By default the window size is restored with respect to the screen at
0,0; this patch uses the screen under mouse cursor. It is still not
optimal, as one would have to have a way to query the workspace for
the current screen. Is there any reasonable way to do that?
5.64 added an isDefault method however this doesn't suffice for usage in
KConfigDialogManager which compares a current value to the default.
KConfigDialogManager currently uses a hack with side effects.
Exposing the value directly solves that.
This is done for window positions, so it should also be done for
window sizes. The combination of both should substantially improve the
use case of maintaining a consistent window arrangement for multiple
display setups (e.g. laptop and laptop + external screen).
Right now it's duplicated twice, and once we add per-screen-arrangement
memory for window sizes as well, it will be duplicated four times unless
we put it into a function.
"QtPrivate::DeprecatedRefClassBehavior::warn|?libKF5ConfigCore.so.5?|?libKF5ConfigCore.so.5?
Using QCharRef with an index pointing outside the valid range of a
QString. The corresponding behavior is deprecated, and will be changed
in a future version of Qt."
We try to access to an invalid index.
We have a huge amount of standard shortcuts. By adding the ability to divide
them into categories, they can be presented to the user in a way that enables
them to easier find the shortcut that they want. Internally some shortcuts were
already grouped but that seems to have stopped for the later additions. (The
related standard actions from KConfigWidgets also seem to have this not exposed
internal grouping.)
See plasma-desktop!26
These functions allow an application to save and restore the positions
of its windows. Positions are stored on a per-screen-arrangement basis.
For example with a single screen connected, the config file would have
entries like this in it:
eDP-1 XPosition=140
eDP-1 YPosition=340
When a second screen is connected, the following gets saved:
eDP-1 HDMI-1 XPosition=3878
eDP-1 HDMI-1 YPosition=29
This ensures that each separate screen arrangement can have its own
saved window position, which is handy for the use case where you have a
laptop that you plug into an external screen some of the time. It also
allows the position to get restored to the correct screen when there are
multiple screens.
This is a necessary first step to getting KDE apps to save their window
positions on X11 The next step would be calling the new functions from
KXMLGui and Kirigami apps, and then porting all apps that manually invoke
KWindowConfig::saveWindowSize() and KWindowConfig::restoreWindowSize()
to also invoke KWindowConfig::saveWindowPosition() and
KWindowConfig::restoreWindowPosition() in the same places.
The functions only work on X11 or other non-Wayland platforms. On
Wayland, the compositor has sole dominion over window positioning so a
compositor-specific solution much be adopted instead, such as
https://bugs.kde.org/show_bug.cgi?id=15329.
CCBUG: 415150
When running unit tests, usually QStandardPaths TestMode is enabled so as
not to mess up the config files in the home dir of the dev running the unit
tests; sGlobalFileName, a global static, was defined only once, which meant
that if KSharedConfig::openConfig() is called before the unit test has had
a chance to call QStandardPaths::setTestModeEnabled(true), then sGlobalFileName
will go on referring to the wrong kdeglobals file (as can be seen in
dfaure's debugging of the issue at [1]). Change the code so as to
detect QStandardPaths TestMode status and change sGlobalFileName as needed.
All unit tests still pass.
[1] https://invent.kde.org/frameworks/kio/-/merge_requests/77#note_74124
While the API might have been once designed to allow random byte strings
to be used as ids and only converted QString to UTF-8, the current
implementation in many places assumes the byte strings passed via
const char* or QByteArray are also in of UTF-8 encoding
(or at least ASCII7-only).
See e.g. KConfigGroup::entryMap() or KConfigGroup::name().
Stating the supported encoding explicitly should avoid any misassumptions.
locationType is used to determine where the config file should be saved.
If it is not passed on to the KConfigSkeleton constructor, the default
(GenericConfigLocation) will be used and any value the application had
set will be ignored.
Summary:
This is useful for unittests. Example:
```
KCONFIGCORE_EXPORT void reloadUrlActionRestrictions();
void someTestMethod()
{
KConfigGroup cg(KSharedConfig::openConfig(), "KDE URL Restrictions");
cg.writeEntry("rule_count", 1);
cg.writeEntry("rule_1", QStringList{"open", {}, {}, {}, "file", "", "", "false"});
cg.sync();
reloadUrlActionRestrictions();
// Some test for code that uses KUrlAuthorized
cg.deleteEntry("rule_1");
cg.deleteEntry("rule_count");
cg.sync();
reloadUrlActionRestrictions();
}
```
This is consistent with the fact that other functions used by
KUrlAuthorized are defined here and exported internally.
Test Plan: Used this in a KIO unittest I'm writing for the future OpenUrlJob
Reviewers: aacid, apol, mdawson
Reviewed By: aacid
Subscribers: kde-frameworks-devel
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D29347
Summary:
- Fix order of member initialisation (compiling with -Wreorder)
- Use Q_UNUSED where needed
- QAtomicInteger::loadRelaxed() has been available since Qt 5.14
Test Plan: make && ctest
Reviewers: #frameworks, dfaure, apol
Reviewed By: dfaure
Subscribers: kde-frameworks-devel
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D27158
Summary:
KApiDox & ECMAddQch have been just teached about Q_DECLARE_FLAGS, so the
underlying typedefs are no longer skipped by doxygen, but can be
documented now, allowing links to be generated for these types e.g. when
used as arguments.
The "#" prefix to the enum name in the description text of all the
Q_DECLARE_FLAGS docs seems needed to properly trigger doxygen autolinks.
GIT_SILENT
Summary:
Fix the following bug, if an entry is set on HOME/.config/kdeglobals and on a specific config file like kcmfonts
When you hit restore defaults button and apply, value will be not wrote on the specific file, but then the value is the one from kdeglobals
This patch ensure we write the key to the specific configuration file on those case with an empty value. KConfig will take default value automatically
Test Plan:
Added a test to ensure flag is working
Tested using some KCM to ensure all is working fine
Reviewers: ervin, dfaure, meven, crossi, hchain
Reviewed By: ervin, dfaure, meven
Subscribers: kde-frameworks-devel
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D28128
Summary:
Add a standard shortcut for this action so it can be synchronized between different apps
and the file dialog; see https://bugs.kde.org/show_bug.cgi?id=262551.
The action in Dolphin also has F8 bound to this, but the definition here only accept two
default shortcuts, and F8 seems kind of random.
CCBUG: 262551
Test Plan: Apply D28803 to KIO and D28804 to Dolphin and test there
Reviewers: dfaure, #frameworks
Reviewed By: dfaure
Subscribers: kde-frameworks-devel
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D28802
Summary:
This is an addendum for commit a2774ff5b41987c3919a9e; making the check
tighter.
Test Plan: make && ctest
Reviewers: #frameworks, dfaure, davidedmundson, cfeck, ervin, meven, bport
Reviewed By: dfaure
Subscribers: kde-frameworks-devel
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D27811
Jacopo De Simoi
David Edmundson
Nate Graham
Laurent Montel
Andreas Cord-Landwehr
David Redondo
Cyril Rossi
l10n daemon script
Ahmad Samir
Friedrich W. H. Kossebau
Arjen Hiemstra
David Faure
Benjamin Port
Ömer Fadıl USTA
Méven Car