PR 36901: pedwarn() + -pedantic-errors + -w (inhibit_warnings) should not emit errors

[Joseph S. Myers]

On Tue, 22 Jul 2008, Manuel López-Ibáñez wrote:

> 2008/7/22 Paolo Carlini <paolo.carlini@oracle.com>:
> > ... hi Manuel and sorry if this is slightly off-topic, but I see that you are "back in action" and wondered if you still remember the issue analyzed in the audit-trail of libstdc++/35637... Shall we maybe open a separate PR?
> >
> 
> I am not back in action, sorry. I am just trying to get committed
> those patches that I have already written, so they are not wasted. I
> don't have enough free time to write new patches or investigate PRs.
> 
> Nonetheless, I opened PR 36901 and I explained my proposed fix. If
> someone has time to tackle it, I will be happy to answer questions and
> provide suggestions.
> 
> It is strange that this hasn't come up more frequently because it
> affects both C and C++. I guess either people don't actually use
> -pedantic-errors or the system headers are typically -pedantic clean.

Or people add explicit in_system_header checks for cases where the system 
headers aren't clean; for example, that's what I did in 
<http://gcc.gnu.org/ml/gcc-patches/2004-05/msg02098.html>, while noting 
that the question of how pedwarns in system headers work should perhaps be 
revisited.

-- 
Joseph S. Myers
joseph@codesourcery.com