PR35058: -Werror= works only with some warning

Manuel López-Ibáñez lopezibanez@gmail.com
Fri Feb 22 17:41:00 GMT 2008


On 22/02/2008, Joseph S. Myers <joseph@codesourcery.com> wrote:
> On Fri, 22 Feb 2008, Manuel López-Ibáñez wrote:
>
>
> > What do you think of the following idea? It makes pedwarn take an
>  > option like warning and it  defines pedwarn0 following warning0. The
>  > option that you pass a pedwarn enables/disables the diagnostic, while
>  > the kind of diagnostic is still controlled by -pedantic-errors. This
>  > also makes -fdiagnostics-show-option work with pedwarns!!
>
>
> I think this is a good idea.  Another function with an extra "is this
>  controlled by -pedantic-errors or not?" argument might also be needed, to
>  replace such functions as pedwarn_c90 (where -std=c89 -pedantic-errors
>  should yield an error, but -std=c99 -pedantic-errors
>  -Wdeclaration-after-statement should only yield a warning).
>

I'm not going to do any of this before the patch here:
http://gcc.gnu.org/ml/gcc-patches/2008-02/msg00962.html
is approved and committed.

Otherwise, it will break the patch or I will have to review C++
pedwarns two times. Once is already too many.

Cheers,

Manuel.



More information about the Gcc-patches mailing list