In commit 45e0a722fb we changed the focus
on the textfield with the intention of fixing the cusor blinking.
This had the unfortunate side effect of breaking waking out of the
screensaver mode with just the keyboard.
That screensaver mode was bound to:
```
property bool blockUI: mainBlock.mainPasswordBox.text.length > 0
```
for that to change, we need to have focus.
Instead this patches cursor blinking more explicity.
Doing it only in the lock screen (as opposed to the toolkit) makes sense
as this only happens due to somewhat unusual usage inside the
lockscreen.
BUG: 431286
This isn't always static, espcially in GTK apps where the xproperty is
set later by gmenuproxyodel.
The Appmenu applet already watches for this role changing, but it wasn't
hooked up.
The wayland implementation does not have the same bug.
BUG: 422786
This makes it look better with the change to Kirigami.SwipeListItem to
make the actions always visible on the desktop, and in the process fixes
a binding loop and some minor visual glitches.
There is an error handling path when we fetch the relevant config()
```
KConfigGroup PanelView::panelConfig(...
{
if (!containment || !screen) {
return KConfigGroup();
}
```
which we indiscrimiately call parent() on.
This patch guards that case, which is presumably screen being
temporarily null.
This code is also Plasma 5.8 compatibility fallback, so arguably we
could get rid of it.
BUG: 425711
An opacity of 0 but still visible still results in nodes in the
scenegraph, which is wasteful. This is shown in gammaray with some
warnings.
Enabled is also bound to visible as if a text field has focus it still
animates the cursor icon even if inivisble, producing wakeups.
BUG: 347772
FIXED-IN: 5.21
For distributions aiming to ship the Wayland session by default, allow
the Wayland session file to be installed as "plasma.desktop". Otherwise,
revert to status quo where the X11 session is installed as "plasma.desktop".
This time around, ensure that CMake stanza is correctly filled out.
This fixes commit b93a06395e.
This reverts commit a6b2891186.
Removing the only item makes no conceptual sense, so let's not permit
it. Likewise, when there's only one item, it can't be moved to the top
since it's already at the top, so let's not show that one either.
BUG: 397978
FIXED-IN: 5.21
When using the lock/logout applet it feels strange, to get only at two
options a confirmation dialog. Additionally it feels odd to answer for
confirmation when neither the option triggered by Shortcuts nor the
Plasma menu asks for confirmation. Suspend and hibernate are the two
options where no data is lost, there is no need to ask for confirmation.
We no longer need --exit-with-session as plasma_shutdown explicitly
quits kwin gracefully.
We also don't want to relaunch ksplash if kwin restarts.
We run the whole script as we need to restart ksmserver as xserver will
be restarted.
For other services, if they get restored through crash handlers or this
doesn't really matter. DBus unique-ness will take care of it.
I find it weird that when I've just trashed some files - presumably to get rid
of them - it then offers me to open that file or the trash folder once done.
It was used to show the KNS dialog to download
wallpaper plugins. But nowadays we can and should
just use the QML components provided by KNewStuff directly.
All occurency in plasma-desktop, plasma-bigscreen and
plasma-nano have open MRs to port away from it (or just remove the import).
Allowing users to manually change the font DPI independent of font sizes
and the global scale is already of questionable value on X11, but on
Wayland where Qt scaling is always used, it is an anachronism. So let's
only show it on X11. On Wayland, people can use the global scaling
feature in the KScreen KCM.
See also https://phabricator.kde.org/T9500
Before the variables would get expanded later and consequently
go through another code path than file locations. This leads to the
text being differently displayed and users getting confused.
Also this allows us to check if the file actually exists.
BUG: 358221
FIXED-IN: 5.21
- Do not use deprecated RunnerContext::type()
- Write the url always in the data, this way we don't
need to do all the parsing twice.
- The concept of the case insensitive search does not make
sense in this runner. If the user want to have case insensitive
file search the baloo/recent docs/places runners should be used.
- Do not show a match when we are sure the file does not exist. This would
only show a KIO error message when launching it, consequently there is
no point in showing the message in the first place.
This solves the problem when the user clicks the padding around the applet or its label. The event is now correctly redirected to the applet.
This fixes both the left button and the middle click.
BUG: 426646
FIXED-IN: 5.21
- Support partial matches
- Allow both translated and non-translated trigger words
- Avoid duplicating the translated strings. This can become useful in
case we also want to add the non-translated trigger words to the
syntaxes (once they get displayed somewhere ;))
- Get rid of QRegExp usage
FEATURE: 428193
FEATURE: 428194
FIXED-IN: 5.21
QDBusServiceWatcher gained wildcard support in Qt 5.14. We can now port
to use it over the deprecated signal on QDBusConnectionInterface.
This not only ports away from deprecated code, but should reduce
plasmashell wakeups for unreltaed enties.
David Edmundson
l10n daemon script
Nate Graham
Neal Gompa
Rik Mills
Alexander Lohnau
Kai Uwe Broulik
Sandro Knauß
Laurent Montel
Ismael Asensio
Konrad Materka