lp:~vanvugt/compiz-core/fix-alt-bugs-2

Branch merges

Propose for merging

No branches dependent on this one.

Merged into lp:compiz-core at revision 3048

Alan Griffiths: Approve on 2012-03-08

Diff: 68 lines (+16/-18)

2 files modified

src/event.cpp (+11/-17)
src/screen.cpp (+5/-1)

api

Related source package recipes

No recipes using this branch. (?)

Create packaging recipe

Related snap packages

Related bugs

Bug #806255: Unity/compiz intercepts Super and Alt keypresses from grabbed windows like VMs.	High	Triaged
Bug #925293: Plugins can't tell the difference between a modifier key-tap, and a modifier key-release (after being used to modify other keys)	High	Fix Released
Bug #943194: [regression] Pressing alt doesn't show the menu title bar in top panel	High	Fix Released

Link a bug report

Related blueprints

3047. By Daniel van Vugt on 2012-03-08

Improved the fix for LP: #943194 so that it now works with GTK-2 menus too.
The first attempt only worked with GTK-3 it seems.

This also required that the fix for LP: #806255 be redesigned to avoid
XGrabKeyboard completely. Because that was part of what broke GTK-2 menus.

3046. By Daniel van Vugt on 2012-03-07

Don't keep an active keyboard grab the whole time a shortcut key is held
down. Doing so was causing multiple bugs with Unity 5, which binds to
the Alt key. Instead use some simple heurisitics to decide if a key has
been "tapped" or not. (LP: #943194) (LP: #943851) (LP: #945373)

Yes, it is a compromise to go back to detecting taps based on timing.
However it is the only existing solution to all of these bugs related to
the Alt key.

3045. By Daniel van Vugt on 2012-03-07

Don't respond to hotkeys if some other window (like a VM or locked
screen saver) already has an active keyboard grab. (LP: #806255)

Also fixed a minor pointer mistake in isBound(), and removed a redundant
call to XFlush.

3044. By Daniel van Vugt on 2012-03-05

Avoid crashing if the decoration window is NULL. Not sure how to
reproduce the crash, but this should avoid it (LP: #937815)

3043. By Daniel van Vugt on 2012-03-05

Avoid pointlessly doing a full copy of a linked list every frame when the
copy is never changed and never different to the source. This cuts an
instant 6-9% from compiz' CPU usage (LP: #946887)

3042. By Daniel van Vugt on 2012-03-05

Don't load plugins if you failed to initialize the screen. That will
cause an assertion failure and probably other issues too. (LP: #946118)

3041. By Alan Griffiths on 2012-03-05

Started as cleaning up some pointless temporaries on the heap and ended by
getting rid of a pointless and dangerous CoreWindow class.

3040. By Daniel van Vugt on 2012-03-05

Let's call it 0.9.7.1 while in-between stable point releases .0 and .2

3039. By Sam Spilsbury on 2012-03-05

Backports rev 2988 from lp:~smspillaz/compiz-core/compiz-core.work_923683
which will prevent the shift race condition from occurring should core decide
to configure a window frame's left and top extents without configuring the
window itself (which can happen if, eg, the frame extents change, but the
window isn't actually moved.

3038. By Sam Spilsbury on 2012-03-05

Some tidy up, always replay the keyboard if something was grabbed and didn't
trigger an action and don't trigger actions which aren't added accidentally
(LP: #943612) (LP: #944631)