xdotool的manual

XDOTOOL(1)                            XDOTOOL(1)

NAME
       xdotool - command-line X11 automation
       tool

SYNOPSIS
       xdotool cmd args...

Notation: Some documentation uses
       [window] to denote an optional window
       argument. This case means that the
       argument, if not present, will default to
       "%1". See "WINDOW STACK" for what "%1"
       means.

DESCRIPTION
       xdotool lets you programatically (or
       manually) simulate keyboard input and
       mouse activity, move and resize windows,
       etc. It does this using X11‘s XTEST
       extension and other Xlib functions.

There is some support for Extended Window
       Manager Hints (aka EWMH or NetWM).  See
       the "EXTENDED WINDOW MANAGER HINTS"
       section for more information.

KEYBOARD COMMANDS
       key [options] keystroke [keystroke ...]
           Options:

--window window
               Send keystrokes to a specific
               window id. You can use "WINDOW
               STACK" references like "%1" and
               "%@" here. If there is a window
               stack, then "%1" is the default,
               otherwise the current window is
               used.

See also: "SENDEVENT NOTES" and
               "WINDOW STACK"

--clearmodifiers
               Clear modifiers before sending
               keystrokes. See CLEARMODIFIERS
               below.

--delay milliseconds
               Delay between keystrokes. Default
               is 12ms.

Type a given keystroke. Examples
           being "alt+r", "Control_L+J",
           "ctrl+alt+n", "BackSpace".

Generally, any valid X Keysym string
           will work. Multiple keys are
           separated by ‘+‘. Aliases exist for
           "alt", "ctrl", "shift", "super", and
           "meta" which all map to Foo_L, such
           as Alt_L and Control_L, etc.

In cases where your keyboard doesn‘t
           actually have the key you want to
           type, xdotool will automatically find
           an unused keycode and use that to
           type the key.

With respect to "COMMAND CHAINING",
           this command consumes the remainder
           of the arguments or until a new
           xdotool command is seen, because no
           xdotool commands are valid
           keystrokes.

Example: Send the keystroke "F2"
            xdotool key F2

Example: Send ‘a‘ with an accent over
           it (not on english keyboards, but
           still works with xdotool)
            xdotool key Aacute

Example: Send ctrl+l and then
           BackSpace as separate keystrokes:
            xdotool key ctrl+l BackSpace

Example: Send ctrl+c to all windows
           matching title ‘gdb‘ (See "COMMAND
           CHAINING")
            xdotool search --name gdb key ctrl+c

keydown [options] keystroke
           Same as above, except only keydown
           (press) events are sent.

keyup keystroke
           Same as above, except only keyup
           (release) events are sent.

type [options] something to type
           Options:

--window windowid
               Send keystrokes to a specific
               window id. See "SENDEVENT NOTES"
               below. The default, if no window
               is given, depends on the window
               stack. If the window stack is
               empty the current window is typed
               at using XTEST. Otherwise, the
               default is "%1" (see "WINDOW
               STACK").

--delay milliseconds
               Delay between keystrokes. Default
               is 12ms.

--clearmodifiers
               Clear modifiers before sending
               keystrokes. See CLEARMODIFIERS
               below.

Types as if you had typed it.
           Supports newlines and tabs (ASCII
           newline and tab). Each keystroke is
           separated by a delay given by the
           --delay option.

With respect to "COMMAND CHAINING",
           this command consumes the remainder
           of the arguments and types them. That
           is, no commands can chain after
           ‘type‘.

Example: to type ‘Hello world!‘ you
           would do:
            xdotool type ‘Hello world!‘

MOUSE COMMANDS
       mousemove [options] x y OR ‘restore‘
           Move the mouse to the specific X and
           Y coordinates on the screen.

You can move the mouse to the
           previous location if you specify
           ‘restore‘ instead of an X and Y
           coordinate. Restoring only works if
           you have moved previously in this
           same command invocation. Further, it
           does not work with the --window
           option.

For example, to click the top-left
           corner of the screen and move the
           mouse to the original position before
           you moved it, use this:
            xdotool mousemove 0 0 click 1
           mousemove xdorestore

--window WINDOW
               Specify a window to move relative
               to. Coordinates 0,0 are at the
               top left of the window you
               choose.

"WINDOW STACK" references are
               valid here, such as %1 and %@.
               Though, using %@ probably doesn‘t
               make sense.

