https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101951
Bug ID: 101951
Summary: Implement F2018:C937
Product: gcc
Version: 12.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: fortran
Assignee: unassigned at gcc dot gnu.org
Reporter: kargl at gcc dot gnu.org
Target Milestone: ---
A small program was posted by Vipul Parekh in the J3
mailing list asking about its conformance.
https://mailman.j3-fortran.org/pipermail/j3/2021-August/013230.html
The program exposes a bug where gfortran was not
handling F2018:C937 (R927) (same as F2008:C640 (R626)).
This patch fixes this problem and I've dejagnu-ified the
test program.
diff --git a/gcc/fortran/match.c b/gcc/fortran/match.c
index 16502da001d..b1a2ac5384a 100644
--- a/gcc/fortran/match.c
+++ b/gcc/fortran/match.c
@@ -4233,7 +4233,7 @@ gfc_match_allocate (void)
if (gfc_match (" :: ") == MATCH_YES)
{
- if (!gfc_notify_std (GFC_STD_F2003, "typespec in ALLOCATE at %L",
+ if (!gfc_notify_std (GFC_STD_F2003, "type-spec in ALLOCATE at %L",
&old_locus))
goto cleanup;
@@ -4266,6 +4266,16 @@ gfc_match_allocate (void)
ts.type = BT_UNKNOWN;
gfc_current_locus = old_locus;
}
+
+ /* F2018:C937 (R927) type-spec shall not specify a type that has a
+ coarray ultimate component. Similar text in F2008:C640 (R626). */
+ if (ts.type == BT_DERIVED
+ && ts.u.derived->attr.coarray_comp)
+ {
+ gfc_error ("Type-spec at %L has a coarray ultimate component",
+ &old_locus);
+ goto cleanup;
+ }
}
for (;;)
diff --git a/gcc/testsuite/gfortran.dg/f2018c937.f90
b/gcc/testsuite/gfortran.dg/f2018c937.f90
new file mode 100644
index 00000000000..2a789a0e2d2
--- /dev/null
+++ b/gcc/testsuite/gfortran.dg/f2018c937.f90
@@ -0,0 +1,10 @@
+! { dg-do compile }
+! { dg-options "-fcoarray=single" }
+! https://mailman.j3-fortran.org/pipermail/j3/2021-August/013230.html
+program foo
+ type :: a_t
+ integer, allocatable :: n[:]
+ end type
+ class(*), allocatable :: a
+ allocate(a_t :: a) ! { dg-error "coarray ultimate component" }
+end program foo
