[FFmpeg-devel] [PATCH v2] avcodec/audiotoolboxenc: return external error if encode failed

[Steven Liu]

Andreas Rheinhardt <andreas.rheinhardt at outlook.com> 于2022年10月5日周三 20:14写道：
>
> Steven Liu:
> > "zhilizhao(赵志立)" <quinkblack at foxmail.com> 于2022年8月22日周一 10:49写道：
> >>
> >>
> >>
> >>> On Aug 19, 2022, at 11:14 PM, James Almer <jamrial at gmail.com> wrote:
> >>>
> >>> On 6/24/2022 4:05 AM, Steven Liu wrote:
> >>>> "zhilizhao(赵志立)" <quinkblack at foxmail.com> 于2022年6月24日周五 14:59写道：
> >>>>>
> >>>>>
> >>>>>
> >>>>>> On Jun 24, 2022, at 1:59 PM, Steven Liu <lq at chinaffmpeg.org> wrote:
> >>>>>>
> >>>>>> because the AudioConverterFillComplexBuffer can return 0 or 1 if
> >>>>>> success.
> >>>>>> so set the ret to 0 it AudioConverterFillComplexBuffer success and
> >>>>>> return ret value for success or return AVERROR_EXTERNAL when
> >>>>>> AudioConverterFillComplexBuffer failed.
> >>>>>> BTW change the error message log level from warning to error.
> >>>>>>
> >>>>>> Signed-off-by: Steven Liu <lq at chinaffmpeg.org>
> >>>>>> ---
> >>>>>> libavcodec/audiotoolboxenc.c | 6 ++++--
> >>>>>> 1 file changed, 4 insertions(+), 2 deletions(-)
> >>>>>>
> >>>>>> diff --git a/libavcodec/audiotoolboxenc.c b/libavcodec/audiotoolboxenc.c
> >>>>>> index f8305ab89b..00293154bf 100644
> >>>>>> --- a/libavcodec/audiotoolboxenc.c
> >>>>>> +++ b/libavcodec/audiotoolboxenc.c
> >>>>>> @@ -554,11 +554,13 @@ static int ffat_encode(AVCodecContext *avctx, AVPacket *avpkt,
> >>>>>>                                      avctx->frame_size,
> >>>>>>                            &avpkt->pts,
> >>>>>>                            &avpkt->duration);
> >>>>>> +        ret = 0;
> >>>>>>     } else if (ret && ret != 1) {
> >>>>>> -        av_log(avctx, AV_LOG_WARNING, "Encode error: %i\n", ret);
> >>>>>> +        av_log(avctx, AV_LOG_ERROR, "Encode error: %i\n", ret);
> >>>>>> +        ret = AVERROR_EXTERNAL;
> >>>>>>     }
> >>>>>>
> >>>>>> -    return 0;
> >>>>>> +    return ret;
> >>>>>> }
> >>>>>>
> >>>>>
> >>>>> LGTM.
> >>>> Applied, Thanks
> >>>
> >>> This is probably the source of the regression described in https://trac.ffmpeg.org/ticket/9866
> >>>
> >>> Can you look at it?
> >>
> >> There is a patch which doesn’t get apply yet:
> >>
> >> http://ffmpeg.org/pipermail/ffmpeg-devel/2022-June/298199.html
> >
> > commit as 627543f58a3166810b9cd9c8b483678c82a99be9
> >>
>
> You should backport this fix to 5.1. See ticket #9960.
Ok, cherry-picked.


Thanks

Steven