path: root/doc/emacs/custom.texi

@c -*- coding: utf-8 -*-
@c This is part of the Emacs manual.
@c Copyright (C) 1985-1987, 1993-1995, 1997, 2000-2019 Free Software
@c Foundation, Inc.
@c See file emacs.texi for copying conditions.
@node Customization
@chapter Customization
@cindex customization

  This chapter describes some simple methods to customize the behavior
of Emacs.

  Apart from the methods described here, see @ref{X Resources} for
information about using X resources to customize Emacs, and see
@ref{Keyboard Macros} for information about recording and replaying
keyboard macros.  Making more far-reaching and open-ended changes
involves writing Emacs Lisp code; see
@iftex
@cite{The Emacs Lisp Reference Manual}.
@end iftex
@ifnottex
@ref{Top, Emacs Lisp, Emacs Lisp, elisp, The Emacs Lisp
Reference Manual}.
@end ifnottex

@menu
* Easy Customization::  Convenient way to browse and change settings.
* Variables::           Many Emacs commands examine Emacs variables
                          to decide what to do; by setting variables,
                          you can control their functioning.
* Key Bindings::        The keymaps say what command each key runs.
                          By changing them, you can redefine keys.
* Init File::           How to write common customizations in the
                          initialization file.
* Authentication::      Keeping persistent authentication information.

@end menu

@node Easy Customization
@section Easy Customization Interface

@cindex settings
@cindex user option
@cindex customizable variable
  Emacs has many @dfn{settings} which you can change.  Most settings
are @dfn{customizable variables} (@pxref{Variables}), which are also
called @dfn{user options}.  There is a huge number of customizable
variables, controlling numerous aspects of Emacs behavior; the
variables documented in this manual are listed in @ref{Variable
Index}.  A separate class of settings are the @dfn{faces}, which
determine the fonts, colors, and other attributes of text
(@pxref{Faces}).

@findex customize
@cindex customization buffer
  To browse and alter settings (both variables and faces), type
@kbd{M-x customize}.  This creates a @dfn{customization buffer}, which
lets you navigate through a logically organized list of settings, edit
and set their values, and save them permanently.

@menu
* Customization Groups::     How settings are classified.
* Browsing Custom::          Browsing and searching for settings.
* Changing a Variable::      How to edit an option's value and set the option.
* Saving Customizations::    Saving customizations for future Emacs sessions.
* Face Customization::       How to edit the attributes of a face.
* Specific Customization::   Customizing specific settings or groups.
* Custom Themes::            Collections of customization settings.
* Creating Custom Themes::   How to create a new custom theme.
@end menu

@node Customization Groups
@subsection Customization Groups
@cindex customization groups

  Customization settings are organized into @dfn{customization
groups}.  These groups are collected into bigger groups, all the way
up to a master group called @code{Emacs}.

  @kbd{M-x customize} creates a customization buffer that shows the
top-level @code{Emacs} group.  It looks like this, in part:

@c we want the buffer example to all be on one page, but unfortunately
@c that's quite a bit of text, so force all space to the bottom.
@c @page
@smallexample
@group
For help using this buffer, see [Easy Customization] in the [Emacs manual].

________________________________________ [ Search ]

 Operate on all settings in this buffer:
 [ Revert... ] [ Apply ] [ Apply and Save ]


Emacs group: Customization of the One True Editor.
      [State]: visible group members are all at standard values.
      See also [Manual].

[Editing]      Basic text editing facilities.
[Convenience]  Convenience features for faster editing.

@var{...more second-level groups...}
@end group
@end smallexample

@noindent
The main part of this buffer shows the @samp{Emacs} customization
group, which contains several other groups (@samp{Editing},
@samp{Convenience}, etc.).  The contents of those groups are not
listed here, only one line of documentation each.

  The @dfn{state} of the group indicates whether setting in that group
has been edited, set, or saved.  @xref{Changing a Variable}.

@cindex editable fields (customization buffer)
@cindex buttons (customization buffer)
@cindex links (customization buffer)
  Most of the customization buffer is read-only, but it includes some
@dfn{editable fields} that you can edit.  For example, at the top of
the customization buffer is an editable field for searching for
settings (@pxref{Browsing Custom}).  There are also @dfn{buttons} and
@dfn{links}, which you can activate by either clicking with the mouse,
or moving point there and typing @kbd{@key{RET}}.  For example, the group
names like @samp{[Editing]} are links; activating one of these links
brings up the customization buffer for that group.

@kindex TAB @r{(customization buffer)}
@kindex S-TAB @r{(customization buffer)}
@findex widget-forward
@findex widget-backward
  In the customization buffer, you can type @kbd{@key{TAB}}
(@code{widget-forward}) to move forward to the next button or editable
field.  @kbd{S-@key{TAB}} (@code{widget-backward}) moves back to the
previous button or editable field.

@node Browsing Custom
@subsection Browsing and Searching for Settings
@findex customize-browse

  From the top-level customization buffer created by @kbd{M-x
customize}, you can follow the links to the subgroups of the
@samp{Emacs} customization group.  These subgroups may contain
settings for you to customize; they may also contain further subgroups,
dealing with yet more specialized subsystems of Emacs.  As you
navigate the hierarchy of customization groups, you should find some
settings that you want to customize.

  If you are interested in customizing a particular setting or
customization group, you can go straight there with the commands
@kbd{M-x customize-option}, @kbd{M-x customize-face}, or @w{@kbd{M-x
customize-group}}.  @xref{Specific Customization}.

@vindex custom-search-field
  If you don't know exactly what groups or settings you want to
customize, you can search for them using the editable search field at
the top of each customization buffer.  Here, you can type in a search
term---either one or more words separated by spaces, or a regular
expression (@pxref{Regexps}).  Then type @kbd{@key{RET}} in the field,
or activate the @samp{Search} button next to it, to switch to a
customization buffer containing groups and settings that match those
terms.  Note, however, that this feature only finds groups and
settings that are loaded in the current Emacs session.

  If you don't want customization buffers to show the search field,
change the variable @code{custom-search-field} to @code{nil}.

  The command @kbd{M-x customize-apropos} is similar to using the
search field, except that it reads the search term(s) using the
minibuffer.  @xref{Specific Customization}.

  @kbd{M-x customize-browse} is another way to browse the available
settings.  This command creates a special customization buffer, which
shows only the names of groups and settings, in a structured layout.
You can show the contents of a group, in the same buffer, by invoking
the @samp{[+]} button next to the group name.  When the group contents
are shown, the button changes to @samp{[-]}; invoking that hides the
group contents again.  Each group or setting in this buffer has a link
which says @samp{[Group]}, @samp{[Option]} or @samp{[Face]}.  Invoking
this link creates an ordinary customization buffer showing just that
group, option, or face; this is the way to change settings that you
find with @kbd{M-x customize-browse}.

@node Changing a Variable
@subsection Changing a Variable

  Here is an example of what a variable, or user option, looks like in
the customization buffer:

@smallexample
[Hide] Kill Ring Max: 60
   [State]: STANDARD.
   Maximum length of kill ring before oldest elements are thrown away.
@end smallexample

  The first line shows that the variable is named
@code{kill-ring-max}, formatted as @samp{Kill Ring Max} for easier
viewing.  Its value is @samp{60}.  The button labeled @samp{[Hide]},
if activated, hides the variable's value and state; this is useful to
avoid cluttering up the customization buffer with very long values
(for this reason, variables that have very long values may start out
hidden).  If you use the @samp{[Hide]} button, it changes to
@samp{[Show Value]}, which you can activate to reveal the value and
state.  On a graphical display, the @samp{[Hide]} and @samp{[Show
Value]} buttons are replaced with graphical triangles pointing
downwards and rightwards respectively.

  The line after the variable name indicates the @dfn{customization
state} of the variable: in this example, @samp{STANDARD} means you
have not changed the variable, so its value is the default one.  The
@samp{[State]} button gives a menu of operations for customizing the
variable.

  Below the customization state is the documentation for the variable.
This is the same documentation that would be shown by the @kbd{C-h v}
command (@pxref{Examining}).  If the documentation is more than one
line long, only one line may be shown.  If so, that line ends with a
@samp{[More]} button; activate this to see the full documentation.

@cindex user options, changing
@cindex customizing variables
@cindex variables, changing
  To enter a new value for @samp{Kill Ring Max}, just move point to
the value and edit it.  For example, type @kbd{M-d} to delete the
@samp{60} and type in another number.  As you begin to alter the text,
the @samp{[State]} line will change:

@smallexample
[State]: EDITED, shown value does not take effect until you
         set or save it.
@end smallexample

@noindent
Editing the value does not make it take effect right away.  To do
that, you must @dfn{set} the variable by activating the @samp{[State]}
button and choosing @samp{Set for Current Session}.  Then the
variable's state becomes:

@smallexample
[State]: SET for current session only.
@end smallexample

@noindent
You don't have to worry about specifying a value that is not valid;
the @samp{Set for Current Session} operation checks for validity and
will not install an unacceptable value.

@kindex M-TAB @r{(customization buffer)}
@kindex C-M-i @r{(customization buffer)}
@findex widget-complete
  While editing certain kinds of values, such as file names, directory
names, and Emacs command names, you can perform completion with
@kbd{C-M-i} (@code{widget-complete}), or the equivalent keys
@kbd{M-@key{TAB}} or @kbd{@key{ESC} @key{TAB}}.  This behaves much
like minibuffer completion (@pxref{Completion}).

  Typing @kbd{@key{RET}} on an editable value field moves point
forward to the next field or button, like @kbd{@key{TAB}}.  You can
thus type @kbd{@key{RET}} when you are finished editing a field, to
move on to the next button or field.  To insert a newline within an
editable field, use @kbd{C-o} or @kbd{C-q C-j}.

  For some variables, there is only a fixed set of legitimate values,
and you are not allowed to edit the value directly.  Instead, a
@samp{[Value Menu]} button appears before the value; activating this
button presents a choice of values.  For a boolean ``on or off''
value, the button says @samp{[Toggle]}, and flips the value.  After
using the @samp{[Value Menu]} or @samp{[Toggle]} button, you must
again set the variable to make the chosen value take effect.

  Some variables have values with complex structure.  For example, the
value of @code{minibuffer-frame-alist} is an association list.  Here
is how it appears in the customization buffer:

@smallexample
[Hide] Minibuffer Frame Alist:
[INS] [DEL] Parameter: width
            Value: 80
[INS] [DEL] Parameter: height
            Value: 2
[INS]
   [ State ]: STANDARD.
   Alist of parameters for the initial minibuffer frame. [Hide]
   @r{[@dots{}more lines of documentation@dots{}]}
@end smallexample

@noindent
In this case, each association in the list consists of two items, one
labeled @samp{Parameter} and one labeled @samp{Value}; both are
editable fields.  You can delete an association from the list with the
@samp{[DEL]} button next to it.  To add an association, use the
@samp{[INS]} button at the position where you want to insert it; the
very last @samp{[INS]} button inserts at the end of the list.

@cindex saving a setting
@cindex settings, how to save
  When you set a variable, the new value takes effect only in the
current Emacs session.  To @dfn{save} the value for future sessions,
use the @samp{[State]} button and select the @samp{Save for Future
Sessions} operation.  @xref{Saving Customizations}.

  You can also restore the variable to its standard value by using the
@samp{[State]} button and selecting the @samp{Erase Customization}
operation.  There are actually four reset operations:

@table @samp
@item Undo Edits
If you have modified but not yet set the variable, this restores the
text in the customization buffer to match the actual value.

@item Revert This Session's Customizations
This restores the value of the variable to the last saved value, if
there was one.  Otherwise it restores the standard value.  It updates
the text accordingly.

@item Erase Customization
This sets the variable to its standard value.  Any saved value that
you have is also eliminated.

@item Set to Backup Value
This sets the variable to a previous value that was set in the
customization buffer in this session.  If you customize a variable
and then reset it, which discards the customized value,
you can get the discarded value back again with this operation.
@end table

@cindex comments on customized settings
  Sometimes it is useful to record a comment about a specific
customization.  Use the @samp{Add Comment} item from the
@samp{[State]} menu to create a field for entering the comment.

  Near the top of the customization buffer are two lines of buttons:

@smallexample
 Operate on all settings in this buffer:
 [Revert...] [Apply] [Apply and Save]
@end smallexample

@noindent
The @samp{[Revert...]} button drops a menu with the first 3 reset
operations described above.  The @samp{[Apply]} button applies the
settings for the current session.  The @samp{[Apply and Save]} button
applies the settings and saves them for future sessions; this button
does not appear if Emacs was started with the @option{-q} or
@option{-Q} option (@pxref{Initial Options}).

@kindex C-c C-c @r{(customization buffer)}
@kindex C-x C-c @r{(customization buffer)}
@findex Custom-set
@findex Custom-save
  The command @kbd{C-c C-c} (@code{Custom-set}) is equivalent to using
the @samp{[Set for Current Session]} button.  The command @kbd{C-x
C-s} (@code{Custom-save}) is like using the @samp{[Save for Future
Sessions]} button.

