FS#26444 - [e-svn] segfaults on strcmp in libc.so.6 when mounting usb drive
Attached to Project:
Arch Linux
Opened by Andrew D. (andrewd18) - Friday, 14 October 2011, 14:27 GMT
Last edited by Allan McRae (Allan) - Saturday, 28 April 2012, 12:13 GMT
Opened by Andrew D. (andrewd18) - Friday, 14 October 2011, 14:27 GMT
Last edited by Allan McRae (Allan) - Saturday, 28 April 2012, 12:13 GMT
|
Details
Description:
Whenever I attempt to mount a USB drive, enlightenment crashes. I have verified this on USB drives with labels that have spaces and labels that don't, to make sure I'm not seeing Enlightenment bug #653 (http://trac.enlightenment.org/e/ticket/653). Steps to reproduce: 1. Start enlightenment. 2. Connect USB flash drive to system. 3. E menu -> Files -> Home. 4. Select the heart for favorites menu. I see a USB drive icon named with the label of the drive, with a U icon in the corner. 5. Double-click the USB drive. Expected Results: * The drive should be mounted and I should see files. Actual Results: * Enlightenment segfaults. GDB Output: Program received signal SIGSEGV, Segmentation fault. 0x00007f9a49544946 in strcmp () from /lib/libc.so.6 #0 0x00007f9a49544946 in strcmp () from /lib/libc.so.6 #1 0x0000000000480cfc in e_fm2_client_data () #2 0x00000000004a83d8 in ?? () #3 0x00007f9a4c2ca3f2 in _ecore_event_call () from /usr/lib/libecore.so.1 #4 0x00007f9a4c2ce97e in ?? () from /usr/lib/libecore.so.1 #5 0x00007f9a4c2cef1f in ecore_main_loop_begin () from /usr/lib/libecore.so.1 #6 0x000000000042fea8 in main () Additional info: * e-svn 62822-1 * glibc 2.14-6.1 installed from allan's repo, as I previously thought I might have bug #24615. I'd be happy to roll back and re-test if necessary. * I do NOT have hal-git installed. |
This task depends upon
Comment by
Dave Reisner (falconindy) - Friday,
14 October 2011, 19:39 GMT
Comment by Andrew D. (andrewd18) -
Friday, 14 October 2011, 19:52 GMT
Comment by
Ronald van Haren (pressh) - Sunday,
06 November 2011, 13:34 GMT
Note that while the segfault happens in strcmp, the problem is
that somewhere in e_fm2_client_data, a NULL is being passed to
strcmp. This isn't an issue with glibc. You should run 'bt full'
from the segfault in gdb and send it upstream.
Thank you. :) I'll file this upstream once I rebuild it with debug
symbols. Feel free to close this, then.
Is it fixed already in the new packages I just uploaded? If not
can you link the upstream bug report here before I close this?