Jenkins-kde-ci: plasma-workspace Plasma-5.5 stable-kf5-qt5 » Linux,gcc - Build # 25 - Still Failing!
GENERAL INFO BUILD FAILURE Build URL: https://build.kde.org/job/plasma-workspace%20Plasma-5.5%20stable-kf5-qt5/PLATFORM=Linux,compiler=gcc/25/ Project: PLATFORM=Linux,compiler=gcc Date of build: Sat, 26 Dec 2015 05:33:23 + Build duration: 4 min 20 sec CHANGE SET Revision a14a7d39c41bd3280dcc56dcd97e846e0278e812 by mpyne: (Use proper deleters for libxcb structs.) change: edit xembed-sni-proxy/sniproxy.cpp ___ Plasma-devel mailing list Plasma-devel@kde.org https://mail.kde.org/mailman/listinfo/plasma-devel
Re: Review Request 126512: Use proper deleters for libxcb structs in xembed-sni-proxy.
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/126512/ --- (Updated Dec. 26, 2015, 5:32 a.m.) Status -- This change has been marked as submitted. Review request for Plasma. Changes --- Submitted with commit a14a7d39c41bd3280dcc56dcd97e846e0278e812 by Michael Pyne to branch Plasma/5.5. Repository: plasma-workspace Description --- Coverity noted we were mismatching new Foo with free(foo), which is undefined behavior (CID 1340556). While I was fixing that I noticed we have the same issue with QScopedPointer<>: when using QSP to track objects returned by libxcb, we must use free() to release memory, not C++ delete. (e.g. see the [XCB API](http://xcb.freedesktop.org/manual/group__XCBAPI.html#ga6727f2bfb24769655e52d1f1c50f58fe)). QScopedPointer will do this if we use QScopedPointerPodDeleter. Diffs - xembed-sni-proxy/sniproxy.cpp ae6eab7 Diff: https://git.reviewboard.kde.org/r/126512/diff/ Testing --- I've only compiled and verified no warnings at this point. I didn't see any relevant crash bugs after a quick web search either. Thanks, Michael Pyne ___ Plasma-devel mailing list Plasma-devel@kde.org https://mail.kde.org/mailman/listinfo/plasma-devel
Re: Review Request 126513: Refactor kcm touchpad to fix some issue related to hotplug touchpad
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/126513/ --- (Updated Dec. 25, 2015, 7:02 p.m.) Review request for Plasma and David Edmundson. Changes --- Make sure we restore touchpad to the state that userRequested before plugin mouse. Also Fix 349545. Bugs: 349545 and 356923 https://bugs.kde.org/show_bug.cgi?id=349545 https://bugs.kde.org/show_bug.cgi?id=356923 Repository: plasma-desktop Description --- Currently, there are some mixed issue in kcm touchpad related to libinput backend and hot plug. There are several issues: 1. only one backend maybe used at runtime, either synaptics or libinput. But libinput backend will only be used if there is a libinput backend present. Which means if libinput touchpad is not present at login, kded will not be able to properly support them. 2. hotplug touchpad will always set touchpad to disabled. 3. hotplug touchpad will not get configuration applied. 4. The libinput devices detection may pick some non-touchpad device. To solve these problems, following changes are made: 1. Remove SynapticsBackend and LibinputBackend and add three new classes called XlibTouchpad/SynapticsTouchpad/LibinputTouchpad. So findTouchpad may pick up both synaptics or libinput device, and no need to depend on a fixed backend. This change also make it easier to support multiple touchpad device easier if needed. 2. hotplug touchpad config not being applied (bug 356923) is mainly because one can't apply settings to a disabled device. Change the statement order in handleReset, apply settings first then set enable status. 3. findTouchpad() is changed to use XListInputDevices, because it provides necessary information of device type in order to filter out non-touchpad device. xf86-input-libinput doesn't have a unique atom for touchpad, so current identityAtom cannot guarantee that findTouchpad always returns a touchpad device for libinput. 4. On my surface pro 4 system, if touchpad is unpluged, the device will be set to disabled automatically first, then touchpadDetached is called, which makes m_enabled in TouchpadDisabled always to be false after unplug the touchapd. This patch makes TouchpadDisabler have two different enabled property. m_userRequestedState only stores the user requested status, so when handleReset is called, it can properly set enabled state to the last user requested state. Diffs (updated) - kcms/touchpad/src/applet/qml/contents/ui/touchpad.qml 8dec7c2 kcms/touchpad/src/applet/touchpadengine.h 16b98c2 kcms/touchpad/src/applet/touchpadengine.cpp eae429e kcms/touchpad/src/backends/x11.cmake c9fcea8 kcms/touchpad/src/backends/x11/libinputproperties.c 9dbf9ea kcms/touchpad/src/backends/x11/libinputtouchpad.h PRE-CREATION kcms/touchpad/src/backends/x11/libinputtouchpad.cpp PRE-CREATION kcms/touchpad/src/backends/x11/propertyinfo.h PRE-CREATION kcms/touchpad/src/backends/x11/propertyinfo.cpp PRE-CREATION kcms/touchpad/src/backends/x11/synapticstouchpad.h PRE-CREATION kcms/touchpad/src/backends/x11/synapticstouchpad.cpp PRE-CREATION kcms/touchpad/src/backends/x11/synclientproperties.h 43d18d2 kcms/touchpad/src/backends/x11/synclientproperties.c 5fd1ed6 kcms/touchpad/src/backends/x11/xlibbackend.h 7cecb4a kcms/touchpad/src/backends/x11/xlibbackend.cpp b55a45f kcms/touchpad/src/backends/x11/xlibtouchpad.h PRE-CREATION kcms/touchpad/src/backends/x11/xlibtouchpad.cpp PRE-CREATION kcms/touchpad/src/kded/kded.h 9b8fe6e kcms/touchpad/src/kded/kded.cpp 409126b kcms/touchpad/src/touchpadbackend.h b225ed9 Diff: https://git.reviewboard.kde.org/r/126513/diff/ Testing --- Without touchpad -> login -> plug touchpad -> settings applied to device using libinput driver. unplug -> replug -> settings applied to device using libinput driver. Disable touchpad -> applet shows up -> unplug touchpad -> applet hides -> replug touchpad -> touchpad is still disabled, and applet shows up -> enable touhcpad -> applet hides. Thanks, Xuetian Weng ___ Plasma-devel mailing list Plasma-devel@kde.org https://mail.kde.org/mailman/listinfo/plasma-devel
Re: Review Request 126515: [DrKonqi] RFC: Support showing a StatusNotifierItem instead of bringing up the dialog right away