@vindex custom-buffer-done-kill
  The @samp{[Exit]} button switches out of the customization buffer,
and buries the buffer at the bottom of the buffer list.  To make it
kill the customization buffer instead, change the variable
@code{custom-buffer-done-kill} to @code{t}.

@node Saving Customizations
@subsection Saving Customizations

  In the customization buffer, you can @dfn{save} a customization
setting by choosing the @samp{Save for Future Sessions} choice from
its @samp{[State]} button.  The @kbd{C-x C-s} (@code{Custom-save})
command, or the @samp{[Apply and Save]} button at the top of the
customization buffer, saves all applicable settings in the buffer.

  Saving works by writing code to a file, usually your initialization
file (@pxref{Init File}).  Future Emacs sessions automatically read
this file at startup, which sets up the customizations again.

@vindex custom-file
  You can choose to save customizations somewhere other than your
initialization file.  To make this work, you must add a couple of
lines of code to your initialization file, to set the variable
@code{custom-file} to the name of the desired file, and to load that
file.  For example:

@example
(setq custom-file "~/.emacs-custom.el")
(load custom-file)
@end example

  You can even specify different customization files for different
Emacs versions, like this:

@example
(cond ((< emacs-major-version 22)
       ;; @r{Emacs 21 customization.}
       (setq custom-file "~/.custom-21.el"))
      ((and (= emacs-major-version 22)
            (< emacs-minor-version 3))
       ;; @r{Emacs 22 customization, before version 22.3.}
       (setq custom-file "~/.custom-22.el"))
      (t
       ;; @r{Emacs version 22.3 or later.}
       (setq custom-file "~/.emacs-custom.el")))

(load custom-file)
@end example

  If Emacs was invoked with the @option{-q} or @option{--no-init-file}
options (@pxref{Initial Options}), it will not let you save your
customizations in your initialization file.  This is because saving
customizations from such a session would wipe out all the other
customizations you might have on your initialization file.

@cindex unsaved customizations, reminder to save
@findex custom-prompt-customize-unsaved-options
  Please note that any customizations you have not chosen to save for
future sessions will be lost when you terminate Emacs.  If you'd like
to be prompted about unsaved customizations at termination time, add
the following to your initialization file:

@example
(add-hook 'kill-emacs-query-functions
          'custom-prompt-customize-unsaved-options)
@end example

@node Face Customization
@subsection Customizing Faces
@cindex customizing faces
@cindex faces, customizing
@cindex fonts and faces

  You can customize faces (@pxref{Faces}), which determine how Emacs
displays different types of text.  Customization groups can contain
both variables and faces.

  For example, in programming language modes, source code comments are
shown with @code{font-lock-comment-face} (@pxref{Font Lock}).  In a
customization buffer, that face appears like this, after you click on
the @samp{[Show All Attributes]} link:

@smallexample
[Hide] Font Lock Comment Face:[sample]
   [State] : STANDARD.
   Font Lock mode face used to highlight comments.
   [ ] Font Family: --
   [ ] Font Foundry: --
   [ ] Width: --
   [ ] Height: --
   [ ] Weight: --
   [ ] Slant: --
   [ ] Underline: --
   [ ] Overline: --
   [ ] Strike-through: --
   [ ] Box around text: --
   [ ] Inverse-video: --
   [X] Foreground: Firebrick     [Choose]  (sample)
   [ ] Background: --
   [ ] Stipple: --
   [ ] Inherit: --
   [Hide Unused Attributes]
@end smallexample

@noindent
The first three lines show the name, @samp{[State]} button, and
documentation for the face.  Below that is a list of @dfn{face
attributes}.  In front of each attribute is a checkbox.  A filled
checkbox, @samp{[X]}, means that the face specifies a value for this
attribute; an empty checkbox, @samp{[ ]}, means that the face does not
specify any special value for the attribute.  You can activate a
checkbox to specify or unspecify its attribute.

  A face does not have to specify every single attribute; in fact,
most faces only specify a few attributes.  In the above example,
@code{font-lock-comment-face} only specifies the foreground color.
Any unspecified attribute is taken from the special face named
@code{default}, whose attributes are all specified.  The
@code{default} face is the face used to display any text that does not
have an explicitly-assigned face; furthermore, its background color
attribute serves as the background color of the frame.

  The @samp{[Hide Unused Attributes]} button, at the end of the
attribute list, hides the unspecified attributes of the face.  When
attributes are being hidden, the button changes to @samp{[Show All
Attributes]}, which reveals the entire attribute list.  The
customization buffer may start out with unspecified attributes hidden,
to avoid cluttering the interface.

  When an attribute is specified, you can change its value in the
usual ways.

  Foreground and background colors can be specified using either color
names or RGB triplets (@pxref{Colors}).  You can also use the
@samp{[Choose]} button to switch to a list of color names; select a
color with @kbd{@key{RET}} in that buffer to put the color name in the
value field.

  Setting, saving and resetting a face work like the same operations for
variables (@pxref{Changing a Variable}).

  A face can specify different appearances for different types of
displays.  For example, a face can make text red on a color display,
but use a bold font on a monochrome display.  To specify multiple
appearances for a face, select @samp{For All Kinds of Displays} in the
menu you get from invoking @samp{[State]}.

@node Specific Customization
@subsection Customizing Specific Items

@table @kbd
@item M-x customize-option @key{RET} @var{option} @key{RET}
@itemx M-x customize-variable @key{RET} @var{option} @key{RET}
Set up a customization buffer for just one user option, @var{option}.

@item M-x customize-face @key{RET} @var{face} @key{RET}
Set up a customization buffer for just one face, @var{face}.

@item M-x customize-group @key{RET} @var{group} @key{RET}
Set up a customization buffer for just one group, @var{group}.

@item M-x customize-apropos @key{RET} @var{regexp} @key{RET}
Set up a customization buffer for all the settings and groups that
match @var{regexp}.

@item M-x customize-changed @key{RET} @var{version} @key{RET}
Set up a customization buffer with all the settings and groups
whose meaning has changed since Emacs version @var{version}.

@item M-x customize-changed-options @key{RET} @var{version} @key{RET}
Set up a customization buffer with all the options whose meaning or
default values have changed since Emacs version @var{version}.

@item M-x customize-saved
Set up a customization buffer containing all settings that you
have saved with customization buffers.

@item M-x customize-unsaved
Set up a customization buffer containing all settings that you have
set but not saved.
@end table

@findex customize-option
  If you want to customize a particular user option, type @kbd{M-x
customize-option}.  This reads the variable name, and sets up the
customization buffer with just that one user option.  When entering
the variable name into the minibuffer, completion is available, but
only for the names of variables that have been loaded into Emacs.

@findex customize-face
@findex customize-group
  Likewise, you can customize a specific face using @kbd{M-x
customize-face}.  You can set up a customization buffer for a specific
customization group using @kbd{M-x customize-group}.

@findex customize-apropos
  @kbd{M-x customize-apropos} prompts for a search term---either one
or more words separated by spaces, or a regular expression---and sets
up a customization buffer for all @emph{loaded} settings and groups
with matching names.  This is like using the search field at the top
of the customization buffer (@pxref{Customization Groups}).

@findex customize-changed
  When you upgrade to a new Emacs version, you might want to consider
customizing new settings, and settings whose meanings or default
values have changed.  To do this, use @kbd{M-x customize-changed} and
specify a previous Emacs version number using the minibuffer.  It
creates a customization buffer which shows all the settings and groups
whose definitions have been changed since the specified version,
loading them if necessary.

@findex customize-saved
@findex customize-unsaved
  If you change settings and then decide the change was a mistake, you
can use two commands to revisit your changes.  Use @kbd{M-x
customize-saved} to customize settings that you have saved.  Use
@kbd{M-x customize-unsaved} to customize settings that you have set
but not saved.

@node Custom Themes
@subsection Custom Themes
@cindex custom themes

  @dfn{Custom themes} are collections of settings that can be enabled
or disabled as a unit.  You can use Custom themes to switch easily
between various collections of settings, and to transfer such
collections from one computer to another.

  A Custom theme is stored as an Emacs Lisp source file.  If the name of
the Custom theme is @var{name}, the theme file is named
@file{@var{name}-theme.el}.  @xref{Creating Custom Themes}, for the
format of a theme file and how to make one.

@findex customize-themes
@vindex custom-theme-directory
@cindex color scheme
  Type @kbd{M-x customize-themes} to switch to a buffer named
@file{*Custom Themes*}, which lists the Custom themes that Emacs knows
about.  By default, Emacs looks for theme files in two locations: the
directory specified by the variable @code{custom-theme-directory}
(which defaults to @file{~/.emacs.d/}), and a directory named
@file{etc/themes} in your Emacs installation (see the variable
@code{data-directory}).  The latter contains several Custom themes
distributed with Emacs that customize Emacs's faces to fit
various color schemes.  (Note, however, that Custom themes need not be
restricted to this purpose; they can be used to customize variables
too.)

@vindex custom-theme-load-path
  If you want Emacs to look for Custom themes in some other directory,
add the directory to the list variable
@code{custom-theme-load-path}.  Its default value is
@code{(custom-theme-directory t)}; here, the symbol
@code{custom-theme-directory} has the special meaning of the value of
the variable @code{custom-theme-directory}, while @code{t} stands for
the built-in theme directory @file{etc/themes}.  The themes listed in
the @file{*Custom Themes*} buffer are those found in the directories
specified by @code{custom-theme-load-path}.

@kindex C-x C-s @r{(Custom Themes buffer)}
  In the @file{*Custom Themes*} buffer, you can activate the checkbox
next to a Custom theme to enable or disable the theme for the current
Emacs session.  When a Custom theme is enabled, all of its settings
(variables and faces) take effect in the Emacs session.  To apply the
choice of theme(s) to future Emacs sessions, type @kbd{C-x C-s}
(@code{custom-theme-save}) or use the @samp{[Save Theme Settings]}
button.

@vindex custom-safe-themes
  When you first enable a Custom theme, Emacs displays the contents of
the theme file and asks if you really want to load it.  Because
loading a Custom theme can execute arbitrary Lisp code, you should
only say yes if you know that the theme is safe; in that case, Emacs
offers to remember in the future that the theme is safe (this is done
by saving the theme file's SHA-256 hash to the variable
@code{custom-safe-themes}; if you want to treat all themes as safe,
change its value to @code{t}).  Themes that come with Emacs (in the
@file{etc/themes} directory) are exempt from this check, and are
always considered safe.

@vindex custom-enabled-themes
  Setting or saving Custom themes actually works by customizing the
variable @code{custom-enabled-themes}.  The value of this variable is
a list of Custom theme names (as Lisp symbols, e.g., @code{tango}).
Instead of using the @file{*Custom Themes*} buffer to set
@code{custom-enabled-themes}, you can customize the variable using the
usual customization interface, e.g., with @kbd{M-x customize-option}.
Note that Custom themes are not allowed to set
@code{custom-enabled-themes} themselves.

  Any customizations that you make through the customization buffer
take precedence over theme settings.  This lets you easily override
individual theme settings that you disagree with.  If settings from
two different themes overlap, the theme occurring earlier in
@code{custom-enabled-themes} takes precedence.  In the customization
buffer, if a setting has been changed from its default by a Custom
theme, its @samp{State} display shows @samp{THEMED} instead of
@samp{STANDARD}.

@findex load-theme
@findex enable-theme
@findex disable-theme
  You can enable a specific Custom theme in the current Emacs session
by typing @kbd{M-x load-theme}.  This prompts for a theme name, loads
the theme from the theme file, and enables it.  If a theme file
has been loaded before, you can enable the theme without loading its
file by typing @kbd{M-x enable-theme}.  To disable a Custom theme,
type @kbd{M-x disable-theme}.

@findex describe-theme
  To see a description of a Custom theme, type @kbd{?} on its line in
the @file{*Custom Themes*} buffer; or type @kbd{M-x describe-theme}
anywhere in Emacs and enter the theme name.

@node Creating Custom Themes
@subsection Creating Custom Themes
@cindex custom themes, creating

@findex customize-create-theme
  You can define a Custom theme using an interface similar to the
customization buffer, by typing @kbd{M-x customize-create-theme}.
This switches to a buffer named @file{*Custom Theme*}.  It also offers
to insert some common Emacs faces into the theme (a convenience, since
Custom themes are often used to customize faces).  If you answer no,
the theme will initially contain no settings.

  Near the top of the @file{*Custom Theme*} buffer, there are editable fields
where you can enter the theme's name and description.  The name can be
anything except @samp{user}.  The description is the one that will be
shown when you invoke @kbd{M-x describe-theme} for the theme.  Its
first line should be a brief one-sentence summary; in the buffer made
by @kbd{M-x customize-themes}, this sentence is displayed next to the
theme name.

