Andy Green wrote: > Yes I didn't commit that patch because I wondered if we find a new world > of previously very low probability problems are now more repeatable.
Oh, that's interesting. When I saw your mail about the patch, I thought it was in andy-tracking. But now I see that it wasn't. So it appears we just ran info some unrelated issues. Funny :) jc's sounded like something that's reproducible. Mine might be in the 100:1 (or worse) class. Should be possible to catch it with an automated test, though. - Werner