> On Dez. 25, 2015, 5:31 nachm., David Edmundson wrote: > > code looks fine. > > > > Though please do the TODOs before merging as historically we have a have a > > habbit of them not getting done > > Also your StatusNotifierItem leaks - but it proably wants to be the > > lifespan of the application anyway. The sad face icon is used wrongly *everywhere* so I guess we can use it here, too. About the context menu, I thought I could add a couple of shortcuts there but I guess just having a "Quit" option (unfortunately you cannot tell KStatusNotifierItem *not* to ask before quitting) is sufficient. I was thinking off adding the app icon as overlay but that probably wouldn't work well, at least not with the black KDE logo. - Kai Uwe --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/126515/#review90104 --- On Dez. 25, 2015, 4:24 nachm., Kai Uwe Broulik wrote: > > --- > This is an automatically generated e-mail. To reply, visit: > https://git.reviewboard.kde.org/r/126515/ > --- > > (Updated Dez. 25, 2015, 4:24 nachm.) > > > Review request for KDE Frameworks, Plasma, KDE Usability, and Martin Gräßlin. > > > Repository: plasma-workspace > > > Description > --- > > This adds a new "--passive" option to DrKonqi where it will only show a > StatusNotifierItem rather than bringing up the crash dialog right away. > > This can be useful for auto-restarting shell services (like plasmashell, > krunner, kded) to improve the perceived quality of the product. > > On Windows RT, for example, the guidelines even explicitly say "rather just > dump the user on the home screen than telling him something went wrong, so he > can just quickly start the app again instead of being annoyed by an error > message". On iOS you also just get dropped on the home screen. Windows > desktop shows a "Searching for a solution" dialog which was *the* major > annoyance when something crashed, rather than the actual crash. > > Video here: https://www.youtube.com/watch?v=t0ZLs-juYKc > > > Diffs > - > > drkonqi/statusnotifier.cpp PRE-CREATION > drkonqi/statusnotifier.h PRE-CREATION > drkonqi/CMakeLists.txt eaeaad4 > drkonqi/main.cpp 7cbaae7 > > Diff: https://git.reviewboard.kde.org/r/126515/diff/ > > > Testing > --- > > I crashed plasmashell, it restarted so fast that you didn't even have a black > screen inbetween, just the panel restarting. Afterwards I got a SNI which > opened DrKonqi when tapped. > > The SNI disappears after 1 minute because if you didn't bother to look after > it by then, you probably forgot what you did to cause the crash anyway :) > > > Thanks, > > Kai Uwe Broulik > > ___ Plasma-devel mailing list Plasma-devel@kde.org https://mail.kde.org/mailman/listinfo/plasma-devel
Jenkins-kde-ci: plasma-workspace Plasma-5.5 stable-kf5-qt5 » Linux,gcc - Build # 24 - Still Failing!
GENERAL INFO BUILD FAILURE Build URL: https://build.kde.org/job/plasma-workspace%20Plasma-5.5%20stable-kf5-qt5/PLATFORM=Linux,compiler=gcc/24/ Project: PLATFORM=Linux,compiler=gcc Date of build: Fri, 25 Dec 2015 18:16:35 + Build duration: 1 min 43 sec CHANGE SET Revision e8c6426ac12c14be68a68b9492274140c28f3285 by David Edmundson: (Guard against applet failing to load in systemtray task) change: edit applets/systemtray/plugin/protocols/plasmoid/plasmoidtask.cpp ___ Plasma-devel mailing list Plasma-devel@kde.org https://mail.kde.org/mailman/listinfo/plasma-devel
Re: Review Request 126517: [ShellManager] Make use of StatusNotifier DrKonqi
> On Dez. 25, 2015, 5:22 nachm., David Edmundson wrote: > > On first thought this is awesome. I like the video. It looks good. > > > > but now I'm going to point something out that's going to ruin your Christmas > > > > If Plasma crashes, you're going to try and show an SNI *in plasma* ... > > which isn't running. > > That's fine for a single crash, not for ones that block startup. > > > > I'm sure we can workaround this somehow - there's a timer in KCrash > > internals where we can know if something is a crash within the first 20 > > seconds of loading. > > Kai Uwe Broulik wrote: > That's why on third (and last) restart attempt it'll bring up DrKonqi > nonetheless. Good point nonetheless. > > David Edmundson wrote: > so it does. > > +1 There: https://www.youtube.com/watch?v=mIoMaJ7e8s4 :) Just instead of that KPassivePopup (I already played around with my other usability idea) with this patch you would get regular DrKonqi. - Kai Uwe --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/126517/#review90102 --- On Dez. 25, 2015, 4:24 nachm., Kai Uwe Broulik wrote: > > --- > This is an automatically generated e-mail. To reply, visit: > https://git.reviewboard.kde.org/r/126517/ > --- > > (Updated Dez. 25, 2015, 4:24 nachm.) > > > Review request for Plasma. > > > Repository: plasma-workspace > > > Description > --- > > This enables the StatusNotifier option introduced in Review 126516, see > Review 126515 for more information. > > When Plasma reached the recursion counter it disables the passive option > since there will most likely be no System Tray to show this error and then > something's really wrong in the first place. > > > Diffs > - > > shell/shellmanager.cpp 0a6ac5e > > Diff: https://git.reviewboard.kde.org/r/126517/diff/ > > > Testing > --- > > Crashed plasma two times, it restarted and I got two SNIs, crashed it another > time, got black screen with DrKonqi dialog. > > > Thanks, > > Kai Uwe Broulik > > ___ Plasma-devel mailing list Plasma-devel@kde.org https://mail.kde.org/mailman/listinfo/plasma-devel
Re: Review Request 126497: Guard against applet failing to load in systemtray task
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/126497/ --- (Updated Dec. 25, 2015, 6:16 p.m.) Status -- This change has been marked as submitted. Review request for Plasma. Changes --- Submitted with commit e8c6426ac12c14be68a68b9492274140c28f3285 by David Edmundson to branch Plasma/5.5. Bugs: 356470 https://bugs.kde.org/show_bug.cgi?id=356470 Repository: plasma-workspace Description --- If an applet fails to load properly m_applet will be null which is a valid state to be in when we destruct the plasmoid task object. BUG: 356470 Diffs - applets/systemtray/plugin/protocols/plasmoid/plasmoidtask.cpp 84d2baf4da0ef6d0381dfc79fa374b5c54cf2189 Diff: https://git.reviewboard.kde.org/r/126497/diff/ Testing --- Thanks, David Edmundson ___ Plasma-devel mailing list Plasma-devel@kde.org https://mail.kde.org/mailman/listinfo/plasma-devel
Jenkins-kde-ci: plasma-workspace Plasma-5.5 stable-kf5-qt5 » Linux,gcc - Build # 23 - Still Failing!
