Bug 348212

Summary: Okular 1.0.0 is crashing an all documents (pdf,ps, epub...)
Product: [Applications] okular Reporter: Jorge Adriano <jorge.adriano>
Component: generalAssignee: Okular developers <okular-devel>
Status: RESOLVED DUPLICATE    
Severity: crash CC: aacid, ivan.cukic, luigi.toscano
Priority: NOR Keywords: drkonqi
Version: unspecified   
Target Milestone: ---   
Platform: openSUSE   
OS: Linux   
Latest Commit: Version Fixed In:
Sentry Crash Report:

Description Jorge Adriano 2015-05-25 17:51:10 UTC
Application: okular (1.0.0)

Qt Version: 5.4.1
Operating System: Linux 4.1.0-rc4-1.g9c4ab6c-desktop x86_64
Distribution: "openSUSE 13.1 (Bottle) (x86_64)"

-- Information about the crash:
- What I was doing when the application crashed:

Opening a document crashes okular. 

Simply running okular, without opening any doc, doesn't crash it. The UI appears and all seems fine. But on opening a document it crashes.

The crash can be reproduced every time.

-- Backtrace:
Application: Okular (okular), signal: Segmentation fault
Using host libthread_db library "/lib64/libthread_db.so.1".
[Current thread is 1 (Thread 0x7f1eb3e50880 (LWP 5797))]

Thread 2 (Thread 0x7f1ea053c700 (LWP 5798)):
#0  0x00007f1eafdd2fdd in poll () from /lib64/libc.so.6
#1  0x00007f1eae44132a in ?? () from /usr/lib64/libxcb.so.1
#2  0x00007f1eae442e1f in xcb_wait_for_event () from /usr/lib64/libxcb.so.1
#3  0x00007f1ea2514f19 in QXcbEventReader::run (this=0x98fdf0) at qxcbconnection.cpp:1105
#4  0x00007f1eb04ccb4f in QThreadPrivate::start (arg=0x98fdf0) at thread/qthread_unix.cpp:337
#5  0x00007f1eab3370db in start_thread () from /lib64/libpthread.so.0
#6  0x00007f1eafddbdad in clone () from /lib64/libc.so.6

