We have a special case for sources starting with "file://" but a "/" also represents an absolute path and shouldn't
conflict with icon theme names.
Kicker sets a custom image as local path and then we would end up trying to load it as a QIcon::fromTheme eventually.
This will cause the implicit size of the icon item to stay at its default as we only check a custom implicit size
for a source QImage or an SVG. Moreover, this potentially introduces scaling artefacts.
Differential Revision: https://phabricator.kde.org/D9812
Summary:
--PlasmaComponents3.ComboBox now appears correctly
its list items in RTL environments by aligning them
to the right
BUG: 387558
Test Plan:
check the combobox list both in RTL and LTR environments
in order to confirm that the list items are aligned correctly
Reviewers: #plasma, mart
Reviewed By: #plasma, mart
Subscribers: mart, plasma-devel, #frameworks
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D9275
This adds a dedicated, green, task bar progress graphic.
Currently there isn't and as a fallback it uses the hover graphic.
BUG: 368215
Differential Revision: https://phabricator.kde.org/D9324
Summary:
minimize file access during url rewrite
we still need it for accessing stuff under code/
Test Plan: plasma still loads, seems to save ~2%
Reviewers: #plasma, davidedmundson
Reviewed By: #plasma, davidedmundson
Subscribers: davidedmundson, ngraham, apol, plasma-devel, #frameworks
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D9124
This reverts commit 4ce50cc61f.
This caused a regression with toggling compositing, causing SVG reloads
with the same path to switch to opague elements. As this revision was
meant as a performance optimization, it can be reverted with no other
consequence for now.
Summary:
like other platforms, on mobile devices make scrollbars visible only when
dragging/flicking, behavior on desktop unchanged
it introduces a runtime dependency on kirigami, but was already kinda there
but broken as the Settings singleton was already used.
if this can't be done, i'll add something in plasmacore which reads as well the
QT_QUICK_CONTROLS_MOBILE environment variable
Test Plan: scrollbar disappearing with the variable set, normal behavior otherwise
Reviewers: #plasma, bshah, davidedmundson
Reviewed By: #plasma, davidedmundson
Subscribers: davidedmundson, plasma-devel, #frameworks
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D9287
Summary:
to minimize metadata lookups, pass the kpackage created in pluginloader
to the applet itself as args. the args structure will be used only by the
internal constructor and applet sublasses will get their args unaltered
Test Plan:
plasma loads correctly, from 3 kpackage instantiation per applet to 2
(last is in the access manager)
Reviewers: #plasma, davidedmundson
Reviewed By: #plasma, davidedmundson
Subscribers: davidedmundson, plasma-devel, #frameworks
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D9123
The style is global per engine, there's no need to set it for every applet created
as the engines are shared.
Since this entire thing is just a hack, just setting a dynamic property to identify
that we've set a style is valid imho.
Also turn it into a plain QtObject since we don't need a fully-fledged Item.
Differential Revision: https://phabricator.kde.org/D4362
Summary:
Setting a window icon is costly enough to be worth avoiding when it's
not actually needed.
This is a resurrection of David's old patch:
https://git.reviewboard.kde.org/r/128484/
The concerns in the old discussion that led to its rejection are no
longer valid today: In the meantime we implemented a way for apps to
announce their .desktop file in a window hint, which KWin (and libtm)
will fall back to to look up an icon if not set. As plasmashell does
this, we can drop the setIcon call here and won't regress.
Test Plan:
The pinned systray popup dialog still gets the Plasma icon in the
Present Windows effect.
Reviewers: #plasma, #frameworks, davidedmundson, graesslin
Subscribers: plasma-devel
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D9209
Summary:
--align selected text correctly for
PlasmaComponents2.ComboBox taking into account
RTL.
BUG: 387415
Test Plan:
use ComboBoxes in RTL and LTR in order
to check a proper behavior
Reviewers: #plasma, davidedmundson
Reviewed By: #plasma, davidedmundson
Subscribers: broulik, plasma-devel, #frameworks
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D9126
Summary:
Plasma::Svg initializes SvgPrivate::scaleFactor to 1.0. On systems with
a higher value for Units:devicePixelRatio, that means Svg::setScaleFactor
gets called with a different value right after instanciation and a fair
amount of work is done twice.
