SoC fortran procedure pointers

[Janus Weil]

Paul,
I applied your original patch to trunk revision 124953 (which you
built it from in the first place) and I'm getting the same errors
there. So I guess my patching was all ok. But I still don't see how
you managed to compile that test program without errors.

Then, I'd like to remove from the patch all the things not connected
to procedure declarations/pointers or abstract interfaces. You already
said that iresolve.c(gfc_resolve_reshape) contains such a part.
I found another curious part in resolve.c(resolve_ref), which I
couldn't make much sense of, and I wonder if this is connected to the
PROCEDURE business in any way, of if it's some other thing. Maybe you
could give me a hint here ...
Thanks,
Janus


2007/6/20, Janus Weil <jaydub66@googlemail.com>:
> 2007/6/20, Paul Thomas <paulthomas2@wanadoo.fr>:
> > Are you sure that the patch applied cleanly?
>
> Well, some parts were rejected at first, so I hand to put them in
> manually. I'm pretty sure I didn't forget anything, but I'll check
> again.
>
> Also I think I found a bug in the patch for parse.c:
>
>       case ST_PROGRAM:
>         p = "PROGRAM";
>         break;
> +     case ST_PROCEDURE:
> +       p = "PROGRAM";
> +       break;
>
> for ST_PROCEDURE we probably want "PROCEDURE" not "PROGRAM". I've
> already corrected this.
>
> Janus
>