  To add a new setting to the theme, use the @samp{[Insert Additional
Face]} or @samp{[Insert Additional Variable]} buttons.  Each button
reads a face or variable name using the minibuffer, with completion,
and inserts a customization entry for the face or variable.  You can
edit the variable values or face attributes in the same way as in a
normal customization buffer.  To remove a face or variable from the
theme, uncheck the checkbox next to its name.

@vindex custom-theme-directory@r{, saving theme files}
  After specifying the Custom theme's faces and variables, type
@kbd{C-x C-s} (@code{custom-theme-write}) or use the buffer's
@samp{[Save Theme]} button.  This saves the theme file, named
@file{@var{name}-theme.el} where @var{name} is the theme name, in the
directory named by @code{custom-theme-directory}.

  From the @file{*Custom Theme*} buffer, you can view and edit an
existing Custom theme by activating the @samp{[Visit Theme]} button
and specifying the theme name.  You can also add the settings of
another theme into the buffer, using the @samp{[Merge Theme]} button.
You can import your non-theme settings into a Custom theme by using
the @samp{[Merge Theme]} button and specifying the special theme named
@samp{user}.

  A theme file is simply an Emacs Lisp source file, and loading the
Custom theme works by loading the Lisp file.  Therefore, you can edit
a theme file directly instead of using the @file{*Custom Theme*}
buffer.  @xref{Custom Themes,,, elisp, The Emacs Lisp Reference
Manual}, for details.

@node Variables
@section Variables
@cindex variable

  A @dfn{variable} is a Lisp symbol which has a value.  The symbol's
name is also called the @dfn{variable name}.  A variable name can
contain any characters that can appear in a file, but most variable
names consist of ordinary words separated by hyphens.

  The name of the variable serves as a compact description of its
role.  Most variables also have a @dfn{documentation string}, which
describes what the variable's purpose is, what kind of value it should
have, and how the value will be used.  You can view this documentation
using the help command @kbd{C-h v} (@code{describe-variable}).
@xref{Examining}.

  Emacs uses many Lisp variables for internal record keeping, but the
most interesting variables for a non-programmer user are those meant
for users to change---these are called @dfn{customizable variables} or
@dfn{user options} (@pxref{Easy Customization}).  In the following
sections, we will describe other aspects of Emacs variables, such as
how to set them outside Customize.

  Emacs Lisp allows any variable (with a few exceptions) to have any
kind of value.  However, many variables are meaningful only if
assigned values of a certain type.  For example, only numbers are
meaningful values for @code{kill-ring-max}, which specifies the
maximum length of the kill ring (@pxref{Earlier Kills}); if you give
@code{kill-ring-max} a string value, commands such as @kbd{C-y}
(@code{yank}) will signal an error.  On the other hand, some variables
don't care about type; for instance, if a variable has one effect for
@code{nil} values and another effect for non-@code{nil} values,
then any value that is not the symbol @code{nil} induces the second
effect, regardless of its type (by convention, we usually use the
value @code{t}---a symbol which stands for ``true''---to specify a
non-@code{nil} value).  If you set a variable using the customization
buffer, you need not worry about giving it an invalid type: the
customization buffer usually only allows you to enter meaningful
values.  When in doubt, use @kbd{C-h v} (@code{describe-variable}) to
check the variable's documentation string to see kind of value it
expects (@pxref{Examining}).

@menu
* Examining::           Examining or setting one variable's value.
* Hooks::               Hook variables let you specify programs for parts
                          of Emacs to run on particular occasions.
* Locals::              Per-buffer values of variables.
* File Variables::      How files can specify variable values.
* Directory Variables:: How variable values can be specified by directory.
* Connection Variables:: Variables which are valid for buffers with a
                           remote default directory.
@end menu

@node Examining
@subsection Examining and Setting Variables
@cindex setting variables

@table @kbd
@item C-h v @var{var} @key{RET}
Display the value and documentation of variable @var{var}
(@code{describe-variable}).

@item M-x set-variable @key{RET} @var{var} @key{RET} @var{value} @key{RET}
Change the value of variable @var{var} to @var{value}.
@end table

  To examine the value of a variable, use @kbd{C-h v}
(@code{describe-variable}).  This reads a variable name using the
minibuffer, with completion, and displays both the value and the
documentation of the variable.  For example,

@example
C-h v fill-column @key{RET}
@end example

@noindent
displays something like this:

@example
fill-column is a variable defined in ‘C source code’.
Its value is 70

  Automatically becomes buffer-local when set.
  This variable is safe as a file local variable if its value
  satisfies the predicate ‘integerp’.

Documentation:
Column beyond which automatic line-wrapping should happen.
Interactively, you can set the buffer local value using C-x f.

You can customize this variable.
@end example

@noindent
The line that says @samp{You can customize the variable} indicates that
this variable is a user option.  @kbd{C-h v} is not restricted to user
options; it allows non-customizable variables too.

@findex set-variable
  The most convenient way to set a specific customizable variable is
with @kbd{M-x set-variable}.  This reads the variable name with the
minibuffer (with completion), and then reads a Lisp expression for the
new value using the minibuffer a second time (you can insert the old
value into the minibuffer for editing via @kbd{M-n}).  For example,

@example
M-x set-variable @key{RET} fill-column @key{RET} 75 @key{RET}
@end example

@noindent
sets @code{fill-column} to 75.

 @kbd{M-x set-variable} is limited to customizable variables, but you
can set any variable with a Lisp expression like this:

@example
(setq fill-column 75)
@end example

@noindent
To execute such an expression, type @kbd{M-:} (@code{eval-expression})
and enter the expression in the minibuffer (@pxref{Lisp Eval}).
Alternatively, go to the @file{*scratch*} buffer, type in the
expression, and then type @kbd{C-j} (@pxref{Lisp Interaction}).

  Setting variables, like all means of customizing Emacs except where
otherwise stated, affects only the current Emacs session.  The only
way to alter the variable in future sessions is to put something in
your initialization file (@pxref{Init File}).

@node Hooks
@subsection Hooks
@cindex hook
@cindex running a hook

  @dfn{Hooks} are an important mechanism for customizing Emacs.  A
hook is a Lisp variable which holds a list of functions, to be called
on some well-defined occasion.  (This is called @dfn{running the
hook}.)  The individual functions in the list are called the @dfn{hook
functions} of the hook.  For example, the hook @code{kill-emacs-hook}
runs just before exiting Emacs (@pxref{Exiting}).

@cindex normal hook
  Most hooks are @dfn{normal hooks}.  This means that when Emacs runs
the hook, it calls each hook function in turn, with no arguments.  We
have made an effort to keep most hooks normal, so that you can use
them in a uniform way.  Every variable whose name ends in @samp{-hook}
is a normal hook.

@cindex abnormal hook
  A few hooks are @dfn{abnormal hooks}.  Their names end in
@samp{-functions}, instead of @samp{-hook} (some old code may also use
the deprecated suffix @samp{-hooks}).  What
makes these hooks abnormal is the way its functions are
called---perhaps they are given arguments, or perhaps the values they
return are used in some way.  For example,
@code{find-file-not-found-functions} is abnormal because as soon as
one hook function returns a non-@code{nil} value, the rest are not
called at all (@pxref{Visiting}).  The documentation of each abnormal
hook variable explains how its functions are used.

@findex add-hook
  You can set a hook variable with @code{setq} like any other Lisp
variable, but the recommended way to add a function to a hook (either
normal or abnormal) is to use @code{add-hook}, as shown by the
following examples.  @xref{Hooks,,, elisp, The Emacs Lisp Reference
Manual}, for details.

  Most major modes run one or more @dfn{mode hooks} as the last step
of initialization.  Mode hooks are a convenient way to customize the
behavior of individual modes; they are always normal.  For example,
here's how to set up a hook to turn on Auto Fill mode in Text mode and
other modes based on Text mode:

@example
(add-hook 'text-mode-hook 'auto-fill-mode)
@end example

@noindent
This works by calling @code{auto-fill-mode}, which enables the minor
mode when no argument is supplied (@pxref{Minor Modes}).  Next,
suppose you don't want Auto Fill mode turned on in @LaTeX{} mode,
which is one of the modes based on Text mode.  You can do this with
the following additional line:

@example
(add-hook 'latex-mode-hook (lambda () (auto-fill-mode -1)))
@end example

@noindent
Here we have used the special macro @code{lambda} to construct an
anonymous function (@pxref{Lambda Expressions,,, elisp, The Emacs Lisp
Reference Manual}), which calls @code{auto-fill-mode} with an argument
of @code{-1} to disable the minor mode.  Because @LaTeX{} mode runs
@code{latex-mode-hook} after running @code{text-mode-hook}, the result
leaves Auto Fill mode disabled.

  Here is a more complex example, showing how to use a hook to
customize the indentation of C code:

@example
@group
(setq my-c-style
  '((c-comment-only-line-offset . 4)
@end group
@group
    (c-cleanup-list . (scope-operator
                       empty-defun-braces
                       defun-close-semi))))
@end group

@group
(add-hook 'c-mode-common-hook
  (lambda () (c-add-style "my-style" my-c-style t)))
@end group
@end example

@cindex Prog mode
@cindex modes for editing programs
  Major mode hooks also apply to other major modes @dfn{derived} from
the original mode (@pxref{Derived Modes,,, elisp, The Emacs Lisp
Reference Manual}).  For instance, HTML mode is derived from Text mode
(@pxref{HTML Mode}); when HTML mode is enabled, it runs
@code{text-mode-hook} before running @code{html-mode-hook}.  This
provides a convenient way to use a single hook to affect several
related modes.  In particular, if you want to apply a hook function to
any programming language mode, add it to @code{prog-mode-hook}; Prog
mode is a major mode that does little else than to let other major
modes inherit from it, exactly for this purpose.

  It is best to design your hook functions so that the order in which
they are executed does not matter.  Any dependence on the order is
asking for trouble.  However, the order is predictable: the hook
functions are executed in the order they appear in the hook.

@findex remove-hook
  If you play with adding various different versions of a hook
function by calling @code{add-hook} over and over, remember that all
the versions you added will remain in the hook variable together.  You
can clear out individual functions by calling @code{remove-hook}, or
do @code{(setq @var{hook-variable} nil)} to remove everything.

@cindex buffer-local hooks
  If the hook variable is buffer-local, the buffer-local variable will
be used instead of the global variable.  However, if the buffer-local
variable contains the element @code{t}, the global hook variable will
be run as well.

@node Locals
@subsection Local Variables

@table @kbd
@item M-x make-local-variable @key{RET} @var{var} @key{RET}
Make variable @var{var} have a local value in the current buffer.

@item M-x kill-local-variable @key{RET} @var{var} @key{RET}
Make variable @var{var} use its global value in the current buffer.

@item M-x make-variable-buffer-local @key{RET} @var{var} @key{RET}
Mark variable @var{var} so that setting it will make it local to the
buffer that is current at that time.
@end table

@cindex local variables
  Almost any variable can be made @dfn{local} to a specific Emacs
buffer.  This means that its value in that buffer is independent of its
value in other buffers.  A few variables are always local in every
buffer.  Every other Emacs variable has a @dfn{global} value which is in
effect in all buffers that have not made the variable local.

@findex make-local-variable
  @kbd{M-x make-local-variable} reads the name of a variable and makes
it local to the current buffer.  Changing its value subsequently in
this buffer will not affect others, and changes in its global value
will not affect this buffer.

@findex make-variable-buffer-local
@cindex per-buffer variables
  @kbd{M-x make-variable-buffer-local} marks a variable so it will
become local automatically whenever it is set.  More precisely, once a
variable has been marked in this way, the usual ways of setting the
variable automatically do @code{make-local-variable} first.  We call
such variables @dfn{per-buffer} variables.  Many variables in Emacs
are normally per-buffer; the variable's document string tells you when
this is so.  A per-buffer variable's global value is normally never
effective in any buffer, but it still has a meaning: it is the initial
value of the variable for each new buffer.

  Major modes (@pxref{Major Modes}) always make variables local to the
buffer before setting the variables.  This is why changing major modes
in one buffer has no effect on other buffers.  Minor modes also work
by setting variables---normally, each minor mode has one controlling
variable which is non-@code{nil} when the mode is enabled
(@pxref{Minor Modes}).  For many minor modes, the controlling variable
is per buffer, and thus always buffer-local.  Otherwise, you can make
it local in a specific buffer like any other variable.

  A few variables cannot be local to a buffer because they are always
local to each display instead (@pxref{Multiple Displays}).  If you try to
make one of these variables buffer-local, you'll get an error message.

@findex kill-local-variable
  @kbd{M-x kill-local-variable} makes a specified variable cease to be
local to the current buffer.  The global value of the variable
henceforth is in effect in this buffer.  Setting the major mode kills
all the local variables of the buffer except for a few variables
specially marked as @dfn{permanent locals}.