Thread 1 (Thread 0x7f1eb3e50880 (LWP 5797)):
[KCrash Handler]
#5  0x00007f1eb3163d6f in at (i=0, this=0x696772614d510028) at /usr/include/qt5/QtCore/qvector.h:389
#6  pointer (this=<optimized out>) at /usr/include/qt5/QtCore/qresultstore.h:107
#7  value (this=<optimized out>) at /usr/include/qt5/QtCore/qresultstore.h:101
#8  resultReference (index=0, this=0xeac4c0) at /usr/include/qt5/QtCore/qfutureinterface.h:247
#9  result (this=0xeac4c0) at /usr/include/qt5/QtCore/qfuture.h:164
#10 result (this=0xeac4b0) at /usr/include/qt5/QtCore/qfuturewatcher.h:122
#11 operator() (__closure=<optimized out>) at /usr/src/debug/kactivities-5.10.0/src/utils/dbusfuture_p.h:169
#12 call (arg=<optimized out>, f=...) at /usr/include/qt5/QtCore/qobjectdefs_impl.h:494
#13 call<QtPrivate::List<>, void> (arg=<optimized out>, f=...) at /usr/include/qt5/QtCore/qobjectdefs_impl.h:551
#14 QtPrivate::QFunctorSlotObject<DBusFuture::continueWith(const QFuture<T>&, _Continuation&&) [with _ReturnType = QString; _Continuation = KActivities::Manager::serviceOwnerChanged(const QString&, const QString&, const QString&)::__lambda43]::__lambda29, 0, QtPrivate::List<>, void>::impl(int, QtPrivate::QSlotObjectBase *, QObject *, void **, bool *) (which=<optimized out>, this_=<optimized out>, r=<optimized out>, a=<optimized out>, ret=<optimized out>) at /usr/include/qt5/QtCore/qobject_impl.h:192
#15 0x00007f1eb06d7b06 in QObject::event (this=0xeac4b0, e=e@entry=0xeb0b30) at kernel/qobject.cpp:1245
#16 0x00007f1eb04c4ab2 in QFutureWatcherBase::event (this=<optimized out>, event=0xeb0b30) at thread/qfuturewatcher.cpp:342
#17 0x00007f1eb162fb7c in QApplicationPrivate::notify_helper (this=this@entry=0x9755a0, receiver=receiver@entry=0xeac4b0, e=e@entry=0xeb0b30) at kernel/qapplication.cpp:3720
#18 0x00007f1eb1634a76 in QApplication::notify (this=0x7ffd74a876c0, receiver=0xeac4b0, e=0xeb0b30) at kernel/qapplication.cpp:3503
#19 0x00007f1eb06a74a5 in QCoreApplication::notifyInternal (this=0x7ffd74a876c0, receiver=0xeac4b0, event=event@entry=0xeb0b30) at kernel/qcoreapplication.cpp:935
#20 0x00007f1eb06a933f in sendEvent (event=0xeb0b30, receiver=<optimized out>) at kernel/qcoreapplication.h:228
#21 QCoreApplicationPrivate::sendPostedEvents (receiver=receiver@entry=0x0, event_type=event_type@entry=0, data=0x96be00) at kernel/qcoreapplication.cpp:1552
#22 0x00007f1eb06a9978 in QCoreApplication::sendPostedEvents (receiver=receiver@entry=0x0, event_type=event_type@entry=0) at kernel/qcoreapplication.cpp:1410
#23 0x00007f1eb06ff393 in postEventSourceDispatch (s=0x9d4810) at kernel/qeventdispatcher_glib.cpp:271
#24 0x00007f1eaae26316 in g_main_context_dispatch () from /usr/lib64/libglib-2.0.so.0
#25 0x00007f1eaae26668 in ?? () from /usr/lib64/libglib-2.0.so.0
#26 0x00007f1eaae2670c in g_main_context_iteration () from /usr/lib64/libglib-2.0.so.0
#27 0x00007f1eb06fe80c in QEventDispatcherGlib::processEvents (this=0x9b0440, flags=...) at kernel/qeventdispatcher_glib.cpp:418
#28 0x00007f1eb06a540b in QEventLoop::exec (this=this@entry=0x7ffd74a875b0, flags=..., flags@entry=...) at kernel/qeventloop.cpp:204
#29 0x00007f1eb06aca76 in QCoreApplication::exec () at kernel/qcoreapplication.cpp:1188
#30 0x00007f1eb0e5666c in QGuiApplication::exec () at kernel/qguiapplication.cpp:1507
#31 0x00007f1eb162c255 in QApplication::exec () at kernel/qapplication.cpp:2956
#32 0x000000000040ec0f in main (argc=1, argv=<optimized out>) at /usr/src/debug/okular-4.99.40/shell/main.cpp:81

Possible duplicates by query: bug 342131.

Reported using DrKonqi
Comment 1 Albert Astals Cid 2015-05-25 18:02:11 UTC
Ivan?
Comment 2 Jorge Adriano 2015-05-25 18:23:57 UTC
May it's my system... let me check a few things out package-wise.
Comment 3 Jorge Adriano 2015-05-25 18:33:03 UTC
Double checked my packages, all seems fine. 
Okular 0.21.3 is working fine, as is evince
Comment 4 Jorge Adriano 2015-05-25 18:33:44 UTC
to be clear,
okular 0.21.3 is fine
okular 1.0.0 is crashing
Comment 5 Ivan Čukić 2015-05-26 08:11:22 UTC
This seems to be happening only on opensuse so far (I'm marking this as a duplicate).

 Jorge, are you able to compile libkactivities with patches that I provide so that we can test?
What output does 'kactivitymanagerd status' give you?

