http://bugzilla.novell.com/show_bug.cgi?id=553819 User extigyro@gmail.com added comment http://bugzilla.novell.com/show_bug.cgi?id=553819#c2 --- Comment #2 from Miro Hadzhiev 2009-11-10 02:38:09 MST --- Tor, I think that this can help. The bugs are connected according to me and the reasons that make me think in that way is that the hang of OO happens when you use the native version and the crash - when the Novel-patched go-openoffice is in use. Especially about the "hang bug", I've just tested the latest OO 3.1.1 under win xp and the 'hang' occurred and again no crash. Under Arch Linux (where both version are present in the Arch repositories) everything is absolutely the same way. Native OO - hang, go-OO - crash. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=553819 User extigyro@gmail.com added comment http://bugzilla.novell.com/show_bug.cgi?id=553819#c4 --- Comment #4 from Miro Hadzhiev 2009-11-10 05:45:12 MST --- In the bug report on the OO bugzilla the OO Team and developers call it the "native" version. And because the bug (the crash) occurs only in go-oo, the OO developers rejected the bug report. And Tor, if you need some more info or log files, ask me right away. I understand that the bug is serious and needs serious attention to be fixed. Kind regards, Miro -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=553819 http://bugzilla.novell.com/show_bug.cgi?id=553819#c6 --- Comment #6 from Miro Hadzhiev 2009-11-28 21:44:13 UTC --- http://www.openoffice.org/issues/show_bug.cgi?id=107280 The "hang" is reported now. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=553819 http://bugzilla.novell.com/show_bug.cgi?id=553819#c8 --- Comment #8 from Tor Lillqvist 2009-12-09 21:58:38 UTC --- Running OOo under valgrind shows up that the code is using freed data: ==12396== Invalid read of size 4 ==12396== at 0x12577149: rtl::Reference::operator->() const (ref.hxx:168) ==12396== by 0x125754B0: webdav_ucp::DAVResourceAccess::LOCK(com::sun::star::ucb::Lock&, com::sun::star::uno::Referencecom::sun::star::ucb::XCommandEnvironment const&) (DAVResourceAccess.cxx:1017) ==12396== by 0x1254F9AE: webdav_ucp::Content::open(com::sun::star::ucb::OpenCommandArgument2 const&, com::sun::star::uno::Referencecom::sun::star::ucb::XCommandEnvironment const&) (webdavcontent.cxx:2481) ==12396== by 0x12548AED: webdav_ucp::Content::execute(com::sun::star::ucb::Command const&, long, com::sun::star::uno::Referencecom::sun::star::ucb::XCommandEnvironment const&) (webdavcontent.cxx:919) ==12396== by 0x50E9DB4: ucbhelper::Content_Impl::executeCommand(com::sun::star::ucb::Command const&) (content.cxx:1809) ==12396== by 0x50E7144: ucbhelper::Content::openWriteableStream() (content.cxx:1256) ==12396== by 0x44FFD71: comphelper::MediaDescriptor::impl_openStreamWithURL(rtl::OUString const&, unsigned char) (mediadescriptor.cxx:837) ==12396== by 0x44FF00E: comphelper::MediaDescriptor::impl_addInputStream(unsigned char) (mediadescriptor.cxx:581) ==12396== by 0x44FEA7C: comphelper::MediaDescriptor::addInputStream() (mediadescriptor.cxx:519) ==12396== by 0xE60F95D: SwAsyncRetrieveInputStreamThread::threadFunction() (in /mnt/sdb/ooo-build-master/build/ooo320-m7/sw/unxlngi6.pro/lib/libswli.so) ==12396== by 0xE60ED4B: ObservableThread::run() (in /mnt/sdb/ooo-build-master/build/ooo320-m7/sw/unxlngi6.pro/lib/libswli.so) ==12396== by 0xE61113A: threadFunc (in /mnt/sdb/ooo-build-master/build/ooo320-m7/sw/unxlngi6.pro/lib/libswli.so) ==12396== by 0x403D117: osl_thread_start_Impl (in /home/tml/ooo32/ure/lib/libuno_sal.so.3) ==12396== by 0x43E06E4: start_thread (in /lib/libpthread-2.10.1.so) ==12396== by 0x43E05FF: ??? (in /lib/libpthread-2.10.1.so) ==12396== Address 0xa96a96c is 12 bytes inside a block of size 36 free'd ==12396== at 0x40268A6: free (in /usr/lib/valgrind/vgpreload_memcheck-x86-linux.so) ==12396== by 0x405BD00: rtl_freeMemory (in /home/tml/ooo32/ure/lib/libuno_sal.so.3) ==12396== by 0x8048CF7: operator delete(void*) (in /home/tml/ooo32/program/soffice.bin) ==12396== by 0x12559667: std::auto_ptr::reset(webdav_ucp::DAVResourceAccess*) (auto_ptr.h:242) ==12396== by 0x1254FBD9: webdav_ucp::Content::open(com::sun::star::ucb::OpenCommandArgument2 const&, com::sun::star::uno::Referencecom::sun::star::ucb::XCommandEnvironment const&) (webdavcontent.cxx:2516) ==12396== by 0x12548AED: webdav_ucp::Content::execute(com::sun::star::ucb::Command const&, long, com::sun::star::uno::Referencecom::sun::star::ucb::XCommandEnvironment const&) (webdavcontent.cxx:919) ==12396== by 0x50E9DB4: ucbhelper::Content_Impl::executeCommand(com::sun::star::ucb::Command const&) (content.cxx:1809) ==12396== by 0x50E7144: ucbhelper::Content::openWriteableStream() (content.cxx:1256) ==12396== by 0x44FFD71: comphelper::MediaDescriptor::impl_openStreamWithURL(rtl::OUString const&, unsigned char) (mediadescriptor.cxx:837) ==12396== by 0x44FF00E: comphelper::MediaDescriptor::impl_addInputStream(unsigned char) (mediadescriptor.cxx:581) ==12396== by 0x44FEA7C: comphelper::MediaDescriptor::addInputStream() (mediadescriptor.cxx:519) ==12396== by 0xE60F95D: SwAsyncRetrieveInputStreamThread::threadFunction() (in /mnt/sdb/ooo-build-master/build/ooo320-m7/sw/unxlngi6.pro/lib/libswli.so) ==12396== by 0xE60ED4B: ObservableThread::run() (in /mnt/sdb/ooo-build-master/build/ooo320-m7/sw/unxlngi6.pro/lib/libswli.so) ==12396== by 0xE61113A: threadFunc (in /mnt/sdb/ooo-build-master/build/ooo320-m7/sw/unxlngi6.pro/lib/libswli.so) ==12396== by 0x403D117: osl_thread_start_Impl (in /home/tml/ooo32/ure/lib/libuno_sal.so.3) ==12396== by 0x43E06E4: start_thread (in /lib/libpthread-2.10.1.so) ==12396== by 0x43E05FF: ??? (in /lib/libpthread-2.10.1.so) (Some temporary debugging printouts in the code above so line numbers might be off from our normal build.) I.e, in webdavcontent.cxx in Content::open() the call that resets the member m_xResAccess, i.e. m_xResAccess.reset(new DAVResourceAccess( *xResAccess.get())) causes something to be freed that the code in DAVResourceAccess::LOCK() then tries to use. I have tried to wrap my head around this for almost a day now but with little success. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=553819 http://bugzilla.novell.com/show_bug.cgi?id=553819#c10 --- Comment #10 from Miro Hadzhiev 2009-12-12 15:04:10 UTC --- Thank you, Tor!! It is really nice that the bug is resolved. P.S. I just want to say that the bug is still even not assigned in Ubuntu (after making even additional comments by me...) which is very disappointing according to me. I really do not see any reason to report bugs there anymore after they are obviously (and for me now) fixing only "Ubuntu-based" bugs. (: -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=553819 http://bugzilla.novell.com/show_bug.cgi?id=553819#c12 Tor Lillqvist changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|FIXED | --- Comment #12 from Tor Lillqvist 2010-01-08 11:12:40 UTC --- No need to file a new bug, let's reopen this one then. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=553819 http://bugzilla.novell.com/show_bug.cgi?id=553819#c13 --- Comment #13 from Marcel Stimberg 2010-01-08 11:42:41 UTC --- Created an attachment (id=335566) --> (http://bugzilla.novell.com/attachment.cgi?id=335566) Backtrace of the SIGSEGV webdav_ucp::DAVResourceAccess::GET Ok, this is the full backtrace of the crash. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=553819 http://bugzilla.novell.com/show_bug.cgi?id=553819#c15 --- Comment #15 from Tor Lillqvist 2010-01-08 12:37:31 UTC --- Could you attach some dozen lines of source code context from the place where it crashes for you? And perhaps run it under a debugger and print out some relevant variable values? (But still, if I can't get the openSUSE build to crash, if it for some reason is Debug (Ubuntu) specific, I think I just need to resolve again as WORKSFORME...) -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

...

