| Commit message (Collapse) | Author | Age |
|---|
| |
|
|
| |
(non-embedded apps).
|
| | |
|
| | |
|
| | |
|
| |
|
|
| |
(still very rough - does not do things like fine-grained stacking, minimization, and such yet).
|
| |
|
|
| |
close right now.
|
| | |
|
| |
|
|
| |
Only a few more to go, and then assembling the WM guts can commence...
|
| |
|
|
| |
an enumeration (WINDOWSTATE -> WINDOWVISIBILITY), so I fixed the usage of that enum in the lumina-desktop project as well.
|
| | |
|
| |
|
|
| |
LuminaX11 library.
|
| |
|
|
| |
(that function is empty for the time being - error in the xcb_ewmh library with char array handling.
|
| |
|
|
| |
LuminaX11 class.
|
| | |
|
| |
|
|
|
|
| |
lumina-open call now, instead of a bunch of individual ones). Also fix up the tray app detection/failover methods a bit to catch/discard tray icons which were registered/destroyed almost simultaneously (or no destroy event was ever caught for the icon). This prevents the situation where a "blank" tray icon may be trying to repaint itself repeatedly (eating up CPU cycles).
NOTE: It seems like these "blank" tray apps are all GTK based, so it might be something in that toolkit which needs fixing to prevent registering a tray window which will never be used (or is instantly destroyed).
|
| | |
|
| | |
|
| |
|
|
| |
given window is "sticky" in which case it always returns the current workspace.
|
| |
|
|
| |
Found out that (at least on FreeBSD 10.2 right now), the xcb_damage_create() routine does absolutely nothing (no damage events ever sent out when the window changes). Switching this back to the Xlib XDamageCreate() routine makes it work perfectly again. Need to ask about this upstream and see if this is a known bug/regression in the XCB libraries.
|
| |
|
|
| |
adjust the theme files accordingly.
|
| |
|
|
| |
installed.
|
| |
|
|
| |
adding new functions for the window manager (direct access/control - not requests).
|
| |
|
|
| |
any XLib functions/libraries/extensions.
|
| | |
|
| |
|
|
| |
functions in lumina-desktop.
|
| |
|
|
| |
LSession to use those new versions
|
| | |
|
| |
|
|
| |
match their own internal specifications. On single-screen setups it works fine now, but on multi-screen setups the app does not seem to return the xsettings back to their original specs properly (Tested with OpenArena and MegaMario). This may result in the screen being "flipped" or placed on different monitors after closing the app.
|
| |
|
|
| |
locations when XINERAMA is used for combining monitors. However, the interface is still completely unresponsive because the Qt backend is completely screwed up by Xinerama (returns 0's for everything related to screens, geometries, etc, preventing menus from opening up even when manually started). This might be something where we just have to say that Xinerama is not supported for Lumina (at least until lumina-wm get finished up).
|
| |
|
|
| |
strange windows (like gimp) which move the main window off-screen before showing it (causing errors/crashes in the program when moving it back).
|
| |
|
|
|
|
| |
1) Switch some X11->XCB functions around (SetAsDesktop(), SetDisableWMActions()).
2) Fix up some QtQuick plugin detection routines (now it will properly see the user's quickplugins if any).
3) Connect the status change signal/slots for QtQuick plugins to try and detect script failures and remove the script (still not working reliably).
|
| |
|
|
| |
of the frame on each side of the given window.
|
| |
|
|
|
|
|
|
| |
GCC warning helped to spot a real bug: near line 1185 of
libLumina/LuminaX11.cpp there is a call: `free(&hints);`, - but the
"hints" is allocated on the stack a few lines above. This looks
like a copy-paste error, since other free(&hints) instances look
like sane (unless I've missed something, of course).
|
| |
|
|
| |
off the debugging for the session/desktop/panel classes by default.
|
| |
|
|
| |
creation to XCB, and place it within the embedding routine in LuminaX11 instead.
|
| |
|
|
| |
tinker with the session cleanup routine quite a bit to streamline the order in which things are closed down (particularly with system tray apps).
|
| |
|
|
| |
properly for some reason on FreeBSD11, and using the QScreen->grabWindow(ID) routine *does* work properly for some reason, so just use the Qt function instead from now on (maybe it has better fallback routines?).
|
| |
|
|
| |
fix FreeBSD 11, but it is a lot more reliable now.
|
| |
|
|
| |
of XLib. This bahaves exactly the same on my FreeBSD 10.x system - still need to test a FreeBSD 11.x system.
|
| |
|
|
| |
the EWMH standards (_NET_WM_NAME, etc..)
|
| |
|
|
| |
for non-valid *.desktop files.
|
| |
|
|
|
|
|
| |
(system tray) to use the new routine.
Also setup the desktop to use the new XCB routine for window icons as well.
Both are tested on 10.x and appear to work perfectly.
|
| |
|
|
| |
QIcon (untested)
|
| |
|
|
| |
"seen" by fluxbox when a window is maximized, and the task manager works with the highly asynchronous XCB events.
|
| |
|
|
| |
with fluxbox maximizing windows underneath panels at the moment.
|
| |
|
|
|
| |
1) Add ability for task manager to minimize/maximize window from details menu
2) Add quick check/movement of new windows to make sure they are not underneath any panels
|
| |
|
|
| |
and fix a stray seg fault in the XCB window class detection routine.
|
| |
|
|
| |
states are properly detected again.
|
| |
|
|
|
|
| |
will be aware of the new status change (send an event instead of changing the property directly).
Also add a quick fix for system tray icons: if the icon size does not match the tray size, send a resize event and redraw a few moments later.
|
| |
|
|
| |
It is functional at the moment, but still has a few rough edges with regards to the X11 background interface (due to the move from XLib to XCB in Qt5). This reulst in some of the window manager interactions not behaving properly (such as sticky status on panels).
|
