[x264-devel] does not link lavf support

[Steven Walters]

On Wed, Jun 22, 2011 at 9:21 AM, Erik Slagter <erik at slagter.name> wrote:
>> You sure are quick to point fingers and haze...
>
> Please note the phrase "I am not blaming the x264 devels..."
>
> In short: yes I overreacted a bit. I was quite annoyed by the attitude
> like "we'll fix it when we feel like" (see previous postings). I'm also
> annoyed by having two versions of ffmpeg/libav and both broke.
>
>> this issue has already been fixed in libav and ffmpeg: (libav's commit
>> was http://git.libav.org/?p=libav.git;a=commit;h=dbafb0e06faa092f60e53d845957fbab7f2a3f2d
>> ) and ffmpeg's was around the same time afaik.
>
> I've been pull'ing both forks like crazy the past few days, I've seen
> this commit come by and it didn't fix the segfault.
>
> Are you saying that after this commit you have a x264 cli that works
> withouth segfaulting (either with or without --input-res)?
>

correct, I am NOT able to reproduce a crash after the indicated
revision, nor after updating to the latest libav (which now has
libswscale at 2.0.0).
I tested on what platforms i have available to me which is win32,
win64, and linux32.

>> this is the first time I've heard of a libswscale crash, and why are
>> you blaming x264 for this?
>> have you looked at the stack trace? have you absolutely discerned that
>> x264 is at fault?
>
> There has been talk here on problems with swscale before. Libav supplies
> a sws_scale.so.1, while ffmpeg supplies a sws_scale_so.0. Furthermore,
> gdb "where" says it's in sws_scale when the segfault happens. So not
> exactly hard evidence, but imho quite a strong hint. I haven't checked
> with all debug options on yet, I am a bit reluctant to do so, as it
> looks like the stack gets trashed.
>
> And I don't "blame" x264 for this, I only suspect a not 100%
> api-adherance by x264 that only now, with recent changes on ffmpeg/libav
> appear.

I interpret this as "You have a bad reputation, so I'll blame you
because of that."

Regardless, as I previously mentioned, I am currently unable to
reproduce the crash.
Until it is proven that this is an issue with how x264cli is using
libswscale, I think your time would be better spent taking this up
with the libav/ffmpeg developers.

> I can handle (major) regressions perfectly fine, as long as I get the
> feeling that something will be done to fix them. I'll be most willingly
> to help fix the issue with debugging or whatever if someone will care.
>
> Thank you for your understanding.
>
>
> _______________________________________________
> x264-devel mailing list
> x264-devel at videolan.org
> http://mailman.videolan.org/listinfo/x264-devel
>
>