Summary:
When `PLASMA_USE_QT_SCALING=1` is set, Plasma uses native Qt scaling. This works fine for integer scale factors, and fixes a lot of bugs (see [[https://bugs.kde.org/show_bug.cgi?id=356446|Bug 356446]]) but it introduces a new one: with non-integer scale factors, text becomes blurry and pixellated because of a bug in `Text.NativeRendering`: https://bugreports.qt.io/browse/QTBUG-67007
QQC2-desktop-style forces the use of `Text.QtRendering` rendering for non-integer scale factors, successfully working around the problem. But PlasmaComponents QML objects don't implement the same workaround, so we see the issue in Plasma. This patch fixes that, and gets us one step closer to being able to use Qt scaling in Plasmashell.
There is no effect when `PLASMA_USE_QT_SCALING=1` is not being used.
FIXED-IN 5.13
BUG: 391691
BUG: 384031
CCBUG: 386216
CCBUG: 391695
CCBUG: 391694
CCBUG: 385547
CCBUG: 391692
CCBUG: 356446
Test Plan:
Before: `PLASMA_USE_QT_SCALING=1` set, 1.2 scale factor: Plasma text looks awful:
{F5749797}
After: `PLASMA_USE_QT_SCALING=1` set, 1.2 scale factor: Plasma text looks amazing!
{F5749798}
Note that we still get sub-pixel anti-aliasing and good kerning. There appear to be no layout regressions.
Without both `PLASMA_USE_QT_SCALING=1` and a non-integer scale factor set, there is no visual change compared to the status quo.
Reviewers: #plasma, davidedmundson
Reviewed By: #plasma, davidedmundson
Subscribers: mart, broulik, #frameworks
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D11244
Summary:
there were some icons which conveyed meaning solely by color
which is discouraged by the hig
https://community.kde.org/KDE_Visual_Design_Group/HIG/Color
this makes use of the little x corner icons that are already
used somewhere in breeze (like muted microphone)
the exception is wireless off that has a bar, so a stronger
this is disabled message than wireless-disconnected
Depends on D10243
...which should go in asap
Test Plan: they all look readable at 100% zoom
Reviewers: #plasma, #vdg, andreask
Reviewed By: andreask
Subscribers: andreask, #frameworks
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D11390
I don't see why this would be silently discarded.
The input shape has nothing to do with the window mask (e.g. rounded border depending on theme).
Differential Revision: https://phabricator.kde.org/D11102
KDirWatch::self() is a singleton so we would get change events also for other files like plasma-appletsrc needlessly reloading plasmarc all the time.
Differential Revision: https://phabricator.kde.org/D11326
Summary:
Using a private flag just added for plasma_install_package, until
there is a new official replacement for kpackage_install_package for
its needs
Test Plan: kdeplasma-addons & Co still build and now have clean cmake log again.
Reviewers: mart, apol
Reviewed By: mart
Subscribers: #frameworks
Tags: #frameworks, #plasma
Differential Revision: https://phabricator.kde.org/D11130
Summary:
We were inconsistent on that and it can get in way with placement on buttons.
Mirrors a change in Breeze icons.
{F5740901}
Reviewers: #frameworks, #plasma, ngraham
Reviewed By: #frameworks, ngraham
Subscribers: ngraham, mart
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D11050
Summary:
Previously IconItem would create icons in new textures for animating. By
keeping them in the atlas not only do we avoid that, but we should
implicitly share the QSGTexture used for non-animating item.
Test Plan:
Moused over icon, it changed correctly
Set core profile, moused over an icon it changed correctly
Reviewers: #plasma
Subscribers: #frameworks
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D11091
Summary:
Currently it asserts. I'm sure it worked at some point.
Fixed to match shaders shipped with Qt.
Test Plan:
Set Core profile
Did crash
Now doesn't
Reviewers: #plasma
Subscribers: #frameworks
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D11090
Summary: The PlasmaComponents TextField makes placeholder text light enough so that it doesn't compete with other UI elements, but the PC3 version does not. This patch fixes that.
