[Gm2] Re: latest cvs fix

[Martin Kalbfuß]

Thanks again :-)

Am Donnerstag, den 01.10.2009, 13:58 +0100 schrieb Gaius Mulley:
> Hi,
> 
> Martin Kalbfuß <address@hidden> writes:
> 
> > I maybe found a bug. I'm not sure.
> > 
> > I have a type:
> > 
> > Vector = ARRAY[0..2] OF REAL
> > 
> > And a procedure:
> > 
> > PROCEDURE ScalarMultiplication(Vector : ARRAY OF REAL; Scalar : REAL;
> > VAR Result : ARRAY OF REAL);
> > 
> > I can do:
> > 
> > AVector := Vector{1.0, 2.0, 3.0};
> > ScalarMultiplication(AVector, 5.0, Result);
> > 
> > But not:
> > 
> > ScalarMultiplication(Vector{1.0, 2.0, 3.0}, 5.0, Result);
> > 
> > The compiler tells that the type cannot converted into a pointer type.
> 
> now fixed in the latest cvs - thanks for the bug report,
> 
> regards,
> Gaius
> 
> 
>   * gm2/gm2-compiler/M2GenGCC.mod:  fixed bug reported by Martin
>     Kalbfuß <address@hidden>,
>     (see testsuite/gm2/iso/run/pass/constructor2.mod) which did
>     not
>     allow a constructor to be passed as an actual parameter to
>     an
>     unbounded array formal parameter.  Procedure CodeUnbounded
>     extended to include a case when a constructor appears as
>     op3.
>   * gm2/gm2-compiler/M2Quads.mod:  ManipulateParameters to merge
>     the rw information and mark a parameter as read if passed by
>     value and write/read if passed by address.
>