------- Comment #3 from pcarlini at suse dot de 2006-10-18 10:06 ------- Let's suppose the issue is confirmed, and we have an useful complete testcase, then, a wild idea (at the moment I'm missing many details about the internals of the safe iterators, sorry): maybe we could explore the possibility to have the _M_invalidate functionality part of _Safe_iterator_base, thus easily use the debug mode mutex. I'm saying that because I'm under the impression that _M_invalidate basically uses info available in the _Safe_iterator_base. I'm only unsure about the tests "this->base() == __victim->base()", I hope can be done on the base objects too...
-- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29496
