The old code was connecting the KIconLoader signal to implicitWidth/HeightChanged
but never actually set a new implicit size.
Also, while at it use "setImplicitSize" - it's also marked internal but public
like setSize and uses qreal instead of QSizeF.
Differential Revision: https://phabricator.kde.org/D4011
numerous icon tests expect the actual production plasma theme to be present
it won't however be resolved because we disable qstandardpaths' default
fallbacks by switching it into testing mode. this ultimately means the
test run needs to have XDG_DATA_DIRS in the environment, if it doesn't
test data will not be present
furthermore QSKIP tests that depend on:
- hicolor icon theme to be present
- production plasma themes
this allows the entire test to pass in less than optimal environments by
means of skipping but also gives a reasonable hint as to why tests are
skipped in the first place. also moving forward if a test isn't guarded
we still have the warning about failure potential allowing someone to
quickly figure out why things failed
REVIEW: 127764
We have a lot of reusable singletons that are just hidden when not needed, such
as the OSD. IconItem, however, will always fade from the previous state, even if
the previous action happened minutes ago.
This fixes for example changing the volume and then later changing the brightness
where you could see the OSD animating from the volume icon.
REVIEW: 122859
Changing the color group of an Svg emits repaintNeeded which will then
cause it to schedule a pixmap update anyway. Includes a unittest.
REVIEW: 127313
source should always return the same value as it was assigned.
Also removed duplicated code (empty sourceString is now handled
as if the source couldn't be converted to any type).
REVIEW: 127282
Add some tests for IconItem.
Also found one issue in animation when quickly changing source, first frame
will be rendered with wrong icon.
REVIEW: 127103
