Re: [Opensync-users] repost: opensync crashing when syncing with evolution

[Jason G. <ju...@lo...>]

Hi Armin,

I'm still experiencing the crash with the new version of Evolution that
was released on the Fedora yum sites recently.  Looking at the stack
trace, it looks to me like it is crashing when opensync is attempting to
free an invalid pointer from xml-vcard.so?  Here's the stack trace:

[sync details omitted]
All clients sent changes or error
All conflicts have been reported
*** glibc detected *** msynctool: munmap_chunk(): invalid pointer:
0x00327b1b ***
======= Backtrace: =========
/lib/libc.so.6(__libc_free+0x17b)[0x3b151f]
/usr/lib/libglib-2.0.so.0(g_free+0x31)[0xc1d5a1]
/usr/local/lib/opensync/formats/xml-vcard.so[0x320aed]
/usr/local/lib/opensync/formats/xml-vcard.so[0x3216fc]
/usr/local/lib/opensync/formats/xml-vcard.so[0x31fbed]
/usr/local/lib/libopensync.so.0(osync_converter_invoke+0xea)[0x33ed9a]
/usr/local/lib/libopensync.so.0(osync_conv_convert_fn+0x16a)[0x33f98a]
/usr/local/lib/libopensync.so.0(osync_change_convert_member_sink
+0x5d)[0x3401ad]
/usr/local/lib/libopensync.so.0(osync_member_commit_change
+0xcb)[0x33abfb]
/usr/local/lib/libosengine.so.0(client_message_handler+0x196)[0xf27186]
/usr/local/lib/libosengine.so.0(_queue_dispatch+0x95)[0xf266f5]
/usr/lib/libglib-2.0.so.0(g_main_context_dispatch+0x16d)[0xc1615d]
/usr/lib/libglib-2.0.so.0[0xc193ef]
/usr/lib/libglib-2.0.so.0(g_main_loop_run+0x1a9)[0xc19799]
/usr/lib/libglib-2.0.so.0[0xc33a1f]
/lib/libpthread.so.0[0xdef3b6]
/lib/libc.so.6(__clone+0x5e)[0x41233e]

On Fri, 2006-06-02 at 22:25 +0200, Armin Bauer wrote:

> Hi,
> 
> the problem I have with evolution is: it crashes here (debian) as well
> or it doesnt work. I tried to trace the bug, but it is definitely inside
> evolution.
> 
> I heard reports that very new versions of eds and evolution work
> correctly but i wasnt yet able to verify them.
> 
> Armin
> 
> Marcos "Japa" Umino wrote:
> > I can easily reproduce the bug, in Fedora 5
> > 
> > I believe another user from h2200 using Gentoo can reproduce it too.
> > 
> > Btw, I can reproduce bug using a evo2file sync too. I want to help.
> > 
> > On 5/21/06, Jason Grant <ju...@lo...> wrote:
> >>
> >>  Regarding the crash reported by a couple of folks around the 13th (see
> >> "Crash with invalid pointer when using evo plugin under fc5"), apart from
> >> the stack trace already provided, are there any more diagnostics that
> >> I can
> >> provide to help with fault diagnosis?  It's easy for me to reproduce the
> >> crash - opensync the only time opensync doesn't crash when syncing
> >> with evo
> >> is when there are no contacts!  So I'm in a good position to provide
> >> whatever info will be helpful.
> >>
> >>  Thanks,
> >>
> >>  Jason.