[Thunar-dev] Experimental Thunar Trash tarball
Brian J. Tarricone
bjt23 at cornell.edu
Thu Jul 27 00:15:17 CEST 2006
-----BEGIN PGP SIGNED MESSAGE-----
On 7/26/2006 3:10 PM, Benedikt Meurer wrote:
> Björn Martensen wrote:
>>>> However, the other two crashes I reported that occur in the treeview
>>>> sidepane are still present.
>>> If the backtrace shows g_slice_alloc() or g_slice_alloc0(), try set the
>>> environment variable G_SLICE to always-malloc and enable malloc()
>>> debugging (man malloc should provide the required info for your libc
>>> implementation), and check if you are still able to reproduce the crash.
>> That helped. No crashes anymore.
>> I set G_SLICE=always-malloc and MALLOC_CHECK_=1 and after exiting thunar
>> and doing the stuff that caused crashes before I now get
>> *** glibc detected *** thunar: free(): invalid pointer: 0x08346b28 ***
>> *** glibc detected *** thunar: free(): invalid pointer: 0x08348fd0 ***
>> *** glibc detected *** thunar: free(): invalid pointer: 0x0834a568 ***
> Ah, good, now you just need to make free() crash so we find out where
> the invalid pointers come from.
You can set MALLOC_CHECK_=2 instead which will cause free() to abort()
on invalid pointers, and run thunar through gdb:
$ gdb Thunar
[... do something to make it crash ...]
[... get your backtrace ...]
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v18.104.22.168 (MingW32)
-----END PGP SIGNATURE-----
More information about the Thunar-dev