This patch introduces a static SvgPrivate::s_lastScaleFactor,
initialized to 1.0 but updated in Svg::setScaleFactor.
SvgPrivate::scaleFactor is then initialized to the static every time
its instanciated. That means after the first time Svg::setScaleFactor
is called with e.g. 2.0, subsequent Svg instances get initialized
with 2.0.
This ends up shaving a decent amount of time off of startup on scaled
systems.
Reviewers: #plasma, davidedmundson, broulik
Subscribers: plasma-devel, #frameworks
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D9119
Summary:
When the component completes `FrameSvg::setRepaintBlocked(false)`
is called, which calls `updateFrameData`, which emits
`repaintNeeded`, which is connected to `doUpdate` anyway.
This saves one `doUpdate` per instanciation. The reason for that
is as follows:
* The concept or repaint blocking is implemented in `FrameSvg`,
which inherits from `Svg`.
* The `repaintNeeded` signal connected to `doUpdate` is the one
in `Svg`.
* `Svg` code emits `repaintNeeded` liberally even when repaints
are blocked, since it doesn't know anything about this.
Hence we check if repaints are blocked in doUpdate and abort,
implementing the QQmlParserStatus interface in the right class.
Reviewers: #plasma, davidedmundson, mart
Subscribers: plasma-devel, #frameworks
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D9138
Summary:
Plasma::Svg initializes SvgPrivate::scaleFactor to 1.0. On systems with
a higher value for Units:devicePixelRatio, that means Svg::setScaleFactor
gets called with a different value right after instanciation and a fair
amount of work is done twice.
This patch introduces a static SvgPrivate::s_lastScaleFactor,
initialized to 1.0 but updated in Svg::setScaleFactor.
SvgPrivate::scaleFactor is then initialized to the static every time
its instanciated. That means after the first time Svg::setScaleFactor
is called with e.g. 2.0, subsequent Svg instances get initialized
with 2.0.
This ends up shaving a decent amount of time off of startup on scaled
systems.
Reviewers: #plasma, davidedmundson, broulik
Subscribers: plasma-devel, #frameworks
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D9119
If a border isn't enabled there's no point in creating a node for itIf a border isn't enabled there's no point in creating a node for it.
Differential Revision: https://phabricator.kde.org/D9121
Summary:
During my plasmashell startup this saves time for 17 calls. Opening
something like Simple Menu saves another ~dozen.
It's worth investigating why Svg::setImagePath gets called repeatedly
with the same args to begin with, but this is worth doing anyway.
Reviewers: #plasma, davidedmundson, mart
Subscribers: plasma-devel, #frameworks
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D9115
There's literally just one place where this thing changes so just emit the signal
there and call it a day.
Differential Revision: https://phabricator.kde.org/D9112
Summary:
At the moment whenever something changed we were emitting colorGroupChanged
and then every color would recompute. This would end up being emitted
over 10 times at plasma startup so far.
This patch makes sure that the property will only be emitted if the color
group actually changes.
Test Plan:
Ran plasma, didn't notice issues.
I don't see all of the changes on the property anymore
Reviewers: #plasma, mart, davidedmundson
Reviewed By: #plasma, mart
Subscribers: davidedmundson, plasma-devel, #frameworks
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D8917
Summary:
ListView.atYEnd is notoriously buggy and sometimes will cause indicators
to blink in and out of existence where not needed.
Reviewers: #plasma, mart
Subscribers: plasma-devel, #frameworks
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D9089
We know we have a QQuickItem so we can just use its methods directly and then also use setSize
which does both at once, saving intermediate signal emission and potential re-layouting
Differential Revision: https://phabricator.kde.org/D9068
This ensures the menu can have nice round corners with Breeze and Oxygen. The same had already been done in ContainmentInterface's menu.
CHANGELOG: Fixed an issue where PlasmaComponents Menu would appear with broken corners
CCBUG: 381799
Differential Revision: https://phabricator.kde.org/D8965
This would create the surface after which this flag has no effect causing menus to have black rounded corners.
CHANGELOG: Fixed an issue where context menus would appear with broken corners
CCBUG: 381799
Differential Revision: https://phabricator.kde.org/D8963
Summary: Using the height property together with paintedHeight had binding loops.
