D6394: Add KF6-TODO to make KComboBox::minimumSizeHint() public
dfaure accepted this revision. This revision is now accepted and ready to land. REPOSITORY R284 KCompletion BRANCH addkf5todo REVISION DETAIL https://phabricator.kde.org/D6394 To: kossebau, #frameworks, dfaure Cc: cfeck, apol
D6047: WIP: Support XDG v6
graesslin added a comment. In https://phabricator.kde.org/D6047#119698, @mart wrote: > trying to implement the kwin part of ping btw, will do in a branch of the branch kwin/xdgv6 no need to hurry with it. IMHO the way ping works is completely broken and won't work. At least Qt holds the connection in a thread, so it will react to the ping even if the application is frozen in the main gui thread. Also please note that ping needs to be different in Wayland compared to X11. In X11 we only ping when trying to close the window. In Wayland we should ping whenever we interact with the window in some way. E.g. when we pass pointer focus, when we pass keyboard focus, maybe even when sending in input events. REPOSITORY R127 KWayland REVISION DETAIL https://phabricator.kde.org/D6047 To: davidedmundson, #plasma Cc: graesslin, mart, plasma-devel, #frameworks, ZrenBot, progwolff, lesliezhai, ali-mohamed, jensreuterberg, abetts, sebas, apol, lukas
D6047: WIP: Support XDG v6
davidedmundson marked 2 inline comments as done. davidedmundson added a comment. Done all comments except ping/pong. Marco, ping me if you ever want this updating. REPOSITORY R127 KWayland REVISION DETAIL https://phabricator.kde.org/D6047 To: davidedmundson, #plasma Cc: graesslin, mart, plasma-devel, #frameworks, ZrenBot, progwolff, lesliezhai, ali-mohamed, jensreuterberg, abetts, sebas, apol, lukas
D6047: WIP: Support XDG v6
davidedmundson updated this revision to Diff 15889. davidedmundson marked 2 inline comments as done. davidedmundson added a comment. Restricted Application edited projects, added Plasma; removed Plasma on Wayland. - cleanups REPOSITORY R127 KWayland CHANGES SINCE LAST UPDATE https://phabricator.kde.org/D6047?vs=15797=15889 BRANCH davidedmundson/xdgv6popup REVISION DETAIL https://phabricator.kde.org/D6047 AFFECTED FILES autotests/client/CMakeLists.txt autotests/client/test_xdg_shell.cpp autotests/client/test_xdg_shell.h autotests/client/test_xdg_shell_v5.cpp autotests/client/test_xdg_shell_v6.cpp src/client/CMakeLists.txt src/client/protocols/xdg-shell-unstable-v6.xml src/client/registry.cpp src/client/registry.h src/client/xdgshell.cpp src/client/xdgshell.h src/client/xdgshell_p.h src/client/xdgshell_v5.cpp src/client/xdgshell_v6.cpp src/server/CMakeLists.txt src/server/display.cpp src/server/xdgshell_interface.cpp src/server/xdgshell_interface.h src/server/xdgshell_interface_p.h src/server/xdgshell_v5_interface.cpp src/server/xdgshell_v5_interface_p.h src/server/xdgshell_v6_interface.cpp src/server/xdgshell_v6_interface_p.h src/tools/mapping.txt tests/CMakeLists.txt tests/xdgtest.cpp To: davidedmundson, #plasma Cc: graesslin, mart, plasma-devel, #frameworks, ZrenBot, progwolff, lesliezhai, ali-mohamed, jensreuterberg, abetts, sebas, apol, lukas
D6047: WIP: Support XDG v6
mart added a comment. trying to implement the kwin part of ping btw, will do in a branch of the branch kwin/xdgv6 REPOSITORY R127 KWayland REVISION DETAIL https://phabricator.kde.org/D6047 To: davidedmundson, #plasma Cc: graesslin, mart, plasma-devel, #frameworks, ZrenBot, progwolff, lesliezhai, ali-mohamed, jensreuterberg, abetts, eliasp, sebas, apol, hein, lukas
D6394: Add KF6-TODO to make KComboBox::minimumSizeHint() public
cfeck added a comment. +1 That shouldn't have landed in "protected" area. And no, MSVC has not changed :) REPOSITORY R284 KCompletion REVISION DETAIL https://phabricator.kde.org/D6394 To: kossebau, #frameworks Cc: cfeck, apol
D6394: Add KF6-TODO to make KComboBox::minimumSizeHint() public