@findex setq-default
  To set the global value of a variable, regardless of whether the
variable has a local value in the current buffer, you can use the Lisp
construct @code{setq-default}.  This construct is used just like
@code{setq}, but it sets variables' global values instead of their local
values (if any).  When the current buffer does have a local value, the
new global value may not be visible until you switch to another buffer.
Here is an example:

@example
(setq-default fill-column 75)
@end example

@noindent
@code{setq-default} is the only way to set the global value of a variable
that has been marked with @code{make-variable-buffer-local}.

@findex default-value
  Lisp programs can use @code{default-value} to look at a variable's
default value.  This function takes a symbol as argument and returns its
default value.  The argument is evaluated; usually you must quote it
explicitly.  For example, here's how to obtain the default value of
@code{fill-column}:

@example
(default-value 'fill-column)
@end example

@node File Variables
@subsection Local Variables in Files
@cindex local variables in files
@cindex file local variables

  A file can specify local variable values to use when editing the
file with Emacs.  Visiting the file or setting a major mode checks for
local variable specifications; it automatically makes these variables
local to the buffer, and sets them to the values specified in the
file.

  File local variables override directory local variables
(@pxref{Directory Variables}), if any are specified for a file's
directory.

@menu
* Specifying File Variables:: Specifying file local variables.
* Safe File Variables::       Making sure file local variables are safe.
@end menu

@node Specifying File Variables
@subsubsection Specifying File Variables

  There are two ways to specify file local variable values: in the first
line, or with a local variables list.  Here's how to specify them in the
first line:

@example
-*- mode: @var{modename}; @var{var}: @var{value}; @dots{} -*-
@end example

@noindent
You can specify any number of variable/value pairs in this way, each
pair with a colon and semicolon.  The special variable/value pair
@code{mode: @var{modename};}, if present, specifies a major mode.  The
@var{value}s are used literally, and not evaluated.

@findex add-file-local-variable-prop-line
@findex delete-file-local-variable-prop-line
@findex copy-dir-locals-to-file-locals-prop-line
  You can use @kbd{M-x add-file-local-variable-prop-line} instead of
adding entries by hand.  This command prompts for a variable and
value, and adds them to the first line in the appropriate way.
@kbd{M-x delete-file-local-variable-prop-line} prompts for a variable,
and deletes its entry from the line.  The command @kbd{M-x
copy-dir-locals-to-file-locals-prop-line} copies the current
directory-local variables to the first line (@pxref{Directory
Variables}).

  Here is an example first line that specifies Lisp mode and sets two
variables with numeric values:

@smallexample
;; -*- mode: Lisp; fill-column: 75; comment-column: 50; -*-
@end smallexample

@noindent
Aside from @code{mode}, other keywords that have special meanings as
file variables are @code{coding}, @code{unibyte}, and @code{eval}.
These are described below.

@cindex shell scripts, and local file variables
@cindex man pages, and local file variables
  In shell scripts, the first line is used to identify the script
interpreter, so you cannot put any local variables there.  To
accommodate this, Emacs looks for local variable specifications in the
@emph{second} line if the first line specifies an interpreter.  The
same is true for man pages which start with the magic string
@samp{'\"} to specify a list of troff preprocessors (not all do,
however).

  Apart from using a @samp{-*-} line, you can define file local
variables using a @dfn{local variables list} near the end of the file.
The start of the local variables list should be no more than 3000
characters from the end of the file, and must be on the last page if
the file is divided into pages.

  If a file has both a local variables list and a @samp{-*-} line,
Emacs processes @emph{everything} in the @samp{-*-} line first, and
@emph{everything} in the local variables list afterward.  The exception
to this is a major mode specification.  Emacs applies this first,
wherever it appears, since most major modes kill all local variables as
part of their initialization.

  A local variables list starts with a line containing the string
@samp{Local Variables:}, and ends with a line containing the string
@samp{End:}.  In between come the variable names and values, one set
per line, like this:

@example
/* Local Variables:  */
/* mode: c           */
/* comment-column: 0 */
/* End:              */
@end example

@noindent
In this example, each line starts with the prefix @samp{/*} and ends
with the suffix @samp{*/}.  Emacs recognizes the prefix and suffix by
finding them surrounding the magic string @samp{Local Variables:}, on
the first line of the list; it then automatically discards them from
the other lines of the list.  The usual reason for using a prefix
and/or suffix is to embed the local variables list in a comment, so it
won't confuse other programs that the file is intended for.  The
example above is for the C programming language, where comments start
with @samp{/*} and end with @samp{*/}.

If some unrelated text might look to Emacs as a local variables list,
you can countermand that by inserting a form-feed character (a page
delimiter, @pxref{Pages}) after that text.  Emacs only looks for
file-local variables in the last page of a file, after the last page
delimiter.

@findex add-file-local-variable
@findex delete-file-local-variable
@findex copy-dir-locals-to-file-locals
  Instead of typing in the local variables list directly, you can use
the command @kbd{M-x add-file-local-variable}.  This prompts for a
variable and value, and adds them to the list, adding the @samp{Local
Variables:} string and start and end markers as necessary.  The
command @kbd{M-x delete-file-local-variable} deletes a variable from
the list.  @kbd{M-x copy-dir-locals-to-file-locals} copies
directory-local variables to the list (@pxref{Directory Variables}).

  As with the @samp{-*-} line, the variables in a local variables list
are used literally, and are not evaluated first.  If you want to split
a long string value across multiple lines of the file, you can use
backslash-newline, which is ignored in Lisp string constants; you
should put the prefix and suffix on each line, even lines that start
or end within the string, as they will be stripped off when processing
the list.  Here is an example:

@example
# Local Variables:
# compile-command: "cc foo.c -Dfoo=bar -Dhack=whatever \
#   -Dmumble=blaah"
# End:
@end example

  Some names have special meanings in a local variables
list:

@itemize
@item
@code{mode} enables the specified major mode.

@item
@code{eval} evaluates the specified Lisp expression (the value
returned by that expression is ignored).

@item
@code{coding} specifies the coding system for character code
conversion of this file.  @xref{Coding Systems}.

@item
@code{unibyte} says to load or compile a file of Emacs Lisp in unibyte
mode, if the value is @code{t}.  @xref{Disabling Multibyte, ,
Disabling Multibyte Characters, elisp, GNU Emacs Lisp Reference
Manual}.

@end itemize

@noindent
These four keywords are not really variables; setting them in any
other context has no special meaning.

  Do not use the @code{mode} keyword for minor modes.  To enable or
disable a minor mode in a local variables list, use the @code{eval}
keyword with a Lisp expression that runs the mode command
(@pxref{Minor Modes}).  For example, the following local variables
list enables Eldoc mode (@pxref{Lisp Doc}) by calling
@code{eldoc-mode} with no argument (calling it with an argument of 1
would do the same), and disables Font Lock mode (@pxref{Font Lock}) by
calling @code{font-lock-mode} with an argument of @minus{}1.

@example
;; Local Variables:
;; eval: (eldoc-mode)
;; eval: (font-lock-mode -1)
;; End:
@end example

@noindent
Note, however, that it is often a mistake to specify minor modes this
way.  Minor modes represent individual user preferences, and it may be
inappropriate to impose your preferences on another user who might
edit the file.  If you wish to automatically enable or disable a minor
mode in a situation-dependent way, it is often better to do it in a
major mode hook (@pxref{Hooks}).

  Use the command @kbd{M-x normal-mode} to reset the local variables
and major mode of a buffer according to the file name and contents,
including the local variables list if any.  @xref{Choosing Modes}.

@node Safe File Variables
@subsubsection Safety of File Variables

  File-local variables can be dangerous; when you visit someone else's
file, there's no telling what its local variables list could do to
your Emacs.  Improper values of the @code{eval} ``variable'', and
other variables such as @code{load-path}, could execute Lisp code you
didn't intend to run.

  Therefore, whenever Emacs encounters file local variable values that
are not known to be safe, it displays the file's entire local
variables list, and asks you for confirmation before setting them.
You can type @kbd{y} or @key{SPC} to put the local variables list into
effect, or @kbd{n} to ignore it.  When Emacs is run in batch mode
(@pxref{Initial Options}), it can't really ask you, so it assumes the
answer @kbd{n}.

  Emacs normally recognizes certain variable/value pairs as safe.
For instance, it is safe to give @code{comment-column} or
@code{fill-column} any integer value.  If a file specifies only
known-safe variable/value pairs, Emacs does not ask for confirmation
before setting them.  Otherwise, you can tell Emacs to record all the
variable/value pairs in this file as safe, by typing @kbd{!} at the
confirmation prompt.  When Emacs encounters these variable/value pairs
subsequently, in the same file or others, it will assume they are
safe.

@vindex safe-local-variable-values
@cindex risky variable
  Some variables, such as @code{load-path}, are considered
particularly @dfn{risky}: there is seldom any reason to specify them
as local variables, and changing them can be dangerous.  If a file
contains only risky local variables, Emacs neither offers nor accepts
@kbd{!} as input at the confirmation prompt.  If some of the local
variables in a file are risky, and some are only potentially unsafe, you
can enter @kbd{!} at the prompt.  It applies all the variables, but only
marks the non-risky ones as safe for the future.  If you really want to
record safe values for risky variables, do it directly by customizing
@samp{safe-local-variable-values} (@pxref{Easy Customization}).

@vindex enable-local-variables
  The variable @code{enable-local-variables} allows you to change the
way Emacs processes local variables.  Its default value is @code{t},
which specifies the behavior described above.  If it is @code{nil},
Emacs simply ignores all file local variables.  @code{:safe} means use
only the safe values and ignore the rest.  Any other value says to
query you about each file that has local variables, without trying to
determine whether the values are known to be safe.

@vindex enable-local-eval
@vindex safe-local-eval-forms
  The variable @code{enable-local-eval} controls whether Emacs
processes @code{eval} variables.  The three possibilities for the
variable's value are @code{t}, @code{nil}, and anything else, just as
for @code{enable-local-variables}.  The default is @code{maybe}, which
is neither @code{t} nor @code{nil}, so normally Emacs does ask for
confirmation about processing @code{eval} variables.

  As an exception, Emacs never asks for confirmation to evaluate any
@code{eval} form if that form occurs within the variable
@code{safe-local-eval-forms}.

@node Directory Variables
@subsection Per-Directory Local Variables
@cindex local variables, for all files in a directory
@cindex directory-local variables
@cindex per-directory local variables

  Sometimes, you may wish to define the same set of local variables to
all the files in a certain directory and its subdirectories, such as
the directory tree of a large software project.  This can be
accomplished with @dfn{directory-local variables}.  File local
variables override directory local variables, so if some of the files
in a directory need specialized settings, you can specify the settings
for the majority of the directory's files in directory variables, and
then define file local variables in a few files which need the general
settings overridden.

@cindex @file{.dir-locals.el} file
  The usual way to define directory-local variables is to put a file
named @file{.dir-locals.el}@footnote{
On MS-DOS, the name of this file should be @file{_dir-locals.el}, due
to limitations of the DOS filesystems.  If the filesystem is limited
to 8+3 file names, the name of the file will be truncated by the OS to
@file{_dir-loc.el}.
}in a directory.  Whenever Emacs visits any file in that directory or
any of its subdirectories, it will apply the directory-local variables
specified in @file{.dir-locals.el}, as though they had been defined as
file-local variables for that file (@pxref{File Variables}).  Emacs
searches for @file{.dir-locals.el} starting in the directory of the
visited file, and moving up the directory tree.  To avoid slowdown,
this search is skipped for remote files.  If needed, the search can be
extended for remote files by setting the variable
@code{enable-remote-dir-locals} to @code{t}.

  You can also use @file{.dir-locals-2.el}; if found, Emacs loads it
in addition to @file{.dir-locals.el}.  This is useful when
@file{.dir-locals.el} is under version control in a shared repository
and can't be used for personal customizations.

  The @file{.dir-locals.el} file should hold a specially-constructed
list, which maps major mode names (symbols) to alists
(@pxref{Association Lists,,, elisp, The Emacs Lisp Reference Manual}).
Each alist entry consists of a variable name and the directory-local
value to assign to that variable, when the specified major mode is
enabled.  Instead of a mode name, you can specify @samp{nil}, which
means that the alist applies to any mode; or you can specify a
subdirectory (a string), in which case the alist applies to all
files in that subdirectory.

  Here's an example of a @file{.dir-locals.el} file:

@example
((nil . ((indent-tabs-mode . t)
         (fill-column . 80)))
 (c-mode . ((c-file-style . "BSD")
            (subdirs . nil)))
 ("src/imported"
  . ((nil . ((change-log-default-name
              . "ChangeLog.local"))))))
@end example

@noindent
This sets @samp{indent-tabs-mode} and @code{fill-column} for any file
in the directory tree, and the indentation style for any C source
file.  The special @code{subdirs} element is not a variable, but a
special keyword which indicates that the C mode settings are only to
be applied in the current directory, not in any subdirectories.
Finally, it specifies a different @file{ChangeLog} file name for any
file in the @file{src/imported} subdirectory.

