Re: [Gm2] TestCase20

[Gaius Mulley]

<address@hidden> writes:

> % gm2 -I../../gm2.lib0 -c -fiso TestCase20a.mod
> TestCase20b.def:10:11: error: error found while compiling the program
> module
> TestCase20b.def:10:11:problem in parameter 1, PROCEDURE DoSeq (uslov:
> Uslov) ;
> TestCase20b.def:10:11:item being passed is 'TrueBolan': BOOLEAN
> TestCase20b.def:10:11:parameter mismatch: identifier with an
> incompatible type is being passed to this procedure
> TestCase20a.mod:12:3: error: error found while compiling the program
> module
> TestCase20a.mod:12:3:problem in parameter 1, PROCEDURE DoSeq (uslov:
> Uslov) ;
> TestCase20a.mod:12:3:item being passed is 'TrueBolan': BOOLEAN
> TestCase20a.mod:12:3:parameter mismatch: identifier with an
> incompatible type is being passed to this procedure
>
> % cat TestCase20*(mod|def)
> MODULE TestCase20a;
>
> IMPORT
>  TestCase20b;
>
> PROCEDURE TrueBolan (): BOOLEAN;
>  BEGIN
>    RETURN TRUE;
>  END TrueBolan;
>
> BEGIN
>  TestCase20b.DoSeq(TrueBolan);
> END TestCase20a.
> DEFINITION MODULE TestCase20b;
>
> IMPORT
>  TestCase20;
>
> CONST
(* should be a TYPE ? *)
>  Uslov = TestCase20.Uslov;

CONST
>  True = TestCase20.Uslov;
>
> PROCEDURE DoSeq(uslov: Uslov);
>
> END TestCase20b.
>
> DEFINITION MODULE TestCase20;
>
> TYPE
>  Uslov = PROCEDURE(): BOOLEAN;
>
> PROCEDURE True(): BOOLEAN;
>
> END TestCase20.

Hi,

I think the code is invalid - shouldn't the Uslov CONST declaration be a
TYPE?  If so the compiler should issue a sensible error message - whereas
currently it gives a misleading error,

regards,
Gaius