kossebau added inline comments. INLINE COMMENTS > apol wrote in kcombobox.h:527 > Why KF6? changing visibility doesn't break ABI, AFAIK. We should just not > reorder them. https://community.kde.org/Policies/Binary_Compatibility_Issues_With_C%2B%2B hints the access rights are mangled into the symbol signature with some compilers. At least MSVC is mentioned here: http://blog.qt.io/blog/2009/08/12/some-thoughts-on-binary-compatibility/ Happy to learn that things have changed meanwhile, this is my latest info on that matter, never heard of anything else :) REPOSITORY R284 KCompletion REVISION DETAIL https://phabricator.kde.org/D6394 To: kossebau, #frameworks Cc: apol
D6394: Add KF6-TODO to make KComboBox::minimumSizeHint() public
apol added inline comments. INLINE COMMENTS > kcombobox.h:527 > > +// TODO KF6: make public like in base classes, so consumers do not need > to cast to base classes > +// when they have a KComboBox (or subclasses) object and want to access > this property Why KF6? changing visibility doesn't break ABI, AFAIK. We should just not reorder them. REPOSITORY R284 KCompletion REVISION DETAIL https://phabricator.kde.org/D6394 To: kossebau, #frameworks Cc: apol
D6394: Add KF6-TODO to make KComboBox::minimumSizeHint() public
kossebau created this revision. Restricted Application added a project: Frameworks. REPOSITORY R284 KCompletion BRANCH addkf5todo REVISION DETAIL https://phabricator.kde.org/D6394 AFFECTED FILES src/kcombobox.h To: kossebau, #frameworks
D6390: WIP: Add remote runners over DBus
davidedmundson created this revision. Restricted Application added projects: Plasma, Frameworks. Restricted Application added subscribers: Frameworks, plasma-devel. REVISION SUMMARY This patch allows adding runners which, rather than being in-process receive a search term and provide results over DBus. A metadata file should be installed with metadata for the service and path, and the remote app should implement the supplied interface. By using metadata files all the current krunner architecture and config management works seamlessly, and it allows the possibility for a runner to be DBus activated if it wants to be. The goal of this patch is threefold: - To make it considerably easier to write runners where the data exists in an already running application. The plasma-browser-integration is a good example, it currently has it's own custom protocol, and ends up sending data on all the tabs only to filter it inside krunner, somewhat wasteful. With this adding searchable tabs to konversation (for example) would be a ten minute patch. - Make it easy to write a runner in any other language. An app just needs to speak DBus with no other dependencies. Potentially even shipped via the store. - This approach could work for runners in sandboxes. TEST PLAN Wrote new sample app Got my runner showing results Brief profiling had roundtrip time as 0ms. REPOSITORY R308 KRunner BRANCH master REVISION DETAIL https://phabricator.kde.org/D6390 AFFECTED FILES src/CMakeLists.txt src/data/org.kde.krunner1.xml src/data/servicetypes/plasma-runner.desktop src/dbusrunner.cpp src/dbusrunner_p.h src/dbusutils_p.h src/runnermanager.cpp To: davidedmundson, #plasma Cc: plasma-devel, #frameworks, ZrenBot, progwolff, lesliezhai, ali-mohamed, jensreuterberg, abetts, sebas, apol, mart, lukas
Re: Kirigami in Frameworks
On Sat, Jun 24, 2017 at 3:27 PM, David Faurewrote: >> the default style for QtQuickControlsStyle1 is "Desktop" >> we could have called this style "Desktop" as well so all would have >> aligned nicely, but that could be quite dangerous, as Qt coulddecide >> any moment that they indeed want to do a style called "Desktop" for >> qqc2, at which point it would conflict, so having the org.kde prefix >> was the safe route. > > Well, we could also rename ours when/if the conflict occurs? since the conflict would be of installed files, it would make released version not installable, which looks to me like too big of a risk. >> One way to silence this could be when the qtquickcontrols2 >> theme installs into qt's qqc2 > > If you mean $QTDIR, that's not my case. The plugin is found via > QML2_IMPORT_PATH I suppose. qtquickcontrols2 will be installed under QML2_IMPORT_PATH and we need to install under that folder >> , to install also an org.kde.desktop >> symlink in QtQuickControls1 that just points to "Desktop" (note that >> style is not in the kirigami repository, kirigami has themes with the >> same name because it's an extension on top of qtquickcontrols2) > > That sounds good, if it can be done. setting a symlink there seems to just work, so i would try to go for that route >> It has some android specific things, like providing a manifest.xml and >> some optional qtandroidextras usage for integration when compiled >> there, but it's intended to be multiplatform, so may make sesie to >> enable it. > > Then I don't think it should be under an android subdir. > A multiplatform app with extra stuff for android integration is still > multiplatform in the first place. sure, renaming it to multipletform then :) >> talking about examples, do you think is better to build them with a >> flag on the top cmake as is now, or provide a separate standalone >> cmake file under examples/ ? > > Much better the way you did it, it's how I see it done in most other > frameworks. > It makes it easy to enable the building of examples once and for all in > kdesrc-buildrc for instance, to make sure they keep compiling and so that they > are available for testing. > In fact, I wonder if the CI shouldn't set BUILD_EXAMPLES=ON in all frameworks > that have that option... (4 currently, 5 with kirigami). ok -- Marco Martin
D6313: WIP: Support device pixel ratio in icon loader and engine
mart added a comment. In https://phabricator.kde.org/D6313#118285, @kvermette wrote: > Behaviorally speaking there's justification for ensuring SVGs are treated the same as PNGs in this case. Looking at the code we aren't shimming the SVGs specifically (unless I'm missing something), but I just wanted to chime in with this and nip special treatment for SVGs in the bud. In maintaining the Aether icon theme I would create links to the higher resolution icons, so you'd see something like a "16x16x2" folder point to my "32x32" 'native' folder, and a "16x16x3"->"48x48" folder, "32x32x2"->"64x64", etc. Just because we *can* scale SVG icons doesn't mean that behavior should be assumed, should the icon set have higher fidelity icons ready for HiDPI. i think for svgs 16x16x2 should be preferred over scaling by 2 the 16x16 version, but still, scaling the 16x16 version should still be preferred over the 32x32 version, so this patch is on the right rirection REPOSITORY R302 KIconThemes REVISION DETAIL https://phabricator.kde.org/D6313 To: broulik, kde-frameworks-devel, #plasma, #vdg Cc: mart, kvermette, cfeck, davidedmundson, plasma-devel, #frameworks, ZrenBot, progwolff, lesliezhai, ali-mohamed, jensreuterberg, abetts, sebas, apol, lukas
D6376: Fix double delete crash during shutdown
rjvbb added a comment. In https://phabricator.kde.org/D6376#119500, @cullmann wrote: > I don't speak about threading races. The whole class is not thread-safe, if threading occurs, all is lost. Really? Only a single thread at a time should trigger audio notifications, is that what you're saying? That would surprise me a little, looking at the code it seems the class exists so that applications can request to play a notification sound by creating an instance and then forget about it. The threading I was thinking of is simply when 2 threads queue an audio notification at the same time, not the cross-thread use of a NotifyByAudio instance. REPOSITORY R289 KNotifications REVISION DETAIL https://phabricator.kde.org/D6376 To: cullmann, #frameworks Cc: aacid, mpyne, rjvbb
D6376: Fix double delete crash during shutdown
cullmann abandoned this revision. REPOSITORY R289 KNotifications REVISION DETAIL https://phabricator.kde.org/D6376 To: cullmann, #frameworks Cc: aacid, mpyne, rjvbb