--screen SCREEN
               Move the mouse to the specified
               screen to move to. This is only
               useful if you have multiple
               screens and ARE NOT using
               Xinerama.

The default is the current
               screen. If you specify --window,
               the --screen flag is ignored.

--polar
               Use polar coordinates. This makes
               ‘x‘ an angle (in degrees, 0-360,
               etc) and ‘y‘ the distance.

Rotation starts at ‘up‘ (0
               degrees) and rotates clockwise:
               90 = right, 180 = down, 270 =
               left.

The origin defaults to the center
               of the current screen. If you
               specify a --window, then the
               origin is the center of that
               window.

--clearmodifiers
               See CLEARMODIFIERS

--sync
               After sending the mouse move
               request, wait until the mouse is
               actually moved. If no movement is
               necessary, we will not wait. This
               is useful for scripts that depend
               on actions being completed before
               moving on.

Note: We wait until the mouse
               moves at all, not necessarily
               that it actually reaches your
               intended destination. Some
               applications lock the mouse
               cursor to certain regions of the
               screen, so waiting for any
               movement is better in the general
               case than waiting for a specific
               target.

mousemove_relative [options] x y
           Move the mouse x,y pixels relative to
           the current position of the mouse
           cursor.

--polar
               Use polar coordinates. This makes
               ‘x‘ an angle (in degrees, 0-360,
               etc) and ‘y‘ the distance.

Rotation starts at ‘up‘ (0
               degrees) and rotates clockwise:
               90 = right, 180 = down, 270 =
               left.

--sync
               After sending the mouse move
               request, wait until the mouse is
               actually moved. If no movement is
               necessary, we will not wait. This
               is useful for scripts that depend
               on actions being completed before
               moving on.

Note that we wait until the mouse
               moves at all, not necessarily
               that it actually reaches your
               intended destination. Some
               applications lock the mouse
               cursor to certain regions of the
               screen, so waiting for any
               movement is better in the general
               case than waiting for a specific
               target.

--clearmodifiers
               See CLEARMODIFIERS

click [options] button
           Send a click, that is, a mousedown
           followed by mouseup for the given
           button with a short delay between the
           two (currently 12ms).

Buttons generally map this way: Left
           mouse is 1, middle is 2, right is 3,
           wheel up is 4, wheel down is 5.

--clearmodifiers
               Clear modifiers before clicking.
               See CLEARMODIFIERS below.

--repeat REPEAT
               Specify how many times to click.
               Default is 1. For a double-click,
               use ‘--repeat 2‘

--delay MILLISECONDS
               Specify how long, in
               milliseconds, to delay between
               clicks. This option is not used
               if the --repeat flag is set to 1
               (default).

--window WINDOW
               Specify a window to send a click
               to. See "SENDEVENT NOTES" below
               for caveats. Uses the current
               mouse position when generating
               the event.

The default, if no window is
               given, depends on the window
               stack. If the window stack is
               empty the current window is typed
               at using XTEST. Otherwise, the
               default is "%1" (see "WINDOW
               STACK").

mousedown [options] button
           Same as click, except only a mouse
           down is sent.

mouseup [options] button
           Same as click, except only a mouse up
           is sent.

getmouselocation [--shell]
           Outputs the x, y, screen, and window
           id of the mouse cursor. Screen
           numbers will be nonzero if you have
           multiple monitors and are not using
           Xinerama.

--shell
               This makes getmouselocation
               output shell data you can eval.
               Example:

% xdotool getmouselocation --shell
                X=880
                Y=443
                SCREEN=0
                WINDOW=16777250

% eval $(xdotool getmouselocation --shell)
                % echo $X,$Y
                714,324

behave_screxdoen_edge [options] where
       command ...
           Bind an action to events when the
           mouse hits the screen edge or corner.

Options are:

--delay MILLISECONDS
               Delay in milliseconds before
               running the command. This allows
               you to require a given edge or
               corner to be held for a short
               period before your command will
               run. If you leave the edge or
               corner before the delay expires
               then the time will reset.

--quiesce MILLISECONDS
               Delay in milliseconds before the
               next command will run. This helps
               prevent accidentally running your
               command extra times; especially
               useful if you have a very short
               --delay (like the default of 0).

