A frame is a rectangle on the screen that contains one or more Emacs windows. A frame initially contains a single main window (plus perhaps a minibuffer window), which you can subdivide vertically or horizontally into smaller windows.
When Emacs communicates directly with a supported window system, such as X Windows, it does not have a terminal frame; instead, it starts with a single window frame, but you can create more, and Emacs can display several such frames at once as is usual for window systems.
See section Emacs Display, for information about the related topic of controlling Emacs redisplay.
To create a new frame, call the function
The argument is an alist specifying frame parameters. Any parameters
not mentioned in alist default according to the value of the
default-frame-alist; parameters not specified even there
default from the standard X resources or whatever is used instead on
The set of possible parameters depends in principle on what kind of window system Emacs uses to display its frames. See section Window Frame Parameters, for documentation of individual parameters you can specify.
make-framebefore it actually creates the frame.
make-frameafter it creates the frame. Each function in
after-make-frame-hookreceives one argument, the frame just created.
A single Emacs can talk to more than one X display.
Initially, Emacs uses just one display--the one chosen with the
DISPLAY environment variable or with the `--display' option
(see section `Initial Options' in The GNU Emacs Manual). To connect to
another display, use the command
make-frame-on-display or specify
display frame parameter when you create the frame.
Emacs treats each X server as a separate terminal, giving each one its own selected frame and its own minibuffer windows.
A few Lisp variables are terminal-local; that is, they have a
separate binding for each terminal. The binding in effect at any time
is the one for the terminal that the currently selected frame belongs
to. These variables include
system-key-alist. They are always terminal-local, and can never
be buffer-local (see section Buffer-Local Variables) or frame-local.
A single X server can handle more than one screen. A display name `host:server.screen' has three parts; the last part specifies the screen number for a given server. When you use two screens belonging to one server, Emacs knows by the similarity in their names that they share a single keyboard, and it treats them as a single terminal.
make-frame(see section Creating Frames).
The optional argument xrm-string, if not
nil, is a
string of resource names and values, in the same format used in the
`.Xresources' file. The values you specify override the resource
values recorded in the X server itself; they apply to all Emacs frames
created on this display. Here's an example of what this string might
"*BorderWidth: 3\n*InternalBorder: 2\n"
See section X Resources.
A frame has many parameters that control its appearance and behavior. Just what parameters a frame has depends on what display mechanism it uses.
Frame parameters exist for the sake of window systems. A terminal frame
has a few parameters, mostly for compatibility's sake; only the
buffer-predicate parameters do something special.
These functions let you read and change the parameter values of a frame.
frame-parametersreturns an alist listing all the parameters of frame and their values.
(parm . value), where parm is a symbol naming a parameter. If you don't mention a parameter in alist, its value doesn't change.
You can specify the parameters for the initial startup frame
initial-frame-alist in your `.emacs' file.
(parameter . value)
Emacs creates the initial frame before it reads your `~/.emacs'
file. After reading that file, Emacs checks
and applies the parameter settings in the altered value to the already
created initial frame.
If these settings affect the frame geometry and appearance, you'll see the frame appear with the wrong ones and then change to the specified ones. If that bothers you, you can specify the same geometry and appearance with X resources; those do take affect before the frame is created. See section `X Resources' in The GNU Emacs Manual.
X resource settings typically apply to all frames. If you want to
specify some X resources solely for the sake of the initial frame, and
you don't want them to apply to subsequent frames, here's how to achieve
this. Specify parameters in
default-frame-alist to override the
X resources for subsequent frames; then, to prevent these from affecting
the initial frame, specify the same parameters in
initial-frame-alist with values that match the X resources.
If these parameters specify a separate minibuffer-only frame with
(minibuffer . nil), and you have not created one, Emacs creates
one for you.
special-display-frame-alist, in section Choosing a Window for Display.
If you use options that specify window appearance when you invoke Emacs,
they take effect by adding elements to
exception is `-geometry', which adds the specified position to
initial-frame-alist instead. See section `Command Arguments' in The GNU Emacs Manual.
Just what parameters a frame has depends on what display mechanism it
uses. Here is a table of the parameters that have special meanings in a
window frame; of these,
meaningful information in terminal frames.
