable `double-click-fuzz' specifies how much the mouse  
 may be moved between clicks that are recognized as a pair.  Its value  
 is measured in pixels.  
   
 ** The new global minor mode `auto-image-file-mode' allows image files  
 to be visited as images.  
   
 ** Two new user-options `grep-command' and `grep-find-command'  
 were added to compile.el.  
   
 ** Withdrawn packages  
   
 *** mldrag.el has been removed.  mouse.el provides the same  
 functionality with aliases for the mldrag functions.  
   
 *** eval-reg.el has been obsoleted by changes to edebug.el and removed.  
   
 *** ph.el has been obsoleted by EUDC and removed.  
74    
75    
76  * Incompatible Lisp changes  * Incompatible Lisp changes
77    
78  There are a few Lisp changes which are not backwards-compatible and  Deleted functions: make-coding-system, register-char-codings,
79  may require changes to existing code. Here is a list for reference.  coding-system-spec
 See the sections below for details.  
   
 ** Since `format' preserves text properties, the idiom  
 `(format "%s" foo)' no longer works to copy and remove properties.  
 Use `copy-sequence' to copy the string, then use `set-text-properties'  
 to remove the properties of the copy.  
   
 ** Since the `keymap' text property now has significance, some code  
 which uses both `local-map' and `keymap' properties (for portability)  
 may, for instance, give rise to duplicate menus when the keymaps from  
 these properties are active.  
   
 ** The change in the treatment of non-ASCII characters in search  
 ranges may affect some code.  
   
 ** A non-nil value for the LOCAL arg of add-hook makes the hook  
 buffer-local even if `make-local-hook' hasn't been called, which might  
 make a difference to some code.  
   
 ** The new treatment of the minibuffer prompt might affect code which  
 operates on the minibuffer.  
   
 ** The new character sets `eight-bit-control' and `eight-bit-graphic'  
 cause `no-conversion' and `emacs-mule-unix' coding systems to produce  
 different results when reading files with non-ASCII characters  
 (previously, both coding systems would produce the same results).  
 Specifically, `no-conversion' interprets each 8-bit byte as a separate  
 character.  This makes `no-conversion' inappropriate for reading  
 multibyte text, e.g. buffers written to disk in their internal MULE  
 encoding (auto-saving does that, for example).  If a Lisp program  
 reads such files with `no-conversion', each byte of the multibyte  
 sequence, including the MULE leading codes such as \201, is treated as  
 a separate character, which prevents them from being interpreted in  
 the buffer as multibyte characters.  
   
 Therefore, Lisp programs that read files which contain the internal  
 MULE encoding should use `emacs-mule-unix'.  `no-conversion' is only  
 appropriate for reading truly binary files.  
   
 ** Code that relies on the obsolete `before-change-function' and  
 `after-change-function' to detect buffer changes will now fail.  Use  
 `before-change-functions' and `after-change-functions' instead.  
   
 ** Code that uses `concat' with integer args now gets an error, as  
 long promised.  
   
 ** The function base64-decode-string now always returns a unibyte  
 string.  
   
 ** Not a Lisp incompatibility as such but, with the introduction of  
 extra private charsets, there is now only one slot free for a new  
 dimension-2 private charset.  User code which tries to add more than  
 one extra will fail unless you rebuild Emacs with some standard  
 charset(s) removed; that is probably inadvisable because it changes  
 the emacs-mule encoding.  Also, files stored in the emacs-mule  
 encoding using Emacs 20 with additional private charsets defined will  
 probably not be read correctly by Emacs 21.  
   
 ** The variable `directory-sep-char' is slated for removal.  
 Not really a change (yet), but a projected one that you should be  
 aware of: The variable `directory-sep-char' is deprecated, and should  
 not be used.  It was always ignored on GNU/Linux and Unix systems and  
 on MS-DOS, but the MS-Windows port tried to support it by adapting the  
 behavior of certain primitives to the value of this variable.  It  
 turned out that such support cannot be reliable, so it was decided to  
 remove this variable in the near future.  Lisp programs are well  
 advised not to set it to anything but '/', because any different value  
 will not have any effect when support for this variable is removed.  
   
   
 * Lisp changes made after edition 2.6 of the Emacs Lisp Manual,  
 (Display-related features are described in a page of their own below.)  
   
 ** Function assq-delete-all replaces function assoc-delete-all.  
   
 ** The new function animate-string, from lisp/play/animate.el  
 allows the animated display of strings.  
   
 ** The new function `interactive-form' can be used to obtain the  
 interactive form of a function.  
   
 ** The keyword :set-after in defcustom allows to specify dependencies  
 between custom options.  Example:  
   
   (defcustom default-input-method nil  
     "*Default input method for multilingual text (a string).  
   This is the input method activated automatically by the command  
   `toggle-input-method' (\\[toggle-input-method])."  
     :group 'mule  
     :type '(choice (const nil) string)  
     :set-after '(current-language-environment))  
   
 This specifies that default-input-method should be set after  
 current-language-environment even if default-input-method appears  
 first in a custom-set-variables statement.  
   
 ** The new hook `kbd-macro-termination-hook' is run at the end of  
 function execute-kbd-macro.  Functions on this hook are called with no  
 args.  The hook is run independent of how the macro was terminated  
 (signal or normal termination).  
   
 ** Functions `butlast' and `nbutlast' for removing trailing elements  
 from a list are now available without requiring the CL package.  
   
 ** The new user-option `even-window-heights' can be set to nil  
 to prevent `display-buffer' from evening out window heights.  
   
 ** The user-option `face-font-registry-alternatives' specifies  
 alternative font registry names to try when looking for a font.  
   
 ** Function `md5' calculates the MD5 "message digest"/"checksum".  
   
 ** Function `delete-frame' runs `delete-frame-hook' before actually  
 deleting the frame.  The hook is called with one arg, the frame  
 being deleted.  
   
 ** `add-hook' now makes the hook local if called with a non-nil LOCAL arg.  
   
 ** The treatment of non-ASCII characters in search ranges has changed.  
 If a range in a regular expression or the arg of  
 skip-chars-forward/backward starts with a unibyte character C and ends  
 with a multibyte character C2, the range is divided into two: one is  
 C..?\377, the other is C1..C2, where C1 is the first character of C2's  
 charset.  
   
 ** The new function `display-message-or-buffer' displays a message in  
 the echo area or pops up a buffer, depending on the length of the  
 message.  
   
 ** The new macro `with-auto-compression-mode' allows evaluating an  
 expression with auto-compression-mode enabled.  
   
 ** In image specifications, `:heuristic-mask' has been replaced  
 with the more general `:mask' property.  
   
 ** Image specifications accept more `:conversion's.  
   
 ** A `?' can be used in a symbol name without escaping it with a  
 backslash.  
   
 ** Reading from the mini-buffer now reads from standard input if Emacs  
 is running in batch mode.  For example,  
   
   (message "%s" (read t))  
   
 will read a Lisp expression from standard input and print the result  
 to standard output.  
   
 ** The argument of `down-list', `backward-up-list', `up-list',  
 `kill-sexp', `backward-kill-sexp' and `mark-sexp' is now optional.  
   
 ** If `display-buffer-reuse-frames' is set, function `display-buffer'  
 will raise frames displaying a buffer, instead of creating a new  
 frame or window.  
   
 ** Two new functions for removing elements from lists/sequences  
 were added  
   
 - Function: remove ELT SEQ  
   
 Return a copy of SEQ with all occurrences of ELT removed.  SEQ must be  
 a list, vector, or string.  The comparison is done with `equal'.  
   
 - Function: remq ELT LIST  
   
 Return a copy of LIST with all occurrences of ELT removed.  The  
 comparison is done with `eq'.  
   
 ** The function `delete' now also works with vectors and strings.  
   
 ** The meaning of the `:weakness WEAK' argument of make-hash-table  
 has been changed: WEAK can now have new values `key-or-value' and  
 `key-and-value', in addition the `nil', `key', `value', and `t'.  
   
 ** Function `aset' stores any multibyte character in any string  
 without signaling "Attempt to change char length of a string".  It may  
 convert a unibyte string to multibyte if necessary.  
   
 ** The value of the `help-echo' text property is called as a function  
 or evaluated, if it is not a string already, to obtain a help string.  
   
 ** Function `make-obsolete' now has an optional arg to say when the  
 function was declared obsolete.  
   
 ** Function `plist-member' is renamed from `widget-plist-member' (which is  
 retained as an alias).  
   
 ** Easy-menu's :filter now works as in XEmacs.  
 It takes the unconverted (i.e. XEmacs) form of the menu and the result  
 is automatically converted to Emacs' form.  
   
 ** The new function `window-list' has been defined  
   
 - Function: window-list &optional FRAME WINDOW MINIBUF  
   
 Return a list of windows on FRAME, starting with WINDOW.  FRAME nil or  
 omitted means use the selected frame.  WINDOW nil or omitted means use  
 the selected window.  MINIBUF t means include the minibuffer window,  
 even if it isn't active.  MINIBUF nil or omitted means include the  
 minibuffer window only if it's active.  MINIBUF neither nil nor t  
 means never include the minibuffer window.  
   
 ** There's a new function `get-window-with-predicate' defined as follows  
   
 - Function: get-window-with-predicate PREDICATE &optional MINIBUF ALL-FRAMES DEFAULT  
   
 Return a window satisfying PREDICATE.  
   
 This function cycles through all visible windows using `walk-windows',  
 calling PREDICATE on each one.  PREDICATE is called with a window as  
 argument.  The first window for which PREDICATE returns a non-nil  
 value is returned.  If no window satisfies PREDICATE, DEFAULT is  
 returned.  
   
 Optional second arg MINIBUF t means count the minibuffer window even  
 if not active.  MINIBUF nil or omitted means count the minibuffer iff  
 it is active.  MINIBUF neither t nor nil means not to count the  
 minibuffer even if it is active.  
   
 Several frames may share a single minibuffer; if the minibuffer  
 counts, all windows on all frames that share that minibuffer count  
 too.  Therefore, if you are using a separate minibuffer frame  
 and the minibuffer is active and MINIBUF says it counts,  
 `walk-windows' includes the windows in the frame from which you  
 entered the minibuffer, as well as the minibuffer window.  
   
 ALL-FRAMES is the optional third argument.  
 ALL-FRAMES nil or omitted means cycle within the frames as specified above.  
 ALL-FRAMES = `visible' means include windows on all visible frames.  
 ALL-FRAMES = 0 means include windows on all visible and iconified frames.  
 ALL-FRAMES = t means include windows on all frames including invisible frames.  
 If ALL-FRAMES is a frame, it means include windows on that frame.  
 Anything else means restrict to the selected frame.  
   
 ** The function `single-key-description' now encloses function key and  
 event names in angle brackets.  When called with a second optional  
 argument non-nil, angle brackets won't be printed.  
   
 ** If the variable `message-truncate-lines' is bound to t around a  
 call to `message', the echo area will not be resized to display that  
 message; it will be truncated instead, as it was done in 20.x.  
 Default value is nil.  
   
 ** The user option `line-number-display-limit' can now be set to nil,  
 meaning no limit.  
   
 ** The new user option `line-number-display-limit-width' controls  
 the maximum width of lines in a buffer for which Emacs displays line  
 numbers in the mode line.  The default is 200.  
   
 ** `select-safe-coding-system' now also checks the most preferred  
 coding-system if buffer-file-coding-system is `undecided' and  
 DEFAULT-CODING-SYSTEM is not specified,  
   
 ** The function `subr-arity' provides information about the argument  
 list of a primitive.  
   
 ** `where-is-internal' now also accepts a list of keymaps.  
   
 ** The text property `keymap' specifies a key map which overrides the  
 buffer's local map and the map specified by the `local-map' property.  
 This is probably what most current uses of `local-map' want, rather  
 than replacing the local map.  
   
 ** The obsolete variables `before-change-function' and  
 `after-change-function' are no longer acted upon and have been  
 removed.  Use `before-change-functions' and `after-change-functions'  
 instead.  
   
 ** The function `apropos-mode' runs the hook `apropos-mode-hook'.  
   
 ** `concat' no longer accepts individual integer arguments,  
 as promised long ago.  
   
 ** The new function `float-time' returns the current time as a float.  
   
 ** The new variable auto-coding-regexp-alist specifies coding systems  
 for reading specific files, analogous to auto-coding-alist, but  
 patterns are checked against file contents instead of file names.  
   
   
 * Lisp changes in Emacs 21.1 (see following page for display-related features)  
   
 ** The new package rx.el provides an alternative sexp notation for  
 regular expressions.  
   
 - Function: rx-to-string SEXP  
   
 Translate SEXP into a regular expression in string notation.  
   
 - Macro: rx SEXP  
   
 Translate SEXP into a regular expression in string notation.  
   
 The following are valid subforms of regular expressions in sexp  
 notation.  
   
 STRING  
      matches string STRING literally.  
   
 CHAR  
      matches character CHAR literally.  
   
 `not-newline'  
      matches any character except a newline.  
                         .  
 `anything'  
      matches any character  
   
 `(any SET)'  
      matches any character in SET.  SET may be a character or string.  
      Ranges of characters can be specified as `A-Z' in strings.  
   
 '(in SET)'  
      like `any'.  
   
 `(not (any SET))'  
      matches any character not in SET  
   
 `line-start'  
      matches the empty string, but only at the beginning of a line  
      in the text being matched  
   
 `line-end'  
      is similar to `line-start' but matches only at the end of a line  
   
 `string-start'  
      matches the empty string, but only at the beginning of the  
      string being matched against.  
   
 `string-end'  
      matches the empty string, but only at the end of the  
      string being matched against.  
   
 `buffer-start'  
      matches the empty string, but only at the beginning of the  
      buffer being matched against.  
   
 `buffer-end'  
      matches the empty string, but only at the end of the  
      buffer being matched against.  
   
 `point'  
      matches the empty string, but only at point.  
   
 `word-start'  
      matches the empty string, but only at the beginning or end of a  
      word.  
   
 `word-end'  
      matches the empty string, but only at the end of a word.  
   
 `word-boundary'  
      matches the empty string, but only at the beginning or end of a  
      word.  
   
 `(not word-boundary)'  
      matches the empty string, but not at the beginning or end of a  
      word.  
   
 `digit'  
      matches 0 through 9.  
   
 `control'  
      matches ASCII control characters.  
   
 `hex-digit'  
      matches 0 through 9, a through f and A through F.  
   
 `blank'  
      matches space and tab only.  
   
 `graphic'  
      matches graphic characters--everything except ASCII control chars,  
      space, and DEL.  
   
 `printing'  
      matches printing characters--everything except ASCII control chars  
      and DEL.  
   
 `alphanumeric'  
      matches letters and digits.  (But at present, for multibyte characters,  
      it matches anything that has word syntax.)  
   
 `letter'  
      matches letters.  (But at present, for multibyte characters,  
      it matches anything that has word syntax.)  
   
 `ascii'  
      matches ASCII (unibyte) characters.  
   
 `nonascii'  
      matches non-ASCII (multibyte) characters.  
   
 `lower'  
      matches anything lower-case.  
   
 `upper'  
      matches anything upper-case.  
   
 `punctuation'  
      matches punctuation.  (But at present, for multibyte characters,  
      it matches anything that has non-word syntax.)  
   
 `space'  
      matches anything that has whitespace syntax.  
   
 `word'  
      matches anything that has word syntax.  
   
 `(syntax SYNTAX)'  
      matches a character with syntax SYNTAX.  SYNTAX must be one  
      of the following symbols.  
   
      `whitespace'               (\\s- in string notation)  
      `punctuation'              (\\s.)  
      `word'                     (\\sw)  
      `symbol'                   (\\s_)  
      `open-parenthesis'         (\\s()  
      `close-parenthesis'        (\\s))  
      `expression-prefix'        (\\s')  
      `string-quote'             (\\s\")  
      `paired-delimiter'         (\\s$)  
      `escape'                   (\\s\\)  
      `character-quote'          (\\s/)  
      `comment-start'            (\\s<)  
      `comment-end'              (\\s>)  
   
 `(not (syntax SYNTAX))'  
      matches a character that has not syntax SYNTAX.  
   
 `(category CATEGORY)'  
      matches a character with category CATEGORY.  CATEGORY must be  
      either a character to use for C, or one of the following symbols.  
   
      `consonant'                        (\\c0 in string notation)  
      `base-vowel'                       (\\c1)  
      `upper-diacritical-mark'           (\\c2)  
      `lower-diacritical-mark'           (\\c3)  
      `tone-mark'                        (\\c4)  
      `symbol'                           (\\c5)  
      `digit'                            (\\c6)  
      `vowel-modifying-diacritical-mark' (\\c7)  
      `vowel-sign'                       (\\c8)  
      `semivowel-lower'                  (\\c9)  
      `not-at-end-of-line'               (\\c<)  
      `not-at-beginning-of-line'         (\\c>)  
      `alpha-numeric-two-byte'           (\\cA)  
      `chinse-two-byte'                  (\\cC)  
      `greek-two-byte'                   (\\cG)  
      `japanese-hiragana-two-byte'       (\\cH)  
      `indian-tow-byte'                  (\\cI)  
      `japanese-katakana-two-byte'       (\\cK)  
      `korean-hangul-two-byte'           (\\cN)  
      `cyrillic-two-byte'                (\\cY)  
      `ascii'                            (\\ca)  
      `arabic'                           (\\cb)  
      `chinese'                          (\\cc)  
      `ethiopic'                         (\\ce)  
      `greek'                            (\\cg)  
      `korean'                           (\\ch)  
      `indian'                           (\\ci)  
      `japanese'                         (\\cj)  
      `japanese-katakana'                (\\ck)  
      `latin'                            (\\cl)  
      `lao'                              (\\co)  
      `tibetan'                          (\\cq)  
      `japanese-roman'                   (\\cr)  
      `thai'                             (\\ct)  
      `vietnamese'                       (\\cv)  
      `hebrew'                           (\\cw)  
      `cyrillic'                         (\\cy)  
      `can-break'                        (\\c|)  
   
 `(not (category CATEGORY))'  
      matches a character that has not category CATEGORY.  
   
 `(and SEXP1 SEXP2 ...)'  
      matches what SEXP1 matches, followed by what SEXP2 matches, etc.  
   
 `(submatch SEXP1 SEXP2 ...)'  
      like `and', but makes the match accessible with `match-end',  
      `match-beginning', and `match-string'.  
   
 `(group SEXP1 SEXP2 ...)'  
      another name for `submatch'.  
   
 `(or SEXP1 SEXP2 ...)'  
      matches anything that matches SEXP1 or SEXP2, etc.  If all  
      args are strings, use `regexp-opt' to optimize the resulting  
      regular expression.  
   
 `(minimal-match SEXP)'  
      produce a non-greedy regexp for SEXP.  Normally, regexps matching  
      zero or more occurrances of something are \"greedy\" in that they  
      match as much as they can, as long as the overall regexp can  
      still match.  A non-greedy regexp matches as little as possible.  
   
 `(maximal-match SEXP)'  
      produce a greedy regexp for SEXP.   This is the default.  
   
 `(zero-or-more SEXP)'  
      matches zero or more occurrences of what SEXP matches.  
   
 `(0+ SEXP)'  
      like `zero-or-more'.  
   
 `(* SEXP)'  
      like `zero-or-more', but always produces a greedy regexp.  
   
 `(*? SEXP)'  
      like `zero-or-more', but always produces a non-greedy regexp.  
   
 `(one-or-more SEXP)'  
      matches one or more occurrences of A.  
   
 `(1+ SEXP)'  
      like `one-or-more'.  
   
 `(+ SEXP)'  
      like `one-or-more', but always produces a greedy regexp.  
   
 `(+? SEXP)'  
      like `one-or-more', but always produces a non-greedy regexp.  
   
 `(zero-or-one SEXP)'  
      matches zero or one occurrences of A.  
   
 `(optional SEXP)'  
      like `zero-or-one'.  
   
 `(? SEXP)'  
      like `zero-or-one', but always produces a greedy regexp.  
   
 `(?? SEXP)'  
      like `zero-or-one', but always produces a non-greedy regexp.  
   
 `(repeat N SEXP)'  
      matches N occurrences of what SEXP matches.  
   
 `(repeat N M SEXP)'  
      matches N to M occurrences of what SEXP matches.  
   
 `(eval FORM)'  
       evaluate FORM and insert result.   If result is a string,  
       `regexp-quote' it.  
   
 `(regexp REGEXP)'  
       include REGEXP in string notation in the result.  
   
 *** The features `md5' and `overlay' are now provided by default.  
   
 *** The special form `save-restriction' now works correctly even if the  
 buffer is widened inside the save-restriction and changes made outside  
 the original restriction.  Previously, doing this would cause the saved  
 restriction to be restored incorrectly.  
   
 *** The functions `find-charset-region' and `find-charset-string' include  
 `eight-bit-control' and/or `eight-bit-graphic' in the returned list  
 when they find 8-bit characters.  Previously, they included `ascii' in a  
 multibyte buffer and `unknown' in a unibyte buffer.  
   
 *** The functions `set-buffer-multibyte', `string-as-multibyte' and  
 `string-as-unibyte' change the byte sequence of a buffer or a string  
 if it contains a character from the `eight-bit-control' character set.  
   
 *** The handling of multibyte sequences in a multibyte buffer is  
 changed.  Previously, a byte sequence matching the pattern  
 [\200-\237][\240-\377]+ was interpreted as a single character  
 regardless of the length of the trailing bytes [\240-\377]+.  Thus, if  
 the sequence was longer than what the leading byte indicated, the  
 extra trailing bytes were ignored by Lisp functions.  Now such extra  
 bytes are independent 8-bit characters belonging to the charset  
 eight-bit-graphic.  
   
 ** Fontsets are now implemented using char-tables.  
   
 A fontset can now be specified for each independent character, for  
 a group of characters or for a character set rather than just for a  
 character set as previously.  
   
 *** The arguments of the function `set-fontset-font' are changed.  
 They are NAME, CHARACTER, FONTNAME, and optional FRAME.  The function  
 modifies fontset NAME to use FONTNAME for CHARACTER.  
   
 CHARACTER may be a cons (FROM . TO), where FROM and TO are non-generic  
 characters.  In that case FONTNAME is used for all characters in the  
 range FROM and TO (inclusive).  CHARACTER may be a charset.  In that  
 case FONTNAME is used for all character in the charset.  
   
 FONTNAME may be a cons (FAMILY . REGISTRY), where FAMILY is the family  
 name of a font and REGISTRY is a registry name of a font.  
   
 *** Variable x-charset-registry has been deleted.  The default charset  
 registries of character sets are set in the default fontset  
 "fontset-default".  
   
 *** The function `create-fontset-from-fontset-spec' ignores the second  
 argument STYLE-VARIANT.  It never creates style-variant fontsets.  
   
 ** The method of composing characters is changed.  Now character  
 composition is done by a special text property `composition' in  
 buffers and strings.  
   
 *** Charset composition is deleted.  Emacs never creates a `composite  
 character' which is an independent character with a unique character  
 code.  Thus the following functions handling `composite characters'  
 have been deleted: composite-char-component,  
 composite-char-component-count, composite-char-composition-rule,  
 composite-char-composition-rule and decompose-composite-char delete.  
 The variables leading-code-composition and min-composite-char have  
 also been deleted.  
   
 *** Three more glyph reference points are added.  They can be used to  
 specify a composition rule.  See the documentation of the variable  
 `reference-point-alist' for more detail.  
   
 *** The function `compose-region' takes new arguments COMPONENTS and  
 MODIFICATION-FUNC.  With COMPONENTS, you can specify not only a  
 composition rule but also characters to be composed.  Such characters  
 may differ between buffer and string text.  
   
 *** The function `compose-string' takes new arguments START, END,  
 COMPONENTS, and MODIFICATION-FUNC.  
   
 *** The function `compose-string' puts text property `composition'  
 directly on the argument STRING instead of returning a new string.  
 Likewise, the function `decompose-string' just removes text property  
 `composition' from STRING.  
   
 *** The new function `find-composition' returns information about  
 a composition at a specified position in a buffer or a string.  
   
 *** The function `decompose-composite-char' is now labeled as  
 obsolete.  
   
 ** The new coding system `mac-roman' is primarily intended for use on  
 the Macintosh but may be used generally for Macintosh-encoded text.  
   
 ** The new character sets `mule-unicode-0100-24ff',  
 `mule-unicode-2500-33ff', and `mule-unicode-e000-ffff' have been  
 introduced for Unicode characters in the range U+0100..U+24FF,  
 U+2500..U+33FF, U+E000..U+FFFF respectively.  
   
 Note that the character sets are not yet unified in Emacs, so  
 characters which belong to charsets such as Latin-2, Greek, Hebrew,  
 etc. and the same characters in the `mule-unicode-*' charsets are  
 different characters, as far as Emacs is concerned.  For example, text  
 which includes Unicode characters from the Latin-2 locale cannot be  
 encoded by Emacs with ISO 8859-2 coding system.  
   
 ** The new coding system `mule-utf-8' has been added.  
 It provides limited support for decoding/encoding UTF-8 text.  For  
 details, please see the documentation string of this coding system.  
   
 ** The new character sets `japanese-jisx0213-1' and  
 `japanese-jisx0213-2' have been introduced for the new Japanese  
 standard JIS X 0213 Plane 1 and Plane 2.  
   
 ** The new character sets `latin-iso8859-14' and `latin-iso8859-15'  
 have been introduced.  
   
 ** The new character sets `eight-bit-control' and `eight-bit-graphic'  
 have been introduced for 8-bit characters in the ranges 0x80..0x9F and  
 0xA0..0xFF respectively.  Note that the multibyte representation of  
 eight-bit-control is never exposed; this leads to an exception in the  
 emacs-mule coding system, which encodes everything else to the  
 buffer/string internal representation.  Note that to search for  
 eight-bit-graphic characters in a multibyte buffer, the search string  
 must be multibyte, otherwise such characters will be converted to  
 their multibyte equivalent.  
   
 ** If the APPEND argument of `write-region' is an integer, it seeks to  
 that offset in the file before writing.  
   
 ** The function `add-minor-mode' has been added for convenience and  
 compatibility with XEmacs (and is used internally by define-minor-mode).  
   
 ** The function `shell-command' now sets the default directory of the  
 `*Shell Command Output*' buffer to the default directory of the buffer  
 from which the command was issued.  
   
 ** The functions `query-replace', `query-replace-regexp',  
 `query-replace-regexp-eval' `map-query-replace-regexp',  
 `replace-string', `replace-regexp', and `perform-replace' take two  
 additional optional arguments START and END that specify the region to  
 operate on.  
   
 ** The new function `count-screen-lines' is a more flexible alternative  
 to `window-buffer-height'.  
   
 - Function: count-screen-lines &optional BEG END COUNT-FINAL-NEWLINE WINDOW  
   
 Return the number of screen lines in the region between BEG and END.  
 The number of screen lines may be different from the number of actual  
 lines, due to line breaking, display table, etc.  
   
 Optional arguments BEG and END default to `point-min' and `point-max'  
 respectively.  
   
 If region ends with a newline, ignore it unless optional third argument  
 COUNT-FINAL-NEWLINE is non-nil.  
   
 The optional fourth argument WINDOW specifies the window used for  
 obtaining parameters such as width, horizontal scrolling, and so  
 on. The default is to use the selected window's parameters.  
   
 Like `vertical-motion', `count-screen-lines' always uses the current  
 buffer, regardless of which buffer is displayed in WINDOW. This makes  
 possible to use `count-screen-lines' in any buffer, whether or not it  
 is currently displayed in some window.  
   
 ** The new function `mapc' is like `mapcar' but doesn't collect the  
 argument function's results.  
   
 ** The functions base64-decode-region and base64-decode-string now  
 signal an error instead of returning nil if decoding fails.  Also,  
 `base64-decode-string' now always returns a unibyte string (in Emacs  
 20, it returned a multibyte string when the result was a valid multibyte  
 sequence).  
   
 ** The function sendmail-user-agent-compose now recognizes a `body'  
 header in the list of headers passed to it.  
   
 ** The new function member-ignore-case works like `member', but  
 ignores differences in case and text representation.  
   
 ** The buffer-local variable cursor-type can be used to specify the  
 cursor to use in windows displaying a buffer.  Values are interpreted  
 as follows:  
   
   t             use the cursor specified for the frame (default)  
   nil           don't display a cursor  
   `bar'         display a bar cursor with default width  
   (bar . WIDTH) display a bar cursor with width WIDTH  
   others        display a box cursor.  
   
 ** The variable open-paren-in-column-0-is-defun-start controls whether  
 an open parenthesis in column 0 is considered to be the start of a  
 defun.  If set, the default, it is considered a defun start.  If not  
 set, an open parenthesis in column 0 has no special meaning.  
   
 ** The new function `string-to-syntax' can be used to translate syntax  
 specifications in string form as accepted by `modify-syntax-entry' to  
 the cons-cell form that is used for the values of the `syntax-table'  
 text property, and in `font-lock-syntactic-keywords'.  
   
 Example:  
   
   (string-to-syntax "()")  
     => (4 . 41)  
   
 ** Emacs' reader supports CL read syntax for integers in bases  
 other than 10.  
   
 *** `#BINTEGER' or `#bINTEGER' reads INTEGER in binary (radix 2).  
 INTEGER optionally contains a sign.  
   
   #b1111  
     => 15  
   #b-1111  
     => -15  
   
 *** `#OINTEGER' or `#oINTEGER' reads INTEGER in octal (radix 8).  
   
   #o666  
     => 438  
   
 *** `#XINTEGER' or `#xINTEGER' reads INTEGER in hexadecimal (radix 16).  
   
   #xbeef  
     => 48815  
   
 *** `#RADIXrINTEGER' reads INTEGER in radix RADIX, 2 <= RADIX <= 36.  
   
   #2R-111  
     => -7  
   #25rah  
     => 267  
   
 ** The function `documentation-property' now evaluates the value of  
 the given property to obtain a string if it doesn't refer to etc/DOC  
 and isn't a string.  
   
 ** If called for a symbol, the function `documentation' now looks for  
 a `function-documentation' property of that symbol.  If it has a non-nil  
 value, the documentation is taken from that value.  If the value is  
 not a string, it is evaluated to obtain a string.  
   
 ** The last argument of `define-key-after' defaults to t for convenience.  
   
 ** The new function `replace-regexp-in-string' replaces all matches  
 for a regexp in a string.  
   
 ** `mouse-position' now runs the abnormal hook  
 `mouse-position-function'.  
   
 ** The function string-to-number now returns a float for numbers  
 that don't fit into a Lisp integer.  
   
 ** The variable keyword-symbols-constants-flag has been removed.  
 Keywords are now always considered constants.  
   
 ** The new function `delete-and-extract-region' deletes text and  
 returns it.  
   
 ** The function `clear-this-command-keys' now also clears the vector  
 returned by function `recent-keys'.  
   
 ** Variables `beginning-of-defun-function' and `end-of-defun-function'  
 can be used to define handlers for the functions that find defuns.  
 Major modes can define these locally instead of rebinding C-M-a  
 etc. if the normal conventions for defuns are not appropriate for the  
 mode.  
   
 ** easy-mmode-define-minor-mode now takes an additional BODY argument  
 and is renamed `define-minor-mode'.  
   
 ** If an abbrev has a hook function which is a symbol, and that symbol  
 has a non-nil `no-self-insert' property, the return value of the hook  
 function specifies whether an expansion has been done or not.  If it  
 returns nil, abbrev-expand also returns nil, meaning "no expansion has  
 been performed."  
   
 When abbrev expansion is done by typing a self-inserting character,  
 and the abbrev has a hook with the `no-self-insert' property, and the  
 hook function returns non-nil meaning expansion has been done,  
 then the self-inserting character is not inserted.  
   
 ** The function `intern-soft' now accepts a symbol as first argument.  
 In this case, that exact symbol is looked up in the specified obarray,  
 and the function's value is nil if it is not found.  
   
 ** The new macro `with-syntax-table' can be used to evaluate forms  
 with the syntax table of the current buffer temporarily set to a  
 specified table.  
   
   (with-syntax-table TABLE &rest BODY)  
   
 Evaluate BODY with syntax table of current buffer set to a copy of  
 TABLE.  The current syntax table is saved, BODY is evaluated, and the  
 saved table is restored, even in case of an abnormal exit.  Value is  
 what BODY returns.  
   
 ** Regular expressions now support intervals \{n,m\} as well as  
 Perl's shy-groups \(?:...\) and non-greedy *? +? and ?? operators.  
 Also back-references like \2 are now considered as an error if the  
 corresponding subgroup does not exist (or is not closed yet).  
 Previously it would have been silently turned into `2' (ignoring the `\').  
   
 ** The optional argument BUFFER of function file-local-copy has been  
 removed since it wasn't used by anything.  
   
 ** The file name argument of function `file-locked-p' is now required  
 instead of being optional.  
   
 ** The new built-in error `text-read-only' is signaled when trying to  
 modify read-only text.  
   
 ** New functions and variables for locales.  
   
 The new variable `locale-coding-system' specifies how to encode and  
 decode strings passed to low-level message functions like strerror and  
 time functions like strftime.  The new variables  
 `system-messages-locale' and `system-time-locale' give the system  
 locales to be used when invoking these two types of functions.  
   
 The new function `set-locale-environment' sets the language  
 environment, preferred coding system, and locale coding system from  
 the system locale as specified by the LC_ALL, LC_CTYPE, and LANG  
 environment variables.  Normally, it is invoked during startup and need  
 not be invoked thereafter.  It uses the new variables  
 `locale-language-names', `locale-charset-language-names', and  
 `locale-preferred-coding-systems' to make its decisions.  
   
