[FFmpeg-devel] Reintroducing FFmpeg to Debian

Michael Niedermayer michaelni at gmx.at
Sat Aug 23 20:56:23 CEST 2014


Hi

On Thu, Aug 21, 2014 at 03:16:50AM +0200, Attila Kinali wrote:
> Servus,
> 
> On Wed, 20 Aug 2014 18:43:18 +0900
> Norbert Preining <preining at logic.at> wrote:
> 
> > By continuing old fights, inspite of the very clearly friendly and
> > open offers and suggestions byu Michael, you and others from AV continue
> > simply to insult and be nasty.
> 
> Sorry, but this is not true. Yes, Michael always offers to mend ways
> and to do this and that. But he never does. It's just lip service.
> He didn't do it before the split and he didn't happen after the split.

Theres something about your statements, the one above as well as the
ones in other mails in this thread that i dont fully understand

You said in this thread that "i'm not a developer
and have not written one line of code for libav. i dont even read
the mailinglists"
(https://lists.debian.org/debian-devel/2014/08/msg00782.html)

and IIRC you also didnt read the ffmpeg mailing lists in the time
before the fork, except some specific threads, nor are you on the
ffmpeg IRC channels and i belive you arent subscribed to the FFmpeg
mailing lists now either. I definitly found a unsubscribe statement
for your address from march 2011 in the logs.

Given that, how can you even know anything about FFmpeg or what its
developers do or say or not do or not say
?

FFmpeg development discussions are mainly on our mailing list. with
a tiny bit on our IRC channels

I know you did read some specific threads which where related to
the fork as you participated in them. But these where possibly the
most heated and aggressive discussions we ever had on
ffmpeg-devel. And certainly are not representative for FFmpeg either
before or after the fork.


> Yes, the people at libav are bitter. Yes, they are angry. But how
> would you feel if people would walk up to you at random OSS events
> and tell you that X just told them you are the bad guy, that you
> steal childrens ice cream? (Yes, this has happend)
> 
> 
> > I am really impressed by the ability of Michael to take this without
> > changing into a more inpolite tone. Which you and others from AV are
> > definitely doing.
> 
> If you mean me by "others", then i would like to ask where i have been
> impolite.
> 
> > * AV maintainers are averse o any cooperation, and just licking their
> >   wounds since several years
> 
> You know, that FFmpeg and libav have been cooperating ever since the
> split? FFmpeg merges all (or almost all?) patches commited to libav

> without further review.

normally (for example in the kernel) code is reviewed before it is
pushed and not when it is merged.
Yet i _try_ to review and test what i merge from libav.
To proof that, here are some examples of bugfixes and decissions
based on such reviews. These wouldnt exist if the code wouldnt be
reviewed.

4 days ago:
libav added some () to a condition:
https://github.com/libav/libav/commit/d456baafb68cd80c0f537f1d843076e4dd853558

on the ffmpeg side the code was instead changed the following way a
year ago
http://git.videolan.org/?p=ffmpeg.git;a=commitdiff;h=1e2ab98460761c86268993e7a7ee690876df5efd
libav decided not to pick this change, and instead made their own
different change as shown above

and in the merge i tested both to double check that our solution is
correct and libavs isnt, and kept our solution.
More precissely libavs breaks decoding of
http://samples.ffmpeg.org/ffmpeg-bugs/roundup/issue414/black_screen_VC-1.mkv


or for example 16days ago
libav fixed a few memory leaks
https://github.com/libav/libav/commit/5b220e1e19c17b202d83d9be0868d152109ae8f0
but they used the wrong deallocation function, that is free() instead
of av_free*() which can lead to memory corruption,
Multiple people noticed that and it was fixed in the same git push
that pushed the merge
http://git.videolan.org/?p=ffmpeg.git;a=commitdiff;h=92deb28945a5f2b58908d383f183cfc1bc1d7fae
also there where multiple other related issues found when reviewing
the change from libav which where fixed in ffmpeg in the same push as
well:
http://git.videolan.org/?p=ffmpeg.git;a=commitdiff;h=12b59e57f3d7a37ef7b29d8a1df5eb886b00b4ba
http://git.videolan.org/?p=ffmpeg.git;a=commitdiff;h=31eaecfee9d84381945f3d5201775b9b00161d7a
http://git.videolan.org/?p=ffmpeg.git;a=commitdiff;h=92a28e9f562124732fa27f0c62118f15a6fee239
http://git.videolan.org/?p=ffmpeg.git;a=commitdiff;h=5f8300afc6537e2e06f8f90989d5f268884bb79c

but either way, id like to suggest again, we move forward and
rather discuss how we can improve the situation, do something about
the split and move toward un-doing it!

Thanks

[...]

-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

Everything should be made as simple as possible, but not simpler.
-- Albert Einstein
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 181 bytes
Desc: Digital signature
URL: <https://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20140823/b231171e/attachment.asc>


More information about the ffmpeg-devel mailing list