[Bug ada/62230] New: Legal program rejected, RM 4.3.2(3, 5/3), ancestor_part can not be expression against error message by compiler itself

["demoonlit at panathenaia dot halfmoon.jp" <gcc-bugzilla at gcc dot gnu dot org>]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62230

            Bug ID: 62230
           Summary: Legal program rejected, RM 4.3.2(3, 5/3),
                    ancestor_part can not be expression against error
                    message by compiler itself
           Product: gcc
           Version: unknown
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: ada
          Assignee: unassigned at gcc dot gnu.org
          Reporter: demoonlit at panathenaia dot halfmoon.jp

Created attachment 33380
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=33380&action=edit
minimal bug triggering source code

GNAT rejects any expressions implying aggregate for ancestor_part.

For example: (Please see the attachment file, also)
   type T1 is limited new Ada.Finalization.Limited_Controlled with null record;
   type T2 is new T1 with null record;
   X : T2 := (T1 with null record); -- OK
   Z : T2 := (T1'(Ada.Finalization.Limited_Controlled with null record) with
null record); -- ERROR

GNAT outputs:
    main.adb:23:09: limited ancestor part must be aggregate or function call

It seems for me that aggregate expression is allowed into ancestor_part
according to the error message outputted by GNAT itself. But, it is not.

And, RM 4.3.2(3, 5/3) allows it explicitly.