On Fri, Mar 13, 2020 at 05:17:48PM +0100, Oleg Nesterov wrote:
> On 03/12, Luc Van Oostenryck wrote:
> >
> > @@ -2098,6 +2103,12 @@ static void handle_preprocessor_line(struct stream *stream, struct token **line,
> > if (false_nesting)
> > goto out;
> > }
> > +
> > + if (expanding) {
> > + if (Wdirective_within_macro && !is_cond)
> ^^^^^^^^^^^^^^^^^^^^^^^
> I am not sure Wdirective_within_macro is really useful after this change,
> I'd suggest to drop my patch but this is up to you.
Yes, it makes much less sense now. I've replaced that by supporting
-pedantic.
> > + warning(start->pos, "directive in macro's argument list");
> > + expanding = 0; // warn only once
> > + }
>
> then you should probably update preprocessor22.c ? See below.
>
> I am not sure about validation/preprocessor/expand-redef.c added by the
> previous patch,
Yes, I hadn't updated the testcases because the patch just a RFC
to see if just allowing the #if/... was OK (but I forgot to mark
is as such) and they also depended on the previous RFC patch about
the expansion of newline.
Series updated and pushed now.
Thanks for the review.
-- Luc
