------- Comment #3 from pault at gcc dot gnu dot org 2009-03-22 12:22 ------- (In reply to comment #2) > This PR should probably be marked as [4.3/4.4 Regression]. > Dear, oh dear! This is annoying. The bug goes away if the first component of the derived type is anything other than a pointer. Also, interchanging the components clears the fault too. The derived type attribute alloc_comp is not getting set for some reason, to judge by the code.
Paul -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=39519
