[FFmpeg-devel] [PATCH] Push hls_ts_options to every chunks (fix #5525)

[ffmpeg at jack.fr.eu.org]

resend_headers seems to be only related to PAT/PMT stuff
Calling avformat_write_header multiple times do not seems that creepy to
me, because we are handling multiples ts chunks (so, basically, multiple
headers must be written)

Pushing the 'system_b' options to every mpegts chunks through
resend_header means that system_b implies PAT/PMT for every chunks (the
current 'resend_headers' behavior), with is not a good solution

Thanks !


On 15/05/2016 00:14, Michael Niedermayer wrote:
> On Mon, May 09, 2016 at 06:17:36PM +0200, Jack wrote:
>> Signed-off-by: jack <ffmpeg at jack.fr.eu.org>
>> ---
>>  libavformat/hlsenc.c | 5 +++++
>>  1 file changed, 5 insertions(+)
>>
>> diff --git a/libavformat/hlsenc.c b/libavformat/hlsenc.c
>> index a9fa5d8..77712d0 100644
>> --- a/libavformat/hlsenc.c
>> +++ b/libavformat/hlsenc.c
>> @@ -624,6 +624,11 @@ static int hls_start(AVFormatContext *s)
>>          err = avformat_write_header(vtt_oc,NULL);
>>          if (err < 0)
>>              return err;
>> +    } else {
>> +        HLSContext *hls = s->priv_data;
>> +        av_dict_copy(&options, hls->format_options, 0);
>> +        avformat_write_header(hls->avf, &options);
>> +        av_dict_free(&options);
> 
> why does resend_headers not resend the headers?
> 
> calling avformat_write_header() multiple times looks odd unless theres
> something iam missing
> 
> [...]
> 
> 
> 
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel at ffmpeg.org
> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>