http://bugs.winehq.org/show_bug.cgi?id=21254
--- Comment #12 from Nikolay Sivov bunglehead@gmail.com 2011-01-23 18:06:20 CST --- Looks really bad. Seems to me that typeinfo data is corrupted somehow with our logic. I don't see any harmful calls between construction and broken call, but at that point data is really corrupted - function count value for it is garbage while before that it's properly initialized to 0. I'll try to valgrind it first, maybe that will show more bugs at least, after that the only way is to go step by step in segment allocation logic probably.