"host:dpy.screen", just like the
niltitle, it appears in the window system's border for the frame, and also in the mode line of windows in that frame if
mode-line-frame-identificationuses `%F' (see section
%-Constructs in the Mode Line). This is normally the case when Emacs is not using a window system, and can only display one frame at a time. See section Frame Titles.
titleparameter is unspecified or
nil. If you don't specify a name, Emacs sets the frame name automatically (see section Frame Titles). If you specify the frame name explicitly when you create the frame, the name is also used (instead of the name of the Emacs executable) when looking up X resources for the frame.
(+ pos)which permits specifying a negative pos value. A negative number -pos, or a list of the form
(- pos), actually specifies the position of the right edge of the window with respect to the right edge of the screen. A positive value of pos counts toward the left. Reminder: if the parameter is a negative integer -pos, then pos is positive. Some window managers ignore program-specified positions. If you want to be sure the position you specify is not ignored, specify a non-
nilvalue for the
user-positionparameter as well.
(+ pos)which permits specifying a negative pos value. A negative number -pos, or a list of the form
(- pos), actually specifies the position of the bottom edge of the window with respect to the bottom edge of the screen. A positive value of pos counts toward the top. Reminder: if the parameter is a negative integer -pos, then pos is positive. Some window managers ignore program-specified positions. If you want to be sure the position you specify is not ignored, specify a non-
nilvalue for the
user-positionparameter as well.
topparameters, use this parameter to say whether the specified position was user-specified (explicitly requested in some way by a human user) or merely program-specified (chosen by a program). A non-
nilvalue says the position was user-specified. Window managers generally heed user-specified positions, and some heed program-specified positions too. But many ignore program-specified positions, placing the window in a default fashion or letting the user place it with the mouse. Some window managers, including
twm, let the user specify whether to obey program-specified positions or ignore them. When you call
make-frame, you should specify a non-
nilvalue for this parameter if the values of the
topparameters represent the user's stated preference; otherwise, use
frame-pixel-height; see section Frame Size And Position.)
frame-pixel-width; see section Frame Size And Position.)
onlymeans this frame is just a minibuffer. If the value is a minibuffer window (in some other frame), the new frame uses that minibuffer.
other-bufferuses this predicate (from the selected frame) to decide which buffers it should consider, if the predicate is not
nil. It calls the predicate with one argument, a buffer, once for each buffer; if the predicate returns a non-
nilvalue, it considers that buffer.
nilfor no scroll bars.
nilmeans yes). (Horizontal scroll bars are not currently implemented.)
nilvalue specifies the default bitmap icon (a picture of a gnu);
nilspecifies a text icon.
nil, the frame's title is used.
foreground-colorframe parameter, you should call
frame-update-face-colorsto update faces accordingly.
background-colorframe parameter, you should call
frame-update-face-colorsto update faces accordingly. See section Functions for Working with Faces.
light, according to whether the background color is a light one or a dark one.
(bar . width). The symbol
boxspecifies an ordinary black box overlaying the character after point; that is the default. The symbol
barspecifies a vertical bar between characters as the cursor.
(bar . width)specifies a bar width pixels wide.
nil, this frame's window is never split automatically.
tfor visible, and
iconfor iconified. See section Visibility of Frames.
You can read or change the size and position of a frame using the
width. Whatever geometry parameters you don't specify are chosen
by the window manager in its usual fashion.
Here are some special features for working with sizes and positions:
Negative parameter values position the bottom edge of the window up from the bottom edge of the screen, or the right window edge to the left of the right edge of the screen. It would probably be better if the values were always counted from the left and top, so that negative arguments would position the frame partly off the top or left edge of the screen, but it seems inadvisable to change that now.
frame-width. When you are using a non-window terminal, the size of the frame is normally the same as the size of the terminal screen.
To set the size based on values measured in pixels, use
frame-char-width to convert
them to units of characters.
If pretend is non-
nil, then Emacs displays lines
lines of output in frame, but does not change its value for the
actual height of the frame. This is only useful for a terminal frame.
