Automatically select a display configuration based on connected devices
This is a compatible Python rewrite of wertarbyte/autorandr. Contributions for bash-completion, fd.o/XDG autostart, Nitrogen, pm-utils, and systemd can be found under contrib.
The original wertarbyte/autorandr
tree is unmaintained, with lots of open pull requests and issues. I forked it
and merged what I thought were the most important changes. If you are searching
for that version, see the legacy
branch.
Note that the Python version is better suited for non-standard configurations,
like if you use --transform
or --reflect
. If you use auto-disper
, you
have to use the bash version, as there is no disper support in the Python
version (yet). Both versions use a compatible configuration file format, so
you can, to some extent, switch between them. I will maintain the legacy
branch until @wertarbyte finds the time to maintain his branch again.
If you are interested in why there are two versions around, see #7, #8 and especially #12 if you are unhappy with this version and would like to contribute to the bash version.
autorandr is available under the terms of the GNU General Public License (version 3).
Contributors to this version of autorandr are:
You can use the autorandr.py
script as a stand-alone binary. If you'd like to
install it as a system-wide application, there is a Makefile included that also
places some configuration files in appropriate directories such that autorandr
is invoked automatically when a monitor is connected or removed, the system
wakes up from suspend, or a user logs into an X11 session. Run make install
as root to install it.
If you prefer to have a system wide install managed by your package manager, you can
make deb
.make rpm
.We appreciate packaging scripts for other distributions, please file a pull request if you write one.
If you prefer pip
over your package manager, you can install autorandr with:
sudo pip install "git+http://github.com/phillipberndt/autorandr#egg=autorandr"
or simply
sudo pip install autorandr
if you prefer to use a stable version.
Save your current display configuration and setup with:
autorandr --save mobile
Connect an additional display, configure your setup and save it:
autorandr --save docked
Now autorandr can detect which hardware setup is active:
$ autorandr
mobile
docked (detected)
To automatically reload your setup:
$ autorandr --change
To manually load a profile:
$ autorandr --load <profile>
or simply:
$ autorandr <profile>
autorandr tries to avoid reloading an identical configuration. To force the (re)configuration:
$ autorandr --load <profile> --force
To prevent a profile from being loaded, place a script call block in its directory. The script is evaluated before the screen setup is inspected, and in case of it returning a value of 0 the profile is skipped. This can be used to query the status of a docking station you are about to leave.
If no suitable profile can be identified, the current configuration is kept.
To change this behaviour and switch to a fallback configuration, specify
--default <profile>
. The system-wide installation of autorandr by default
calls autorandr with a parameter --default default
. There are three special,
virtual configurations called horizontal
, vertical
and common
. They
automatically generate a configuration that incorporates all screens
connected to the computer. You can symlink default
to one of these
names in your configuration directory to have autorandr use any of them
as the default configuration without you having to change the system-wide
configuration.
You can store default values for any option in an INI-file located at
~/.config/autorandr/settings.ini
. In a config
section, you may place any
default values in the form option-name=option-argument
.
A common and effective use of this is to specify default skip-options
, for
instance skipping the gamma
setting if using
redshift
as a daemon. To implement
the equivalent of --skip-options gamma
, your settings.ini
file should look
like this:
[config]
skip-options=gamma
Three more scripts can be placed in the configuration directory
(as defined by the XDG spec,
usually ~/.config/autorandr
or ~/.autorandr
if you have an old installation
for user configuration and /etc/xdg/autorandr
for system wide configuration):
postswitch
is executed after a mode switch has taken place. This can be
used to notify window managers or other applications about the switch.preswitch
is executed before a mode switch takes place.postsave
is executed after a profile was stored or altered.predetect
is executed before autorandr attempts to run xrandr. These scripts must be executable and can be placed directly in the configuration directory, where they will always be executed, or in the profile subdirectories, where they will only be executed on changes regarding that specific profile.
Instead (or in addition) to these scripts, you can also place as many executable
files as you like in subdirectories called script_name.d
(e.g. postswitch.d
).
The order of execution of scripts in these directories is by file name, you can
force a certain ordering by naming them 10-wallpaper
, 20-restart-wm
, etc.
If a script with the same name occurs multiple times, user configuration takes precedence over system configuration (as specified by the XDG spec) and profile configuration over general configuration.
As a concrete example, suppose you have the files
/etc/xdg/autorandr/postswitch
~/.config/autorandr/postswitch
~/.config/autorandr/postswitch.d/notify-herbstluftwm
~/.config/autorandr/docked/postswitch
and switch from mobile
to docked
. Then
~/.config/autorandr/docked/postswitch
is executed, since the profile specific
configuration takes precedence, and
~/.config/autorandr/postswitch.d/notify-herbstluftwm
is executed, since
it has a unique name.
If you switch back from docked
to mobile
, ~/.config/autorandr/postswitch
is executed instead of the docked
specific postswitch
.
If you experience issues with xrandr being executed too early after connecting
a new monitor, then you can use a predetect
script to delay the execution.
Write e.g. sleep 1
into that file to make autorandr wait a second before
running xrandr
.
Some of autorandr's state is exposed as environment variables
prefixed with AUTORANDR_
, such as:
AUTORANDR_CURRENT_PROFILE
AUTORANDR_CURRENT_PROFILES
AUTORANDR_PROFILE_FOLDER
AUTORANDR_MONITORS
with the intention that they can be used within the hook scripts.
For instance, you might display which profile has just been activated by
including the following in a postswitch
script:
notify-send -i display "Display profile" "$AUTORANDR_CURRENT_PROFILE"
The one kink is that during preswitch
, AUTORANDR_CURRENT_PROFILE
is
reporting the upcoming profile rather than the current one.
The EDID strings in the ~/.config/autorandr/*/setup
files may contain an
asterisk to enable wildcard matching: Such EDIDs are matched against connected
monitors using the usual file name globbing rules. This can be used to create
profiles matching multiple (or any) monitors.
In order for udev
to detect drm
events from the native NVidia driver, the
kernel parameter nvidia-drm.modeset
must be set to 1. For example, add a file
/etc/modprobe.d/nvidia-drm-modeset.conf
:
options nvidia_drm modeset=1
Before running autorandr will check the environment for the WAYLAND_DISPLAY
variable to check if the program is running in a Wayland session. This is to
avoid issues between usage of xrandr in Wayland environments.
If you need to run autorandr in a Wayland environment, one workaround is to
unset the WAYLAND_DISPLAY
variable before running the program, such as:
WAYLAND_DISPLAY= autorandr
autorandr 1.15
autorandr 1.14
autorandr 1.13.3
autorandr 1.13.2
autorandr 1.13.1
autorandr 1.13
--skip-options set
to skip setting propertiesautorandr 1.12.1
--match-edid
(see #273)autorandr 1.12
--list
to list all profiles--cycle
to cycle all detected profilesautorandr 1.11
autorandr 1.10.1
autorandr 1.10
autorandr 1.9
--force
autorandr 1.8.1
autorandr 1.8
autorandr 1.7
autorandr 1.6
AUTORANDR_MONITORS
variable to hooks (by @bricewge, #106)autorandr 1.5
autorandr 1.4
off
, which disables all screensautorandr 1.3
--load
--batch
mode (See #87)autorandr 1.2
--panning
unless it is required (See #72)clone-largest
virtual profileautorandr 1.1
--no-block
from udev rule (See #61)predetect
#
) in config/setup filesautorandr 1.0
/etc/xdg/autostart
by default