GENERAL INFO BUILD FAILURE Build URL: https://build.kde.org/job/plasma-workspace%20Plasma-5.5%20stable-kf5-qt5/PLATFORM=Linux,compiler=gcc/23/ Project: PLATFORM=Linux,compiler=gcc Date of build: Fri, 25 Dec 2015 18:13:15 + Build duration: 2 min 11 sec CHANGE SET Revision 198400e3bdaaeb82403ee3c90c45bf3eae3ad5c2 by David Edmundson: (destroy the container window we create) change: edit xembed-sni-proxy/sniproxy.cpp ___ Plasma-devel mailing list Plasma-devel@kde.org https://mail.kde.org/mailman/listinfo/plasma-devel
Re: Review Request 126498: Use the right graphics for min and max sliders
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/126498/ --- (Updated Dec. 25, 2015, 6:15 p.m.) Status -- This change has been marked as submitted. Review request for Plasma. Changes --- Submitted with commit 426a24c051d187875abc0e373cdc252f151b1d21 by David Edmundson to branch Plasma/5.5. Repository: plasma-desktop Description --- We never noticed on Breeze because they look the same. BUG:356790 REVIEW: 126498 Diffs - desktoppackage/contents/configuration/panelconfiguration/Ruler.qml 16b0c7137070d237778e705396b6e271859a947e Diff: https://git.reviewboard.kde.org/r/126498/diff/ Testing --- Thanks, David Edmundson ___ Plasma-devel mailing list Plasma-devel@kde.org https://mail.kde.org/mailman/listinfo/plasma-devel
Re: Review Request 126498: Use the right graphics for min and max sliders
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/126498/ --- (Updated Dec. 25, 2015, 6:14 p.m.) Review request for Plasma. Repository: plasma-desktop Description (updated) --- We never noticed on Breeze because they look the same. BUG:356790 REVIEW: 126498 Diffs - desktoppackage/contents/configuration/panelconfiguration/Ruler.qml 16b0c7137070d237778e705396b6e271859a947e Diff: https://git.reviewboard.kde.org/r/126498/diff/ Testing --- Thanks, David Edmundson ___ Plasma-devel mailing list Plasma-devel@kde.org https://mail.kde.org/mailman/listinfo/plasma-devel
Re: Review Request 126512: Use proper deleters for libxcb structs in xembed-sni-proxy.
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/126512/#review90106 --- Ship it! Ship It! - David Edmundson On Dec. 25, 2015, 4:53 a.m., Michael Pyne wrote: > > --- > This is an automatically generated e-mail. To reply, visit: > https://git.reviewboard.kde.org/r/126512/ > --- > > (Updated Dec. 25, 2015, 4:53 a.m.) > > > Review request for Plasma. > > > Repository: plasma-workspace > > > Description > --- > > Coverity noted we were mismatching new Foo with free(foo), which is > undefined behavior (CID 1340556). While I was fixing that I noticed we > have the same issue with QScopedPointer<>: when using QSP to track > objects returned by libxcb, we must use free() to release memory, not > C++ delete. (e.g. see > the [XCB > API](http://xcb.freedesktop.org/manual/group__XCBAPI.html#ga6727f2bfb24769655e52d1f1c50f58fe)). > > QScopedPointer will do this if we use QScopedPointerPodDeleter. > > > Diffs > - > > xembed-sni-proxy/sniproxy.cpp ae6eab7 > > Diff: https://git.reviewboard.kde.org/r/126512/diff/ > > > Testing > --- > > I've only compiled and verified no warnings at this point. I didn't see any > relevant crash bugs after a quick web search either. > > > Thanks, > > Michael Pyne > > ___ Plasma-devel mailing list Plasma-devel@kde.org https://mail.kde.org/mailman/listinfo/plasma-devel
Re: Review Request 126517: [ShellManager] Make use of StatusNotifier DrKonqi
> On Dec. 25, 2015, 5:22 p.m., David Edmundson wrote: > > On first thought this is awesome. I like the video. It looks good. > > > > but now I'm going to point something out that's going to ruin your Christmas > > > > If Plasma crashes, you're going to try and show an SNI *in plasma* ... > > which isn't running. > > That's fine for a single crash, not for ones that block startup. > > > > I'm sure we can workaround this somehow - there's a timer in KCrash > > internals where we can know if something is a crash within the first 20 > > seconds of loading. > > Kai Uwe Broulik wrote: > That's why on third (and last) restart attempt it'll bring up DrKonqi > nonetheless. Good point nonetheless. so it does. +1 - David --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/126517/#review90102 --- On Dec. 25, 2015, 4:24 p.m., Kai Uwe Broulik wrote: > > --- > This is an automatically generated e-mail. To reply, visit: > https://git.reviewboard.kde.org/r/126517/ > --- > > (Updated Dec. 25, 2015, 4:24 p.m.) > > > Review request for Plasma. > > > Repository: plasma-workspace > > > Description > --- > > This enables the StatusNotifier option introduced in Review 126516, see > Review 126515 for more information. > > When Plasma reached the recursion counter it disables the passive option > since there will most likely be no System Tray to show this error and then > something's really wrong in the first place. > > > Diffs > - > > shell/shellmanager.cpp 0a6ac5e > > Diff: https://git.reviewboard.kde.org/r/126517/diff/ > > > Testing > --- > > Crashed plasma two times, it restarted and I got two SNIs, crashed it another > time, got black screen with DrKonqi dialog. > > > Thanks, > > Kai Uwe Broulik > > ___ Plasma-devel mailing list Plasma-devel@kde.org https://mail.kde.org/mailman/listinfo/plasma-devel
Re: Review Request 126515: [DrKonqi] RFC: Support showing a StatusNotifierItem instead of bringing up the dialog right away
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/126515/#review90104 --- code looks fine. Though please do the TODOs before merging as historically we have a have a habbit of them not getting done Also your StatusNotifierItem leaks - but it proably wants to be the lifespan of the application anyway. - David Edmundson On Dec. 25, 2015, 4:24 p.m., Kai Uwe Broulik wrote: > > --- > This is an automatically generated e-mail. To reply, visit: > https://git.reviewboard.kde.org/r/126515/ > --- > > (Updated Dec. 25, 2015, 4:24 p.m.) > > > Review request for KDE Frameworks, Plasma, KDE Usability, and Martin Gräßlin. > > > Repository: plasma-workspace > > > Description > --- > > This adds a new "--passive" option to DrKonqi where it will only show a > StatusNotifierItem rather than bringing up the crash dialog right away. > > This can be useful for auto-restarting shell services (like plasmashell, > krunner, kded) to improve the perceived quality of the product. > > On Windows RT, for example, the guidelines even explicitly say "rather just > dump the user on the home screen than telling him something went wrong, so he > can just quickly start the app again instead of being annoyed by an error > message". On iOS you also just get dropped on the home screen. Windows > desktop shows a "Searching for a solution" dialog which was *the* major > annoyance when something crashed, rather than the actual crash. > > Video here: https://www.youtube.com/watch?v=t0ZLs-juYKc > > > Diffs > - > > drkonqi/statusnotifier.cpp PRE-CREATION > drkonqi/statusnotifier.h PRE-CREATION > drkonqi/CMakeLists.txt eaeaad4 > drkonqi/main.cpp 7cbaae7 > > Diff: https://git.reviewboard.kde.org/r/126515/diff/ > > > Testing > --- > > I crashed plasmashell, it restarted so fast that you didn't even have a black > screen inbetween, just the panel restarting. Afterwards I got a SNI which > opened DrKonqi when tapped. > > The SNI disappears after 1 minute because if you didn't bother to look after > it by then, you probably forgot what you did to cause the crash anyway :) > > > Thanks, > > Kai Uwe Broulik > > ___ Plasma-devel mailing list Plasma-devel@kde.org https://mail.kde.org/mailman/listinfo/plasma-devel
