This branch contains features not yet merged into the mainline. Included features not merge into the mainline:

Redshift reads environment variables to figure out which adjustment method to use if not method has been selected.

This feature will probably not be merge into the mainline.

It is possible to use external commands in the configuration file, with fallback values if the command fails.

For example, changing the line crtc=crtc-index in redshift.conf to

crtc=fallback-crtc-index$(get-crtc-by-edid long-edid-value)

will run the command sh -c "get-crtc-by-edid long-edid-value" and read its output and use that for the crtc setting. But if get-crtc-by-edid is killed by a signal or exit with a value other than zero, or otherwise cannot be executed successfully, it with use the value fallback-crtc-index.

If built with --enable-fakegdi it is possible to use the wingdi adjustment method without Windows. If built with this options, when wingdi is used the Windows GDI calls will be translated to X RandR calls. This is intended for testing updates to the gamma-w32gdi without have to run Windows.

: in the argument for the options -l and -m has been removed in favour of using separate arguments.

This as not been tested on Windows. The option display=N has been added to the wingdi method. It is used to select monitor.

Redshift suggests specifying CRTC if it thinks you specified screen instead by mistake.

The option -P or configuration preserve-calibrations can be used if not running in one shot mode. When used, the current calibrations will not be overridden, instread they will be applied on top of the adjustments made by Redshift.

By specifying adjustment method settings multiple times in redshift.conf as separate sections, you can use multiple CRTC:s and multiple screens or graphics cards concurrently. The gamma settings can be selected specified for the CRTC:s individually. It is also possible to use all instead of an index.

Similar to the previous section, it is possible to specify multiple X displays. However, it is not possible to use display=all.

It is possible to, at compile to, set {MIN,MAX}_{BRIGHTNESS,GAMMA} and TRANSITION_{LOW,HIGH}. The upper brightness and gamma boundaries can be removed by defining NO_MAX_{BRIGHTNESS,GAMMA}.

If redshift fails to use DRM it tries to figure out why and informs the user. If is not able to send a gamma ramp update it will check why and take appropriate actions, rather than ignoring it and assuming that it is because you are in a display server.

When using RandR or DRM it is possible to select monitor by its EDID value. A value used to aid plug-and-play, and are in but rare cases unique. Namely, thay may, and most often do, contain the monitor's serial number or random data, week and year of manufacture, some other kind of serial number, and prototype calibration information. Becuase of this addition it is also possible to use ignorable=1 which causes redshift to skip a display, graphics card, screen, CRTC or monitor it cannot find.

Quartz support using Core Graphics has been added for Mac OS X support. This have however only been tested on GNU/Linux using the --enable-fakequartz configure option and it is plausible that it will not run on or compile for Mac OS X.

If built with --enable-fakequartz it is possible to use the new quartz adjustment method without Mac OS X. If built with this options, when quartz is used the Core Graphics calls will be translated to X RandR calls. This is intended for testing updates to the gamma-quartz without have to run Mac OS X, as will as to be able to test the original implemention.

The program will not crash if it fails to get the user's entry in the passwd file. It also adds support for homeless users.

If redshift-gtk receives SIGUSR1 it is relayed to redshift, and if it receives SIGINT it exits cleanly which also fixes a bug where redshift-gtk consumes 100 % of one CPU if Control-c is pressed.

redshift-gtk no longer prints an error message if AppIndicator3 cannot be loaded.

The user's location is printed with N/S–E/W suffixes instead of being either negative if the user is on the south hemisphere or on the west side of the prime meridian. This is intended to help user's on the these locations when the miss the minus sign and ge the working solar elevations.

The crtc, screen, card and display options for the adjustment methods can now be comma-separated lists.

The [hooks] section has been added to redshift.conf In this section valid key names are day, night and twilight. Valid value as sh-syntax commands. day hooks will be executed when redshift enters or starts at full daytime solar elevation, and night hooks will be executed when redshift enters or starts at full night solar elevation. twilight hooks will be executed when redshift enters or starts at a transtition period between daytime and night.

The following information has been added to the info dialogue in redshift-gtk:

• Time of the previous sunrise/sunset
• Time of the next sunrise/sunset
• The time twilight (the transition period) started/ended
• The time twilight starts/ends

The time points are calculated an interative method and it will say that the will never occur / has never occurred if no satisfying point in time can be found within the span of one year.

When SIGUSR2 is received settings are reloaded. These settings are currently limited to settings that do not affect location providers and adjustment methods. Transition to the new settings are used if the line reload-transition=0 does not appear in redshift.conf.