This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [gfortran, patch] PR 23675: Character function of module-variable length
Erik Edelmann wrote:
> I agree that it's better to do this at resolution. However, if
> we do it in resolve_charlen we will do it for all character
> lengths, which is not necessary and therefore a waste of
> compilation time (AFAICS it's needed only for lengths of use
> associated function result variables). In the attached patch I
> do it in resolve_function() instead. It's not a big deal, of
> course -- character lengths usually have simple expressions that
> don't take long time to traverse, but to me it just feels better
> this way (but I can of course change it if you want me to).
I have no preference, if we find other cases where we have to call
gfc_set_symbols_referenced we can still rethink this.
> 2006-01-03 Erik Edelmann <eedelman@gcc.gnu.org>
>
> PR fortran/23675
> * expr.c (gfc_expr_set_symbols_referenced): New function.
> * gfortran.h: Add a function prototype for it.
> * resolve.c (resolve_function): Use it for
> use associated character functions lengths.
>
>
> 2006-01-03 Erik Edelmann <eedelman@gcc.gnu.org>
>
> PR fortran/23675
> gfortran.dg/char_result_11.f90: New.
Ok.
- Tobi