[FFmpeg-devel] for ffmpeg org

[Michael Bradshaw]

On Wed, Feb 8, 2012 at 1:19 PM, John Dexter <jdxsolutions at gmail.com> wrote:

> On 8 February 2012 20:05, Carl Eugen Hoyos <cehoyos at ag.or.at> wrote:
> > John Dexter <jdxsolutions <at> gmail.com> writes:
> >
> >> the lack of enthusiasm towards making
> >> ffmpeg Windows-friendly
> >
> > Please understand that there _only_ is a lack of patches!
> > (Of course not every hack will be committed but protesting
> > about us not committing hypothetical, non-existing patches
> > is not exactly helpful.)
>
> Well I was responding directly to the comment "Most of the FFmpeg devs
> develop on Linux, so there probably won't be a lot of motivation to
> port the code just to use Visual Studio."
>
> I suppose I read that as "we don't want the changes" but maybe I
> mis-read, and it simply means "don't expect us to help, but we'll
> accept the changes"? In which case, I apologise... the 'vibe' seems
> quite anti-VC++ but maybe that is my mistake.
>

Sorry, maybe I didn't word it the best way.  Personally, I would *love* to
see FFmpeg build with VC++, seeing as I do my primary development in Visual
Studio.  And I do think the community would like it if FFmpeg built with
VC++.  I certainly did't mean that to come off as anti-VC++.  However, what
I was really trying to say was that FFmpeg works, and compiling FFmpeg with
VC++ isn't necessarily a critical feature; it's a nice-to-have feature.
 The core development team who is most familiar with FFmpeg (of which I am
not) is usually more focused on the critical features, and while I'm sure
they would like all the nice-to-haves, there are other things that I think
have a higher priority on their list.  So if FFmpeg is ever to build with
VC++ or Visual Studio, it'll probably need to be a community effort and we
can't just expect the main dev team to do it for us, seeing as nothing
is inherently broken and there are other tasks of higher priority. But it's
not a small task.  I was indeed saying "the changes would be great
(assuming they didn't break or completely mutilate anything), but don't
expect someone else to do it for you; you need to be the motivated
individual who gets it done, and you can't expect someone else to be that
motivated individual."

Another reason why there may not be a ton of motivation is the fact that
VC++ can consume the libs built with MinGW/GCC.  This solves the problem
for a lot of the people who just want to use the libs and don't necessarily
care about the fact they can't step into them to debug.  Again, I'm not
saying this means that no effort should be made to get FFmpeg building with
VC++, nor that there is no desire for this; I'm saying you have to find the
people willing to expend that effort, and seeing as there are many people
involved in this project, each with his or her own priorities, goals, and
responsibilities, that'll be the difficult part.

We can't expect the main dev team to do this for us.  It'd be awesome if I
woke up tomorrow and it was done, but they're already chugging away at
progressing FFmpeg as whole.  The status of FFmpeg with VC++ isn't stopping
them from progressing with their work, and I don't think it's fair that we
expect them to drop everything they're doing to do something we aren't
willing to do ourselves.

Sorry, that was a lot longer than I intended.



Side note: It would be cool if there was a list of the current problems
with building FFmpeg with VC++ and Intel (named initializers with VC++, for
example).  That way various solutions could be explored and weighed.