*** This bug has been marked as a duplicate of bug 347817 ***
Comment 6 Ivan Čukić 2015-05-26 08:12:52 UTC
p.s. I've blindly (since I can not reproduce the issue) created a patch that should deal with this, but it needs testing and someone to test the further changes.
Comment 7 Luigi Toscano 2015-05-26 08:32:59 UTC
OpenSUSE have not been provided the package in a long time. You marked it as coming from OpenSUSE RPMs, if this is the case please use the stable (kdelibs4-based) package instead of an unsupported branch.
Comment 8 Ivan Čukić 2015-05-26 09:27:52 UTC
Luigi, can you elaborate a bit? What is unsupported?
Comment 9 Luigi Toscano 2015-05-26 09:44:08 UTC
(In reply to Ivan Čukić from comment #8)
> Luigi, can you elaborate a bit? What is unsupported?

The frameworks branch of Okular is not supported.
Comment 10 Ivan Čukić 2015-05-26 09:48:56 UTC
Well, even if it isn't supported by openSUSE, it will be supported at some point (when it goes into master) and we should get this fixed before that happens. (it also affects other products which *are* currently supported like the new plasma and kwin)
Comment 11 Luigi Toscano 2015-05-26 10:20:06 UTC
Fine for fixing the issue, but the message for all users is "please do not expect the frameworks branch of Okular to work unless you are willing to fix it". Users relying on packages should use the stable versions or the versions which will get stable soon, and Okular is not in either of these categories right now.
Comment 12 Jorge Adriano 2015-05-26 11:14:54 UTC
(In reply to Luigi Toscano from comment #11)
> Fine for fixing the issue, but the message for all users is "please do not
> expect the frameworks branch of Okular to work unless you are willing to fix
> it". Users relying on packages should use the stable versions or the
> versions which will get stable soon, and Okular is not in either of these
> categories right now.

Hi Luigi, 

Indeed, I am aware of that. I have KDE 4.12 installed + wolfie's 5.x branch on the side
so I can 'drive test' some of the newer apps that interest me the most, like Okular.  Should a 5.x app fail, like in this case, I just move back to the corresponding in 4.12. So no drama. 

http://download.opensuse.org/repositories/home:/wolfi323:/branches:/KDE:/Frameworks5/openSUSE_13.1/
Comment 13 Jorge Adriano 2015-05-26 12:05:21 UTC
(In reply to Ivan Čukić from comment #5)
> This seems to be happening only on opensuse so far (I'm marking this as a
> duplicate).
> 
>  Jorge, are you able to compile libkactivities with patches that I provide
> so that we can test?
> What output does 'kactivitymanagerd status' give you?
> 
> *** This bug has been marked as a duplicate of bug 347817 ***

OK this seems a bit messy, so it may be a packaging problem. 

Like I said in another comment, I moved from the stable Frameworks 5 branch to wolfie's so I could test/use the newer Okular, which is what interests me the most. And on moving, I was careful to 'switch system packages' to wolfie's to make sure I didn't end up with a mix of stable/unstable. The repository is: http://download.opensuse.org/repositories/home:/wolfi323:/branches:/KDE:/Frameworks5/openSUSE_13.1/

I should also note that my current working desktop is 4.14, as the packages allow (for the most part), to keep both. 

So I am running  4.14, and /usr/bin/kactivitymanagerd refers to the 4.14 version, and this doesn't take the "status argument". There is also /usr/bin/kactivitymanagerd5, and its output is:

--------------------------------------------
kactivitymanagerd5 status
The service is running, version: 
--------------------------------------------
Like that ^, with no version is specified by the output. 

However, this made me realise the installed kactivitymanagerd5 is being provided by the stable Frameworks 5  repository and not wolfie's, as the latter installs binaries in /opt/kf5/bin/. Searching in wolfie's repository, it doesn't appear to provide any kactivitymanagerd (which I guess is why it wasn't replaced I guess, no dependency issues are detected though). I don't know if this omission hints at some packaging problem...

As for compiling it libkactivities myself, quite frankly I'd be a bit wary of trying that. Not much time on my hands and afraid to mess up mess up my working system. I think contacting openSUSE's wolfi323 could be the way to go. He should probably be interested in integrating the potential fix in his unstable branch
https://build.opensuse.org/project/show/home:wolfi323
Comment 14 Ivan Čukić 2015-05-26 13:31:39 UTC
If anyone looking at this is an opensuse packager, please look at https://community.kde.org/Frameworks/Coinstallability (or, if you know an oS KDE packager, please forward the link)