Description of problem: running recoll I believe Version-Release number of selected component: unrtf-0.21.5-2.fc20 Additional info: reporter: libreport-2.2.2 backtrace_rating: 4 cmdline: unrtf --nopict --html /mnt/2TBDrive4/TB3/home/uminaTB3/Desktop/TB2/AllDocumentFiles/Waters.doc crash_function: hash_get_string executable: /usr/bin/unrtf kernel: 3.14.4-200.fc20.x86_64 runlevel: N 5 type: CCpp uid: 1000 Truncated backtrace: Thread no. 1 (2 frames) #0 hash_get_string at hash.c:212 #1 optimize_word at word.c:236 Potential duplicate: bug 957215
Created attachment 908659 [details] File: backtrace
Created attachment 908660 [details] File: cgroup
Created attachment 908661 [details] File: core_backtrace
Created attachment 908662 [details] File: dso_list
Created attachment 908663 [details] File: environ
Created attachment 908664 [details] File: exploitable
Created attachment 908665 [details] File: limits
Created attachment 908666 [details] File: maps
Created attachment 908667 [details] File: open_fds
Created attachment 908668 [details] File: proc_pid_status
Created attachment 908669 [details] File: var_log_messages
Another user experienced a similar problem: recoll uses unrtf when establishing its document index by searching the disks. I do not know which type of file causes the abort, but there may be some large files it is not capable of handling. reporter: libreport-2.2.2 backtrace_rating: 4 cmdline: unrtf --nopict --html /mnt/2TBDrive4/TB3/home/uminaTB3/Desktop/TB2/MessyDesktop1/repeater/vertex/ri3umv311.rtf crash_function: hash_get_string executable: /usr/bin/unrtf kernel: 3.14.7-200.fc20.x86_64 package: unrtf-0.21.5-2.fc20 reason: unrtf killed by SIGSEGV runlevel: N 5 type: CCpp uid: 1000
Created attachment 927503 [details] A RTF file causing the problem I'm experiencing this bug, too. Attached is a file that causes the segfault. I didn't report this bug to the upstream bug tracker as it seems rather deserted.
(gdb) run TXT.rtf Starting program: /usr/bin/unrtf TXT.rtf Program received signal SIGSEGV, Segmentation fault. hash_get_string (value=18446744072300266155) at hash.c:212 212 if (hi->value == value) (gdb) bt #0 hash_get_string (value=18446744072300266155) at hash.c:212 #1 0x000000000040bf2a in optimize_word (w=0x682780) at word.c:236 #2 0x000000000040c00b in optimize_word (w=0x6824e0) at word.c:267 #3 0x000000000040c00b in optimize_word (w=0x65e050) at word.c:267 #4 0x00000000004017c3 in main (argc=2, argv=0x7fffffffdc38) at main.c:266
The upstream bug tracker is not deserted any more :) I could check that the problem file was processed without problem by the current code from the unrtf repository (http://savannah.gnu.org/projects/unrtf). Expect unrtf 0.21.8 to be released shortly, with fixes for many similar issues.
You're right, I think 0.21.8 is coming soon. Rahul and I still need to push 0.21.7 to Fedora 19 and 20 (I don't have permissions in pkgdb). However, 0.21.7 is available in Fedora 21's updates-testing repository. You could update to Fedora 21 and test that out: https://admin.fedoraproject.org/updates/unrtf-0.21.7-1.fc21
Oh, sorry Jean-Francois, I just realized you're one of the upstream authors :) So of course you'd know when 0.21.8 is coming out :)
No, my bad, I should have mentioned it. We had a big push recently to fix all known crashes in unrtf, after a little prodding from oss-security...
Ken - Provided commit access. Hope that helps!
Thanks Rahul! Mark, would you mind testing against 0.21.9? I've submitted it for testing today, and it should be signed and pushed out to Fedora's updates-testing mirrors soon. You can grab the Fedora 20 build directly from http://koji.fedoraproject.org/koji/buildinfo?buildID=601815 immediately if you wish. If you'd rather wait for signed builds, you can wait a few days for the Fedora infra admins to sign the package and push it out to the Fedora mirrors: https://admin.fedoraproject.org/updates/unrtf-0.21.9-1.fc20 . When that's done, you can run "yum --enablerepo=updates-testing update unrtf", and yum will then update your system to unrtf-0.21.9-1.fc20 . Either way, we'd really welcome your feedback and confirmation that this does in fact fix your issue.
Ken, you meant me by "Mark"? Anyway, I tested unrtf-0.21.9-1.fc20 and the bug is gone. No more segfaults with any of my RTF files. Thanks!
unrtf-0.21.9-1.fc20 has been submitted as an update for Fedora 20. https://admin.fedoraproject.org/updates/FEDORA-2015-0129/unrtf-0.21.9-1.fc20
Wow, my bad! Please accept my apologies for getting your name wrong Mike. Thanks for letting us know that 0.21.9 fixes your bugs. I will be sure to push it to Fedora's stable updates repo as soon as it's spent the required waiting period (7 days) in updates-testing.
unrtf-0.21.9-1.fc20 has been pushed to the Fedora 20 stable repository. If problems still persist, please make note of it in this bug report.