[packman] ffmpeg-4 fails to compile

Hans-Peter Jansen hpj at urpla.net
Sat Jul 10 13:32:48 CEST 2021


Hi İsmail,

always nice to hear from you!

Am Samstag, 10. Juli 2021, 12:37:28 CEST schrieb İsmail Dönmez:
> Hi,
> 
> On Sat, Jul 10, 2021 at 12:33 PM Hans-Peter Jansen <hpj at urpla.net> wrote:
> > Hi Packmans,
> > 
> > in order to avoid duplicate work and to give you a heads up, I submitted a
> > fixed version of ffmpeg-4 to multimedia:libs
> > 
> > https://build.opensuse.org/request/show/905444
> 
> I don't think your patch is wrong, but I'd suggest backporting
> init_fail on top the old patch:
> https://github.com/cisco-open-source/ffmpeg/blob/master/libavcodec/ac3enc.c#
> L2490

Hmm, are you sure about it? Seems, you're looking into an old source.

I would have referred you to 

https://git.ffmpeg.org/gitweb/ffmpeg.git/blob/HEAD:/libavcodec/ac3enc.c

but that's failing for me.

Here's the head/master of GH mirror:

https://github.com/FFmpeg/FFmpeg/blob/master/libavcodec/ac3enc.c

I cannot see a valid approach to backport the validate_options() call to any 
goto label. The more interesting calls all you this pattern now:


    set_bandwidth(s);

    bit_alloc_init(s);

    ret = s->mdct_init(s);
    if (ret)
        return ret;

    ret = allocate_buffers(s);
    if (ret)
        return ret;

    ff_audiodsp_init(&s->adsp);
    ff_me_cmp_init(&s->mecc, avctx);
    ff_ac3dsp_init(&s->ac3dsp, avctx->flags & AV_CODEC_FLAG_BITEXACT);

    dprint_options(s);

    ff_thread_once(&init_static_once, exponent_init);

    return 0;
}


What do you think?

Cheers,
Pete





More information about the Packman mailing list