[patch, fortran] PR32633 - bogus error with implicit SAVE

[Tobias Burnus]

Hi Daniel,

I think you missed the following in the new patch:

>     case FL_PROCEDURE:
> -      conf2 (save);

(And
>  	case PROC_DUMMY:
> -	  conf2 (save);
but I think this is never reached.)

Additionally, I miss a do-compile test case which fails currently and it
fixed by this patch - just to make sure we will never regress on this
again; the test case should go beyond the test case from PR32633.
I glanced at the patch and it seems to be otherwise ok, but I would like
to go over it again (or someone else should do it).
Tobias