 ** syntax tables now understand nested comments.  
 To declare a comment syntax as allowing nesting, just add an `n'  
 modifier to either of the characters of the comment end and the comment  
 start sequences.  
   
 ** The function `pixmap-spec-p' has been renamed `bitmap-spec-p'  
 because `bitmap' is more in line with the usual X terminology.  
   
 ** New function `propertize'  
   
 The new function `propertize' can be used to conveniently construct  
 strings with text properties.  
   
 - Function: propertize STRING &rest PROPERTIES  
   
 Value is a copy of STRING with text properties assigned as specified  
 by PROPERTIES.  PROPERTIES is a sequence of pairs PROPERTY VALUE, with  
 PROPERTY being the name of a text property and VALUE being the  
 specified value of that property.  Example:  
   
   (propertize "foo" 'face 'bold 'read-only t)  
   
 ** push and pop macros.  
   
 Simple versions of the push and pop macros of Common Lisp  
 are now defined in Emacs Lisp.  These macros allow only symbols  
 as the place that holds the list to be changed.  
   
 (push NEWELT LISTNAME)  add NEWELT to the front of LISTNAME's value.  
 (pop LISTNAME)          return first elt of LISTNAME, and remove it  
                         (thus altering the value of LISTNAME).  
   
 ** New dolist and dotimes macros.  
   
 Simple versions of the dolist and dotimes macros of Common Lisp  
 are now defined in Emacs Lisp.  
   
 (dolist (VAR LIST [RESULT]) BODY...)  
       Execute body once for each element of LIST,  
       using the variable VAR to hold the current element.  
       Then return the value of RESULT, or nil if RESULT is omitted.  
   
 (dotimes (VAR COUNT [RESULT]) BODY...)  
       Execute BODY with VAR bound to successive integers running from 0,  
       inclusive, to COUNT, exclusive.  
       Then return the value of RESULT, or nil if RESULT is omitted.  
   
 ** Regular expressions now support Posix character classes such as  
 [:alpha:], [:space:] and so on.  These must be used within a character  
 class--for instance, [-[:digit:].+] matches digits or a period  
 or a sign.  
   
 [:digit:]  matches 0 through 9  
 [:cntrl:]  matches ASCII control characters  
 [:xdigit:]  matches 0 through 9, a through f and A through F.  
 [:blank:]  matches space and tab only  
 [:graph:]  matches graphic characters--everything except ASCII control chars,  
            space, and DEL.  
 [:print:]  matches printing characters--everything except ASCII control chars  
            and DEL.  
 [:alnum:]  matches letters and digits.  
            (But at present, for multibyte characters,  
             it matches anything that has word syntax.)  
 [:alpha:]  matches letters.  
            (But at present, for multibyte characters,  
             it matches anything that has word syntax.)  
 [:ascii:]  matches ASCII (unibyte) characters.  
 [:nonascii:]  matches non-ASCII (multibyte) characters.  
 [:lower:]  matches anything lower-case.  
 [:punct:]  matches punctuation.  
            (But at present, for multibyte characters,  
             it matches anything that has non-word syntax.)  
 [:space:]  matches anything that has whitespace syntax.  
 [:upper:]  matches anything upper-case.  
 [:word:]   matches anything that has word syntax.  
   
 ** Emacs now has built-in hash tables.  
   
 The following functions are defined for hash tables:  
   
 - Function: make-hash-table ARGS  
   
 The argument list ARGS consists of keyword/argument pairs.  All arguments  
 are optional.  The following arguments are defined:  
   
 :test TEST  
   
 TEST must be a symbol specifying how to compare keys.  Default is `eql'.  
 Predefined are `eq', `eql' and `equal'.  If TEST is not predefined,  
 it must have been defined with `define-hash-table-test'.  
   
 :size SIZE  
   
 SIZE must be an integer > 0 giving a hint to the implementation how  
 many elements will be put in the hash table.  Default size is 65.  
   
 :rehash-size REHASH-SIZE  
   
 REHASH-SIZE specifies by how much to grow a hash table once it becomes  
 full.  If REHASH-SIZE is an integer, add that to the hash table's old  
 size to get the new size.  Otherwise, REHASH-SIZE must be a float >  
 1.0, and the new size is computed by multiplying REHASH-SIZE with the  
 old size.  Default rehash size is 1.5.  
   
 :rehash-threshold THRESHOLD  
   
 THRESHOLD must be a float > 0 and <= 1.0 specifying when to resize the  
 hash table.  It is resized when the ratio of (number of entries) /  
 (size of hash table) is >= THRESHOLD.  Default threshold is 0.8.  
   
 :weakness WEAK  
   
 WEAK must be either nil, one of the symbols `key, `value',  
 `key-or-value', `key-and-value', or t, meaning the same as  
 `key-and-value'.  Entries are removed from weak tables during garbage  
 collection if their key and/or value are not referenced elsewhere  
 outside of the hash table.  Default are non-weak hash tables.  
   
 - Function: makehash &optional TEST  
   
 Similar to make-hash-table, but only TEST can be specified.  
   
 - Function: hash-table-p TABLE  
   
 Returns non-nil if TABLE is a hash table object.  
   
 - Function: copy-hash-table TABLE  
   
 Returns a copy of TABLE.  Only the table itself is copied, keys and  
 values are shared.  
   
 - Function: hash-table-count TABLE  
   
 Returns the number of entries in TABLE.  
   
 - Function: hash-table-rehash-size TABLE  
   
 Returns the rehash size of TABLE.  
   
 - Function: hash-table-rehash-threshold TABLE  
   
 Returns the rehash threshold of TABLE.  
   
 - Function: hash-table-rehash-size TABLE  
   
 Returns the size of TABLE.  
   
 - Function: hash-table-test TABLE  
   
 Returns the test TABLE uses to compare keys.  
   
 - Function: hash-table-weakness TABLE  
   
 Returns the weakness specified for TABLE.  
   
 - Function: clrhash TABLE  
   
 Clear TABLE.  
   
 - Function: gethash KEY TABLE &optional DEFAULT  
   
 Look up KEY in TABLE and return its associated VALUE or DEFAULT if  
 not found.  
   
 - Function: puthash KEY VALUE TABLE  
   
 Associate KEY with VALUE in TABLE.  If KEY is already associated with  
 another value, replace the old value with VALUE.  
   
 - Function: remhash KEY TABLE  
   
 Remove KEY from TABLE if it is there.  
   
 - Function: maphash FUNCTION TABLE  
   
 Call FUNCTION for all elements in TABLE.  FUNCTION must take two  
 arguments KEY and VALUE.  
   
 - Function: sxhash OBJ  
   
 Return a hash code for Lisp object OBJ.  
   
 - Function: define-hash-table-test NAME TEST-FN HASH-FN  
   
 Define a new hash table test named NAME.  If NAME is specified as  
 a test in `make-hash-table', the table created will use TEST-FN for  
 comparing keys, and HASH-FN to compute hash codes for keys.  Test  
 and hash function are stored as symbol property `hash-table-test'  
 of NAME with a value of (TEST-FN HASH-FN).  
   
 TEST-FN must take two arguments and return non-nil if they are the same.  
   
 HASH-FN must take one argument and return an integer that is the hash  
 code of the argument.  The function should use the whole range of  
 integer values for hash code computation, including negative integers.  
   
 Example: The following creates a hash table whose keys are supposed to  
 be strings that are compared case-insensitively.  
   
   (defun case-fold-string= (a b)  
     (compare-strings a nil nil b nil nil t))  
   
   (defun case-fold-string-hash (a)  
     (sxhash (upcase a)))  
   
   (define-hash-table-test 'case-fold 'case-fold-string=  
                           'case-fold-string-hash))  
   
   (make-hash-table :test 'case-fold)  
   
 ** The Lisp reader handles circular structure.  
   
 It now works to use the #N= and #N# constructs to represent  
 circular structures.  For example, #1=(a . #1#) represents  
 a cons cell which is its own cdr.  
   
 ** The Lisp printer handles circular structure.  
   
 If you bind print-circle to a non-nil value, the Lisp printer outputs  
 #N= and #N# constructs to represent circular and shared structure.  
   
 ** If the second argument to `move-to-column' is anything but nil or  
 t, that means replace a tab with spaces if necessary to reach the  
 specified column, but do not add spaces at the end of the line if it  
 is too short to reach that column.  
   
 ** perform-replace has a new feature:  the REPLACEMENTS argument may  
 now be a cons cell (FUNCTION . DATA).  This means to call FUNCTION  
 after each match to get the replacement text.  FUNCTION is called with  
 two arguments: DATA, and the number of replacements already made.  
   
 If the FROM-STRING contains any upper-case letters,  
 perform-replace also turns off `case-fold-search' temporarily  
 and inserts the replacement text without altering case in it.  
   
 ** The function buffer-size now accepts an optional argument  
 to specify which buffer to return the size of.  
   
 ** The calendar motion commands now run the normal hook  
 calendar-move-hook after moving point.  
   
 ** The new variable small-temporary-file-directory specifies a  
 directory to use for creating temporary files that are likely to be  
 small.  (Certain Emacs features use this directory.)  If  
 small-temporary-file-directory is nil, they use  
 temporary-file-directory instead.  
   
 ** The variable `inhibit-modification-hooks', if non-nil, inhibits all  
 the hooks that track changes in the buffer.  This affects  
 `before-change-functions' and `after-change-functions', as well as  
 hooks attached to text properties and overlay properties.  
   
 ** assq-delete-all is a new function that deletes all the  
 elements of an alist which have a car `eq' to a particular value.  
   
 ** make-temp-file provides a more reliable way to create a temporary file.  
   
 make-temp-file is used like make-temp-name, except that it actually  
 creates the file before it returns.  This prevents a timing error,  
 ensuring that no other job can use the same name for a temporary file.  
   
 ** New exclusive-open feature in `write-region'  
   
 The optional seventh arg is now called MUSTBENEW.  If non-nil, it insists  
 on a check for an existing file with the same name.  If MUSTBENEW  
 is `excl', that means to get an error if the file already exists;  
 never overwrite. If MUSTBENEW is neither nil nor `excl', that means  
 ask for confirmation before overwriting, but do go ahead and  
 overwrite the file if the user gives confirmation.  
   
 If the MUSTBENEW argument in `write-region' is `excl',  
 that means to use a special feature in the `open' system call  
 to get an error if the file exists at that time.  
 The error reported is `file-already-exists'.  
   
 ** Function `format' now handles text properties.  
   
 Text properties of the format string are applied to the result string.  
 If the result string is longer than the format string, text properties  
 ending at the end of the format string are extended to the end of the  
 result string.  
   
 Text properties from string arguments are applied to the result  
 string where arguments appear in the result string.  
   
 Example:  
   
   (let ((s1 "hello, %s")  
         (s2 "world"))  
      (put-text-property 0 (length s1) 'face 'bold s1)  
      (put-text-property 0 (length s2) 'face 'italic s2)  
      (format s1 s2))  
   
 results in a bold-face string with an italic `world' at the end.  
   
 ** Messages can now be displayed with text properties.  
   
 Text properties are handled as described above for function `format'.  
 The following example displays a bold-face message with an italic  
 argument in it.  
   
   (let ((msg "hello, %s!")  
         (arg "world"))  
      (put-text-property 0 (length msg) 'face 'bold msg)  
      (put-text-property 0 (length arg) 'face 'italic arg)  
      (message msg arg))  
   
 ** Sound support  
   
 Emacs supports playing sound files on GNU/Linux and the free BSDs  
 (Voxware driver and native BSD driver, aka as Luigi's driver).  
   
 Currently supported file formats are RIFF-WAVE (*.wav) and Sun Audio  
 (*.au).  You must configure Emacs with the option `--with-sound=yes'  
 to enable sound support.  
   
 Sound files can be played by calling (play-sound SOUND).  SOUND is a  
 list of the form `(sound PROPERTY...)'.  The function is only defined  
 when sound support is present for the system on which Emacs runs.  The  
 functions runs `play-sound-functions' with one argument which is the  
 sound to play, before playing the sound.  
   
 The following sound properties are supported:  
   
 - `:file FILE'  
   
 FILE is a file name.  If FILE isn't an absolute name, it will be  
 searched relative to `data-directory'.  
   
 - `:data DATA'  
   
 DATA is a string containing sound data.  Either :file or :data  
 may be present, but not both.  
   
 - `:volume VOLUME'  
   
 VOLUME must be an integer in the range 0..100 or a float in the range  
 0..1.  This property is optional.  
   
 - `:device DEVICE'  
   
 DEVICE is a string specifying the system device on which to play the  
 sound.  The default device is system-dependent.  
   
 Other properties are ignored.  
   
 An alternative interface is called as  
 (play-sound-file FILE &optional VOLUME DEVICE).  
   
 ** `multimedia' is a new Finder keyword and Custom group.  
   
 ** keywordp is a new predicate to test efficiently for an object being  
 a keyword symbol.  
   
 ** Changes to garbage collection  
   
 *** The function garbage-collect now additionally returns the number  
 of live and free strings.  
   
 *** There is a new variable `strings-consed' holding the number of  
 strings that have been consed so far.  
   
   
 * Lisp-level Display features added after release 2.6 of the Emacs  
 Lisp Manual  
   
 ** The user-option `resize-mini-windows' controls how Emacs resizes  
 mini-windows.  
   
 ** The function `pos-visible-in-window-p' now has a third optional  
 argument, PARTIALLY.  If a character is only partially visible, nil is  
 returned, unless PARTIALLY is non-nil.  
   
 ** On window systems, `glyph-table' is no longer used.  
   
 ** Help strings in menu items are now used to provide `help-echo' text.  
   
 ** The function `image-size' can be used to determine the size of an  
 image.  
   
 - Function: image-size SPEC &optional PIXELS FRAME  
   
 Return the size of an image as a pair (WIDTH . HEIGHT).  
   
 SPEC is an image specification.  PIXELS non-nil means return sizes  
 measured in pixels, otherwise return sizes measured in canonical  
 character units (fractions of the width/height of the frame's default  
 font).  FRAME is the frame on which the image will be displayed.  
 FRAME nil or omitted means use the selected frame.  
   
 ** The function `image-mask-p' can be used to determine if an image  
 has a mask bitmap.  
   
 - Function: image-mask-p SPEC &optional FRAME  
   
 Return t if image SPEC has a mask bitmap.  
 FRAME is the frame on which the image will be displayed.  FRAME nil  
 or omitted means use the selected frame.  
   
 ** The function `find-image' can be used to find a usable image  
 satisfying one of a list of specifications.  
   
 ** The STRING argument of `put-image' and `insert-image' is now  
 optional.  
   
 ** Image specifications may contain the property `:ascent center' (see  
 below).  
   
   
 * New Lisp-level Display features in Emacs 21.1  
   
 ** The function tty-suppress-bold-inverse-default-colors can be used  
 to make Emacs avoid displaying text with bold black foreground on TTYs.  
   
 Some terminals, notably PC consoles, emulate bold text by displaying  
 text in brighter colors.  On such a console, a bold black foreground  
 is displayed in a gray color.  If this turns out to be hard to read on  
 your monitor---the problem occurred with the mode line on  
 laptops---you can instruct Emacs to ignore the text's boldness, and to  
 just display it black instead.  
   
 This situation can't be detected automatically.  You will have to put  
 a line like  
   
   (tty-suppress-bold-inverse-default-colors t)  
   
 in your `.emacs'.  
   
 ** New face implementation.  
   
 Emacs faces have been reimplemented from scratch.  They don't use XLFD  
 font names anymore and face merging now works as expected.  
   
 *** New faces.  
   
 Each face can specify the following display attributes:  
   
    1. Font family or fontset alias name.  
   
    2. Relative proportionate width, aka character set width or set  
    width (swidth), e.g. `semi-compressed'.  
   
    3. Font height in 1/10pt  
   
    4. Font weight, e.g. `bold'.  
   
    5. Font slant, e.g. `italic'.  
   
    6. Foreground color.  
   
    7. Background color.  
   
    8. Whether or not characters should be underlined, and in what color.  
   
    9. Whether or not characters should be displayed in inverse video.  
   
    10. A background stipple, a bitmap.  
   
    11. Whether or not characters should be overlined, and in what color.  
   
    12. Whether or not characters should be strike-through, and in what  
    color.  
   
    13. Whether or not a box should be drawn around characters, its  
    color, the width of the box lines, and 3D appearance.  
   
 Faces are frame-local by nature because Emacs allows to define the  
 same named face (face names are symbols) differently for different  
 frames.  Each frame has an alist of face definitions for all named  
 faces.  The value of a named face in such an alist is a Lisp vector  
 with the symbol `face' in slot 0, and a slot for each of the face  
 attributes mentioned above.  
   
 There is also a global face alist `face-new-frame-defaults'.  Face  
 definitions from this list are used to initialize faces of newly  
 created frames.  
   
 A face doesn't have to specify all attributes.  Those not specified  
 have a nil value.  Faces specifying all attributes are called  
 `fully-specified'.  
   
 *** Face merging.  
   
 The display style of a given character in the text is determined by  
 combining several faces.  This process is called `face merging'.  Any  
 aspect of the display style that isn't specified by overlays or text  
 properties is taken from the `default' face.  Since it is made sure  
 that the default face is always fully-specified, face merging always  
 results in a fully-specified face.  
   
 *** Face realization.  
   
 After all face attributes for a character have been determined by  
 merging faces of that character, that face is `realized'.  The  
 realization process maps face attributes to what is physically  
 available on the system where Emacs runs.  The result is a `realized  
 face' in form of an internal structure which is stored in the face  
 cache of the frame on which it was realized.  
   
 Face realization is done in the context of the charset of the  
 character to display because different fonts and encodings are used  
 for different charsets.  In other words, for characters of different  
 charsets, different realized faces are needed to display them.  
   
 Except for composite characters, faces are always realized for a  
 specific character set and contain a specific font, even if the face  
 being realized specifies a fontset.  The reason is that the result of  
 the new font selection stage is better than what can be done with  
 statically defined font name patterns in fontsets.  
   
 In unibyte text, Emacs' charsets aren't applicable; function  
 `char-charset' reports ASCII for all characters, including those >  
 0x7f.  The X registry and encoding of fonts to use is determined from  
 the variable `face-default-registry' in this case.  The variable is  
 initialized at Emacs startup time from the font the user specified for  
 Emacs.  
   
 Currently all unibyte text, i.e. all buffers with  
 `enable-multibyte-characters' nil are displayed with fonts of the same  
 registry and encoding `face-default-registry'.  This is consistent  
 with the fact that languages can also be set globally, only.  
   
 **** Clearing face caches.  
   
 The Lisp function `clear-face-cache' can be called to clear face caches  
 on all frames.  If called with a non-nil argument, it will also unload  
 unused fonts.  
   
 *** Font selection.  
   
 Font selection tries to find the best available matching font for a  
 given (charset, face) combination.  This is done slightly differently  
 for faces specifying a fontset, or a font family name.  
   
 If the face specifies a fontset name, that fontset determines a  
 pattern for fonts of the given charset.  If the face specifies a font  
 family, a font pattern is constructed.  Charset symbols have a  
 property `x-charset-registry' for that purpose that maps a charset to  
 an XLFD registry and encoding in the font pattern constructed.  
   
 Available fonts on the system on which Emacs runs are then matched  
 against the font pattern.  The result of font selection is the best  
 match for the given face attributes in this font list.  
   
 Font selection can be influenced by the user.  
   
 The user can specify the relative importance he gives the face  
 attributes width, height, weight, and slant by setting  
 face-font-selection-order (faces.el) to a list of face attribute  
 names.  The default is (:width :height :weight :slant), and means  
 that font selection first tries to find a good match for the font  
 width specified by a face, then---within fonts with that width---tries  
 to find a best match for the specified font height, etc.  
   
 Setting `face-font-family-alternatives' allows the user to specify  
 alternative font families to try if a family specified by a face  
 doesn't exist.  
   
 Setting `face-font-registry-alternatives' allows the user to specify  
 all alternative font registry names to try for a face specifying a  
 registry.  
   
 Please note that the interpretations of the above two variables are  
 slightly different.  
   
 Setting face-ignored-fonts allows the user to ignore specific fonts.  
   
   
 **** Scalable fonts  
   
 Emacs can make use of scalable fonts but doesn't do so by default,  
 since the use of too many or too big scalable fonts may crash XFree86  
 servers.  
   
 To enable scalable font use, set the variable  
 `scalable-fonts-allowed'.  A value of nil, the default, means never use  
 scalable fonts.  A value of t means any scalable font may be used.  
 Otherwise, the value must be a list of regular expressions.  A  
 scalable font may then be used if it matches a regular expression from  
 that list.  Example:  
   
   (setq scalable-fonts-allowed '("muleindian-2$"))  
   
 allows the use of scalable fonts with registry `muleindian-2'.  
   
 *** Functions and variables related to font selection.  
   
 - Function: x-family-fonts &optional FAMILY FRAME  
   
 Return a list of available fonts of family FAMILY on FRAME.  If FAMILY  
 is omitted or nil, list all families.  Otherwise, FAMILY must be a  
 string, possibly containing wildcards `?' and `*'.  
   
 If FRAME is omitted or nil, use the selected frame.  Each element of  
 the result is a vector [FAMILY WIDTH POINT-SIZE WEIGHT SLANT FIXED-P  
 FULL REGISTRY-AND-ENCODING].  FAMILY is the font family name.  
 POINT-SIZE is the size of the font in 1/10 pt.  WIDTH, WEIGHT, and  
 SLANT are symbols describing the width, weight and slant of the font.  
 These symbols are the same as for face attributes.  FIXED-P is non-nil  
 if the font is fixed-pitch.  FULL is the full name of the font, and  
 REGISTRY-AND-ENCODING is a string giving the registry and encoding of  
 the font.  The result list is sorted according to the current setting  
 of the face font sort order.  
   
 - Function: x-font-family-list  
   
 Return a list of available font families on FRAME.  If FRAME is  
 omitted or nil, use the selected frame.  Value is a list of conses  
 (FAMILY . FIXED-P) where FAMILY is a font family, and FIXED-P is  
 non-nil if fonts of that family are fixed-pitch.  
   
 - Variable: font-list-limit  
   
 Limit for font matching.  If an integer > 0, font matching functions  
 won't load more than that number of fonts when searching for a  
 matching font.  The default is currently 100.  
   
 *** Setting face attributes.  
   
 For the most part, the new face implementation is interface-compatible  
 with the old one.  Old face attribute related functions are now  
 implemented in terms of the new functions `set-face-attribute' and  
 `face-attribute'.  
   
 Face attributes are identified by their names which are keyword  
 symbols.  All attributes can be set to `unspecified'.  
   
 The following attributes are recognized:  
   
 `:family'  
   
 VALUE must be a string specifying the font family, e.g. ``courier'',  
 or a fontset alias name.  If a font family is specified, wild-cards `*'  
 and `?' are allowed.  
   
 `:width'  
   
 VALUE specifies the relative proportionate width of the font to use.  
 It must be one of the symbols `ultra-condensed', `extra-condensed',  
 `condensed', `semi-condensed', `normal', `semi-expanded', `expanded',  
 `extra-expanded', or `ultra-expanded'.  
   
 `:height'  
   
 VALUE must be either an integer specifying the height of the font to use  
 in 1/10 pt, a floating point number specifying the amount by which to  
 scale any underlying face, or a function, which is called with the old  
 height (from the underlying face), and should return the new height.  
   
 `:weight'  
   
 VALUE specifies the weight of the font to use.  It must be one of the  
 symbols `ultra-bold', `extra-bold', `bold', `semi-bold', `normal',  
 `semi-light', `light', `extra-light', `ultra-light'.  
   
 `:slant'  
   
 VALUE specifies the slant of the font to use.  It must be one of the  
 symbols `italic', `oblique', `normal', `reverse-italic', or  
 `reverse-oblique'.  
   
 `:foreground', `:background'  
   
 VALUE must be a color name, a string.  
   
 `:underline'  
   
 VALUE specifies whether characters in FACE should be underlined.  If  
 VALUE is t, underline with foreground color of the face.  If VALUE is  
 a string, underline with that color.  If VALUE is nil, explicitly  
 don't underline.  
   
 `:overline'  
   
 VALUE specifies whether characters in FACE should be overlined.  If  
 VALUE is t, overline with foreground color of the face.  If VALUE is a  
 string, overline with that color.  If VALUE is nil, explicitly don't  
 overline.  
   
 `:strike-through'  
   
 VALUE specifies whether characters in FACE should be drawn with a line  
 striking through them.  If VALUE is t, use the foreground color of the  
 face.  If VALUE is a string, strike-through with that color.  If VALUE  
 is nil, explicitly don't strike through.  
   
 `:box'  
   
 VALUE specifies whether characters in FACE should have a box drawn  
 around them.  If VALUE is nil, explicitly don't draw boxes.  If  
 VALUE is t, draw a box with lines of width 1 in the foreground color  
 of the face.  If VALUE is a string, the string must be a color name,  
 and the box is drawn in that color with a line width of 1.  Otherwise,  
 VALUE must be a property list of the form `(:line-width WIDTH  
 :color COLOR :style STYLE)'.  If a keyword/value pair is missing from  
 the property list, a default value will be used for the value, as  
 specified below.  WIDTH specifies the width of the lines to draw; it  
 defaults to 1.  COLOR is the name of the color to draw in, default is  
 the foreground color of the face for simple boxes, and the background  
 color of the face for 3D boxes.  STYLE specifies whether a 3D box  
 should be draw.  If STYLE is `released-button', draw a box looking  
 like a released 3D button.  If STYLE is `pressed-button' draw a box  
 that appears like a pressed button.  If STYLE is nil, the default if  
 the property list doesn't contain a style specification, draw a 2D  
 box.  
   
 `:inverse-video'  
   
 VALUE specifies whether characters in FACE should be displayed in  
 inverse video. VALUE must be one of t or nil.  
   
 `:stipple'  
   
 If VALUE is a string, it must be the name of a file of pixmap data.  
 The directories listed in the `x-bitmap-file-path' variable are  
 searched.  Alternatively, VALUE may be a list of the form (WIDTH  
 HEIGHT DATA) where WIDTH and HEIGHT are the size in pixels, and DATA  
 is a string containing the raw bits of the bitmap.  VALUE nil means  
 explicitly don't use a stipple pattern.  
   
 For convenience, attributes `:family', `:width', `:height', `:weight',  
 and `:slant' may also be set in one step from an X font name:  
   
 `:font'  
   
 Set font-related face attributes from VALUE.  VALUE must be a valid  
 XLFD font name.  If it is a font name pattern, the first matching font  
 is used--this is for compatibility with the behavior of previous  
 versions of Emacs.  
   
 For compatibility with Emacs 20, keywords `:bold' and `:italic' can  
 be used to specify that a bold or italic font should be used.  VALUE  
 must be t or nil in that case.  A value of `unspecified' is not allowed."  
   
 Please see also the documentation of `set-face-attribute' and  
 `defface'.  
   
 `:inherit'  
   
 VALUE is the name of a face from which to inherit attributes, or a list  
 of face names.  Attributes from inherited faces are merged into the face  
 like an underlying face would be, with higher priority than underlying faces.  
   
 *** Face attributes and X resources  
   
 The following X resource names can be used to set face attributes  
 from X resources:  
   
   Face attribute        X resource              class  
 -----------------------------------------------------------------------  
   :family               attributeFamily .       Face.AttributeFamily  
   :width                attributeWidth          Face.AttributeWidth  
   :height               attributeHeight         Face.AttributeHeight  
   :weight               attributeWeight         Face.AttributeWeight  
   :slant                attributeSlant          Face.AttributeSlant  
    foreground           attributeForeground     Face.AttributeForeground  
   :background           attributeBackground .   Face.AttributeBackground  
   :overline             attributeOverline       Face.AttributeOverline  
   :strike-through       attributeStrikeThrough  Face.AttributeStrikeThrough  
   :box                  attributeBox            Face.AttributeBox  
   :underline            attributeUnderline      Face.AttributeUnderline  
   :inverse-video        attributeInverse        Face.AttributeInverse  
   :stipple              attributeStipple        Face.AttributeStipple  
         or              attributeBackgroundPixmap  
                                                 Face.AttributeBackgroundPixmap  
   :font                 attributeFont           Face.AttributeFont  
   :bold                 attributeBold           Face.AttributeBold  
   :italic               attributeItalic .       Face.AttributeItalic  
   :font                 attributeFont           Face.AttributeFont  
   
 *** Text property `face'.  
   
 The value of the `face' text property can now be a single face  
 specification or a list of such specifications.  Each face  
 specification can be  
   
 1. A symbol or string naming a Lisp face.  
   
 2. A property list of the form (KEYWORD VALUE ...) where each  
    KEYWORD is a face attribute name, and VALUE is an appropriate value  
    for that attribute.  Please see the doc string of `set-face-attribute'  
    for face attribute names.  
   
 3. Conses of the form (FOREGROUND-COLOR . COLOR) or  
    (BACKGROUND-COLOR . COLOR) where COLOR is a color name.  This is  
    for compatibility with previous Emacs versions.  
   
 ** Support functions for colors on text-only terminals.  
   
 The function `tty-color-define' can be used to define colors for use  
 on TTY and MSDOS frames.  It maps a color name to a color number on  
 the terminal.  Emacs defines a couple of common color mappings by  
 default.  You can get defined colors with a call to  
 `defined-colors'.  The function `tty-color-clear' can be  
 used to clear the mapping table.  
   
 ** Unified support for colors independent of frame type.  
   
 The new functions `defined-colors', `color-defined-p', `color-values',  
 and `display-color-p' work for any type of frame.  On frames whose  
 type is neither x nor w32, these functions transparently map X-style  
 color specifications to the closest colors supported by the frame  
 display.  Lisp programs should use these new functions instead of the  
 old `x-defined-colors', `x-color-defined-p', `x-color-values', and  
 `x-display-color-p'.  (The old function names are still available for  
 compatibility; they are now aliases of the new names.)  Lisp programs  
 should no more look at the value of the variable window-system to  
 modify their color-related behavior.  
   
 The primitives `color-gray-p' and `color-supported-p' also work for  
 any frame type.  
   
 ** Platform-independent functions to describe display capabilities.  
   
 The new functions `display-mouse-p', `display-popup-menus-p',  
 `display-graphic-p', `display-selections-p', `display-screens',  
 `display-pixel-width', `display-pixel-height', `display-mm-width',  
 `display-mm-height', `display-backing-store', `display-save-under',  
 `display-planes', `display-color-cells', `display-visual-class', and  
 `display-grayscale-p' describe the basic capabilities of a particular  
 display.  Lisp programs should call these functions instead of testing  
 the value of the variables `window-system' or `system-type', or calling  
 platform-specific functions such as `x-display-pixel-width'.  
   
 The new function `display-images-p' returns non-nil if a particular  
 display can display image files.  
   
 ** The minibuffer prompt is now actually inserted in the minibuffer.  
   
 This makes it possible to scroll through the prompt, if you want to.  
 To disallow this completely (like previous versions of emacs), customize  
 the variable `minibuffer-prompt-properties', and turn on the  
 `Inviolable' option.  
   
 The function `minibuffer-prompt-end' returns the current position of the  
 end of the minibuffer prompt, if the minibuffer is current.  
 Otherwise, it returns `(point-min)'.  
   
 ** New `field' abstraction in buffers.  
   
 There is now code to support an abstraction called `fields' in emacs  
 buffers.  A field is a contiguous region of text with the same `field'  
 property (which can be a text property or an overlay).  
   
 Many emacs functions, such as forward-word, forward-sentence,  
 forward-paragraph, beginning-of-line, etc., stop moving when they come  
 to the boundary between fields; beginning-of-line and end-of-line will  
 not let the point move past the field boundary, but other movement  
 commands continue into the next field if repeated.  Stopping at field  
 boundaries can be suppressed programmatically by binding  
 `inhibit-field-text-motion' to a non-nil value around calls to these  
 functions.  
   
 Now that the minibuffer prompt is inserted into the minibuffer, it is in  
 a separate field from the user-input part of the buffer, so that common  
 editing commands treat the user's text separately from the prompt.  
   
 The following functions are defined for operating on fields:  
   
 - Function: constrain-to-field NEW-POS OLD-POS &optional ESCAPE-FROM-EDGE ONLY-IN-LINE INHIBIT-CAPTURE-PROPERTY  
   
 Return the position closest to NEW-POS that is in the same field as OLD-POS.  
   
 A field is a region of text with the same `field' property.  
 If NEW-POS is nil, then the current point is used instead, and set to the  
 constrained position if that is different.  
   
 If OLD-POS is at the boundary of two fields, then the allowable  
 positions for NEW-POS depends on the value of the optional argument  
 ESCAPE-FROM-EDGE: If ESCAPE-FROM-EDGE is nil, then NEW-POS is  
 constrained to the field that has the same `field' char-property  
 as any new characters inserted at OLD-POS, whereas if ESCAPE-FROM-EDGE  
 is non-nil, NEW-POS is constrained to the union of the two adjacent  
 fields.  Additionally, if two fields are separated by another field with  
 the special value `boundary', then any point within this special field is  
 also considered to be `on the boundary'.  
   
 If the optional argument ONLY-IN-LINE is non-nil and constraining  
 NEW-POS would move it to a different line, NEW-POS is returned  
 unconstrained.  This useful for commands that move by line, like  
 C-n or C-a, which should generally respect field boundaries  
 only in the case where they can still move to the right line.  
   
 If the optional argument INHIBIT-CAPTURE-PROPERTY is non-nil, and OLD-POS has  
 a non-nil property of that name, then any field boundaries are ignored.  
   
 Field boundaries are not noticed if `inhibit-field-text-motion' is non-nil.  
   
