Summary:
Before icons loaded internally with QIcon::fromTheme were being colored with the
colors from the current global color scheme instead of the ones from the current
Plasma Theme. Leading to visual bugs when the two differ. This happened because
KIconLoader uses the global color scheme by default.
A prominent case is the notification send by the network plasmoid when one
successfully connected to a wireless network. It sets the icon
"network-wireless-on" which is not included in Breeze icons (but is included in
Breeze Plasma Theme). If the current icon theme is indeed Breeze, IconItem
resorts to using QIcon::fromTheme and we end up with a wrong colored
"network-wireless" icon.
BUG: 417780
Test Plan:
{F8125752}
{F8125753}
Reviewers: #plasma, cblack, ngraham, mart
Reviewed By: #plasma, cblack, ngraham, mart
Subscribers: mart, wbauer, cblack, kde-frameworks-devel
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D27589
Summary:
Right now the widget tooltip's icon/image is top-aligned, which in practice makes it
vertically centered when the tooltip has only one line of text, but not vertically
centered when there are two lines of text.
This patch makes it vertally centered all the time, improving the appearance.
Test Plan:
Before: {F8095695}
After: {F8095696}
{F8095697}
Reviewers: #vdg, #plasma, filipf, ndavis
Reviewed By: #vdg, filipf, ndavis
Subscribers: kde-frameworks-devel
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D27299
Summary:
move all the logic in an attached property QObject
it's a bit less heavy to create so should make things slightly better
when this is used in delegates
Test Plan: Plasmoid background removal (with colors inversion) still works
Reviewers: bshah, #plasma
Subscribers: apol, kde-frameworks-devel
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D27200
Summary:
KDeclarative::setupBindings() is deprecated, replaced calls with its
implementation (setupContext + static setupEngine call), but see comment
about whether we can be sure this is done only once per engine, as
requested by the setupEngine API. It's 100% equivalent to the old code
though, can't be worse.
KPackage::contentsHash is deprecated, replaced call with implementation.
Depends on D25433 for the KActionCollection::changed signal.
Test Plan: Builds
Reviewers: mart, davidedmundson, broulik, apol
Reviewed By: apol
Subscribers: apol, kde-frameworks-devel
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D25434
Summary:
When we destroy a colorscope, all children will try to reparent. This patch
makes it so they never pick an object if it's being removed.
Test Plan:
The logout screen greeter and kscreenlocker doesn't crash anymore
upon close.
Reviewers: #plasma, mart
Reviewed By: #plasma, mart
Subscribers: kde-frameworks-devel
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D25953
Otherwise colorScope doesn't attach to things like Plasma.Svg.
Also remove superfluous cast as parentItem() returns a QQuickItem *.
Differential Revision: https://phabricator.kde.org/D25887
Summary:
add the ability for the user to override the background hints anddecide
about applets having background or a shadow
A new handle button will be added to manually enable/disable background for
plasmoids that support it
if they don't, the ImmutableBackground flag is set in the hints.
PlasmaCore.ColorScope has been expanded to work more like Kirigami.Theme (which will be replaced by) even if it's oòld behavior still works. This makes the complementary colorscope in applets with shadow actually work (and at the same time gives a clearer porting path for kf6)
Test Plan: Tested the functionality and correct save/restore
Reviewers: #plasma, davidedmundson
Reviewed By: #plasma, davidedmundson
Subscribers: davidedmundson, ndavis, kde-frameworks-devel
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D25590
Test Plan: Used in model
Reviewers: #plasma, broulik
Reviewed By: #plasma, broulik
Subscribers: broulik, kde-frameworks-devel
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D21419
Summary:
Needed e.g. to fix the missing update of the mask on changing the size of
the Plasma panel if only listening to mask changes.
Reviewers: #plasma, apol
Reviewed By: apol
Subscribers: kde-frameworks-devel
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D21201
Summary:
In that method we can assue a high chance that the mask is also affected.
Instead of adding complicated additional code to track if the mask
really changed, we rely on the cached value from FrameSvg.
This method is not too often called in a Plasma session, so that
further optimization work to avoid unneccesary emits seems not yet
worth it.
