When using the X Window System, you can create multiple windows at the X level in a single Emacs session. Each X window that belongs to Emacs displays a frame which can contain one or several Emacs windows. A frame initially contains a single general-purpose Emacs window which you can subdivide vertically or horizontally into smaller windows. A frame normally contains its own echo area and minibuffer, but you can make frames that don't have these--they use the echo area and minibuffer of another frame.
Editing you do in one frame also affects the other frames. For instance, if you put text in the kill ring in one frame, you can yank it in another frame. If you exit Emacs through C-x C-c in one frame, it terminates all the frames. To delete just one frame, use C-x 5 0.
To avoid confusion, we reserve the word "window" for the subdivisions that Emacs implements, and never use it to refer to a frame.
Emacs compiled for MS-DOS emulates some aspects of the window system so that you can use many of the features described in this chapter. See section Keyboard and Mouse on MS-DOS, for more information.
The mouse commands for selecting and copying a region are mostly
compatible with the
xterm program. You can use the same mouse
commands for copying between Emacs and other X client programs.
If you select a region with any of these mouse commands, and then immediately afterward type the DELETE function key, it deletes the region that you selected. The BACKSPACE function key and the ASCII character DEL do not do this; if you type any other key in between the mouse command and DELETE, it does not do this.
mouse-set-point). This is normally the left button.
mouse-set-region). You can specify both ends of the region with this single command. If you move the mouse off the top or bottom of the window while dragging, the window scrolls at a steady rate until you move the mouse back into the window. This way, you can select regions that don't fit entirely on the screen. The number of lines scrolled per step depends on how far away from the window edge the mouse has gone; the variable
mouse-scroll-min-linesspecifies a minimum step size.
mouse-yank-at-click). This is normally the middle button.
mouse-save-then-kill, has several functions depending on where you click and the status of the region. The most basic case is when you click Mouse-1 in one place and then Mouse-3 in another. This selects the text between those two positions as the region. It also copies the new region to the kill ring, so that you can copy it to someplace else. If you click Mouse-1 in the text, scroll with the scroll bar, and then click Mouse-3, it remembers where point was before scrolling (where you put it with Mouse-1), and uses that position as the other end of the region. This is so that you can select a region that doesn't fit entirely on the screen. More generally, if you do not have a highlighted region, Mouse-3 selects the text between point and the click position as the region. It does this by setting the mark where point was, and moving point to where you click. If you have a highlighted region, or if the region was set just before by dragging button 1, Mouse-3 adjusts the nearer end of the region by moving it to where you click. The adjusted region's text also replaces the old region's text in the kill ring. If you originally specified the region using a double or triple Mouse-1, so that the region is defined to consist of entire words or lines, then adjusting the region with Mouse-3 also proceeds by entire words or lines. If you use Mouse-3 a second time consecutively, at the same place, that kills the region already selected.
The simplest way to kill text with the mouse is to press Mouse-1 at one end, then press Mouse-3 twice at the other end. See section Deletion and Killing. To copy the text into the kill ring without deleting it from the buffer, press Mouse-3 just once--or just drag across the text with Mouse-1. Then you can copy it elsewhere by yanking it.
To yank the killed or copied text somewhere else, move the mouse there
and press Mouse-2. See section Yanking. However, if
mouse-yank-at-point is non-
nil, Mouse-2 yanks at
point. Then it does not matter where you click, or even which of the
frame's windows you click on. The default value is
variable also affects yanking the secondary selection.
To copy text to another X window, kill it or save it in the kill ring. Under X, this also sets the primary selection. Then use the "paste" or "yank" command of the program operating the other window to insert the text from the selection.
To copy text from another X window, use the "cut" or "copy" command of the program operating the other window, to select the text you want. Then yank it in Emacs with C-y or Mouse-2.
These cutting and pasting commands also work on MS-Windows.
When Emacs puts text into the kill ring, or rotates text to the front
of the kill ring, it sets the primary selection in the X server.
This is how other X clients can access the text. Emacs also stores the
text in the cut buffer, but only if the text is short enough
x-cut-buffer-max specifies the maximum number of characters);
putting long strings in the cut buffer can be slow.
The commands to yank the first entry in the kill ring actually check first for a primary selection in another program; after that, they check for text in the cut buffer. If neither of those sources provides text to yank, the kill ring contents are used.