 - Function: delete-field &optional POS  
   
 Delete the field surrounding POS.  
 A field is a region of text with the same `field' property.  
 If POS is nil, the value of point is used for POS.  
   
 - Function: field-beginning &optional POS ESCAPE-FROM-EDGE  
   
 Return the beginning of the field surrounding POS.  
 A field is a region of text with the same `field' property.  
 If POS is nil, the value of point is used for POS.  
 If ESCAPE-FROM-EDGE is non-nil and POS is at the beginning of its  
 field, then the beginning of the *previous* field is returned.  
   
 - Function: field-end &optional POS ESCAPE-FROM-EDGE  
   
 Return the end of the field surrounding POS.  
 A field is a region of text with the same `field' property.  
 If POS is nil, the value of point is used for POS.  
 If ESCAPE-FROM-EDGE is non-nil and POS is at the end of its field,  
 then the end of the *following* field is returned.  
   
 - Function: field-string &optional POS  
   
 Return the contents of the field surrounding POS as a string.  
 A field is a region of text with the same `field' property.  
 If POS is nil, the value of point is used for POS.  
   
 - Function: field-string-no-properties &optional POS  
   
 Return the contents of the field around POS, without text-properties.  
 A field is a region of text with the same `field' property.  
 If POS is nil, the value of point is used for POS.  
   
 ** Image support.  
   
 Emacs can now display images.  Images are inserted into text by giving  
 strings or buffer text a `display' text property containing one of  
 (AREA IMAGE) or IMAGE.  The display of the `display' property value  
 replaces the display of the characters having that property.  
   
 If the property value has the form (AREA IMAGE), AREA must be one of  
 `(margin left-margin)', `(margin right-margin)' or `(margin nil)'.  If  
 AREA is `(margin nil)', IMAGE will be displayed in the text area of a  
 window, otherwise it will be displayed in the left or right marginal  
 area.  
   
 IMAGE is an image specification.  
   
 *** Image specifications  
   
 Image specifications are lists of the form `(image PROPS)' where PROPS  
 is a property list whose keys are keyword symbols.  Each  
 specifications must contain a property `:type TYPE' with TYPE being a  
 symbol specifying the image type, e.g. `xbm'.  Properties not  
 described below are ignored.  
   
 The following is a list of properties all image types share.  
   
 `:ascent ASCENT'  
   
 ASCENT must be a number in the range 0..100, or the symbol `center'.  
 If it is a number, it specifies the percentage of the image's height  
 to use for its ascent.  
   
 If not specified, ASCENT defaults to the value 50 which means that the  
 image will be centered with the base line of the row it appears in.  
   
 If ASCENT is `center' the image is vertically centered around a  
 centerline which is the vertical center of text drawn at the position  
 of the image, in the manner specified by the text properties and  
 overlays that apply to the image.  
   
 `:margin MARGIN'  
   
 MARGIN must be either a number >= 0 specifying how many pixels to put  
 as margin around the image, or a pair (X . Y) with X specifying the  
 horizontal margin and Y specifying the vertical margin.  Default is 0.  
   
 `:relief RELIEF'  
   
 RELIEF is analogous to the `:relief' attribute of faces.  Puts a relief  
 around an image.  
   
 `:conversion ALGO'  
   
 Apply an image algorithm to the image before displaying it.  
   
 ALGO `laplace' or `emboss' means apply a Laplace or ``emboss''  
 edge-detection algorithm to the image.  
   
 ALGO `(edge-detection :matrix MATRIX :color-adjust ADJUST)' means  
 apply a general edge-detection algorithm.  MATRIX must be either a  
 nine-element list or a nine-element vector of numbers.  A pixel at  
 position x/y in the transformed image is computed from original pixels  
 around that position.  MATRIX specifies, for each pixel in the  
 neighborhood of x/y, a factor with which that pixel will influence the  
 transformed pixel; element 0 specifies the factor for the pixel at  
 x-1/y-1, element 1 the factor for the pixel at x/y-1 etc. as shown  
 below.  
   
   (x-1/y-1  x/y-1  x+1/y-1  
    x-1/y    x/y    x+1/y  
    x-1/y+1  x/y+1  x+1/y+1)  
   
 The resulting pixel is computed from the color intensity of the color  
 resulting from summing up the RGB values of surrounding pixels,  
 multiplied by the specified factors, and dividing that sum by the sum  
 of the factors' absolute values.  
   
 Laplace edge-detection currently uses a matrix of  
   
   (1  0  0  
    0  0  0  
    9  9 -1)  
   
 Emboss edge-detection uses a matrix of  
   
   ( 2 -1  0  
    -1  0  1  
     0  1 -2)  
   
 ALGO `disabled' means transform the image so that it looks  
 ``disabled''.  
   
 `:mask MASK'  
   
 If MASK is `heuristic' or `(heuristic BG)', build a clipping mask for  
 the image, so that the background of a frame is visible behind the  
 image.  If BG is not specified, or if BG is t, determine the  
 background color of the image by looking at the 4 corners of the  
 image, assuming the most frequently occurring color from the corners is  
 the background color of the image.  Otherwise, BG must be a list `(RED  
 GREEN BLUE)' specifying the color to assume for the background of the  
 image.  
   
 If MASK is nil, remove a mask from the image, if it has one.  Images  
 in some formats include a mask which can be removed by specifying  
 `:mask nil'.  
   
 `:file FILE'  
   
 Load image from FILE.  If FILE is not absolute after expanding it,  
 search for the image in `data-directory'.  Some image types support  
 building images from data.  When this is done, no `:file' property  
 may be present in the image specification.  
   
 `:data DATA'  
   
 Get image data from DATA.  (As of this writing, this is not yet  
 supported for image type `postscript').  Either :file or :data may be  
 present in an image specification, but not both.  All image types  
 support strings as DATA, some types allow additional types of DATA.  
   
 *** Supported image types  
   
 **** XBM, image type `xbm'.  
   
 XBM images don't require an external library.  Additional image  
 properties supported are  
   
 `:foreground FG'  
   
 FG must be a string specifying the image foreground color, or nil  
 meaning to use the default.  Default is the frame's foreground.  
   
 `:background BG'  
   
 BG must be a string specifying the image foreground color, or nil  
 meaning to use the default.  Default is the frame's background color.  
   
 XBM images can be constructed from data instead of file.  In this  
 case, the image specification must contain the following properties  
 instead of a `:file' property.  
   
 `:width WIDTH'  
   
 WIDTH specifies the width of the image in pixels.  
   
 `:height HEIGHT'  
   
 HEIGHT specifies the height of the image in pixels.  
   
 `:data DATA'  
   
 DATA must be either  
   
    1. a string large enough to hold the bitmap data, i.e. it must  
    have a size >= (WIDTH + 7) / 8 * HEIGHT  
   
    2. a bool-vector of size >= WIDTH * HEIGHT  
   
    3. a vector of strings or bool-vectors, one for each line of the  
    bitmap.  
   
    4. a string that's an in-memory XBM file.  Neither width nor  
    height may be specified in this case because these are defined  
    in the file.  
   
 **** XPM, image type `xpm'  
   
 XPM images require the external library `libXpm', package  
 `xpm-3.4k.tar.gz', version 3.4k or later.  Make sure the library is  
 found when Emacs is configured by supplying appropriate paths via  
 `--x-includes' and `--x-libraries'.  
   
 Additional image properties supported are:  
   
 `:color-symbols SYMBOLS'  
   
 SYMBOLS must be a list of pairs (NAME . COLOR), with NAME being the  
 name of color as it appears in an XPM file, and COLOR being an X color  
 name.  
   
 XPM images can be built from memory instead of files.  In that case,  
 add a `:data' property instead of a `:file' property.  
   
 The XPM library uses libz in its implementation so that it is able  
 to display compressed images.  
   
 **** PBM, image type `pbm'  
   
 PBM images don't require an external library.  Color, gray-scale and  
 mono images are supported.  Additional image properties supported for  
 mono images are  
   
 `:foreground FG'  
   
 FG must be a string specifying the image foreground color, or nil  
 meaning to use the default.  Default is the frame's foreground.  
   
 `:background FG'  
   
 BG must be a string specifying the image foreground color, or nil  
 meaning to use the default.  Default is the frame's background color.  
   
 **** JPEG, image type `jpeg'  
   
 Support for JPEG images requires the external library `libjpeg',  
 package `jpegsrc.v6a.tar.gz', or later.  Additional image properties  
 are:  
   
 **** TIFF, image type `tiff'  
   
 Support for TIFF images requires the external library `libtiff',  
 package `tiff-v3.4-tar.gz', or later.  There are no additional image  
 properties defined.  
   
 **** GIF, image type `gif'  
   
 Support for GIF images requires the external library `libungif', package  
 `libungif-4.1.0', or later.  
   
 Additional image properties supported are:  
   
 `:index INDEX'  
   
 INDEX must be an integer >= 0.  Load image number INDEX from a  
 multi-image GIF file.  An error is signaled if INDEX is too large.  
80    
81  This could be used to implement limited support for animated GIFs.  ** The character codes for characters from the
82  For example, the following function displays a multi-image GIF file  eight-bit-control/eight-bit-graphic charsets aren't now in the range
83  at point-min in the current buffer, switching between sub-images  128-255.
 every 0.1 seconds.  
84    
 (defun show-anim (file max)  
   "Display multi-image GIF file FILE which contains MAX subimages."  
   (display-anim (current-buffer) file 0 max t))  
   
 (defun display-anim (buffer file idx max first-time)  
   (when (= idx max)  
     (setq idx 0))  
   (let ((img (create-image file nil nil :index idx)))  
     (save-excursion  
       (set-buffer buffer)  
       (goto-char (point-min))  
       (unless first-time (delete-char 1))  
       (insert-image img "x"))  
     (run-with-timer 0.1 nil 'display-anim buffer file (1+ idx) max nil)))  
   
 **** PNG, image type `png'  
   
 Support for PNG images requires the external library `libpng',  
 package `libpng-1.0.2.tar.gz', or later.  There are no additional image  
 properties defined.  
   
 **** Ghostscript, image type `postscript'.  
   
 Additional image properties supported are:  
   
 `:pt-width WIDTH'  
   
 WIDTH is width of the image in pt (1/72 inch).  WIDTH must be an  
 integer.  This is a required property.  
   
 `:pt-height HEIGHT'  
   
 HEIGHT specifies the height of the image in pt (1/72 inch).  HEIGHT  
 must be a integer.  This is an required property.  
   
 `:bounding-box BOX'  
   
 BOX must be a list or vector of 4 integers giving the bounding box of  
 the PS image, analogous to the `BoundingBox' comment found in PS  
 files.  This is an required property.  
   
 Part of the Ghostscript interface is implemented in Lisp.  See  
 lisp/gs.el.  
   
 *** Lisp interface.  
   
 The variable `image-types' contains a list of those image types  
 which are supported in the current configuration.  
   
 Images are stored in an image cache and removed from the cache when  
 they haven't been displayed for `image-cache-eviction-delay seconds.  
 The function `clear-image-cache' can be used to clear the image cache  
 manually.  Images in the cache are compared with `equal', i.e. all  
 images with `equal' specifications share the same image.  
   
 *** Simplified image API, image.el  
   
 The new Lisp package image.el contains functions that simplify image  
 creation and putting images into text.  The function `create-image'  
 can be used to create images.  The macro `defimage' can be used to  
 define an image based on available image types.  The functions  
 `put-image' and `insert-image' can be used to insert an image into a  
 buffer.  
   
 ** Display margins.  
   
 Windows can now have margins which are used for special text  
 and images.  
   
 To give a window margins, either set the buffer-local variables  
 `left-margin-width' and `right-margin-width', or call  
 `set-window-margins'.  The function `window-margins' can be used to  
 obtain the current settings.  To make `left-margin-width' and  
 `right-margin-width' take effect, you must set them before displaying  
 the buffer in a window, or use `set-window-buffer' to force an update  
 of the display margins.  
   
 You can put text in margins by giving it a `display' text property  
 containing a pair of the form `(LOCATION . VALUE)', where LOCATION is  
 one of `left-margin' or `right-margin' or nil.  VALUE can be either a  
 string, an image specification or a stretch specification (see later  
 in this file).  
   
 ** Help display  
   
 Emacs displays short help messages in the echo area, when the mouse  
 moves over a tool-bar item or a piece of text that has a text property  
 `help-echo'.  This feature also applies to strings in the mode line  
 that have a `help-echo' property.  
   
 If the value of the `help-echo' property is a function, that function  
 is called with three arguments WINDOW, OBJECT and POSITION.  WINDOW is  
 the window in which the help was found.  
   
 If OBJECT is a buffer, POS is the position in the buffer where the  
 `help-echo' text property was found.  
   
 If OBJECT is an overlay, that overlay has a `help-echo' property, and  
 POS is the position in the overlay's buffer under the mouse.  
   
 If OBJECT is a string (an overlay string or a string displayed with  
 the `display' property), POS is the position in that string under the  
 mouse.  
   
 If the value of the `help-echo' property is neither a function nor a  
 string, it is evaluated to obtain a help string.  
   
 For tool-bar and menu-bar items, their key definition is used to  
 determine the help to display.  If their definition contains a  
 property `:help FORM', FORM is evaluated to determine the help string.  
 For tool-bar items without a help form, the caption of the item is  
 used as help string.  
   
 The hook `show-help-function' can be set to a function that displays  
 the help string differently.  For example, enabling a tooltip window  
 causes the help display to appear there instead of in the echo area.  
   
 ** Vertical fractional scrolling.  
   
 The display of text in windows can be scrolled smoothly in pixels.  
 This is useful, for example, for making parts of large images visible.  
   
 The function `window-vscroll' returns the current value of vertical  
 scrolling, a non-negative fraction of the canonical character height.  
 The function `set-window-vscroll' can be used to set the vertical  
 scrolling value.  Here is an example of how these function might be  
 used.  
   
   (global-set-key [A-down]  
     #'(lambda ()  
         (interactive)  
         (set-window-vscroll (selected-window)  
                             (+ 0.5 (window-vscroll)))))  
   (global-set-key [A-up]  
     #'(lambda ()  
         (interactive)  
         (set-window-vscroll (selected-window)  
                             (- (window-vscroll) 0.5)))))  
   
 ** New hook `fontification-functions'.  
   
 Functions from `fontification-functions' are called from redisplay  
 when it encounters a region of text that is not yet fontified.  This  
 variable automatically becomes buffer-local when set.  Each function  
 is called with one argument, POS.  
   
 At least one of the hook functions should fontify one or more  
 characters starting at POS in the current buffer.  It should mark them  
 as fontified by giving them a non-nil value of the `fontified' text  
 property.  It may be reasonable for these functions to check for the  
 `fontified' property and not put it back on, but they do not have to.  
   
 ** Tool bar support.  
   
 Emacs supports a tool bar at the top of a frame under X.  The frame  
 parameter `tool-bar-lines' (X resource "toolBar", class "ToolBar")  
 controls how may lines to reserve for the tool bar.  A zero value  
 suppresses the tool bar.  If the value is non-zero and  
 `auto-resize-tool-bars' is non-nil the tool bar's size will be changed  
 automatically so that all tool bar items are visible.  
   
 *** Tool bar item definitions  
   
 Tool bar items are defined using `define-key' with a prefix-key  
 `tool-bar'.  For example `(define-key global-map [tool-bar item1] ITEM)'  
 where ITEM is a list `(menu-item CAPTION BINDING PROPS...)'.  
   
 CAPTION is the caption of the item, If it's not a string, it is  
 evaluated to get a string.  The caption is currently not displayed in  
 the tool bar, but it is displayed if the item doesn't have a `:help'  
 property (see below).  
   
 BINDING is the tool bar item's binding.  Tool bar items with keymaps as  
 binding are currently ignored.  
   
 The following properties are recognized:  
   
 `:enable FORM'.  
   
 FORM is evaluated and specifies whether the tool bar item is enabled  
 or disabled.  
   
 `:visible FORM'  
   
 FORM is evaluated and specifies whether the tool bar item is displayed.  
   
 `:filter FUNCTION'  
   
 FUNCTION is called with one parameter, the same list BINDING in which  
 FUNCTION is specified as the filter.  The value FUNCTION returns is  
 used instead of BINDING to display this item.  
   
 `:button (TYPE SELECTED)'  
   
 TYPE must be one of `:radio' or `:toggle'.  SELECTED is evaluated  
 and specifies whether the button is selected (pressed) or not.  
   
 `:image IMAGES'  
   
 IMAGES is either a single image specification or a vector of four  
 image specifications.  If it is a vector, this table lists the  
 meaning of each of the four elements:  
   
    Index        Use when item is  
    ----------------------------------------  
      0          enabled and selected  
      1          enabled and deselected  
      2          disabled and selected  
      3          disabled and deselected  
   
 If IMAGE is a single image specification, a Laplace edge-detection  
 algorithm is used on that image to draw the image in disabled state.  
   
 `:help HELP-STRING'.  
   
 Gives a help string to display for the tool bar item.  This help  
 is displayed when the mouse is moved over the item.  
   
 The function `toolbar-add-item' is a convenience function for adding  
 toolbar items generally, and `tool-bar-add-item-from-menu' can be used  
 to define a toolbar item with a binding copied from an item on the  
 menu bar.  
   
 The default bindings use a menu-item :filter to derive the tool-bar  
 dynamically from variable `tool-bar-map' which may be set  
 buffer-locally to override the global map.  
   
 *** Tool-bar-related variables.  
   
 If `auto-resize-tool-bar' is non-nil, the tool bar will automatically  
 resize to show all defined tool bar items.  It will never grow larger  
 than 1/4 of the frame's size.  
   
 If `auto-raise-tool-bar-buttons' is non-nil, tool bar buttons will be  
 raised when the mouse moves over them.  
   
 You can add extra space between tool bar items by setting  
 `tool-bar-button-margin' to a positive integer specifying a number of  
 pixels, or a pair of integers (X . Y) specifying horizontal and  
 vertical margins .  Default is 1.  
   
 You can change the shadow thickness of tool bar buttons by setting  
 `tool-bar-button-relief' to an integer.  Default is 3.  
   
 *** Tool-bar clicks with modifiers.  
   
 You can bind commands to clicks with control, shift, meta etc. on  
 a tool bar item.  If  
   
   (define-key global-map [tool-bar shell]  
     '(menu-item "Shell" shell  
                 :image (image :type xpm :file "shell.xpm")))  
   
 is the original tool bar item definition, then  
   
   (define-key global-map [tool-bar S-shell] 'some-command)  
   
 makes a binding to run `some-command' for a shifted click on the same  
 item.  
   
 ** Mode line changes.  
   
 *** Mouse-sensitive mode line.  
   
 The mode line can be made mouse-sensitive by displaying strings there  
 that have a `local-map' text property.  There are three ways to display  
 a string with a `local-map' property in the mode line.  
   
 1. The mode line spec contains a variable whose string value has  
 a `local-map' text property.  
   
 2. The mode line spec contains a format specifier (e.g. `%12b'), and  
 that format specifier has a `local-map' property.  
   
 3. The mode line spec contains a list containing `:eval FORM'.  FORM  
 is evaluated.  If the result is a string, and that string has a  
 `local-map' property.  
   
 The same mechanism is used to determine the `face' and `help-echo'  
 properties of strings in the mode line.  See `bindings.el' for an  
 example.  
   
 *** If a mode line element has the form `(:eval FORM)', FORM is  
 evaluated and the result is used as mode line element.  
   
 *** You can suppress mode-line display by setting the buffer-local  
 variable mode-line-format to nil.  
   
 *** A headerline can now be displayed at the top of a window.  
   
 This mode line's contents are controlled by the new variable  
 `header-line-format' and `default-header-line-format' which are  
 completely analogous to `mode-line-format' and  
 `default-mode-line-format'.  A value of nil means don't display a top  
 line.  
   
 The appearance of top mode lines is controlled by the face  
 `header-line'.  
   
 The function `coordinates-in-window-p' returns `header-line' for a  
 position in the header-line.  
   
 ** Text property `display'  
   
 The `display' text property is used to insert images into text,  
 replace text with other text, display text in marginal area, and it is  
 also used to control other aspects of how text displays.  The value of  
 the `display' property should be a display specification, as described  
 below, or a list or vector containing display specifications.  
   
 *** Replacing text, displaying text in marginal areas  
   
 To replace the text having the `display' property with some other  
 text, use a display specification of the form `(LOCATION STRING)'.  
   
 If LOCATION is `(margin left-margin)', STRING is displayed in the left  
 marginal area, if it is `(margin right-margin)', it is displayed in  
 the right marginal area, and if LOCATION is `(margin nil)' STRING  
 is displayed in the text.  In the latter case you can also use the  
 simpler form STRING as property value.  
   
 *** Variable width and height spaces  
   
 To display a space of fractional width or height, use a display  
 specification of the form `(LOCATION STRECH)'.  If LOCATION is  
 `(margin left-margin)', the space is displayed in the left marginal  
 area, if it is `(margin right-margin)', it is displayed in the right  
 marginal area, and if LOCATION is `(margin nil)' the space is  
 displayed in the text.  In the latter case you can also use the  
 simpler form STRETCH as property value.  
   
 The stretch specification STRETCH itself is a list of the form `(space  
 PROPS)', where PROPS is a property list which can contain the  
 properties described below.  
   
 The display of the fractional space replaces the display of the  
 characters having the `display' property.  
   
 - :width WIDTH  
   
 Specifies that the space width should be WIDTH times the normal  
 character width.  WIDTH can be an integer or floating point number.  
   
 - :relative-width FACTOR  
   
 Specifies that the width of the stretch should be computed from the  
 first character in a group of consecutive characters that have the  
 same `display' property.  The computation is done by multiplying the  
 width of that character by FACTOR.  
   
 - :align-to HPOS  
   
 Specifies that the space should be wide enough to reach HPOS.  The  
 value HPOS is measured in units of the normal character width.  
   
 Exactly one of the above properties should be used.  
   
 - :height HEIGHT  
   
 Specifies the height of the space, as HEIGHT, measured in terms of the  
 normal line height.  
   
 - :relative-height FACTOR  
   
 The height of the space is computed as the product of the height  
 of the text having the `display' property and FACTOR.  
   
 - :ascent ASCENT  
   
 Specifies that ASCENT percent of the height of the stretch should be  
 used for the ascent of the stretch, i.e. for the part above the  
 baseline.  The value of ASCENT must be a non-negative number less or  
 equal to 100.  
   
 You should not use both `:height' and `:relative-height' together.  
   
 *** Images  
   
 A display specification for an image has the form `(LOCATION  
 . IMAGE)', where IMAGE is an image specification.  The image replaces,  
 in the display, the characters having this display specification in  
 their `display' text property.  If LOCATION is `(margin left-margin)',  
 the image will be displayed in the left marginal area, if it is  
 `(margin right-margin)' it will be displayed in the right marginal  
 area, and if LOCATION is `(margin nil)' the image will be displayed in  
 the text.  In the latter case you can also use the simpler form IMAGE  
 as display specification.  
   
 *** Other display properties  
   
 - (space-width FACTOR)  
   
 Specifies that space characters in the text having that property  
 should be displayed FACTOR times as wide as normal; FACTOR must be an  
 integer or float.  
   
 - (height HEIGHT)  
   
 Display text having this property in a font that is smaller or larger.  
   
 If HEIGHT is a list of the form `(+ N)', where N is an integer, that  
 means to use a font that is N steps larger.  If HEIGHT is a list of  
 the form `(- N)', that means to use a font that is N steps smaller.  A  
 ``step'' is defined by the set of available fonts; each size for which  
 a font is available counts as a step.  
   
 If HEIGHT is a number, that means to use a font that is HEIGHT times  
 as tall as the frame's default font.  
   
 If HEIGHT is a symbol, it is called as a function with the current  
 height as argument.  The function should return the new height to use.  
   
 Otherwise, HEIGHT is evaluated to get the new height, with the symbol  
 `height' bound to the current specified font height.  
   
 - (raise FACTOR)  
   
 FACTOR must be a number, specifying a multiple of the current  
 font's height.  If it is positive, that means to display the characters  
 raised.  If it is negative, that means to display them lower down.  The  
 amount of raising or lowering is computed without taking account of the  
 `height' subproperty.  
   
 *** Conditional display properties  
   
 All display specifications can be conditionalized.  If a specification  
 has the form `(when CONDITION . SPEC)', the specification SPEC applies  
 only when CONDITION yields a non-nil value when evaluated.  During the  
 evaluation, `object' is bound to the string or buffer having the  
 conditional display property; `position' and `buffer-position' are  
 bound to the position within `object' and the buffer position where  
 the display property was found, respectively.  Both positions can be  
 different when object is a string.  
   
 The normal specification consisting of SPEC only is equivalent to  
 `(when t . SPEC)'.  
   
 ** New menu separator types.  
   
 Emacs now supports more than one menu separator type.  Menu items with  
 item names consisting of dashes only (including zero dashes) are  
 treated like before.  In addition, the following item names are used  
 to specify other menu separator types.  
   
 - `--no-line' or `--space', or `--:space', or `--:noLine'  
   
 No separator lines are drawn, but a small space is inserted where the  
 separator occurs.  
   
 - `--single-line' or `--:singleLine'  
   
 A single line in the menu's foreground color.  
   
 - `--double-line' or `--:doubleLine'  
   
 A double line in the menu's foreground color.  
   
 - `--single-dashed-line' or `--:singleDashedLine'  
   
 A single dashed line in the menu's foreground color.  
   
 - `--double-dashed-line' or `--:doubleDashedLine'  
   
 A double dashed line in the menu's foreground color.  
   
 - `--shadow-etched-in' or `--:shadowEtchedIn'  
   
 A single line with 3D sunken appearance.  This is the form  
 displayed for item names consisting of dashes only.  
   
 - `--shadow-etched-out' or `--:shadowEtchedOut'  
   
 A single line with 3D raised appearance.  
   
 - `--shadow-etched-in-dash' or `--:shadowEtchedInDash'  
   
 A single dashed line with 3D sunken appearance.  
   
 - `--shadow-etched-out-dash' or `--:shadowEtchedOutDash'  
   
 A single dashed line with 3D raise appearance.  
   
 - `--shadow-double-etched-in' or `--:shadowDoubleEtchedIn'  
   
 Two lines with 3D sunken appearance.  
   
 - `--shadow-double-etched-out' or `--:shadowDoubleEtchedOut'  
   
 Two lines with 3D raised appearance.  
   
 - `--shadow-double-etched-in-dash' or `--:shadowDoubleEtchedInDash'  
   
 Two dashed lines with 3D sunken appearance.  
   
 - `--shadow-double-etched-out-dash' or `--:shadowDoubleEtchedOutDash'  
   
 Two dashed lines with 3D raised appearance.  
   
 Under LessTif/Motif, the last four separator types are displayed like  
 the corresponding single-line separators.  
   
 ** New frame parameters for scroll bar colors.  
   
 The new frame parameters `scroll-bar-foreground' and  
 `scroll-bar-background' can be used to change scroll bar colors.  
 Their value must be either a color name, a string, or nil to specify  
 that scroll bars should use a default color.  For toolkit scroll bars,  
 default colors are toolkit specific.  For non-toolkit scroll bars, the  
 default background is the background color of the frame, and the  
 default foreground is black.  
   
 The X resource name of these parameters are `scrollBarForeground'  
 (class ScrollBarForeground) and `scrollBarBackground' (class  
 `ScrollBarBackground').  
   
 Setting these parameters overrides toolkit specific X resource  
 settings for scroll bar colors.  
   
 ** You can set `redisplay-dont-pause' to a non-nil value to prevent  
 display updates from being interrupted when input is pending.  
   
 ** Changing a window's width may now change its window start if it  
 starts on a continuation line.  The new window start is computed based  
 on the window's new width, starting from the start of the continued  
 line as the start of the screen line with the minimum distance from  
 the original window start.  
   
 ** The variable `hscroll-step' and the functions  
 `hscroll-point-visible' and `hscroll-window-column' have been removed  
 now that proper horizontal scrolling is implemented.  
   
 ** Windows can now be made fixed-width and/or fixed-height.  
   
 A window is fixed-size if its buffer has a buffer-local variable  
 `window-size-fixed' whose value is not nil.  A value of `height' makes  
 windows fixed-height, a value of `width' makes them fixed-width, any  
 other non-nil value makes them both fixed-width and fixed-height.  
   
 The following code makes all windows displaying the current buffer  
 fixed-width and fixed-height.  
   
   (set (make-local-variable 'window-size-fixed) t)  
   
 A call to enlarge-window on a window gives an error if that window is  
 fixed-width and it is tried to change the window's width, or if the  
 window is fixed-height, and it is tried to change its height.  To  
 change the size of a fixed-size window, bind `window-size-fixed'  
 temporarily to nil, for example  
   
   (let ((window-size-fixed nil))  
      (enlarge-window 10))  
   
 Likewise, an attempt to split a fixed-height window vertically,  
 or a fixed-width window horizontally results in a error.  
   
 ** The cursor-type frame parameter is now supported on MS-DOS  
 terminals.  When Emacs starts, it by default changes the cursor shape  
 to a solid box, as it does on Unix.  The `cursor-type' frame parameter  
 overrides this as it does on Unix, except that the bar cursor is  
 horizontal rather than vertical (since the MS-DOS display doesn't  
 support a vertical-bar cursor).  
   
   
   
 * Emacs 20.7 is a bug-fix release with few user-visible changes  
   
 ** It is now possible to use CCL-based coding systems for keyboard  
 input.  
   
 ** ange-ftp now handles FTP security extensions, like Kerberos.  
   
 ** Rmail has been extended to recognize more forms of digest messages.  
   
 ** Now, most coding systems set in keyboard coding system work not  
 only for character input, but also in incremental search.  The  
 exceptions are such coding systems that handle 2-byte character sets  
 (e.g euc-kr, euc-jp) and that use ISO's escape sequence  
 (e.g. iso-2022-jp).  They are ignored in incremental search.  
   
 ** Support for Macintosh PowerPC-based machines running GNU/Linux has  
 been added.  
   
   
 * Emacs 20.6 is a bug-fix release with one user-visible change  
   
 ** Support for ARM-based non-RISCiX machines has been added.  
   
   
   
 * Emacs 20.5 is a bug-fix release with no user-visible changes.  
   
 ** Not new, but not mentioned before:  
 M-w when Transient Mark mode is enabled disables the mark.  
   
 * Changes in Emacs 20.4  
   
 ** Init file may be called .emacs.el.  
   
 You can now call the Emacs init file `.emacs.el'.  
 Formerly the name had to be `.emacs'.  If you use the name  
 `.emacs.el', you can byte-compile the file in the usual way.  
   
 If both `.emacs' and `.emacs.el' exist, the latter file  
 is the one that is used.  
   
 ** shell-command, and shell-command-on-region, now return  
 the exit code of the command (unless it is asynchronous).  
 Also, you can specify a place to put the error output,  
 separate from the command's regular output.  
 Interactively, the variable shell-command-default-error-buffer  
 says where to put error output; set it to a buffer name.  
 In calls from Lisp, an optional argument ERROR-BUFFER specifies  
 the buffer name.  
   
 When you specify a non-nil error buffer (or buffer name), any error  
 output is inserted before point in that buffer, with \f\n to separate  
 it from the previous batch of error output.  The error buffer is not  
 cleared, so error output from successive commands accumulates there.  
   
 ** Setting the default value of enable-multibyte-characters to nil in  
 the .emacs file, either explicitly using setq-default, or via Custom,  
 is now essentially equivalent to using --unibyte: all buffers  
 created during startup will be made unibyte after loading .emacs.  
   
 ** C-x C-f now handles the wildcards * and ? in file names.  For  
 example, typing C-x C-f c*.c RET visits all the files whose names  
 match c*.c.  To visit a file whose name contains * or ?, add the  
 quoting sequence /: to the beginning of the file name.  
   
 ** The M-x commands keep-lines, flush-lines and count-matches  
 now have the same feature as occur and query-replace:  
 if the pattern contains any upper case letters, then  
 they never ignore case.  
   
 ** The end-of-line format conversion feature previously mentioned  
 under `* Emacs 20.1 changes for MS-DOS and MS-Windows' actually  
 applies to all operating systems.  Emacs recognizes from the contents  
 of a file what convention it uses to separate lines--newline, CRLF, or  
 just CR--and automatically converts the contents to the normal Emacs  
 convention (using newline to separate lines) for editing.  This is a  
 part of the general feature of coding system conversion.  
   
 If you subsequently save the buffer, Emacs converts the text back to  
 the same format that was used in the file before.  
   
 You can turn off end-of-line conversion by setting the variable  
 `inhibit-eol-conversion' to non-nil, e.g. with Custom in the MULE group.  
   
 ** The character set property `prefered-coding-system' has been  
 renamed to `preferred-coding-system', for the sake of correct spelling.  
 This is a fairly internal feature, so few programs should be affected.  
   
 ** Mode-line display of end-of-line format is changed.  
 The indication of the end-of-line format of the file visited by a  
 buffer is now more explicit when that format is not the usual one for  
 your operating system.  For example, the DOS-style end-of-line format  
 is displayed as "(DOS)" on Unix and GNU/Linux systems.  The usual  
 end-of-line format is still displayed as a single character (colon for  
 Unix, backslash for DOS and Windows, and forward slash for the Mac).  
   
 The values of the variables eol-mnemonic-unix, eol-mnemonic-dos,  
 eol-mnemonic-mac, and eol-mnemonic-undecided, which are strings,  
 control what is displayed in the mode line for each end-of-line  
 format.  You can now customize these variables.  
   