Test Plan:
Before:
{F5734041}
After:
{F5734042}
Reviewers: mart, hein, #plasma
Reviewed By: mart, #plasma
Subscribers: #frameworks
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D10922
Summary:
If it didn't make sense to have Headings be 20% transparent because nearby or child textual content would be darker (see D10899), then it //really// doesn't make sense to have Titles be 20% transparent, because then **all** the text is darker than the titles, which is really odd.
As with D10899, this brings Plasma in line with Kirigami, where the large titles are already 100% opaque.
Test Plan:
Widget Explorer title, before:
{F5732725}
Widget Explorer title, after:
{F5732726}
As with D10899, it's a very subtle change, so I don't expect anyone to //consciously// notice. But people's unconscious brains will probably say , "wow, text looks sharper and more readable now!", though they won't be able to articulate why.
Reviewers: #plasma, mart
Reviewed By: #plasma, mart
Subscribers: #frameworks
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D10902
Summary:
We [[https://cgit.kde.org/kirigami.git/commit/?id=95417dc1f9bea79cabffb95fa0a79621167a6106|recently removed Kirigami Headings' slight transparency]], after concluding that it looked inappropriate to have Headings with lighter text than their textual content. `PlasmaExtras.Heading` does the same thing, so we should make the same change here, both based on the inherent merits of the change, and also to maintain consistency with Kirigami.
This also slightly improves matters for people who complain about poor text contrast throughout Plasma.
Test Plan:
Browsed System Settings; headings are now ever so slightly darker (or lighter, for users of Dark themes)
Also, here are some befores-and-afters for a widget I'm working on that uses Headings. Before:
{F5732641}
After:
{F5732643}
It's subtle; you'll need to flip between the two with the arrow keys to see the difference, and may need to also zoom in ({key meta +})
Reviewers: #plasma, mart
Reviewed By: #plasma, mart
Subscribers: #frameworks
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D10899
Summary:
automoc by itself can only detect metadata files referenced by direct
usage of the Q_PLUGIN_METADATA macro. It does not do any C++ preprocessor
evaluation. Instead it needs to be helped with regexp-based filter rules
for detecting any names of files used as additional moc input.
See docs for AUTOMOC_DEPEND_FILTERS for further details.
In the near future all the boilerplate code should be replaced with a yet to
be designed ECM macro. For now explicit code is used to collect use cases
for the macro and still to fix the issue already.
Test Plan:
Changing a JSON file (or for those generated from .desktop files
changing that one) and running make results in the related *.moc file
being regenerated and the related object file being recompiled.
qtplugininfo shows that the created plugin binary has up-to-date JSON
content.
Reviewers: #frameworks, #build_system, apol
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D10732
When an applet is placed on the desktop and already shows its full representation, like Weather applet,
don't create the expander item.
The DefaultCompactApplet shows the dialog when plasmoid.expanded which will be always the case and you end up with an empty
window randomly showing up on startup.
Differential Revision: https://phabricator.kde.org/D10764
Summary:
make svgitem schedule a polish right when it's done loading
ScrollView now forces to layout items as soon is created, to preload items
before the window is shown
remove useless include
Test Plan:
possible to preload (almost) the whole systray with less hacks
on that part (see D10692)
Reviewers: #plasma, davidedmundson, broulik
Reviewed By: #plasma, broulik
Subscribers: davidedmundson, broulik, plasma-devel, #frameworks
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D10722
Summary: this is already done by KDEFrameworkCompilerSettings so it's just duplicated now
Test Plan: still builds and runs
Reviewers: #plasma, kossebau
Reviewed By: kossebau
Subscribers: kossebau, plasma-devel, #frameworks
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D10748
Summary:
This changes between linear and nearest neighbour filtering.
Docs for QtQuick say:
In Qt Quick 2, this property has minimal impact on performance.
Generally we round to icon size anyway and resize the underlying images
anyway, so this will have no impact.
