http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54208
janus at gcc dot gnu.org changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |janus at gcc dot gnu.org
--- Comment #2 from janus at gcc dot gnu.org 2012-08-09 19:12:44 UTC ---
The problem is apparently that, when trying to reduce the constant expression,
we only resolve the expression itself (i.e. "ubound(arr,1)") but not the array
spec of the symbol 'arr'.
I find it surprising that this sort of thing has not been reported/fixed before
(it may have been reported at least: I have not checked bugzilla for similar
PRs).
Anyway, the following patchlet gets rid of the error, but may possibly
introduce regressions (unchecked):
Index: gcc/fortran/resolve.c
===================================================================
--- gcc/fortran/resolve.c (revision 190186)
+++ gcc/fortran/resolve.c (working copy)
@@ -5393,6 +5393,8 @@ resolve_variable (gfc_expr *e)
&& gfc_current_ns->parent->parent == sym->ns)))
sym->attr.host_assoc = 1;
+ gfc_resolve_array_spec (sym->as, 0);
+
resolve_procedure:
if (t == SUCCESS && resolve_procedure_expression (e) == FAILURE)
t = FAILURE;
