https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113377
--- Comment #7 from anlauf at gcc dot gnu.org --- (In reply to Mikael Morin from comment #6) > (In reply to anlauf from comment #4) > > > > Note that the following scalar example also fails: > > > "Fortunately", it is invalid. :-) > > From 15.5.2.12 (Argument presence and restrictions on arguments not present): > > An optional dummy argument that is not present is subject to the following > restrictions. > (...) > (8) If it is allocatable, it shall not be allocated, deallocated, or > supplied as an actual argument corresponding to an optional nonallocatable > dummy argument. > > In comment #4, j from one is non-present, allocatable, optional and passed > to j from two which is optional nonallocatable. Thanks for clarifying this. This helps to reduce the testcases I am looking at. For the next step, I've actually stepped back a little: it helps to fix the non-elemental cases first. I am currently working with: diff --git a/gcc/fortran/trans-expr.cc b/gcc/fortran/trans-expr.cc index 9dd1f4086f4..52fdbd5ca66 100644 --- a/gcc/fortran/trans-expr.cc +++ b/gcc/fortran/trans-expr.cc @@ -6526,6 +6648,10 @@ gfc_conv_procedure_call (gfc_se * se, gfc_symbol * sym, gfc_init_se (&argse, NULL); argse.want_pointer = 1; gfc_conv_expr (&argse, e); + if (e->symtree->n.sym->attr.dummy + && POINTER_TYPE_P (TREE_TYPE (argse.expr))) + argse.expr = gfc_build_addr_expr (NULL_TREE, + argse.expr); cond = fold_convert (TREE_TYPE (argse.expr), null_pointer_node); cond = fold_build2_loc (input_location, NE_EXPR, @@ -7256,6 +7382,7 @@ gfc_conv_procedure_call (gfc_se * se, gfc_symbol * sym, && e->symtree->n.sym->attr.optional && (((e->rank != 0 && elemental_proc) || e->representation.length || e->ts.type == BT_CHARACTER + || (e->rank == 0 && fsym && fsym->as == NULL) || (e->rank != 0 && (fsym == NULL || (fsym->as I'll attach a testcase exercising this for integer dummies, but I have a full set for other types at hand.