[rkward-tracker] [ rkward-Bugs-3433496 ] FIXED IN SVN: Crashed while typing text

Wed Nov 21 13:08:49 UTC 2012

Bugs item #3433496, was opened at 2011-11-04 05:48
Message generated for change (Settings changed) made by tfry
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=459007&aid=3433496&group_id=50231

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: CRASH
>Status: Closed
Resolution: Fixed
Priority: 5
Private: No
Submitted By: Milan Bouchet-Valat (nalimilan)
Assigned to: Thomas Friedrichsmeier (tfry)
Summary: FIXED IN SVN: Crashed while typing text

Initial Comment:
RKWard crashed today while I was typing text in the editor. It's the first time it happens, and I can't tell what I was doing special. I think I had just removed a selected text by hitting the backspace key.

I it can help, RKWard had been running for quite a long time, I had done many operations.

I'm using version 0.5.6z+0.5.7+devel1 on Fedora 15. Below is the trace from the relevant thread, and attach is the whole stacktrace.

Thread 1 (Thread 0x7fc8b1d60840 (LWP 22000)):
[KCrash Handler]
#6  0x000000000044499d in RKCodeCompletionModel::data (this=<optimized out>, index=<optimized out>, role=0) at /home/milan/Dev/RKWard/rkward/rkward/windows/rkcommandeditorwindow.cpp:1006
#7  0x00007fc89c0afad7 in KateCompletionModel::Item::Item(bool, KateCompletionModel*, HierarchicalModelHandler const&, QPair<KTextEditor::CodeCompletionModel*, QModelIndex>) () from /usr/lib64/libkatepartinterfaces.so.4
#8  0x00007fc89c0b3fd2 in KateCompletionModel::createItem(HierarchicalModelHandler const&, QModelIndex const&, bool) () from /usr/lib64/libkatepartinterfaces.so.4
#9  0x00007fc89c0b4e33 in KateCompletionModel::createItems(HierarchicalModelHandler const&, QModelIndex const&, bool) () from /usr/lib64/libkatepartinterfaces.so.4
#10 0x00007fc89c0b8018 in KateCompletionModel::createGroups() () from /usr/lib64/libkatepartinterfaces.so.4
#11 0x00007fc89c0b84f4 in KateCompletionModel::setCompletionModels(QList<KTextEditor::CodeCompletionModel*> const&) () from /usr/lib64/libkatepartinterfaces.so.4
#12 0x00007fc89c0a79b1 in KateCompletionWidget::startCompletion(KTextEditor::Range const&, QList<KTextEditor::CodeCompletionModel*> const&, KTextEditor::CodeCompletionModel::InvocationType) () from /usr/lib64/libkatepartinterfaces.so.4
#13 0x00007fc89c0a9198 in KateCompletionWidget::startCompletion(KTextEditor::Range const&, KTextEditor::CodeCompletionModel*, KTextEditor::CodeCompletionModel::InvocationType) () from /usr/lib64/libkatepartinterfaces.so.4
#14 0x000000000044b357 in RKCommandEditorWindow::tryCompletion (this=0x19ace60) at /home/milan/Dev/RKWard/rkward/rkward/windows/rkcommandeditorwindow.cpp:570
#15 0x000000000044ceea in RKCommandEditorWindow::qt_metacall (this=0x19ace60, _c=QMetaObject::InvokeMetaMethod, _id=<optimized out>, _a=0x7fff31bfb610) at /home/milan/Dev/RKWard/rkward/build/rkward/windows/rkcommandeditorwindow.moc:234
#16 0x00000035e4f6cfea in QMetaObject::activate (sender=0x1ad3ec0, m=<optimized out>, local_signal_index=<optimized out>, argv=0x0) at kernel/qobject.cpp:3278
#17 0x00000035e4f708d9 in QObject::event (this=0x1ad3ec0, e=<optimized out>) at kernel/qobject.cpp:1181
#18 0x0000003c069b9324 in notify_helper (e=0x7fff31bfbd10, receiver=0x1ad3ec0, this=0xe23e30) at kernel/qapplication.cpp:4481
#19 QApplicationPrivate::notify_helper (this=0xe23e30, receiver=0x1ad3ec0, e=0x7fff31bfbd10) at kernel/qapplication.cpp:4453
#20 0x0000003c069be1b1 in QApplication::notify (this=0x7fff31bfc050, receiver=0x1ad3ec0, e=0x7fff31bfbd10) at kernel/qapplication.cpp:4360
#21 0x0000003c056431c6 in KApplication::notify(QObject*, QEvent*) () from /usr/lib64/libkdeui.so.5
#22 0x00000035e4f5a20c in QCoreApplication::notifyInternal (this=0x7fff31bfc050, receiver=0x1ad3ec0, event=0x7fff31bfbd10) at kernel/qcoreapplication.cpp:787
#23 0x00000035e4f86c02 in sendEvent (event=0x7fff31bfbd10, receiver=<optimized out>) at kernel/qcoreapplication.h:215
#24 QTimerInfoList::activateTimers (this=0xe279e0) at kernel/qeventdispatcher_unix.cpp:603
#25 0x00000035e4f84444 in timerSourceDispatch (source=<optimized out>) at kernel/qeventdispatcher_glib.cpp:184
#26 0x00000035df6427ed in g_main_dispatch (context=0xe274e0) at gmain.c:2441
#27 g_main_context_dispatch (context=0xe274e0) at gmain.c:3014
#28 0x00000035df642fc8 in g_main_context_iterate (context=0xe274e0, block=<optimized out>, dispatch=1, self=<optimized out>) at gmain.c:3092
#29 0x00000035df64325c in g_main_context_iteration (context=0xe274e0, may_block=1) at gmain.c:3155
#30 0x00000035e4f84dcf in QEventDispatcherGlib::processEvents (this=0xe234d0, flags=<optimized out>) at kernel/qeventdispatcher_glib.cpp:422
#31 0x0000003c06a5c12e in QGuiEventDispatcherGlib::processEvents (this=<optimized out>, flags=<optimized out>) at kernel/qguieventdispatcher_glib.cpp:207
#32 0x00000035e4f59722 in QEventLoop::processEvents (this=<optimized out>, flags=...) at kernel/qeventloop.cpp:149
#33 0x00000035e4f5991f in QEventLoop::exec (this=0x7fff31bfbfa0, flags=...) at kernel/qeventloop.cpp:201
#34 0x00000035e4f5da67 in QCoreApplication::exec () at kernel/qcoreapplication.cpp:1064
#35 0x0000000000433154 in main (argc=<optimized out>, argv=<optimized out>) at /home/milan/Dev/RKWard/rkward/rkward/main.cpp:177