 ** In the previous version of Emacs, tar-mode didn't work well if a  
 filename contained non-ASCII characters.  Now this is fixed.  Such a  
 filename is decoded by file-name-coding-system if the default value of  
 enable-multibyte-characters is non-nil.  
   
 ** The command temp-buffer-resize-mode toggles a minor mode  
 in which temporary buffers (such as help buffers) are given  
 windows just big enough to hold the whole contents.  
   
 ** If you use completion.el, you must now run the function  
 dynamic-completion-mode to enable it.  Just loading the file  
 doesn't have any effect.  
   
 ** In Flyspell mode, the default is now to make just one Ispell process,  
 not one per buffer.  
   
 ** If you use iswitchb but do not call (iswitchb-default-keybindings) to  
 use the default keybindings, you will need to add the following line:  
   (add-hook 'minibuffer-setup-hook 'iswitchb-minibuffer-setup)  
   
 ** Auto-show mode is no longer enabled just by loading auto-show.el.  
 To control it, set `auto-show-mode' via Custom or use the  
 `auto-show-mode' command.  
   
 ** Handling of X fonts' ascent/descent parameters has been changed to  
 avoid redisplay problems.  As a consequence, compared with previous  
 versions the line spacing and frame size now differ with some font  
 choices, typically increasing by a pixel per line.  This change  
 occurred in version 20.3 but was not documented then.  
   
 ** If you select the bar cursor style, it uses the frame's  
 cursor-color, rather than the cursor foreground pixel.  
   
 ** In multibyte mode, Rmail decodes incoming MIME messages using the  
 character set specified in the message.  If you want to disable this  
 feature, set the variable rmail-decode-mime-charset to nil.  
   
 ** Not new, but not mentioned previously in NEWS: when you use #! at  
 the beginning of a file to make it executable and specify an  
 interpreter program, Emacs looks on the second line for the -*- mode  
 and variable specification, as well as on the first line.  
   
 ** Support for IBM codepage encoding of non-ASCII characters.  
   
 The new command M-x codepage-setup creates a special coding system  
 that can be used to convert text between a specific IBM codepage and  
 one of the character sets built into Emacs which matches that  
 codepage.  For example, codepage 850 corresponds to Latin-1 character  
 set, codepage 855 corresponds to Cyrillic-ISO character set, etc.  
   
 Windows codepages 1250, 1251 and some others, where Windows deviates  
 from the corresponding ISO character set, are also supported.  
   
 IBM box-drawing characters and other glyphs which don't have  
 equivalents in the corresponding ISO character set, are converted to  
 a character defined by dos-unsupported-char-glyph on MS-DOS, and to  
 `?' on other systems.  
   
 IBM codepages are widely used on MS-DOS and MS-Windows, so this  
 feature is most useful on those platforms, but it can also be used on  
 Unix.  
   
 Emacs compiled for MS-DOS automatically loads the support for the  
 current codepage when it starts.  
   
 ** Mail changes  
   
 *** When mail is sent using compose-mail (C-x m), and if  
 `mail-send-nonascii' is set to the new default value `mime',  
 appropriate MIME headers are added.  The headers are added only if  
 non-ASCII characters are present in the body of the mail, and no other  
 MIME headers are already present.  For example, the following three  
 headers are added if the coding system used in the *mail* buffer is  
 latin-1:  
   
   MIME-version: 1.0  
   Content-type: text/plain; charset=iso-8859-1  
   Content-Transfer-Encoding: 8bit  
   
 *** The new variable default-sendmail-coding-system specifies the  
 default way to encode outgoing mail.  This has higher priority than  
 default-buffer-file-coding-system but has lower priority than  
 sendmail-coding-system and the local value of  
 buffer-file-coding-system.  
   
 You should not set this variable manually.  Instead, set  
 sendmail-coding-system to specify a fixed encoding for all outgoing  
 mail.  
   
 *** When you try to send a message that contains non-ASCII characters,  
 if the coding system specified by those variables doesn't handle them,  
 Emacs will ask you to select a suitable coding system while showing a  
 list of possible coding systems.  
   
 ** CC Mode changes  
   
 *** c-default-style can now take an association list that maps major  
 modes to style names.  When this variable is an alist, Java mode no  
 longer hardcodes a setting to "java" style.  See the variable's  
 docstring for details.  
   
 *** It's now possible to put a list as the offset on a syntactic  
 symbol.  The list is evaluated recursively until a non-nil offset is  
 found.  This is useful to combine several lineup functions to act in a  
 prioritized order on a single line.  However, none of the supplied  
 lineup functions use this feature currently.  
   
 *** New syntactic symbol catch-clause, which is used on the "catch" and  
 "finally" lines in try-catch constructs in C++ and Java.  
   
 *** New cleanup brace-catch-brace on c-cleanup-list, which does for  
 "catch" lines what brace-elseif-brace does for "else if" lines.  
   
 *** The braces of Java anonymous inner classes are treated separately  
 from the braces of other classes in auto-newline mode.  Two new  
 symbols inexpr-class-open and inexpr-class-close may be used on  
 c-hanging-braces-alist to control the automatic newlines used for  
 anonymous classes.  
   
 *** Support for the Pike language added, along with new Pike specific  
 syntactic symbols: inlambda, lambda-intro-cont  
   
 *** Support for Java anonymous classes via new syntactic symbol  
 inexpr-class.  New syntactic symbol inexpr-statement for Pike  
 support and gcc-style statements inside expressions.  New lineup  
 function c-lineup-inexpr-block.  
   
 *** New syntactic symbol brace-entry-open which is used in brace lists  
 (i.e. static initializers) when a list entry starts with an open  
 brace.  These used to be recognized as brace-list-entry's.  
 c-electric-brace also recognizes brace-entry-open braces  
 (brace-list-entry's can no longer be electrified).  
   
 *** New command c-indent-line-or-region, not bound by default.  
   
 *** `#' is only electric when typed in the indentation of a line.  
   
 *** Parentheses are now electric (via the new command c-electric-paren)  
 for auto-reindenting lines when parens are typed.  
   
 *** In "gnu" style, inline-open offset is now set to zero.  
   
 *** Uniform handling of the inclass syntactic symbol.  The indentation  
 associated with it is now always relative to the class opening brace.  
 This means that the indentation behavior has changed in some  
 circumstances, but only if you've put anything besides 0 on the  
 class-open syntactic symbol (none of the default styles do that).  
   
 ** Gnus changes.  
   
 *** 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.  
   
 *** Canceling 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.  
   
 *** `gnus-posting-styles' has been re-activated.  
   
 ** Changes to TeX and LaTeX mode  
   
 *** The new variable `tex-start-options-string' can be used to give  
 options for the TeX run.  The default value causes TeX to run in  
 nonstopmode.  For an interactive TeX run set it to nil or "".  
   
 *** The command `tex-feed-input' sends input to the Tex Shell.  In a  
 TeX buffer it is bound to the keys C-RET, C-c RET, and C-c C-m (some  
 of these keys may not work on all systems).  For instance, if you run  
 TeX interactively and if the TeX run stops because of an error, you  
 can continue it without leaving the TeX buffer by typing C-RET.  
   
 *** The Tex Shell Buffer is now in `compilation-shell-minor-mode'.  
 All error-parsing commands of the Compilation major mode are available  
 but bound to keys that don't collide with the shell.  Thus you can use  
 the Tex Shell for command line executions like a usual shell.  
   
 *** The commands `tex-validate-region' and `tex-validate-buffer' check  
 the matching of braces and $'s.  The errors are listed in a *Occur*  
 buffer and you can use C-c C-c or mouse-2 to go to a particular  
 mismatch.  
   
 ** Changes to RefTeX mode  
   
 *** The table of contents buffer can now also display labels and  
 file boundaries in addition to sections. Use `l', `i', and `c' keys.  
   
 *** Labels derived from context (the section heading) are now  
 lowercase by default.  To make the label legal in LaTeX, latin-1  
 characters will lose their accent.  All Mule characters will be  
 removed from the label.  
   
 *** The automatic display of cross reference information can also use  
 a window instead of the echo area.  See variable `reftex-auto-view-crossref'.  
   
 *** kpsewhich can be used by RefTeX to find TeX and BibTeX files.  See the  
 customization group `reftex-finding-files'.  
   
 *** The option `reftex-bibfile-ignore-list' has been renamed to  
 `reftex-bibfile-ignore-regexps' and indeed can be fed with regular  
 expressions.  
   
 *** Multiple Selection buffers are now hidden buffers.  
   
 ** New/deleted modes and packages  
   
 *** The package snmp-mode.el provides major modes for editing SNMP and  
 SNMPv2 MIBs.  It has entries on `auto-mode-alist'.  
   
 *** The package sql.el provides a major mode, M-x sql-mode, for  
 editing SQL files, and M-x sql-interactive-mode for interacting with  
 SQL interpreters.  It has an entry on `auto-mode-alist'.  
   
 *** M-x highlight-changes-mode provides a minor mode displaying buffer  
 changes with a special face.  
   
 *** ispell4.el has been deleted.  It got in the way of ispell.el and  
 this was hard to fix reliably.  It has long been obsolete -- use  
 Ispell 3.1 and ispell.el.  
   
 * MS-DOS changes in Emacs 20.4  
   
 ** Emacs compiled for MS-DOS now supports MULE features better.  
 This includes support for display of all ISO 8859-N character sets,  
 conversion to and from IBM codepage encoding of non-ASCII characters,  
 and automatic setup of the MULE environment at startup.  For details,  
 check out the section `MS-DOS and MULE' in the manual.  
   
 The MS-DOS installation procedure automatically configures and builds  
 Emacs with input method support if it finds an unpacked Leim  
 distribution when the config.bat script is run.  
   
 ** Formerly, the value of lpr-command did not affect printing on  
 MS-DOS unless print-region-function was set to nil, but now it  
 controls whether an external program is invoked or output is written  
 directly to a printer port.  Similarly, in the previous version of  
 Emacs, the value of ps-lpr-command did not affect PostScript printing  
 on MS-DOS unless ps-printer-name was set to something other than a  
 string (eg. t or `pipe'), but now it controls whether an external  
 program is used.  (These changes were made so that configuration of  
 printing variables would be almost identical across all platforms.)  
   
 ** In the previous version of Emacs, PostScript and non-PostScript  
 output was piped to external programs, but because most print programs  
 available for MS-DOS and MS-Windows cannot read data from their standard  
 input, on those systems the data to be output is now written to a  
 temporary file whose name is passed as the last argument to the external  
 program.  
   
 An exception is made for `print', a standard program on Windows NT,  
 and `nprint', a standard program on Novell Netware.  For both of these  
 programs, the command line is constructed in the appropriate syntax  
 automatically, using only the value of printer-name or ps-printer-name  
 as appropriate--the value of the relevant `-switches' variable is  
 ignored, as both programs have no useful switches.  
   
 ** The value of the variable dos-printer (cf. dos-ps-printer), if it has  
 a value, overrides the value of printer-name (cf. ps-printer-name), on  
 MS-DOS and MS-Windows only.  This has been true since version 20.3, but  
 was not documented clearly before.  
   
 ** All the Emacs games now work on MS-DOS terminals.  
 This includes Tetris and Snake.  
   
 * Lisp changes in Emacs 20.4  
   
 ** New functions line-beginning-position and line-end-position  
 return the position of the beginning or end of the current line.  
 They both accept an optional argument, which has the same  
 meaning as the argument to beginning-of-line or end-of-line.  
   
 ** find-file and allied functions now have an optional argument  
 WILDCARD.  If this is non-nil, they do wildcard processing,  
 and visit all files that match the wildcard pattern.  
   
 ** Changes in the file-attributes function.  
   
 *** The file size returned by file-attributes may be an integer or a float.  
 It is an integer if the size fits in a Lisp integer, float otherwise.  
   
 *** The inode number returned by file-attributes may be an integer (if  
 the number fits in a Lisp integer) or a cons cell containing two  
 integers.  
   
 ** The new function directory-files-and-attributes returns a list of  
 files in a directory and their attributes.  It accepts the same  
 arguments as directory-files and has similar semantics, except that  
 file names and attributes are returned.  
   
 ** The new function file-attributes-lessp is a helper function for  
 sorting the list generated by directory-files-and-attributes.  It  
 accepts two arguments, each a list of a file name and its attributes.  
 It compares the file names of each according to string-lessp and  
 returns the result.  
   
 ** The new function file-expand-wildcards expands a wildcard-pattern  
 to produce a list of existing files that match the pattern.  
   
 ** New functions for base64 conversion:  
   
 The function base64-encode-region converts a part of the buffer  
 into the base64 code used in MIME.  base64-decode-region  
 performs the opposite conversion.  Line-breaking is supported  
 optionally.  
   
 Functions base64-encode-string and base64-decode-string do a similar  
 job on the text in a string.  They return the value as a new string.  
   
 **  
 The new function process-running-child-p  
 will tell you if a subprocess has given control of its  
 terminal to its own child process.  
   
 ** interrupt-process and such functions have a new feature:  
 when the second argument is `lambda', they send a signal  
 to the running child of the subshell, if any, but if the shell  
 itself owns its terminal, no signal is sent.  
   
 ** There are new widget types `plist' and `alist' which can  
 be used for customizing variables whose values are plists or alists.  
   
 ** easymenu.el now understands `:key-sequence' and `:style button'.  
 :included is an alias for :visible.  
   
 easy-menu-add-item now understands the values returned by  
 easy-menu-remove-item and easy-menu-item-present-p.  This can be used  
 to move or copy menu entries.  
   
 ** Multibyte editing changes  
   
 *** The definitions of sref and char-bytes are changed.  Now, sref is  
 an alias of aref and char-bytes always returns 1.  This change is to  
 make some Emacs Lisp code which works on 20.2 and earlier also  
 work on the latest Emacs.  Such code uses a combination of sref and  
 char-bytes in a loop typically as below:  
         (setq char (sref str idx)  
               idx (+ idx (char-bytes idx)))  
 The byte-compiler now warns that this is obsolete.  
   
 If you want to know how many bytes a specific multibyte character  
 (say, CH) occupies in a multibyte buffer, use this code:  
         (charset-bytes (char-charset ch))  
   
 *** In multibyte mode, when you narrow a buffer to some region, and the  
 region is preceded or followed by non-ASCII codes, inserting or  
 deleting at the head or the end of the region may signal this error:  
   
     Byte combining across boundary of accessible buffer text inhibited  
   
 This is to avoid some bytes being combined together into a character  
 across the boundary.  
   
 *** The functions find-charset-region and find-charset-string include  
 `unknown' in the returned list in the following cases:  
     o The current buffer or the target string is unibyte and  
       contains 8-bit characters.  
     o The current buffer or the target string is multibyte and  
       contains invalid characters.  
   
 *** The functions decode-coding-region and encode-coding-region remove  
 text properties of the target region.  Ideally, they should correctly  
 preserve text properties, but for the moment, it's hard.  Removing  
 text properties is better than preserving them in a less-than-correct  
 way.  
   
 *** prefer-coding-system sets EOL conversion of default coding systems.  
 If the argument to prefer-coding-system specifies a certain type of  
 end of line conversion, the default coding systems set by  
 prefer-coding-system will specify that conversion type for end of line.  
   
 *** The new function thai-compose-string can be used to properly  
 compose Thai characters in a string.  
   
 ** The primitive `define-prefix-command' now takes an optional third  
 argument NAME, which should be a string.  It supplies the menu name  
 for the created keymap.  Keymaps created in order to be displayed as  
 menus should always use the third argument.  
   
 ** The meanings of optional second arguments for read-char,  
 read-event, and read-char-exclusive are flipped.  Now the second  
 arguments are INHERIT-INPUT-METHOD.  These functions use the current  
 input method (if any) if and only if INHERIT-INPUT-METHOD is non-nil.  
   
 ** The new function clear-this-command-keys empties out the contents  
 of the vector that (this-command-keys) returns.  This is useful in  
 programs that read passwords, to prevent the passwords from echoing  
 inadvertently as part of the next command in certain cases.  
   
 ** The new macro `with-temp-message' displays a temporary message in  
 the echo area, while executing some Lisp code.  Like `progn', it  
 returns the value of the last form, but it also restores the previous  
 echo area contents.  
   
    (with-temp-message MESSAGE &rest BODY)  
   
 ** The function `require' now takes an optional third argument  
 NOERROR.  If it is non-nil, then there is no error if the  
 requested feature cannot be loaded.  
   
 ** In the function modify-face, an argument of (nil) for the  
 foreground color, background color or stipple pattern  
 means to clear out that attribute.  
   
 ** The `outer-window-id' frame property of an X frame  
 gives the window number of the outermost X window for the frame.  
   
 ** Temporary buffers made with with-output-to-temp-buffer are now  
 read-only by default, and normally use the major mode Help mode  
 unless you put them in some other non-Fundamental mode before the  
 end of with-output-to-temp-buffer.  
   
 ** The new functions gap-position and gap-size return information on  
 the gap of the current buffer.  
   
 ** The new functions position-bytes and byte-to-position provide a way  
 to convert between character positions and byte positions in the  
 current buffer.  
   
 ** vc.el defines two new macros, `edit-vc-file' and `with-vc-file', to  
 facilitate working with version-controlled files from Lisp programs.  
 These macros check out a given file automatically if needed, and check  
 it back in after any modifications have been made.  
   
 * Installation Changes in Emacs 20.3  
   
 ** The default value of load-path now includes most subdirectories of  
 the site-specific directories /usr/local/share/emacs/site-lisp and  
 /usr/local/share/emacs/VERSION/site-lisp, in addition to those  
 directories themselves.  Both immediate subdirectories and  
 subdirectories multiple levels down are added to load-path.  
   
 Not all subdirectories are included, though.  Subdirectories whose  
 names do not start with a letter or digit are excluded.  
 Subdirectories named RCS or CVS are excluded.  Also, a subdirectory  
 which contains a file named `.nosearch' is excluded.  You can use  
 these methods to prevent certain subdirectories from being searched.  
   
 Emacs finds these subdirectories and adds them to load-path when it  
 starts up.  While it would be cleaner to find the subdirectories each  
 time Emacs loads a file, that would be much slower.  
   
 This feature is an incompatible change.  If you have stored some Emacs  
 Lisp files in a subdirectory of the site-lisp directory specifically  
 to prevent them from being used, you will need to rename the  
 subdirectory to start with a non-alphanumeric character, or create a  
 `.nosearch' file in it, in order to continue to achieve the desired  
 results.  
   
 ** Emacs no longer includes an old version of the C preprocessor from  
 GCC.  This was formerly used to help compile Emacs with C compilers  
 that had limits on the significant length of an identifier, but in  
 fact we stopped supporting such compilers some time ago.  
   
 * Changes in Emacs 20.3  
   
 ** The new command C-x z (repeat) repeats the previous command  
 including its argument.  If you repeat the z afterward,  
 it repeats the command additional times; thus, you can  
 perform many repetitions with one keystroke per repetition.  
   
 ** Emacs now supports "selective undo" which undoes only within a  
 specified region.  To do this, set point and mark around the desired  
 region and type C-u C-x u (or C-u C-_).  You can then continue undoing  
 further, within the same region, by repeating the ordinary undo  
 command C-x u or C-_.  This will keep undoing changes that were made  
 within the region you originally specified, until either all of them  
 are undone, or it encounters a change which crosses the edge of that  
 region.  
   
 In Transient Mark mode, undoing when a region is active requests  
 selective undo.  
   
 ** If you specify --unibyte when starting Emacs, then all buffers are  
 unibyte, except when a Lisp program specifically creates a multibyte  
 buffer.  Setting the environment variable EMACS_UNIBYTE has the same  
 effect.  The --no-unibyte option overrides EMACS_UNIBYTE and directs  
 Emacs to run normally in multibyte mode.  
   
 The option --unibyte does not affect the reading of Emacs Lisp files,  
 though.  If you want a Lisp file to be read in unibyte mode, use  
 -*-unibyte: t;-*- on its first line.  That will force Emacs to  
 load that file in unibyte mode, regardless of how Emacs was started.  
   
 ** toggle-enable-multibyte-characters no longer has a key binding and  
 no longer appears in the menu bar.  We've realized that changing the  
 enable-multibyte-characters variable in an existing buffer is  
 something that most users not do.  
   
 ** You can specify a coding system to use for the next cut or paste  
 operations through the window system with the command C-x RET X.  
 The coding system can make a difference for communication with other  
 applications.  
   
 C-x RET x specifies a coding system for all subsequent cutting and  
 pasting operations.  
   
 ** You can specify the printer to use for commands that do printing by  
 setting the variable `printer-name'.  Just what a printer name looks  
 like depends on your operating system.  You can specify a different  
 printer for the Postscript printing commands by setting  
 `ps-printer-name'.  
   
 ** Emacs now supports on-the-fly spell checking by the means of a  
 minor mode.  It is called M-x flyspell-mode.  You don't have to remember  
 any other special commands to use it, and you will hardly notice it  
 except when you make a spelling error.  Flyspell works by highlighting  
 incorrect words as soon as they are completed or as soon as the cursor  
 hits a new word.  
   
 Flyspell mode works with whichever dictionary you have selected for  
 Ispell in Emacs.  In TeX mode, it understands TeX syntax so as not  
 to be confused by TeX commands.  
   
 You can correct a misspelled word by editing it into something  
 correct.  You can also correct it, or accept it as correct, by  
 clicking on the word with Mouse-2; that gives you a pop-up menu  
 of various alternative replacements and actions.  
   
 Flyspell mode also proposes "automatic" corrections.  M-TAB replaces  
 the current misspelled word with a possible correction.  If several  
 corrections are made possible, M-TAB cycles through them in  
 alphabetical order, or in order of decreasing likelihood if  
 flyspell-sort-corrections is nil.  
   
 Flyspell mode also flags an error when a word is repeated, if  
 flyspell-mark-duplications-flag is non-nil.  
   
 ** Changes in input method usage.  
   
 Now you can use arrow keys (right, left, down, up) for selecting among  
 the alternatives just the same way as you do by C-f, C-b, C-n, and C-p  
 respectively.  
   
 You can use the ENTER key to accept the current conversion.  
   
 If you type TAB to display a list of alternatives, you can select one  
 of the alternatives with Mouse-2.  
   
 The meaning of the variable `input-method-verbose-flag' is changed so  
 that you can set it to t, nil, `default', or `complex-only'.  
   
   If the value is nil, extra guidance is never given.  
   
   If the value is t, extra guidance is always given.  
   
   If the value is `complex-only', extra guidance is always given only  
   when you are using complex input methods such as chinese-py.  
   
   If the value is `default' (this is the default), extra guidance is  
   given in the following case:  
     o When you are using a complex input method.  
     o When you are using a simple input method but not in the minibuffer.  
   
 If you are using Emacs through a very slow line, setting  
 input-method-verbose-flag to nil or to complex-only is a good choice,  
 and if you are using an input method you are not familiar with,  
 setting it to t is helpful.  
   
 The old command select-input-method is now called set-input-method.  
   
 In the language environment "Korean", you can use the following  
 keys:  
         Shift-SPC       toggle-korean-input-method  
         C-F9            quail-hangul-switch-symbol-ksc  
         F9              quail-hangul-switch-hanja  
 These key bindings are canceled when you switch to another language  
 environment.  
   
 ** The minibuffer history of file names now records the specified file  
 names, not the entire minibuffer input.  For example, if the  
 minibuffer starts out with /usr/foo/, you might type in /etc/passwd to  
 get  
   
      /usr/foo//etc/passwd  
   
 which stands for the file /etc/passwd.  
   
 Formerly, this used to put /usr/foo//etc/passwd in the history list.  
 Now this puts just /etc/passwd in the history list.  
   
 ** If you are root, Emacs sets backup-by-copying-when-mismatch to t  
 at startup, so that saving a file will be sure to preserve  
 its owner and group.  
   
 ** find-func.el can now also find the place of definition of Emacs  
 Lisp variables in user-loaded libraries.  
   
 ** C-x r t (string-rectangle) now deletes the existing rectangle  
 contents before inserting the specified string on each line.  
   
 ** There is a new command delete-whitespace-rectangle  
 which deletes whitespace starting from a particular column  
 in all the lines on a rectangle.  The column is specified  
 by the left edge of the rectangle.  
   
 ** 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 new command M-x speedbar displays a frame in which directories,  
 files, and tags can be displayed, manipulated, and jumped to.  The  
 frame defaults to 20 characters in width, and is the same height as  
 the frame that it was started from.  Some major modes define  
 additional commands for the speedbar, including Rmail, GUD/GDB, and  
 info.  
   
 ** query-replace-regexp is now bound to C-M-%.  
   
 ** In Transient Mark mode, when the region is active, M-x  
 query-replace and the other replace commands now operate on the region  
 contents only.  
   
 ** M-x write-region, when used interactively, now asks for  
 confirmation before overwriting an existing file.  When you call  
 the function from a Lisp program, a new optional argument CONFIRM  
 says whether to ask for confirmation in this case.  
   
 ** If you use find-file-literally and the file is already visited  
 non-literally, the command asks you whether to revisit the file  
 literally.  If you say no, it signals an error.  
   
 ** Major modes defined with the "derived mode" feature  
 now use the proper name for the mode hook: WHATEVER-mode-hook.  
 Formerly they used the name WHATEVER-mode-hooks, but that is  
 inconsistent with Emacs conventions.  
   
 ** shell-command-on-region (and shell-command) reports success or  
 failure if the command produces no output.  
   
 ** Set focus-follows-mouse to nil if your window system or window  
 manager does not transfer focus to another window when you just move  
 the mouse.  
   
 ** mouse-menu-buffer-maxlen has been renamed to  
 mouse-buffer-menu-maxlen to be consistent with the other related  
 function and variable names.  
   
 ** The new variable auto-coding-alist specifies coding systems for  
 reading specific files.  This has higher priority than  
 file-coding-system-alist.  
   
 ** If you set the variable unibyte-display-via-language-environment to  
 t, then Emacs displays non-ASCII characters are displayed by  
 converting them to the equivalent multibyte characters according to  
 the current language environment.  As a result, they are displayed  
 according to the current fontset.  
   
 ** C-q's handling of codes in the range 0200 through 0377 is changed.  
   
 The codes in the range 0200 through 0237 are inserted as one byte of  
 that code regardless of the values of nonascii-translation-table and  
 nonascii-insert-offset.  
   
 For the codes in the range 0240 through 0377, if  
 enable-multibyte-characters is non-nil and nonascii-translation-table  
 nor nonascii-insert-offset can't convert them to valid multibyte  
 characters, they are converted to Latin-1 characters.  
   
 ** If you try to find a file that is not read-accessible, you now get  
 an error, rather than an empty buffer and a warning.  
   
 ** In the minibuffer history commands M-r and M-s, an upper case  
 letter in the regular expression forces case-sensitive search.  
   
 ** In the *Help* buffer, cross-references to commands and variables  
 are inferred and hyperlinked.  Use C-h m in Help mode for the relevant  
 command keys.  
   
 ** M-x apropos-command, with a prefix argument, no longer looks for  
 user option variables--instead it looks for noninteractive functions.  
   
 Meanwhile, the command apropos-variable normally searches for  
 user option variables; with a prefix argument, it looks at  
 all variables that have documentation.  
   
 ** When you type a long line in the minibuffer, and the minibuffer  
 shows just one line, automatically scrolling works in a special way  
 that shows you overlap with the previous line of text.  The variable  
 minibuffer-scroll-overlap controls how many characters of overlap  
 it should show; the default is 20.  
   
 Meanwhile, Resize Minibuffer mode is still available; in that mode,  
 the minibuffer grows taller (up to a point) as needed to show the whole  
 of your input.  
   
 ** The new command M-x customize-changed-options lets you customize  
 all the options whose meanings or 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 which were changed since that version.  
 Newly added options are included as well.  
   
 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.  
   
 This function is also bound to the Changed Options entry in the  
 Customize menu.  
   
 ** When you run M-x grep with a prefix argument, it figures out  
 the tag around point and puts that into the default grep command.  
   
 ** The new command M-* (pop-tag-mark) pops back through a history of  
 buffer positions from which M-. or other tag-finding commands were  
 invoked.  
   
 ** The new variable comment-padding specifies the number of spaces  
 that `comment-region' will insert before the actual text of the comment.  
 The default is 1.  
   
 ** In Fortran mode the characters `.', `_' and `$' now have symbol  
 syntax, not word syntax.  Fortran mode now supports `imenu' and has  
 new commands fortran-join-line (M-^) and fortran-narrow-to-subprogram  
 (C-x n d).  M-q can be used to fill a statement or comment block  
 sensibly.  
   
 ** GUD now supports jdb, the Java debugger, and pdb, the Python debugger.  
   
 ** If you set the variable add-log-keep-changes-together to a non-nil  
 value, the command `C-x 4 a' will automatically notice when you make  
 two entries in one day for one file, and combine them.  
   
 ** You can use the command M-x diary-mail-entries to mail yourself a  
 reminder about upcoming diary entries.  See the documentation string  
 for a sample shell script for calling this function automatically  
 every night.  
   
 ** Desktop changes  
   
 *** All you need to do to enable use of the Desktop package, is to set  
 the variable desktop-enable to t with Custom.  
   
 *** Minor modes are now restored.  Which minor modes are restored  
 and how modes are restored is controlled by `desktop-minor-mode-table'.  
   
 ** There is no need to do anything special, now, to enable Gnus to  
 read and post multi-lingual articles.  
   
 ** Outline mode has now support for showing hidden outlines when  
 doing an isearch.  In order for this to happen search-invisible should  
 be set to open (the default).  If an isearch match is inside a hidden  
 outline the outline is made visible.  If you continue pressing C-s and  
 the match moves outside the formerly invisible outline, the outline is  
 made invisible again.  
   
 ** Mail reading and sending changes  
   
 *** The Rmail e command now switches to displaying the whole header of  
 the message before it lets you edit the message.  This is so that any  
 changes you make in the header will not be lost if you subsequently  
 toggle.  
   
 *** The w command in Rmail, which writes the message body into a file,  
 now works in the summary buffer as well.  (The command to delete the  
 summary buffer is now Q.)  The default file name for the w command, if  
 the message has no subject, is stored in the variable  
 rmail-default-body-file.  
   
 *** Most of the commands and modes that operate on mail and netnews no  
 longer depend on the value of mail-header-separator.  Instead, they  
 handle whatever separator the buffer happens to use.  
   
 *** If you set mail-signature to a value which is not t, nil, or a string,  
 it should be an expression.  When you send a message, this expression  
 is evaluated to insert the signature.  
   
 *** The new Lisp library feedmail.el (version 8) enhances processing of  
 outbound email messages.  It works in coordination with other email  
 handling packages (e.g., rmail, VM, gnus) and is responsible for  
 putting final touches on messages and actually submitting them for  
 transmission.  Users of the emacs program "fakemail" might be  
 especially interested in trying feedmail.  
   
 feedmail is not enabled by default.  See comments at the top of  
 feedmail.el for set-up instructions.  Among the bigger features  
 provided by feedmail are:  
   
 **** you can park outgoing messages into a disk-based queue and  
 stimulate sending some or all of them later (handy for laptop users);  
 there is also a queue for draft messages  
   
 **** you can get one last look at the prepped outbound message and  
 be prompted for confirmation  
   
 **** does smart filling of address headers  
   
 **** can generate a MESSAGE-ID: line and a DATE: line; the date can be  
 the time the message was written or the time it is being sent; this  
 can make FCC copies more closely resemble copies that recipients get  
   
 **** you can specify an arbitrary function for actually transmitting  
 the message; included in feedmail are interfaces for /bin/[r]mail,  
 /usr/lib/sendmail, and elisp smtpmail; it's easy to write a new  
 function for something else (10-20 lines of elisp)  
   
 ** Dired changes  
   
 *** The Dired function dired-do-toggle, which toggles marked and unmarked  
 files, is now bound to "t" instead of "T".  
   
 *** dired-at-point has been added to ffap.el.  It allows one to easily  
 run Dired on the directory name at point.  
   
 *** Dired has a new command: %g.  It searches the contents of  
 files in the directory and marks each file that contains a match  
 for a specified regexp.  
   
 ** VC Changes  
   
 *** New option vc-ignore-vc-files lets you turn off version control  
 conveniently.  
   
 *** VC Dired has been completely rewritten.  It is now much  
 faster, especially for CVS, and works very similar to ordinary  
 Dired.  
   
 VC Dired is invoked by typing C-x v d and entering the name of the  
 directory to display.  By default, VC Dired gives you a recursive  
 listing of all files at or below the given directory which are  
 currently locked (for CVS, all files not up-to-date are shown).  
   
 You can change the listing format by setting vc-dired-recurse to nil,  
 then it shows only the given directory, and you may also set  
 vc-dired-terse-display to nil, then it shows all files under version  
 control plus the names of any subdirectories, so that you can type `i'  
 on such lines to insert them manually, as in ordinary Dired.  
   
 All Dired commands operate normally in VC Dired, except for `v', which  
 is redefined as the version control prefix.  That means you may type  
 `v l', `v =' etc. to invoke `vc-print-log', `vc-diff' and the like on  
 the file named in the current Dired buffer line.  `v v' invokes  
 `vc-next-action' on this file, or on all files currently marked.  
   
 The new command `v t' (vc-dired-toggle-terse-mode) allows you to  
 toggle between terse display (only locked files) and full display (all  
 VC files plus subdirectories).  There is also a special command,  
 `* l', to mark all files currently locked.  
   
 Giving a prefix argument to C-x v d now does the same thing as in  
 ordinary Dired: it allows you to supply additional options for the ls  
 command in the minibuffer, to fine-tune VC Dired's output.  
   
 *** Under CVS, if you merge changes from the repository into a working  
 file, and CVS detects conflicts, VC now offers to start an ediff  
 session to resolve them.  
   
