Version: 0.4 (using KDE 3.4.0, compiled sources) Compiler: gcc version 3.4.2 [FreeBSD] 20040728 OS: FreeBSD (i386) release 5.3-RELEASE kpdf crashes and dumps core with a signal 10 when loading any pdf document. This is consistant. Disabling "Enable background generation", or building with --disable-multithreaded-kpdf, fixes the problem.
Hmm, Disabling "Enable background generation" and --disable-multithreaded-kpdf do different things, i have to fix that. Does it work with --disable-multithreaded-kpdf and "Enable background generation" activated?
It works fine with --disable-multithreaded-kpdf and "Enable background generation" activated.
In an attempt to debug kpdf, I discover to my surprise that the problem goes away when compiling with the -g option. In any case, here's a backtrace from gdb (without the -g): Program received signal SIGBUS, Bus error. [Switching to Thread 3 (LWP 100175)] 0x297d23e6 in TextPage::coalesce () from /usr/local/lib/kde3/libkpdfpart.so (gdb) bt #0 0x297d23e6 in TextPage::coalesce () from /usr/local/lib/kde3/libkpdfpart.so #1 0x2981c9ce in KPDFOutputDev::endPage () from /usr/local/lib/kde3/libkpdfpart.so #2 0x2977fdd0 in Gfx::~Gfx () from /usr/local/lib/kde3/libkpdfpart.so #3 0x297ae0d7 in Page::displaySlice () from /usr/local/lib/kde3/libkpdfpart.so #4 0x297ae3c6 in Page::display () from /usr/local/lib/kde3/libkpdfpart.so #5 0x297abdb4 in PDFDoc::displayPage () from /usr/local/lib/kde3/libkpdfpart.so #6 0x2981a36f in PDFPixmapGeneratorThread::run () from /usr/local/lib/kde3/libkpdfpart.so #7 0x28bc9a1d in QThreadInstance::start () from /usr/X11R6/lib/libqt-mt.so.3 #8 0x292ebe66 in pthread_create () from /usr/lib/libpthread.so.1 #9 0x293afcdb in _ctx_start () from /lib/libc.so.5
Correction to earlier posts. It now seems that yesterday I had built kpdf with -g, during an initial attempt at debugging. Consequently, the --disable-multithreaded-kpdf flag is irrelevant. Sorry about the confusion.
Let me see if i understand it, with -g it works and without it ( --multithreaded-kpdf does not matter) you get the crash and the backtrace you posted?
That is correct. In addition, setting "Enable background generation" prevents the problem.
I'm marking that bug as duplicate of bug 93352 as it has the same backtrace given in it and the same problem (crashes will all pdf), can you please try if removing -march helps you too as it seems it did with the other users? *** This bug has been marked as a duplicate of 93352 ***