[gfortran] ping patch PR fortran/19362

[Tobias Schlüter]

François-Xavier Coudert wrote:
> We have a nice patch for PR fortran/19362 that has been waiting for a long 
> time: http://gcc.gnu.org/ml/gcc-patches/2005-01/msg00761.html
> 
> Could someone review it? I propose a testcase to go with it, since the 
> original patch did not include one.

While the patch will certainly make things work, I don't know how we're
dealing with types well enough to tell if it isn't working around another bug
someplace else and I therefore don't think I can approve it.

What I'm worried about is that we have to determine in the frontend that the
assignment is indeed legal, so I don't see why we should build two different
types at all.

- Tobi