Re: [gfortran] Fix PR17615: Wrong symbol used when genrating code

[Paul Brook <paul at codesourcery dot com>]

On Friday 24 September 2004 17:35, Tobias Schlüter wrote:
> In PR 17615 we choked when generating code for a function called through a
> generic interface. The problem was that in one place we didn't look at the
> actual function that was meant to be called, but at the generic interface
> instead. This of course had to break under some circumstances.
>
> Bubblestrapped and tested. I will also add a testcase.
>
> - Tobi
>
> 2004-09-24  Tobias Schlueter  <tobias.schlueter@physik.uni-muenchen.de>
>
>  PR fortran/17615
>  * trans-expr.c (gfc_trans_arrayfunc_assign): Look at resolved
>  function to determine return type.
>
> Index: trans-expr.c
> ===================================================================
> RCS file: /cvs/gcc/gcc/gcc/fortran/trans-expr.c,v
> retrieving revision 1.28
> diff -u -p -r1.28 trans-expr.c
> --- trans-expr.c        17 Sep 2004 21:54:57 -0000      1.28
> +++ trans-expr.c        24 Sep 2004 16:28:01 -0000
> @@ -2007,7 +2007,7 @@ gfc_trans_arrayfunc_assign (gfc_expr * e
>    /* The frontend doesn't seem to bother filling in expr->symtree for
> intrinsic functions.  */
>    gcc_assert (expr2->value.function.isym
> -         || (gfc_return_by_reference (expr2->symtree->n.sym)
> +             || (gfc_return_by_reference (expr2->value.function.esym)
>               && expr2->symtree->n.sym->result->attr.dimension));

Why do you change the first occurrence of expr2->symtree, but not the second?

Paul