------- Comment #2 from drow at gcc dot gnu dot org 2006-07-26 13:11 ------- Subject: Re: ext/pb_ds/regression/tree_data_map_rand.cc fails with a particular random seed.
On Wed, Jul 26, 2006 at 11:26:01AM -0000, pcarlini at suse dot de wrote: > Out of curiosity, can you try whether changing the optimization level has any > effect? I'm asking because we are also seeing random failures of > ext/pb_ds/regression/priority_queue_rand.cc on ia64-linux which go away at -O0 > and -O1... In that case, I'm suspecting some sort of miscompilation. Fails at -O1 in the same way. Fails differently at -O0. Failure at /space/fsf/commit/gcc/libstdc++-v3/testsuite/util/regression/rand/assoc/detail/insert_fn_imps.hpp: 71 container: aabaaab bccadab aacdbc bbbbdab ....more... native container: ...more... Uncaught exception: insert Failed at index 297 Test failed with seed 1153519516 -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28457