 Alternatively, you can use the new command `vc-resolve-conflicts' to  
 resolve conflicts in a file at any time.  It works in any buffer that  
 contains conflict markers as generated by rcsmerge (which is what CVS  
 uses as well).  
   
 *** You can now transfer changes between branches, using the new  
 command vc-merge (C-x v m).  It is implemented for RCS and CVS.  When  
 you invoke it in a buffer under version-control, you can specify  
 either an entire branch or a pair of versions, and the changes on that  
 branch or between the two versions are merged into the working file.  
 If this results in any conflicts, they may be resolved interactively,  
 using ediff.  
   
 ** Changes in Font Lock  
   
 *** The face and variable previously known as font-lock-reference-face  
 are now called font-lock-constant-face to better reflect their typical  
 use for highlighting constants and labels.  (Its face properties are  
 unchanged.)  The variable font-lock-reference-face remains for now for  
 compatibility reasons, but its value is font-lock-constant-face.  
   
 ** Frame name display changes  
   
 *** The command set-frame-name lets you set the name of the current  
 frame.  You can use the new command select-frame-by-name to select and  
 raise a frame; this is mostly useful on character-only terminals, or  
 when many frames are invisible or iconified.  
   
 *** On character-only terminal (not a window system), changing the  
 frame name is now reflected on the mode line and in the Buffers/Frames  
 menu.  
   
 ** Comint (subshell) changes  
   
 *** In Comint modes, the commands to kill, stop or interrupt a  
 subjob now also kill pending input.  This is for compatibility  
 with ordinary shells, where the signal characters do this.  
   
 *** There are new commands in Comint mode.  
   
 C-c C-x fetches the "next" line from the input history;  
 that is, the line after the last line you got.  
 You can use this command to fetch successive lines, one by one.  
   
 C-c SPC accumulates lines of input.  More precisely, it arranges to  
 send the current line together with the following line, when you send  
 the following line.  
   
 C-c C-a if repeated twice consecutively now moves to the process mark,  
 which separates the pending input from the subprocess output and the  
 previously sent input.  
   
 C-c M-r now runs comint-previous-matching-input-from-input;  
 it searches for a previous command, using the current pending input  
 as the search string.  
   
 *** New option compilation-scroll-output can be set to scroll  
 automatically in compilation-mode windows.  
   
 ** 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 settings 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.  
   
 ** Changes to hippie-expand.  
   
 *** New customization variable `hippie-expand-dabbrev-skip-space'. If  
 non-nil, trailing spaces may be included in the abbreviation to search for,  
 which then gives the same behavior as the original `dabbrev-expand'.  
   
 *** New customization variable `hippie-expand-dabbrev-as-symbol'. If  
 non-nil, characters of syntax '_' is considered part of the word when  
 expanding dynamically.  
   
 *** New customization variable `hippie-expand-no-restriction'. If  
 non-nil, narrowed buffers are widened before they are searched.  
   
 *** New customization variable `hippie-expand-only-buffers'. If  
 non-empty, buffers searched are restricted to the types specified in  
 this list. Useful for example when constructing new special-purpose  
 expansion functions with `make-hippie-expand-function'.  
   
 *** Text properties of the expansion are no longer copied.  
   
 ** Changes in BibTeX mode.  
   
 *** Any titleword matching a regexp in the new variable  
 bibtex-autokey-titleword-ignore (case sensitive) is ignored during  
 automatic key generation.  This replaces variable  
 bibtex-autokey-titleword-first-ignore, which only checked for matches  
 against the first word in the title.  
   
 *** Autokey generation now uses all words from the title, not just  
 capitalized words.  To avoid conflicts with existing customizations,  
 bibtex-autokey-titleword-ignore is set up such that words starting with  
 lowerkey characters will still be ignored.  Thus, if you want to use  
 lowercase words from the title, you will have to overwrite the  
 bibtex-autokey-titleword-ignore standard setting.  
   
 *** Case conversion of names and title words for automatic key  
 generation is more flexible.  Variable bibtex-autokey-preserve-case is  
 replaced by bibtex-autokey-titleword-case-convert and  
 bibtex-autokey-name-case-convert.  
   
 ** Changes in vcursor.el.  
   
 *** Support for character terminals is available: there is a new keymap  
 and the vcursor will appear as an arrow between buffer text.  A  
 variable `vcursor-interpret-input' allows input from the vcursor to be  
 entered exactly as if typed.  Numerous functions, including  
 `vcursor-compare-windows', have been rewritten to improve consistency  
 in the selection of windows and corresponding keymaps.  
   
 *** vcursor options can now be altered with M-x customize under the  
 Editing group once the package is loaded.  
   
 *** Loading vcursor now does not define keys by default, as this is  
 generally a bad side effect.  Use M-x customize to set  
 vcursor-key-bindings to t to restore the old behavior.  
   
 *** vcursor-auto-disable can be `copy', which turns off copying from the  
 vcursor, but doesn't disable it, after any non-vcursor command.  
   
 ** Ispell changes.  
   
 *** You can now spell check comments and strings in the current  
 buffer with M-x ispell-comments-and-strings.  Comments and strings  
 are identified by syntax tables in effect.  
   
 *** Generic region skipping implemented.  
 A single buffer can be broken into a number of regions where text will  
 and will not be checked.  The definitions of the regions can be user  
 defined.  New applications and improvements made available by this  
 include:  
   
     o URLs are automatically skipped  
     o EMail message checking is vastly improved.  
   
 *** Ispell can highlight the erroneous word even on non-window terminals.  
   
 ** Changes to RefTeX mode  
   
 RefTeX has been updated in order to make it more usable with very  
 large projects (like a several volume math book).  The parser has been  
 re-written from scratch.  To get maximum speed from RefTeX, check the  
 section `Optimizations' in the manual.  
   
 *** New recursive parser.  
   
 The old version of RefTeX created a single large buffer containing the  
 entire multifile document in order to parse the document.  The new  
 recursive parser scans the individual files.  
   
 *** Parsing only part of a document.  
   
 Reparsing of changed document parts can now be made faster by enabling  
 partial scans.  To use this feature, read the documentation string of  
 the variable `reftex-enable-partial-scans' and set the variable to t.  
   
     (setq reftex-enable-partial-scans t)  
   
 *** Storing parsing information in a file.  
   
 This can improve startup times considerably.  To turn it on, use  
   
     (setq reftex-save-parse-info t)  
   
 *** Using multiple selection buffers  
   
 If the creation of label selection buffers is too slow (this happens  
 for large documents), you can reuse these buffers by setting  
   
     (setq reftex-use-multiple-selection-buffers t)  
   
 *** References to external documents.  
   
 The LaTeX package `xr' allows to cross-reference labels in external  
 documents.  RefTeX can provide information about the external  
 documents as well.  To use this feature, set up the \externaldocument  
 macros required by the `xr' package and rescan the document with  
 RefTeX.  The external labels can then be accessed with the `x' key in  
 the selection buffer provided by `reftex-reference' (bound to `C-c )').  
 The `x' key also works in the table of contents buffer.  
   
 *** Many more labeled LaTeX environments are recognized by default.  
   
 The built-in command list now covers all the standard LaTeX commands,  
 and all of the major packages included in the LaTeX distribution.  
   
 Also, RefTeX now understands the \appendix macro and changes  
 the enumeration of sections in the *toc* buffer accordingly.  
   
 *** Mouse support for selection and *toc* buffers  
   
 The mouse can now be used to select items in the selection and *toc*  
 buffers.  See also the new option `reftex-highlight-selection'.  
   
 *** New keymaps for selection and table of contents modes.  
   
 The selection processes for labels and citation keys, and the table of  
 contents buffer now have their own keymaps: `reftex-select-label-map',  
 `reftex-select-bib-map', `reftex-toc-map'.  The selection processes  
 have a number of new keys predefined.  In particular, TAB lets you  
 enter a label with completion.  Check the on-the-fly help (press `?'  
 at the selection prompt) or read the Info documentation to find out  
 more.  
   
 *** Support for the varioref package  
   
 The `v' key in the label selection buffer toggles \ref versus \vref.  
   
 *** New hooks  
   
 Three new hooks can be used to redefine the way labels, references,  
 and citations are created. These hooks are  
 `reftex-format-label-function', `reftex-format-ref-function',  
 `reftex-format-cite-function'.  
   
 *** Citations outside LaTeX  
   
 The command `reftex-citation' may also be used outside LaTeX (e.g. in  
 a mail buffer).  See the Info documentation for details.  
   
 *** Short context is no longer fontified.  
   
 The short context in the label menu no longer copies the  
 fontification from the text in the buffer.  If you prefer it to be  
 fontified, use  
   
    (setq reftex-refontify-context t)  
   
 ** file-cache-minibuffer-complete now accepts a prefix argument.  
 With a prefix argument, it does not try to do completion of  
 the file name within its directory; it only checks for other  
 directories that contain the same file name.  
   
 Thus, given the file name Makefile, and assuming that a file  
 Makefile.in exists in the same directory, ordinary  
 file-cache-minibuffer-complete will try to complete Makefile to  
 Makefile.in and will therefore never look for other directories that  
 have Makefile.  A prefix argument tells it not to look for longer  
 names such as Makefile.in, so that instead it will look for other  
 directories--just as if the name were already complete in its present  
 directory.  
   
 ** New modes and packages  
   
 *** There is a new alternative major mode for Perl, Cperl mode.  
 It has many more features than Perl mode, and some people prefer  
 it, but some do not.  
   
 *** There is a new major mode, M-x vhdl-mode, for editing files of VHDL  
 code.  
   
 *** M-x which-function-mode enables a minor mode that displays the  
 current function name continuously in the mode line, as you move  
 around in a buffer.  
   
 Which Function mode is effective in major modes which support Imenu.  
   
 *** Gametree is a major mode for editing game analysis trees.  The author  
 uses it for keeping notes about his postal Chess games, but it should  
 be helpful for other two-player games as well, as long as they have an  
 established system of notation similar to Chess.  
   
 *** The new minor mode checkdoc-minor-mode provides Emacs Lisp  
 documentation string checking for style and spelling.  The style  
 guidelines are found in the Emacs Lisp programming manual.  
   
 *** The net-utils package makes some common networking features  
 available in Emacs.  Some of these functions are wrappers around  
 system utilities (ping, nslookup, etc); others are implementations of  
 simple protocols (finger, whois) in Emacs Lisp.  There are also  
 functions to make simple connections to TCP/IP ports for debugging and  
 the like.  
   
 *** highlight-changes-mode is a minor mode that uses colors to  
 identify recently changed parts of the buffer text.  
   
 *** The new package `midnight' lets you specify things to be done  
 within Emacs at midnight--by default, kill buffers that you have not  
 used in a considerable time.  To use this feature, customize  
 the user option `midnight-mode' to t.  
   
 *** The file generic-x.el defines a number of simple major modes.  
   
   apache-generic-mode: For Apache and NCSA httpd configuration files  
   samba-generic-mode: Samba configuration files  
   fvwm-generic-mode: For fvwm initialization files  
   x-resource-generic-mode: For X resource files  
   hosts-generic-mode: For hosts files (.rhosts, /etc/hosts, etc)  
   mailagent-rules-generic-mode: For mailagent .rules files  
   javascript-generic-mode: For JavaScript files  
   vrml-generic-mode: For VRML files  
   java-manifest-generic-mode: For Java MANIFEST files  
   java-properties-generic-mode: For Java property files  
   mailrc-generic-mode: For .mailrc files  
   
   Platform-specific modes:  
   
   prototype-generic-mode: For Solaris/Sys V prototype files  
   pkginfo-generic-mode: For Solaris/Sys V pkginfo files  
   alias-generic-mode: For C shell alias files  
   inf-generic-mode: For MS-Windows INF files  
   ini-generic-mode: For MS-Windows INI files  
   reg-generic-mode: For MS-Windows Registry files  
   bat-generic-mode: For MS-Windows BAT scripts  
   rc-generic-mode: For MS-Windows Resource files  
   rul-generic-mode: For InstallShield scripts  
   
 * Lisp changes in Emacs 20.3 since the Emacs Lisp Manual was published  
   
 ** If you want a Lisp file to be read in unibyte mode,  
 use -*-unibyte: t;-*- on its first line.  
 That will force Emacs to read that file in unibyte mode.  
 Otherwise, the file will be loaded and byte-compiled in multibyte mode.  
   
 Thus, each lisp file is read in a consistent way regardless of whether  
 you started Emacs with --unibyte, so that a Lisp program gives  
 consistent results regardless of how Emacs was started.  
   
 ** The new function assoc-default is useful for searching an alist,  
 and using a default value if the key is not found there.  You can  
 specify a comparison predicate, so this function is useful for  
 searching comparing a string against an alist of regular expressions.  
   
 ** The functions unibyte-char-to-multibyte and  
 multibyte-char-to-unibyte convert between unibyte and multibyte  
 character codes, in a way that is appropriate for the current language  
 environment.  
   
 ** The functions read-event, read-char and read-char-exclusive now  
 take two optional arguments.  PROMPT, if non-nil, specifies a prompt  
 string.  SUPPRESS-INPUT-METHOD, if non-nil, says to disable the  
 current input method for reading this one event.  
   
 ** Two new variables print-escape-nonascii and print-escape-multibyte  
 now control whether to output certain characters as  
 backslash-sequences.  print-escape-nonascii applies to single-byte  
 non-ASCII characters; print-escape-multibyte applies to multibyte  
 characters.  Both of these variables are used only when printing  
 in readable fashion (prin1 uses them, princ does not).  
   
 * Lisp changes in Emacs 20.3 before the Emacs Lisp Manual was published  
   
 ** Compiled Emacs Lisp files made with the modified "MBSK" version  
 of Emacs 20.2 do not work in Emacs 20.3.  
   
 ** Buffer positions are now measured in characters, as they were  
 in Emacs 19 and before.  This means that (forward-char 1)  
 always increases point by 1.  
   
 The function chars-in-region now just subtracts its arguments.  It is  
 considered obsolete.  The function char-boundary-p has been deleted.  
   
 See below for additional changes relating to multibyte characters.  
   
 ** defcustom, defface and defgroup now accept 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 "20.3")  
   
    (defgroup foo-group nil "The foo group."  
      :version "20.3")  
   
 If an entire new group is added or the variables in it have the  
 default values changed, then just add a `:version' to that group. It  
 is recommended that new packages added to the distribution contain a  
 `:version' in the top level group.  
   
 This information is used to control the customize-changed-options command.  
   
 ** 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 previous Emacs versions by explicitly setting these variables  
 to themselves.  
   
 If you set the variable keyword-symbols-constant-flag to nil,  
 this error is suppressed, and you can set these symbols to any  
 values whatever.  
   
 ** There is a new debugger command, R.  
 It evaluates an expression like e, but saves the result  
 in the buffer *Debugger-record*.  
   
 ** Frame-local variables.  
   
 You can now make a variable local to various frames.  To do this, call  
 the function make-variable-frame-local; this enables frames to have  
 local bindings for that variable.  
   
 These frame-local bindings are actually frame parameters: you create a  
 frame-local binding in a specific frame by calling  
 modify-frame-parameters and specifying the variable name as the  
 parameter name.  
   
 Buffer-local bindings take precedence over frame-local bindings.  
 Thus, if the current buffer has a buffer-local binding, that binding is  
 active; otherwise, if the selected frame has a frame-local binding,  
 that binding is active; otherwise, the default binding is active.  
   
 It would not be hard to implement window-local bindings, but it is not  
 clear that this would be very useful; windows tend to come and go in a  
 very transitory fashion, so that trying to produce any specific effect  
 through a window-local binding would not be very robust.  
   
 ** `sregexq' and `sregex' are two new functions for constructing  
 "symbolic regular expressions."  These are Lisp expressions that, when  
 evaluated, yield conventional string-based regexps.  The symbolic form  
 makes it easier to construct, read, and maintain complex patterns.  
 See the documentation in sregex.el.  
   
 ** parse-partial-sexp's return value has an additional element which  
 is used to pass information along if you pass it to another call to  
 parse-partial-sexp, starting its scan where the first call ended.  
 The contents of this field are not yet finalized.  
   
 ** eval-region now accepts a fourth optional argument READ-FUNCTION.  
 If it is non-nil, that function is used instead of `read'.  
   
 ** unload-feature by default removes the feature's functions from  
 known hooks to avoid trouble, but a package providing FEATURE can  
 define a hook FEATURE-unload-hook to be run by unload-feature instead.  
   
 ** read-from-minibuffer no longer returns the argument DEFAULT-VALUE  
 when the user enters empty input.  It now returns the null string, as  
 it did in Emacs 19.  The default value is made available in the  
 history via M-n, but it is not applied here as a default.  
   
 The other, more specialized minibuffer-reading functions continue to  
 return the default value (not the null string) when the user enters  
 empty input.  
   
 ** The new variable read-buffer-function controls which routine to use  
 for selecting buffers.  For example, if you set this variable to  
 `iswitchb-read-buffer', iswitchb will be used to read buffer names.  
 Other functions can also be used if they accept the same arguments as  
 `read-buffer' and return the selected buffer name as a string.  
   
 ** The new function read-passwd reads a password from the terminal,  
 echoing a period for each character typed.  It takes three arguments:  
 a prompt string, a flag which says "read it twice to make sure", and a  
 default password to use if the user enters nothing.  
   
 ** The variable fill-nobreak-predicate gives major modes a way to  
 specify not to break a line at certain places.  Its value is a  
 function which is called with no arguments, with point located at the  
 place where a break is being considered.  If the function returns  
 non-nil, then the line won't be broken there.  
   
 ** window-end now takes an optional second argument, UPDATE.  
 If this is non-nil, then the function always returns an accurate  
 up-to-date value for the buffer position corresponding to the  
 end of the window, even if this requires computation.  
   
 ** other-buffer now takes an optional argument FRAME  
 which specifies which frame's buffer list to use.  
 If it is nil, that means use the selected frame's buffer list.  
   
 ** The new variable buffer-display-time, always local in every buffer,  
 holds the value of (current-time) as of the last time that a window  
 was directed to display this buffer.  
   
 ** It is now meaningful to compare two window-configuration objects  
 with `equal'.  Two window-configuration objects are equal if they  
 describe equivalent arrangements of windows, in the same frame--in  
 other words, if they would give the same results if passed to  
 set-window-configuration.  
   
 ** compare-window-configurations is a new function that compares two  
 window configurations loosely.  It ignores differences in saved buffer  
 positions and scrolling, and considers only the structure and sizes of  
 windows and the choice of buffers to display.  
   
 ** The variable minor-mode-overriding-map-alist allows major modes to  
 override the key bindings of a minor mode.  The elements of this alist  
 look like the elements of minor-mode-map-alist: (VARIABLE . KEYMAP).  
   
 If the VARIABLE in an element of minor-mode-overriding-map-alist has a  
 non-nil value, the paired KEYMAP is active, and totally overrides the  
 map (if any) specified for the same variable in minor-mode-map-alist.  
   
 minor-mode-overriding-map-alist is automatically local in all buffers,  
 and it is meant to be set by major modes.  
   
 ** The function match-string-no-properties is like match-string  
 except that it discards all text properties from the result.  
   
 ** 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 new variable temporary-file-directory specifies the directory  
 to use for creating temporary files.  The default value is determined  
 in a reasonable way for your operating system; on GNU and Unix systems  
 it is based on the TMP and TMPDIR environment variables.  
   
 ** Menu changes  
   
 *** easymenu.el now uses the new menu item format and supports the  
 keywords :visible and :filter.  The existing keyword :keys is now  
 better supported.  
   
 The variable `easy-menu-precalculate-equivalent-keybindings' controls  
 a new feature which calculates keyboard equivalents for the menu when  
 you define the menu.  The default is t.  If you rarely use menus, you  
 can set the variable to nil to disable this precalculation feature;  
 then the calculation is done only if you use the menu bar.  
   
 *** A new format for menu items is supported.  
   
 In a keymap, a key binding that has the format  
  (STRING . REAL-BINDING) or (STRING HELP-STRING . REAL-BINDING)  
 defines a menu item. Now a menu item definition may also be a list that  
 starts with the symbol `menu-item'.  
   
 The format is:  
  (menu-item ITEM-NAME) or  
  (menu-item ITEM-NAME REAL-BINDING . ITEM-PROPERTY-LIST)  
 where ITEM-NAME is an expression which evaluates to the menu item  
 string, and ITEM-PROPERTY-LIST has the form of a property list.  
 The supported properties include  
   
 :enable FORM      Evaluate FORM to determine whether the  
                   item is enabled.  
 :visible FORM     Evaluate FORM to determine whether the  
                   item should appear in the menu.  
 :filter FILTER-FN  
                   FILTER-FN is a function of one argument,  
                   which will be REAL-BINDING.  
                   It should return a binding to use instead.  
 :keys DESCRIPTION  
                   DESCRIPTION is a string that describes an equivalent keyboard  
                   binding for REAL-BINDING.  DESCRIPTION is expanded with  
                   `substitute-command-keys' before it is used.  
 :key-sequence KEY-SEQUENCE  
                   KEY-SEQUENCE is a key-sequence for an equivalent  
                   keyboard binding.  
 :key-sequence nil  
                   This means that the command normally has no  
                   keyboard equivalent.  
 :help HELP        HELP is the extra help string (not currently used).  
 :button (TYPE . SELECTED)  
                   TYPE is :toggle or :radio.  
                   SELECTED is a form, to be evaluated, and its  
                   value says whether this button is currently selected.  
   
 Buttons are at the moment only simulated by prefixes in the menu.  
 Eventually ordinary X-buttons may be supported.  
   
 (menu-item ITEM-NAME) defines unselectable item.  
   
 ** New event types  
   
 *** The new event type `mouse-wheel' is generated by a wheel on a  
 mouse (such as the MS Intellimouse).  The event contains a delta that  
 corresponds to the amount and direction that the wheel is rotated,  
 which is typically used to implement a scroll or zoom.  The format is:  
   
   (mouse-wheel POSITION DELTA)  
   
 where POSITION is a list describing the position of the event in the  
 same format as a mouse-click event, and DELTA is a signed number  
 indicating the number of increments by which the wheel was rotated.  A  
 negative DELTA indicates that the wheel was rotated backwards, towards  
 the user, and a positive DELTA indicates that the wheel was rotated  
 forward, away from the user.  
   
 As of now, this event type is generated only on MS Windows.  
   
 *** The new event type `drag-n-drop' is generated when a group of  
 files is selected in an application outside of Emacs, and then dragged  
 and dropped onto an Emacs frame.  The event contains a list of  
 filenames that were dragged and dropped, which are then typically  
 loaded into Emacs.  The format is:  
   
   (drag-n-drop POSITION FILES)  
   
 where POSITION is a list describing the position of the event in the  
 same format as a mouse-click event, and FILES is the list of filenames  
 that were dragged and dropped.  
   
 As of now, this event type is generated only on MS Windows.  
   
 ** Changes relating to multibyte characters.  
   
 *** The variable enable-multibyte-characters is now read-only;  
 any attempt to set it directly signals an error.  The only way  
 to change this value in an existing buffer is with set-buffer-multibyte.  
   
 *** In a string constant, `\ ' now stands for "nothing at all".  You  
 can use it to terminate a hex escape which is followed by a character  
 that could otherwise be read as part of the hex escape.  
   
 *** String indices are now measured in characters, as they were  
 in Emacs 19 and before.  
   
 The function chars-in-string has been deleted.  
 The function concat-chars has been renamed to `string'.  
   
 *** The function set-buffer-multibyte sets the flag in the current  
 buffer that says whether the buffer uses multibyte representation or  
 unibyte representation.  If the argument is nil, it selects unibyte  
 representation.  Otherwise it selects multibyte representation.  
   
 This function does not change the contents of the buffer, viewed  
 as a sequence of bytes.  However, it does change the contents  
 viewed as characters; a sequence of two bytes which is treated as  
 one character when the buffer uses multibyte representation  
 will count as two characters using unibyte representation.  
   
 This function sets enable-multibyte-characters to record which  
 representation is in use.  It also adjusts various data in the buffer  
 (including its markers, overlays and text properties) so that they are  
 consistent with the new representation.  
   
 *** string-make-multibyte takes a string and converts it to multibyte  
 representation.  Most of the time, you don't need to care  
 about the representation, because Emacs converts when necessary;  
 however, it makes a difference when you compare strings.  
   
 The conversion of non-ASCII characters works by adding the value of  
 nonascii-insert-offset to each character, or by translating them  
 using the table nonascii-translation-table.  
   
 *** string-make-unibyte takes a string and converts it to unibyte  
 representation.  Most of the time, you don't need to care about the  
 representation, but it makes a difference when you compare strings.  
   
 The conversion from multibyte to unibyte representation  
 loses information; the only time Emacs performs it automatically  
 is when inserting a multibyte string into a unibyte buffer.  
   
 *** string-as-multibyte takes a string, and returns another string  
 which contains the same bytes, but treats them as multibyte.  
   
 *** string-as-unibyte takes a string, and returns another string  
 which contains the same bytes, but treats them as unibyte.  
   
 *** The new function compare-strings lets you compare  
 portions of two strings.  Unibyte strings are converted to multibyte,  
 so that a unibyte string can match a multibyte string.  
 You can specify whether to ignore case or not.  
   
 *** assoc-ignore-case now uses compare-strings so that  
 it can treat unibyte and multibyte strings as equal.  
   
 *** Regular expression operations and buffer string searches now  
 convert the search pattern to multibyte or unibyte to accord with the  
 buffer or string being searched.  
   
 One consequence is that you cannot always use \200-\377 inside of  
 [...] to match all non-ASCII characters.  This does still work when  
 searching or matching a unibyte buffer or string, but not when  
 searching or matching a multibyte string.  Unfortunately, there is no  
 obvious choice of syntax to use within [...] for that job.  But, what  
 you want is just to match all non-ASCII characters, the regular  
 expression [^\0-\177] works for it.  
   
 *** Structure of coding system changed.  
   
 All coding systems (including aliases and subsidiaries) are named  
 by symbols; the symbol's `coding-system' property is a vector  
 which defines the coding system.  Aliases share the same vector  
 as the principal name, so that altering the contents of this  
 vector affects the principal name and its aliases.  You can define  
 your own alias name of a coding system by the function  
 define-coding-system-alias.  
   
 The coding system definition includes a property list of its own.  Use  
 the new functions `coding-system-get' and `coding-system-put' to  
 access such coding system properties as post-read-conversion,  
 pre-write-conversion, character-translation-table-for-decode,  
 character-translation-table-for-encode, mime-charset, and  
 safe-charsets.  For instance, (coding-system-get 'iso-latin-1  
 'mime-charset) gives the corresponding MIME-charset parameter  
 `iso-8859-1'.  
   
 Among the coding system properties listed above, safe-charsets is new.  
 The value of this property is a list of character sets which this  
 coding system can correctly encode and decode.  For instance:  
 (coding-system-get 'iso-latin-1 'safe-charsets) => (ascii latin-iso8859-1)  
   
 Here, "correctly encode" means that the encoded character sets can  
 also be handled safely by systems other than Emacs as far as they  
 are capable of that coding system.  Though, Emacs itself can encode  
 the other character sets and read it back correctly.  
   
 *** The new function select-safe-coding-system can be used to find a  
 proper coding system for encoding the specified region or string.  
 This function requires a user interaction.  
   
 *** The new functions find-coding-systems-region and  
 find-coding-systems-string are helper functions used by  
 select-safe-coding-system.  They return a list of all proper coding  
 systems to encode a text in some region or string.  If you don't want  
 a user interaction, use one of these functions instead of  
 select-safe-coding-system.  
   
 *** The explicit encoding and decoding functions, such as  
 decode-coding-region and encode-coding-string, now set  
 last-coding-system-used to reflect the actual way encoding or decoding  
 was done.  
   
 *** The new function detect-coding-with-language-environment can be  
 used to detect a coding system of text according to priorities of  
 coding systems used by some specific language environment.  
   
 *** The functions detect-coding-region and detect-coding-string always  
 return a list if the arg HIGHEST is nil.  Thus, if only ASCII  
 characters are found, they now return a list of single element  
 `undecided' or its subsidiaries.  
   
 *** The new functions coding-system-change-eol-conversion and  
 coding-system-change-text-conversion can be used to get a different  
 coding system than what specified only in how end-of-line or text is  
 converted.  
   
 *** The new function set-selection-coding-system can be used to set a  
 coding system for communicating with other X clients.  
   
 *** The function `map-char-table' now passes as argument only valid  
 character codes, plus generic characters that stand for entire  
 character sets or entire subrows of a character set.  In other words,  
 each time `map-char-table' calls its FUNCTION argument, the key value  
 either will be a valid individual character code, or will stand for a  
 range of characters.  
   
 *** The new function `char-valid-p' can be used for checking whether a  
 Lisp object is a valid character code or not.  
   
 *** The new function `charset-after' returns a charset of a character  
 in the current buffer at position POS.  
   
 *** Input methods are now implemented using the variable  
 input-method-function.  If this is non-nil, its value should be a  
 function; then, whenever Emacs reads an input event that is a printing  
 character with no modifier bits, it calls that function, passing the  
 event as an argument.  Often this function will read more input, first  
 binding input-method-function to nil.  
   
 The return value should be a list of the events resulting from input  
 method processing.  These events will be processed sequentially as  
 input, before resorting to unread-command-events.  Events returned by  
 the input method function are not passed to the input method function,  
 not even if they are printing characters with no modifier bits.  
   
 The input method function is not called when reading the second and  
 subsequent events of a key sequence.  
   
 *** You can customize any language environment by using  
 set-language-environment-hook and exit-language-environment-hook.  
   
 The hook `exit-language-environment-hook' should be used to undo  
 customizations that you made with set-language-environment-hook.  For  
 instance, if you set up a special key binding for a specific language  
 environment by set-language-environment-hook, you should set up  
 exit-language-environment-hook to restore the normal key binding.  
   
 * Changes in Emacs 20.1  
   
 ** Emacs has a new facility for customization of its many user  
 options.  It is called M-x customize.  With this facility you can look  
 at the many user options in an organized way; they are grouped into a  
 tree structure.  
   
 M-x customize also knows what sorts of values are legitimate for each  
 user option and ensures that you don't use invalid values.  
   
 With M-x customize, you can set options either for the present Emacs  
 session or permanently.  (Permanent settings are stored automatically  
 in your .emacs file.)  
   
 ** Scroll bars are now on the left side of the window.  
 You can change this with M-x customize-option scroll-bar-mode.  
   
 ** The mode line no longer includes the string `Emacs'.  
 This makes more space in the mode line for other information.  
   
 ** When you select a region with the mouse, it is highlighted  
 immediately afterward.  At that time, if you type the DELETE key, it  
 kills the region.  
   
 The BACKSPACE key, and the ASCII character DEL, do not do this; they  
 delete the character before point, as usual.  
   
 ** In an incremental search the whole current match is highlighted  
 on terminals which support this.  (You can disable this feature  
 by setting search-highlight to nil.)  
   
 ** In the minibuffer, in some cases, you can now use M-n to  
 insert the default value into the minibuffer as text.  In effect,  
 the default value (if the minibuffer routines know it) is tacked  
 onto the history "in the future".  (The more normal use of the  
 history list is to use M-p to insert minibuffer input used in the  
 past.)  
   
 ** 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.  
   
 ** Scrolling changes  
   
 *** Scroll commands to scroll a whole screen now preserve the screen  
 position of the cursor, if scroll-preserve-screen-position is non-nil.  
   
 In this mode, if you scroll several screens back and forth, finishing  
 on the same screen where you started, the cursor goes back to the line  
 where it started.  
   
 *** If you set scroll-conservatively to a small number, then when you  
 move point a short distance off the screen, Emacs will scroll the  
 screen just far enough to bring point back on screen, provided that  
 does not exceed `scroll-conservatively' lines.  
   
 *** The new variable scroll-margin says how close point can come to the  
 top or bottom of a window.  It is a number of screen lines; if point  
 comes within that many lines of the top or bottom of the window, Emacs  
 recenters the window.  
   
 ** International character set support (MULE)  
   
 Emacs now supports a wide variety of international character sets,  
 including European variants of the Latin alphabet, as well as Chinese,  
 Devanagari (Hindi and Marathi), Ethiopian, Greek, IPA, Japanese,  
 Korean, Lao, Russian, Thai, Tibetan, and Vietnamese scripts.  These  
 features have been merged from the modified version of Emacs known as  
 MULE (for "MULti-lingual Enhancement to GNU Emacs")  
   
 Users of these scripts have established many more-or-less standard  
 coding systems for storing files.  Emacs uses a single multibyte  
 character encoding within Emacs buffers; it can translate from a wide  
 variety of coding systems when reading a file and can translate back  
 into any of these coding systems when saving a file.  
   
 Keyboards, even in the countries where these character sets are used,  
 generally don't have keys for all the characters in them.  So Emacs  
 supports various "input methods", typically one for each script or  
 language, to make it possible to type them.  
   
 The Emacs internal multibyte encoding represents a non-ASCII  
 character as a sequence of bytes in the range 0200 through 0377.  
   
 The new prefix key C-x RET is used for commands that pertain  
 to multibyte characters, coding systems, and input methods.  
   
 You can disable multibyte character support as follows:  
   
   (setq-default enable-multibyte-characters nil)  
   
 Calling the function standard-display-european turns off multibyte  
 characters, unless you specify a non-nil value for the second  
 argument, AUTO.  This provides compatibility for people who are  
 already using standard-display-european to continue using unibyte  
 characters for their work until they want to change.  
   
 *** Input methods  
   
 An input method is a kind of character conversion which is designed  
 specifically for interactive input.  In Emacs, typically each language  
 has its own input method (though sometimes several languages which use  
 the same characters can share one input method).  Some languages  
 support several input methods.  
   