Reviewers: #plasma, mart
Reviewed By: #plasma, mart
Subscribers: broulik, plasma-devel, #frameworks
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D8870
Summary:
We were reporting that the margins were changing whenever anything changed
in a FrameSvg, this patch makes sure we are only reporting the changes
when it actually changes.
This also fixes a binding loop in Desktop.qml from plasma-desktop.
BUG: 382233
Reviewers: #plasma, davidedmundson
Reviewed By: #plasma, davidedmundson
Subscribers: broulik, davidedmundson, plasma-devel, #frameworks
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D8809
frame->fooHeight is the same as frame->fooFixedHeight if the border is enabled, we only meddle
with margins, not heights. Thus we can just do the lookup once and re-use the value.
Also store the hintFooMargin instead of doing the same exact lookup once for fixed
and once for non-fixed margins.
Differential Revision: https://phabricator.kde.org/D8135
Summary:
The OSDs are positioned by the Wayland compositor. This fixed the issue
found in D8268.
Reviewers: #frameworks, #plasma
Subscribers: plasma-devel
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D8304
Summary:
Module internal types are not implicitly available anymore simply by placing
them inside the module's main directory. It is now necessary to mention them
as internal types in the qmldir file.
See qtdeclarative commits 22a2cc43387ec3b9f74a6c01f8665378a4541147 (introduced
the change) and 930aea8b9ca59a24838cf7f279653e3b2ee40cee (explanation)
Test Plan:
Built whole stack with Qt 5.10. The digital clock plasmoid does not
load due to the DaysCalendar type not being available. With these changes to
qmldir and a reload of the plasmoid it works.
Reviewers: #plasma, broulik
Reviewed By: #plasma, broulik
Subscribers: broulik, #frameworks, plasma-devel
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D8077
Documentation says "Children are stored internally as a linked list, so iterating over the children via the index is suboptimal."
Differential Revision: https://phabricator.kde.org/D8046
For a panel it makes sense to have panel options easily available on every applet
but on the desktop you typically have some free space to access the containment actions.
This is especially confusing if you have a Folder View on your Folder View desktop
and end up with a "Folder View Options" menu that affects the desktop behind it rather than the applet.
Differential Revision: https://phabricator.kde.org/D8009
Summary:
Plasma code is full of Label {opacity: 0.6}. We shouldn't be hardcoding
style into a UI, especially with magic numbers.
This adds a new component that does the same thing but will allow
plasmoids to write more semantic code.
Reviewers: #plasma, hein
Reviewed By: #plasma, hein
Subscribers: hein, sebas, broulik, plasma-devel, #frameworks
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D4831
Summary:
Software rendering (at time of writing Qt5.10) doesn't seem to like our
tiling/stretching in the 9-tiles.
Also when using QPainter it's arguably faster to create and cache
pixmaps of the whole frame, which is what the slow path of FrameSVGItem
does.
Test Plan: Ran FrameSVG test then plasma shell with scene device set
Reviewers: #plasma, broulik
Reviewed By: #plasma, broulik
Subscribers: plasma-devel, #frameworks
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D8013
Summary:
IconItem uses a gorgeous shader to fade between two states. This doesn't
work with software rendering.
Given software rendering is a second class citizen, just skip the
animation.
Test Plan:
Hovered over kickoff.
Went from one to the other, without temporarily disappearing
Reviewers: #plasma, broulik
Reviewed By: #plasma, broulik
Subscribers: plasma-devel, #frameworks
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D8014
Summary:
This draws a line underneath a Plasma checkbox label when it has
keyboard focus. This is needed as when an item is checked, the focus
shadow is practically invisible.
Visuals when clicking is completely unchanged
This brings it more in line with our QStyle.
Test Plan: Attached runtime tests
Reviewers: #plasma, #vdg, mart
Reviewed By: #plasma, mart
Subscribers: plasma-devel, #frameworks
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D7819
The "parent" used to be resolved by the place where the utils.js was imported from.
However, it was changed into a singleton and now this didn't work.
BUG: 384776
Summary:
The name (with a correct role) is already exported as a tab
By having the static text internally we get the name twice when we do a
flat review of the controls.
Reviewers: #plasma
Subscribers: plasma-devel, #frameworks
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D7815
this was actually breaking keyboard navigation
see https://bugreports.qt.io/browse/QTBUG-63176
for our use case in pasmashell, it's better to just
rely on focus of the internal items or of the
ListView
reviewed-by: davidedmundson
Summary: Some widgets are missed up in RTL mode, this should fix it.