Reviewers: #plasma, davidedmundson
Reviewed By: #plasma, davidedmundson
Subscribers: kde-frameworks-devel
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D21179
Just because the item is visible doesn't mean the window itself is. Keep track of the window it's in.
Use itemChange instead of connects and move the check for starting to startRedirect so we don't have to check the conditions
for that in every change handler (visible && enabled && window->visible). It also saves three connects in the constructor.
Also, don't unredirect if we didn't redirect in the first place to avoid warnings printed on console.
BUG: 406303
Differential Revision: https://phabricator.kde.org/D20805
Summary:
When just the smooth property is changed, this also results in a call of
QQuickItem::updatePaintNode reimplementations. So always just set the
filtering flag to match the current value of the smooth property.
Reviewers: #plasma, broulik, mart
Reviewed By: #plasma, broulik
Subscribers: kde-frameworks-devel
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D20579
Summary:
QQuickItem already has a "smooth" property, which was shadowed by the
custom ones with the same name. As the property was not part of public
symbols, but instead is dynamically looked up without the class name part
of the property, we just remove the custom versions and instead switch
all custom code to use the inherited property.
While doing that the code is also fixed to properly update the
textureNodes to the current value of the "smooth" property.
As well as not set the filtering only on the texture, when it is passed
to a texturenode which will set its own filtering state to the texture
right before the bind call and thus wipe any previous direct setting
of the filter mode on the texture.
This also changed the default to smooth=true for SvgItem, though
effectively smooth was always true, as the texturenode was hardcoded
to get a QSGTexture::Linear filtering (which, as said above, is forced
onto its texture, no matter what was otherwise set before).
Reviewers: #plasma, davidedmundson, mart
Reviewed By: #plasma, davidedmundson
Subscribers: kde-frameworks-devel
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D20510
Summary:
This property allows e.g. the Plasma shell to query the shape a panel has
when defining the area for BlurBehind.
Reviewers: #plasma, mart
Reviewed By: #plasma, mart
Subscribers: kde-frameworks-devel
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D20202
Summary:
FrameSvg updates the margins e.g. on a Theme change already in its
internal event handling before the methods of FrameSvgItem are invoked.
Due to that CheckMarginsChange guards before this change could not detect
any changes, and thus missed to trigger an update via the margin objects
to the QtQuick world.
So when changing the theme e.g. in a Plasma shell or in the Plasma Theme
Explorer, existing FrameSvg items are using outdated margins, resulting
in broken display.
By keeping a copy of the last seen margins as part of the item and not only
per method, changes outside the FrameSvgItem items can be properly detected.
Test Plan:
Switch Plasma themes between Air, Breeze & Oxygen, see how different margins
are reflected in the widgets sizes instantly.
Switch Plasma themes in the Plasma Theme Explorer, with "Show Margins"
enabled, see how margins are always correctly updated intead of taking the
value of the last theme used.
Reviewers: #plasma, mart
Reviewed By: #plasma, mart
Subscribers: kde-frameworks-devel
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D19743
Summary:
The old logic results in broken rendering if the sample was bigger than the
area to render, due to normalized texture rect being 0 in the respective
dimansion. As well as stretched mapping of the sample for fractional
relationships between the area to render and the sample size.
Test Plan:
Borders of backgrounds of FluffyBunny theme render properly, no other
regressions seen (but unsure what else might rely on that artifact).
Reviewers: #plasma, mart
Reviewed By: #plasma, mart
Subscribers: apol, sitter, kde-frameworks-devel
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D19787
When changing an IconItem from using a local image (e.g. PNG file) to using a Plasma themed icon,
the implicit size wasn't updated since the old image was still loaded in m_imageIcon.
Clearing it when clearing the other icon stores fixes this.
BUG: 405298
Differential Revision: https://phabricator.kde.org/D19674
Plasma Dialog used to change the item's visiblity but this is no longer the case, so it would animate when a window is shown and
the source has changed in the meantime.
Differential Revision: https://phabricator.kde.org/D19421
Especially since the Theme isn't modified in any way but just used to read some data.
While the Private part of Theme is already shared, creating a Theme instance still has some setup cost.
Differential Revision: https://phabricator.kde.org/D18149