 The simplest kind of input method works by mapping ASCII letters into  
 another alphabet.  This is how the Greek and Russian input methods  
 work.  
   
 A more powerful technique is composition: converting sequences of  
 characters into one letter.  Many European input methods use  
 composition to produce a single non-ASCII letter from a sequence which  
 consists of a letter followed by diacritics.  For example, a' is one  
 sequence of two characters that might be converted into a single  
 letter.  
   
 The input methods for syllabic scripts typically use mapping followed  
 by conversion.  The input methods for Thai and Korean work this way.  
 First, letters are mapped into symbols for particular sounds or tone  
 marks; then, sequences of these which make up a whole syllable are  
 mapped into one syllable sign--most often a "composite character".  
   
 None of these methods works very well for Chinese and Japanese, so  
 they are handled specially.  First you input a whole word using  
 phonetic spelling; then, after the word is in the buffer, Emacs  
 converts it into one or more characters using a large dictionary.  
   
 Since there is more than one way to represent a phonetically spelled  
 word using Chinese characters, Emacs can only guess which one to use;  
 typically these input methods give you a way to say "guess again" if  
 the first guess is wrong.  
   
 *** The command C-x RET m (toggle-enable-multibyte-characters)  
 turns multibyte character support on or off for the current buffer.  
   
 If multibyte character support is turned off in a buffer, then each  
 byte is a single character, even codes 0200 through 0377--exactly as  
 they did in Emacs 19.34.  This includes the features for support for  
 the European characters, ISO Latin-1 and ISO Latin-2.  
   
 However, there is no need to turn off multibyte character support to  
 use ISO Latin-1 or ISO Latin-2; the Emacs multibyte character set  
 includes all the characters in these character sets, and Emacs can  
 translate automatically to and from either one.  
   
 *** Visiting a file in unibyte mode.  
   
 Turning off multibyte character support in the buffer after visiting a  
 file with multibyte code conversion will display the multibyte  
 sequences already in the buffer, byte by byte.  This is probably not  
 what you want.  
   
 If you want to edit a file of unibyte characters (Latin-1, for  
 example), you can do it by specifying `no-conversion' as the coding  
 system when reading the file.  This coding system also turns off  
 multibyte characters in that buffer.  
   
 If you turn off multibyte character support entirely, this turns off  
 character conversion as well.  
   
 *** Displaying international characters on X Windows.  
   
 A font for X typically displays just one alphabet or script.  
 Therefore, displaying the entire range of characters Emacs supports  
 requires using many fonts.  
   
 Therefore, Emacs now supports "fontsets".  Each fontset is a  
 collection of fonts, each assigned to a range of character codes.  
   
 A fontset has a name, like a font.  Individual fonts are defined by  
 the X server; fontsets are defined within Emacs itself.  But once you  
 have defined a fontset, you can use it in a face or a frame just as  
 you would use a font.  
   
 If a fontset specifies no font for a certain character, or if it  
 specifies a font that does not exist on your system, then it cannot  
 display that character.  It will display an empty box instead.  
   
 The fontset height and width are determined by the ASCII characters  
 (that is, by the font in the fontset which is used for ASCII  
 characters).  
   
 *** Defining fontsets.  
   
 Emacs does not use any fontset by default.  Its default font is still  
 chosen as in previous versions.  You can tell Emacs to use a fontset  
 with the `-fn' option or the `Font' X resource.  
   
 Emacs creates a standard fontset automatically according to the value  
 of standard-fontset-spec.  This fontset's short name is  
 `fontset-standard'.  Bold, italic, and bold-italic variants of the  
 standard fontset are created automatically.  
   
 If you specify a default ASCII font with the `Font' resource or `-fn'  
 argument, a fontset is generated from it.  This works by replacing the  
 FOUNDARY, FAMILY, ADD_STYLE, and AVERAGE_WIDTH fields of the font name  
 with `*' then using this to specify a fontset.  This fontset's short  
 name is `fontset-startup'.  
   
 Emacs checks resources of the form Fontset-N where N is 0, 1, 2...  
 The resource value should have this form:  
         FONTSET-NAME, [CHARSET-NAME:FONT-NAME]...  
 FONTSET-NAME should have the form of a standard X font name, except:  
         * most fields should be just the wild card "*".  
         * the CHARSET_REGISTRY field should be "fontset"  
         * the CHARSET_ENCODING field can be any nickname of the fontset.  
 The construct CHARSET-NAME:FONT-NAME can be repeated any number  
 of times; each time specifies the font for one character set.  
 CHARSET-NAME should be the name of a character set, and FONT-NAME  
 should specify an actual font to use for that character set.  
   
 Each of these fontsets has an alias which is made from the  
 last two font name fields, CHARSET_REGISTRY and CHARSET_ENCODING.  
 You can refer to the fontset by that alias or by its full name.  
   
 For any character sets that you don't mention, Emacs tries to choose a  
 font by substituting into FONTSET-NAME.  For instance, with the  
 following resource,  
         Emacs*Fontset-0: -*-fixed-medium-r-normal-*-24-*-*-*-*-*-fontset-24  
 the font for ASCII is generated as below:  
         -*-fixed-medium-r-normal-*-24-*-ISO8859-1  
 Here is the substitution rule:  
     Change CHARSET_REGISTRY and CHARSET_ENCODING to that of the charset  
     defined in the variable x-charset-registries.  For instance, ASCII has  
     the entry (ascii . "ISO8859-1") in this variable.  Then, reduce  
     sequences of wild cards -*-...-*- with a single wildcard -*-.  
     (This is to prevent use of auto-scaled fonts.)  
   
 The function which processes the fontset resource value to create the  
 fontset is called create-fontset-from-fontset-spec.  You can also call  
 that function explicitly to create a fontset.  
   
 With the X resource Emacs.Font, you can specify a fontset name just  
 like an actual font name.  But be careful not to specify a fontset  
 name in a wildcard resource like Emacs*Font--that tries to specify the  
 fontset for other purposes including menus, and they cannot handle  
 fontsets.  
   
 *** The command M-x set-language-environment sets certain global Emacs  
 defaults for a particular choice of language.  
   
 Selecting a language environment typically specifies a default input  
 method and which coding systems to recognize automatically when  
 visiting files.  However, it does not try to reread files you have  
 already visited; the text in those buffers is not affected.  The  
 language environment may also specify a default choice of coding  
 system for new files that you create.  
   
 It makes no difference which buffer is current when you use  
 set-language-environment, because these defaults apply globally to the  
 whole Emacs session.  
   
 For example, M-x set-language-environment RET Latin-1 RET  
 chooses the Latin-1 character set.  In the .emacs file, you can do this  
 with (set-language-environment "Latin-1").  
   
 *** The command C-x RET f (set-buffer-file-coding-system)  
 specifies the file coding system for the current buffer.  This  
 specifies what sort of character code translation to do when saving  
 the file.  As an argument, you must specify the name of one of the  
 coding systems that Emacs supports.  
   
 *** The command C-x RET c (universal-coding-system-argument)  
 lets you specify a coding system when you read or write a file.  
 This command uses the minibuffer to read a coding system name.  
 After you exit the minibuffer, the specified coding system  
 is used for *the immediately following command*.  
   
 So if the immediately following command is a command to read or  
 write a file, it uses the specified coding system for that file.  
   
 If the immediately following command does not use the coding system,  
 then C-x RET c ultimately has no effect.  
   
 For example,  C-x RET c iso-8859-1 RET C-x C-f temp RET  
 visits the file `temp' treating it as ISO Latin-1.  
   
 *** You can specify the coding system for a file using the -*-  
 construct.  Include `coding: CODINGSYSTEM;' inside the -*-...-*-  
 to specify use of coding system CODINGSYSTEM.  You can also  
 specify the coding system in a local variable list at the end  
 of the file.  
   
 *** The command C-x RET t (set-terminal-coding-system) specifies  
 the coding system for terminal output.  If you specify a character  
 code for terminal output, all characters output to the terminal are  
 translated into that character code.  
   
 This feature is useful for certain character-only terminals built in  
 various countries to support the languages of those countries.  
   
 By default, output to the terminal is not translated at all.  
   
 *** The command C-x RET k (set-keyboard-coding-system) specifies  
 the coding system for keyboard input.  
   
 Character code translation of keyboard input is useful for terminals  
 with keys that send non-ASCII graphic characters--for example,  
 some terminals designed for ISO Latin-1 or subsets of it.  
   
 By default, keyboard input is not translated at all.  
   
 Character code translation of keyboard input is similar to using an  
 input method, in that both define sequences of keyboard input that  
 translate into single characters.  However, input methods are designed  
 to be convenient for interactive use, while the code translations are  
 designed to work with terminals.  
   
 *** The command C-x RET p (set-buffer-process-coding-system)  
 specifies the coding system for input and output to a subprocess.  
 This command applies to the current buffer; normally, each subprocess  
 has its own buffer, and thus you can use this command to specify  
 translation to and from a particular subprocess by giving the command  
 in the corresponding buffer.  
   
 By default, process input and output are not translated at all.  
   
 *** The variable file-name-coding-system specifies the coding system  
 to use for encoding file names before operating on them.  
 It is also used for decoding file names obtained from the system.  
   
 *** The command C-\ (toggle-input-method) activates or deactivates  
 an input method.  If no input method has been selected before, the  
 command prompts for you to specify the language and input method you  
 want to use.  
   
 C-u C-\ (select-input-method) lets you switch to a different input  
 method.  C-h C-\ (or C-h I) describes the current input method.  
   
 *** Some input methods remap the keyboard to emulate various keyboard  
 layouts commonly used for particular scripts.  How to do this  
 remapping properly depends on your actual keyboard layout.  To specify  
 which layout your keyboard has, use M-x quail-set-keyboard-layout.  
   
 *** The command C-h C (describe-coding-system) displays  
 the coding systems currently selected for various purposes, plus  
 related information.  
   
 *** The command C-h h (view-hello-file) displays a file called  
 HELLO, which has examples of text in many languages, using various  
 scripts.  
   
 *** The command C-h L (describe-language-support) displays  
 information about the support for a particular language.  
 You specify the language as an argument.  
   
 *** The mode line now contains a letter or character that identifies  
 the coding system used in the visited file.  It normally follows the  
 first dash.  
   
 A dash indicates the default state of affairs: no code conversion  
 (except CRLF => newline if appropriate).  `=' means no conversion  
 whatsoever.  The ISO 8859 coding systems are represented by digits  
 1 through 9.  Other coding systems are represented by letters:  
   
     A alternativnyj (Russian)  
     B big5 (Chinese)  
     C cn-gb-2312 (Chinese)  
     C iso-2022-cn (Chinese)  
     D in-is13194-devanagari (Indian languages)  
     E euc-japan (Japanese)  
     I iso-2022-cjk or iso-2022-ss2 (Chinese, Japanese, Korean)  
     J junet (iso-2022-7) or old-jis (iso-2022-jp-1978-irv)  (Japanese)  
     K euc-korea (Korean)  
     R koi8 (Russian)  
     Q tibetan  
     S shift_jis (Japanese)  
     T lao  
     T tis620 (Thai)  
     V viscii or vscii (Vietnamese)  
     i iso-2022-lock (Chinese, Japanese, Korean)  
     k iso-2022-kr (Korean)  
     v viqr (Vietnamese)  
     z hz (Chinese)  
   
 When you are using a character-only terminal (not a window system),  
 two additional characters appear in between the dash and the file  
 coding system.  These two characters describe the coding system for  
 keyboard input, and the coding system for terminal output.  
   
 *** The new variable rmail-file-coding-system specifies the code  
 conversion to use for RMAIL files.  The default value is nil.  
   
 When you read mail with Rmail, each message is decoded automatically  
 into Emacs' internal format.  This has nothing to do with  
 rmail-file-coding-system.  That variable controls reading and writing  
 Rmail files themselves.  
   
 *** The new variable sendmail-coding-system specifies the code  
 conversion for outgoing mail.  The default value is nil.  
   
 Actually, there are three different ways of specifying the coding system  
 for sending mail:  
   
 - If you use C-x RET f in the mail buffer, that takes priority.  
 - Otherwise, if you set sendmail-coding-system non-nil, that specifies it.  
 - Otherwise, the default coding system for new files is used,  
   if that is non-nil.  That comes from your language environment.  
 - Otherwise, Latin-1 is used.  
   
 *** The command C-h t (help-with-tutorial) accepts a prefix argument  
 to specify the language for the tutorial file.  Currently, English,  
 Japanese, Korean and Thai are supported.  We welcome additional  
 translations.  
   
 ** An easy new way to visit a file with no code or format conversion  
 of any kind: Use M-x find-file-literally.  There is also a command  
 insert-file-literally which inserts a file into the current buffer  
 without any conversion.  
   
 ** C-q's handling of octal character codes is changed.  
 You can now specify any number of octal digits.  
 RET terminates the digits and is discarded;  
 any other non-digit terminates the digits and is then used as input.  
   
 ** There are new commands for looking up Info documentation for  
 functions, variables and file names used in your programs.  
   
 Type M-x info-lookup-symbol to look up a symbol in the buffer at point.  
 Type M-x info-lookup-file to look up a file in the buffer at point.  
   
 Precisely which Info files are used to look it up depends on the major  
 mode.  For example, in C mode, the GNU libc manual is used.  
   
 ** M-TAB in most programming language modes now runs the command  
 complete-symbol.  This command performs completion on the symbol name  
 in the buffer before point.  
   
 With a numeric argument, it performs completion based on the set of  
 symbols documented in the Info files for the programming language that  
 you are using.  
   
 With no argument, it does completion based on the current tags tables,  
 just like the old binding of M-TAB (complete-tag).  
   
 ** File locking works with NFS now.  
   
 The lock file for FILENAME is now a symbolic link named .#FILENAME,  
 in the same directory as FILENAME.  
   
 This means that collision detection between two different machines now  
 works reasonably well; it also means that no file server or directory  
 can become a bottleneck.  
   
 The new method does have drawbacks.  It means that collision detection  
 does not operate when you edit a file in a directory where you cannot  
 create new files.  Collision detection also doesn't operate when the  
 file server does not support symbolic links.  But these conditions are  
 rare, and the ability to have collision detection while using NFS is  
 so useful that the change is worth while.  
   
 When Emacs or a system crashes, this may leave behind lock files which  
 are stale.  So you may occasionally get warnings about spurious  
 collisions.  When you determine that the collision is spurious, just  
 tell Emacs to go ahead anyway.  
   
 ** If you wish to use Show Paren mode to display matching parentheses,  
 it is no longer sufficient to load paren.el.  Instead you must call  
 show-paren-mode.  
   
 ** If you wish to use Delete Selection mode to replace a highlighted  
 selection when you insert new text, it is no longer sufficient to load  
 delsel.el.  Instead you must call the function delete-selection-mode.  
   
 ** If you wish to use Partial Completion mode to complete partial words  
 within symbols or filenames, it is no longer sufficient to load  
 complete.el.  Instead you must call the function partial-completion-mode.  
   
 ** If you wish to use uniquify to rename buffers for you,  
 it is no longer sufficient to load uniquify.el.  You must also  
 set uniquify-buffer-name-style to one of the non-nil legitimate values.  
   
 ** Changes in View mode.  
   
 *** Several new commands are available in View mode.  
 Do H in view mode for a list of commands.  
   
 *** There are two new commands for entering View mode:  
 view-file-other-frame and view-buffer-other-frame.  
   
 *** Exiting View mode does a better job of restoring windows to their  
 previous state.  
   
 *** New customization variable view-scroll-auto-exit. If non-nil,  
 scrolling past end of buffer makes view mode exit.  
   
 *** New customization variable view-exits-all-viewing-windows.  If  
 non-nil, view-mode will at exit restore all windows viewing buffer,  
 not just the selected window.  
   
 *** New customization variable view-read-only.  If non-nil, visiting a  
 read-only file automatically enters View mode, and toggle-read-only  
 turns View mode on or off.  
   
 *** New customization variable view-remove-frame-by-deleting controls  
 how to remove a not needed frame at view mode exit. If non-nil,  
 delete the frame, if nil make an icon of it.  
   
 ** C-x v l, the command to print a file's version control log,  
 now positions point at the entry for the file's current branch version.  
   
 ** C-x v =, the command to compare a file with the last checked-in version,  
 has a new feature.  If the file is currently not locked, so that it is  
 presumably identical to the last checked-in version, the command now asks  
 which version to compare with.  
   
 ** When using hideshow.el, incremental search can temporarily show hidden  
 blocks if a match is inside the block.  
   
 The block is hidden again if the search is continued and the next match  
 is outside the block.  By customizing the variable  
 isearch-hide-immediately you can choose to hide all the temporarily  
 shown blocks only when exiting from incremental search.  
   
 By customizing the variable hs-isearch-open you can choose what kind  
 of blocks to temporarily show during isearch: comment blocks, code  
 blocks, all of them or none.  
   
 ** The new command C-x 4 0 (kill-buffer-and-window) kills the  
 current buffer and deletes the selected window.  It asks for  
 confirmation first.  
   
 ** C-x C-w, which saves the buffer into a specified file name,  
 now changes the major mode according to that file name.  
 However, the mode will not be changed if  
 (1) a local variables list or the `-*-' line specifies a major mode, or  
 (2) the current major mode is a "special" mode,  
     not suitable for ordinary files, or  
 (3) the new file name does not particularly specify any mode.  
   
 This applies to M-x set-visited-file-name as well.  
   
 However, if you set change-major-mode-with-file-name to nil, then  
 these commands do not change the major mode.  
   
 ** M-x occur changes.  
   
 *** If the argument to M-x occur contains upper case letters,  
 it performs a case-sensitive search.  
   
 *** In the *Occur* buffer made by M-x occur,  
 if you type g or M-x revert-buffer, this repeats the search  
 using the same regular expression and the same buffer as before.  
   
 ** In Transient Mark mode, the region in any one buffer is highlighted  
 in just one window at a time.  At first, it is highlighted in the  
 window where you set the mark.  The buffer's highlighting remains in  
 that window unless you select to another window which shows the same  
 buffer--then the highlighting moves to that window.  
   
 ** The feature to suggest key bindings when you use M-x now operates  
 after the command finishes.  The message suggesting key bindings  
 appears temporarily in the echo area.  The previous echo area contents  
 come back after a few seconds, in case they contain useful information.  
   
 ** Each frame now independently records the order for recently  
 selected buffers, so that the default for C-x b is now based on the  
 buffers recently selected in the selected frame.  
   
 ** Outline mode changes.  
   
 *** Outline mode now uses overlays (this is the former noutline.el).  
   
 *** Incremental searches skip over invisible text in Outline mode.  
   
 ** When a minibuffer window is active but not the selected window, if  
 you try to use the minibuffer, you used to get a nested minibuffer.  
 Now, this not only gives an error, it also cancels the minibuffer that  
 was already active.  
   
 The motive for this change is so that beginning users do not  
 unknowingly move away from minibuffers, leaving them active, and then  
 get confused by it.  
   
 If you want to be able to have recursive minibuffers, you must  
 set enable-recursive-minibuffers to non-nil.  
   
 ** Changes in dynamic abbrevs.  
   
 *** Expanding dynamic abbrevs with M-/ is now smarter about case  
 conversion.  If the expansion has mixed case not counting the first  
 character, and the abbreviation matches the beginning of the expansion  
 including case, then the expansion is copied verbatim.  
   
 The expansion is also copied verbatim if the abbreviation itself has  
 mixed case.  And using SPC M-/ to copy an additional word always  
 copies it verbatim except when the previous copied word is all caps.  
   
 *** The values of `dabbrev-case-replace' and `dabbrev-case-fold-search'  
 are no longer Lisp expressions.  They have simply three possible  
 values.  
   
 `dabbrev-case-replace' has these three values: nil (don't preserve  
 case), t (do), or `case-replace' (do like M-x query-replace).  
 `dabbrev-case-fold-search' has these three values: nil (don't ignore  
 case), t (do), or `case-fold-search' (do like search).  
   
 ** Minibuffer history lists are truncated automatically now to a  
 certain length.  The variable history-length specifies how long they  
 can be.  The default value is 30.  
   
 ** Changes in Mail mode.  
   
 *** The key C-x m no longer runs the `mail' command directly.  
 Instead, it runs the command `compose-mail', which invokes the mail  
 composition mechanism you have selected with the variable  
 `mail-user-agent'.  The default choice of user agent is  
 `sendmail-user-agent', which gives behavior compatible with the old  
 behavior.  
   
 C-x 4 m now runs compose-mail-other-window, and C-x 5 m runs  
 compose-mail-other-frame.  
   
 *** While composing a reply to a mail message, from Rmail, you can use  
 the command C-c C-r to cite just the region from the message you are  
 replying to.  This copies the text which is the selected region in the  
 buffer that shows the original message.  
   
 *** The command C-c C-i inserts a file at the end of the message,  
 with separator lines around the contents.  
   
 *** The command M-x expand-mail-aliases expands all mail aliases  
 in suitable mail headers.  Emacs automatically extracts mail alias  
 definitions from your mail alias file (e.g., ~/.mailrc).  You do not  
 need to expand mail aliases yourself before sending mail.  
   
 *** New features in the mail-complete command.  
   
 **** The mail-complete command now inserts the user's full name,  
 for local users or if that is known.  The variable mail-complete-style  
 controls the style to use, and whether to do this at all.  
 Its values are like those of mail-from-style.  
   
 **** The variable mail-passwd-command lets you specify a shell command  
 to run to fetch a set of password-entries that add to the ones in  
 /etc/passwd.  
   
 **** The variable mail-passwd-file now specifies a list of files to read  
 to get the list of user ids.  By default, one file is used:  
 /etc/passwd.  
   
 ** You can "quote" a file name to inhibit special significance of  
 special syntax, by adding `/:' to the beginning.  Thus, if you have a  
 directory named `/foo:', you can prevent it from being treated as a  
 reference to a remote host named `foo' by writing it as `/:/foo:'.  
   
 Emacs uses this new construct automatically when necessary, such as  
 when you start it with a working directory whose name might otherwise  
 be taken to be magic.  
   
 ** There is a new command M-x grep-find which uses find to select  
 files to search through, and grep to scan them.  The output is  
 available in a Compile mode buffer, as with M-x grep.  
   
 M-x grep now uses the -e option if the grep program supports that.  
 (-e prevents problems if the search pattern starts with a dash.)  
   
 ** In Dired, the & command now flags for deletion the files whose names  
 suggest they are probably not needed in the long run.  
   
 In Dired, * is now a prefix key for mark-related commands.  
   
 new key         dired.el binding                old key  
 -------         ----------------                -------  
   * c           dired-change-marks              c  
   * m           dired-mark                      m  
   * *           dired-mark-executables          *  (binding deleted)  
   * /           dired-mark-directories          /  (binding deleted)  
   * @           dired-mark-symlinks             @  (binding deleted)  
   * u           dired-unmark                    u  
   * DEL         dired-unmark-backward           DEL  
   * ?           dired-unmark-all-files          C-M-?  
   * !           dired-unmark-all-marks  
   * %           dired-mark-files-regexp         % m  
   * C-n         dired-next-marked-file          M-}  
   * C-p         dired-prev-marked-file          M-{  
   
 ** Rmail changes.  
   
 *** When Rmail cannot convert your incoming mail into Babyl format, it  
 saves the new mail in the file RMAILOSE.n, where n is an integer  
 chosen to make a unique name.  This way, Rmail will not keep crashing  
 each time you run it.  
   
 *** In Rmail, the variable rmail-summary-line-count-flag now controls  
 whether to include the line count in the summary.  Non-nil means yes.  
   
 *** In Rmail summary buffers, d and C-d (the commands to delete  
 messages) now take repeat counts as arguments.  A negative argument  
 means to move in the opposite direction.  
   
 *** In Rmail, the t command now takes an optional argument which lets  
 you specify whether to show the message headers in full or pruned.  
   
 *** In Rmail, the new command w (rmail-output-body-to-file) writes  
 just the body of the current message into a file, without the headers.  
 It takes the file name from the message subject, by default, but you  
 can edit that file name in the minibuffer before it is actually used  
 for output.  
   
 ** Gnus changes.  
   
 *** nntp.el has been totally rewritten in an asynchronous fashion.  
   
 *** Article prefetching functionality has been moved up into  
 Gnus.  
   
 *** Scoring can now be performed with logical operators like  
 `and', `or', `not', and parent redirection.  
   
 *** Article washing status can be displayed in the  
 article mode line.  
   
 *** gnus.el has been split into many smaller files.  
   
 *** Suppression of duplicate articles based on Message-ID.  
   
 (setq gnus-suppress-duplicates t)  
   
 *** New variables for specifying what score and adapt files  
 are to be considered home score and adapt files.  See  
 `gnus-home-score-file' and `gnus-home-adapt-files'.  
   
 *** Groups can inherit group parameters from parent topics.  
   
 *** Article editing has been revamped and is now usable.  
   
 *** Signatures can be recognized in more intelligent fashions.  
 See `gnus-signature-separator' and `gnus-signature-limit'.  
   
 *** Summary pick mode has been made to look more nn-like.  
 Line numbers are displayed and the `.' command can be  
 used to pick articles.  
   
 *** Commands for moving the .newsrc.eld from one server to  
 another have been added.  
   
     `M-x gnus-change-server'  
   
 *** A way to specify that "uninteresting" fields be suppressed when  
 generating lines in buffers.  
   
 *** Several commands in the group buffer can be undone with  
 `C-M-_'.  
   
 *** Scoring can be done on words using the new score type `w'.  
   
 *** Adaptive scoring can be done on a Subject word-by-word basis:  
   
     (setq gnus-use-adaptive-scoring '(word))  
   
 *** Scores can be decayed.  
   
     (setq gnus-decay-scores t)  
   
 *** Scoring can be performed using a regexp on the Date header.  The  
 Date is normalized to compact ISO 8601 format first.  
   
 *** A new command has been added to remove all data on articles from  
 the native server.  
   
    `M-x gnus-group-clear-data-on-native-groups'  
   
 *** A new command for reading collections of documents  
 (nndoc with nnvirtual on top) has been added -- `C-M-d'.  
   
 *** Process mark sets can be pushed and popped.  
   
 *** A new mail-to-news backend makes it possible to post  
 even when the NNTP server doesn't allow posting.  
   
 *** A new backend for reading searches from Web search engines  
 (DejaNews, Alta Vista, InReference) has been added.  
   
     Use the `G w' command in the group buffer to create such  
     a group.  
   
 *** Groups inside topics can now be sorted using the standard  
 sorting functions, and each topic can be sorted independently.  
   
     See the commands under the `T S' submap.  
   
 *** Subsets of the groups can be sorted independently.  
   
     See the commands under the `G P' submap.  
   
 *** Cached articles can be pulled into the groups.  
   
     Use the `Y c' command.  
   
 *** Score files are now applied in a more reliable order.  
   
 *** Reports on where mail messages end up can be generated.  
   
     `M-x nnmail-split-history'  
   
 *** More hooks and functions have been added to remove junk  
 from incoming mail before saving the mail.  
   
     See `nnmail-prepare-incoming-header-hook'.  
   
 *** The nnml mail backend now understands compressed article files.  
   
 *** To enable Gnus to read/post multi-lingual articles, you must execute  
 the following code, for instance, in your .emacs.  
   
         (add-hook 'gnus-startup-hook 'gnus-mule-initialize)  
   
 Then, when you start Gnus, it will decode non-ASCII text automatically  
 and show appropriate characters.  (Note: if you are using gnus-mime  
 from the SEMI package, formerly known as TM, you should NOT add this  
 hook to gnus-startup-hook; gnus-mime has its own method of handling  
 this issue.)  
   
 Since it is impossible to distinguish all coding systems  
 automatically, you may need to specify a choice of coding system for a  
 particular news group.  This can be done by:  
   
         (gnus-mule-add-group NEWSGROUP 'CODING-SYSTEM)  
   
 Here NEWSGROUP should be a string which names a newsgroup or a tree  
 of newsgroups.  If NEWSGROUP is "XXX.YYY", all news groups under  
 "XXX.YYY" (including "XXX.YYY.ZZZ") will use the specified coding  
 system.  CODING-SYSTEM specifies which coding system to use (for both  
 for reading and posting).  
   
 CODING-SYSTEM can also be a cons cell of the form  
   (READ-CODING-SYSTEM . POST-CODING-SYSTEM)  
 Then READ-CODING-SYSTEM is used when you read messages from the  
 newsgroups, while POST-CODING-SYSTEM is used when you post messages  
 there.  
   
 Emacs knows the right coding systems for certain newsgroups by  
 default.  Here are some of these default settings:  
   
         (gnus-mule-add-group "fj" 'iso-2022-7)  
         (gnus-mule-add-group "alt.chinese.text" 'hz-gb-2312)  
         (gnus-mule-add-group "alt.hk" 'hz-gb-2312)  
         (gnus-mule-add-group "alt.chinese.text.big5" 'cn-big5)  
         (gnus-mule-add-group "soc.culture.vietnamese" '(nil . viqr))  
   
 When you reply by mail to an article, these settings are ignored;  
 the mail is encoded according to sendmail-coding-system, as usual.  
   
 ** CC mode changes.  
   
 *** If you edit primarily one style of C (or C++, Objective-C, Java)  
 code, you may want to make the CC Mode style variables have global  
 values so that you can set them directly in your .emacs file.  To do  
 this, set c-style-variables-are-local-p to nil in your .emacs file.  
 Note that this only takes effect if you do it *before* cc-mode.el is  
 loaded.  
   
 If you typically edit more than one style of C (or C++, Objective-C,  
 Java) code in a single Emacs session, you may want to make the CC Mode  
 style variables have buffer local values.  By default, all buffers  
 share the same style variable settings; to make them buffer local, set  
 c-style-variables-are-local-p to t in your .emacs file.  Note that you  
 must do this *before* CC Mode is loaded.  
   
 *** The new variable c-indentation-style holds the C style name  
 of the current buffer.  
   
 *** The variable c-block-comments-indent-p has been deleted, because  
 it is no longer necessary.  C mode now handles all the supported styles  
 of block comments, with no need to say which one you will use.  
   
 *** There is a new indentation style "python", which specifies the C  
 style that the Python developers like.  
   
 *** There is a new c-cleanup-list option: brace-elseif-brace.  
 This says to put ...} else if (...) {... on one line,  
 just as brace-else-brace says to put ...} else {... on one line.  
   
 ** VC Changes [new]  
   
 *** In vc-retrieve-snapshot (C-x v r), if you don't specify a snapshot  
 name, it retrieves the *latest* versions of all files in the current  
 directory and its subdirectories (aside from files already locked).  
   
 This feature is useful if your RCS directory is a link to a common  
 master directory, and you want to pick up changes made by other  
 developers.  
   
 You can do the same thing for an individual file by typing C-u C-x C-q  
 RET in a buffer visiting that file.  
   
 *** VC can now handle files under CVS that are being "watched" by  
 other developers.  Such files are made read-only by CVS.  To get a  
 writable copy, type C-x C-q in a buffer visiting such a file.  VC then  
 calls "cvs edit", which notifies the other developers of it.  
   
 *** vc-version-diff (C-u C-x v =) now suggests reasonable defaults for  
 version numbers, based on the current state of the file.  
   
 ** Calendar changes.  
   
 *** A new function, list-holidays, allows you list holidays or  
 subclasses of holidays for ranges of years.  Related menu items allow  
 you do this for the year of the selected date, or the  
 following/previous years.  
   
 *** There is now support for the Baha'i calendar system.  Use `pb' in  
 the *Calendar* buffer to display the current Baha'i date.  The Baha'i  
 calendar, or "Badi calendar" is a system of 19 months with 19 days  
 each, and 4 intercalary days (5 during a Gregorian leap year).  The  
 calendar begins May 23, 1844, with each of the months named after a  
 supposed attribute of God.  
   
 ** ps-print changes  
   
 There are some new user variables and subgroups for customizing the page  
 layout.  
   
 *** Headers & Footers (subgroup)  
   
 Some printer systems print a header page and force the first page to  
 be printed on the back of the header page when using duplex.  If your  
 printer system has this behavior, set variable  
 `ps-banner-page-when-duplexing' to t.  
   
 If variable `ps-banner-page-when-duplexing' is non-nil, it prints a  
 blank page as the very first printed page.  So, it behaves as if the  
 very first character of buffer (or region) were a form feed ^L (\014).  
   
 The variable `ps-spool-config' specifies who is responsible for  
 setting duplex mode and page size.  Valid values are:  
   
  lpr-switches    duplex and page size are configured by `ps-lpr-switches'.  
                  Don't forget to set `ps-lpr-switches' to select duplex  
                  printing for your printer.  
   
  setpagedevice   duplex and page size are configured by ps-print using the  
                  setpagedevice PostScript operator.  
   
  nil             duplex and page size are configured by ps-print *not* using  
                  the setpagedevice PostScript operator.  
   
 The variable `ps-spool-tumble' specifies how the page images on  
 opposite sides of a sheet are oriented with respect to each other.  If  
 `ps-spool-tumble' is nil, ps-print produces output suitable for  
 bindings on the left or right.  If `ps-spool-tumble' is non-nil,  
 ps-print produces output suitable for bindings at the top or bottom.  
 This variable takes effect only if `ps-spool-duplex' is non-nil.  
 The default value is nil.  
   