----------------------------------------------------------------------

Comment By: Milan Bouchet-Valat (nalimilan)
Date: 2011-11-17 09:47

Message:
OK, sorry, I wonder how I got so confused. I'll install the new revision,
and report if it ever happens again.

----------------------------------------------------------------------

Comment By: Thomas Friedrichsmeier (tfry)
Date: 2011-11-17 09:33

Message:
Hi!

In the latest revision of rkcommandeditorwindow.cpp (r4021, November 7),
line 1027 is not inside RKCodeCompletionModel::data (). So the backtrace
can't be from that version. The revision where I think I have fixed the
crash was r4017 from November 4. Here, line 1027 is inside the function,
but that is very unlikely to trigger a crash (unless memory is thrashed).
The previous unfix revision was r4008, October 28. Here, line 1027 is
inside the function, and it is exactly the line where a crash is to be
expected, if my analysis of the cause of the problem was correct.

So, you were definitely running a version from before November 7. And my
guess is that you were running version without the commit in question
(before November 4), indeed.

Regards
Thomas

----------------------------------------------------------------------

Comment By: Milan Bouchet-Valat (nalimilan)
Date: 2011-11-17 02:37

Message:
I've triggered this crash again today while replacing text in three
scripts. I'm not sure of the SVN revision since I updated my copy without
installing it, but the installed version dates back to the end of last week
at least. Is the fix supposed to be present?

Dumping the new trace here:
Application: RKWard (rkward.frontend), signal: Segmentation fault
82	T_PSEUDO (SYSCALL_SYMBOL, SYSCALL_NAME, SYSCALL_NARGS)
[Current thread is 1 (Thread 0x7f8ce38f5840 (LWP 21495))]