You can specify the variables @code{mode}, @code{eval}, and
@code{unibyte} in your @file{.dir-locals.el}, and they have the same
meanings as they would have in file local variables.  @code{coding}
cannot be specified as a directory local variable.  @xref{File
Variables}.

@findex add-dir-local-variable
@findex delete-dir-local-variable
@findex copy-file-locals-to-dir-locals
  Instead of editing the @file{.dir-locals.el} file by hand, you can
use the command @kbd{M-x add-dir-local-variable}.  This prompts for a
mode or subdirectory, and for variable and value, and adds the
entry defining the directory-local variable.  @kbd{M-x
delete-dir-local-variable} deletes an entry.  @kbd{M-x
copy-file-locals-to-dir-locals} copies the file-local variables in the
current file into @file{.dir-locals.el}.

@findex dir-locals-set-class-variables
@findex dir-locals-set-directory-class
  Another method of specifying directory-local variables is to define
a group of variables/value pairs in a @dfn{directory class}, using the
@code{dir-locals-set-class-variables} function; then, tell Emacs which
directories correspond to the class by using the
@code{dir-locals-set-directory-class} function.  These function calls
normally go in your initialization file (@pxref{Init File}).  This
method is useful when you can't put @file{.dir-locals.el} in a
directory for some reason.  For example, you could apply settings to
an unwritable directory this way:

@example
(dir-locals-set-class-variables 'unwritable-directory
   '((nil . ((some-useful-setting . value)))))

(dir-locals-set-directory-class
   "/usr/include/" 'unwritable-directory)
@end example

  If a variable has both a directory-local and file-local value
specified, the file-local value takes effect.  Unsafe directory-local
variables are handled in the same way as unsafe file-local variables
(@pxref{Safe File Variables}).

  Directory-local variables also take effect in certain buffers that
do not visit a file directly but perform work within a directory, such
as Dired buffers (@pxref{Dired}).

@node Connection Variables
@subsection Per-Connection Local Variables
@cindex local variables, for all remote connections
@cindex connection-local variables
@cindex per-connection local variables

  Most of the variables reflect the situation on the local machine.
Often, they must use a different value when you operate in buffers
with a remote default directory.  Think about the shell to be applied
when calling @code{shell} -- it might be @file{/bin/bash} on your
local machine, and @file{/bin/ksh} on a remote machine.

  This can be accomplished with @dfn{connection-local variables}.
Directory and file local variables override connection-local
variables.  Unsafe connection-local variables are handled in the same
way as unsafe file-local variables (@pxref{Safe File Variables}).

@findex connection-local-set-profile-variables
@findex connection-local-set-profiles
  Connection-local variables are declared as a group of
variables/value pairs in a @dfn{profile}, using the
@code{connection-local-set-profile-variables} function.  The function
@code{connection-local-set-profiles} activates profiles for a given
criteria, identifying a remote machine:

@example
(connection-local-set-profile-variables 'remote-ksh
   '((shell-file-name . "/bin/ksh")
     (shell-command-switch . "-c")))

(connection-local-set-profile-variables 'remote-bash
   '((shell-file-name . "/bin/bash")
     (shell-command-switch . "-c")))

(connection-local-set-profiles
   '(:application tramp :machine "remotemachine") 'remote-ksh)
@end example

  This code declares two different profiles, @code{remote-ksh} and
@code{remote-bash}. The profile @code{remote-ksh} is applied to all
buffers which have a remote default directory matching the regexp
@code{"remotemachine} as host name.  Such a criteria can also
discriminate for the properties @code{:protocol} (this is the Tramp
method) or @code{:user} (a remote user name).  The @code{nil} criteria
matches all buffers with a remote default directory.

@node Key Bindings
@section Customizing Key Bindings
@cindex key bindings

  This section describes @dfn{key bindings}, which map keys to
commands, and @dfn{keymaps}, which record key bindings.  It also
explains how to customize key bindings, which is done by editing your
init file (@pxref{Init Rebinding}).

@cindex reserved key bindings
@cindex keys, reserved
  Since most modes define their own key bindings, activating a mode
might override your custom key bindings.  A small number of keys are
reserved for user-defined bindings, and should not be used by modes,
so key bindings using those keys are safer in this regard.  The
reserved key sequences are those consisting of @kbd{C-c} followed by a
letter (either upper or lower case), and function keys @key{F5}
through @key{F9} without modifiers (@pxref{Modifier Keys}).

@menu
* Keymaps::             Generalities.  The global keymap.
* Prefix Keymaps::      Keymaps for prefix keys.
* Local Keymaps::       Major and minor modes have their own keymaps.
* Minibuffer Maps::     The minibuffer uses its own local keymaps.
* Rebinding::           How to redefine one key's meaning conveniently.
* Init Rebinding::      Rebinding keys with your initialization file.
* Modifier Keys::       Using modifier keys.
* Function Keys::       Rebinding terminal function keys.
* Named ASCII Chars::   Distinguishing @key{TAB} from @kbd{C-i}, and so on.
* Mouse Buttons::       Rebinding mouse buttons in Emacs.
* Disabling::           Disabling a command means confirmation is required
                          before it can be executed.  This is done to protect
                          beginners from surprises.
@end menu

@node Keymaps
@subsection Keymaps
@cindex keymap

  As described in @ref{Commands}, each Emacs command is a Lisp
function whose definition provides for interactive use.  Like every
Lisp function, a command has a function name, which usually consists
of lower-case letters and hyphens.

  A @dfn{key sequence} (@dfn{key}, for short) is a sequence of
@dfn{input events} that have a meaning as a unit.  Input events
include characters, function keys, and mouse buttons---all the inputs
that you can send to the computer.  A key sequence gets its meaning
from its @dfn{binding}, which says what command it runs.

  The bindings between key sequences and command functions are
recorded in data structures called @dfn{keymaps}.  Emacs has many of
these, each used on particular occasions.

@cindex global keymap
  The @dfn{global} keymap is the most important keymap because it is
always in effect.  The global keymap defines keys for Fundamental mode
(@pxref{Major Modes}); most of these definitions are common to most or
all major modes.  Each major or minor mode can have its own keymap
which overrides the global definitions of some keys.

  For example, a self-inserting character such as @kbd{g} is
self-inserting because the global keymap binds it to the command
@code{self-insert-command}.  The standard Emacs editing characters
such as @kbd{C-a} also get their standard meanings from the global
keymap.  Commands to rebind keys, such as @kbd{M-x global-set-key},
work by storing the new binding in the proper place in the global map
(@pxref{Rebinding}).

@cindex function key
  Most modern keyboards have function keys as well as character keys.
Function keys send input events just as character keys do, and keymaps
can have bindings for them.  Key sequences can mix function keys and
characters.  For example, if your keyboard has a @key{Home} function
key, Emacs can recognize key sequences like @kbd{C-x @key{Home}}.  You
can even mix mouse events with keyboard events, such as
@kbd{S-down-mouse-1}.

  On text terminals, typing a function key actually sends the computer
a sequence of characters; the precise details of the sequence depend
on the function key and on the terminal type.  (Often the sequence
starts with @kbd{@key{ESC} [}.)  If Emacs understands your terminal
type properly, it automatically handles such sequences as single input
events.

@node Prefix Keymaps
@subsection Prefix Keymaps

  Internally, Emacs records only single events in each keymap.
Interpreting a key sequence of multiple events involves a chain of
keymaps: the first keymap gives a definition for the first event,
which is another keymap, which is used to look up the second event in
the sequence, and so on.  Thus, a prefix key such as @kbd{C-x} or
@key{ESC} has its own keymap, which holds the definition for the event
that immediately follows that prefix.

  The definition of a prefix key is usually the keymap to use for
looking up the following event.  The definition can also be a Lisp
symbol whose function definition is the following keymap; the effect is
the same, but it provides a command name for the prefix key that can be
used as a description of what the prefix key is for.  Thus, the binding
of @kbd{C-x} is the symbol @code{Control-X-prefix}, whose function
definition is the keymap for @kbd{C-x} commands.  The definitions of
@kbd{C-c}, @kbd{C-x}, @kbd{C-h}, and @key{ESC} as prefix keys appear in
the global map, so these prefix keys are always available.

  Aside from ordinary prefix keys, there is a fictitious ``prefix key''
which represents the menu bar; see @ref{Menu Bar,,,elisp, The Emacs Lisp
Reference Manual}, for special information about menu bar key bindings.
Mouse button events that invoke pop-up menus are also prefix keys; see
@ref{Menu Keymaps,,,elisp, The Emacs Lisp Reference Manual}, for more
details.

  Some prefix keymaps are stored in variables with names:

@itemize @bullet
@item
@vindex ctl-x-map
@code{ctl-x-map} is the variable name for the map used for characters that
follow @kbd{C-x}.
@item
@vindex help-map
@code{help-map} is for characters that follow @kbd{C-h}.
@item
@vindex esc-map
@code{esc-map} is for characters that follow @key{ESC}.  Thus, all Meta
characters are actually defined by this map.
@item
@vindex ctl-x-4-map
@code{ctl-x-4-map} is for characters that follow @kbd{C-x 4}.
@item
@vindex mode-specific-map
@code{mode-specific-map} is for characters that follow @kbd{C-c}.
@end itemize

@node Local Keymaps
@subsection Local Keymaps

@cindex local keymap
@cindex minor mode keymap
  So far, we have explained the ins and outs of the global map.  Major
modes customize Emacs by providing their own key bindings in
@dfn{local keymaps}.  For example, C mode overrides @key{TAB} to make
it indent the current line for C code.  Minor modes can also have
local keymaps; whenever a minor mode is in effect, the definitions in
its keymap override both the major mode's local keymap and the global
keymap.  In addition, portions of text in the buffer can specify their
own keymaps, which override all other keymaps.

  A local keymap can redefine a key as a prefix key by defining it as
a prefix keymap.  If the key is also defined globally as a prefix, its
local and global definitions (both keymaps) effectively combine: both
definitions are used to look up the event that follows the prefix key.
For example, if a local keymap defines @kbd{C-c} as a prefix keymap,
and that keymap defines @kbd{C-z} as a command, this provides a local
meaning for @kbd{C-c C-z}.  This does not affect other sequences that
start with @kbd{C-c}; if those sequences don't have their own local
bindings, their global bindings remain in effect.

  Another way to think of this is that Emacs handles a multi-event key
sequence by looking in several keymaps, one by one, for a binding of the
whole key sequence.  First it checks the minor mode keymaps for minor
modes that are enabled, then it checks the major mode's keymap, and then
it checks the global keymap.  This is not precisely how key lookup
works, but it's good enough for understanding the results in ordinary
circumstances.

@node Minibuffer Maps
@subsection Minibuffer Keymaps

@cindex minibuffer keymaps
@vindex minibuffer-local-map
@vindex minibuffer-local-ns-map
@vindex minibuffer-local-completion-map
@vindex minibuffer-local-must-match-map
@vindex minibuffer-local-filename-completion-map
@vindex minibuffer-local-filename-must-match-map
  The minibuffer has its own set of local keymaps; they contain various
completion and exit commands.

@itemize @bullet
@item
@code{minibuffer-local-map} is used for ordinary input (no completion).
@item
@code{minibuffer-local-ns-map} is similar, except that @key{SPC} exits
just like @key{RET}.
@item
@code{minibuffer-local-completion-map} is for permissive completion.
@item
@code{minibuffer-local-must-match-map} is for strict completion and
for cautious completion.
@item
@code{minibuffer-local-filename-completion-map} and
@code{minibuffer-local-filename-must-match-map} are like the two
previous ones, but they are specifically for file name completion.
They do not bind @key{SPC}.
@end itemize

@node Rebinding
@subsection Changing Key Bindings Interactively
@cindex key rebinding, this session
@cindex redefining keys, this session
@cindex binding keys

  The way to redefine an Emacs key is to change its entry in a keymap.
You can change the global keymap, in which case the change is
effective in all major modes (except those that have their own
overriding local bindings for the same key).  Or you can change a
local keymap, which affects all buffers using the same major mode.

  In this section, we describe how to rebind keys for the present
Emacs session.  @xref{Init Rebinding}, for a description of how to
make key rebindings affect future Emacs sessions.

@findex global-set-key
@findex local-set-key
@findex global-unset-key
@findex local-unset-key
@table @kbd
@item M-x global-set-key @key{RET} @var{key} @var{cmd} @key{RET}
Define @var{key} globally to run @var{cmd}.
@item M-x local-set-key @key{RET} @var{key} @var{cmd} @key{RET}
Define @var{key} locally (in the major mode now in effect) to run
@var{cmd}.
@item M-x global-unset-key @key{RET} @var{key}
Make @var{key} undefined in the global map.
@item M-x local-unset-key @key{RET} @var{key}
Make @var{key} undefined locally (in the major mode now in effect).
@end table

  For example, the following binds @kbd{C-z} to the @code{shell}
command (@pxref{Interactive Shell}), replacing the normal global
definition of @kbd{C-z}:

@example
M-x global-set-key @key{RET} C-z shell @key{RET}
@end example

@noindent
The @code{global-set-key} command reads the command name after the
key.  After you press the key, a message like this appears so that you
can confirm that you are binding the key you want:

@example
Set key C-z to command:
@end example

  You can redefine function keys and mouse events in the same way; just
type the function key or click the mouse when it's time to specify the
key to rebind.

