------- Comment #1 from paolo dot carlini at oracle dot com 2008-07-24 22:06 ------- I suppose you can believe me if I tell you that I care about libstdc++-v3.
Then, commit 138087 cannot have anything to do with these failures, really, believe me. Actually, I suspect 138092 is ""at fault"". Are you 101% sure about 138087? Can you double check, please? Then, libstdc++ is currently not broken at all. These two testcases, involving an optional, extension allocator, in an optional mode, are known to be very brittle, already failed on and off on various targets, many times in the past. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36924