Using a smaller height than the terminal actually implements may be
useful to reproduce behavior observed on a smaller screen, or if the
terminal malfunctions when using its whole screen. Setting the frame
height "for real" does not always work, because knowing the correct
actual size may be necessary for correct cursor positioning on a
The older functions
set-screen-width were used to specify the height and width of the
screen, in Emacs versions that did not support multiple frames. They
are semi-obsolete, but still work; they apply to the selected frame.
x-parse-geometryconverts a standard X window geometry string to an alist that you can use as part of the argument to
The alist describes which parameters were specified in geom, and
gives the values specified for them. Each element looks like
(parameter . value). The possible parameter
For the size parameters, the value must be an integer. The position
top are not totally accurate,
because some values indicate the position of the right or bottom edges
instead. These are the value possibilities for the position
Here is an example:
(x-parse-geometry "35x70+0-0") => ((height . 70) (width . 35) (top - 0) (left . 0))
Every frame has a
name parameter; this serves as the default
for the frame title which window systems typically display at the top of
the frame. You can specify a name explicitly by setting the
Normally you don't specify the name explicitly, and Emacs computes the
frame name automatically based on a template stored in the variable
frame-title-format. Emacs recomputes the name each time the
frame is redisplayed.
mode-line-format. See section The Data Structure of the Mode Line.
twhen there are two or more frames (not counting minibuffer-only frames or invisible frames). The default value of
multiple-framesso as to put the buffer name in the frame title only when there is more than one frame.
Frames remain potentially visible until you explicitly delete them. A deleted frame cannot appear on the screen, but continues to exist as a Lisp object until there are no references to it. There is no way to cancel the deletion of a frame aside from restoring a saved frame configuration (see section Frame Configurations); this is similar to the way windows behave.
nilif the frame frame has not been deleted.
Some window managers provide a command to delete a window. These work
by sending a special message to the program that operates the window.
When Emacs gets one of these commands, it generates a
delete-frame event, whose normal definition is a command that
calls the function
delete-frame. See section Miscellaneous Window System Events.
frame-listreturns a list of all the frames that have not been deleted. It is analogous to
buffer-listfor buffers. The list that you get is newly created, so modifying the list doesn't have any effect on the internals of Emacs.
next-framelets you cycle conveniently through all the frames from an arbitrary starting point. It returns the "next" frame after frame in the cycle. If frame is omitted or
nil, it defaults to the selected frame.
The second argument, minibuf, says which frames to consider:
next-frame, but cycles through all frames in the opposite direction.
previous-window, in section Cyclic Ordering of Windows.
Each window is part of one and only one frame; you can get the frame
All the non-minibuffer windows in a frame are arranged in a cyclic order. The order runs from the frame's top window, which is at the upper left corner, down and to the right, until it reaches the window at the lower right corner (always the minibuffer window, if the frame has one), and then it moves back to the top. See section Cyclic Ordering of Windows.
At any time, exactly one window on any frame is selected within the
frame. The significance of this designation is that selecting the
frame also selects this window. You can get the frame's current
selected window with
Conversely, selecting a window for Emacs with
makes that window selected within its frame. See section Selecting Windows.
Another function that (usually) returns one of the windows in a given
minibuffer-window. See section Minibuffer Miscellany.
Normally, each frame has its own minibuffer window at the bottom, which
is used whenever that frame is selected. If the frame has a minibuffer,
you can get it with
minibuffer-window (see section Minibuffer Miscellany).
However, you can also create a frame with no minibuffer. Such a frame
must use the minibuffer window of some other frame. When you create the
frame, you can specify explicitly the minibuffer window to use (in some
other frame). If you don't, then the minibuffer is found in the frame
which is the value of the variable
value should be a frame that does have a minibuffer.
If you use a minibuffer-only frame, you might want that frame to raise
when you enter the minibuffer. If so, set the variable
t. See section Raising and Lowering Frames.
At any time, one frame in Emacs is the selected frame. The selected window always resides on the selected frame.
Some window systems and window managers direct keyboard input to the window object that the mouse is in; others require explicit clicks or commands to shift the focus to various window objects. Either way, Emacs automatically keeps track of which frame has the focus.
Lisp programs can also switch frames "temporarily" by calling the
select-frame. This does not alter the window system's
concept of focus; rather, it escapes from the window manager's control
until that control is somehow reasserted.
When using a text-only terminal, only the selected terminal frame is
actually displayed on the terminal.
switch-frame is the only way
to switch frames, and the change lasts until overridden by a subsequent
switch-frame. Each terminal screen except for the
initial one has a number, and the number of the selected frame appears
in the mode line before the buffer name (see section Variables Used in the Mode Line).
Emacs cooperates with the window system by arranging to select frames as
the server and window manager request. It does so by generating a
special kind of input event, called a focus event, when
appropriate. The command loop handles a focus event by calling
handle-switch-frame. See section Focus Events.