Thread 4 (Thread 0x7f8cdc88a700 (LWP 21500)):
#0  0x00000035dd6d7423 in __GI___poll (fds=<optimized out>, nfds=<optimized
out>, timeout=<optimized out>) at ../sysdeps/unix/sysv/linux/poll.c:87
#1  0x00000035df642d24 in g_main_context_poll (n_fds=1, fds=0x7f8cd8001110,
priority=<optimized out>, timeout=-1, context=0xaa0b30) at gmain.c:3405
#2  g_main_context_iterate (context=0xaa0b30, block=<optimized out>,
dispatch=1, self=<optimized out>) at gmain.c:3087
#3  0x00000035df64360d in g_main_loop_run (loop=0xaa0b10) at gmain.c:3300
#4  0x00007f8cdcbb0564 in gdbus_shared_thread_func (data=<optimized out>)
at gdbusprivate.c:276
#5  0x00000035df6683a6 in g_thread_create_proxy (data=0xaa0c20) at
gthread.c:1955
#6  0x00000035dde07b31 in start_thread (arg=0x7f8cdc88a700) at
pthread_create.c:305
#7  0x00000035dd6dfd2d in clone () at
../sysdeps/unix/sysv/linux/x86_64/clone.S:115

Thread 3 (Thread 0x7f8ccf3ec700 (LWP 21501)):
#0  0xffffffffff60010a in ?? ()
#1  0x00007f8ccf3eb7d0 in ?? ()
#2  0x00007ffff41e06e2 in ?? ()
Backtrace stopped: previous frame identical to this frame (corrupt stack?)

Thread 2 (Thread 0x7f8ccd907700 (LWP 21505)):
#0  pthread_cond_wait@@GLIBC_2.3.2 () at
../nptl/sysdeps/unix/sysv/linux/x86_64/pthread_cond_wait.S:165
#1  0x0000003958d84d62 in QTWTF::TCMalloc_PageHeap::scavengerThread
(this=0x395907dc20) at
../3rdparty/javascriptcore/JavaScriptCore/wtf/FastMalloc.cpp:2359
#2  0x0000003958d84d99 in QTWTF::TCMalloc_PageHeap::runScavengerThread
(context=<optimized out>) at
../3rdparty/javascriptcore/JavaScriptCore/wtf/FastMalloc.cpp:1464
#3  0x00000035dde07b31 in start_thread (arg=0x7f8ccd907700) at
pthread_create.c:305
#4  0x00000035dd6dfd2d in clone () at
../sysdeps/unix/sysv/linux/x86_64/clone.S:115