 The variable `ps-header-frame-alist' specifies a header frame  
 properties alist.  Valid frame properties are:  
   
   fore-color    Specify the foreground frame color.  
                 Value should be a float number between 0.0 (black  
                 color) and 1.0 (white color), or a string which is a  
                 color name, or a list of 3 float numbers which  
                 correspond to the Red Green Blue color scale, each  
                 float number between 0.0 (dark color) and 1.0 (bright  
                 color).  The default is 0 ("black").  
   
   back-color    Specify the background frame color (similar to fore-color).  
                 The default is 0.9 ("gray90").  
   
   shadow-color  Specify the shadow color (similar to fore-color).  
                 The default is 0 ("black").  
   
   border-color  Specify the border color (similar to fore-color).  
                 The default is 0 ("black").  
   
   border-width  Specify the border width.  
                 The default is 0.4.  
   
 Any other property is ignored.  
   
 Don't change this alist directly; instead use Custom, or the  
 `ps-value', `ps-get', `ps-put' and `ps-del' functions (see there for  
 documentation).  
   
 Ps-print can also print footers.  The footer variables are:  
 `ps-print-footer', `ps-footer-offset', `ps-print-footer-frame',  
 `ps-footer-font-family', `ps-footer-font-size', `ps-footer-line-pad',  
 `ps-footer-lines', `ps-left-footer', `ps-right-footer' and  
 `ps-footer-frame-alist'.  These variables are similar to those  
 controlling headers.  
   
 *** Color management (subgroup)  
   
 If `ps-print-color-p' is non-nil, the buffer's text will be printed in  
 color.  
   
 *** Face Management (subgroup)  
   
 If you need to print without worrying about face background colors,  
 set the variable `ps-use-face-background' which specifies if face  
 background should be used.  Valid values are:  
   
  t              always use face background color.  
  nil            never use face background color.  
  (face...)      list of faces whose background color will be used.  
   
 *** N-up printing (subgroup)  
   
 The variable `ps-n-up-printing' specifies the number of pages per  
 sheet of paper.  
   
 The variable `ps-n-up-margin' specifies the margin in points (pt)  
 between the sheet border and the n-up printing.  
   
 If variable `ps-n-up-border-p' is non-nil, a border is drawn around  
 each page.  
   
 The variable `ps-n-up-filling' specifies how the page matrix is filled  
 on each sheet of paper.  Following are the valid values for  
 `ps-n-up-filling' with a filling example using a 3x4 page matrix:  
   
    `left-top'   1  2  3  4         `left-bottom'    9  10 11 12  
                 5  6  7  8                          5  6  7  8  
                 9  10 11 12                         1  2  3  4  
   
    `right-top'  4  3  2  1         `right-bottom'   12 11 10 9  
                 8  7  6  5                          8  7  6  5  
                 12 11 10 9                          4  3  2  1  
   
    `top-left'   1  4  7  10        `bottom-left'    3  6  9  12  
                 2  5  8  11                         2  5  8  11  
                 3  6  9  12                         1  4  7  10  
   
    `top-right'  10 7  4  1         `bottom-right'   12 9  6  3  
                 11 8  5  2                          11 8  5  2  
                 12 9  6  3                          10 7  4  1  
   
 Any other value is treated as `left-top'.  
   
 *** Zebra stripes (subgroup)  
   
 The variable `ps-zebra-color' controls the zebra stripes grayscale or  
 RGB color.  
   
 The variable `ps-zebra-stripe-follow' specifies how zebra stripes  
 continue on next page.  Visually, valid values are (the character `+'  
 to the right of each column indicates that a line is printed):  
   
                    `nil'        `follow'        `full'        `full-follow'  
    Current Page --------     -----------     ---------     ----------------  
                 1  XXXXX +   1  XXXXXXXX +   1  XXXXXX +   1  XXXXXXXXXXXXX +  
                 2  XXXXX +   2  XXXXXXXX +   2  XXXXXX +   2  XXXXXXXXXXXXX +  
                 3  XXXXX +   3  XXXXXXXX +   3  XXXXXX +   3  XXXXXXXXXXXXX +  
                 4        +   4           +   4         +   4                +  
                 5        +   5           +   5         +   5                +  
                 6        +   6           +   6         +   6                +  
                 7  XXXXX +   7  XXXXXXXX +   7  XXXXXX +   7  XXXXXXXXXXXXX +  
                 8  XXXXX +   8  XXXXXXXX +   8  XXXXXX +   8  XXXXXXXXXXXXX +  
                 9  XXXXX +   9  XXXXXXXX +   9  XXXXXX +   9  XXXXXXXXXXXXX +  
                 10       +   10          +  
                 11       +   11          +  
                 --------     -----------     ---------     ----------------  
       Next Page --------     -----------     ---------     ----------------  
                 12 XXXXX +   12          +   10 XXXXXX +   10               +  
                 13 XXXXX +   13 XXXXXXXX +   11 XXXXXX +   11               +  
                 14 XXXXX +   14 XXXXXXXX +   12 XXXXXX +   12               +  
                 15       +   15 XXXXXXXX +   13        +   13 XXXXXXXXXXXXX +  
                 16       +   16          +   14        +   14 XXXXXXXXXXXXX +  
                 17       +   17          +   15        +   15 XXXXXXXXXXXXX +  
                 18 XXXXX +   18          +   16 XXXXXX +   16               +  
                 19 XXXXX +   19 XXXXXXXX +   17 XXXXXX +   17               +  
                 20 XXXXX +   20 XXXXXXXX +   18 XXXXXX +   18               +  
                 21       +   21 XXXXXXXX +  
                 22       +   22          +  
                 --------     -----------     ---------     ----------------  
   
 Any other value is treated as `nil'.  
   
   
 *** Printer management (subgroup)  
   
 The variable `ps-printer-name-option' determines the option used by  
 some utilities to indicate the printer name; it's used only when  
 `ps-printer-name' is a non-empty string.  If you're using the lpr  
 utility to print, for example, `ps-printer-name-option' should be set  
 to "-P".  
   
 The variable `ps-manual-feed' indicates if the printer requires manual  
 paper feeding.  If it's nil, automatic feeding takes place.  If it's  
 non-nil, manual feeding takes place.  
   
 The variable `ps-end-with-control-d' specifies whether C-d (\x04)  
 should be inserted at end of the generated PostScript.  Non-nil means  
 do so.  
   
 *** Page settings (subgroup)  
   
 If variable `ps-warn-paper-type' is nil, it's *not* treated as an  
 error if the PostScript printer doesn't have a paper with the size  
 indicated by `ps-paper-type'; the default paper size will be used  
 instead.  If `ps-warn-paper-type' is non-nil, an error is signaled if  
 the PostScript printer doesn't support a paper with the size indicated  
 by `ps-paper-type'.  This is used when `ps-spool-config' is set to  
 `setpagedevice'.  
   
 The variable `ps-print-upside-down' determines the orientation for  
 printing pages: nil means `normal' printing, non-nil means  
 `upside-down' printing (that is, the page is rotated by 180 degrees).  
   
 The variable `ps-selected-pages' specifies which pages to print.  If  
 it's nil, all pages are printed.  If it's a list, list elements may be  
 integers specifying a single page to print, or cons cells (FROM . TO)  
 specifying to print from page FROM to TO.  Invalid list elements, that  
 is integers smaller than one, or elements whose FROM is greater than  
 its TO, are ignored.  
   
 The variable `ps-even-or-odd-pages' specifies how to print even/odd  
 pages.  Valid values are:  
   
    nil          print all pages.  
   
    `even-page'  print only even pages.  
   
    `odd-page'   print only odd pages.  
   
    `even-sheet' print only even sheets.  
                 That is, if `ps-n-up-printing' is 1, it behaves like  
                 `even-page', but for values greater than 1, it'll  
                 print only the even sheet of paper.  
   
    `odd-sheet'  print only odd sheets.  
                 That is, if `ps-n-up-printing' is 1, it behaves like  
                 `odd-page'; but for values greater than 1, it'll print  
                 only the odd sheet of paper.  
   
 Any other value is treated as nil.  
   
 If you set `ps-selected-pages' (see there for documentation), pages  
 are filtered by `ps-selected-pages', and then by  
 `ps-even-or-odd-pages'.  For example, if we have:  
   
    (setq ps-selected-pages '(1 4 (6 . 10) (12 . 16) 20))  
   
 and we combine this with `ps-even-or-odd-pages' and  
 `ps-n-up-printing', we get:  
   
 `ps-n-up-printing' = 1:  
    `ps-even-or-odd-pages'       PAGES PRINTED  
         nil                     1, 4, 6, 7, 8, 9, 10, 12, 13, 14, 15, 16, 20  
         even-page               4, 6, 8, 10, 12, 14, 16, 20  
         odd-page                1, 7, 9, 13, 15  
         even-sheet              4, 6, 8, 10, 12, 14, 16, 20  
         odd-sheet               1, 7, 9, 13, 15  
   
 `ps-n-up-printing' = 2:  
    `ps-even-or-odd-pages'       PAGES PRINTED  
         nil                     1/4, 6/7, 8/9, 10/12, 13/14, 15/16, 20  
         even-page               4/6, 8/10, 12/14, 16/20  
         odd-page                1/7, 9/13, 15  
         even-sheet              6/7, 10/12, 15/16  
         odd-sheet               1/4, 8/9, 13/14, 20  
   
 *** Miscellany (subgroup)  
   
 The variable `ps-error-handler-message' specifies where error handler  
 messages should be sent.  
   
 It is also possible to add a user-defined PostScript prologue code in  
 front of all generated prologue code by setting the variable  
 `ps-user-defined-prologue'.  
   
 The variable `ps-line-number-font' specifies the font for line numbers.  
   
 The variable `ps-line-number-font-size' specifies the font size in  
 points for line numbers.  
   
 The variable `ps-line-number-color' specifies the color for line  
 numbers.  See `ps-zebra-color' for documentation.  
   
 The variable `ps-line-number-step' specifies the interval in which  
 line numbers are printed.  For example, if `ps-line-number-step' is set  
 to 2, the printing will look like:  
   
    1 one line  
      one line  
    3 one line  
      one line  
    5 one line  
      one line  
      ...  
   
 Valid values are:  
   
 integer         an integer specifying the interval in which line numbers are  
                 printed.  If it's smaller than or equal to zero, 1  
                 is used.  
   
 `zebra'         specifies that only the line number of the first line in a  
                 zebra stripe is to be printed.  
   
 Any other value is treated as `zebra'.  
   
 The variable `ps-line-number-start' specifies the starting point in  
 the interval given by `ps-line-number-step'.  For example, if  
 `ps-line-number-step' is set to 3, and `ps-line-number-start' is set to  
 3, the output will look like:  
   
      one line  
      one line  
    3 one line  
      one line  
      one line  
    6 one line  
      one line  
      one line  
    9 one line  
      one line  
      ...  
   
 The variable `ps-postscript-code-directory' specifies the directory  
 where the PostScript prologue file used by ps-print is found.  
   
 The variable `ps-line-spacing' determines the line spacing in points,  
 for ordinary text, when generating PostScript (similar to  
 `ps-font-size').  
   
 The variable `ps-paragraph-spacing' determines the paragraph spacing,  
 in points, for ordinary text, when generating PostScript (similar to  
 `ps-font-size').  
   
 The variable `ps-paragraph-regexp' specifies the paragraph delimiter.  
   
 The variable `ps-begin-cut-regexp' and `ps-end-cut-regexp' specify the  
 start and end of a region to cut out when printing.  
   
 ** hideshow changes.  
   
 *** now supports hiding of blocks of single line comments (like // for  
 C++, ; for lisp).  
   
 *** Support for java-mode added.  
   
 *** When doing `hs-hide-all' it is now possible to also hide the comments  
 in the file if `hs-hide-comments-when-hiding-all' is set.  
   
 *** The new function `hs-hide-initial-comment' hides the comments at  
 the beginning of the files.  Finally those huge RCS logs don't stay in your  
 way!  This is run by default when entering the `hs-minor-mode'.  
   
 *** Now uses overlays instead of `selective-display', so is more  
 robust and a lot faster.  
   
 *** A block beginning can span multiple lines.  
   
 *** The new variable `hs-show-hidden-short-form' if t, directs hideshow  
 to show only the beginning of a block when it is hidden.  See the  
 documentation for more details.  
   
 ** Changes in Enriched mode.  
   
 *** When you visit a file in enriched-mode, Emacs will make sure it is  
 filled to the current fill-column.  This behavior is now independent  
 of the size of the window.  When you save the file, the fill-column in  
 use is stored as well, so that the whole buffer need not be refilled  
 the next time unless the fill-column is different.  
   
 *** use-hard-newlines is now a minor mode.  When it is enabled, Emacs  
 distinguishes between hard and soft newlines, and treats hard newlines  
 as paragraph boundaries.  Otherwise all newlines inserted are marked  
 as soft, and paragraph boundaries are determined solely from the text.  
   
 ** Font Lock mode  
   
 *** Custom support  
   
 The variables font-lock-face-attributes, font-lock-display-type and  
 font-lock-background-mode are now obsolete; the recommended way to specify the  
 faces to use for Font Lock mode is with M-x customize-group on the new custom  
 group font-lock-highlighting-faces.  If you set font-lock-face-attributes in  
 your ~/.emacs file, Font Lock mode will respect its value.  However, you should  
 consider converting from setting that variable to using M-x customize.  
   
 You can still use X resources to specify Font Lock face appearances.  
   
 *** Maximum decoration  
   
 Fontification now uses the maximum level of decoration supported by  
 default.  Previously, fontification used a mode-specific default level  
 of decoration, which is typically the minimum level of decoration  
 supported.  You can set font-lock-maximum-decoration to nil  
 to get the old behavior.  
   
 *** New support  
   
 Support is now provided for Java, Objective-C, AWK and SIMULA modes.  
   
 Note that Font Lock mode can be turned on without knowing exactly what modes  
 support Font Lock mode, via the command global-font-lock-mode.  
   
 *** Configurable support  
   
 Support for C, C++, Objective-C and Java can be more easily configured for  
 additional types and classes via the new variables c-font-lock-extra-types,  
 c++-font-lock-extra-types, objc-font-lock-extra-types and, you guessed it,  
 java-font-lock-extra-types.  These value of each of these variables should be a  
 list of regexps matching the extra type names.  For example, the default value  
 of c-font-lock-extra-types is ("\\sw+_t") which means fontification follows the  
 convention that C type names end in _t.  This results in slower fontification.  
   
 Of course, you can change the variables that specify fontification in whatever  
 way you wish, typically by adding regexps.  However, these new variables make  
 it easier to make specific and common changes for the fontification of types.  
   
 *** Adding highlighting patterns to existing support  
   
 You can use the new function font-lock-add-keywords to add your own  
 highlighting patterns, such as for project-local or user-specific constructs,  
 for any mode.  
   
 For example, to highlight `FIXME:' words in C comments, put:  
   
  (font-lock-add-keywords 'c-mode '(("\\<FIXME:" 0 font-lock-warning-face t)))  
   
 in your ~/.emacs.  
   
 *** New faces  
   
 Font Lock now defines two new faces, font-lock-builtin-face and  
 font-lock-warning-face.  These are intended to highlight builtin keywords,  
 distinct from a language's normal keywords, and objects that should be brought  
 to user attention, respectively.  Various modes now use these new faces.  
   
 *** Changes to fast-lock support mode  
   
 The fast-lock package, one of the two Font Lock support modes, can now process  
 cache files silently.  You can use the new variable fast-lock-verbose, in the  
 same way as font-lock-verbose, to control this feature.  
   
 *** Changes to lazy-lock support mode  
   
 The lazy-lock package, one of the two Font Lock support modes, can now fontify  
 according to the true syntactic context relative to other lines.  You can use  
 the new variable lazy-lock-defer-contextually to control this feature.  If  
 non-nil, changes to the buffer will cause subsequent lines in the buffer to be  
 refontified after lazy-lock-defer-time seconds of idle time.  If nil, then only  
 the modified lines will be refontified; this is the same as the previous Lazy  
 Lock mode behaviour and the behaviour of Font Lock mode.  
   
 This feature is useful in modes where strings or comments can span lines.  
 For example, if a string or comment terminating character is deleted, then if  
 this feature is enabled subsequent lines in the buffer will be correctly  
 refontified to reflect their new syntactic context.  Previously, only the line  
 containing the deleted character would be refontified and you would have to use  
 the command M-g M-g (font-lock-fontify-block) to refontify some lines.  
   
 As a consequence of this new feature, two other variables have changed:  
   
 Variable `lazy-lock-defer-driven' is renamed `lazy-lock-defer-on-scrolling'.  
 Variable `lazy-lock-defer-time' can now only be a time, i.e., a number.  
 Buffer modes for which on-the-fly deferral applies can be specified via the  
 new variable `lazy-lock-defer-on-the-fly'.  
   
 If you set these variables in your ~/.emacs, then you may have to change those  
 settings.  
   
 ** Ada mode changes.  
   
 *** There is now better support for using find-file.el with Ada mode.  
 If you switch between spec and body, the cursor stays in the same  
 procedure (modulo overloading).  If a spec has no body file yet, but  
 you try to switch to its body file, Ada mode now generates procedure  
 stubs.  
   
 *** There are two new commands:  
  - `ada-make-local'   : invokes gnatmake on the current buffer  
  - `ada-check-syntax' : check syntax of current buffer.  
   
 The user options `ada-compiler-make', `ada-make-options',  
 `ada-language-version', `ada-compiler-syntax-check', and  
 `ada-compile-options' are used within these commands.  
   
 *** Ada mode can now work with Outline minor mode.  The outline level  
 is calculated from the indenting, not from syntactic constructs.  
 Outlining does not work if your code is not correctly indented.  
   
 *** The new function `ada-gnat-style' converts the buffer to the style of  
 formatting used in GNAT.  It places two blanks after a comment start,  
 places one blank between a word end and an opening '(', and puts one  
 space between a comma and the beginning of a word.  
   
 ** Scheme mode changes.  
   
 *** Scheme mode indentation now uses many of the facilities of Lisp  
 mode; therefore, the variables to customize it are the variables used  
 for Lisp mode which have names starting with `lisp-'.  The variables  
 with names starting with `scheme-' which used to do this no longer  
 have any effect.  
   
 If you want to use different indentation for Scheme and Lisp, this is  
 still possible, but now you must do it by adding a hook to  
 scheme-mode-hook, which could work by setting the `lisp-' indentation  
 variables as buffer-local variables.  
   
 *** DSSSL mode is a variant of Scheme mode, for editing DSSSL scripts.  
 Use M-x dsssl-mode.  
   
 ** Changes to the emacsclient program  
   
 *** If a socket can't be found, and environment variables LOGNAME or  
 USER are set, emacsclient now looks for a socket based on the UID  
 associated with the name.  That is an emacsclient running as root  
 can connect to an Emacs server started by a non-root user.  
   
 *** The emacsclient program now accepts an option --no-wait which tells  
 it to return immediately without waiting for you to "finish" the  
 buffer in Emacs.  
   
 *** The new option --alternate-editor allows to specify an editor to  
 use if Emacs is not running.  The environment variable  
 ALTERNATE_EDITOR can be used for the same effect; the command line  
 option takes precedence.  
   
 ** M-x eldoc-mode enables a minor mode in which the echo area  
 constantly shows the parameter list for function being called at point  
 (in Emacs Lisp and Lisp Interaction modes only).  
   
 ** C-x n d now runs the new command narrow-to-defun,  
 which narrows the accessible parts of the buffer to just  
 the current defun.  
   
 ** Emacs now handles the `--' argument in the standard way; all  
 following arguments are treated as ordinary file names.  
   
 ** On MSDOS and Windows, the bookmark file is now called _emacs.bmk,  
 and the saved desktop file is now called _emacs.desktop (truncated if  
 necessary).  
   
 ** When you kill a buffer that visits a file,  
 if there are any registers that save positions in the file,  
 these register values no longer become completely useless.  
 If you try to go to such a register with C-x j, then you are  
 asked whether to visit the file again.  If you say yes,  
 it visits the file and then goes to the same position.  
   
 ** When you visit a file that changes frequently outside Emacs--for  
 example, a log of output from a process that continues to run--it may  
 be useful for Emacs to revert the file without querying you whenever  
 you visit the file afresh with C-x C-f.  
   
 You can request this behavior for certain files by setting the  
 variable revert-without-query to a list of regular expressions.  If a  
 file's name matches any of these regular expressions, find-file and  
 revert-buffer revert the buffer without asking for permission--but  
 only if you have not edited the buffer text yourself.  
   
 ** set-default-font has been renamed to set-frame-font  
 since it applies only to the current frame.  
   
 ** In TeX mode, you can use the variable tex-main-file to specify the  
 file for tex-file to run TeX on.  (By default, tex-main-file is nil,  
 and tex-file runs TeX on the current visited file.)  
   
 This is useful when you are editing a document that consists of  
 multiple files.  In each of the included files, you can set up a local  
 variable list which specifies the top-level file of your document for  
 tex-main-file.  Then tex-file will run TeX on the whole document  
 instead of just the file you are editing.  
   
 ** RefTeX mode  
   
 RefTeX mode is a new minor mode with special support for \label, \ref  
 and \cite macros in LaTeX documents.  RefTeX distinguishes labels of  
 different environments (equation, figure, ...) and has full support for  
 multifile documents.  To use it, select a buffer with a LaTeX document and  
 turn the mode on with M-x reftex-mode.  Here are the main user commands:  
   
 C-c (    reftex-label  
    Creates a label semi-automatically.  RefTeX is context sensitive and  
    knows which kind of label is needed.  
   
 C-c )    reftex-reference  
    Offers in a menu all labels in the document, along with context of the  
    label definition.  The selected label is referenced as \ref{LABEL}.  
   
 C-c [    reftex-citation  
    Prompts for a regular expression and displays a list of matching BibTeX  
    database entries.  The selected entry is cited with a \cite{KEY} macro.  
   
 C-c &    reftex-view-crossref  
    Views the cross reference of a \ref or \cite command near point.  
   
 C-c =    reftex-toc  
    Shows a table of contents of the (multifile) document.  From there you  
    can quickly jump to every section.  
   
 Under X, RefTeX installs a "Ref" menu in the menu bar, with additional  
 commands.  Press `?' to get help when a prompt mentions this feature.  
 Full documentation and customization examples are in the file  
 reftex.el.  You can use the finder to view the file documentation:  
 C-h p --> tex --> reftex.el  
   
 ** Changes in BibTeX mode.  
   
 *** Info documentation is now available.  
   
 *** Don't allow parentheses in string constants anymore.  This confused  
 both the BibTeX program and Emacs BibTeX mode.  
   
 *** Renamed variable bibtex-mode-user-optional-fields to  
 bibtex-user-optional-fields.  
   
 *** Removed variable bibtex-include-OPTannote  
 (use bibtex-user-optional-fields instead).  
   
 *** New interactive functions to copy and kill fields and complete  
 entries to the BibTeX kill ring, from where they can be yanked back by  
 appropriate functions.  
   
 *** New interactive functions for repositioning and marking of  
 entries. They are bound by default to C-M-l and C-M-h.  
   
 *** New hook bibtex-clean-entry-hook. It is called after entry has  
 been cleaned.  
   
 *** New variable bibtex-field-delimiters, which replaces variables  
 bibtex-field-{left|right}-delimiter.  
   
 *** New variable bibtex-entry-delimiters to determine how entries  
 shall be delimited.  
   
 *** Allow preinitialization of fields. See documentation of  
 bibtex-user-optional-fields, bibtex-entry-field-alist, and  
 bibtex-include-OPTkey for details.  
   
 *** Book and InBook entries require either an author or an editor  
 field. This is now supported by bibtex.el. Alternative fields are  
 prefixed with `ALT'.  
   
 *** New variable bibtex-entry-format, which replaces variable  
 bibtex-clean-entry-zap-empty-opts and allows specification of many  
 formatting options performed on cleaning an entry (see variable  
 documentation).  
   
 *** Even more control on how automatic keys are generated. See  
 documentation of bibtex-generate-autokey for details. Transcriptions  
 for foreign languages other than German are now handled, too.  
   
 *** New boolean user option bibtex-comma-after-last-field to decide if  
 comma should be inserted at end of last field.  
   
 *** New boolean user option bibtex-align-at-equal-sign to determine if  
 alignment should be made at left side of field contents or at equal  
 signs. New user options to control entry layout (e.g. indentation).  
   
 *** New function bibtex-fill-entry to realign entries.  
   
 *** New function bibtex-reformat to reformat region or buffer.  
   
 *** New function bibtex-convert-alien to convert a BibTeX database  
 from alien sources.  
   
 *** New function bibtex-complete-key (similar to bibtex-complete-string)  
 to complete prefix to a key defined in buffer. Mainly useful in  
 crossref entries.  
   
 *** New function bibtex-count-entries to count entries in buffer or  
 region.  
   
 *** Added support for imenu.  
   
 *** The function `bibtex-validate' now checks current region instead  
 of buffer if mark is active. Now it shows all errors of buffer in a  
 `compilation mode' buffer. You can use the normal commands (e.g.  
 `next-error') for compilation modes to jump to errors.  
   
 *** New variable `bibtex-string-file-path' to determine where the files  
 from `bibtex-string-files' are searched.  
   
 ** Iso Accents mode now supports Latin-3 as an alternative.  
   
 ** The command next-error now opens blocks hidden by hideshow.  
   
 ** The function using-unix-filesystems has been replaced by the  
 functions add-untranslated-filesystem and remove-untranslated-filesystem.  
 Each of these functions takes the name of a drive letter or directory  
 as an argument.  
   
 When a filesystem is added as untranslated, all files on it are read  
 and written in binary mode (no cr/lf translation is performed).  
   
 ** browse-url changes  
   
 *** New methods for: Grail (browse-url-generic), MMM (browse-url-mmm),  
 Lynx in a separate xterm (browse-url-lynx-xterm) or in an Emacs window  
 (browse-url-lynx-emacs), remote W3 (browse-url-w3-gnudoit), generic  
 non-remote-controlled browsers (browse-url-generic) and associated  
 customization variables.  
   
 *** New commands `browse-url-of-region' and `browse-url'.  
   
 *** URLs marked up with <URL:...> (RFC1738) work if broken across  
 lines.  Browsing methods can be associated with URL regexps  
 (e.g. mailto: URLs) via `browse-url-browser-function'.  
   
 ** Changes in Ediff  
   
 *** Clicking Mouse-2 on a brief command description in Ediff control panel  
 pops up the Info file for this command.  
   
 *** There is now a variable, ediff-autostore-merges, which controls whether  
 the result of a merge is saved in a file. By default, this is done only when  
 merge is done from a session group (eg, when merging files in two different  
 directories).  
   
 *** Since Emacs 19.31 (this hasn't been announced before), Ediff can compare  
 and merge groups of files residing in different directories, or revisions of  
 files in the same directory.  
   
 *** Since Emacs 19.31, Ediff can apply multi-file patches interactively.  
 The patches must be in the context format or GNU unified format.  (The bug  
 related to the GNU format has now been fixed.)  
   
 ** Changes in Viper  
   
 *** The startup file is now .viper instead of .vip  
 *** All variable/function names have been changed to start with viper-  
     instead of vip-.  
 *** C-\ now simulates the meta-key in all Viper states.  
 *** C-z in Insert state now escapes to Vi for the duration of the next  
 Viper command. In Vi and Insert states, C-z behaves as before.  
 *** C-c \ escapes to Vi for one command if Viper is in Insert or Emacs states.  
 *** _ is no longer the meta-key in Vi state.  
 *** The variable viper-insert-state-cursor-color can be used to change cursor  
 color when Viper is in insert state.  
 *** If search lands the cursor near the top or the bottom of the window,  
 Viper pulls the window up or down to expose more context. The variable  
 viper-adjust-window-after-search controls this behavior.  
   
 ** Etags changes.  
   
 *** In C, C++, Objective C and Java, Etags tags global variables by  
 default.  The resulting tags files are inflated by 30% on average.  
 Use --no-globals to turn this feature off.  Etags can also tag  
 variables which are members of structure-like constructs, but it does  
 not by default.  Use --members to turn this feature on.  
   
 *** C++ member functions are now recognized as tags.  
   
 *** Java is tagged like C++.  In addition, "extends" and "implements"  
 constructs are tagged.  Files are recognised by the extension .java.  
   
 *** Etags can now handle programs written in Postscript.  Files are  
 recognised by the extensions .ps and .pdb (Postscript with C syntax).  
 In Postscript, tags are lines that start with a slash.  
   
 *** Etags now handles Objective C and Objective C++ code.  The usual C and  
 C++ tags are recognized in these languages; in addition, etags  
 recognizes special Objective C syntax for classes, class categories,  
 methods and protocols.  
   
 *** Etags also handles Cobol.  Files are recognised by the extension  
 .cobol.  The tagged lines are those containing a word that begins in  
 column 8 and ends in a full stop, i.e. anything that could be a  
 paragraph name.  
   
 *** Regexps in Etags now support intervals, as in ed or grep.  The syntax of  
 an interval is \{M,N\}, and it means to match the preceding expression  
 at least M times and as many as N times.  
   
 ** The format for specifying a custom format for time-stamp to insert  
 in files has changed slightly.  
   
 With the new enhancements to the functionality of format-time-string,  
 time-stamp-format will change to be eventually compatible with it.  
 This conversion is being done in two steps to maintain compatibility  
 with old time-stamp-format values.  
   
 In the new scheme, alternate case is signified by the number-sign  
 (`#') modifier, rather than changing the case of the format character.  
 This feature is as yet incompletely implemented for compatibility  
 reasons.  
   
 In the old time-stamp-format, all numeric fields defaulted to their  
 natural width.  (With format-time-string, each format has a  
 fixed-width default.)  In this version, you can specify the colon  
 (`:') modifier to a numeric conversion to mean "give me the historical  
 time-stamp-format width default."  Do not use colon if you are  
 specifying an explicit width, as in "%02d".  
   
 Numbers are no longer truncated to the requested width, except in the  
 case of "%02y", which continues to give a two-digit year.  Digit  
 truncation probably wasn't being used for anything else anyway.  
   
 The new formats will work with old versions of Emacs.  New formats are  
 being recommended now to allow time-stamp-format to change in the  
 future to be compatible with format-time-string.  The new forms being  
 recommended now will continue to work then.  
   
 See the documentation string for the variable time-stamp-format for  
 details.  
   
 ** There are some additional major modes:  
   
 dcl-mode, for editing VMS DCL files.  
 m4-mode, for editing files of m4 input.  
 meta-mode, for editing MetaFont and MetaPost source files.  
   
 ** In Shell mode, the command shell-copy-environment-variable lets you  
 copy the value of a specified environment variable from the subshell  
 into Emacs.  
   
 ** New Lisp packages include:  
   
 *** battery.el displays battery status for laptops.  
   
 *** M-x bruce (named after Lenny Bruce) is a program that might  
 be used for adding some indecent words to your email.  
   
 *** M-x crisp-mode enables an emulation for the CRiSP editor.  
   
 *** M-x dirtrack arranges for better tracking of directory changes  
 in shell buffers.  
   
 *** The new library elint.el provides for linting of Emacs Lisp code.  
 See the documentation for `elint-initialize', `elint-current-buffer'  
 and `elint-defun'.  
   
 *** M-x expand-add-abbrevs defines a special kind of abbrev which is  
 meant for programming constructs.  These abbrevs expand like ordinary  
 ones, when you type SPC, but only at the end of a line and not within  
 strings or comments.  
   
 These abbrevs can act as templates: you can define places within an  
 abbrev for insertion of additional text.  Once you expand the abbrev,  
 you can then use C-x a p and C-x a n to move back and forth to these  
 insertion points.  Thus you can conveniently insert additional text  
 at these points.  
   
 *** filecache.el remembers the location of files so that you  
 can visit them by short forms of their names.  
   
 *** find-func.el lets you find the definition of the user-loaded  
 Emacs Lisp function at point.  
   
 *** M-x handwrite converts text to a "handwritten" picture.  
   
 *** M-x iswitchb-buffer is a command for switching to a buffer, much like  
 switch-buffer, but it reads the argument in a more helpful way.  
   
 *** M-x landmark implements a neural network for landmark learning.  
   
 *** M-x locate provides a convenient interface to the `locate' program.  
   
 *** M4 mode is a new mode for editing files of m4 input.  
   
 *** mantemp.el creates C++ manual template instantiations  
 from the GCC error messages which indicate which instantiations are needed.  
   
 *** mouse-copy.el provides a one-click copy and move feature.  
 You can drag a region with M-mouse-1, and it is automatically  
 inserted at point.  M-Shift-mouse-1 deletes the text from its  
 original place after inserting the copy.  
   
 *** mouse-drag.el lets you do scrolling by dragging Mouse-2  
 on the buffer.  
   
 You click the mouse and move; that distance either translates into the  
 velocity to scroll (with mouse-drag-throw) or the distance to scroll  
 (with mouse-drag-drag).  Horizontal scrolling is enabled when needed.  
   
 Enable mouse-drag with:  
     (global-set-key [down-mouse-2] 'mouse-drag-throw)  
 -or-  
     (global-set-key [down-mouse-2] 'mouse-drag-drag)  
   
 *** mspools.el is useful for determining which mail folders have  
 mail waiting to be read in them.  It works with procmail.  
   
 *** Octave mode is a major mode for editing files of input for Octave.  
 It comes with a facility for communicating with an Octave subprocess.  
   
 *** ogonek  
   
 The ogonek package provides functions for changing the coding of  
 Polish diacritic characters in buffers.  Codings known from various  
 platforms are supported such as ISO8859-2, Mazovia, IBM Latin2, and  
 TeX.  For example, you can change the coding from Mazovia to  
 ISO8859-2.  Another example is a change of coding from ISO8859-2 to  
 prefix notation (in which `/a' stands for the aogonek character, for  
 instance) and vice versa.  
   