  You can rebind a key that contains more than one event in the same
way.  Emacs keeps reading the key to rebind until it is a complete key
(that is, not a prefix key).  Thus, if you type @kbd{C-f} for
@var{key}, that's the end; it enters the minibuffer immediately to
read @var{cmd}.  But if you type @kbd{C-x}, since that's a prefix, it
reads another character; if that is @kbd{4}, another prefix character,
it reads one more character, and so on.  For example,

@example
M-x global-set-key @key{RET} C-x 4 $ spell-other-window @key{RET}
@end example

@noindent
redefines @kbd{C-x 4 $} to run the (fictitious) command
@code{spell-other-window}.

  You can remove the global definition of a key with
@code{global-unset-key}.  This makes the key @dfn{undefined}; if you
type it, Emacs will just beep.  Similarly, @code{local-unset-key} makes
a key undefined in the current major mode keymap, which makes the global
definition (or lack of one) come back into effect in that major mode.

  If you have redefined (or undefined) a key and you subsequently wish
to retract the change, undefining the key will not do the job---you need
to redefine the key with its standard definition.  To find the name of
the standard definition of a key, go to a Fundamental mode buffer in a
fresh Emacs and use @kbd{C-h c}.  The documentation of keys in this
manual also lists their command names.

  If you want to prevent yourself from invoking a command by mistake, it
is better to disable the command than to undefine the key.  A disabled
command is less work to invoke when you really want to.
@xref{Disabling}.

@node Init Rebinding
@subsection Rebinding Keys in Your Init File
@cindex rebinding major mode keys
@cindex key rebinding, permanent
@cindex rebinding keys, permanently
@c This node is referenced in the tutorial.  When renaming or deleting
@c it, the tutorial needs to be adjusted.  (TUTORIAL.de)

  If you have a set of key bindings that you like to use all the time,
you can specify them in your initialization file by writing Lisp code.
@xref{Init File}, for a description of the initialization file.

@findex kbd
  There are several ways to write a key binding using Lisp.  The
simplest is to use the @code{kbd} function, which converts a textual
representation of a key sequence---similar to how we have written key
sequences in this manual---into a form that can be passed as an
argument to @code{global-set-key}.  For example, here's how to bind
@kbd{C-z} to the @code{shell} command (@pxref{Interactive Shell}):

@example
(global-set-key (kbd "C-z") 'shell)
@end example

@noindent
The single-quote before the command name, @code{shell}, marks it as a
constant symbol rather than a variable.  If you omit the quote, Emacs
would try to evaluate @code{shell} as a variable.  This probably
causes an error; it certainly isn't what you want.

  Here are some additional examples, including binding function keys
and mouse events:

@example
(global-set-key (kbd "C-c y") 'clipboard-yank)
(global-set-key (kbd "C-M-q") 'query-replace)
(global-set-key (kbd "<f5>") 'flyspell-mode)
(global-set-key (kbd "C-<f5>") 'display-line-numbers-mode)
(global-set-key (kbd "C-<right>") 'forward-sentence)
(global-set-key (kbd "<mouse-2>") 'mouse-save-then-kill)
@end example

  Instead of using @code{kbd}, you can use a Lisp string or vector to
specify the key sequence.  Using a string is simpler, but only works
for @acronym{ASCII} characters and Meta-modified @acronym{ASCII}
characters.  For example, here's how to bind @kbd{C-x M-l} to
@code{make-symbolic-link} (@pxref{Copying and Naming}):

@example
(global-set-key "\C-x\M-l" 'make-symbolic-link)
@end example

  To bind a key sequence including @key{TAB}, @key{RET}, @key{ESC}, or
@key{DEL}, the string should contain the Emacs Lisp escape sequence
@samp{\t}, @samp{\r}, @samp{\e}, or @samp{\d} respectively.  Here is
an example which binds @kbd{C-x @key{TAB}} to @code{indent-rigidly}
(@pxref{Indentation}):

@example
(global-set-key "\C-x\t" 'indent-rigidly)
@end example

  When the key sequence includes function keys or mouse button events,
or non-@acronym{ASCII} characters such as @code{C-=} or @code{H-a},
you can use a vector to specify the key sequence.  Each element in the
vector stands for an input event; the elements are separated by spaces
and surrounded by a pair of square brackets.  If a vector element is a
character, write it as a Lisp character constant: @samp{?} followed by
the character as it would appear in a string.  Function keys are
represented by symbols (@pxref{Function Keys}); simply write the
symbol's name, with no other delimiters or punctuation.  Here are some
examples:

@example
(global-set-key [?\C-=] 'make-symbolic-link)
(global-set-key [?\M-\C-=] 'make-symbolic-link)
(global-set-key [?\H-a] 'make-symbolic-link)
(global-set-key [f7] 'make-symbolic-link)
(global-set-key [C-mouse-1] 'make-symbolic-link)
@end example

@noindent
You can use a vector for the simple cases too:

@example
(global-set-key [?\C-z ?\M-l] 'make-symbolic-link)
@end example

  Language and coding systems may cause problems with key bindings for
non-@acronym{ASCII} characters.  @xref{Init Non-ASCII}.

  As described in @ref{Local Keymaps}, major modes and minor modes can
define local keymaps.  These keymaps are constructed when the mode is
used for the first time in a session.  If you wish to change one of
these keymaps, you must use the @dfn{mode hook} (@pxref{Hooks}).

@findex define-key
  For example, Texinfo mode runs the hook @code{texinfo-mode-hook}.
Here's how you can use the hook to add local bindings for @kbd{C-c n}
and @kbd{C-c p} in Texinfo mode:

@example
(add-hook 'texinfo-mode-hook
          (lambda ()
            (define-key texinfo-mode-map "\C-cp"
                        'backward-paragraph)
            (define-key texinfo-mode-map "\C-cn"
                        'forward-paragraph)))
@end example

@node Modifier Keys
@subsection Modifier Keys
@cindex modifier keys, and key rebinding

  The default key bindings in Emacs are set up so that modified
alphabetical characters are case-insensitive.  In other words,
@kbd{C-A} does the same thing as @kbd{C-a}, and @kbd{M-A} does the
same thing as @kbd{M-a}.  This concerns only alphabetical characters,
and does not apply to shifted versions of other keys; for
instance, @kbd{C-@@} is not the same as @kbd{C-2}.

  A @key{Control}-modified alphabetical character is always considered
case-insensitive: Emacs always treats @kbd{C-A} as @kbd{C-a},
@kbd{C-B} as @kbd{C-b}, and so forth.  The reason for this is
historical.

  For all other modifiers, you can make the modified alphabetical
characters case-sensitive when you customize Emacs.  For instance, you
could make @kbd{M-a} and @kbd{M-A} run different commands.

  Although only the @key{Control} and @key{Meta} modifier keys are
commonly used, Emacs supports three other modifier keys.  These are
called @key{Super}, @key{Hyper}, and @key{Alt}.  Few terminals provide
ways to use these modifiers; the key labeled @key{Alt} on most
keyboards usually issues the @key{Meta} modifier, not @key{Alt}.  The
standard key bindings in Emacs do not include any characters with
these modifiers.  However, you can customize Emacs to assign meanings
to them.  The modifier bits are labeled as @samp{s-}, @samp{H-} and
@samp{A-} respectively.

  Even if your keyboard lacks these additional modifier keys, you can
enter it using @kbd{C-x @@}: @kbd{C-x @@ h} adds the Hyper flag to
the next character, @kbd{C-x @@ s} adds the Super flag, and
@kbd{C-x @@ a} adds the Alt flag.  For instance, @kbd{C-x @@ h
C-a} is a way to enter @kbd{Hyper-Control-a}.  (Unfortunately, there
is no way to add two modifiers by using @kbd{C-x @@} twice for the
same character, because the first one goes to work on the @kbd{C-x}.)

@node Function Keys
@subsection Rebinding Function Keys

  Key sequences can contain function keys as well as ordinary
characters.  Just as Lisp characters (actually integers) represent
keyboard characters, Lisp symbols represent function keys.  If the
function key has a word as its label, then that word is also the name of
the corresponding Lisp symbol.  Here are the conventional Lisp names for
common function keys:

@table @asis
@item @code{left}, @code{up}, @code{right}, @code{down}
Cursor arrow keys.

@item @code{begin}, @code{end}, @code{home}, @code{next}, @code{prior}
Other cursor repositioning keys.

@item @code{select}, @code{print}, @code{execute}, @code{backtab}
@itemx @code{insert}, @code{undo}, @code{redo}, @code{clearline}
@itemx @code{insertline}, @code{deleteline}, @code{insertchar}, @code{deletechar}
Miscellaneous function keys.

@item @code{f1}, @code{f2}, @dots{} @code{f35}
Numbered function keys (across the top of the keyboard).

@item @code{kp-add}, @code{kp-subtract}, @code{kp-multiply}, @code{kp-divide}
@itemx @code{kp-backtab}, @code{kp-space}, @code{kp-tab}, @code{kp-enter}
@itemx @code{kp-separator}, @code{kp-decimal}, @code{kp-equal}
@itemx @code{kp-prior}, @code{kp-next}, @code{kp-end}, @code{kp-home}
@itemx @code{kp-left}, @code{kp-up}, @code{kp-right}, @code{kp-down}
@itemx @code{kp-insert }, @code{kp-delete}
Keypad keys (to the right of the regular keyboard), with names or punctuation.

@item @code{kp-0}, @code{kp-1}, @dots{} @code{kp-9}
Keypad keys with digits.

@item @code{kp-f1}, @code{kp-f2}, @code{kp-f3}, @code{kp-f4}
Keypad PF keys.
@end table

  These names are conventional, but some systems (especially when using
X) may use different names.  To make certain what symbol is used for a
given function key on your terminal, type @kbd{C-h c} followed by that
key.

  @xref{Init Rebinding}, for examples of binding function keys.

@cindex keypad
  Many keyboards have a numeric keypad on the right-hand side.
The numeric keys in the keypad double up as cursor motion keys,
toggled by a key labeled @samp{Num Lock}.  By default, Emacs
translates these keys to the corresponding keys on the main keyboard.
For example, when @samp{Num Lock} is on, the key labeled @samp{8} on
the numeric keypad produces @code{kp-8}, which is translated to
@kbd{8}; when @samp{Num Lock} is off, the same key produces
@code{kp-up}, which is translated to @key{UP}.  If you rebind a key
such as @kbd{8} or @key{UP}, it affects the equivalent keypad key too.
However, if you rebind a @samp{kp-} key directly, that won't affect
its non-keypad equivalent.  Note that the modified keys are not
translated: for instance, if you hold down the @key{Meta} key while
pressing the @samp{8} key on the numeric keypad, that generates
@kbd{M-@key{kp-8}}.

  Emacs provides a convenient method for binding the numeric keypad
keys, using the variables @code{keypad-setup},
@code{keypad-numlock-setup}, @code{keypad-shifted-setup}, and
@code{keypad-numlock-shifted-setup}.  These can be found in the
@samp{keyboard} customization group (@pxref{Easy Customization}).  You
can rebind the keys to perform other tasks, such as issuing numeric
prefix arguments.

@node Named ASCII Chars
@subsection Named @acronym{ASCII} Control Characters

  @key{TAB}, @key{RET}, @key{BS}, @key{LFD}, @key{ESC}, and @key{DEL}
started out as names for certain @acronym{ASCII} control characters,
used so often that they have special keys of their own.  For instance,
@key{TAB} was another name for @kbd{C-i}.  Later, users found it
convenient to distinguish in Emacs between these keys and the corresponding
control characters typed with the @key{Ctrl} key.  Therefore, on most
modern terminals, they are no longer the same: @key{TAB} is different
from @kbd{C-i}.

  Emacs can distinguish these two kinds of input if the keyboard does.
It treats the special keys as function keys named @code{tab},
@code{return}, @code{backspace}, @code{linefeed}, @code{escape}, and
@code{delete}.  These function keys translate automatically into the
corresponding @acronym{ASCII} characters @emph{if} they have no
bindings of their own.  As a result, neither users nor Lisp programs
need to pay attention to the distinction unless they care to.

  If you do not want to distinguish between (for example) @key{TAB} and
@kbd{C-i}, make just one binding, for the @acronym{ASCII} character @key{TAB}
(octal code 011).  If you do want to distinguish, make one binding for
this @acronym{ASCII} character, and another for the function key @code{tab}.