Focus events normally do their job by invoking this command. Don't call it for any other reason.
last-event-framewill be focus-frame. Also, switch-frame events specifying frame will instead select focus-frame.
If focus-frame is
nil, that cancels any existing
redirection for frame, which therefore once again receives its own
One use of focus redirection is for frames that don't have minibuffers. These frames use minibuffers on other frames. Activating a minibuffer on another frame redirects focus to that frame. This puts the focus on the minibuffer's frame, where it belongs, even though the mouse remains in the frame that activated the minibuffer.
Selecting a frame can also change focus redirections. Selecting frame
foo had been selected, changes any redirections
foo so that they point to
bar instead. This
allows focus redirection to work properly when the user switches from
one frame to another using
This means that a frame whose focus is redirected to itself is treated
differently from a frame whose focus is not redirected.
select-frame affects the former but not the latter.
The redirection lasts until
redirect-frame-focus is called to
nilsays that it does. When this is so, the command
other-framemoves the mouse to a position consistent with the new selected frame.
A window frame may be visible, invisible, or iconified. If it is visible, you can see its contents. If it is iconified, the frame's contents do not appear on the screen, but an icon does. If the frame is invisible, it doesn't show on the screen, not even as an icon.
Visibility is meaningless for terminal frames, since only the selected one is actually displayed in any case.
tif frame is visible,
nilif it is invisible, and
iconif it is iconified.
The visibility status of a frame is also available as a frame parameter. You can read or change it as such. See section Window Frame Parameters.
The user can iconify and deiconify frames with the window manager. This happens below the level at which Emacs can exert any control, but Emacs does provide events that you can use to keep track of such changes. See section Miscellaneous Window System Events.
Most window systems use a desktop metaphor. Part of this metaphor is the idea that windows are stacked in a notional third dimension perpendicular to the screen surface, and thus ordered from "highest" to "lowest". Where two windows overlap, the one higher up covers the one underneath. Even a window at the bottom of the stack can be seen if no other window overlaps it.
A window's place in this ordering is not fixed; in fact, users tend to change the order frequently. Raising a window means moving it "up", to the top of the stack. Lowering a window means moving it to the bottom of the stack. This motion is in the notional third dimension only, and does not change the position of the window on the screen.
You can raise and lower Emacs frame Windows with these functions:
nil, activation of the minibuffer raises the frame that the minibuffer window is in.
You can also enable auto-raise (raising automatically when a frame is selected) or auto-lower (lowering automatically when it is deselected) for any frame using frame parameters. See section Window Frame Parameters.
A frame configuration records the current arrangement of frames, all their properties, and the window configuration of each one. (See section Window Configurations.)
Sometimes it is useful to track the mouse, which means to display something to indicate where the mouse is and move the indicator as the mouse moves. For efficient mouse tracking, you need a way to wait until the mouse actually moves.
The convenient way to track the mouse is to ask for events to represent mouse motion. Then you can wait for motion by waiting for an event. In addition, you can easily handle any other sorts of events that may occur. That is useful, because normally you don't want to track the mouse forever--only until some other event, such as the release of a button.
read-eventto read the motion events and modify the display accordingly. See section Motion Events, for the format of mouse motion events.
The value of
track-mouse is that of the last form in body.
You should design body to return when it sees the up-event that
indicates the release of the button, or whatever kind of event means
it is time to stop tracking.
The usual purpose of tracking mouse motion is to indicate on the screen the consequences of pushing or releasing a button at the current position.
In many cases, you can avoid the need to track the mouse by using
mouse-face text property (see section Properties with Special Meanings).
That works at a much lower level and runs more smoothly than
Lisp-level mouse tracking.
give access to the current position of the mouse.
(frame x . y), where x and y are integers giving the position in characters relative to the top left corner of the inside of frame.
mouse-positionexcept that it returns coordinates in units of pixels rather than units of characters.
set-mouse-positionexcept that x and y are in units of pixels rather than units of characters. These coordinates are not required to be within the frame.
If frame is not visible, this function does nothing. The return value is not significant.
When using a window system, a Lisp program can pop up a menu so that the user can choose an alternative with the mouse.