Event timeline

* Mouse hits an edge or corner.
            * If delay is nonzero, the mouse must stay in this edge or corner until delay time expires.
            * If still in the edge/corner, trigger.
            * If quiesce is nonzero, then there is a cool-down period where the next
              trigger cannot occur

Valid ‘where‘ values are:

left
           top-left
           top
           top-right
           right
           bottom-left
           bottom
           bottom-right

Examples:
            # Activate google-chrome when you
           move the mouse to the bottom-left
           corner:
            xdotool behave_screen_edge bottom-
           left \
              search --class google-chrome
           windowactivate

# Go to the next workspace (right). Known to work in GNOME (metacity and compiz)
            xdotool behave_screen_edge --delay 500 bottom-right key XF86Forward

# Activate firefox and do a web search in a new tab for text in your clipboard
            xdotool behave_screen_edge --delay 1000 top-left \
                search --classname Navigator \
                windowactivate --sync key --delay 250 ctrl+t ctrl+k ctrl+v Return

WINDOW COMMANDS
       search [options] pattern
           Search for windows with titles,
           names, or classes with a regular
           expression pattern. The output is
           line-delimited list of X window
           identifiers. If you are using
           "COMMAND CHAINING", the search
           command will only write window ids to
           stdout if it is the last (or only)
           command in the chain; otherwise, it
           is silent.

The result is saved to the window
           stack for future chained commands.
           See "WINDOW STACK" and "COMMAND
           CHAINING" for details.

The default options are "--name
           --class --classname" (unless you
           specify one one or more of --name
           --class or --classname).

The options available are:

--class
               Match against the window class.

--classname
               Match against the window
               classname.

--maxdepth N
               Set recursion/child search depth.
               Default is -1, meaning infinite.
               0 means no depth, only root
               windows will be searched. If you
               only want toplevel windows, set
               maxdepth of 1 (or 2, depending on
               how your window manager does
               decorations).

--name
               Match against the window name.
               This is the same string that is
               displayed in the window titlebar.

--onlyvisible
               Show only visible windows in the
               results. This means ones with map
               state IsViewable.

--pid PID
               Match windows that belong to a
               specific process id. This may not
               work for some X applications that
               do not set this metadata on its
               windows.

--screen N
               Select windows only on a specific
               screen. Default is to search all
               screens. Only meaningful if you
               have multiple displays and are
               not using Xinerama.

--desktop N
               Only match windows on a certain
               desktop. ‘N‘ is a number. The
               default is to search all
               desktops.

--limit N
               Stop searching after finding N
               matching windows. Specifying a
               limit will help speed up your
               search if you only want a few
               results.

The default is no search limit
               (which is equivalent to ‘--limit
               0‘)

--title
               DEPRECATED. See --name.

--all
               Require that all conditions be
               met. For example:

xdotool search --all --pid 1424 --name "Hello World"

This will match only windows that
               have "Hello World" as a name and
               are owned by pid 1424.

--any
               Match windows that match any
               condition (logically, ‘or‘). This
               is on by default. For example:

xdotool search --any --pid 1424 --name "Hello World"

This will match any windows owned
               by pid 1424 or windows with name
               "Hello World"

--sync
               Block until there are results.
               This is useful when you are
               launching an application want
               want to wait until the
               application window is visible.
               For example:

google-chrome &
                xdotool search --sync --onlyvisible --class "google-chrome"

selectwindow
           Get the window id (for a client) by
           clicking on it. Useful for having
           scripts query you humans for what
           window to act on. For example,
           killing a window by clicking on it:

xdotool selectwindow windowkill

behave window action command ...
           Bind an action to an event on a
           window. This lets you run additional
           xdotool commands whenever a matched
           event occurs.

The command run as a result of the
           behavior is run with %1 being the
           window that was acted upon. Examples
           follow after the event list.

The following are valid events:

mouse-enter
               Fires when the mouse enters a
               window. This is similar to ‘mouse
               over‘ events in javascript, if
               that helps.

mouse-leave
               Fires when the mouse leaves a
               window. This is the opposite of
               ‘mouse-enter‘

mouse-click
               Fires when the mouse is clicked.
               Specifically, when the mouse
               button is released.

focus
               Fires when the window gets input
               focus.

blur
               Fires when the window loses
               focus.