Re: Review Request 126517: [ShellManager] Make use of StatusNotifier DrKonqi
> On Dez. 25, 2015, 5:22 nachm., David Edmundson wrote: > > On first thought this is awesome. I like the video. It looks good. > > > > but now I'm going to point something out that's going to ruin your Christmas > > > > If Plasma crashes, you're going to try and show an SNI *in plasma* ... > > which isn't running. > > That's fine for a single crash, not for ones that block startup. > > > > I'm sure we can workaround this somehow - there's a timer in KCrash > > internals where we can know if something is a crash within the first 20 > > seconds of loading. That's why on third (and last) restart attempt it'll bring up DrKonqi nonetheless. Good point nonetheless. - Kai Uwe --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/126517/#review90102 --- On Dez. 25, 2015, 4:24 nachm., Kai Uwe Broulik wrote: > > --- > This is an automatically generated e-mail. To reply, visit: > https://git.reviewboard.kde.org/r/126517/ > --- > > (Updated Dez. 25, 2015, 4:24 nachm.) > > > Review request for Plasma. > > > Repository: plasma-workspace > > > Description > --- > > This enables the StatusNotifier option introduced in Review 126516, see > Review 126515 for more information. > > When Plasma reached the recursion counter it disables the passive option > since there will most likely be no System Tray to show this error and then > something's really wrong in the first place. > > > Diffs > - > > shell/shellmanager.cpp 0a6ac5e > > Diff: https://git.reviewboard.kde.org/r/126517/diff/ > > > Testing > --- > > Crashed plasma two times, it restarted and I got two SNIs, crashed it another > time, got black screen with DrKonqi dialog. > > > Thanks, > > Kai Uwe Broulik > > ___ Plasma-devel mailing list Plasma-devel@kde.org https://mail.kde.org/mailman/listinfo/plasma-devel
Re: Review Request 126517: [ShellManager] Make use of StatusNotifier DrKonqi
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/126517/#review90102 --- On first thought this is awesome. I like the video. It looks good. but now I'm going to point something out that's going to ruin your Christmas If Plasma crashes, you're going to try and show an SNI *in plasma* ... which isn't running. That's fine for a single crash, not for ones that block startup. I'm sure we can workaround this somehow - there's a timer in KCrash internals where we can know if something is a crash within the first 20 seconds of loading. - David Edmundson On Dec. 25, 2015, 4:24 p.m., Kai Uwe Broulik wrote: > > --- > This is an automatically generated e-mail. To reply, visit: > https://git.reviewboard.kde.org/r/126517/ > --- > > (Updated Dec. 25, 2015, 4:24 p.m.) > > > Review request for Plasma. > > > Repository: plasma-workspace > > > Description > --- > > This enables the StatusNotifier option introduced in Review 126516, see > Review 126515 for more information. > > When Plasma reached the recursion counter it disables the passive option > since there will most likely be no System Tray to show this error and then > something's really wrong in the first place. > > > Diffs > - > > shell/shellmanager.cpp 0a6ac5e > > Diff: https://git.reviewboard.kde.org/r/126517/diff/ > > > Testing > --- > > Crashed plasma two times, it restarted and I got two SNIs, crashed it another > time, got black screen with DrKonqi dialog. > > > Thanks, > > Kai Uwe Broulik > > ___ Plasma-devel mailing list Plasma-devel@kde.org https://mail.kde.org/mailman/listinfo/plasma-devel
Re: Review Request 126515: [DrKonqi] RFC: Support showing a StatusNotifierItem instead of bringing up the dialog right away
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/126515/#review90100 --- Usability: I envisioned this to be used for auto-restarting shell services and not to be used by applications. Another interesting thought could be enabling this by default for all applications but for regular applications trigger a desktop notification (with report bug / restart app) together with the tray icon. Mockup: http://wstaw.org/m/2015/12/25/drkonqipassive.png - Kai Uwe Broulik On Dez. 25, 2015, 4:24 nachm., Kai Uwe Broulik wrote: > > --- > This is an automatically generated e-mail. To reply, visit: > https://git.reviewboard.kde.org/r/126515/ > --- > > (Updated Dez. 25, 2015, 4:24 nachm.) > > > Review request for KDE Frameworks, Plasma, KDE Usability, and Martin Gräßlin. > > > Repository: plasma-workspace > > > Description > --- > > This adds a new "--passive" option to DrKonqi where it will only show a > StatusNotifierItem rather than bringing up the crash dialog right away. > > This can be useful for auto-restarting shell services (like plasmashell, > krunner, kded) to improve the perceived quality of the product. > > On Windows RT, for example, the guidelines even explicitly say "rather just > dump the user on the home screen than telling him something went wrong, so he > can just quickly start the app again instead of being annoyed by an error > message". On iOS you also just get dropped on the home screen. Windows > desktop shows a "Searching for a solution" dialog which was *the* major > annoyance when something crashed, rather than the actual crash. > > Video here: https://www.youtube.com/watch?v=t0ZLs-juYKc > > > Diffs > - > > drkonqi/statusnotifier.cpp PRE-CREATION > drkonqi/statusnotifier.h PRE-CREATION > drkonqi/CMakeLists.txt eaeaad4 > drkonqi/main.cpp 7cbaae7 > > Diff: https://git.reviewboard.kde.org/r/126515/diff/ > > > Testing > --- > > I crashed plasmashell, it restarted so fast that you didn't even have a black > screen inbetween, just the panel restarting. Afterwards I got a SNI which > opened DrKonqi when tapped. > > The SNI disappears after 1 minute because if you didn't bother to look after > it by then, you probably forgot what you did to cause the crash anyway :) > > > Thanks, > > Kai Uwe Broulik > > ___ Plasma-devel mailing list Plasma-devel@kde.org https://mail.kde.org/mailman/listinfo/plasma-devel
Review Request 126515: [DrKonqi] RFC: Support showing a StatusNotifierItem instead of bringing up the dialog right away
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/126515/ --- Review request for KDE Frameworks, Plasma, KDE Usability, and Martin Gräßlin. Repository: plasma-workspace Description --- This adds a new "--passive" option to DrKonqi where it will only show a StatusNotifierItem rather than bringing up the crash dialog right away. This can be useful for auto-restarting shell services (like plasmashell, krunner, kded) to improve the perceived quality of the product. On Windows RT, for example, the guidelines even explicitly say "rather just dump the user on the home screen than telling him something went wrong, so he can just quickly start the app again instead of being annoyed by an error message". On iOS you also just get dropped on the home screen. Windows desktop shows a "Searching for a solution" dialog which was *the* major annoyance when something crashed, rather than the actual crash. Video here: https://www.youtube.com/watch?v=t0ZLs-juYKc Diffs - drkonqi/statusnotifier.cpp PRE-CREATION drkonqi/statusnotifier.h PRE-CREATION drkonqi/CMakeLists.txt eaeaad4 drkonqi/main.cpp 7cbaae7 Diff: https://git.reviewboard.kde.org/r/126515/diff/ Testing --- I crashed plasmashell, it restarted so fast that you didn't even have a black screen inbetween, just the panel restarting. Afterwards I got a SNI which opened DrKonqi when tapped. The SNI disappears after 1 minute because if you didn't bother to look after it by then, you probably forgot what you did to cause the crash anyway :) Thanks, Kai Uwe Broulik ___ Plasma-devel mailing list Plasma-devel@kde.org https://mail.kde.org/mailman/listinfo/plasma-devel