The secondary selection is another way of selecting text using X. It does not use point or the mark, so you can use it to kill text without setting point or the mark.
mouse-set-secondary). The highlighting appears and changes as you drag. If you move the mouse off the top or bottom of the window while dragging, the window scrolls at a steady rate until you move the mouse back into the window. This way, you can mark regions that don't fit entirely on the screen.
mouse-secondary-save-then-kill). A second click at the same place kills the secondary selection just made.
mouse-yank-secondary). This places point at the end of the yanked text.
Double or triple clicking of M-Mouse-1 operates on words and lines, much like Mouse-1.
mouse-yank-at-point is non-
yanks at point. Then it does not matter precisely where you click; all
that matters is which window you click on. See section Mouse Commands for Editing.
Some Emacs buffers display lists of various sorts. These include lists of files, of buffers, of possible completions, of matches for a pattern, and so on.
Since yanking text into these buffers is not very useful, most of them define Mouse-2 specially, as a command to use or view the item you click on.
For example, if you click Mouse-2 on a file name in a Dired buffer, you visit that file. If you click Mouse-2 on an error message in the `*Compilation*' buffer, you go to the source code for that error message. If you click Mouse-2 on a completion in the `*Completions*' buffer, you choose that completion.
You can usually tell when Mouse-2 has this special sort of meaning because the sensitive text highlights when you move the mouse over it.
Mouse clicks modified with the CTRL and SHIFT keys bring up menus.
You can use mouse clicks on window mode lines to select and manipulate windows.
C-Mouse-2 on a scroll bar splits the corresponding window vertically. See section Splitting Windows.
The prefix key C-x 5 is analogous to C-x 4, with parallel subcommands. The difference is that C-x 5 commands create a new frame rather than just a new window in the selected frame (See section Displaying in Another Window). If an existing visible or iconified frame already displays the requested material, these commands use the existing frame, after raising or deiconifying as necessary.
The various C-x 5 commands differ in how they find or create the buffer to select:
find-file-other-frame. See section Visiting Files.
dired-other-frame. See section Dired, the Directory Editor.
mail-other-frame. It is the other-frame variant of C-x m. See section Sending Mail.
find-tag-other-frame, the multiple-frame variant of M-.. See section Tags Tables.
find-file-read-only-other-frame. See section Visiting Files.
You can control the appearance of new frames you create by setting the
frame parameters in
default-frame-alist. You can use the
initial-frame-alist to specify parameters that affect
only the initial frame. See section `Initial Parameters' in The Emacs Lisp Reference Manual, for more information.
An Emacs frame can have a speedbar, which is a vertical window that serves as a scrollable menu of files you could visit and tags within those files. To create a speedbar, type M-x speedbar; this creates a speedbar window for the selected frame. From then on, you can click on a file name in the speedbar to visit that file in the corresponding Emacs frame, or click on a tag name to jump to that tag in the Emacs frame.
Initially the speedbar lists the immediate contents of the current directory, one file per line. Each line also has a box, `[+]' or `<+>', that you can click on with Mouse-2 to "open up" the contents of that item. If the line names a directory, opening it adds the contents of that directory to the speedbar display, underneath the directory's own line. If the line lists an ordinary file, opening it up adds a list of the tags in that file to the speedbar display. When a file is opened up, the `[+]' changes to `[-]'; you can click on that box to "close up" that file (hide its contents).
Some major modes, including Rmail mode, Info, and GUD, have specialized ways of putting useful items into the speedbar for you to select. For example, in Rmail mode, the speedbar shows a list of Rmail files, and lets you move the current message to another Rmail file by clicking on its `<M>' box.
A speedbar belongs to one Emacs frame, and always operates on that frame. If you use multiple frames, you can make a speedbar for some or all of the frames; type M-x speedbar in any given frame to make a speedbar for it.
A single Emacs can talk to more than one X Windows display.
Initially, Emacs uses just one display--the one specified with the
DISPLAY environment variable or with the `--display' option
(see section Initial Options). To connect to another display, use the
A single X server can handle more than one screen. When you open frames on two screens belonging to one server, Emacs knows they share a single keyboard, and it treats all the commands arriving from these screens as a single stream of input.
When you open frames on different X servers, Emacs makes a separate input stream for each server. This way, two users can type simultaneously on the two displays, and Emacs will not garble their input. Each server also has its own selected frame. The commands you enter with a particular X server apply to that server's selected frame.
Despite these features, people using the same Emacs job from different displays can still interfere with each other if they are not careful. For example, if any one types C-x C-c, that exits the Emacs job for all of them!
You can make certain chosen buffers, for which Emacs normally creates
a second window when you have just one window, appear in special frames
of their own. To do this, set the variable
special-display-buffer-names to a list of buffer names; any
buffer whose name is in that list automatically gets a special frame,
when an Emacs command wants to display it "in another window."
For example, if you set the variable this way,
(setq special-display-buffer-names '("*Completions*" "*grep*" "*tex-shell*"))
then completion lists,
grep output and the TeX mode shell
buffer get individual frames of their own. These frames, and the
windows in them, are never automatically split or reused for any other
buffers. They continue to show the buffers they were created for,
unless you alter them by hand. Killing the special buffer deletes its
More generally, you can set
special-display-regexps to a list
of regular expressions; then a buffer gets its own frame if its name
matches any of those regular expressions. (Once again, this applies only
to buffers that normally get displayed for you in a separate window.)
special-display-frame-alist specifies the frame
parameters for these frames. It has a default value, so you don't need
to set it.
For those who know Lisp, an element of
can also be a list. Then the first element is the buffer name or
regular expression; the rest of the list specifies how to create the
frame. It can be an association list specifying frame parameter values;
these values take precedence over parameter values specified in
special-display-frame-alist. Alternatively, it can have this
where function is a symbol. Then the frame is constructed by calling function; its first argument is the buffer, and its remaining arguments are args.
An analogous feature lets you specify buffers which should be
displayed in the selected window. See section Forcing Display in the Same Window. The
same-window feature takes precedence over the special-frame feature;
therefore, if you add a buffer name to
special-display-buffer-names and it has no effect, check to see
whether that feature is also in use for the same buffer name.
This section describes commands for altering the display style and window management behavior of the selected frame.
modelineface also, so that it remains in inverse video compared with the default.
auto-raise-modehas no effect on it.
auto-lower-modehas no effect on auto-lower implemented by the X window manager. To control that, you must use the appropriate window manager features.
In Emacs versions that use an X toolkit, the color-setting and font-setting functions don't affect menus and the menu bar, since they are displayed by their own widget classes. To change the appearance of the menus and menu bar, you must use X resources (see section X Resources). See section Window Color Options, regarding colors. See section Font Specification Options, regarding choice of font.
For information on frame parameters and customization, see section `Frame Parameters' in The Emacs Lisp Reference Manual.
When using X, Emacs normally makes a scroll bar at the left of each Emacs window. The scroll bar runs the height of the window, and shows a moving rectangular inner box which represents the portion of the buffer currently displayed. The entire height of the scroll bar represents the entire length of the buffer.
You can use Mouse-2 (normally, the middle button) in the scroll bar to move or drag the inner box up and down. If you move it to the top of the scroll bar, you see the top of the buffer. If you move it to the bottom of the scroll bar, you see the bottom of the buffer.
The left and right buttons in the scroll bar scroll by controlled increments. Mouse-1 (normally, the left button) moves the line at the level where you click up to the top of the window. Mouse-3 (normally, the right button) moves the line at the top of the window down to the level where you click. By clicking repeatedly in the same place, you can scroll by the same distance over and over.
Aside from scrolling, you can also click C-Mouse-2 in the scroll bar to split a window vertically. The split occurs on the line where you click.
You can enable or disable Scroll Bar mode with the command M-x scroll-bar-mode. With no argument, it toggles the use of scroll bars. With an argument, it turns use of scroll bars on if and only if the argument is positive. This command applies to all frames, including frames yet to be created. You can use the X resource `verticalScrollBars' to control the initial setting of Scroll Bar mode. See section X Resources.
To enable or disable scroll bars for just the selected frame, use the M-x toggle-scroll-bar command.
You can turn display of menu bars on or off with M-x menu-bar-mode. With no argument, this command toggles Menu Bar mode, a minor mode. With an argument, the command turns Menu Bar mode on if the argument is positive, off if the argument is not positive. You can use the X resource `menuBarLines' to control the initial setting of Menu Bar mode. See section X Resources. Expert users often turn off the menu bar, especially on text-only terminals, where this makes one additional line available for text.
See section The Menu Bar for information on how to invoke commands with the menu bar.
When using Emacs with X, you can set up multiple styles of displaying characters. The aspects of style that you can control are the type font, the foreground color, the background color, and whether to underline. Emacs on MS-DOS supports faces partially by letting you control the foreground and background colors of each face (see section Emacs and MS-DOS).
The way you control display style is by defining named faces. Each face can specify a type font, a foreground color, a background color, and an underline flag; but it does not have to specify all of them.
The style of display used for a given character in the text is determined by combining several faces. Any aspect of the display style that isn't specified by overlays or text properties comes from the frame itself.
