[Interest] [Qt4.8] UI events coming through despite QEventLoop::ExcludeUserInputEvents

[Andreas Pakulat]

Hi,

On Mon, Nov 10, 2014 at 11:35 AM, René J.V. <rjvbertin at gmail.com> wrote:

> Hi,
>
> What about this backtrace snippet I got this morning?
>
> #0  0x00007ffff51d1c6d in poll () at ../sysdeps/unix/syscall-template.S:81
> #1  0x00007fffeeda0fe4 in ?? () from /lib/x86_64-linux-gnu/libglib-2.0.so.0
> #2  0x00007fffeeda10ec in g_main_context_iteration () from
> /lib/x86_64-linux-gnu/libglib-2.0.so.0
> #3  0x00007ffff5b9772e in QEventDispatcherGlib::processEvents
> (this=0x616bb0, flags=...)
>     at kernel/qeventdispatcher_glib.cpp:436
> #4  0x00007ffff6615d86 in QGuiEventDispatcherGlib::processEvents
> (this=<optimized out>, flags=...)
>     at kernel/qguieventdispatcher_glib.cpp:204
> #5  0x00007ffff5b655af in QEventLoop::processEvents (this=this at entry=0x7fffffffd0e0,
> flags=...)
>     at kernel/qeventloop.cpp:149
> #6  0x00007ffff5b658ed in QEventLoop::exec (this=0x7fffffffd0e0,
> flags=...) at kernel/qeventloop.cpp:204
> #7  0x00007ffff5ffdaf3 in KJob::exec (this=0xfc3f00) at
> ../../kdecore/jobs/kjob.cpp:210
> #8  0x00007fff8f787b0b in KMKernel::cleanup (this=0xf10c60) at
> ../../kmail/kmkernel.cpp:1299
> #9  0x00007fff8fbd2cc6 in KMailPart::~KMailPart (this=0xd32c40,
> vtt=0x7fff8fdd7300 <VTT for KMailPart>)
>     at ../../kmail/kmail_part.cpp:119
> #10 0x00007fff8fbd2e73 in ~KMailPart (this=<optimized out>,
> this=<optimized out>) at ../../kmail/kmail_part.cpp:114
> #11 KMailPart::~KMailPart (this=0xd32c40) at ../../kmail/kmail_part.cpp:114
> #12 0x00007ffff7bcdc30 in KontactInterface::Plugin::~Plugin
> (this=0xbd32a0, vtt=<optimized out>)
>     at ../../kontactinterface/plugin.cpp:92
> #13 0x00007fff94837eda in ~KMailPlugin (this=<optimized out>,
> vtt=<optimized out>, this=<optimized out>,
>     vtt=<optimized out>) at
> ../../../../kontact/plugins/kmail/kmail_plugin.cpp:176
> #14 KMailPlugin::~KMailPlugin (this=0xbd32a0) at
> ../../../../kontact/plugins/kmail/kmail_plugin.cpp:174
> #15 0x00007fff94837f4c in KMailPlugin::~KMailPlugin (this=0xbd32a0)
>     at ../../../../kontact/plugins/kmail/kmail_plugin.cpp:174
> #16 0x00007ffff777d762 in Kontact::MainWindow::~MainWindow (this=0x955020,
>     vtt=0x7ffff799fff0 <VTT for Kontact::MainWindow>) at
> ../../../kontact/src/mainwindow.cpp:296
> #17 0x00007ffff777de07 in Kontact::MainWindow::~MainWindow (this=0x955020)
>     at ../../../kontact/src/mainwindow.cpp:271
> #18 0x00007ffff777debc in Kontact::MainWindow::~MainWindow (this=0x955020)
>     at ../../../kontact/src/mainwindow.cpp:271
> #19 0x00007ffff5b81388 in QObject::event (this=this at entry=0x955020,
> e=e at entry=0x67e4c90) at kernel/qobject.cpp:1203
> #20 0x00007ffff65be1f3 in QWidget::event (this=this at entry=0x955020,
> event=event at entry=0x67e4c90)
> ---Type <return> to continue, or q <return> to quit---q
>
> I sadly don't have qt4 sources available for the debugger on this machine
> so can't check in more detail, but it sure looks like the event processing
> function called by the QEventLoop::exec function called from KJob::exec
> calls a helper function that appears to be dedicated to UI events.
>

Without Qt sources thats impossible to say for sure, but it also receives
flags and it calls another function which again has no 'gui' in its name,
which then calls into glib.

Also the above does not look like a crash, its just a backtrace taken at a
point where glib polls some filedescriptor (which one is not visible and
neither is visible what happens inside glib due to missing debug symbols)

Andreas
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.qt-project.org/pipermail/interest/attachments/20141110/6964c436/attachment.html>