Examples:

# Print the cursor location whenever the mouse enters a currently-visible
            # window:
            xdotool search --onlyvisible . behave %@ mouse-enter getmouselocation

# Print the window title and pid whenever an xterm gets focus
            xdotool search --class xterm behave %@ focus getwindowname getwindowpid

# Emulate focus-follows-mouse
            xdotool search . behave %@ mouse-enter windowfocus

getwindowpid [window]
           Output the PID owning a given window.
           This requires effort from the
           application owning a window and may
           not work for all windows. This uses
           _NET_WM_PID property of the window.
           See "EXTENDED WINDOW MANAGER HINTS"
           below for more information.

If no window is given, the default is
           ‘%1‘. If no windows are on the stack,
           then this is an error. See "WINDOW
           STACK" for more details.

Example: Find the PID for all xterms:
            xdotool search --class xterm
           getwindowpid %@

getwindowname [window]
           Output the name of a given window,
           also known as the title. This is the
           text displayed in the window‘s
           titlebar by your window manager.

If no window is given, the default is
           ‘%1‘. If no windows are on the stack,
           then this is an error. See "WINDOW
           STACK" for more details.

getwindowgeometry [options] [window]
           Output the geometry (location and
           position) of a window. The values
           include: x, y, width, height, and
           screen number.

--shell
               Output values suitable for ‘eval‘
               in a shell.

getwindowfocus [-f]
           Prints the window id of the currently
           focused window. Saves the result to
           the window stack. See "WINDOW STACK"
           for more details.

If the current window has no WM_CLASS
           property, we assume it is not a
           normal top-level window and traverse
           up the parents until we find a window
           with a WM_CLASS set and return that
           window id.

If you really want the window
           currently having focus and don‘t care
           if it has a WM_CLASS setting, then
           use ‘getwindowfocus -f‘

windowsize [options] [window] width
       height
           Set the window size of the given
           window. If no window is given, %1 is
           the default.  See "WINDOW STACK" and
           "COMMAND CHAINING" for more details.

Percentages are valid for width and
           height. They are relative to the
           geometry of the screen the window is
           on. For example, to make a window the
           full width of the screen, but half
           height:

xdotool windowsize I<window> 100% 50%

Percentages are valid with --usehints
           and still mean pixel-width relative
           to the screen size.

The options available are:

--usehints
               Use window sizing hints (when
               available) to set width and
               height.  This is useful on
               terminals for setting the size
               based on row/column of text
               rather than pixels.

--sync
               After sending the window size
               request, wait until the window is
               actually resized. If no change is
               necessary, we will not wait. This
               is useful for scripts that depend
               on actions being completed before
               moving on.

Note: Because many window
               managers may ignore or alter the
               original resize request, we will
               wait until the size changes from
               its original size, not necessary
               to the requested size.

Example: To set a terminal to be
           80x24 characters, you would use:
            xdotool windowsize --usehints
           some_windowid 80 24

windowmove [options] [window] x y
           Move the window to the given
           position. If no window is given, %1
           is the default. See "WINDOW STACK"
           and "COMMAND CHAINING" for more
           details.

If the given x coordinate is
           literally ‘x‘, then the window‘s
           current x position will be unchanged.
           The same applies for ‘y‘.

Examples:

xdotool getactivewindow windowmove 100 100    # Moves to 100,100
            xdotool getactivewindow windowmove x 100      # Moves to x,100
            xdotool getactivewindow windowmove 100 y      # Moves to 100,y
            xdotool getactivewindow windowmove 100 y      # Moves to 100,y

--sync
               After sending the window move
               request, wait until the window is
               actually moved. If no movement is
               necessary, we will not wait. This
               is useful for scripts that depend
               on actions being completed before
               moving on.

--relative
               Make movement relative to the
               current window position.

windowfocus [options] [window]
           Focus a window. If no window is
           given, %1 is the default. See "WINDOW
           STACK" and "COMMAND CHAINING" for
           more details.

Uses XSetInputFocus which may be
           ignored by some window managers or
           programs.

--sync
               After sending the window focus
               request, wait until the window is
               actually focused. This is useful
               for scripts that depend on
               actions being completed before
               moving on.

windowmap [options] [window]
           Map a window. In X11 terminology,
           mapping a window means making it
           visible on the screen. If no window
           is given, %1 is the default. See
           "WINDOW STACK" and "COMMAND CHAINING"
           for more details.

