has caused a regression (segfault) in some cases. (LP: #958540)

Alan Griffiths authored Mar 19, 2012 and Daniel van Vugt committed Mar 19, 2012

(LP: #808007)

r3050, and looks like it might be contributing to incorrect tap detection
in some cases with empty workspaces (LP: #953089)

r3050, and looks like it might be contributing to incorrect tap detection
in some cases with empty workspaces (LP: #953089)

has caused a regression (segfault) in some cases. (LP: #958540)

(LP: #945816)

More accurately; Changing any hotkey whose default or previous value was
a single modifier key would disable all other combinations with that modifier.
This was a regression caused by lp:compiz-core r3050.

Łukasz 'sil2100' Zemczak authored Mar 15, 2012 and Daniel van Vugt committed Mar 15, 2012

environment. Modified a few tests to use xvfb.

(LP: #945816)

More accurately; Changing any hotkey whose default or previous value was
a single modifier key would disable all other combinations with that modifier.
This was a regression caused by lp:compiz-core r3050.

Corrected CMake target dependencies so that parallel builds now work, as in "make -j8" (LP: #938417)

as in "make -j8" (LP: #938417)

Indeed, the script doesn't use anything bash specific - we can use sh instead. Removed a duplicated set of arguments.

As noted by Daniel, we were lacking a check if Xvfb is installed. Added one, informing the user when Xvfb is not installed.

Alan Griffiths authored Mar 12, 2012 and Daniel van Vugt committed Mar 12, 2012

/1/ it contained two copies of CompSize (a member and a base);
/2/ it allowed non-const use of base CompSize member functions (which breaks
    invariants);
/3/ it could be copied/assigned leading to double-delete of mData; and,
/4/ it could be "sliced" leading to leak of mData.

None of the core code did any of 2-4, but they'll no longer compile.

Eliminated the timing hack. Instead use individual passive grabs on
<modifier>+each_key to detect whether <modifier> is being tapped, or used
to modify another key.

Eliminated the timing hack. Instead use individual passive grabs on
<modifier>+each_key to detect whether <modifier> is being tapped, or used
to modifier another key.