Bug 169851

Thanks for the report. The line it crashes on, though:
  JSObject *err = static_cast<JSObject *>(cons->construct(exec,args));
... is not really something that ought to ever crash --- it's invoking 
a builtin constructor, that's there all the time (from context it's a TypeError 
due to trying to access a field inside undefined). I presume it's not reproducible?

P.S. Sound like you could perhaps comment on bug #169722.

You are right, I can't reproduce it for now. When konqueror crashed and I was asked to resume the session upon restart, it crashed every time I used the back button again. Tried it three times. It was the same with #169817. Smells like stack corruption to me.

I use lots of tabs for browsing, maybe it's somehow related. If I get another crash log, I'll post it here.

I'll try to see if I can reproduce #169722.

Is there a good reason why I was able to crash it every time once it resumed the "broken" session? I guess the konqueror session saves URL, window size and other stuff in there. Where does this get saved to? Maybe the session data gets corrupted and that's why I was able to trigger it over and over again once it happens.

(In reply to comment #3)
> Is there a good reason why I was able to crash it every time once it resumed
> the "broken" session? I guess the konqueror session saves URL, window size and
> other stuff in there. Where does this get saved to? Maybe the session data gets
> corrupted and that's why I was able to trigger it over and over again once it
> happens.
> 

Well the reason is konqueror saves everything it can, even the buffer (which is normally the largest part), for every historyitems (back &forward arrows for each konq view). I plan to make the crash session manager smarter smarter for 4.2 so that it can detect  when a session is crashy and try to somehow deal with it but it's not done yet.

You can see the saved session as files (one per konqueror process) inside ~/.kde/share/apps/konqueror/autosave.

Thanks for the session path. Konqueror didn't crash on me for two days, I'll close the ticket and will reopen it in case I should get another backtrace.

A coworker generated a HTML code documentation with "DoxyS" under Windows and mentioned "make sure you have Javascript turned on, it makes heavy use of that".
When I openeded the index page, konqueror crashed :-)

Here's the backtrace:
Application: KDE Daemon (kded4), signal SIGSEGV
[Thread debugging using libthread_db enabled]
[New Thread 0xb8082770 (LWP 3446)]
[KCrash handler]
#6  QString::operator= (this=<value optimized out>, 
    other=<value optimized out>) at ../../src/corelib/arch/qatomic_i386.h:122
#7  0x00f510a8 in KCookieJar::stripDomain (this=0x9de76a8, _fqdn=@0xbfeb2fb8, 
    _domain=@0xbfeb3008)
    at /usr/src/debug/kdelibs-4.1.0/kioslave/http/kcookiejar/kcookiejar.cpp:520
#8  0x00f5171c in KCookieJar::stripDomain (this=0x9de76a8, cookie=@0x9f55020)
    at /usr/src/debug/kdelibs-4.1.0/kioslave/http/kcookiejar/kcookiejar.cpp:527
#9  0x00f51941 in KCookieJar::addCookie (this=0x9de76a8, cookie=@0x9f55020)
    at /usr/src/debug/kdelibs-4.1.0/kioslave/http/kcookiejar/kcookiejar.cpp:935
#10 0x00f564b6 in KCookieServer::checkCookies (this=0x9e76170, 
    cookieList=0xbfeb3190)
    at /usr/src/debug/kdelibs-4.1.0/kioslave/http/kcookiejar/kcookieserver.cpp:181
#11 0x00f57083 in KCookieServer::addCookies (this=0x9e76170, url=@0x9f00e58, 
    cookieHeader=@0x9eef728, windowId=<value optimized out>, 
    useDOMFormat=false)
    at /usr/src/debug/kdelibs-4.1.0/kioslave/http/kcookiejar/kcookieserver.cpp:151
#12 0x00f573f5 in KCookieServer::addCookies (this=0x9e76170, arg1=@0x9f00e58, 
    arg2=@0x9eef728, arg3=33)
    at /usr/src/debug/kdelibs-4.1.0/kioslave/http/kcookiejar/kcookieserver.cpp:418
#13 0x00f5f01f in KCookieServerAdaptor::qt_metacall (this=0x9dd6710, 
    _c=QMetaObject::InvokeMetaMethod, _id=0, _a=0xbfeb3388)
    at /usr/src/debug/kdelibs-4.1.0/i386-redhat-linux-gnu/kioslave/http/kcookiejar/kcookieserveradaptor.cpp:39
#14 0x002cd74e in QDBusConnectionPrivate::deliverCall (
    this=<value optimized out>, object=<value optimized out>, 
    msg=<value optimized out>, metaTypes=<value optimized out>, 
    slotIdx=<value optimized out>) at qdbusintegrator.cpp:865
#15 0x002cea81 in QDBusConnectionPrivate::activateCall (
    this=<value optimized out>, object=<value optimized out>, 
    flags=<value optimized out>, msg=<value optimized out>)
    at qdbusintegrator.cpp:777
#16 0x002cee91 in QDBusConnectionPrivate::activateObject (
    this=<value optimized out>, node=<value optimized out>, 
    msg=<value optimized out>, pathStartPos=<value optimized out>)
    at qdbusintegrator.cpp:1318
#17 0x002cf3fa in QDBusActivateObjectEvent::placeMetaCall (this=)
    at qdbusintegrator.cpp:1412
#18 0x07afc0f9 in QObject::event (this=<value optimized out>, 
    e=<value optimized out>) at kernel/qobject.cpp:1140
#19 0x02f5630c in QApplicationPrivate::notify_helper (
    this=<value optimized out>, receiver=<value optimized out>, 
    e=<value optimized out>) at kernel/qapplication.cpp:3800
#20 0x02f5e16e in QApplication::notify (this=<value optimized out>, 
    receiver=<value optimized out>, e=<value optimized out>)
    at kernel/qapplication.cpp:3392
#21 0x038dbddd in KApplication::notify (this=<value optimized out>, 
    receiver=<value optimized out>, event=<value optimized out>)
    at /usr/src/debug/kdelibs-4.1.0/kdeui/kernel/kapplication.cpp:311
#22 0x07aec731 in QCoreApplication::notifyInternal (
    this=<value optimized out>, receiver=<value optimized out>, 
    event=<value optimized out>) at kernel/qcoreapplication.cpp:591
#23 0x07aed3a5 in QCoreApplicationPrivate::sendPostedEvents (
    receiver=<value optimized out>, event_type=<value optimized out>, 
    data=<value optimized out>)
    at ../../src/corelib/kernel/qcoreapplication.h:215
#24 0x07aed5bd in QCoreApplication::sendPostedEvents (receiver=)
    at kernel/qcoreapplication.cpp:1095
#25 0x07b1725f in postEventSourceDispatch (s=)
    at ../../src/corelib/kernel/qcoreapplication.h:220
#26 0x0061d0d8 in g_main_context_dispatch () from /lib/libglib-2.0.so.0
#27 0x00620783 in ?? () from /lib/libglib-2.0.so.0
#28 0x00620941 in g_main_context_iteration () from /lib/libglib-2.0.so.0
#29 0x07b16ea8 in QEventDispatcherGlib::processEvents (
    this=<value optimized out>, flags=<value optimized out>)
    at kernel/qeventdispatcher_glib.cpp:325
#30 0x02feed95 in QGuiEventDispatcherGlib::processEvents (this=)
    at kernel/qguieventdispatcher_glib.cpp:204
#31 0x07aeae1a in QEventLoop::processEvents (this=<value optimized out>, 
    flags=<value optimized out>) at kernel/qeventloop.cpp:149
#32 0x07aeafda in QEventLoop::exec (this=<value optimized out>, 
    flags=<value optimized out>) at kernel/qeventloop.cpp:200
#33 0x07aed685 in QCoreApplication::exec () at kernel/qcoreapplication.cpp:849
#34 0x02f56187 in QApplication::exec () at kernel/qapplication.cpp:3330
#35 0x00395366 in kdemain (argc=<value optimized out>, 
    argv=<value optimized out>)
    at /usr/src/debug/kdelibs-4.1.0/kded/kded.cpp:847
#36 0x08048752 in main (argc=)
    at /usr/src/debug/kdelibs-4.1.0/i386-redhat-linux-gnu/kded/kded4_dummy.cpp:3

