[vlc-devel] [PATCH] configure: Support systems with popcount(3) and popcountll(3)

[Thomas Guillem]

As said by Rémi in an other review, vlc_common.h can not depend on config.h (where HAVE_POPCOUNT is defined) since it's a public header .

The other issue is that we can't really rename these functions for 3.0. The API is frozen, and we don't want to break the build of third party modules that use popcount*.

What I can accept is just to put a #ifdef NETBSD arround popcount* functions.

On Sun, Mar 11, 2018, at 21:26, Kamil Rytarowski wrote:
> Ping!
> 
> I still need popcount(3) and popcountll(3) disabled for NetBSD or
> renamed to vlc_popcount*().
> 
> On 05.03.2018 11:38, Kamil Rytarowski wrote:
> > "include: revector bit ops"
> > https://github.com/videolan/vlc/commit/924d56751b3fbf31d55dec678df0541a39bdb4ad
> > 
> > This one is unrelated to vlc-3.0. I need the submitted patch to get
> > merged in the 3.0 branch. Alternatively some other way to correct NetBSD
> > build.
> > 
> > On 05.03.2018 09:25, Thomas Guillem wrote:
> >> Is this patch still needed after the "include: revector bit ops" commit ?
> >>  
> >> On Mon, Mar 5, 2018, at 09:08, Kamil Rytarowski wrote:
> >>> On 25.02.2018 16:32, Thomas Guillem wrote:
> >>>>
> >>>> On Sun, Feb 25, 2018, at 16:11, Kamil Rytarowski wrote:
> >>>>> On 25.02.2018 15:56, Thomas Guillem wrote:
> >>>>>>
> >>>>>> On Sun, Feb 25, 2018, at 15:36, Rémi Denis-Courmont wrote:
> >>>>>>> Le sunnuntaina 25. helmikuuta 2018, 13.12.55 EET Thomas Guillem a écrit :
> >>>>>>>> Rémi, could explain what this patch is breaking ?
> >>>>>>>
> >>>>>>> Did you find a platform that it does not break? Which one?
> >>>>>>
> >>>>>> Linux, debian stable.
> >>>>>>
> >>>>>> I don't see the problem with this patch.
> >>>>>>
> >>>>>> If you don't want to explain the problem, I can push it and wait for all buildbots.
> >>>>>>
> >>>>>
> >>>>> Thanks!
> >>>>>
> >>>>> I don't run !NetBSD systems at all, but I pledge to submit a patch
> >>>>> proposal for a potential fallout (assuming that a target OS is
> >>>>> sufficiently Open-Source to sense its internals).
> >>>>
> >>>> First I'll do more tests, I'll be able to test all archs tomorrow.
> >>>>
> >>>>>
> >>>>> Email had 1 attachment:
> >>>>> + signature.asc
> >>>>>   1k (application/pgp-signature)
> >>>
> >>> Ping?
> >>>
> >>> Email had 1 attachment:
> >>> + signature.asc
> >>>   1k (application/pgp-signature)
> > 
> > 
> 
> 
> Email had 1 attachment:
> + signature.asc
>   1k (application/pgp-signature)