 To use this package load it using  
     M-x load-library [enter] ogonek  
 Then, you may get an explanation by calling one of  
     M-x ogonek-jak        -- in Polish  
     M-x ogonek-how        -- in English  
 The info specifies the commands and variables provided as well as the  
 ways of customization in `.emacs'.  
   
 *** Interface to ph.  
   
 Emacs provides a client interface to CCSO Nameservers (ph/qi)  
   
 The CCSO nameserver is used in many universities to provide directory  
 services about people.  ph.el provides a convenient Emacs interface to  
 these servers.  
   
 *** uce.el is useful for replying to unsolicited commercial email.  
   
 *** vcursor.el implements a "virtual cursor" feature.  
 You can move the virtual cursor with special commands  
 while the real cursor does not move.  
   
 *** webjump.el is a "hot list" package which you can set up  
 for visiting your favorite web sites.  
   
 *** M-x winner-mode is a minor mode which saves window configurations,  
 so you can move back to other configurations that you have recently used.  
   
 ** movemail change  
   
 Movemail no longer needs to be installed setuid root in order for POP  
 mail retrieval to function properly.  This is because it no longer  
 supports the RPOP (reserved-port POP) protocol; instead, it uses the  
 user's POP password to authenticate to the mail server.  
   
 This change was made earlier, but not reported in NEWS before.  
   
 * Emacs 20.1 changes for MS-DOS and MS-Windows.  
   
 ** Changes in handling MS-DOS/MS-Windows text files.  
   
 Emacs handles three different conventions for representing  
 end-of-line: CRLF for MSDOS, LF for Unix and GNU, and CR (used on the  
 Macintosh).  Emacs determines which convention is used in a specific  
 file based on the contents of that file (except for certain special  
 file names), and when it saves the file, it uses the same convention.  
   
 To save the file and change the end-of-line convention, you can use  
 C-x RET f (set-buffer-file-coding-system) to specify a different  
 coding system for the buffer.  Then, when you save the file, the newly  
 specified coding system will take effect.  For example, to save with  
 LF, specify undecided-unix (or some other ...-unix coding system); to  
 save with CRLF, specify undecided-dos.  
   
 * Lisp Changes in Emacs 20.1  
   
 ** Byte-compiled files made with Emacs 20 will, in general, work in  
 Emacs 19 as well, as long as the source code runs in Emacs 19.  And  
 vice versa: byte-compiled files made with Emacs 19 should also run in  
 Emacs 20, as long as the program itself works in Emacs 20.  
   
 ** Windows-specific functions and variables have been renamed  
 to start with w32- instead of win32-.  
   
 In hacker language, calling something a "win" is a form of praise.  We  
 don't want to praise a non-free Microsoft system, so we don't call it  
 "win".  
   
 ** Basic Lisp changes  
   
 *** A symbol whose name starts with a colon now automatically  
 evaluates to itself.  Therefore such a symbol can be used as a constant.  
   
 *** The defined purpose of `defconst' has been changed.  It should now  
 be used only for values that should not be changed whether by a program  
 or by the user.  
   
 The actual behavior of defconst has not been changed.  
   
 *** There are new macros `when' and `unless'  
   
 (when CONDITION BODY...)  is short for  (if CONDITION (progn BODY...))  
 (unless CONDITION BODY...)  is short for  (if CONDITION nil BODY...)  
   
 *** Emacs now defines functions caar, cadr, cdar and cddr with their  
 usual Lisp meanings.  For example, caar returns the car of the car of  
 its argument.  
   
 *** equal, when comparing strings, now ignores their text properties.  
   
 *** The new function `functionp' tests whether an object is a function.  
   
 *** arrayp now returns t for char-tables and bool-vectors.  
   
 *** Certain primitives which use characters (as integers) now get an  
 error if the integer is not a valid character code.  These primitives  
 include insert-char, char-to-string, and the %c construct in the  
 `format' function.  
   
 *** The `require' function now insists on adding a suffix, either .el  
 or .elc, to the file name.  Thus, (require 'foo) will not use a file  
 whose name is just foo.  It insists on foo.el or foo.elc.  
   
 *** The `autoload' function, when the file name does not contain  
 either a directory name or the suffix .el or .elc, insists on  
 adding one of these suffixes.  
   
 *** string-to-number now takes an optional second argument BASE  
 which specifies the base to use when converting an integer.  
 If BASE is omitted, base 10 is used.  
   
 We have not implemented other radices for floating point numbers,  
 because that would be much more work and does not seem useful.  
   
 *** substring now handles vectors as well as strings.  
   
 *** The Common Lisp function eql is no longer defined normally.  
 You must load the `cl' library to define it.  
   
 *** The new macro `with-current-buffer' lets you evaluate an expression  
 conveniently with a different current buffer.  It looks like this:  
   
   (with-current-buffer BUFFER BODY-FORMS...)  
   
 BUFFER is the expression that says which buffer to use.  
 BODY-FORMS say what to do in that buffer.  
   
 *** The new primitive `save-current-buffer' saves and restores the  
 choice of current buffer, like `save-excursion', but without saving or  
 restoring the value of point or the mark.  `with-current-buffer'  
 works using `save-current-buffer'.  
   
 *** The new macro `with-temp-file' lets you do some work in a new buffer and  
 write the output to a specified file.  Like `progn', it returns the value  
 of the last form.  
   
 *** The new macro `with-temp-buffer' lets you do some work in a new buffer,  
 which is discarded after use.  Like `progn', it returns the value of the  
 last form.  If you wish to return the buffer contents, use (buffer-string)  
 as the last form.  
   
 *** The new function split-string takes a string, splits it at certain  
 characters, and returns a list of the substrings in between the  
 matches.  
   
 For example, (split-string "foo bar lose" " +") returns ("foo" "bar" "lose").  
   
 *** The new macro with-output-to-string executes some Lisp expressions  
 with standard-output set up so that all output feeds into a string.  
 Then it returns that string.  
   
 For example, if the current buffer name is `foo',  
   
 (with-output-to-string  
   (princ "The buffer is ")  
   (princ (buffer-name)))  
   
 returns "The buffer is foo".  
   
 ** Non-ASCII characters are now supported, if enable-multibyte-characters  
 is non-nil.  
   
 These characters have character codes above 256.  When inserted in the  
 buffer or stored in a string, they are represented as multibyte  
 characters that occupy several buffer positions each.  
   
 *** When enable-multibyte-characters is non-nil, a single character in  
 a buffer or string can be two or more bytes (as many as four).  
   
 Buffers and strings are still made up of unibyte elements;  
 character positions and string indices are always measured in bytes.  
 Therefore, moving forward one character can increase the buffer  
 position by 2, 3 or 4.  The function forward-char moves by whole  
 characters, and therefore is no longer equivalent to  
   (lambda (n) (goto-char (+ (point) n))).  
   
 ASCII characters (codes 0 through 127) are still single bytes, always.  
 Sequences of byte values 128 through 255 are used to represent  
 non-ASCII characters.  These sequences are called "multibyte  
 characters".  
   
 The first byte of a multibyte character is always in the range 128  
 through 159 (octal 0200 through 0237).  These values are called  
 "leading codes".  The second and subsequent bytes are always in the  
 range 160 through 255 (octal 0240 through 0377).  The first byte, the  
 leading code, determines how many bytes long the sequence is.  
   
 *** The function forward-char moves over characters, and therefore  
 (forward-char 1) may increase point by more than 1 if it moves over a  
 multibyte character.  Likewise, delete-char always deletes a  
 character, which may be more than one buffer position.  
   
 This means that some Lisp programs, which assume that a character is  
 always one buffer position, need to be changed.  
   
 However, all ASCII characters are always one buffer position.  
   
 *** The regexp [\200-\377] no longer matches all non-ASCII characters,  
 because when enable-multibyte-characters is non-nil, these characters  
 have codes that are not in the range octal 200 to octal 377.  However,  
 the regexp [^\000-\177] does match all non-ASCII characters,  
 guaranteed.  
   
 *** The function char-boundary-p returns non-nil if position POS is  
 between two characters in the buffer (not in the middle of a  
 character).  
   
 When the value is non-nil, it says what kind of character follows POS:  
   
  0 if POS is at an ASCII character or at the end of range,  
  1 if POS is before a 2-byte length multi-byte form,  
  2 if POS is at a head of 3-byte length multi-byte form,  
  3 if POS is at a head of 4-byte length multi-byte form,  
  4 if POS is at a head of multi-byte form of a composite character.  
   
 *** The function char-bytes returns how many bytes the character CHAR uses.  
   
 *** Strings can contain multibyte characters.  The function  
 `length' returns the string length counting bytes, which may be  
 more than the number of characters.  
   
 You can include a multibyte character in a string constant by writing  
 it literally.  You can also represent it with a hex escape,  
 \xNNNNNNN..., using as many digits as necessary.  Any character which  
 is not a valid hex digit terminates this construct.  If you want to  
 follow it with a character that is a hex digit, write backslash and  
 newline in between; that will terminate the hex escape.  
   
 *** The function concat-chars takes arguments which are characters  
 and returns a string containing those characters.  
   
 *** The function sref access a multibyte character in a string.  
 (sref STRING INDX) returns the character in STRING at INDEX.  INDEX  
 counts from zero.  If INDEX is at a position in the middle of a  
 character, sref signals an error.  
   
 *** The function chars-in-string returns the number of characters  
 in a string.  This is less than the length of the string, if the  
 string contains multibyte characters (the length counts bytes).  
   
 *** The function chars-in-region returns the number of characters  
 in a region from BEG to END.  This is less than (- END BEG) if the  
 region contains multibyte characters (the length counts bytes).  
   
 *** The function string-to-list converts a string to a list of  
 the characters in it.  string-to-vector converts a string  
 to a vector of the characters in it.  
   
 *** The function store-substring alters part of the contents  
 of a string.  You call it as follows:  
   
    (store-substring STRING IDX OBJ)  
   
 This says to alter STRING, by storing OBJ starting at index IDX in  
 STRING.  OBJ may be either a character or a (smaller) string.  
 This function really does alter the contents of STRING.  
 Since it is impossible to change the length of an existing string,  
 it is an error if OBJ doesn't fit within STRING's actual length.  
   
 *** char-width returns the width (in columns) of the character CHAR,  
 if it were displayed in the current buffer and the selected window.  
   
 *** string-width returns the width (in columns) of the text in STRING,  
 if it were displayed in the current buffer and the selected window.  
   
 *** truncate-string-to-width shortens a string, if necessary,  
 to fit within a certain number of columns.  (Of course, it does  
 not alter the string that you give it; it returns a new string  
 which contains all or just part of the existing string.)  
   
 (truncate-string-to-width STR END-COLUMN &optional START-COLUMN PADDING)  
   
 This returns the part of STR up to column END-COLUMN.  
   
 The optional argument START-COLUMN specifies the starting column.  
 If this is non-nil, then the first START-COLUMN columns of the string  
 are not included in the resulting value.  
   
 The optional argument PADDING, if non-nil, is a padding character to be added  
 at the beginning and end the resulting string, to extend it to exactly  
 WIDTH columns.  If PADDING is nil, that means do not pad; then, if STRING  
 is narrower than WIDTH, the value is equal to STRING.  
   
 If PADDING and START-COLUMN are both non-nil, and if there is no clean  
 place in STRING that corresponds to START-COLUMN (because one  
 character extends across that column), then the padding character  
 PADDING is added one or more times at the beginning of the result  
 string, so that its columns line up as if it really did start at  
 column START-COLUMN.  
   
 *** When the functions in the list after-change-functions are called,  
 the third argument is the number of bytes in the pre-change text, not  
 necessarily the number of characters.  It is, in effect, the  
 difference in buffer position between the beginning and the end of the  
 changed text, before the change.  
   
 *** The characters Emacs uses are classified in various character  
 sets, each of which has a name which is a symbol.  In general there is  
 one character set for each script, not for each language.  
   
 **** The function charsetp tests whether an object is a character set name.  
   
 **** The variable charset-list holds a list of character set names.  
   
 **** char-charset, given a character code, returns the name of the character  
 set that the character belongs to.  (The value is a symbol.)  
   
 **** split-char, given a character code, returns a list containing the  
 name of the character set, followed by one or two byte-values  
 which identify the character within that character set.  
   
 **** make-char, given a character set name and one or two subsequent  
 byte-values, constructs a character code.  This is roughly the  
 opposite of split-char.  
   
 **** find-charset-region returns a list of the character sets  
 of all the characters between BEG and END.  
   
 **** find-charset-string returns a list of the character sets  
 of all the characters in a string.  
   
 *** Here are the Lisp facilities for working with coding systems  
 and specifying coding systems.  
   
 **** The function coding-system-list returns a list of all coding  
 system names (symbols).  With optional argument t, it returns a list  
 of all distinct base coding systems, not including variants.  
 (Variant coding systems are those like latin-1-dos, latin-1-unix  
 and latin-1-mac which specify the end-of-line conversion as well  
 as what to do about code conversion.)  
   
 **** coding-system-p tests a symbol to see if it is a coding system  
 name.  It returns t if so, nil if not.  
   
 **** file-coding-system-alist specifies which coding systems to use  
 for certain file names.  It works like network-coding-system-alist,  
 except that the PATTERN is matched against the file name.  
   
 Each element has the format (PATTERN . VAL), where PATTERN determines  
 which file names the element applies to.  PATTERN should be a regexp  
 to match against a file name.  
   
 VAL is a coding system, a cons cell containing two coding systems, or  
 a function symbol.  If VAL is a coding system, it is used for both  
 decoding what received from the network stream and encoding what sent  
 to the network stream.  If VAL is a cons cell containing two coding  
 systems, the car specifies the coding system for decoding, and the cdr  
 specifies the coding system for encoding.  
   
 If VAL is a function symbol, the function must return a coding system  
 or a cons cell containing two coding systems, which is used as above.  
   
 **** The variable network-coding-system-alist specifies  
 the coding system to use for network sockets.  
   
 Each element has the format (PATTERN . VAL), where PATTERN determines  
 which network sockets the element applies to.  PATTERN should be  
 either a port number or a regular expression matching some network  
 service names.  
   
 VAL is a coding system, a cons cell containing two coding systems, or  
 a function symbol.  If VAL is a coding system, it is used for both  
 decoding what received from the network stream and encoding what sent  
 to the network stream.  If VAL is a cons cell containing two coding  
 systems, the car specifies the coding system for decoding, and the cdr  
 specifies the coding system for encoding.  
   
 If VAL is a function symbol, the function must return a coding system  
 or a cons cell containing two coding systems, which is used as above.  
   
 **** process-coding-system-alist specifies which coding systems to use  
 for certain subprocess.  It works like network-coding-system-alist,  
 except that the PATTERN is matched against the program name used to  
 start the subprocess.  
   
 **** The variable default-process-coding-system specifies the coding  
 systems to use for subprocess (and net connection) input and output,  
 when nothing else specifies what to do.  The value is a cons cell  
 (OUTPUT-CODING . INPUT-CODING).  OUTPUT-CODING applies to output  
 to the subprocess, and INPUT-CODING applies to input from it.  
   
 **** The variable coding-system-for-write, if non-nil, specifies the  
 coding system to use for writing a file, or for output to a synchronous  
 subprocess.  
   
 It also applies to any asynchronous subprocess or network connection,  
 but in a different way: the value of coding-system-for-write when you  
 start the subprocess or connection affects that subprocess or  
 connection permanently or until overridden.  
   
 The variable coding-system-for-write takes precedence over  
 file-coding-system-alist, process-coding-system-alist and  
 network-coding-system-alist, and all other methods of specifying a  
 coding system for output.  But most of the time this variable is nil.  
 It exists so that Lisp programs can bind it to a specific coding  
 system for one operation at a time.  
   
 **** coding-system-for-read applies similarly to input from  
 files, subprocesses or network connections.  
   
 **** The function process-coding-system tells you what  
 coding systems(s) an existing subprocess is using.  
 The value is a cons cell,  
  (DECODING-CODING-SYSTEM . ENCODING-CODING-SYSTEM)  
 where DECODING-CODING-SYSTEM is used for decoding output from  
 the subprocess, and ENCODING-CODING-SYSTEM is used for encoding  
 input to the subprocess.  
   
 **** The function set-process-coding-system can be used to  
 change the coding systems in use for an existing subprocess.  
   
 ** Emacs has a new facility to help users manage the many  
 customization options.  To make a Lisp program work with this facility,  
 you need to use the new macros defgroup and defcustom.  
   
 You use defcustom instead of defvar, for defining a user option  
 variable.  The difference is that you specify two additional pieces of  
 information (usually): the "type" which says what values are  
 legitimate, and the "group" which specifies the hierarchy for  
 customization.  
   
 Thus, instead of writing  
   
     (defvar foo-blurgoze nil  
       "*Non-nil means that foo will act very blurgozely.")  
   
 you would now write this:  
   
     (defcustom foo-blurgoze nil  
       "*Non-nil means that foo will act very blurgozely."  
       :type 'boolean  
       :group foo)  
   
 The type `boolean' means that this variable has only  
 two meaningful states: nil and non-nil.  Other type values  
 describe other possibilities; see the manual for Custom  
 for a description of them.  
   
 The "group" argument is used to specify a group which the option  
 should belong to.  You define a new group like this:  
   
     (defgroup ispell nil  
       "Spell checking using Ispell."  
       :group 'processes)  
   
 The "group" argument in defgroup specifies the parent group.  The root  
 group is called `emacs'; it should not contain any variables itself,  
 but only other groups.  The immediate subgroups of `emacs' correspond  
 to the keywords used by C-h p.  Under these subgroups come  
 second-level subgroups that belong to individual packages.  
   
 Each Emacs package should have its own set of groups.  A simple  
 package should have just one group; a more complex package should  
 have a hierarchy of its own groups.  The sole or root group of a  
 package should be a subgroup of one or more of the "keyword"  
 first-level subgroups.  
   
 ** New `widget' library for inserting UI components in buffers.  
   
 This library, used by the new custom library, is documented in a  
 separate manual that accompanies Emacs.  
   
 ** easy-mmode  
   
 The easy-mmode package provides macros and functions that make  
 developing minor modes easier.  Roughly, the programmer has to code  
 only the functionality of the minor mode.  All the rest--toggles,  
 predicate, and documentation--can be done in one call to the macro  
 `easy-mmode-define-minor-mode' (see the documentation).  See also  
 `easy-mmode-define-keymap'.  
   
 ** Text property changes  
   
 *** The `intangible' property now works on overlays as well as on a  
 text property.  
   
 *** The new functions next-char-property-change and  
 previous-char-property-change scan through the buffer looking for a  
 place where either a text property or an overlay might change.  The  
 functions take two arguments, POSITION and LIMIT.  POSITION is the  
 starting position for the scan.  LIMIT says where to stop the scan.  
   
 If no property change is found before LIMIT, the value is LIMIT.  If  
 LIMIT is nil, scan goes to the beginning or end of the accessible part  
 of the buffer.  If no property change is found, the value is the  
 position of the beginning or end of the buffer.  
   
 *** In the `local-map' text property or overlay property, the property  
 value can now be a symbol whose function definition is a keymap.  This  
 is an alternative to using the keymap itself.  
   
 ** Changes in invisibility features  
   
 *** Isearch can now temporarily show parts of the buffer which are  
 hidden by an overlay with a invisible property, when the search match  
 is inside that portion of the buffer.  To enable this the overlay  
 should have a isearch-open-invisible property which is a function that  
 would be called having the overlay as an argument, the function should  
 make the overlay visible.  
   
 During incremental search the overlays are shown by modifying the  
 invisible and intangible properties, if beside this more actions are  
 needed the overlay should have a isearch-open-invisible-temporary  
 which is a function. The function is called with 2 arguments: one is  
 the overlay and the second is nil when it should show the overlay and  
 t when it should hide it.  
   
 *** add-to-invisibility-spec, remove-from-invisibility-spec  
   
 Modes that use overlays to hide portions of a buffer should set the  
 invisible property of the overlay to the mode's name (or another symbol)  
 and modify the `buffer-invisibility-spec' to include that symbol.  
 Use  `add-to-invisibility-spec' and `remove-from-invisibility-spec' to  
 manipulate the `buffer-invisibility-spec'.  
 Here is an example of how to do this:  
   
  ;; If we want to display an ellipsis:  
  (add-to-invisibility-spec '(my-symbol . t))  
  ;; If you don't want ellipsis:  
  (add-to-invisibility-spec 'my-symbol)  
   
   ...  
  (overlay-put  (make-overlay beginning end)  'invisible 'my-symbol)  
   
  ...  
  ;; When done with the overlays:  
  (remove-from-invisibility-spec '(my-symbol . t))  
  ;; Or respectively:  
  (remove-from-invisibility-spec 'my-symbol)  
   
 ** Changes in syntax parsing.  
   
 *** The syntax-directed buffer-scan functions (such as  
 `parse-partial-sexp', `forward-word' and similar functions) can now  
 obey syntax information specified by text properties, if the variable  
 `parse-sexp-lookup-properties' is non-nil.  
   
 If the value of `parse-sexp-lookup-properties' is nil, the behavior  
 is as before: the syntax-table of the current buffer is always  
 used to determine the syntax of the character at the position.  
   
 When `parse-sexp-lookup-properties' is non-nil, the syntax of a  
 character in the buffer is calculated thus:  
   
         a) if the `syntax-table' text-property of that character  
            is a cons, this cons becomes the syntax-type;  
   
            Valid values of `syntax-table' text-property are: nil, a valid  
            syntax-table, and a valid syntax-table element, i.e.,  
            a cons cell of the form (SYNTAX-CODE . MATCHING-CHAR).  
   
         b) if the character's `syntax-table' text-property  
            is a syntax table, this syntax table is used  
            (instead of the syntax-table of the current buffer) to  
            determine the syntax type of the character.  
   
         c) otherwise the syntax-type is determined by the syntax-table  
            of the current buffer.  
   
 *** The meaning of \s in regular expressions is also affected by the  
 value of `parse-sexp-lookup-properties'.  The details are the same as  
 for the syntax-directed buffer-scan functions.  
   
 *** There are two new syntax-codes, `!' and `|' (numeric values 14  
 and 15).  A character with a code `!' starts a comment which is ended  
 only by another character with the same code (unless quoted).  A  
 character with a code `|' starts a string which is ended only by  
 another character with the same code (unless quoted).  
   
 These codes are mainly meant for use as values of the `syntax-table'  
 text property.  
   
 *** The function `parse-partial-sexp' has new semantics for the sixth  
 arg COMMENTSTOP.  If it is `syntax-table', parse stops after the start  
 of a comment or a string, or after end of a comment or a string.  
   
 *** The state-list which the return value from `parse-partial-sexp'  
 (and can also be used as an argument) now has an optional ninth  
 element: the character address of the start of last comment or string;  
 nil if none.  The fourth and eighth elements have special values if the  
 string/comment is started by a "!"  or "|" syntax-code.  
   
 *** Since new features of `parse-partial-sexp' allow a complete  
 syntactic parsing, `font-lock' no longer supports  
 `font-lock-comment-start-regexp'.  
   
 ** Changes in face features  
   
 *** The face functions are now unconditionally defined in Emacs, even  
 if it does not support displaying on a device that supports faces.  
   
 *** The function face-documentation returns the documentation string  
 of a face (or nil if it doesn't have one).  
   
 *** The function face-bold-p returns t if a face should be bold.  
 set-face-bold-p sets that flag.  
   
 *** The function face-italic-p returns t if a face should be italic.  
 set-face-italic-p sets that flag.  
   
 *** You can now specify foreground and background colors for text  
 by adding elements of the form (foreground-color . COLOR-NAME)  
 and (background-color . COLOR-NAME) to the list of faces in  
 the `face' property (either the character's text property or an  
 overlay property).  
   
 This means that you no longer need to create named faces to use  
 arbitrary colors in a Lisp package.  
   
 ** Changes in file-handling functions  
   
 *** File-access primitive functions no longer discard an extra redundant  
 directory name from the beginning of the file name.  In other words,  
 they no longer do anything special with // or /~.  That conversion  
 is now done only in substitute-in-file-name.  
   
 This makes it possible for a Lisp program to open a file whose name  
 begins with ~.  
   
 *** If copy-file is unable to set the date of the output file,  
 it now signals an error with the condition file-date-error.  
   
 *** The inode number returned by file-attributes may be an integer (if  
 the number fits in a Lisp integer) or a list of integers.  
   
 *** insert-file-contents can now read from a special file,  
 as long as the arguments VISIT and REPLACE are nil.  
   
 *** The RAWFILE arg to find-file-noselect, if non-nil, now suppresses  
 character code conversion as well as other things.  
   
 Meanwhile, this feature does work with remote file names  
 (formerly it did not).  
   
 *** Lisp packages which create temporary files should use the TMPDIR  
 environment variable to decide which directory to put them in.  
   
 *** interpreter-mode-alist elements now specify regexps  
 instead of constant strings.  
   
 *** expand-file-name no longer treats `//' or `/~' specially.  It used  
 to delete all the text of a file name up through the first slash of  
 any `//' or `/~' sequence.  Now it passes them straight through.  
   
 substitute-in-file-name continues to treat those sequences specially,  
 in the same way as before.  
   
 *** The variable `format-alist' is more general now.  
 The FROM-FN and TO-FN in a format definition can now be strings  
 which specify shell commands to use as filters to perform conversion.  
   
 *** The new function access-file tries to open a file, and signals an  
 error if that fails.  If the open succeeds, access-file does nothing  
 else, and returns nil.  
   
 *** The function insert-directory now signals an error if the specified  
 directory cannot be listed.  
   
 ** Changes in minibuffer input  
   
 *** The functions read-buffer, read-variable, read-command, read-string  
 read-file-name, read-from-minibuffer and completing-read now take an  
 additional argument which specifies the default value.  If this  
 argument is non-nil, it should be a string; that string is used in two  
 ways:  
   
   It is returned if the user enters empty input.  
   It is available through the history command M-n.  
   
 *** The functions read-string, read-from-minibuffer,  
 read-no-blanks-input and completing-read now take an additional  
 argument INHERIT-INPUT-METHOD.  If this is non-nil, then the  
 minibuffer inherits the current input method and the setting of  
 enable-multibyte-characters from the previously current buffer.  
   
 In an interactive spec, you can use M instead of s to read an  
 argument in this way.  
   
 *** All minibuffer input functions discard text properties  
 from the text you enter in the minibuffer, unless the variable  
 minibuffer-allow-text-properties is non-nil.  
   
 ** Echo area features  
   
 *** Clearing the echo area now runs the normal hook  
 echo-area-clear-hook.  Note that the echo area can be used while the  
 minibuffer is active; in that case, the minibuffer is still active  
 after the echo area is cleared.  
   
 *** The function current-message returns the message currently displayed  
 in the echo area, or nil if there is none.  
   
 ** Keyboard input features  
   
 *** tty-erase-char is a new variable that reports which character was  
 set up as the terminal's erase character when time Emacs was started.  
   
 *** num-nonmacro-input-events is the total number of input events  
 received so far from the terminal.  It does not count those generated  
 by keyboard macros.  
   
 ** Frame-related changes  
   
 *** make-frame runs the normal hook before-make-frame-hook just before  
 creating a frame, and just after creating a frame it runs the abnormal  
 hook after-make-frame-functions with the new frame as arg.  
   
 *** The new hook window-configuration-change-hook is now run every time  
 the window configuration has changed.  The frame whose configuration  
 has changed is the selected frame when the hook is run.  
   
 *** Each frame now independently records the order for recently  
 selected buffers, in its buffer-list frame parameter, so that the  
 value of other-buffer is now based on the buffers recently displayed  
 in the selected frame.  
   
 *** The value of the frame parameter vertical-scroll-bars  
 is now `left', `right' or nil.  A non-nil value specifies  
 which side of the window to put the scroll bars on.  
   
 ** X Windows features  
   
 *** You can examine X resources for other applications by binding  
 x-resource-class around a call to x-get-resource.  The usual value of  
 x-resource-class is "Emacs", which is the correct value for Emacs.  
   
 *** In menus, checkboxes and radio buttons now actually work.  
 The menu displays the current status of the box or button.  
   
 *** The function x-list-fonts now takes an optional fourth argument  
 MAXIMUM which sets a limit on how many matching fonts to return.  
 A smaller value of MAXIMUM makes the function faster.  
   
 If the only question is whether *any* font matches the pattern,  
 it is good to supply 1 for this argument.  
   
 ** Subprocess features  
   
 *** A reminder: it is no longer necessary for subprocess filter  
 functions and sentinels to do save-match-data, because Emacs does this  
 automatically.  
   
 *** The new function shell-command-to-string executes a shell command  
 and returns the output from the command as a string.  
   
 *** The new function process-contact returns t for a child process,  
 and (HOSTNAME SERVICE) for a net connection.  
   
 ** An error in running pre-command-hook or post-command-hook  
 does clear the variable to nil.  The documentation was wrong before.  
   
 ** In define-key-after, if AFTER is t, the new binding now always goes  
 at the end of the keymap.  If the keymap is a menu, this means it  
 goes after the other menu items.  
   
 ** If you have a program that makes several changes in the same area  
 of the buffer, you can use the macro combine-after-change-calls  
 around that Lisp code to make it faster when after-change hooks  
 are in use.  
   
 The macro arranges to call the after-change functions just once for a  
 series of several changes--if that seems safe.  
   
 Don't alter the variables after-change-functions and  
 after-change-function within the body of a combine-after-change-calls  
 form.  
   
 ** If you define an abbrev (with define-abbrev) whose EXPANSION  
 is not a string, then the abbrev does not expand in the usual sense,  
 but its hook is still run.  
   
 ** Normally, the Lisp debugger is not used (even if you have enabled it)  
 for errors that are handled by condition-case.  
   
 If you set debug-on-signal to a non-nil value, then the debugger is called  
 regardless of whether there is a handler for the condition.  This is  
 useful for debugging problems that happen inside of a condition-case.  
   
 This mode of operation seems to be unreliable in other ways.  Errors that  
 are normal and ought to be handled, perhaps in timers or process  
 filters, will instead invoke the debugger.  So don't say you weren't  
 warned.  
   
 ** The new variable ring-bell-function lets you specify your own  
 way for Emacs to "ring the bell".  
   
 ** If run-at-time's TIME argument is t, the action is repeated at  
 integral multiples of REPEAT from the epoch; this is useful for  
 functions like display-time.  
   
 ** You can use the function locate-library to find the precise file  
 name of a Lisp library.  This isn't new, but wasn't documented before.  
   
 ** Commands for entering view mode have new optional arguments that  
 can be used from Lisp.  Low-level entrance to and exit from view mode  
 is done by functions view-mode-enter and view-mode-exit.  
   
 ** batch-byte-compile-file now makes Emacs return a nonzero status code  
 if there is an error in compilation.  
   
 ** pop-to-buffer, switch-to-buffer-other-window and  
 switch-to-buffer-other-frame now accept an additional optional  
 argument NORECORD, much like switch-to-buffer.  If it is non-nil,  
 they don't put the buffer at the front of the buffer list.  
   
 ** If your .emacs file leaves the *scratch* buffer non-empty,  
 Emacs does not display the startup message, so as to avoid changing  
 the *scratch* buffer.  
   
 ** The new function regexp-opt returns an efficient regexp to match a string.  
 The arguments are STRINGS and (optionally) PAREN.  This function can be used  
 where regexp matching or searching is intensively used and speed is important,  
 e.g., in Font Lock mode.  
   
 ** The variable buffer-display-count is local to each buffer,  
 and is incremented each time the buffer is displayed in a window.  
 It starts at 0 when the buffer is created.  
   
 ** The new function compose-mail starts composing a mail message  
 using the user's chosen mail composition agent (specified with the  
 variable mail-user-agent).  It has variants compose-mail-other-window  
 and compose-mail-other-frame.  
   
 ** The `user-full-name' function now takes an optional parameter which  
 can either be a number (the UID) or a string (the login name).  The  
 full name of the specified user will be returned.  
   
 ** Lisp packages that load files of customizations, or any other sort  
 of user profile, should obey the variable init-file-user in deciding  
 where to find it.  They should load the profile of the user name found  
 in that variable.  If init-file-user is nil, meaning that the -q  
 option was used, then Lisp packages should not load the customization  
 files at all.  
   
 ** format-time-string now allows you to specify the field width  
 and type of padding.  This works as in printf: you write the field  
 width as digits in the middle of a %-construct.  If you start  
 the field width with 0, it means to pad with zeros.  
   
 For example, %S normally specifies the number of seconds since the  
 minute; %03S means to pad this with zeros to 3 positions, %_3S to pad  
 with spaces to 3 positions.  Plain %3S pads with zeros, because that  
 is how %S normally pads to two positions.  
   
 ** thing-at-point now supports a new kind of "thing": url.  
   
 ** imenu.el changes.  
   
 You can now specify a function to be run when selecting an  
 item from menu created by imenu.  
   
 An example of using this feature: if we define imenu items for the  
 #include directives in a C file, we can open the included file when we  
 select one of those items.  
   
 * For older news, see the file ONEWS  
85    
86  ----------------------------------------------------------------------  ----------------------------------------------------------------------
87  Copyright information:  Copyright information:
88    
89  Copyright (C) 1999, 2000, 2001 Free Software Foundation, Inc.  Copyright (C) 2002 Free Software Foundation, Inc.
90    
91     Permission is granted to anyone to make or distribute verbatim copies     Permission is granted to anyone to make or distribute verbatim copies
92     of this document as received, in any medium, provided that the     of this document as received, in any medium, provided that the

Legend:
Removed from v.1.613  
changed lines
  Added in v.1.613.2.1

savannah-hackers-public@gnu.org
ViewVC Help
Powered by ViewVC 1.1.26