GitLab

GIT_SILENT

Might be easier to read this way.

GIT_SILENT

GIT_SILENT

GIT_SILENT

engine can be null, d->engine is guaranteed to not be null

The only remaining non-deprecated methods of KDeclarative are static, so
we no longer need an instance to access them.

The engine passed to QmlObject is potentially shared, either because it
comes from QmlObjectSharedEngine or because the caller is already using
it. Since raw pointers do not provide any information about that,
deprecate the raw pointer constructors and replace them with a
constructor taking a std::shared_ptr.

This allows QmlObject to properly track if its the first user of the
QQmlEngine and if so, call KDeclarative::setupEngine on it. It also
allows it to know if it is the *last* user of the engine, and in that
case properly cleanup the QNAM factory, without affecting the shared
engine case.

BUG: 451790

Not sure, but seems like they were stuck in a class' constructor purely
for historical reasons, guarded by ifdefs at some point even.

The label will be displayed in the tooltip or beside the full date.

Sub-labels are required to show alternate calendar in the calendar
applet.

Two new signals are added:

1. `subLabelReady`: Emitted when the plugin has loaded the sub-labels.
2. `alternateDateReady`: Emitted when the plugin has loaded the
   alternate dates.

`SubLabel` struct contains yearLabel, monthLabel and dayLabel, and the
plugin can specify the priority among Low, Default, High and Urgent for
a SubLabel.

CCBUG: 429882

We can add this here so we don't have to implement it a million times in
each individual KCM.

This is now done as part of the plasma integration plugin.

Task: https://phabricator.kde.org/T12135

BUG: 450507

As the docs (https://doc.qt.io/qt-5/qqmlengine.html#setNetworkAccessManagerFactory) state, the engine does not take ownership
of the factory. Consequently we have to clean it up ourselves.

These functions had divergent code, and openInfoCenter was not working
properly because it was following the pattern set by openSystemSettings
which did not work because System Settings' desktop file does not follow
rDNS style naming, but other ones do.

BUG: 443984
FIXED-IN: 5.92

when there is an header on the kcm, show the separator at the proper
position, ecountering perfectly the scrollbar separator

Also show them as soon there is a scrollbar, as the scrollbar separator
looks bad when floating

Before:

![Screenshot_20220204_110037](/uploads/de468d039eb50735003395347e91f703/Screenshot_20220204_110037.png)

After:

![Screenshot_20220204_105935](/uploads/4eb458a766b84f1306a58a1724b28a69/Screenshot_20220204_105935.png)

This property should always be of type QKeySequence and nothing else.

Unfortunately though,

1) QKeySequence is not exported in QML, and even if it were, being a
QGadget makes it virtually impossible to use as a property type because
of mind-blowing variety of internal errors from QML engine, such as:

    Invalid property assignment: unsupported type "QKeySequence*"
    TypeError: Property 'KeySequence' of object [null] is not a function
    TypeError: Type error
    Error: Cannot assign QKeySequence to QKeySequence*

2) Qt factory methods of QKeySequence are not marked to be Q_INVOKABLE
from QML side, I had to introduce one in helper class, so that we can
always construct a valid object (gadget) with an empty sequence instead
of messing with null / undefined / empty string values as before.

Fixed whitespaces and formatting; renamed _helper id to just helper;
replaced some == checks with strict ===, and replaced dynamic/indirect
parent.keySequence lookup with root.keySequence access (because parent
is a generic Item, which doesn't have keySequence property).

Among other things, according to MDN docs[1] `+` and `+=` are the
preferred ways of string concatenation in JavaScript.

[1]: https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/String/concat

It allows reusing components from org.kde.kcm module in other code which
is not running in a KCM module. A marginal use-case, but a possible
one. For example, according to lxr[1] search, the Timer applet is the
only applet that uses those controls, but still it shouldn't spam logs
with meaningless 'undefined reference' messages, since mostly
(without going into applet's general usability concerns) it seems to be
working perfectly fine.

[1]: https://lxr.kde.org/search?%21v=kf5-qt5&_filestring=&_string=org.kde.kcm

QOpenGLContext is an incomplete type for QT_NO_OPENGL builds so, whilst
references and pointers are fine, they cannot be dereferenced (implict
in the case of references), nor can automatic storage variables be
declared.
Co-authored-by: Alex Richardson <arichardson.kde@gmail.com>

Orca can read the name and the description of the button now.

BUG: 449282

So arrow keys can still navigate the items after the menu is closed.

We seldom use `Keys.onMenuPressed` in QML, which causes inconsistency
between QtWidget apps and QtQuick apps, so it's time to fix it.

Instead of /usr/include/KF5/kdeclarative_version.h.

For more details see:
kservice!79

GIT_SILENT

Task: https://phabricator.kde.org/T12126

Laurent Montel authored Dec 30, 2021 and Volker Krause committed Jan 07, 2022

Core5Compat have been changed to make their linking to it PUBLIC."