summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2016-12-08Bump to xrandr3.Jérémie Courrèges-Anglas
2016-12-08Store RootWindow in global_screen structureMathieu OTHACEHE
Every screen stores a copy of xrandr RootWindow even if they all share the same RootWindow. This patch stores the RootWindow in a global location for all screens. Further work is required to remove RootWindow from per screen structure.
2016-12-08Select xrandr primary screen at startup if possibleMathieu OTHACEHE
If xrandr support is enabled and a screen is marked primary, use it as current screen at startup. Otherwise, use the first screen in sorted screen list as current screen at startup.
2016-12-08Simplify scanwins functionMathieu OTHACEHE
Use find_screen_by_attr function, remove screen argument and a useless debug message.
2016-12-08Add a function to find a screen matching a given XWindowAttributesMathieu OTHACEHE
It allows to find the screen associated to a Window, using it's x and y position.
2016-12-08Add primary screen detectionMathieu OTHACEHE
The user may specify that a screen is primary with xrandr. Save this information in ratpoison for future use. The use of XRRGetOutputPrimary forces us to bump xrandr supported revision from >=1.2 to >=1.3.
2016-12-08Move function declaration to the right headerMathieu OTHACEHE
find_screen is defined in screen.c so it makes more sense to declare it in screen.h
2016-12-08Remove useless argument from is_rp_window_for_screen functionMathieu OTHACEHE
The screen argument is useless as this function operates on all screens. To find if a window belong to a specific ratpoison screen, function is_rp_window_for_given_screen may be used.
2016-12-01Bump to ratpoison-1.4.9-xrandr2Jérémie Courrèges-Anglas
2016-12-01Fix and simplify sfrestoreJérémie Courrèges-Anglas
Use a scratch buffer for each screen, this simplifies a lot allocations and iterations, and allows restoring screens that contain more than one frame.
2016-12-01sfrestore: in case of error, mention the screen number, not its xrandr idJérémie Courrèges-Anglas
2016-12-01sfrestore: partial fix for frame allocation/deallocationJérémie Courrèges-Anglas
There is still a problem with allocation of screens vs frames. A screen can have multiple frames (for now this triggers a segfault).
2016-12-01sselect: print the screen number in case of failureJérémie Courrèges-Anglas
2016-12-01Callers shouldn't pass a negative number to screen_number()Jérémie Courrèges-Anglas
2016-12-01fdump: don't segfault if screen is not foundJérémie Courrèges-Anglas
2016-12-01Tighten screen sorting: comparing a screen with itself should return 0Jérémie Courrèges-Anglas
2016-12-01Kill gcc uninitialized warningMathieu OTHACEHE
It is useless to initialize screen_count, but it prevents a gcc warning.
2016-12-01Add height dimension to screen sortingMathieu OTHACEHE
Screen are only sorted according to their width offset from origin ("left" in screen struct). Also sort screen that share a same width offset according to their height offset ("top" in screen struct).
2016-12-01Fix sfrestoreMathieu OTHACEHE
Use screen numset in sfrestore
2016-12-01Fix sfdumpMathieu OTHACEHE
Use screen numset in sfdump
2016-12-01Fix fdump and sselectMathieu OTHACEHE
Use screen numset in fdump and sselect.
2016-12-01Assign a number to each screen using a numsetMathieu OTHACEHE
A number is added to rp_screen to identify every screen. The screen numbers are affected at ratpoison start, when screens are sorted. On screen addition, a new number is requested. On screen deletion, the screen number is released.
2016-11-24Kill gcc unused function warningMathieu OTHACEHE
The xrandr_rotation_string function is used only if DEBUG is enabled.
2016-11-24Tweak previous.Jérémie Courrèges-Anglas
Use nicer names that the long ac_cv_* variables. Also, the reason why the third parameter of AC_CHECK_LIB was significant is that, if using the default (empty), the macro adds -lXrandr to LIBS. Later, LIBS is used for compile tests, which fail if libXrandr is not in the default search path of ld(1), for example in /usr/X11R6/lib...
2016-11-23Fix rotations when using xrandrJérémie Courrèges-Anglas
2016-11-23define HAVE_XRANDR/link with libXrandr when we actually want toJérémie Courrèges-Anglas
2016-11-23Only call xrandr_fill_screen is xrandr is availableJérémie Courrèges-Anglas
2016-11-23Keep on handling ConfigureNotify if we don't have xrandrJérémie Courrèges-Anglas
2016-11-23Kill superflous var.Jérémie Courrèges-Anglas
2016-11-23Reinstate code that should stay: we may not have xrandrJérémie Courrèges-Anglas
2016-11-23Sprinkle more debug in xrandr code.Jérémie Courrèges-Anglas
2016-11-22Use a different tarball name for xrandr support.Jérémie Courrèges-Anglas
2016-11-22autotools: xkb is an external package -> --with[out]-xkbJérémie Courrèges-Anglas
2016-11-22Temporarily mark sfrestore as brokenJérémie Courrèges-Anglas
2016-11-22Style nits, typo and non-existing functions.Jérémie Courrèges-Anglas
2016-11-21Only use xrandr if Xrandr init succeedsJérémie Courrèges-Anglas
2016-11-21Replace extra definition with a declarationJérémie Courrèges-Anglas
2016-11-21autotools: detect precisely Xrandr supportJérémie Courrèges-Anglas
Bail out if xrandr support can't be found but the user explicitely requested it.
2016-11-21libXext isn't usedJérémie Courrèges-Anglas
2016-11-17Add xrandr supportMathieu OTHACEHE
Drop deprecated xinerama support and replace it with xrandr. Signed-off-by: Mathieu OTHACEHE <m.othacehe@gmail.com>
2016-11-11Debug statement that snuck in previous.Jérémie Courrèges-Anglas
2016-11-11Introduce multiple completion styles.Mathieu OTHACEHE
The default, legacy style is named BASIC. A new completion style named SUBSTRING is added. Use SUBSTRING for window name completion in select command. Use BASIC everywhere else.
2016-08-28NEWS entries for 1.4.9Jérémie Courrèges-Anglas
2016-06-12Simplify config file open checksJérémie Courrèges-Anglas
2016-06-12Read the system config even if the homedir can't be found.Jérémie Courrèges-Anglas
2016-06-12No need for getopt_long to return the index for long option parametersJérémie Courrèges-Anglas
2016-06-12Zero out struct sigaction instead of setting known fields.Jérémie Courrèges-Anglas
2016-06-12sigaction should always replace the signal handlerJérémie Courrèges-Anglas
The code is confusing, there is no reason to skip setting the handler if it is already set to SIG_IGN, especially now that SIG_IGN isn't used in ratpoison.
2016-06-12sigaction has other benefits than working on SVR4 systemsJérémie Courrèges-Anglas
2016-06-12Print a more readable error message if sigaction failsJérémie Courrèges-Anglas