= {_M_start = 0x0, _M_finish = 0x0, _M_end_of_storage = {<_STL::allocatorrtl::OUString> = {<No data fields>}, _M_data = 0x0}}, <No data fields>} (gdb) print rResource $2 = (webdav_ucp::DAVResource &) @0xbfffc628: {uri = {pData = 0xb52f14},

http://bugzilla.novell.com/show_bug.cgi?id=553819 http://bugzilla.novell.com/show_bug.cgi?id=553819#c17 --- Comment #17 from Marcel Stimberg 2010-01-08 14:31:02 UTC --- It seems to be more difficult to trigger than before, but as I said it is still fairly reliably triggered for me (what works most of the time is scrolling while it is loading, so the images are not on the screen and immediately scrolling back. The scrollbar than "hangs" and then it crashes). While testing with gdb, I once had a crash in a different GET function of DAVResourceAccess.cxx, so it seems to crash in either one of these (in the respective line with m_xSession->GET): uno::Reference< io::XStream > DAVResourceAccess::GET( const std::vector< rtl::OUString > & rHeaderNames, DAVResource & rResource, const uno::Reference< ucb::XCommandEnvironment > & xEnv, sal_Bool bAllowEmpty ) throw( DAVException ) { initialize(); uno::Reference< io::XStream > xStream; int errorCount = 0; bool bRetry; do { bRetry = false; try { DAVRequestHeaders aHeaders; getUserRequestHeaders( xEnv, getRequestURI(), rtl::OUString::createFromAscii( "GET" ), aHeaders ); xStream = m_xSession->GET( getRequestURI(), rHeaderNames, rResource, DAVRequestEnvironment( getRequestURI(), new DAVAuthListener_Impl( xEnv, m_aURL ), aHeaders, xEnv ), bAllowEmpty ); } catch ( DAVException & e ) { errorCount++; bRetry = handleException( e, errorCount ); if ( !bRetry ) throw; } } while ( bRetry ); return xStream; } //========================================================================= void DAVResourceAccess::GET( uno::Reference< io::XOutputStream > & rStream, const std::vector< rtl::OUString > & rHeaderNames, DAVResource & rResource, const uno::Reference< ucb::XCommandEnvironment > & xEnv ) throw( DAVException ) { initialize(); bool bRetry; int errorCount = 0; do { bRetry = false; try { DAVRequestHeaders aHeaders; getUserRequestHeaders( xEnv, getRequestURI(), rtl::OUString::createFromAscii( "GET" ), aHeaders ); m_xSession->GET( getRequestURI(), rStream, rHeaderNames, rResource, DAVRequestEnvironment( getRequestURI(), new DAVAuthListener_Impl( xEnv, m_aURL ), aHeaders, xEnv ) ); } catch ( DAVException & e ) { errorCount++; bRetry = handleException( e, errorCount ); if ( !bRetry ) throw; } } while ( bRetry ); } This is what I tried in gdb after a crash in the first function (like in my stacktrace from comment #13): (gdb) print rHeaderNames $1 = (const _STL::vectorrtl::OUString > &) @0xbfffc650: {<_STL::_Vector_basertl::OUString properties = {<_STL::_Vector_base >> = {_M_start = 0x0, _M_finish = 0x0, _M_end_of_storage = {<_STL::allocator> = {<No data fields>}, _M_data = 0x0}}, <No data fields>}} (gdb) print xEnv $3 = (const com::sun::star::uno::Referencecom::sun::star::ucb::XCommandEnvironment &) @0x8a7cd38: {com::sun::star::uno::BaseReference = { _pInterface = 0x8818364}, <No data fields>} (gdb) print m_aURL $4 = {pData = 0xa} (gdb) print aHeaders $5 = {<_STL::_Vector_base<_STL::pair, _STL::allocator<_STL::pair > >> = {_M_start = 0x0, _M_finish = 0x0, _M_end_of_storage = {<_STL::allocator<_STL::pair >> = {<No data fields>}, _M_data = 0x0}}, <No data fields>} (gdb) print xEnv $6 = (const com::sun::star::uno::Referencecom::sun::star::ucb::XCommandEnvironment &) @0x8a7cd38: {com::sun::star::uno::BaseReference = { _pInterface = 0x8818364}, <No data fields>} (gdb) print bAllowEmpty $7 = 0 '\000' (gdb) print m_xSession $8 = {m_pBody = 0x6f0044} pData ($4) looks a bit strange for a pointer, doesn't it? I therefore tried this: (gdb) print m_aURL.pData $9 = (rtl_uString *) 0xa (gdb) printf "%s", m_aURL.pData Cannot access memory at address 0xa Hope this is of any help :-/ -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.