[FFmpeg-devel] [PATCH] avfilter: align data frame when needed

wm4 nfxjfg at googlemail.com
Sat May 6 12:56:14 EEST 2017


On Sat, 6 May 2017 11:20:13 +0200
Nicolas George <george at nsup.org> wrote:

> Le sextidi 16 floréal, an CCXXV, Kieran Kunhya a écrit :
> > I sent a patch about this nearly three years ago, it wasn't applied for
> > whatever reason.
> > This is pretty common if you don't use av_malloc.  
> 
> Thanks for pointing it.

Well, nobody expected that apparently you didn't know that you have to
use av_malloc for everything in FFmpeg exactly because of alignment
requirements. Usually every API user learns this the hard way. (I wonder
if you ever used the FFmpeg API outside of ffmpeg.c...)

> As for the rest, I can observe that my hope that some people will start
> acting like adult

I can't say your behavior was particularly adult like. You raised a fuzz
because a bug was found and we wanted it to be fixed.

Adults fix their bugs and don't blame them on others.

> and spend their time in constructive ways instead of
> spending it flaming scapegoats.

This isn't about "scapegoats" - we just tried to get rid of that bug.
(Still trying...)

Apropos "scapegoat" - you're blaming everyone here, without
acknowledging any responsibility yourself.

What we really don't want in this project is that every time a bug is
pointed out, someone freaks out and raises a shit storm.

> And I am quite fed up to be the
> designated scapegoat just because I happen to see flaws in other
> people's design.

"Seeing flaws in other people's design" is ok, but your behavior
otherwise was a little questionable.

By the way, I can't help to observe that you refused to fix the bug
since it was not your fault (even though your change triggered it).
Fine, but what exactly makes you expect others to fix that bug? These
"others" are even less responsible for the bug than you.

That would still have been fine, but then you rejected bug fix patches
by others. Now I'm the first to say that we should fix bugs properly,
instead of painting them over with hacks, but let me remind you that no
better solution was in sight (or anyone working on it), and the crash
still happened, so applying whatever short-term fixes we had would have
been fine. (Or alternatively, reverting the patch that triggered it.)

> This project is in sore need of an Architect (
> http://esr.ibiblio.org/?p=7478
> ), but I do not see many around here.

That's... rich. I hope you don't claim to be that Architect. If you
think that's an insult, be aware that you just insulted everyone else
here. (Fortunately you added a "many" so that the reader can chose
whether he should be affected by the insult or not. How funny of you.)

> So I have taken the moral high ground, and just posted a proposal for an
> actual fix, despite my statement of not intending to do so.

This is amazing. When I saw your patches I thought you started to
understand (although I wasn't impressed by the lack of implementation),
but here you go and turn that around into yet another passive
aggressive bout. Very adult-like.

Maybe we can avoid all this bullshit the next time?


More information about the ffmpeg-devel mailing list