There's not KJS in there, but I suspect the stack somehow gets corrupted...

Here's another while tracing some PHP/libxml2 issue which returned an empty page instead of the Horde web application:

Application: Konqueror (konqueror), signal SIGSEGV
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
[Thread debugging using libthread_db enabled]
[New Thread 0xb7ee2770 (LWP 9743)]
[New Thread 0xb673cb90 (LWP 17618)]
[KCrash handler]
#6  0x02f60f07 in QWidget::testAttribute (this=)
    at ../../src/gui/kernel/qwidget.h:990
#7  0x02f5d154 in QApplication::setActiveWindow (act=<value optimized out>)
    at ../../src/gui/kernel/qwidget.h:952
#8  0x02fc6acc in QApplication::x11ProcessEvent (this=<value optimized out>, 
    event=<value optimized out>) at kernel/qapplication_x11.cpp:3087
#9  0x02fef6cc in x11EventSourceDispatch (s=<value optimized out>, 
    callback=<value optimized out>, user_data=<value optimized out>)
    at kernel/qguieventdispatcher_glib.cpp:148
#10 0x0061d0d8 in g_main_context_dispatch () from /lib/libglib-2.0.so.0
#11 0x00620783 in ?? () from /lib/libglib-2.0.so.0
#12 0x00620941 in g_main_context_iteration () from /lib/libglib-2.0.so.0
#13 0x07b16ea8 in QEventDispatcherGlib::processEvents (
    this=<value optimized out>, flags=<value optimized out>)
    at kernel/qeventdispatcher_glib.cpp:325
#14 0x02feed95 in QGuiEventDispatcherGlib::processEvents (this=)
    at kernel/qguieventdispatcher_glib.cpp:204
#15 0x07aeae1a in QEventLoop::processEvents (this=<value optimized out>, 
    flags=<value optimized out>) at kernel/qeventloop.cpp:149
#16 0x07aeafda in QEventLoop::exec (this=<value optimized out>, 
    flags=<value optimized out>) at kernel/qeventloop.cpp:200
#17 0x07aed685 in QCoreApplication::exec () at kernel/qcoreapplication.cpp:849
#18 0x02f56187 in QApplication::exec () at kernel/qapplication.cpp:3330
#19 0x020e900f in kdemain () from /usr/lib/libkdeinit4_konqueror.so
#20 0x08048792 in _start ()

Hope they are any good.

Comment #6 is probably because the local page set a cookie --- you were opening a local file, right --- and it's the cookie jar (inside kded4) crashing, not konqueror, I should fix it either way..

Comment #7 is some internal UI stuff in Qt, and really doesn't tell me a thing, I am afraid.

Yes, I was opening a local file in comment #6.

The backtrace of comment #7 really looks strange, but that's the one I can reproduce easily. It involves a segfaulted http server process which then aborts the connections and this crashes konqueror later on. Multiple backtraces from the same issue almost look the same. My gut tells me konqueror still uses a widget that is already freed. Konqueror just shows a blank page and states "Page loaded." in the status bar though the connection must have been aborted.

See bug 170147 for discussion of the kcookiejar crash.

I didn't see a Konqueror crash since KDE 4.1.1 for a long time, so I guess we can close this one :-)