[PATH][Ada] Fix multilib handling of gnat.dg when RUNTESTFLAGS is used

[Laurent GUERBY]

On Wed, 2009-05-27 at 09:43 +0200, Arnaud Charlet wrote:
> > This regression is known (detected by Eric Botcazou) and a potential
> > solution was discussed (with patches) a few weeks ago:
> > 
> > http://gcc.gnu.org/ml/gcc-patches/2009-05/msg00313.html
> > http://gcc.gnu.org/ml/gcc-patches/2009-05/msg00334.html
> > 
> > I sent a PING for my latest patch yesterday:
> > 
> > http://gcc.gnu.org/ml/gcc-patches/2009-05/msg01639.html
> 
> Your latest patch is incorrect as I said.
> 
> I suggest a possible alternative, can you take it into account and verify
> it works for you?

I'll check but your change should work since the same path
will be taken during runtest than with my patch. My understanding is
that you were waiting for Vincent answer to your question:

"but maybe there was a good reason why Vincent did things differently?"

Before deciding.

> If it does, I'd suggest sending the updated patch and get Vincent's
> attention.

Ok will do shortly.

Laurent