[Evolution] Bug#395416: evolution-data-server: E-D-S 1.6.3 and 1.8.1 crash due to thread race condition

[Patrick Ohly]

On So, 2006-10-29 at 10:28 +0100, Øystein Gisnås wrote:
> It would be useful if you could pin down what exact actions are
> required to reproduce this. The less actions and more detail, the
> better.

Unfortunately the only way I know how to trigger this reliably is by
running the SyncEvolution "TestEvolution CalendarSource" test repeatedly
until E-D-S crashes. The stack backtraces that I quote in the Gnome
bugzilla entry came from such a crash. That doesn't mean that it cannot
occur during normal interactive use, it's just less likely.

If you want to reproduce it, the SyncEvolution 0.4 tar ball [1] contains
all sources that you need. Compile it normally, then follow the
instructions in "HACKING" to set up testing. You can skip the parts
about configuring two sources, one should be enough and you don't need a
SyncML server either.

-- 
Bye, Patrick Ohly
--  
Patrick.Ohly at gmx.de
http://www.estamos.de/