--sync
               After requesting the window map,
               wait until the window is actually
               mapped (visible). This is useful
               for scripts that depend on
               actions being completed before
               moving on.

windowminimize [options] [window]
           Minimize a window. In X11
           terminology, this is called
           ‘iconify.‘  If no window is given, %1
           is the default. See "WINDOW STACK"
           and "COMMAND CHAINING" for more
           details.

--sync
               After requesting the window
               minimize, wait until the window
               is actually minimized. This is
               useful for scripts that depend on
               actions being completed before
               moving on.

windowraise [window_id=%1]
           Raise the window to the top of the
           stack. This may not work on all
           window managers. If no window is
           given, %1 is the default. See "WINDOW
           STACK" and "COMMAND CHAINING" for
           more details.

windowreparent [source_window]
       destination_window
           Reparent a window. This moves the
           source_window to be a child window of
           destination_window. If no source is
           given, %1 is the default.  "WINDOW
           STACK" window references (like %1)
           are valid for both source_window and
           destination_window See "WINDOW STACK"
           and "COMMAND CHAINING" for more
           details.

windowkill [window]
           Kill a window. This action will
           destroy the window and kill the
           client controlling it. If no window
           is given, %1 is the default. See
           WINDOW STACK and "COMMAND CHAINING"
           for more details.

windowunmap [options] [window_id=%1]
           Unmap a window, making it no longer
           appear on your screen. If no window
           is given, %1 is the default. See
           "WINDOW STACK" and "COMMAND CHAINING"
           for more details.

--sync
               After requesting the window
               unmap, wait until the window is
               actually unmapped (hidden). This
               is useful for scripts that depend
               on actions being completed before
               moving on.

set_window [options] [windowid=%1]
           Set properties about a window. If no
           window is given, %1 is the default.
           See "WINDOW STACK" and "COMMAND
           CHAINING" for more details.

Options:

--name newname
               Set window WM_NAME (the window
               title, usually)

--icon-name newiconname
               Set window WM_ICON_NAME (the
               window title when minimized,
               usually)

--role newrole
               Set window WM_WINDOW_ROLE

--classname newclassname
               Set window class name (not to be
               confused with window class)

--class newclass
               Set window class (not to be
               confused with window class name)

--overrideredirect value
               Set window‘s override_redirect
               value. This value is a hint to
               the window manager for whether or
               not it should be managed. If the
               redirect value is 0, then the
               window manager will draw borders
               and treat this window normally.
               If the value is 1, the window
               manager will ignore this window.

If you change this value, your
               window manager may not notice the
               change until the window is mapped
               again, so you may want to issue
               ‘windowunmap‘ and ‘windowmap‘ to
               make the window manager take
               note.

DESKTOP AND WINDOW COMMANDS
       These commands follow the EWMH standard.
       See the section "EXTENDED WINDOW MANAGER
       HINTS" for more information.

windowactivate [options] [window]
           Activate the window. This command is
           different from windowfocus: if the
           window is on another desktop, we will
           switch to that desktop. It also uses
           a different method for bringing the
           window up. I recommend trying this
           command before using windowfocus, as
           it will work on more window managers.

If no window is given, %1 is the
           default. See "WINDOW STACK" and
           "COMMAND CHAINING" for more details.

--sync
               After sending the window
               activation, wait until the window
               is actually activated. This is
               useful for scripts that depend on
               actions being completed before
               moving on.

getactivewindow
           Output the current active window.
           This command is often more reliable
           than getwindowfocus. The result is
           saved to the window stack. See
           "WINDOW STACK" for more details.

set_num_desktops number
           Changes the number of desktops or
           workspaces.

get_num_desktops
           Output the current number of
           desktops.

get_desktop_viewport [--shell]
           Report the current viewport‘s
           position. If --shell is given, the
           output is friendly to shell eval.

Viewports are sometimes used instead
           of ‘virtual desktops‘ on some window
           managers. A viewport is simply a view
           on a very large desktop area.

set_desktop_viewport x y
           Move the viewport to the given
           position. Not all requests will be
           obeyed - some windowmangers only obey
           requests that align to workspace
           boundaries, such as the screen size.

For example, if your screen is
           1280x800, you can move to the 2nd
           workspace by doing:
            xdotool set_desktop_viewport 1280 0

set_desktop [options] desktop_number
           Change the current view to the
           specified desktop.