Review Request 126517: [ShellManager] Make use of StatusNotifier DrKonqi
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/126517/ --- Review request for Plasma. Repository: plasma-workspace Description --- This enables the StatusNotifier option introduced in Review 126516, see Review 126515 for more information. When Plasma reached the recursion counter it disables the passive option since there will most likely be no System Tray to show this error and then something's really wrong in the first place. Diffs - shell/shellmanager.cpp 0a6ac5e Diff: https://git.reviewboard.kde.org/r/126517/diff/ Testing --- Crashed plasma two times, it restarted and I got two SNIs, crashed it another time, got black screen with DrKonqi dialog. Thanks, Kai Uwe Broulik ___ Plasma-devel mailing list Plasma-devel@kde.org https://mail.kde.org/mailman/listinfo/plasma-devel
Review Request 126514: [Milou/KRunner] Support runner actions
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/126514/ --- Review request for Plasma, KDE Usability and Vishesh Handa. Repository: milou Description --- This brings back support for additional runner actions. A runner can specify supplemental actions to a search result. It is deliberately no ToolButton as that would look weird, the Highlight blends in much better, although I'm not too happy with the huge Plasma tooltip. I'm not sure how keyboard interaction should work as both Tab and arrow keys would conflict with navigation. Diffs - lib/CMakeLists.txt 1123b60 lib/qml/ResultDelegate.qml 6d60bf2 lib/qml/ResultsView.qml b36cc19 lib/sourcesmodel.h 6c2d205 lib/sourcesmodel.cpp 65c69b3 Diff: https://git.reviewboard.kde.org/r/126514/diff/ Testing --- Hardly any runner makes use of this, though. I got the kill runner to send sigterm and sigkill, and also some mysterious ktp runer I couldn't find where it is offers "Start Chat" and "Open Chat Log" actions. The Solid runner supposedly also adds actions but I couldn't get that to work. Unfortunately I had to add QtWidgets for QAction :( File Attachments Runner action https://git.reviewboard.kde.org/media/uploaded/files/2015/12/25/3b70cc42-f3f9-42e5-aa5c-3b30067ae06e__krunneractions.png Thanks, Kai Uwe Broulik ___ Plasma-devel mailing list Plasma-devel@kde.org https://mail.kde.org/mailman/listinfo/plasma-devel
Re: RFC: split platformtheme plugin from frameworkintegration and move to kde/workspace
On Tuesday, 2015-12-08, 07:49:54, Martin Graesslin wrote: > that good as ours. That's not the point. We want GTK applications to pick > our file dialog in Plasma and we want our application's to pick GNOME's > file dialog in GNOME. Our subjective feeling of superior user experience is > pointless if the user is used to the platform's file dialog. It is probably old news to the more active contributors but I just came across this [1] while working through my huge FOSS mailinglist backlog, Looks like GTK+ will finally have the capability to use the plaftform's file dialogs! Looking forward to having the KDE dialog when using Gimp or Inkscape :) Cheers, Kevin [1] https://blogs.gnome.org/alexl/2015/11/05/native-file-choosers-in-gtk/ -- Kevin Krammer, KDE developer, xdg-utils developer KDE user support, developer mentoring signature.asc Description: This is a digitally signed message part. ___ Plasma-devel mailing list Plasma-devel@kde.org https://mail.kde.org/mailman/listinfo/plasma-devel
icons stylesheet
Hi Marco, how the scripts apply-stylesheet.sh work? the owncloud and system updates icons from breeze-icons are "broken" and I don't know how to change it. cu Andreas Kainz ___ Plasma-devel mailing list Plasma-devel@kde.org https://mail.kde.org/mailman/listinfo/plasma-devel
Re: Review Request 126373: change the taskbar color from blue to gray
> On Dez. 20, 2015, 8:45 nachm., Eike Hein wrote: > > General PSA: I will soon push a fixed-up version of tasks.svgz to the repo > > that cleans up all the random margins and inconsistent corner styles in it > > that currently mess up alignments inside and outside of task buttons as > > well as corner appearance. Whatever the outcome of this, it needs to be > > rebased against those fixes first. > > Uri Herrera wrote: > The margins are not random. I literally had to use kruler and kmag to > properly align that. I even used a layered file with a red rectangle in it > because that area consists of a deadzone. I pointed that out when I made the > reveiw request originally. > > Eike Hein wrote: > The inside margins are currently inconsistent at the top and bottom > causing the icon to be vertically misaligned. The outside margins are > inconsistent as well - I think it's trying to position the button well inside > the panel, but this is an incorrect approach to the problem (the optical > imbalance with consistent margins is caused by how the panel inside margins > function, which is something we need to address in the theming system - my > approach to fix it was rejected, I'll be trying again though). The outside > margins and the corners (top ones rounded, bottom ones not) subtly break > things when the panel is anywhere but the bottom. > > I'm working on a series of changes to make the panel and Task Manager > finally pixel-perfect - fix the theme, fix the panel controller when > resizing, fix the panel default size, and fix something in FrameSVG/the panel > containment. > > Johan Ouwerkerk wrote: > In any case, the top version (new draft?) has markedly better alignment > of the text with icons... The other two versions seem vertically imbalanced > (more margin on the bottom than at the top). > Ideally, the icon is vertically aligned with the baseline of the text > such that ascenders and descenders extend equally far beyond the > corresponding border of the icon, hough this may be rather tricky when > considering languages with different baseline conventions (hanging or > variable or none). > > Eike Hein wrote: > The top version on the screenshot still shows all of the problems I'm > currently fixing FWIW. > > Eike Hein wrote: > I've now pushed the cleaned up tasks.svgz to git master. This fixes: > > * Fix inconsistent inside margins causing vertical misalignment inside > Task Buttons. > * Fix inconsistent outside margins causing incorrect alignment inside > panel. > * Fix outside margins and button corner appearance in locations other > than South. > * Align button frames with things like the pager edges. > * Fix a few instances of incorrect color scheme application. > * Remove many unused elements. > > This is what this gets us: http://i.imgur.com/7KaeJH5.png > > This took *hours* to fix. If anyone breaks these margins I will *eat* > them. Please base future changes on this version of the file! > > Johan Ouwerkerk wrote: > Much better! Appreciated. > > Perhaps a test application that sets window name in a variety of scripts > might be useful to check how it performs on various scripts...? > > Eike Hein wrote: > I test CJK and Arabic fairly regularly. > > andreas kainz wrote: > as I'm to jung to die can you change the behavior from your new > tasks.svgz file to > > 3. add blue boarder as all other tasks have (see first line in > taskbar3.png) > > as suggested above. Looking good, Eike, I'm just not sure about the gray background for minimized tasks. It makes them more prominent than the non-minimized ones - I liked the frameless variant for those better. - Kai Uwe --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/126373/#review89779 --- On Dez. 16, 2015, 7:23 nachm., andreas kainz wrote: > > --- > This is an automatically generated e-mail. To reply, visit: > https://git.reviewboard.kde.org/r/126373/ > --- > > (Updated Dez. 16, 2015, 7:23 nachm.) > > > Review request for Plasma, Marco Martin and Uri Herrera. > > > Repository: plasma-framework > > > Description > --- > > Problem > === > with the new taskbar we look that the look and feel is consistent between > plasma and the applications therefore Uri change the selected application > taskbar button to a blue background. The problem is that the blue background > couldn't be the same blue than e.g. in the dolphin sidebar cause when you > select an element in the sidebar the text color change from "black" to white > which isn't possible in the system tray and than the contrast isn't that good > in the
Re: Review Request 126373: change the taskbar color from blue to gray
> On Dez. 20, 2015, 8:45 nachm., Eike Hein wrote: > > General PSA: I will soon push a fixed-up version of tasks.svgz to the repo > > that cleans up all the random margins and inconsistent corner styles in it > > that currently mess up alignments inside and outside of task buttons as > > well as corner appearance. Whatever the outcome of this, it needs to be > > rebased against those fixes first. > > Uri Herrera wrote: > The margins are not random. I literally had to use kruler and kmag to > properly align that. I even used a layered file with a red rectangle in it > because that area consists of a deadzone. I pointed that out when I made the > reveiw request originally. > > Eike Hein wrote: > The inside margins are currently inconsistent at the top and bottom > causing the icon to be vertically misaligned. The outside margins are > inconsistent as well - I think it's trying to position the button well inside > the panel, but this is an incorrect approach to the problem (the optical > imbalance with consistent margins is caused by how the panel inside margins > function, which is something we need to address in the theming system - my > approach to fix it was rejected, I'll be trying again though). The outside > margins and the corners (top ones rounded, bottom ones not) subtly break > things when the panel is anywhere but the bottom. > > I'm working on a series of changes to make the panel and Task Manager > finally pixel-perfect - fix the theme, fix the panel controller when > resizing, fix the panel default size, and fix something in FrameSVG/the panel > containment. > > Johan Ouwerkerk wrote: > In any case, the top version (new draft?) has markedly better alignment > of the text with icons... The other two versions seem vertically imbalanced > (more margin on the bottom than at the top). > Ideally, the icon is vertically aligned with the baseline of the text > such that ascenders and descenders extend equally far beyond the > corresponding border of the icon, hough this may be rather tricky when > considering languages with different baseline conventions (hanging or > variable or none). > > Eike Hein wrote: > The top version on the screenshot still shows all of the problems I'm > currently fixing FWIW. > > Eike Hein wrote: > I've now pushed the cleaned up tasks.svgz to git master. This fixes: > > * Fix inconsistent inside margins causing vertical misalignment inside > Task Buttons. > * Fix inconsistent outside margins causing incorrect alignment inside > panel. > * Fix outside margins and button corner appearance in locations other > than South. > * Align button frames with things like the pager edges. > * Fix a few instances of incorrect color scheme application. > * Remove many unused elements. > > This is what this gets us: http://i.imgur.com/7KaeJH5.png > > This took *hours* to fix. If anyone breaks these margins I will *eat* > them. Please base future changes on this version of the file! > > Johan Ouwerkerk wrote: > Much better! Appreciated. > > Perhaps a test application that sets window name in a variety of scripts > might be useful to check how it performs on various scripts...? > > Eike Hein wrote: > I test CJK and Arabic fairly regularly. as I'm to jung to die can you change the behavior from your new tasks.svgz file to 3. add blue boarder as all other tasks have (see first line in taskbar3.png) as suggested above. - andreas --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/126373/#review89779 --- On Dez. 16, 2015, 7:23 nachm., andreas kainz wrote: > > --- > This is an automatically generated e-mail. To reply, visit: > https://git.reviewboard.kde.org/r/126373/ > --- > > (Updated Dez. 16, 2015, 7:23 nachm.) > > > Review request for Plasma, Marco Martin and Uri Herrera. > > > Repository: plasma-framework > > > Description > --- > > Problem > === > with the new taskbar we look that the look and feel is consistent between > plasma and the applications therefore Uri change the selected application > taskbar button to a blue background. The problem is that the blue background > couldn't be the same blue than e.g. in the dolphin sidebar cause when you > select an element in the sidebar the text color change from "black" to white > which isn't possible in the system tray and than the contrast isn't that good > in the panel for selected application (gray text on blue background). In > addition the blue is very visible. see screenshot before. > > Solution > > I like that we use the same color language but when you look at the dolphin > toolbar on top selecte
Re: Review Request 121010: Keep track of notifications that were covered by a fullscreen window