Reviewers: #plasma, #frameworks, mart
Reviewed By: #plasma, mart
Subscribers: mart, broulik, davidedmundson, plasma-devel
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D7715
The appendCategory(ConfigCategory*) allows to add a category created by e.g. an Instantiator
that can use bindings rather than fixed arguments.
The removeCategory methods allow to remove a category when e.g. the model in an Instantiator resets.
CCBUG: 372090
Differential Revision: https://phabricator.kde.org/D7602
We'll eventually delete the Plasma::Svg anyway if we failed to find an appropriate icon.
Just need to make sure we don't use isValid with the old source set but this is only done with
m_usesPlasmaTheme, hence the new check there, also optimizes the non-themed case.
Differential Revision: https://phabricator.kde.org/D7584
On my machine I end up creating 490 margins and 490 fixed margin objects.
After this change I only create 171 margin objects and 20 fixed margin objects.
Differential Revision: https://phabricator.kde.org/D7582
Summary:
using the qml url interceptor, rewrite imports in order
to load always the plasma styled qqc2 controls in plasmoids
and plasma views. As is per-engine, is possible to load
controls with qstyle theme in configuration dialogs
and plasma style in plasmoids.
Note: this replaces just the style org.kde.desktop
to Plasma and not every style (so with Material
set as QT_QUICK_CONTROLS_STYLE environment variable
plasma would load controls with material style) because
we can control both org.kde.desktop and Plasma styles:
they must have the same subdirectories as unlike files,
their existence is checked before the rewrite (in our case,
"private"), so in order to work they must exist in both
styles
Test Plan:
a qqc2 button in a plasmoid has the plasma style, in
a config dialog has the qwidget style. rewrite works
only from org.kde.desktop to Plasma and not from other
styles, due to the existence of the "private" subdirectory
Reviewers: #plasma, davidedmundson
Reviewed By: #plasma, davidedmundson
Subscribers: davidedmundson, plasma-devel, #frameworks
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D6964
Summary:
It's used by ToolButton et al.
We install it in the theme folder, but not in the plasma components dir.
Test Plan: Used a toolbutton without error
Reviewers: #plasma, mart
Reviewed By: #plasma, mart
Subscribers: plasma-devel, #frameworks
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D6693
Summary: createObjectFromSource can return a null on error. This needs guarding.
Reviewers: #plasma, mart
Reviewed By: #plasma, mart
Subscribers: plasma-devel, #frameworks
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D7532
I don't get the comment on why it is /dev/null instead of QString() but using SimpleConfig at least saves 0.3ms per call for me.
Differential Revision: https://phabricator.kde.org/D7510
It's not being set to anything anymore and people typically don't have 256 color displays these days.
Differential Revision: https://phabricator.kde.org/D7516
We're passing it an absolute path to a file in our install directory anyway,
so we probably don't want inheritance.
Differential Revision: https://phabricator.kde.org/D7362
Summary:
imagePath can be an absolute path into an iconTheme
-> Do not try to find it in the Plasma theme
imagePath can be empty
-> Do not try to look it up at all
Reviewers: #plasma, mart
Reviewed By: #plasma, mart
Subscribers: #frameworks
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D7250
Summary:
on multiscreen systems, the notification dialog gets resized to 0x0
right after componentcomplete by someone else (either kwin, notification
applet or Qt code) causing 382340, resizing a window to a null size is not correct anyways
and should never happen.
with this, all other resize events are correct and notifications are ok on multiscreen
systems
BUG:382340
Test Plan:
system with primary monitor and panel in the external monitor
at the right of a panel, notifications are fine
Reviewers: #plasma, davidedmundson
Reviewed By: #plasma, davidedmundson
Subscribers: davidedmundson, broulik, plasma-devel, #frameworks
Tags: #plasma, #frameworks
Differential Revision: https://phabricator.kde.org/D7127
Summary:
By default, Qt is built without accessibility there, which is what
would indirectly include the needed QColor with a "normal" Qt build
here.
Reviewers: #frameworks, broulik
Reviewed By: broulik
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D7117
The applet context menu contains containment actions which might be outdated if we don't call this.
Differential Revision: https://phabricator.kde.org/D6894