--relative
               Use relative movements instead of
               absolute. This lets you move
               relative to the current desktop.

get_desktop
           Output the current desktop in view.

set_desktop_for_window [window]
       desktop_number
           Move a window to a different desktop.
           If no window is given, %1 is the
           default. See "WINDOW STACK" and
           "COMMAND CHAINING" for more details.

get_desktop_for_window [window]
           Output the desktop currently
           containing the given window. Move a
           window to a different desktop. If no
           window is given, %1 is the default.
           See WINDOW STACK and "COMMAND
           CHAINING" for more details.

MISCELLANEOUS COMMANDS
       exec [options] command [...]
           Execute a program. This is often
           useful when combined with
           behave_screen_edge to do things like
           locking your screen.

Options:

--sync
               Block until the child process
               exits. The child process exit
               status is then passed to the
               parent process (xdotool) which
               copies it.

Examples:
            # Lock the screen when the mouse
           sits in the top-right corner
            xdotool behave_screen_edge --delay
           1000 top-right \
              exec gnome-screensaver-command
           --lock
            # Substitute ‘xscreensaver-command
           -lock‘ if you use that program.

# The following will fail to move the mouse because we use ‘--sync‘ and
            # /bin/false exits nonzero:
            xdotool exec --sync /bin/false mousemove 0 0

# This succeeds, though, since we do not use --sync on the exec command.
            xdotool exec /bin/false mousemove 0 0

sleep seconds
           Sleep for a specified period.
           Fractions of seconds (like 1.3, or
           0.4) are valid, here.

SCRIPTS
       xdotool can read a list of commands via
       stdin or a file if you want. A script
       will fail when any command fails.

Truthfully, ‘script‘ mode isn‘t fully
       fleshed out and may fall below your
       expectations. If you have suggestions,
       please email the list or file a bug (See
       CONTACT).

Scripts can use positional arguments
       (Represented by $1, $2, ...) and
       environment variables (like $HOME or
       $WINDOWID). Quoting arguments should work
       as expected.

Scripts are processed for parameter and
       environment variable expansion and then
       run as if you had invoked xdotool with
       the entire script on one line (using
       COMMAND CHAINING).

·   Read commands from a file:

xdotool filename

·   Read commands from stdin:

xdotool -

·   Read commands from a redirected file

xdotool - < myfile

You can also write scripts that only
       execute xdotool. Example:

#!/usr/local/bin/xdotool
        search --onlyvisible --classname $1

windowsize %@ $2 $3
        windowraise %@

windowmove %1 0 0
        windowmove %2 $2 0
        windowmove %3 0 $3
        windowmove %4 $2 $3

This script will take all windows matched
       by the classname query given by arg1 ($1)
       and sizes/moves them into a 2x2 grid with
       windows sized by the 2nd and 3rd
       parameters.

Here‘s an example usage:

% ./myscript xterm 600 400

Running it like this will take 4 visible
       xterms, raise them, and move them into a
       2x2 tile grid with each window 600x400
       pixels in size.

CLEARMODIFIERS
       Any command taking the --clearmodifiers
       flag will attempt to clear any active
       input modifiers during the command and
       restore them afterwards.

For example, if you were to run this
       command:
        xdotool key a

The result would be ‘a‘ or ‘A‘ depending
       on whether or not you were holding the
       shift key on your keyboard. Often it is
       undesirable to have any modifiers active,
       so you can tell xdotool to clear any
       active modifiers.

The order of operations if you hold shift
       while running ‘xdotool key
       --clearmodifiers a‘ is this:

1. Query for all active modifiers (finds
       shift, in this case)
       2. Try to clear shift by sending ‘key up‘
       for the shift key
       3. Runs normal ‘xdotool key a‘
       4. Restore shift key by sending ‘key
       down‘ for shift

The --clearmodifiers flag can currently
       clear of the following:

·   any key in your active keymap that
           has a modifier associated with it.
           (See xmodmap(1)‘s ‘xmodmap -pm‘
           output)

·   mouse buttons (1, 2, 3, 4, and 5)

·   caps lock

SENDEVENT NOTES
       If you are trying to send key input to a
       specific window, and it does not appear
       to be working, then it‘s likely your
       application is ignoring the events
       xdotool is generating. This is fairly
       common.

Sending keystrokes to a specific window
       uses a different API than simply typing
       to the active window. If you specify
       ‘xdotool type --window 12345 hello‘
       xdotool will generate key events and send
       them directly to window 12345.  However,
       X11 servers will set a special flag on
       all events generated in this way (see
       XEvent.xany.send_event in X11‘s manual).
       Many programs observe this flag and
       reject these events.

