Open GoogleCodeExporter opened 9 years ago
This was my patch.
Original comment by norman.juchler@dacuda.com
on 27 Sep 2011 at 3:04
Attachments:
I had same thought on using condition waiting, but then I realized it would be
easier fixing it the way I did in the end :)
There's a catch with conditional variable: pthread_cond_wait unlocks the mutex,
so child thread should first lock it, then do the job, then signal, then unlock
it so pthread_cond_wait could lock it back. Otherwise you can't guarantee that
signal will be sent *after* you start waiting, which could lead to another
deadlock. You have everything except the lock in child thread. Hope I'm right
on that point, it drives me crazy sometimes :-P
Another thing I was trying to solve with the patch is potentionally
inconsistent value of use_minidump_write_mutex_ since it's being set to true
before mutex is actually locked.
Original comment by mike.dld
on 27 Sep 2011 at 3:32
And umm, could you please clean your patch so it doesn't contain unchanged
parts (ignoring whitespace)? Would be nice ;)
Original comment by mike.dld
on 27 Sep 2011 at 3:39
Lol, could you please just fix it, since 2015 already
Original comment by dron...@gmail.com
on 30 Mar 2015 at 3:57
Original issue reported on code.google.com by
mike.dld
on 16 Mar 2011 at 2:06Attachments: