This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH][RFA] Enabling -W[no-]error=... for cpplib warnings
On Mon, Nov 9, 2009 at 5:16 AM, Simon Baldwin <simonb@google.com> wrote:
> 2009/11/8 Gabriel Dos Reis <dosreis@gmail.com>
>>
>> On Tue, Nov 3, 2009 at 6:25 AM, Simon Baldwin <simonb@google.com> wrote:
>> > Thank you for the review -- inlined responses below.
>> >
>> > I've attached a revised and corrected version of the patch to the
>> > mail; bootstrap C/C++/Fortran and appropriate suite re-done. ?Please
>> > take a second look. ?Thanks.
>>
>> My earlier comments were about the name of the command line. ?It is
>> misleading. ?We are not warning #warning. ?Please, change the
>> name to something like -Wcpp or better. ?Wwarning-directive
>> sounds as if we were warning about #warning themselves.
>
> Neither -Wcomment nor -Wsystem-headers behave in the way this reading
> of their names might suggest, yet both are accepted.
I believe building on 'exceptions' to make the list of exceptions long is
wrong headed and a wrong excuse.
>
> However, -Wcpp sounds worse to me.? The preprocessor can issue several
> warnings, and -Wcpp sounds like it controls them all.? In fact, only
> the output of #warning is what's affected here.? -Wwarning-directive
> has the advantage of being relatively short and easy to remember,
well, it is not shorter than 'cpp', and the issue is not 'easy to
remember' -- both
would be.
> and
> also contains "warning" and so is likely to be found in a man or text
> search.
-W already standard for 'warn'.
>
> Please let me know precisely what you would like this command line
> flag to be called, and I will change it to that.
I believe I did.
-- Gaby