It is important to note that for key and
       mouse events, we only use XSendEvent when
       a specific window is targeted. Otherwise,
       we use XTEST.

Some programs can be configured to accept
       events even if they are generated by
       xdotool. Seek the documentation of your
       application for help.

Specific application notes (from the
       author‘s testing): * Firefox 3 seems to
       ignore all input when it does not have
       focus.  * xterm can be configured while
       running with ctrl+leftclick, ‘Allow
       SendEvents‘ * gnome-terminal appears to
       accept generated input by default.

WINDOW STACK
       Certain commands (search,
       getactivewindow, getwindowfocus) will
       find windows for you. These results
       generally printed to stdout, but they are
       also saved to memory for future use
       during the lifetime of the xdotool
       process. See "COMMAND CHAINING" for more
       information.

The only modifications support for the
       window stack are to replace it. That is,
       two of two sequential searches, only the
       last one‘s results will be the window
       stack.

COMMAND CHAINING
       xdotool supports running multiple
       commands on a single invocation.
       Generally, you‘ll start with a search
       command (see "WINDOW STACK") and then
       perform a set of actions on those
       results.

To query the window stack, you can use
       special notation "%N" where N is a number
       or the ‘@‘ symbol. If %N is given, the
       Nth window will be selected from the
       window stack. Generally you will only
       want the first window or all windows.
       Note that the order of windows in the
       window stack corresponds to the window
       stacking order, i.e. the bottom-most
       window will be reported first (see
       XQueryTree(3)). Thus the order of the
       windows in the window stack may not be
       consistent across invocations.

The notation described above is used as
       the "window" argument for any given
       command.

For example, to resize all xterms to
       80x24:

xdotool search --class xterm -- windowsize --usehints %@ 80 24

Resize move the current window:

xdotool getactivewindow windowmove 0 0

In all cases, the default window
       argument, if omitted, will default to
       "%1". It is obviously an error if you
       omit the window argument and the window
       stack is empty. If you try to use the
       window stack and it is empty, it is also
       an error.

To activate the first firefox window
       found:

xdotool search --class firefox windowactivate

These would error:

xdotool windowactivate
        xdotool windowactivate %1
        xdotool windowactivate %@

When xdotool exits, the current window
       stack is lost.

Additinally, commands that modify the
       "WINDOW STACK" will not print the results
       if they are not the last command. For
       example:

# Output the active window:
        % xdotool getactivewindow
        20971533

# Output the pid of the active window, but not the active window id:
        % xdotool getactivewindow getwindowpid
        4686

EXTENDED WINDOW MANAGER HINTS
       The following pieces of the EWMH standard
       are supported:

_NET_SUPPORTED
           Asks the window manager what is
           supported

_NET_CURRENT_DESKTOP
           Query and set the current desktop.
           Support for this enables these
           commands: "set_desktop",
           "get_desktop".

_NET_WM_DESKTOP
           Query and set what desktop a window
           is living in. Support for this
           enables these commands:
           "set_desktop_for_window",
           "get_desktop_for_window".

_NET_ACTIVE_WINDOW
           Allows you to query and set the
           active window by asking the window
           manager to bring it forward. Support
           for this enables these commands:
           "windowactivate", "getactivewindow".

_NET_WM_PID
           This feature is application
           dependent, not window-manager
           dependent. Query the PID owning a
           given window. Support for this
           enables these commands:
           "getwindowpid".

SUPPORTED FEATURES
       xdotool (and libxdo) will try to function
       under all circumstances. However, there
       may be some cases where functionality is
       not provided by your X server or by your
       window manager. In these cases, xdotool
       will try to detect and tell you if an
       action requires a feature not currently
       supported by your system.

For window-manager specific features, see
       "EXTENDED WINDOW MANAGER HINTS".

XTEST
           If your X server does not support
           XTEST, then some typing and mouse
           movement features may not work.
           Specifically, typing and mouse
           actions that act on the "current
           window" (window 0 in libxdo) are
           unlikely to work.

In most cases, XTEST is a feature you
           can enable on your X server if it is
           not enabled by default.

You can see the list of supported X
           extensions by typing ‘xdpyinfo‘ and
           looking the text ‘number of
           extensions: ...‘

BUGS
       Typing unusual symbols under non-us
       keybindings is known to occasionally send
       the wrong character.

