https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119948
--- Comment #2 from kargls at comcast dot net ---
(In reply to Damian Rouson from comment #0)
>
> % gfortran -c source-allocate-pure-function-result-component.f90
> source-allocate-pure-function-result.f90:17:20:
>
> 17 | allocate(test%i, source = 0)
> | 1
> Error: Bad allocate-object at (1) for a PURE procedure
This one is interesting. The error checks for an impure variable,
but it is not set up to deal with result-name.
>From match.cc(lines 4684-4689):
bool impure = gfc_impure_variable (tail->expr->symtree->n.sym);
if (impure && gfc_pure (NULL))
{
gfc_error ("Bad allocate-object at %C for a PURE procedure");
goto cleanup;
}
The function gfc_impure_variable is contained in resolve.cc. The result-name
is not marked as pure (, yet?), so the function returns true.
(gdb) p tail->expr->symtree->n.sym->attr.result
$9 = 1
(gdb) p tail->expr->symtree->n.sym->attr.pure
$10 = 0
(gdb) p tail->expr->symtree->n.sym->ns->proc_name->name
$13 = 0x804e174b0 "construct_test"
(gdb) p tail->expr->symtree->n.sym->ns->proc_name->result->name
$15 = 0x804e15198 "test"
So, this patch seems to do the right things.
diff --git a/gcc/fortran/resolve.cc b/gcc/fortran/resolve.cc
index f03708efef7..18bc0254cb3 100644
--- a/gcc/fortran/resolve.cc
+++ b/gcc/fortran/resolve.cc
@@ -18460,6 +18460,17 @@ gfc_impure_variable (gfc_symbol *sym)
if (sym->attr.use_assoc || sym->attr.in_common)
return 1;
+ /* If sym is for a result-name, then its pure attribute may not be set.
+ Check the namespace to see if this a procedure, and if it is a
+ pure function. FIXME: what about elemental? */
+ if (sym->attr.result && !sym->attr.pure)
+ {
+ if (sym->ns->proc_name
+ && sym->ns->proc_name->attr.pure
+ && sym->ns->proc_name->attr.function)
+ return 0;
+ }
+
/* Check if the symbol's ns is inside the pure procedure. */
for (ns = gfc_current_ns; ns; ns = ns->parent)
{
