Hmm; good point. I found it odd that the new logging framework code doesn't crash given the same trace under the same conditions, though. Also, since the crash seems to be in the core, I'd be a little surprised if it was the threading / logging code in there that was doing it.
I *did* just notice that the script path in that trace doesn't match the path to the bro executable; if the problem appears on a real interface as well, will re-run and generate a correct trace 
--Gilbert