SEE ALSO
       xprop(1), xwininfo(1),

Project site:
       <http://www.semicomplete.com/projects/xdotool>

Google Code:
       <http://semicomplete.googlecode.com/>

EWMH specification:
       <http://standards.freedesktop.org/wm-spec/wm-spec-1.3.html>

CONTACT
       Please send questions to
       [email protected] File bugs
       and feature requests at the following
       URL:

<http://code.google.com/p/semicomplete/issues/list>

Alternately, if you prefer email, feel
       free to file bugs by emailing the list.
       What works for you :)

AUTHOR
       xdotool was written by Jordan Sissel.

This manual page was written originally
       by Daniel Kahn Gillmor
       <[email protected]> for the Debian
       project (but may be used by others). It
       is maintained by Jordan Sissel.

Patches, ideas, and other contributions
       by many, nice folks. See the CHANGELIST
       file for who provided what.

2013-01-10         XDOTOOL(1)

时间: 2024-10-18 00:24:17

xdotool的manual的相关文章

git manual

git(1) p,li,dt,dd,div,pre,h1,h2,h3,h4,h5,h6 { } body { margin: 1em 5% 1em 5% } a { color: blue; text-decoration: underline } a:visited { color: fuchsia } em { font-style: italic; color: navy } strong { font-weight: bold; color: #083194 } tt { color:

YASM User Manual

This document is the user manual for the Yasm assembler. It is intended as both an introduction and a general-purpose reference for all Yasm users. 1.?Introduction Yasm is a BSD-licensed assembler that is designed from the ground up to allow for mult

[MySQL Reference Manual] 7 备份和恢复

7. 备份和恢复 本章主要会介绍: 1.备份的类型:逻辑备份,物理备份,全备和增量4种 2.创建备份的方法 3.还原方法,包括还原到时间点 4.备份计划,压缩和加密 5.表维护,恢复损坏的表 7. 备份和恢复... 1 7.1备份和还原类型... 1 7.1.1 物理备份VS逻辑备份... 1 7.1.2 Online VS OFFLINE. 1 7.1.3 本地VS远程... 1 7.1.4 快照备份... 1 7.1.5 全备VS增量备份... 1 7.1.6 完全恢复VS时间点(增量)恢复

MySQLSyntaxErrorException: You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version for the right syntax to use near ...

下面是我update数据库时打印出来的异常: ### Error updating database. Cause: com.mysql.jdbc.exceptions.jdbc4.MySQLSyntaxErrorException: You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version for the right syntax to use nea

P6 EPPM Manual Installation Guide (Oracle Database)

Contents Oracle Database Manual Configuration Overview ,,★★5 Oracle Database Installation ,,★★6 Creating the Database Structure for Oracle and Loading Application Data ,,★★7 Creating the P6 EPPM Database Structure for Oracle ,,★★7 Copying the Script

《The One Page Linux Manual》的中文版

对于初学者来说,Linux等开源技术的学习曲线比较"陡".美国新泽西州立大学(The State University of New Jersey)的老师为了方便Linux入门课程讲授,将最常用的命令总结了一下,撰写一个<The One Page Linux Manual>. 下载:http://www.cs.rutgers.edu/resources/howto/One_Page_Linux_Manual.pdf 此文档简单易用,一目了然.初学者将此文档正反打印到一页A4

Linux User&#39;s Manual IOSTAT

IOSTAT(1) Linux User's Manual IOSTAT(1) NAME iostat - Report Central Processing Unit (CPU) statistics and input/out- put statistics for devices, partitions and network filesystems (NFS). SYNOPSIS iostat [ -c ] [ -d ] [ -N ] [ -n ] [ -h ] [ -k | -m ]

You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version

1.错误描述 org.hibernate.exception.SQLGrammarException: error executing work at org.hibernate.exception.internal.SQLExceptionTypeDelegate.convert(SQLExceptionTypeDelegate.java:80) at org.hibernate.exception.internal.StandardSQLExceptionConverter.convert(

Manual close is not allowed over a Spring managed SqlSession

Manual close is not allowed over a Spring managed SqlSession 整合spring与MyBatis时出现如下警告: [org.springframework.beans.factory.support.DisposableBeanAdapter.invokeCustomDestroyMethod(DisposableBeanAdapter.java:360)]-[WARN] Invocation of destroy method 'clo