Summary:
When the size of the combo box adjusts to its contents, the dropdown
listbox isn't large enough to fit the contents and a vertical scrollbar
so instead of using Metrics::MenuButton_IndicatorWidth. I defined
a new Metrics::ComboBox_IndicatorWidth which is two pixels larger than
Metrics::MenuButton_IndicatorWidth.
To test this I used "kcmshell5 kcm_kscreen -style breeze" and opened
the resolutions QComboBox (If there are less than 15 possible
resolutions, a slider will appear instead of a QComboBox, so in that
case it's possible to use a virtual machine with QXL video that for
me shows plenty of possible resolutions).
Before this patch, most of the larger resolutions are shortened to
something like "192...200" or "160...200" and can't be read (in the
dropdown listbox, when selected they can be read correctly in the
closed combobox), after applying this patch, all the contents of the
listbox can be read correctly.
Reviewers: #plasma, #plasma:_design, hpereiradacosta
Reviewed By: hpereiradacosta
Subscribers: hpereiradacosta, davidedmundson, plasma-devel
Tags: #plasma, #plasma:_design
Differential Revision: https://phabricator.kde.org/D3192
Summary:
When the size of the combo box adjusts to its contents, the dropdown
listbox isn't large enough to fit the contents and a vertical scrollbar
so instead of using Metrics::MenuButton_IndicatorWidth. I defined
a new Metrics::ComboBox_IndicatorWidth which is two pixels larger than
Metrics::MenuButton_IndicatorWidth.
To test this I used "kcmshell5 kcm_kscreen -style breeze" and opened
the resolutions QComboBox (If there are less than 15 possible
resolutions, a slider will appear instead of a QComboBox, so in that
case it's possible to use a virtual machine with QXL video that for
me shows plenty of possible resolutions).
Before this patch, most of the larger resolutions are shortened to
something like "192...200" or "160...200" and can't be read (in the
dropdown listbox, when selected they can be read correctly in the
closed combobox), after applying this patch, all the contents of the
listbox can be read correctly.
Reviewers: #plasma, #plasma:_design, hpereiradacosta
Reviewed By: hpereiradacosta
Subscribers: hpereiradacosta, davidedmundson, plasma-devel
Tags: #plasma, #plasma:_design
Differential Revision: https://phabricator.kde.org/D3192
Summary:
This is a proof of conceptof lighter scrollbars, according to
ideas from the VDG:
show only a tiny placeholder normally, and show the full scrollbar
only on mouse over.
Test Plan:
tested on few applications but i know well the
code is just an early hacky proof of concept,
so i'll redesign it as requested.
being a big visible thing, it will probably need
an option as well to make it behave as it used to.
in action https://www.youtube.com/watch?v=O72qxQHjHcw
Reviewers: #plasma, #breeze, #vdg, hpereiradacosta
Reviewed By: hpereiradacosta
Subscribers: plasma-devel, abetts
Tags: #plasma
Differential Revision: https://phabricator.kde.org/D3131
Summary:
So far breeze hard disabled the window moving on Wayland. With this
change the required functionality gets added.
For that Breeze creates an additional Seat and a Pointer on it to track
all pointer button events on the window. That way the kstyle gets the
latest serial which needs to be passed to the move requests. This is not
available through QtWayland's native interface, thus Breeze needs to
interact with Wayland directly.
When the move is triggered Breeze gets the ShellSurface for the window
and triggers the move on the own Seat object with the tracked serial.
Test Plan: Tested with KWin and Weston, move triggered in both.
Reviewers: #plasma_on_wayland
Subscribers: plasma-devel
Tags: #plasma_on_wayland
Differential Revision: https://phabricator.kde.org/D3019
Since Qt 5.6.0, Qt5 applications started crashing on exit. All signs
point to this delete-on-destroy hack which was added to avoid outliving
the plugin lifetime.
This method is wrong because the returned style is owned by the caller
(QApplication, QProxyStyle, etc) and will cleaned up when those users
are destructed.
REVIEW: 128760
Since Qt 5.6.0, Qt5 applications started crashing on exit. All signs
point to this delete-on-destroy hack which was added to avoid outliving
the plugin lifetime.
This method is wrong because the returned style is owned by the caller
(QApplication, QProxyStyle, etc) and will cleaned up when those users
are destructed.
Review: 128760
This enables painting the focus underline also for QtQuick Controls and fixes
an off-by-one error which caused the line to be rendered outside the
QtQuick Controls canvas and was thus invisible.
REVIEW: 128654
Hugo Pereira Da Costa
Antonio Larrosa
Marco Martin
l10n daemon script
Aleix Pol
Jonathan Riddell
Martin Gräßlin
Peter Wu
Andreas Hartmetz
David Edmundson
Ken Vermette
Kai Uwe Broulik