Patrick, Thanks for the tip! This isn't boost, but new gcc (migrated from 4.6.3 to 4.7.1). Since it's not syncevolution related, I'll try to investigate what is the problem. Sorry for taking your time.
Regards, Vlad. On Thu, 2012-09-13 at 17:28 +0200, Patrick Ohly wrote: > On Thu, 2012-09-13 at 16:36 +0200, Vladimir Elisseev wrote: > > Oke, I've compiled with debugging symbols, so see output of gdb below. > > Hopefully it helps. Let me know if you need any other commands (I > > mentioned already I'm not familiar with gdb, sorry.) > > I see roughly where it crashes, but I have no idea why. It looks like a > boost::signals2 internal boost::shared_ptr is empty at a time when > boost::signals2 expects it to be set. > > SyncEvolution is careful to work with real singletons her (instances > initialized in a function before returning a pointer to them), so it is > not a problem of using a global instance which has not been initialized > yet at a time when another global instance constructor tries to use it - > I hope Boost itself is not making that mistake. > > Did the version of Boost perhaps change since the last time that you > built SyncEvolution? Can you double-check by recompiling SyncEvolution > 1.2.99.4 with the current set of system libraries? > > Alternatively (or in addition), can you compile with "-O0 -g" and > produce another stack backtrace? It might be a bit more informative. > _______________________________________________ SyncEvolution mailing list SyncEvolution@syncevolution.org http://lists.syncevolution.org/listinfo/syncevolution
