[KScreen] [Bug 365040] kscreen KCM cannot be used if KSCREEN_BACKEND is not set

[Allan Sandfeld]

https://bugs.kde.org/show_bug.cgi?id=365040

--- Comment #5 from Allan Sandfeld  ---
carewolf@twilight% systemsettings5 
   org.kde.kactivities.lib.core: Killing the
consumer
kscreen: Failed to request backend: "org.freedesktop.DBus.Error.ServiceUnknown"
: "The name org.kde.KScreen was not provided by any .service files"
org.kde.kactivities.lib.core: Killing the consumer
org.kde.kactivities.lib.core: Killing the consumer
carewolf@twilight% export KSCREEN_BACKEND=KSC_XRandR.so
   carewolf@twilight% systemsettings5  
 
org.kde.kactivities.lib.core: Killing the consumer
carewolf@twilight%

-- 
You are receiving this mail because:
You are watching all bug changes.

[KScreen] [Bug 365040] kscreen KCM cannot be used if KSCREEN_BACKEND is not set

[Allan Sandfeld]

https://bugs.kde.org/show_bug.cgi?id=365040

Allan Sandfeld  changed:

   What|Removed |Added

 CC||k...@carewolf.com

--- Comment #4 from Allan Sandfeld  ---
I have this problem currently after doing a KDE5 custom build to /opt/kde5.

I first thought it was due to conflicts with system KDE but after removing that
kscreen still didn't work. Fortunately I found this bug and setting
KSCREEN_BACKEND manually makes the kcm owrk.

-- 
You are receiving this mail because:
You are watching all bug changes.

[KScreen] [Bug 365040] kscreen KCM cannot be used if KSCREEN_BACKEND is not set

[Andreas Cord-Landwehr via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=365040

Andreas Cord-Landwehr  changed:

   What|Removed |Added

 Status|UNCONFIRMED |RESOLVED
 Resolution|--- |WORKSFORME

--- Comment #3 from Andreas Cord-Landwehr  ---
After some major cleanup of my .config and .local directories (grown since the
first Qt5 based pre-alpha versions), I cannot reproduce this issue anymore. So,
I will close it.

-- 
You are receiving this mail because:
You are watching all bug changes.

[KScreen] [Bug 365040] kscreen KCM cannot be used if KSCREEN_BACKEND is not set

[Andreas Cord-Landwehr via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=365040

--- Comment #2 from Andreas Cord-Landwehr  ---
You are right; I looked at the wrong code location. Here is a probably more
verbose output. I now generated a backtrace for getting the debug message:

kcmshell5(30793)/(kscreen) KScreen::BackendManager::startBackend: starting
external backend launcher for ""