The argument position specifies where on the screen to put the menu. It can be either a mouse button event (which says to put the menu where the user actuated the button) or a list of this form:
((xoffset yoffset) window)
where xoffset and yoffset are coordinates, measured in pixels, counting from the top left corner of window's frame.
If position is
t, it means to use the current mouse
position. If position is
nil, it means to precompute the
key binding equivalents for the keymaps specified in menu,
without actually displaying or popping up the menu.
The argument menu says what to display in the menu. It can be a keymap or a list of keymaps (see section Menu Keymaps). Alternatively, it can have the following form:
(title pane1 pane2...)
where each pane is a list of form
(title (line . item)...)
Each line should be a string, and each item should be the value to return if that line is chosen.
Usage note: Don't use
x-popup-menu to display a menu
if you could do the job with a prefix key defined with a menu keymap.
If you use a menu keymap to implement a menu, C-h c and C-h
a can see the individual items in that menu and provide help for them.
If instead you implement the menu by defining a command that calls
x-popup-menu, the help facilities cannot know what happens inside
that command, so they cannot give any help for the menu's items.
The menu bar mechanism, which lets you switch between submenus by
moving the mouse, cannot look within the definition of a command to see
that it calls
x-popup-menu. Therefore, if you try to implement a
x-popup-menu, it cannot work with the menu bar in
an integrated fashion. This is why all menu bar submenus are
implemented with menu keymaps within the parent menu, and never with
x-popup-menu. See section The Menu Bar,
If you want a menu bar submenu to have contents that vary, you should
still use a menu keymap to implement it. To make the contents vary, add
a hook function to
menu-bar-update-hook to update the contents of
the menu keymap as necessary.
A dialog box is a variant of a pop-up menu--it looks a little
different, it always appears in the center of a frame, and it has just
one level and one pane. The main use of dialog boxes is for asking
questions that the user can answer with "yes", "no", and a few other
alternatives. The functions
dialog boxes instead of the keyboard, when called from commands invoked
by mouse clicks.
(title (string . value)...)
which looks like the list that specifies a single pane for
The return value is value from the chosen alternative.
An element of the list may be just a string instead of a cons cell
(string . value). That makes a box that cannot
nil appears in the list, it separates the left-hand items from
the right-hand items; items that precede the
nil appear on the
left, and items that follow the
nil appear on the right. If you
don't include a
nil in the list, then approximately half the
items appear on each side.
Dialog boxes always appear in the center of a frame; the argument
position specifies which frame. The possible values are as in
x-popup-menu, but the precise coordinates don't matter; only the
In some configurations, Emacs cannot display a real dialog box; so instead it displays the same items in a pop-up menu in the center of the frame.
These variables specify which shape to use for the mouse pointer in various situations, when using the X Window System:
These variables affect newly created frames. They do not normally affect existing frames; however, if you set the mouse color of a frame, that also updates its pointer shapes based on the current values of these variables. See section Window Frame Parameters.
The values you can use, to specify either of these pointer shapes, are defined in the file `lisp/term/x-win.el'. Use M-x apropos RET x-pointer RET to see a list of them.
The X server records a set of selections which permit transfer of data between application programs. The various selections are distinguished by selection types, represented in Emacs by symbols. X clients including Emacs can read or set the selection for any given type.
nil, it means to clear out the selection. Otherwise, data may be a string, a symbol, an integer (or a cons of two integers or list of two integers), an overlay, or a cons of two markers pointing to the same buffer. An overlay or a pair of markers stands for text in the overlay or between the markers.
The argument data may also be a vector of valid non-vector selection values.
Each possible type has its own selection value, which changes
independently. The usual values of type are
SECONDARY; these are symbols with upper-case names, in accord
with X Window System conventions. The default is
The data-type argument specifies the form of data conversion to
use, to convert the raw data obtained from another X client into Lisp
data. Meaningful values include
INTEGER. (These are symbols with
upper-case names in accord with X conventions.) The default for
The X server also has a set of numbered cut buffers which can store text or other data being moved between applications. Cut buffers are considered obsolete, but Emacs supports them for the sake of X clients that still use them.
The argument pattern should be a string, perhaps with wildcard characters: the `*' character matches any substring, and the `?' character matches any single character. Pattern matching of font names ignores case.
If you specify face and frame, face should be a face name (a symbol) and frame should be a frame.
The optional argument maximum sets a limit on how many fonts to
return. If this is non-
nil, then the return value is truncated
after the first maximum matching fonts. Specifying a small value
for maximum can make this function much faster, in cases where
many fonts match the pattern.
