Re: [Patch,Fortran] PR 40873 - -fwhole-program decl fixes

[Mikael Morin <mikael dot morin at sfr dot fr>]

Le 27.07.2010 07:54, Tobias Burnus a écrit :
> Mikael Morin wrote:
> > Le 26.07.2010 22:32, Tobias Burnus a écrit :
> > > Build and and currently regtesting on x86-64-linux. If there is no
> > > failure:
> > > OK for the trunk?
> > The extra expr.c hunk (pr45081 fix) which slipped through, is ok (and
> > obvious).
> But regresses in terms of diagnostics - I will leave it out and have a
> separate look later.
>
> > For the interface thing see below.        /* F2003, 12.3.1.1 (3b);
> > F2008, 12.4.2.2 (3b) */
> > > -      if (def_sym->result->attr.pointer
> > > -          || def_sym->result->attr.allocatable)
> > > +      if ((def_sym->result->attr.pointer
> > > +           || def_sym->result->attr.allocatable)
> > > +&&  (sym->attr.if_source != IFSRC_IFBODY
> > The above makes sense, but the 4 following lines are odd.
> >
> > If I understand it correctly if there is a function
> > definition/interface mismatch (between pointer or allocatable
> > attributes), we generate an error asking for explicit interface (even
> > if the symbol comes from an interface block which _is_ an explicit
> > interface).
> > Well, maybe it's better than nothing after all.
>
> That was my idea - I think interface mismatches are rare enough that one
> does not need to replicate all the messages - and the error might be
> slightly odd, but should give a good hint.
OK, we can wait for someone to open a PR(if any) to decide to do 
something here.

> > > +           || def_sym->result->attr.pointer
> > > +            != sym->result->attr.pointer)
> > There is an extra parenthesis here
> I concur.
>
> > > +           || (def_sym->result->attr.allocatable
> > And another one here
>
> I concur. With those changes - which I spotted after I had send the
> email (but did not want to send a reply before testing finished),
> regtesting succeeded.
>
>
> > >             &&  def_sym->ts.u.cl->length != NULL)
> > >           {
> > >             gfc_charlen *cl = sym->ts.u.cl;
> > > @@ -1992,14 +1998,14 @@ resolve_global_procedure (gfc_symbol *sy
> > >       }
> > >
> > >         /* F2003, 12.3.1.1 (4); F2008, 12.4.2.2 (4) */
> > > -      if (def_sym->attr.elemental)
> > > +      if (def_sym->attr.elemental&&  !sym->attr.elemental)
> > Same as above, I would put
> > if (def_sym->attr.elemental&&  sym->attr.if_source != IFSRC_IFBODY)
> > like in the previous cases.
> > It is odd to guess interface explicitness out of symbol elementalness.
>
> Well, I want to print also an error if the INTERFACE has no ELEMENTAL -
> I could split the message in one for IFSRC_IFBODY and one for wrong
> interfaces, if you think it makes sense.
I was afraid that one could come up with a case where both def_sym and 
sym were declared as elemental (without interface body), thus skipping 
the check. I can't produce one, however, so it's probably fine as is.
> > > -      if (def_sym->attr.is_bind_c)
> > > +      if (def_sym->attr.is_bind_c&&  !sym->attr.is_bind_c)
> > Same here.
>
> Ditto.
Same here. No failing testcase.
> Thanks for the review!
>
> Tobias
Thanks for all your whole-file work.
Mikael