GNOME Bugzilla – Bug 393029
syncing with PDA via gnome-pilot crashes evo-exchange during authentication
Last modified: 2007-02-26 14:55:37 UTC
Steps to reproduce: 1. Start Evolution with an Exchange account (evolution GUI must be up) 2. Set gnome-pilot ECalendar conduit to sync with Exchange calendar (type of sync should not matter, but Copy to PDA was used) 3. Sync PDA with Evolution. Crash will occur Stack trace: Memory status: size: 73932800 vsize: 0 resident: 73932800 share: 0 rss: 20750336 rss_rlim: 0 CPU usage: start_time: 1167970015 rtime: 0 utime: 611 stime: 0 cutime:517 cstime: 0 timeout: 94 it_real_value: 0 frequency: 0 Backtrace was generated from '/usr/libexec/evolution-exchange' Using host libthread_db library "/lib/i686/libthread_db.so.1". [Thread debugging using libthread_db enabled] [New Thread -1236232512 (LWP 2927)] [New Thread -1268532320 (LWP 3969)] [New Thread -1260139616 (LWP 3956)] [New Thread -1236571232 (LWP 2929)] 0xbfffe410 in ?? ()
+ Trace 99524
Other information:
Created attachment 79431 [details] [review] patch for this bug Simple fix for the crash. The code was going into the if statement only if ecomp was NULL; this is the opposite of what was intended as ecomp is then accessed inside the if statement without being set.
This bug was originally reported by Jonathan Steinhart (jsteinhart@gmail.com) on the gnome-pilot e-mail list. Here is what he originally wrote: On an up-to-date Ubuntu Edgy system, with evolution/evolution-exchange 2.8.1 and gnome-pilot-conduits 2.0.14. Sync performs fine with ecalendar disabled; but when set to "Synchronize", evolution-exchange (which I use to access my company email/calendar) crashes right after this output from gpilotd: ecalconduit-Message: --------------------------------------------------------- ecalconduit-Message: pre_sync: Calendar Conduit v.0.1.6 ecalconduit-Message: Using timezone: /softwarestudio.org/Olson_20011030_5/America/New_York ecalconduit-Message: Authenticating calendar Worked fine (i.e. would correctly sync my exchange calendar to my palm) up until a few months ago - unfortunately I don't remember exactly when it stopped, so I can't correlate the breakage with any other events. Any have an idea of why this is failing (or barring that, where to look for more data on the problem)? Thanks in advance!
Verified that the attached patch alleviates this bug on my system.
(In reply to comment #1) > Created an attachment (id=79431) [edit] > patch for this bug > > Simple fix for the crash. The code was going into the if statement only if > ecomp was NULL; this is the opposite of what was intended as ecomp is then > accessed inside the if statement without being set. > Patch looks good. Can you please commit it for both STABLE and SVN Head? Thanks.
Committed to SVN Head.