http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49917

--- Comment #1 from Paolo Carlini <paolo.carlini at oracle dot com> 2011-07-30 
20:06:34 UTC ---
Looks like the test should a dg-do run (not implicitly a dg-do compile!) and
then changed to assign if i != 2, right? Then I believe it would work as a
regression test for c++/13865.

Reply via email to