What's wrong with breaking a lesser used platform?  It might motivate
someone to actually fix the problem :)  At the very least we'll hear about
it and then we can fix it.

donald

On Wed, Sep 23, 2015 at 7:35 PM, Martin Winter <
mwin...@opensourcerouting.org> wrote:

> Donald,
>
>
> On 23 Sep 2015, at 5:27, Donald Sharp wrote:
>
> NO -—Werror-enable
>>>>
>>> *********************
>>> Please be aware that the configure option —-Werror-enable fails at this
>>> time with a few errors
>>> on some platforms. For all of these tests, I had to keep it to
>>> —-Werror-disable (the old default)
>>> -—Werror-enable works on the following platforms:
>>>       Ubuntu 12.04
>>>       Ubuntu 14.04
>>>       CentOS 7
>>>       Debian 8
>>> —-Werror-enable fails on these platforms:
>>>       CentOS 6                (see
>>> https://ci1.netdef.org/browse/TESTING-WERROR-CI006BUILD-1 )
>>>               cc1: error: unrecognized command line option
>>> "-Wno-unused-result"
>>>       NetBSD 6                (see
>>> https://ci1.netdef.org/browse/TESTING-WERROR-CI007BUILD-1 )
>>>               pim_cmd.c: In function 'show_mroute_count':
>>>               pim_cmd.c:2353:10: error: format '%d' expects type 'int',
>>> but argument 3 has type 'long unsigned int'
>>>       FreeBSD 8               (see
>>> https://ci1.netdef.org/browse/TESTING-WERROR-CI009BUILD-1 )
>>>               same as NetBSD 6
>>>       FreeBSD 9               (see
>>> https://ci1.netdef.org/browse/TESTING-WERROR-CI004BUILD-1 )
>>>               same as NetBSD 6
>>>       FreeBSD 10              (see
>>> https://ci1.netdef.org/browse/TESTING-WERROR-CI003BUILD-1 )
>>>               same as NetBSD 6
>>>
>>>
>>> I'll submit patches for these build errors here in the next couple of
>> days,
>> but I would strongly argue against not including this, We should not be
>> accepting code that causes warnings.  Forcing users to ensure that their
>> code is correct is a good thing, without this we will always have a moving
>> target of broken code in some manner.
>>
>
> Great if we can fix them all. My main worry was that there are more
> platforms
> which I don’t test (yet) and still people depend on Quagga. Build (by
> default)
> will fail on these platforms.
> We (as OSR) have a whole set of patches to get Solaris build back working
> (with no
> warnings - still 2..3 more issues to go before it’s ready) and wanted to
> add NetBSD 7 and OpenBSD to my build system. I expect at least some of them
> failing.
>
> I fully support having this in, but mainly ask for a bit more time to
> concentrate some of the work in trying to solve as many of these issues.
> Maybe give me another 1..2 months (i.e. push it to next batch) before
> enforcing it ?
>
> Thanks for pointing out these build issues Martin!
>>
>
> Thanks for fixing them!
>
> - Martin Winter
>   mwin...@opensourcerouting.org
>
>
_______________________________________________
Quagga-dev mailing list
Quagga-dev@lists.quagga.net
https://lists.quagga.net/mailman/listinfo/quagga-dev

Reply via email to