Description of problem: OpenOffice.org 2.3 Impress crashes if you try to add an OGG video. Version-Release number of selected component (if applicable): openoffice.org-impress-2.3.0-6.5.el5 How reproducible: Always Steps to Reproduce: 1. Open OpenOffice.org2.3 Impress 2. Create a new blank presentation 3. Go to Insert -> Movie and Sound, choose any ogg video. Actual results: OpenOffice Impress crashes without a bug report. When run from the command line, this error shows up: libGL warning: 3D driver claims to not support visual 0x5b /usr/lib/openoffice.org/program/soffice: line 147: 25373 Segmentation fault "$sd_prog/$sd_binary" "$@" Expected results: The video would be inserted as an object in the presentation. Additional info: Reproduceable in Open Office 2.3 on a Mac, but does not crash in Open Office 2.3 in Fedora 8.
Hmm, so what video driver is X using ? This might be a X problem or it might be some visual misuse by OOo. I'd expect to get a crash-reporter on a "normal" crash, so its odd to get a crash without the crashreporter appearing. Can you run (as an ordinary user, not root) /usr/lib/openoffice.org/program/crash_report -stack /dev/null to collect details about your driver and your screen depth and paste in that information here
Machine is using the i810 driver (and intel driver while using Fedora) Crash log from a long long time ago (but same OS and installation): (I) x.org loaded video driver of... (II) Loading /usr/lib/xorg/modules/drivers/i810_drv.so (III) Desktop is: GNOME (IV) libgcj version is: libgcj-4.1.2-42.el5-i386 (V) kernel is: Linux 2.6.18-92.1.1.el5 #1 SMP Thu May 22 09:01:29 EDT 2008 i686 i686 i386 (VI) OpenOffice.org core rpm version is: openoffice.org-core-2.3.0-6.5.el5-i386 (VII) depth of root window: 24 planes (VIII) accessibility is: false ...start sestatus details ... SELinux status: enabled SELinuxfs mount: /selinux Current mode: enforcing Mode from config file: enforcing Policy version: 21 Policy from config file: targeted ...end sestatus details ... ...start stackreport details ... ...end stackreport details ... ...start sample ldd details ... linux-gate.so.1 => (0x00db0000) libuno_sal.so.3 => /usr/lib/openoffice.org/program/libuno_sal.so.3 (0x001cd000) libuno_salhelpergcc3.so.3 => /usr/lib/openoffice.org/program/libuno_salhelpergcc3.so.3 (0x00f64000) libstore.so.3 => /usr/lib/openoffice.org/program/libstore.so.3 (0x00ac9000) libdl.so.2 => /lib/libdl.so.2 (0x00ef7000) libpthread.so.0 => /lib/libpthread.so.0 (0x00835000) libstlport_gcc.so => /usr/lib/openoffice.org/program/libstlport_gcc.so (0x00b18000) libstdc++.so.6 => /usr/lib/libstdc++.so.6 (0x0039f000) libm.so.6 => /lib/libm.so.6 (0x00544000) libgcc_s.so.1 => /lib/libgcc_s.so.1 (0x00a23000) libc.so.6 => /lib/libc.so.6 (0x0056b000) libcrypt.so.1 => /lib/libcrypt.so.1 (0x00959000) /lib/ld-linux.so.2 (0x00ae7000) ...end sample ldd details ...
Update: I tried switching my driver on my Lenovo T60 from i810 to intel, repeated the steps, and it still crashes.
I strongly suspect that its a class of bug that needs running in X sync mode with SAL_SYNCHRONIZE=1 and gdb /usr/lib/openoffice.org/program/soffice.bin breakpoint on _exit/exit and bt from there
I'm going to try and setup a suitable RHEL-5 env here to reproduce this, but if someone else wants to have a go then install the (massive) matching debuginfo for OOo export SAL_SYNCHRONIZE=1 ooimpress & get OOo the state where you are about to insert the .ogg and in another terminal ps auxw|grep simpress.bin and note the pid gdb (gdb) attach <pid> (gdb) break exit (gdb) break _exit (gdb) cont and go ahead and insert the .ogg and hopefully gdb will catch the crash (gdb) bt to then get the bt
Created attachment 309786 [details] gdb trace of issue I followed your steps. Attached is the gdb results.
Confirmed, indeed, the immediate location of the crash is in gstreamer-plugins-base
Created attachment 309847 [details] backport from a more recent gstream that makes the problem go away Here's a backport that lets RHEL-5.2 OOo play the "Truth Happens" without blowing up
FWIW the old Fedora 6 bug 239192 looks the same as this
Do you have an upstream bug reference for that? We're supposed to upgrade gstreamer-plugins-good to the latest stable for 5.3 anyway.
nope, I don't know an upstream reference, but as its fixed in F-9 gstreamer it should come for free with a bump to the latest stable
Add to 5.4 radar, pretty sure this isn't making 5.3 at this point.
Should be fixed with bug 450105, we'll see.
This request was evaluated by Red Hat Product Management for inclusion, but this component is not scheduled to be updated in the current Red Hat Enterprise Linux release. If you would like this request to be reviewed for the next minor release, ask your support representative to set the next rhel-x.y flag to "?".
This is supposed to be fixed already. Reopen if not.