I have few questions regarding ABRT's duplicate stack detection algorithm. On ABRT's plans page, it is mentioned that - 

How to decide if 2 backtraces are the same or not?
My questions are -

1. What all informations are parsed from a backtrace ? is it only list of function name or some other informations as well ? 
2.  If a backtrace has multiple threads, what kind of ordering is performed before creating a hash? Two different crashes may have similar threads but in backtrace threads might be jumbled up ? 

Thanks in advance.