A fontset is a list of fonts, each assigned to a range of character codes. An individual font cannot display the whole range of characters that Emacs supports, but a fontset can. Fontsets have names, just as fonts do, and you can use a fontset name in place of a font name when you specify the "font" for a frame or a face. Here is information about defining a fontset under Lisp program control.
Whitespace characters before and after the commas are ignored.
The first part of the string, fontpattern, should have the form of a standard X font name, except that the last two fields should be `fontset-alias'.
The new fontset has two names, one long and one short. The long name is
fontpattern in its entirety. The short name is
`fontset-alias'. You can refer to the fontset by either
name. If a fontset with the same name already exists, an error is
signaled, unless noerror is non-
nil, in which case this
function does nothing.
If optional argument style-variant-p is non-
nil, that says
to create bold, italic and bold-italic variants of the fontset as well.
These variant fontsets do not have a short name, only a long one, which
is made by altering fontpattern to indicate the bold or italic
The specification string also says which fonts to use in the fontset. See below for the details.
The construct `charset:font' specifies which font to use (in this fontset) for one particular character set. Here, charset is the name of a character set, and font is the font to use for that character set. You can use this construct any number of times in the specification string.
For the remaining character sets, those that you don't specify explicitly, Emacs chooses a font based on fontpattern: it replaces `fontset-alias' with a value that names one character set. For the ASCII character set, `fontset-alias' is replaced with `ISO8859-1'.
In addition, when several consecutive fields are wildcards, Emacs collapses them into a single wildcard. This is to prevent use of auto-scaled fonts. Fonts made by scaling larger fonts are not usable for editing, and scaling a smaller font is not useful because it is better to use the smaller font in its own size, which Emacs does.
Thus if fontpattern is this,
the font specification for ASCII characters would be this:
and the font specification for Chinese GB2312 characters would be this:
You may not have any Chinese font matching the above font specification. Most X distributions include only Chinese fonts that have `song ti' or `fangsong ti' in the family field. In such a case, `Fontset-n' can be specified as below:
Emacs.Fontset-0: -*-fixed-medium-r-normal-*-24-*-*-*-*-*-fontset-24,\ chinese-gb2312:-*-*-medium-r-normal-*-24-*-gb2312*-*
Then, the font specifications for all but Chinese GB2312 characters have `fixed' in the family field, and the font specification for Chinese GB2312 characters has a wild card `*' in the family field.
tif so; otherwise,
nil. The argument frame says which frame's display to ask about; if frame is omitted or
nil, the selected frame is used.
Note that this does not tell you whether the display you are using really supports that color. You can ask for any defined color on any kind of display, and you will get some result--that is how the X server works. Here's an approximate way to test whether your display supports the color color:
(defun x-color-supported-p (color &optional frame) (and (x-color-defined-p color frame) (or (x-display-color-p frame) (member color '("black" "white")) (and (> (x-display-planes frame) 1) (equal color "gray")))))
(x-color-values "black") => (0 0 0) (x-color-values "white") => (65280 65280 65280) (x-color-values "red") => (65280 0 0) (x-color-values "pink") => (65280 49152 51968) (x-color-values "hungry") => nil
The color values are returned for frame's display. If frame
is omitted or
nil, the information is returned for the selected
x-get-resourceretrieves a resource value from the X Windows defaults database.
Resources are indexed by a combination of a key and a class. This function searches using a key of the form `instance.attribute' (where instance is the name under which Emacs was invoked), and using `Emacs.class' as the class.
The optional arguments component and subclass add to the key and the class, respectively. You must specify both of them or neither. If you specify them, the key is `instance.component.attribute', and the class is `Emacs.class.subclass'.
x-get-resourceshould look up. The default value is
"Emacs". You can examine X resources for application names other than "Emacs" by binding this variable to some other string, around a call to
See section `X Resources' in The GNU Emacs Manual.
This section describes functions you can use to get information about
the capabilities and origin of an X display that Emacs is using. Each
of these functions lets you specify the display you are interested in:
the display argument can be either a display name, or a frame
(meaning use the display that frame is on). If you omit the
display argument, or specify
nil, that means to use the
selected frame's display.
nilif the display supports the SaveUnder feature.
tif the screen can display shades of gray.
tif the screen is a color screen.