[Patch, Fortran] PR 36322/36463

[Paul Richard Thomas]

Janus,

On Fri, Oct 24, 2008 at 11:44 AM, Janus Weil <jaydub66@googlemail.com> wrote:
> Hi all,
>
> here is a patch which fixes most of the problems in PR36322 and
> PR36463. I think it's enough for PR3622 to be closed and for PR36463
> not to be called a regression any more (I will keep this one open and
> fix the remaining trouble in a follow-up patch). I had several test
> cases before, but compressed them all into one. The patch is
> regression-tested on i686-pc-linux-gnu. Ok for trunk?

I now do not have enough time to deal with this until early next week
- if nobody has reviewed it by then, I would be glad so to do.

I have a question, though:  Is this patch not substantially
duplicating the interface mapping that goes on in trans-expr.c?(This
only occurred to me a few minutes ago)

Paul