xbmc_12.0~git20121119.22795bc-1_amd64.changes ACCEPTED into experimental

[Andres Mejia]

On Thu, Nov 22, 2012 at 1:11 AM, Reinhard Tartler <siretart at gmail.com> wrote:
> On Tue, Nov 20, 2012 at 1:12 AM, Andres Mejia <amejia004 at gmail.com> wrote:
>> On Mon, Nov 19, 2012 at 7:07 PM, Andres Mejia <amejia004 at gmail.com> wrote:
>>> Hi all,
>>>
>>> FYI, I uploaded a new version of XBMC. I'm notifying you all because
>>> of a major change. This version of XBMC will be built and run using
>>> XBMC's internal copy of FFMpeg (the 10.2 branch of FFMpeg that is).
>>> Due to the number of changes done between FFMpeg and Libav
>>> (particularly with libavfilter), it would have taken an enormous
>>> amount of work to try and get XBMC building and running using Libav.
>>> Also, I did forget to note this major change in the changelog, but I
>>> did add a NEWS entry for this. Sorry about that.
>>>
>>> Also as an aside, I think we should discuss how to get ffmpeg back
>>> into Debian again. As I said some time ago, either Libav, FFMpeg, or
>>> both have to get their libraries and header paths renamed.
>>
>> I also want to note this now, apparently chromium has been using it's
>> own internal copy of FFMpeg for some time now. Therefore, XBMC is not
>> the first package where using an internal ffmpeg is being done.
>>
>
> Are you sure that all involved developers, including chromium & xbmc
> upstream as well as their packagers, would agree on the same FFmpeg
> snapshot?

I don't believe anyone ever mentioned that these two projects should
use the same snapshot of ffmpeg.

> TBH, I'm very skeptical. And with FFmpeg releasing like crazy after
> the fork, I don't see the project suitable at all for a distro
> scenario such as Debian. For instance, you indicate that XBMC uses the
> 0.10 branch; the latest upstream release, however, is 1.0, the release
> before is 0.11. Chromium maintains it (defacto) own branch, which
> admittedly is based on FFmpeg master. Therefore, tracking upstream
> releases does not seem to accommodate either project.
>
> To conclude, I understand that xbmc is causing new challenges here,
> but TBH, I think the better solution would be to investigate what's
> missing in libav's libavfilter. For instance, I think you mean the
> audio filter still, which has landed in libav 9. I imagine that there
> is still some stuff missing, but if it is useful, I'm sure it can be
> submitted libav upstream.
>
> --
> regards,
>     Reinhard

I don't work on the components of xbmc that uses ffmpeg. Elupus is who
you want to speak to. You can either try the xbmc-dev channels or the
xbmc forums to reach out to him to see what (if anything) libav could
do so xbmc could work with libav.

-- 
~ Andres