http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48699
--- Comment #16 from Tobias Burnus <burnus at gcc dot gnu.org> 2011-05-30 09:22:25 UTC --- (In reply to comment #13) > Moreover, I think that this test case is in fact invalid, cf. PR 48887. We are talking about comment 4, aren't we? (That test case fails with ifort 12 as "sm2" is not allocatable. It (and the one in PR 48887) compiles with crayftn, but I have not checked whether the resulting code makes sense.)