  With an ordinary @acronym{ASCII} terminal, there is no way to distinguish
between @key{TAB} and @kbd{C-i} (and likewise for other such pairs),
because the terminal sends the same character in both cases.

@node Mouse Buttons
@subsection Rebinding Mouse Buttons
@cindex mouse button events
@cindex rebinding mouse buttons
@cindex click events
@cindex drag events
@cindex down events
@cindex button-down events

  Emacs uses Lisp symbols to designate mouse buttons, too.  The ordinary
mouse events in Emacs are @dfn{click} events; these happen when you
press a button and release it without moving the mouse.  You can also
get @dfn{drag} events, when you move the mouse while holding the button
down.  Drag events happen when you finally let go of the button.

  The symbols for basic click events are @code{mouse-1} for the leftmost
button, @code{mouse-2} for the next, and so on.  Here is how you can
redefine the second mouse button to split the current window:

@example
(global-set-key [mouse-2] 'split-window-below)
@end example

  The symbols for drag events are similar, but have the prefix
@samp{drag-} before the word @samp{mouse}.  For example, dragging the
first button generates a @code{drag-mouse-1} event.

  You can also define bindings for events that occur when a mouse button
is pressed down.  These events start with @samp{down-} instead of
@samp{drag-}.  Such events are generated only if they have key bindings.
When you get a button-down event, a corresponding click or drag event
will always follow.

@cindex double clicks
@cindex triple clicks
  If you wish, you can distinguish single, double, and triple clicks.  A
double click means clicking a mouse button twice in approximately the
same place.  The first click generates an ordinary click event.  The
second click, if it comes soon enough, generates a double-click event
instead.  The event type for a double-click event starts with
@samp{double-}: for example, @code{double-mouse-3}.

  This means that you can give a special meaning to the second click at
the same place, but it must act on the assumption that the ordinary
single click definition has run when the first click was received.

  This constrains what you can do with double clicks, but user interface
designers say that this constraint ought to be followed in any case.  A
double click should do something similar to the single click, only
more so.  The command for the double-click event should perform the
extra work for the double click.

  If a double-click event has no binding, it changes to the
corresponding single-click event.  Thus, if you don't define a
particular double click specially, it executes the single-click command
twice.

  Emacs also supports triple-click events whose names start with
@samp{triple-}.  Emacs does not distinguish quadruple clicks as event
types; clicks beyond the third generate additional triple-click events.
However, the full number of clicks is recorded in the event list, so
if you know Emacs Lisp you can distinguish if you really want to
(@pxref{Click Events,,, elisp, The Emacs Lisp Reference Manual}).
We don't recommend distinct meanings for more than three clicks, but
sometimes it is useful for subsequent clicks to cycle through the same
set of three meanings, so that four clicks are equivalent to one
click, five are equivalent to two, and six are equivalent to three.

  Emacs also records multiple presses in drag and button-down events.
For example, when you press a button twice, then move the mouse while
holding the button, Emacs gets a @samp{double-drag-} event.  And at the
moment when you press it down for the second time, Emacs gets a
@samp{double-down-} event (which is ignored, like all button-down
events, if it has no binding).

@vindex double-click-time
  The variable @code{double-click-time} specifies how much time can
elapse between clicks and still allow them to be grouped as a multiple
click.  Its value is in units of milliseconds.  If the value is
@code{nil}, double clicks are not detected at all.  If the value is
@code{t}, then there is no time limit.  The default is 500.

@vindex double-click-fuzz
  The variable @code{double-click-fuzz} specifies how much the mouse
can move between clicks and still allow them to be grouped as a multiple
click.  Its value is in units of pixels on windowed displays and in
units of 1/8 of a character cell on text-mode terminals; the default is
3.

  The symbols for mouse events also indicate the status of the modifier
keys, with the usual prefixes @samp{C-}, @samp{M-}, @samp{H-},
@samp{s-}, @samp{A-}, and @samp{S-}.  These always precede @samp{double-}
or @samp{triple-}, which always precede @samp{drag-} or @samp{down-}.

  A frame includes areas that don't show text from the buffer, such as
the mode line and the scroll bar.  You can tell whether a mouse button
comes from a special area of the screen by means of dummy prefix
keys.  For example, if you click the mouse in the mode line, you get
the prefix key @code{mode-line} before the ordinary mouse-button symbol.
Thus, here is how to define the command for clicking the first button in
a mode line to run @code{scroll-up-command}:

@example
(global-set-key [mode-line mouse-1] 'scroll-up-command)
@end example

  Here is the complete list of these dummy prefix keys and their
meanings:

@table @code
@item mode-line
The mouse was in the mode line of a window.
@item vertical-line
The mouse was in the vertical line separating side-by-side windows.  (If
you use scroll bars, they appear in place of these vertical lines.)
@item vertical-scroll-bar
The mouse was in a vertical scroll bar.  (This is the only kind of
scroll bar Emacs currently supports.)
@item menu-bar
The mouse was in the menu bar.
@item header-line
The mouse was in a header line.
@ignore
@item horizontal-scroll-bar
The mouse was in a horizontal scroll bar.  Horizontal scroll bars do
horizontal scrolling, and people don't use them often.
@end ignore
@end table

  You can put more than one mouse button in a key sequence, but it isn't
usual to do so.

@node Disabling
@subsection Disabling Commands
@cindex disabled command

  Disabling a command means that invoking it interactively asks for
confirmation from the user.  The purpose of disabling a command is to
prevent users from executing it by accident; we do this for commands
that might be confusing to the uninitiated.

  Attempting to invoke a disabled command interactively in Emacs
displays a window containing the command's name, its documentation,
and some instructions on what to do immediately; then Emacs asks for
input saying whether to execute the command as requested, enable it
and execute it, or cancel.  If you decide to enable the command, you
must then answer another question---whether to do this permanently, or
just for the current session.  (Enabling permanently works by
automatically editing your initialization file.)  You can also type
@kbd{!} to enable @emph{all} commands, for the current session only.

  The direct mechanism for disabling a command is to put a
non-@code{nil} @code{disabled} property on the Lisp symbol for the
command.  Here is the Lisp program to do this:

@example
(put 'delete-region 'disabled t)
@end example

  If the value of the @code{disabled} property is a string, that string
is included in the message displayed when the command is used:

@example
(put 'delete-region 'disabled
     "It's better to use `kill-region' instead.\n")
@end example

@findex disable-command
@findex enable-command
  You can make a command disabled either by editing the initialization
file directly, or with the command @kbd{M-x disable-command}, which
edits the initialization file for you.  Likewise, @kbd{M-x
enable-command} edits the initialization file to enable a command
permanently.  @xref{Init File}.

  If Emacs was invoked with the @option{-q} or @option{--no-init-file}
options (@pxref{Initial Options}), it will not edit your
initialization file.  Doing so could lose information because Emacs
has not read your initialization file.

  Whether a command is disabled is independent of what key is used to
invoke it; disabling also applies if the command is invoked using
@kbd{M-x}.  However, disabling a command has no effect on calling it
as a function from Lisp programs.

@node Init File
@section The Emacs Initialization File
@cindex init file
@cindex .emacs file
@cindex ~/.emacs file
@cindex Emacs initialization file
@cindex startup (init file)

  When Emacs is started, it normally tries to load a Lisp program from
an @dfn{initialization file}, or @dfn{init file} for short.  This
file, if it exists, specifies how to initialize Emacs for you.  Emacs
looks for your init file using the filenames @file{~/.emacs},
@file{~/.emacs.el}, or @file{~/.emacs.d/init.el}; you can choose to
use any one of these three names (@pxref{Find Init}).  Here, @file{~/}
stands for your home directory.

  You can use the command line switch @samp{-q} to prevent loading
your init file, and @samp{-u} (or @samp{--user}) to specify a
different user's init file (@pxref{Initial Options}).

@cindex @file{default.el}, the default init file
  There can also be a @dfn{default init file}, which is the library
named @file{default.el}, found via the standard search path for
libraries.  The Emacs distribution contains no such library; your site
may create one for local customizations.  If this library exists, it is
loaded whenever you start Emacs (except when you specify @samp{-q}).
But your init file, if any, is loaded first; if it sets
@code{inhibit-default-init} non-@code{nil}, then @file{default} is not
loaded.

@cindex site init file
@cindex @file{site-start.el}, the site startup file
  Your site may also have a @dfn{site startup file}; this is named
@file{site-start.el}, if it exists.  Like @file{default.el}, Emacs
finds this file via the standard search path for Lisp libraries.
Emacs loads this library before it loads your init file.  To inhibit
loading of this library, use the option @samp{--no-site-file}.
@xref{Initial Options}.  We recommend against using
@file{site-start.el} for changes that some users may not like.  It is
better to put them in @file{default.el}, so that users can more easily
override them.

@cindex @file{site-lisp} directories
  You can place @file{default.el} and @file{site-start.el} in any of
the directories which Emacs searches for Lisp libraries.  The variable
@code{load-path} (@pxref{Lisp Libraries}) specifies these directories.
Many sites put these files in a subdirectory named @file{site-lisp} in
the Emacs installation directory, such as
@file{/usr/local/share/emacs/site-lisp}.

  Byte-compiling your init file is not recommended (@pxref{Byte
Compilation,, Byte Compilation, elisp, the Emacs Lisp Reference
Manual}).  It generally does not speed up startup very much, and often
leads to problems when you forget to recompile the file.  A better
solution is to use the Emacs server to reduce the number of times you
have to start Emacs (@pxref{Emacs Server}).  If your init file defines
many functions, consider moving them to a separate (byte-compiled)
file that you load in your init file.

  If you are going to write actual Emacs Lisp programs that go beyond
minor customization, you should read the @cite{Emacs Lisp Reference Manual}.
@ifnottex
@xref{Top, Emacs Lisp, Emacs Lisp, elisp, the Emacs Lisp Reference
Manual}.
@end ifnottex

@menu
* Init Syntax::         Syntax of constants in Emacs Lisp.
* Init Examples::       How to do some things with an init file.
* Terminal Init::       Each terminal type can have an init file.
* Find Init::           How Emacs finds the init file.
* Init Non-ASCII::      Using non-@acronym{ASCII} characters in an init file.
* Early Init File::     Another init file, which is read early on.
@end menu

@node Init Syntax
@subsection Init File Syntax

  The init file contains one or more Lisp expressions.  Each of these
consists of a function name followed by arguments, all surrounded by
parentheses.  For example, @code{(setq fill-column 60)} calls the
function @code{setq} to set the variable @code{fill-column}
(@pxref{Filling}) to 60.

  You can set any Lisp variable with @code{setq}, but with certain
variables @code{setq} won't do what you probably want in the
@file{.emacs} file.  Some variables automatically become buffer-local
when set with @code{setq}; what you want in @file{.emacs} is to set
the default value, using @code{setq-default}.  Some customizable minor
mode variables do special things to enable the mode when you set them
with Customize, but ordinary @code{setq} won't do that; to enable the
mode in your @file{.emacs} file, call the minor mode command.  The
following section has examples of both of these methods.

  The second argument to @code{setq} is an expression for the new
value of the variable.  This can be a constant, a variable, or a
function call expression.  In @file{.emacs}, constants are used most
of the time.  They can be:

@table @asis
@item Numbers:
Numbers are written in decimal, with an optional initial minus sign.

@item Strings:
@cindex Lisp string syntax
@cindex string syntax
Lisp string syntax is the same as C string syntax with a few extra
features.  Use a double-quote character to begin and end a string constant.

In a string, you can include newlines and special characters literally.
But often it is cleaner to use backslash sequences for them: @samp{\n}
for newline, @samp{\b} for backspace, @samp{\r} for carriage return,
@samp{\t} for tab, @samp{\f} for formfeed (control-L), @samp{\e} for
escape, @samp{\\} for a backslash, @samp{\"} for a double-quote, or
@samp{\@var{ooo}} for the character whose octal code is @var{ooo}.
Backslash and double-quote are the only characters for which backslash
sequences are mandatory.

@samp{\C-} can be used as a prefix for a control character, as in
@samp{\C-s} for @acronym{ASCII} control-S, and @samp{\M-} can be used as a prefix for
a Meta character, as in @samp{\M-a} for @kbd{@key{Meta}-A} or
@samp{\M-\C-a} for @kbd{@key{Ctrl}-@key{Meta}-A}.

@xref{Init Non-ASCII}, for information about including
non-@acronym{ASCII} in your init file.

@item Characters:
@cindex Lisp character syntax
@cindex character syntax
Lisp character constant syntax consists of a @samp{?} followed by
either a character or an escape sequence starting with @samp{\}.
Examples: @code{?x}, @code{?\n}, @code{?\"}, @code{?\)}.  Note that
strings and characters are not interchangeable in Lisp; some contexts
require one and some contexts require the other.

@xref{Init Non-ASCII}, for information about binding commands to
keys which send non-@acronym{ASCII} characters.