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/121010/#review90090 --- Ping. (Is a roundtrip to the X Server that bad for the comparatively rare event of getting a notification?) - Kai Uwe Broulik On Nov. 6, 2014, 11:47 vorm., Kai Uwe Broulik wrote: > > --- > This is an automatically generated e-mail. To reply, visit: > https://git.reviewboard.kde.org/r/121010/ > --- > > (Updated Nov. 6, 2014, 11:47 vorm.) > > > Review request for Plasma, KDE Usability and Martin Gräßlin. > > > Repository: plasma-workspace > > > Description > --- > > This patch improves fullscreen handling in notifications. > > The type of the NotificationPopup is changed to Dialog.Notification because > that's what it's for, I think there's been a bit of a miscommunication about > that (I think there needs to be another layer for the OSD because that one > should be above all..). > The Notification type keeps the window above (sorta fixing Bug 338946 because > the close button now works as expected now) everything but not a fullscreen > window if it's active. This patch checks when a notification arrives whether > the active window is fullscreen (concealing the notification) and adds it to > the history. To indicate that you missed notifications while you were > fullscreen, the icon glows blue (until you interact with a notification or > open the plasmoid). > > > Diffs > - > > applets/notifications/package/contents/ui/NotificationIcon.qml 1449c19 > applets/notifications/package/contents/ui/NotificationPopup.qml 212f1e6 > applets/notifications/package/contents/ui/Notifications.qml acdccdd > applets/notifications/package/contents/ui/main.qml 3c2fa13 > applets/notifications/plugin/notificationshelper.h af8f6fa > applets/notifications/plugin/notificationshelper.cpp 425f0d6 > > Diff: https://git.reviewboard.kde.org/r/121010/diff/ > > > Testing > --- > > Fullscreened my IDE, "sleep 2; knotificationdbustest", waited, > unfullscreened, icon was blue and the history a few notifications in it. > > > File Attachments > > > Pending notification in tray > > https://git.reviewboard.kde.org/media/uploaded/files/2014/11/06/3f2141b2-1452-4e5c-8b00-782c571ad34b__notificationpendingthing.png > > > Thanks, > > Kai Uwe Broulik > > ___ Plasma-devel mailing list Plasma-devel@kde.org https://mail.kde.org/mailman/listinfo/plasma-devel
Re: Review Request 126462: [Task Manager] Drop application name from "Recent Documents" heading
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/126462/ --- (Updated Dec. 25, 2015, 10:44 a.m.) Status -- This change has been marked as submitted. Review request for Plasma and Eike Hein. Changes --- Submitted with commit 7e6bfdea58e1a5a89103114293f50403677f996b by Kai Uwe Broulik to branch master. Repository: plasma-desktop Description --- It makes the menu overly huge and is not too useful imho. Diffs - applets/taskmanager/plugin/backend.cpp ef7bb2a Diff: https://git.reviewboard.kde.org/r/126462/diff/ Testing --- No more title Thanks, Kai Uwe Broulik ___ Plasma-devel mailing list Plasma-devel@kde.org https://mail.kde.org/mailman/listinfo/plasma-devel
Re: Review Request 126513: Refactor kcm touchpad to fix some issue related to hotplug touchpad
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/126513/#review90089 --- This might also fix bug#349545 - Rajeesh K V On Dec. 25, 2015, 9:27 a.m., Xuetian Weng wrote: > > --- > This is an automatically generated e-mail. To reply, visit: > https://git.reviewboard.kde.org/r/126513/ > --- > > (Updated Dec. 25, 2015, 9:27 a.m.) > > > Review request for Plasma and David Edmundson. > > > Bugs: 356923 > https://bugs.kde.org/show_bug.cgi?id=356923 > > > Repository: plasma-desktop > > > Description > --- > > Currently, there are some mixed issue in kcm touchpad related to libinput > backend and hot plug. > There are several issues: > 1. only one backend maybe used at runtime, either synaptics or libinput. But > libinput backend will only be used if there is a libinput backend present. > Which means if libinput touchpad is not present at login, kded will not be > able to properly support them. > 2. hotplug touchpad will always set touchpad to disabled. > 3. hotplug touchpad will not get configuration applied. > 4. The libinput devices detection may pick some non-touchpad device. > > To solve these problems, following changes are made: > 1. Remove SynapticsBackend and LibinputBackend and add three new classes > called XlibTouchpad/SynapticsTouchpad/LibinputTouchpad. So findTouchpad may > pick up both synaptics or libinput device, and no need to depend on a fixed > backend. This change also make it easier to support multiple touchpad device > easier if needed. > > 2. hotplug touchpad config not being applied (bug 356923) is mainly because > one can't apply settings to a disabled device. Change the statement order in > handleReset, apply settings first then set enable status. > > 3. findTouchpad() is changed to use XListInputDevices, because it provides > necessary information of device type in order to filter out non-touchpad > device. xf86-input-libinput doesn't have a unique atom for touchpad, so > current identityAtom cannot guarantee that findTouchpad always returns a > touchpad device for libinput. > > 4. On my surface pro 4 system, if touchpad is unpluged, the device will be > set to disabled automatically first, then touchpadDetached is called, which > makes m_enabled in TouchpadDisabled always to be false after unplug the > touchapd. This patch makes TouchpadDisabler have two different enabled > property. m_userRequestedState only stores the user requested status, so when > handleReset is called, it can properly set enabled state to the last user > requested state. > > > Diffs > - > > kcms/touchpad/src/applet/qml/contents/ui/touchpad.qml 8dec7c2 > kcms/touchpad/src/applet/touchpadengine.h 16b98c2 > kcms/touchpad/src/applet/touchpadengine.cpp eae429e > kcms/touchpad/src/backends/x11.cmake c9fcea8 > kcms/touchpad/src/backends/x11/libinputproperties.c 9dbf9ea > kcms/touchpad/src/backends/x11/libinputtouchpad.h PRE-CREATION > kcms/touchpad/src/backends/x11/libinputtouchpad.cpp PRE-CREATION > kcms/touchpad/src/backends/x11/propertyinfo.h PRE-CREATION > kcms/touchpad/src/backends/x11/propertyinfo.cpp PRE-CREATION > kcms/touchpad/src/backends/x11/synapticstouchpad.h PRE-CREATION > kcms/touchpad/src/backends/x11/synapticstouchpad.cpp PRE-CREATION > kcms/touchpad/src/backends/x11/synclientproperties.h 43d18d2 > kcms/touchpad/src/backends/x11/synclientproperties.c 5fd1ed6 > kcms/touchpad/src/backends/x11/xlibbackend.h 7cecb4a > kcms/touchpad/src/backends/x11/xlibbackend.cpp b55a45f > kcms/touchpad/src/backends/x11/xlibtouchpad.h PRE-CREATION > kcms/touchpad/src/backends/x11/xlibtouchpad.cpp PRE-CREATION > kcms/touchpad/src/kded/kded.h 9b8fe6e > kcms/touchpad/src/kded/kded.cpp 409126b > kcms/touchpad/src/touchpadbackend.h b225ed9 > > Diff: https://git.reviewboard.kde.org/r/126513/diff/ > > > Testing > --- > > Without touchpad -> login -> plug touchpad -> settings applied to device > using libinput driver. > unplug -> replug -> settings applied to device using libinput driver. > > > Disable touchpad -> applet shows up -> unplug touchpad -> applet hides -> > replug touchpad -> touchpad is still disabled, and applet shows up -> enable > touhcpad -> applet hides. > > > Thanks, > > Xuetian Weng > > ___ Plasma-devel mailing list Plasma-devel@kde.org https://mail.kde.org/mailman/listinfo/plasma-devel
Re: Review Request 126513: Refactor kcm touchpad to fix some issue related to hotplug touchpad
