https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83522
--- Comment #13 from Paul Thomas <pault at gcc dot gnu.org> --- > Thus the patch is contrary to the standard and must be removed as soon > as possible. > > Paul By the way, the original problem is fixed on trunk by one of my recent deferred length character patches. You will note that some of the failing testcases (eg. deferred_character_2[2,3].f90 test this explicitly. Cheers Paul