Hello world, the attached patch, committed as obvious, fixes a regression introduced by yesterday's patch. Apparently, clobber statements are even more finicky that I thought and do not work for saved variables either.
Regards Thomas 2018-09-23 Thomas Koenig <tkoe...@gcc.gnu.org> PR fortran/87395 * gfc_conv_procedure_call: Reformat comments slightly. Do not add clobber on INTENT(OUT) for saved variables. 2018-09-23 Thomas Koenig <tkoe...@gcc.gnu.org> PR fortran/87395 * gfortran.dg/intent_out_10.f90: New test.
Index: trans-expr.c =================================================================== --- trans-expr.c (Revision 264506) +++ trans-expr.c (Arbeitskopie) @@ -5281,7 +5281,10 @@ gfc_conv_procedure_call (gfc_se * se, gfc_symbol * add_clobber = fsym && fsym->attr.intent == INTENT_OUT && !fsym->attr.allocatable && !fsym->attr.pointer && !e->symtree->n.sym->attr.pointer - && !e->symtree->n.sym->attr.dummy /* See PR 41453. */ + /* See PR 41453. */ + && !e->symtree->n.sym->attr.dummy + /* FIXME - PR 87395 and PR 41453 */ + && e->symtree->n.sym->attr.save == SAVE_NONE && e->ts.type != BT_CHARACTER && e->ts.type != BT_DERIVED && e->ts.type != BT_CLASS && !sym->attr.elemental;
! { dg-do compile } ! PR 87395 - this used to ICE module mo integer, save :: x contains subroutine foo x = 42 call bar(x) contains subroutine bar(y) integer, intent(out) :: y end subroutine bar end subroutine foo end module mo