[Bug ada/78066] Misleading error message about redundant Dynamic_Predicate incorrectly mentions "Valid"

[ebotcazou at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78066

Eric Botcazou <ebotcazou at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2016-10-22
                 CC|                            |ebotcazou at gcc dot gnu.org
     Ever confirmed|0                           |1
      Known to fail|                            |6.2.1, 7.0

--- Comment #1 from Eric Botcazou <ebotcazou at gcc dot gnu.org> ---
> The exact repetition of the range constraint on type T
> in its Dynamic_Predicate leads to infinite recursion.
> This is as predicted by a warning. However, the warning mentions
> attribute Valid, whereas the source has a membership test.

It's because the compiler (cleverly?) synthesizes 'Valid out of the test.

> Also, the workaround shown in the commented line seems to be
> checking membership in the same set of values, but compilation
> and program do not exhibit the same behavior.
> 
> (Is infinite recursion acceptable behavior when constraint
> and Dynamic_Predicate are identical?)

Of course not.