Enriched mode, the mode for editing formatted text, includes several commands and menus for specifying faces. See section Faces in Formatted Text, for how to specify the font for text in the buffer. See section Colors in Formatted Text, for how to specify the foreground and background color.
To alter the appearance of a face, use the customization buffer. See section Customizing Faces. You can also use X resources to specify attributes of particular faces (see section X Resources).
To see what faces are currently defined, and what they look like, type M-x list-faces-display. It's possible for a given face to look different in different frames; this command shows the appearance in the frame in which you type it. Here's a list of the standardly defined faces:
When Transient Mark mode is enabled, the text of the region is
highlighted when the mark is active. This uses the face named
region; you can control the style of highlighting by changing the
style of this face (see section Customizing Faces). See section Transient Mark Mode,
for more information about Transient Mark mode and activation and
deactivation of the mark.
One easy way to use faces is to turn on Font Lock mode. This minor mode, which is always local to a particular buffer, arranges to choose faces according to the syntax of the text you are editing. It can recognize comments and strings in most languages; in several languages, it can also recognize and properly highlight various other important constructs. See section Font Lock mode, for more information about Font Lock mode and syntactic highlighting.
You can print out the buffer with the highlighting that appears
on your screen using the command
See section Postscript Hardcopy.
Font Lock mode is a minor mode, always local to a particular buffer, which highlights (or "fontifies") using various faces according to the syntax of the text you are editing. It can recognize comments and strings in most languages; in several languages, it can also recognize and properly highlight various other important constructs--for example, names of functions being defined or reserved keywords.
The command M-x font-lock-mode turns Font Lock mode on or off
according to the argument, and toggles the mode when it has no argument.
turn-on-font-lock unconditionally enables Font Lock
mode. This is useful in mode-hook functions. For example, to enable
Font Lock mode whenever you edit a C file, you can do this:
(add-hook 'c-mode-hook 'turn-on-font-lock)
To turn on Font Lock mode automatically in all modes which support it,
use the function
global-font-lock-mode, like this:
In Font Lock mode, when you edit the text, the highlighting updates
automatically in the line that you changed. Most changes don't affect
the highlighting of subsequent lines, but occasionally they do. To
rehighlight a range of lines, use the command M-g M-g
In certain major modes, M-g M-g refontifies the entire current
function. (The variable
how to find the current function.) In other major modes, M-g M-g
refontifies 16 lines above and below point.
With a prefix argument n, M-g M-g refontifies n lines above and below point, regardless of the mode.
To get the full benefit of Font Lock mode, you need to choose a default font which has bold, italic, and bold-italic variants; or else you need to have a color or gray-scale screen.
font-lock-maximum-decoration specifies the
preferred level of fontification, for modes that provide multiple
levels. Level 1 is the least amount of fontification; some modes
support levels as high as 3. The normal default is "as high as
possible." You can specify an integer, which applies to all modes, or
you can specify different numbers for particular major modes; for
example, to use level 1 for C/C++ modes, and the default level
otherwise, use this:
(setq font-lock-maximum-decoration '((c-mode . 1) (c++-mode . 1)))
Fontification can be too slow for large buffers, so you can suppress
it. The variable
font-lock-maximum-size specifies a buffer size,
beyond which buffer fontification is suppressed.
Comment and string fontification (or "syntactic" fontification) relies on analysis of the syntactic structure of the buffer text. For the purposes of speed, some modes including C mode and Lisp mode rely on a special convention: an open-parenthesis in the leftmost column always defines the beginning of a defun, and is thus always outside any string or comment. (See section Defuns.) If you don't follow this convention, then Font Lock mode can misfontify the text after an open-parenthesis in the leftmost column that is inside a string or comment.
buffer-local) specifies how Font Lock mode can find a position
guaranteed to be outside any comment or string. In modes which use the
leftmost column parenthesis convention, the default value of the variable
beginning-of-defun---that tells Font Lock mode to use the
convention. If you set this variable to
nil, Font Lock no longer
relies on the convention. This avoids incorrect results, but the price
is that, in some cases, fontification for a changed text must rescan
buffer text from the beginning of the buffer.
Font Lock highlighting patterns already exist for many modes, but you
may want to fontify additional patterns. You can use the function
font-lock-add-keywords, to add your own highlighting patterns for
a particular mode. For example, to highlight `FIXME:' words in C
comments, use this:
(font-lock-add-keywords 'c-mode '(("\\<\\(FIXME\\):" 1 font-lock-warning-face t)))
Font Lock support modes make Font Lock mode faster for large buffers. There are two support modes: Fast Lock mode and Lazy Lock mode. They use two different methods of speeding up Font Lock mode.