Reviewers: #plasma, apol
Reviewed By: apol
Subscribers: plasma-devel, #frameworks
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D10653
Summary:
when preloading, preload the expander as well, and reparent the
full representation to it, so we have a scene schanged less.
also do a create() of that window (may be worth to do create()
automatically in every Dialog?)
Test Plan: the first open is visibly a lot faster now
Reviewers: #plasma, davidedmundson
Reviewed By: #plasma, davidedmundson
Subscribers: davidedmundson, plasma-devel, #frameworks
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D10681
Summary:
-- when the combobox isnt inside the main Item
but rather a sub-Item it doesnt inherit the
LayoutMirroring properties. This is a workaround
to handle that behavior.
QTBUG: https://bugreports.qt.io/browse/QTBUG-66446
Test Plan:
checkout that LayoutMirroring works properly in
such case
Reviewers: #plasma, mart, safaalfulaij
Reviewed By: safaalfulaij
Subscribers: safaalfulaij, plasma-devel, #frameworks
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D10592
Summary:
preload popups of some applets after init in the background
based on a value of X-Plasma-PreloadWeight in the desktop file
if present, otherwise some default values based on the applet
type (Provides)
Save the weight in the config, if an applet is never opened,
slowly decrease the weight, when it reaches 0 don't preload it
next start, increase every time it gets opened, so at the moment
it's quite aggressive about preloading, in order to not do it
a lot of plasma startups without touching the applet are needed
Applet with a very big weigth will be preloaded immediately,
therefore having an impact on the time it will take to have
a panel visible and usable, while lesser weigths will preload
after a random number of seconds between 2 and 10, so will load
in the background after everything is started
Test Plan:
Plasma starts up correctly, applets load correctly and can be added
correctly both those expanded or collapsed.
plasmashell appears correctly usable without too big hiccups even
while it's loading things in the background
some numbers:
without preloading, plasma takes around 64 mb of memory after startup
when preloading everything about 94, so it's a cost of about 30 mb
which is not negligible.
don't have precise timing, but if everything gets preloaded immediately,
the time to get an usable desktop appears to be at least doubled,
while the delayed preloading (except just a couple of applets) doesn't
seem to have a big impact on the time needed to get an usable desktop
Reviewers: #plasma
Subscribers: davidedmundson, broulik, apol, ngraham, plasma-devel, #frameworks
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D10512
Ensures the fading node's texture has the same filtering as the regular node,
otherwise during transition the icon could show scaling artefacts.
Differential Revision: https://phabricator.kde.org/D10533
Summary:
On wayland (at least) it's initialized when the window is assigned, so
we need to listen to the itemChanged to adopt it.
Test Plan: manual testing
Reviewers: #plasma, hein
Reviewed By: #plasma, hein
Subscribers: hein, plasma-devel, #frameworks
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D10420
Summary: Much like it is specified by kirigami itself
Test Plan: Now I don't get a warning
Reviewers: #plasma, mart
Reviewed By: #plasma, mart
Subscribers: plasma-devel, #frameworks
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D10225
Check that the sizes of the color channels in the fbconfig match those
in the window visual.
This fixes a 2-10-10-10 fbconfig being choosen for the ARGB32 visual.
Differential Revision: https://phabricator.kde.org/D10173
Signed-off-by: Fredrik Höglund <fredrik@kde.org>
Acked-by: David Edmundson <kde@davidedmundson.co.uk>
Apparently causes background contrast to not work on startup, let's just revert this
and figure out a proper solution later.
BUG: 389352
This reverts commit f3cd5d55fa.
Unset the Label's default height which completely confuses the renderer either shifting the item vertically (e.g. when toggling
system tray popup), or letting it get too wide (e.g. when moving from a two-line tooltip to a one-line tooltip).
Also let the Layouts figure out sizing rather than trying to overrule it with flaky values.
BUG: 386260
BUG: 389371
Differential Revision: https://phabricator.kde.org/D10077
* Add QStringLiteral et al where appropriate
* Also add override keyword to avoid warnings
* Minor style changes and cleanups
Differential Revision: https://phabricator.kde.org/D10092