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/126513/ --- (Updated Dec. 25, 2015, 9:27 a.m.) Review request for Plasma and David Edmundson. Changes --- update testing Bugs: 356923 https://bugs.kde.org/show_bug.cgi?id=356923 Repository: plasma-desktop Description --- Currently, there are some mixed issue in kcm touchpad related to libinput backend and hot plug. There are several issues: 1. only one backend maybe used at runtime, either synaptics or libinput. But libinput backend will only be used if there is a libinput backend present. Which means if libinput touchpad is not present at login, kded will not be able to properly support them. 2. hotplug touchpad will always set touchpad to disabled. 3. hotplug touchpad will not get configuration applied. 4. The libinput devices detection may pick some non-touchpad device. To solve these problems, following changes are made: 1. Remove SynapticsBackend and LibinputBackend and add three new classes called XlibTouchpad/SynapticsTouchpad/LibinputTouchpad. So findTouchpad may pick up both synaptics or libinput device, and no need to depend on a fixed backend. This change also make it easier to support multiple touchpad device easier if needed. 2. hotplug touchpad config not being applied (bug 356923) is mainly because one can't apply settings to a disabled device. Change the statement order in handleReset, apply settings first then set enable status. 3. findTouchpad() is changed to use XListInputDevices, because it provides necessary information of device type in order to filter out non-touchpad device. xf86-input-libinput doesn't have a unique atom for touchpad, so current identityAtom cannot guarantee that findTouchpad always returns a touchpad device for libinput. 4. On my surface pro 4 system, if touchpad is unpluged, the device will be set to disabled automatically first, then touchpadDetached is called, which makes m_enabled in TouchpadDisabled always to be false after unplug the touchapd. This patch makes TouchpadDisabler have two different enabled property. m_userRequestedState only stores the user requested status, so when handleReset is called, it can properly set enabled state to the last user requested state. Diffs - kcms/touchpad/src/applet/qml/contents/ui/touchpad.qml 8dec7c2 kcms/touchpad/src/applet/touchpadengine.h 16b98c2 kcms/touchpad/src/applet/touchpadengine.cpp eae429e kcms/touchpad/src/backends/x11.cmake c9fcea8 kcms/touchpad/src/backends/x11/libinputproperties.c 9dbf9ea kcms/touchpad/src/backends/x11/libinputtouchpad.h PRE-CREATION kcms/touchpad/src/backends/x11/libinputtouchpad.cpp PRE-CREATION kcms/touchpad/src/backends/x11/propertyinfo.h PRE-CREATION kcms/touchpad/src/backends/x11/propertyinfo.cpp PRE-CREATION kcms/touchpad/src/backends/x11/synapticstouchpad.h PRE-CREATION kcms/touchpad/src/backends/x11/synapticstouchpad.cpp PRE-CREATION kcms/touchpad/src/backends/x11/synclientproperties.h 43d18d2 kcms/touchpad/src/backends/x11/synclientproperties.c 5fd1ed6 kcms/touchpad/src/backends/x11/xlibbackend.h 7cecb4a kcms/touchpad/src/backends/x11/xlibbackend.cpp b55a45f kcms/touchpad/src/backends/x11/xlibtouchpad.h PRE-CREATION kcms/touchpad/src/backends/x11/xlibtouchpad.cpp PRE-CREATION kcms/touchpad/src/kded/kded.h 9b8fe6e kcms/touchpad/src/kded/kded.cpp 409126b kcms/touchpad/src/touchpadbackend.h b225ed9 Diff: https://git.reviewboard.kde.org/r/126513/diff/ Testing (updated) --- Without touchpad -> login -> plug touchpad -> settings applied to device using libinput driver. unplug -> replug -> settings applied to device using libinput driver. Disable touchpad -> applet shows up -> unplug touchpad -> applet hides -> replug touchpad -> touchpad is still disabled, and applet shows up -> enable touhcpad -> applet hides. Thanks, Xuetian Weng ___ Plasma-devel mailing list Plasma-devel@kde.org https://mail.kde.org/mailman/listinfo/plasma-devel
Re: Review Request 126513: Refactor kcm touchpad to fix some issue related to hotplug touchpad
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/126513/ --- (Updated Dec. 25, 2015, 9:25 a.m.) Review request for Plasma and David Edmundson. Changes --- Some fixes on changes. Add missing flush() in applyConfig. Expose workingTouchpadFound to applet, don't disable applet if no touchpad found during initialization. Bugs: 356923 https://bugs.kde.org/show_bug.cgi?id=356923 Repository: plasma-desktop Description --- Currently, there are some mixed issue in kcm touchpad related to libinput backend and hot plug. There are several issues: 1. only one backend maybe used at runtime, either synaptics or libinput. But libinput backend will only be used if there is a libinput backend present. Which means if libinput touchpad is not present at login, kded will not be able to properly support them. 2. hotplug touchpad will always set touchpad to disabled. 3. hotplug touchpad will not get configuration applied. 4. The libinput devices detection may pick some non-touchpad device. To solve these problems, following changes are made: 1. Remove SynapticsBackend and LibinputBackend and add three new classes called XlibTouchpad/SynapticsTouchpad/LibinputTouchpad. So findTouchpad may pick up both synaptics or libinput device, and no need to depend on a fixed backend. This change also make it easier to support multiple touchpad device easier if needed. 2. hotplug touchpad config not being applied (bug 356923) is mainly because one can't apply settings to a disabled device. Change the statement order in handleReset, apply settings first then set enable status. 3. findTouchpad() is changed to use XListInputDevices, because it provides necessary information of device type in order to filter out non-touchpad device. xf86-input-libinput doesn't have a unique atom for touchpad, so current identityAtom cannot guarantee that findTouchpad always returns a touchpad device for libinput. 4. On my surface pro 4 system, if touchpad is unpluged, the device will be set to disabled automatically first, then touchpadDetached is called, which makes m_enabled in TouchpadDisabled always to be false after unplug the touchapd. This patch makes TouchpadDisabler have two different enabled property. m_userRequestedState only stores the user requested status, so when handleReset is called, it can properly set enabled state to the last user requested state. Diffs (updated) - kcms/touchpad/src/applet/qml/contents/ui/touchpad.qml 8dec7c2 kcms/touchpad/src/applet/touchpadengine.h 16b98c2 kcms/touchpad/src/applet/touchpadengine.cpp eae429e kcms/touchpad/src/backends/x11.cmake c9fcea8 kcms/touchpad/src/backends/x11/libinputproperties.c 9dbf9ea kcms/touchpad/src/backends/x11/libinputtouchpad.h PRE-CREATION kcms/touchpad/src/backends/x11/libinputtouchpad.cpp PRE-CREATION kcms/touchpad/src/backends/x11/propertyinfo.h PRE-CREATION kcms/touchpad/src/backends/x11/propertyinfo.cpp PRE-CREATION kcms/touchpad/src/backends/x11/synapticstouchpad.h PRE-CREATION kcms/touchpad/src/backends/x11/synapticstouchpad.cpp PRE-CREATION kcms/touchpad/src/backends/x11/synclientproperties.h 43d18d2 kcms/touchpad/src/backends/x11/synclientproperties.c 5fd1ed6 kcms/touchpad/src/backends/x11/xlibbackend.h 7cecb4a kcms/touchpad/src/backends/x11/xlibbackend.cpp b55a45f kcms/touchpad/src/backends/x11/xlibtouchpad.h PRE-CREATION kcms/touchpad/src/backends/x11/xlibtouchpad.cpp PRE-CREATION kcms/touchpad/src/kded/kded.h 9b8fe6e kcms/touchpad/src/kded/kded.cpp 409126b kcms/touchpad/src/touchpadbackend.h b225ed9 Diff: https://git.reviewboard.kde.org/r/126513/diff/ Testing --- Without touchpad -> login -> plug touchpad -> settings applied to device using libinput driver. unplug -> replug -> settings applied to device using libinput driver. Thanks, Xuetian Weng ___ Plasma-devel mailing list Plasma-devel@kde.org https://mail.kde.org/mailman/listinfo/plasma-devel
