https://gcc.gnu.org/bugzilla/show_bug.cgi?id=127755
Bug ID: 127755
Summary: Bogus compile errors for correct use of an EVENT_TYPE
pointer component
Product: gcc
Version: 16.2.0
URL: https://docs.google.com/document/d/1OoRq_b-Anv9YIEyFPv
G1HZ7tALRM2ceY4h7UdCcQBJU/edit?tab=t.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: fortran
Assignee: unassigned at gcc dot gnu.org
Reporter: dobonachea at lbl dot gov
Target Milestone: ---
Created attachment 65767
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=65767&action=edit
Test demonstrating correct use of an EVENT_TYPE pointer component
GFortran 16.2.0 on Linux/X64 is issuing bogus errors for the attached program,
which I believe fully conforms to the Fortran 2023 standard. The program works
as expected with Intel ifx 2026.1.1.
Here are the relevant F2023 16.10.2.10 constraints for EVENT_TYPE:
> C1603 A named entity with declared type EVENT_TYPE, or which has a noncoarray
> potential subobject
> component with declared type EVENT_TYPE, shall be a variable. A component
> that is of such a type
> shall be a data component.
>
> C1604 A named variable with declared type EVENT_TYPE shall be a coarray. A
> named variable with a
> noncoarray potential subobject component of type EVENT_TYPE shall be a
> coarray.
>
> C1605 An event variable shall not appear in a variable definition context
> except as the event-variable in an
> EVENT POST or EVENTWAIT statement, as an allocate-object, or as an actual
> argument in a reference
> to a procedure with an explicit interface if the corresponding dummy argument
> has INTENT (INOUT).
>
> C1606 A variable with a nonpointer subobject of type EVENT_TYPE shall not
> appear in a variable definition
> context except as an allocate-object in an ALLOCATE statement without a
> SOURCE= specifier, as an
> allocate-object in a DEALLOCATE statement, or as an actual argument in a
> reference to a procedure
> with an explicit interface if the corresponding dummy argument has INTENT
> (INOUT).
FWIW, here is a standards conformance analysis for this test program generated
using Google Gemini, which I believe is accurate:
https://docs.google.com/document/d/1OoRq_b-Anv9YIEyFPvG1HZ7tALRM2ceY4h7UdCcQBJU/edit?tab=t.0
GFortran Demonstration:
---------------------------------------------------------
cgpu$ gfortran --version
GNU Fortran (GCC) 16.2.0
Copyright (C) 2026 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
cgpu$ cat event_test5.F90
program event_test
use iso_fortran_env, only: event_type
implicit none
type :: event_ptr
type(event_type), pointer :: p
end type
type(event_type), target :: data_ready[*]
type(event_ptr) :: foo[*]
foo%p => data_ready
event post(foo[this_image()]%p)
event post(foo%p)
event wait(foo%p, UNTIL_COUNT=2)
print *, this_image(), "success"
end program
cgpu$ gfortran -fcoarray=lib event_test5.F90 -lcaf_shmem && ./a.out
event_test5.F90:6:34:
6 | type(event_type), pointer :: p
| 1
Error: Component p at (1) of type EVENT_TYPE must have a codimension or be a
subcomponent of a coarray, which is not possible as the component has the
pointer attribute
event_test5.F90:12:3:
12 | foo%p => data_ready
| 1~~~~
Error: LOCK_EVENT in variable definition context (pointer assignment) at (1)
event_test5.F90:15:13:
15 | event post(foo%p)
| 1
Error: Event variable argument at (1) must be a coarray or coindexed
event_test5.F90:16:13:
16 | event wait(foo%p, UNTIL_COUNT=2)
| 1
Error: Event variable argument at (1) must be a coarray but not coindexed
---------------------------------------------------------
Every one of these compiler errors from GFortran is wrong:
1. "Error: Component p at (1) of type EVENT_TYPE must have a codimension or be
a subcomponent of a coarray"
There is no such requirement in the standard.
This appears to be referring to C1604, but it does not apply to a pointer
component (which is not a potential subobject component).
2. " 12 | foo%p => data_ready"
"Error: LOCK_EVENT in variable definition context (pointer assignment) at "
There is no "LOCK" here, I assume "LOCK_EVENT" is a typo for "EVENT_TYPE".
However pointer assignment is a pointer association context (F2023 19.6.8) not
a variable definition context (F2023 19.6.7), therefore C1605/C1606 do not
apply.
3. "15 | event post(foo%p)"
"Error: Event variable argument at (1) must be a coarray or coindexed"
There is no such requirement in the standard. See 11.7.7 EVENT POST statement:
> R1174 event-post-stmt is EVENT POST ( event-variable [ , sync-stat-list ] )
> R1175 event-variable is scalar-variable
> C1183 (R1175) An event-variable shall be of type EVENT_TYPE from the
> intrinsic module ISO_FORTRAN_ENV (16.10.2).
The event-variable is only required to be a scalar variable of type EVENT_TYPE.
4. "event wait(foo%p, UNTIL_COUNT=2)"
"Error: Event variable argument at (1) must be a coarray but not coindexed"
There is no such requirement in the standard. See 11.7.8 EVENT WAIT statement:
> R1176 event-wait-stmt is EVENT WAIT ( event-variable [ , event-wait-spec-list
> ] )
> C1184 (R1176) The event-variable in an event-wait-stmt shall not be coindexed.
The event-variable is only required to be a noncoindexed scalar variable of
type EVENT_TYPE.