To make Font Lock mode faster for buffers visiting large files, you can use Fast Lock mode. Fast Lock mode saves the font information for each file in a separate cache file; each time you visit the file, it rereads the font information from the cache file instead of refontifying the text from scratch.
The command M-x fast-lock-mode turns Fast Lock mode on or off, according to the argument (with no argument, it toggles). You can also arrange to enable Fast Lock mode whenever you use Font Lock mode, like this:
(setq font-lock-support-mode 'fast-lock-mode)
It is not worth writing a cache file for small buffers. Therefore,
fast-lock-minimum-size specifies a minimum file size
for caching font information.
fast-lock-cache-directories specifies where to put
the cache files. Its value is a list of directories to try;
means the same directory as the file being edited. The default value is
("." "~/.emacs-flc"), which means to use the same directory if
possible, and otherwise the directory `~/.emacs-flc'.
fast-lock-save-others specifies whether Fast Lock
mode should save cache files for files that you do not own. A
nil value means yes (and that is the default).
To make Font Lock mode faster for large buffers, you can use Lazy Lock mode to reduce the amount of text that is fontified. In Lazy Lock mode, buffer fontification is demand-driven; it happens to portions of the buffer that are about to be displayed. And fontification of your changes is deferred; it happens only when Emacs has been idle for a certain short period of time.
The command M-x lazy-lock-mode turns Lazy Lock mode on or off, according to the argument (with no argument, it toggles). You can also arrange to enable Lazy Lock mode whenever you use Font Lock mode, like this:
(setq font-lock-support-mode 'lazy-lock-mode)
It is not worth avoiding buffer fontification for small buffers.
Therefore, the variable
lazy-lock-minimum-size specifies a
minimum buffer size for demand-driven buffer fontification. Buffers
smaller than that are fontified all at once, as in plain Font Lock mode.
When you alter the buffer, Lazy Lock mode defers fontification of the
text you changed. The variable
how many seconds Emacs must be idle before it starts fontifying your
changes. If the value is 0, then changes are fontified immediately, as
in plain Font Lock mode.
Lazy Lock mode normally fontifies newly visible portions of the buffer
before they are first displayed. However, if the value of
lazy-lock-defer-on-scrolling is non-
nil, newly visible
text is fontified only when Emacs is idle for
In some modes, including C mode and Emacs Lisp mode, changes in one
line's contents can alter the context for subsequent lines, and thus
change how they ought to be fontified. Ordinarily, you must type
M-g M-g to refontify the subsequent lines. However, if you set
lazy-lock-defer-contextually to non-
Lock mode does this automatically, after
When Emacs is idle for a long time, Lazy Lock fontifies additional portions of the buffer, not yet displayed, in case you will display them later. This is called stealth fontification.
lazy-lock-stealth-time specifies how many seconds
Emacs has to be idle before stealth fontification starts. A value of
nil means no stealth fontification. The variables
specify the granularity and verbosity of stealth fontification.
Here is a simple guide to help you choose one of the Font Lock support modes.
font-lock-support-mode specifies which of these
support modes to use; for example, to specify that Fast Lock mode is
used for C/C++ modes, and Lazy Lock mode otherwise, set the variable
(setq font-lock-support-mode '((c-mode . fast-lock-mode) (c++-mode . fast-lock-mode) (t . lazy-lock-mode)))
Use M-x highlight-changes-mode to enable a minor mode that uses faces (colors, typically) to indicate which parts of the buffer were changed most recently.
The following commands let you create, delete and operate on frames:
iconify-or-deiconify-frame). The normal meaning of C-z, to suspend Emacs, is not useful under a window system, so it has a different binding in that case. If you type this command on an Emacs frame's icon, it deiconifies the frame.
delete-frame). This is not allowed if there is only one frame.
If your terminal does not have a window system that Emacs supports, then it can display only one Emacs frame at a time. However, you can still create multiple Emacs frames, and switch between them. Switching frames on these terminals is much like switching between different window configurations.
Use C-x 5 2 to create a new frame and switch to it; use C-x 5 o to cycle through the existing frames; use C-x 5 0 to delete the current frame.
Each frame has a number to distinguish it. If your terminal can display only one frame at a time, the selected frame's number n appears near the beginning of the mode line, in the form `Fn'.
`Fn' is actually the frame's name. You can also specify a different name if you wish, and you can select a frame by its name. Use the command M-x set-frame-name RET name RET to specify a new name for the selected frame, and use M-x select-frame-by-name RET name RET to select a frame according to its name. The name you specify appears in the mode line when the frame is selected.