@item True:
@code{t} stands for ``true''.

@item False:
@code{nil} stands for ``false''.

@item Other Lisp objects:
@cindex Lisp object syntax
Write a single-quote (@code{'}) followed by the Lisp object you want.
@end table

@node Init Examples
@subsection Init File Examples

  Here are some examples of doing certain commonly desired things with
Lisp expressions:

@itemize @bullet
@item
Add a directory to the variable @code{load-path}.  You can then put
Lisp libraries that are not included with Emacs in this directory, and
load them with @kbd{M-x load-library}.  @xref{Lisp Libraries}.

@example
(add-to-list 'load-path "/path/to/lisp/libraries")
@end example

@item
Make @key{TAB} in C mode just insert a tab if point is in the middle of a
line.

@example
(setq c-tab-always-indent nil)
@end example

Here we have a variable whose value is normally @code{t} for ``true''
and the alternative is @code{nil} for ``false''.

@item
Make searches case sensitive by default (in all buffers that do not
override this).

@example
(setq-default case-fold-search nil)
@end example

This sets the default value, which is effective in all buffers that do
not have local values for the variable (@pxref{Locals}).  Setting
@code{case-fold-search} with @code{setq} affects only the current
buffer's local value, which is probably not what you want to do in an
init file.

@item
@vindex user-mail-address@r{, in init file}
Specify your own email address, if Emacs can't figure it out correctly.

@example
(setq user-mail-address "cheney@@torture.gov")
@end example

Various Emacs packages, such as Message mode, consult
@code{user-mail-address} when they need to know your email address.
@xref{Mail Headers}.

@item
Make Text mode the default mode for new buffers.

@example
(setq-default major-mode 'text-mode)
@end example

Note that @code{text-mode} is used because it is the command for
entering Text mode.  The single-quote before it makes the symbol a
constant; otherwise, @code{text-mode} would be treated as a variable
name.

@need 1500
@item
Set up defaults for the Latin-1 character set,
which supports most of the languages of Western Europe.

@example
(set-language-environment "Latin-1")
@end example

@need 1500
@item
Turn off Line Number mode, a global minor mode.

@example
(line-number-mode 0)
@end example

@need 1500
@item
Turn on Auto Fill mode automatically in Text mode and related modes
(@pxref{Hooks}).

@example
(add-hook 'text-mode-hook 'auto-fill-mode)
@end example

@item
Load the installed Lisp library named @file{foo} (actually a file
@file{foo.elc} or @file{foo.el} in a standard Emacs directory).

@example
(load "foo")
@end example

When the argument to @code{load} is a relative file name, not starting
with @samp{/} or @samp{~}, @code{load} searches the directories in
@code{load-path} (@pxref{Lisp Libraries}).

@item
Load the compiled Lisp file @file{foo.elc} from your home directory.

@example
(load "~/foo.elc")
@end example

Here a full file name is used, so no searching is done.

@item
@cindex loading Lisp libraries automatically
@cindex autoload Lisp libraries
Tell Emacs to find the definition for the function @code{myfunction}
by loading a Lisp library named @file{mypackage} (i.e., a file
@file{mypackage.elc} or @file{mypackage.el}):

@example
(autoload 'myfunction "mypackage" "Do what I say." t)
@end example

@noindent
Here the string @code{"Do what I say."} is the function's
documentation string.  You specify it in the @code{autoload}
definition so it will be available for help commands even when the
package is not loaded.  The last argument, @code{t}, indicates that
this function is interactive; that is, it can be invoked interactively
by typing @kbd{M-x myfunction @key{RET}} or by binding it to a key.
If the function is not interactive, omit the @code{t} or use
@code{nil}.

@item
Rebind the key @kbd{C-x l} to run the function @code{make-symbolic-link}
(@pxref{Init Rebinding}).

@example
(global-set-key "\C-xl" 'make-symbolic-link)
@end example

or

@example
(define-key global-map "\C-xl" 'make-symbolic-link)
@end example

Note once again the single-quote used to refer to the symbol
@code{make-symbolic-link} instead of its value as a variable.

@item
Do the same thing for Lisp mode only.

@example
(define-key lisp-mode-map "\C-xl" 'make-symbolic-link)
@end example

@item
Redefine all keys which now run @code{next-line} in Fundamental mode
so that they run @code{forward-line} instead.

@findex substitute-key-definition
@example
(substitute-key-definition 'next-line 'forward-line
                           global-map)
@end example

@item
Make @kbd{C-x C-v} undefined.

@example
(global-unset-key "\C-x\C-v")
@end example

One reason to undefine a key is so that you can make it a prefix.
Simply defining @kbd{C-x C-v @var{anything}} will make @kbd{C-x C-v} a
prefix, but @kbd{C-x C-v} must first be freed of its usual non-prefix
definition.

@item
Make @samp{$} have the syntax of punctuation in Text mode.
Note the use of a character constant for @samp{$}.

@example
(modify-syntax-entry ?\$ "." text-mode-syntax-table)
@end example

@item
Enable the use of the command @code{narrow-to-region} without confirmation.

@example
(put 'narrow-to-region 'disabled nil)
@end example

@item
Adjusting the configuration to various platforms and Emacs versions.

Users typically want Emacs to behave the same on all systems, so the
same init file is right for all platforms.  However, sometimes it
happens that a function you use for customizing Emacs is not available
on some platforms or in older Emacs versions.  To deal with that
situation, put the customization inside a conditional that tests whether
the function or facility is available, like this:

@example
(if (fboundp 'blink-cursor-mode)
    (blink-cursor-mode 0))

(if (boundp 'coding-category-utf-8)
    (set-coding-priority '(coding-category-utf-8)))
@end example

@noindent
You can also simply disregard the errors that occur if the
function is not defined.

@example
(ignore-errors (set-face-background 'region "grey75"))
@end example

A @code{setq} on a variable which does not exist is generally
harmless, so those do not need a conditional.
@end itemize

@node Terminal Init
@subsection Terminal-specific Initialization

@vindex term-file-aliases
  Each terminal type can have a Lisp library to be loaded into Emacs when
it is run on that type of terminal.  For a terminal type named
@var{termtype}, the library is called @file{term/@var{termtype}}.
(If there is an entry of the form @code{(@var{termtype} . @var{alias})}
in the @code{term-file-aliases} association list, Emacs uses
@var{alias} in place of @var{termtype}.)  The library is
found by searching the directories @code{load-path} as usual and trying the
suffixes @samp{.elc} and @samp{.el}.  Normally it appears in the
subdirectory @file{term} of the directory where most Emacs libraries are
kept.

  The usual purpose of the terminal-specific library is to map the
escape sequences used by the terminal's function keys onto more
meaningful names, using @code{input-decode-map} (or
@code{function-key-map} before it).  See the file
@file{term/lk201.el} for an example of how this is done.  Many function
keys are mapped automatically according to the information in the
Termcap data base; the terminal-specific library needs to map only the
function keys that Termcap does not specify.

  When the terminal type contains a hyphen, only the part of the name
before the first hyphen is significant in choosing the library name.
Thus, terminal types @samp{aaa-48} and @samp{aaa-30-rv} both use
the library @file{term/aaa}.  The code in the library can use
@code{(getenv "TERM")} to find the full terminal type name.

@vindex term-file-prefix
  The library's name is constructed by concatenating the value of the
variable @code{term-file-prefix} and the terminal type.  Your @file{.emacs}
file can prevent the loading of the terminal-specific library by setting
@code{term-file-prefix} to @code{nil}.

@vindex tty-setup-hook
  Emacs runs the hook @code{tty-setup-hook} at the end of
initialization, after both your @file{.emacs} file and any
terminal-specific library have been read in.  Add hook functions to this
hook if you wish to override part of any of the terminal-specific
libraries and to define initializations for terminals that do not have a
library.  @xref{Hooks}.

@node Find Init
@subsection How Emacs Finds Your Init File

  Normally Emacs uses your home directory to find @file{~/.emacs};
that's what @samp{~} means in a file name.  @xref{General Variables, HOME}.
If neither @file{~/.emacs} nor @file{~/.emacs.el} is found, Emacs looks for
@file{~/.emacs.d/init.el} (which, like @file{~/.emacs.el}, can be
byte-compiled).

  However, if you run Emacs from a shell started by @code{su}, Emacs
tries to find your own @file{.emacs}, not that of the user you are
currently pretending to be.  The idea is that you should get your own
editor customizations even if you are running as the super user.

  More precisely, Emacs first determines which user's init file to use.
It gets your user name from the environment variables @env{LOGNAME} and
@env{USER}; if neither of those exists, it uses effective user-ID@.
If that user name matches the real user-ID, then Emacs uses @env{HOME};
otherwise, it looks up the home directory corresponding to that user
name in the system's data base of users.
@c  LocalWords:  backtab

@node Init Non-ASCII
@subsection Non-@acronym{ASCII} Characters in Init Files
@cindex international characters in @file{.emacs}
@cindex non-@acronym{ASCII} characters in @file{.emacs}
@cindex non-@acronym{ASCII} keys, binding
@cindex rebinding non-@acronym{ASCII} keys

  Language and coding systems may cause problems if your init file
contains non-@acronym{ASCII} characters, such as accented letters, in
strings or key bindings.

  If you want to use non-@acronym{ASCII} characters in your init file,
you should put a @w{@samp{-*-coding: @var{coding-system}-*-}} tag on
the first line of the init file, and specify a coding system that
supports the character(s) in question.  @xref{Recognize Coding}.  This
is because the defaults for decoding non-@acronym{ASCII} text might
not yet be set up by the time Emacs reads those parts of your init
file which use such strings, possibly leading Emacs to decode those
strings incorrectly.  You should then avoid adding Emacs Lisp code
that modifies the coding system in other ways, such as calls to
@code{set-language-environment}.

  To bind non-@acronym{ASCII} keys, you must use a vector (@pxref{Init
Rebinding}).  The string syntax cannot be used, since the
non-@acronym{ASCII} characters will be interpreted as meta keys.  For
instance:

@example
(global-set-key [?@var{char}] 'some-function)
@end example

@noindent
Type @kbd{C-q}, followed by the key you want to bind, to insert @var{char}.

@node Early Init File
@subsection The Early Init File
@cindex early init file

  Most customizations for Emacs should be put in the normal init file,
@file{.emacs} or @file{~/.emacs.d/init.el}.  However, it is sometimes desirable
to have customizations that take effect during Emacs startup earlier than the
normal init file is processed.  Such customizations can be put in the early
init file, @file{~/.emacs.d/early-init.el}.  This file is loaded before the
package system and GUI is initialized, so in it you can customize variables
that affect frame appearance as well as the package initialization process,
such as @code{package-enable-at-startup}, @code{package-load-list}, and
@code{package-user-dir}.  Note that variables like @code{package-archives}
which only affect the installation of new packages, and not the process of
making already-installed packages available, may be customized in the regular
init file.  @xref{Package Installation}.

  We do not recommend that you move into @file{early-init.el}
customizations that can be left in the normal init files.  That is
because the early init file is read before the GUI is initialized, so
customizations related to GUI features will not work reliably in
@file{early-init.el}.  By contrast, the normal init files are read
after the GUI is initialized.  If you must have customizations in the
early init file that rely on GUI features, make them run off hooks
provided by the Emacs startup, such as @code{window-setup-hook} or
@code{tty-setup-hook}.  @xref{Hooks}.

  For more information on the early init file, @pxref{Init File,,,
elisp, The Emacs Lisp Reference Manual}.

@node Authentication
@section Keeping Persistent Authentication Information

  Some Emacs packages, which connect to other services, require
authentication (@pxref{Passwords}), e.g., see @ref{Top, Gnus,, gnus, The
Gnus Manual}, or @ref{Top, Tramp,, tramp, The Tramp Manual}.  Because
it might be annoying to provide the same user name and password again
and again, Emacs offers to keep this information persistent via the
@file{auth-source} library.

@cindex @file{~/.authinfo} file
@cindex @file{~/.authinfo.gpg} file
@cindex ~/.netrc file
  By default, the authentication information is taken from the file
@file{~/.authinfo} or @file{~/.authinfo.gpg} or @file{~/.netrc}.
These files have a syntax similar to netrc files as known from the
@command{ftp} program, like this:

@example
machine @var{mymachine} login @var{myloginname} password @var{mypassword} port @var{myport}
@end example

  Similarly, the @file{auth-source} library supports multiple storage
backend, currently either the classic netrc backend, JSON files, the
Secret Service API, and pass, the standard unix password manager.

@vindex auth-sources
  All these alternatives can be customized via the user option
@code{auth-sources}, see @ref{Help for users, Emacs auth-source,,
auth, Emacs auth-source}.

@vindex auth-source-save-behavior
  When a password is entered interactively, which is not found via the
configured backend, some of the backends offer to save it
persistently.  This can be changed by customizing the user option
@code{auth-source-save-behavior}.