Thread 1 (Thread 0x7f8ce38f5840 (LWP 21495)):
[KCrash Handler]
#6  0x000000000044509d in RKCodeCompletionModel::data (this=<optimized
out>, index=<optimized out>, role=0) at
/home/milan/Dev/rkward_trunk/rkward/rkward/windows/rkcommandeditorwindow.cpp:1027
#7  0x00007f8ccde0aad7 in KateCompletionModel::Item::Item(bool,
KateCompletionModel*, HierarchicalModelHandler const&,
QPair<KTextEditor::CodeCompletionModel*, QModelIndex>) () from
/usr/lib64/libkatepartinterfaces.so.4
#8  0x00007f8ccde0efd2 in
KateCompletionModel::createItem(HierarchicalModelHandler const&,
QModelIndex const&, bool) () from /usr/lib64/libkatepartinterfaces.so.4
#9  0x00007f8ccde0fe33 in
KateCompletionModel::createItems(HierarchicalModelHandler const&,
QModelIndex const&, bool) () from /usr/lib64/libkatepartinterfaces.so.4
#10 0x00007f8ccde13018 in KateCompletionModel::createGroups() () from
/usr/lib64/libkatepartinterfaces.so.4
#11 0x00007f8ccde134f4 in
KateCompletionModel::setCompletionModels(QList<KTextEditor::CodeCompletionModel*>
const&) () from /usr/lib64/libkatepartinterfaces.so.4
#12 0x00007f8ccde029b1 in
KateCompletionWidget::startCompletion(KTextEditor::Range const&,
QList<KTextEditor::CodeCompletionModel*> const&,
KTextEditor::CodeCompletionModel::InvocationType) () from
/usr/lib64/libkatepartinterfaces.so.4
#13 0x00007f8ccde04198 in
KateCompletionWidget::startCompletion(KTextEditor::Range const&,
KTextEditor::CodeCompletionModel*,
KTextEditor::CodeCompletionModel::InvocationType) () from
/usr/lib64/libkatepartinterfaces.so.4
#14 0x000000000044bc57 in RKCommandEditorWindow::tryCompletion
(this=0x241f540) at
/home/milan/Dev/rkward_trunk/rkward/rkward/windows/rkcommandeditorwindow.cpp:591
#15 0x000000000044d7da in RKCommandEditorWindow::qt_metacall
(this=0x241f540, _c=QMetaObject::InvokeMetaMethod, _id=<optimized out>,
_a=0x7ffff40fb820) at
/home/milan/Dev/rkward_trunk/rkward/build/rkward/windows/rkcommandeditorwindow.moc:234
#16 0x000000395096d59a in QMetaObject::activate(QObject*, QMetaObject
const*, int, void**) () from /usr/lib64/libQtCore.so.4
#17 0x0000003950970e89 in QObject::event(QEvent*) () from
/usr/lib64/libQtCore.so.4
#18 0x00000039541bcc64 in QApplicationPrivate::notify_helper(QObject*,
QEvent*) () from /usr/lib64/libQtGui.so.4
#19 0x00000039541c1af1 in QApplication::notify(QObject*, QEvent*) () from
/usr/lib64/libQtGui.so.4
#20 0x00000039532431c6 in KApplication::notify(QObject*, QEvent*) () from
/usr/lib64/libkdeui.so.5
#21 0x000000395095a8dc in QCoreApplication::notifyInternal(QObject*,
QEvent*) () from /usr/lib64/libQtCore.so.4
#22 0x0000003950987192 in ?? () from /usr/lib64/libQtCore.so.4
#23 0x0000003950984a18 in ?? () from /usr/lib64/libQtCore.so.4
#24 0x00000035df6427ed in g_main_dispatch (context=0xa520e0) at
gmain.c:2441
#25 g_main_context_dispatch (context=0xa520e0) at gmain.c:3014
#26 0x00000035df642fc8 in g_main_context_iterate (context=0xa520e0,
block=<optimized out>, dispatch=1, self=<optimized out>) at gmain.c:3092
#27 0x00000035df64325c in g_main_context_iteration (context=0xa520e0,
may_block=1) at gmain.c:3155
#28 0x000000395098537f in
QEventDispatcherGlib::processEvents(QFlags<QEventLoop::ProcessEventsFlag>)
() from /usr/lib64/libQtCore.so.4
#29 0x000000395425fc0e in ?? () from /usr/lib64/libQtGui.so.4
#30 0x0000003950959df2 in
QEventLoop::processEvents(QFlags<QEventLoop::ProcessEventsFlag>) () from
/usr/lib64/libQtCore.so.4
#31 0x0000003950959fef in
QEventLoop::exec(QFlags<QEventLoop::ProcessEventsFlag>) () from
/usr/lib64/libQtCore.so.4
#32 0x000000395095e107 in QCoreApplication::exec() () from
/usr/lib64/libQtCore.so.4
#33 0x0000000000433694 in main (argc=<optimized out>, argv=<optimized out>)
at /home/milan/Dev/rkward_trunk/rkward/rkward/main.cpp:178

----------------------------------------------------------------------

Comment By: Thomas Friedrichsmeier (tfry)
Date: 2011-11-04 12:44

Message:
> No, I hadn't removed any object, at least not for about one minute. I
was
> working on code that replaced a matrix, but the matrix already existed.

Well, it *could* still have been the issue, if the matrix has any names on
it, and those names were in the completion list, and some of the
changed...

Anyway, I have committed a fix for the theoretical scenario I described in
SVN. This may or may not have been the source of the problem, but the code
in RKCodeCompletionModel::data() is very straight-forward, now, and I can't
see any way it could crash (at least not as long as the katepart calls the
function with anything even remotely valid).

So I think, it's safe to mark this as fixed. However, if you ever run into
this again, or you have any more ideas on what exactly might have led to
the crash, then please let me know.

Regards
Thomas

----------------------------------------------------------------------

Comment By: Milan Bouchet-Valat (nalimilan)
Date: 2011-11-04 12:26

Message:
No, I hadn't removed any object, at least not for about one minute. I was
working on code that replaced a matrix, but the matrix already existed.

----------------------------------------------------------------------

Comment By: Thomas Friedrichsmeier (tfry)
Date: 2011-11-04 12:17

Message:
Hi!

One potential scenario that could possibly lead to a crash with this
backtrace is this: An object, which is part of a current completion list is
being deleted, while the completion list is shown. At least theoretically,
that could lead to a crash. But I'm not quite sure this can be achieved in
practice. Anyway, does that sound like it might have happened?

I'll look into adding some safety mechanism for this case.

Regards
Thomas

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=459007&aid=3433496&group_id=50231