At this point, I have (added an assert in
"/opt/kde/src/kde/workspace/libkscreen/src/backendmanager.cpp", line=291):
#5  0x7fffda2f3bc6 in KScreen::BackendManager::startBackend(QString const&,
QMap const&) (this=0x6af270, backend=..., arguments=...)
at /opt/kde/src/kde/workspace/libkscreen/src/backendmanager.cpp:291
#6  0x7fffda2f3aca in KScreen::BackendManager::requestBackend()
(this=0x6af270) at
/opt/kde/src/kde/workspace/libkscreen/src/backendmanager.cpp:276
#7  0x7fffda2fe5cb in KScreen::ConfigOperationPrivate::requestBackend()
(this=0x9265f0) at
/opt/kde/src/kde/workspace/libkscreen/src/configoperation.cpp:45
#8  0x7fffda30019e in KScreen::GetConfigOperation::start() (this=0x92bfd0)
at /opt/kde/src/kde/workspace/libkscreen/src/getconfigoperation.cpp:178
#9  0x7fffda315dcd in
KScreen::ConfigOperation::qt_static_metacall(QObject*, QMetaObject::Call, int,
void**) (_o=0x92bfd0, _c=QMetaObject::InvokeMetaMethod, _id=1, _a=0x92bd10)
at /opt/kde/build/kde/workspace/libkscreen/src/moc_configoperation.cpp:81
#10 0x74d8ad07 in QMetaCallEvent::placeMetaCall(QObject*)
(this=0x93afb0, object=0x92bfd0) at kernel/qobject.cpp:495
#11 0x74d8bda0 in QObject::event(QEvent*) (this=0x92bfd0, e=0x93afb0)
at kernel/qobject.cpp:1256
#12 0x7594a4f2 in QApplicationPrivate::notify_helper(QObject*, QEvent*)
(this=0x61a250, receiver=0x92bfd0, e=0x93afb0) at kernel/qapplication.cpp:3804
#13 0x759478bd in QApplication::notify(QObject*, QEvent*)
(this=0x7fffd1f0, receiver=0x92bfd0, e=0x93afb0) at
kernel/qapplication.cpp:3163
#14 0x74d51bfa in QCoreApplication::notifyInternal2(QObject*, QEvent*)
(receiver=0x92bfd0, event=0x93afb0) at kernel/qcoreapplication.cpp:1015
#15 0x74d55c30 in QCoreApplication::sendEvent(QObject*, QEvent*)
(receiver=0x92bfd0, event=0x93afb0) at
../../include/QtCore/../../src/corelib/kernel/qcoreapplication.h:225
#16 0x74d52fe2 in QCoreApplicationPrivate::sendPostedEvents(QObject*,
int, QThreadData*) (receiver=0x0, event_type=0, data=0x6190d0) at
kernel/qcoreapplication.cpp:1650
#17 0x74d52962 in QCoreApplication::sendPostedEvents(QObject*, int)
(receiver=0x0, event_type=0) at kernel/qcoreapplication.cpp:1508
#18 0x74dcbb2c in postEventSourceDispatch(GSource*, GSourceFunc,
gpointer) (s=0x672fb0) at kernel/qeventdispatcher_glib.cpp:270
#19 0x7fffed56a1a7 in g_main_context_dispatch (context=0x7fffdc0016f0) at
/build/glib2.0-wnDt2X/glib2.0-2.48.1/./glib/gmain.c:3154
#20 0x7fffed56a1a7 in g_main_context_dispatch
(context=context@entry=0x7fffdc0016f0) at
/build/glib2.0-wnDt2X/glib2.0-2.48.1/./glib/gmain.c:3769
#21 0x7fffed56a400 in g_main_context_iterate
(context=context@entry=0x7fffdc0016f0, block=block@entry=1,
dispatch=dispatch@entry=1, self=)
at /build/glib2.0-wnDt2X/glib2.0-2.48.1/./glib/gmain.c:3840
---Type  to continue, or q  to quit---
#22 0x7fffed56a4ac in g_main_context_iteration (context=0x7fffdc0016f0,
may_block=1) at /build/glib2.0-wnDt2X/glib2.0-2.48.1/./glib/gmain.c:3901
#23 0x74dcc305 in
QEventDispatcherGlib::processEvents(QFlags)
(this=0x67ceb0, flags=...) at kernel/qeventdispatcher_glib.cpp:417
#24 0x7fffe78cce86 in
QPAEventDispatcherGlib::processEvents(QFlags)
(this=0x67ceb0, flags=...) at eventdispatchers/qeventdispatcher_glib.cpp:115
#25 0x74d4e570 in
QEventLoop::processEvents(QFlags)
(this=0x7fffd050, flags=...) at kernel/qeventloop.cpp:128
#26 0x74d4e86c in
QEventLoop::exec(QFlags) (this=0x7fffd050,
flags=...) at kernel/qeventloop.cpp:204
#27 0x74d52344 in QCoreApplication::exec() () at
kernel/qcoreapplication.cpp:1285
#28 0x7517d598 in QGuiApplication::exec() () at
kernel/qguiapplication.cpp:1607
#29 0x759472a1 in QApplication::exec() () at
kernel/qapplication.cpp:2979
#30 0x77bd4210 in kdemain(int, char**) (_argc=2, _argv=0x7fffdb18)
at /opt/kde/src/kde/workspace/kde-cli-tools/kcmshell/main.cpp:294
#31 0x004009a8 in main(int, char**) (argc=2, argv=0x7fffdb18) at
/opt/kde/build/kde/workspace/kde-cli-tools/kcmshell/kcmshell5_dummy.cpp:3

-- 
You are receiving this mail because:
You are watching all bug changes.

[KScreen] [Bug 365040] kscreen KCM cannot be used if KSCREEN_BACKEND is not set

[Sebastian Kügler via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=365040

--- Comment #1 from Sebastian Kügler  ---
I'm a bit puzzled.

kscreen-doctor and the kscreen KCM use the same method to load the backend
(GetConfigOperation). GetConfigOperation internally asks the backendmanager for
the preferred backend, if that's xrandr, it loads the plugin out of process,
any other backend loads in process.

On X11, it should never actually hit configoperation.cpp:137 since that's the
in-process codepath, are you sure it goes through that function?

backendmanager.cpp decides in its ctor which method (InProcess / OutOfProcess
it uses), after choosing the backend plugin.


Moreover, it works on my system as expected, no KSCREEN_BACKEND env var is set.

-- 
You are receiving this mail because:
You are watching all bug changes.