Bug#802786: libsigc++-2.0-dev: c++ only option in cflags

[Emilio Pozuelo Monfort]

Control: severity -1 serious

On 23/10/15 17:43, Michael Biebl wrote:
> Am 23.10.2015 um 17:01 schrieb Michael Biebl:
>> cc1: error: command line option '-std=c++11' is valid for C++/ObjC++ but
>>> not for C [-Werror]
>>>
>>> Removing -std=c++11 from your .pc files makes these packages build
>>> again.
> 
> Afaics, this is a warning by default. So the mere existence of
> -std=c++11 does not make said package ftbfs, only if you combine that
> with -Werror.
> It would also be good to have an actual list of packages which FTBFS
> because of that.
> I recompiled all rdeps of libsigc++2.0-dev/libglibmm2.4-dev before
> applying that change and did not notice any build failures because of that.
> 
> The alternative to dropping the .pc file hack is to have ~90 RC bugs
> [1]. As said, it would be good to have an actual list of packages
> failing to build because of that change. This would make it easier to
> evaluate the situation.
> If there are only one or two, dropping -Werror for those package seems
> like the better alternative then having to touch 90 packages.
> 
> CCing Emilio in his role as release manager. I talked to him about this
> issue before making this change.
> 
> Let's hear his input.

We should keep this bug open and drop -std=c++11 once we have GCC 6 (which
should default to c++11 IIRC) and will happen during this cycle. I'd rather have
this problem here affecting one rdep than have 80 RC bugs.

In the meantime, let's make this bug RC so we don't forget to fix it before the
Stretch release. Bumping the severity.

Cheers,
Emilio