On Friday 13 December 2013 13:39:43 Kingsley G. Morse Jr. wrote: > Hi Lisandro, > > Thank you for your quick and thoughtful reply. > > I like your idea of getting a better backtrace. > > I installed libqt4-dbg. > > Then I duplicated the bug by running kdenlive > by itself, and then in gdb. > > Both times, I did > > kdenlive -> Project -> Add clip -> > file_with_a_transparent_background.png Mm, same as I did. But I can't reproduce it. Do you happen to have a small png with transparency that does trigger the bug? I'm suspecting that this might be related to the arch, thought. > The backtraces are below. > > I'd be interested in your thoughts on > > 1.) What version of Qt4 works for you? 4:4.8.5+git192-g085f851+dfsg-2, currently in sid. > 2.) Does either of the new backtraces tell > you anything new? Yes, see my comments below. > 3.) Does kdenlive needs to be run a different > way to > > a.) use libqt4-dbg and > > b.) give us a better backtrace? Not that I know of. > Here's the backtrace from running kdenlive with > gdb. > > Program received signal SIGFPE, Arithmetic exception. > 0xb7e6e823 in ?? () > (gdb) bt > #0 0xb7e6e823 in ?? () > #1 0xb7e995dc in ?? () > #2 0xb7e9ad93 in ?? () > #3 0xb7e9c8e7 in ?? () > #4 0xb62745ca in QMetaCallEvent::placeMetaCall (this=0xb9277518, > object=0xb82fe168) at kernel/qobject.cpp:524 I'm no expert here, but you loaded the Qt debugging symbols and we are still missing the ones at the top of the stack. So this suggest the problem could be somewhere else. Moreover this is coming from a metacall, so it must be somthing that gets triggered after a signal/slot connection, but not due to it. [snip] > Here's the backtrace from running kdenlive without > gdb. > > Application: Kdenlive (kdenlive), signal: Floating point exception > Using host libthread_db library Now this also tell us something. Would you mind telling me which is your CPU brand and model? cat /proc/cpuinfo could help. > "/lib/i386-linux-gnu/i686/cmov/libthread_db.so.1". [Current thread is 1 > (Thread 0xb2361980 (LWP 24444))] [snip] > #60 0xb3ebefff in ?? () from > /usr/lib/i386-linux-gnu/libnvidia-glcore.so.304.88 Backtrace stopped: > previous frame inner to this frame (corrupt stack?) Eeeek, are you using the nvidia propietary driver? If so, try using nouveau to reproduce the bug. > Thread 1 (Thread 0xb2361980 (LWP 24444)): > [KCrash Handler] > #6 0xb76b9823 in ?? () > #7 0xb76e45dc in ?? () > #8 0xb76e5d93 in ?? () > #9 0xb76e78e7 in ?? () > #10 0xb5abf5ca in QMetaCallEvent::placeMetaCall (this=0xba62bd88, > object=0xb96b9b88) at kernel/qobject.cpp:524 #11 0xb5ac749b in Same as with GDB, but no info too. Soes Dr Konqui starts up when you trigger the bug? If so, you can generate the backtrace there *and* there is an option that tells you which files are involved. Please send me that list so I can tell you which other dbg packages we need. Kinds regards, Lisandro. -- Some consider UNIX to be the second most important invention to come out of AT&T Bell Labs after the transistor. Dennis Ritchie, creator of the C programming language and co-creator of UNIX http://www.linfo.org/q_unix.html Lisandro Damián Nicanor Pérez Meyer http://perezmeyer.com.ar/ http://perezmeyer.blogspot.com/
Attachment:
signature.asc
Description: This is a digitally signed message part.