Re: [FFmpeg-devel] [PATCH] avformat: set start_time_realtime when demuxing

[Andreas Rheinhardt <andreas.rheinhardt@outlook.com>]

Gyan Doshi:
> 
> 
> On 2022-06-15 03:18 am, Andreas Rheinhardt wrote:
>> Gyan Doshi:
>>> Add new flag AVFMT_START_REALTIME for the couple of demuxers that set it
>>> internally.
>>> ---
>>>   doc/APIchanges         | 3 +++
>>>   libavformat/avformat.h | 1 +
>>>   libavformat/demux.c    | 3 +++
>>>   libavformat/rtsp.c     | 3 ++-
>>>   libavformat/version.h  | 2 +-
>>>   5 files changed, 10 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/doc/APIchanges b/doc/APIchanges
>>> index 5857e67ae6..4aa9e36777 100644
>>> --- a/doc/APIchanges
>>> +++ b/doc/APIchanges
>>> @@ -14,6 +14,9 @@ libavutil:     2021-04-27
>>>     API changes, most recent first:
>>>   +2022-06-xx - xxxxxxxxxx - lavf 59.26.100 - avformat.h
>>> +  Add AVFMT_START_REALTIME flag.
>>> +
>>>   2022-06-12 - xxxxxxxxxx - lavf 59.25.100 - avio.h
>>>     Add avio_vprintf(), similar to avio_printf() but allow to use it
>>>     from within a function taking a variable argument list as input.
>>> diff --git a/libavformat/avformat.h b/libavformat/avformat.h
>>> index f12fa7d904..4e247d68fc 100644
>>> --- a/libavformat/avformat.h
>>> +++ b/libavformat/avformat.h
>>> @@ -499,6 +499,7 @@ typedef struct AVProbeData {
>>>                                           The user or muxer can
>>> override this through
>>>                                          
>>> AVFormatContext.avoid_negative_ts
>>>                                           */
>>> +#define AVFMT_START_REALTIME 0x80000 /**< Demuxer sets
>>> start_time_realtime */
>> This is incorrect: If this were merged, start_time_realtime would always
>> be set by libavformat; whether it is done by the generic code or by
>> demuxer specific code is an implementation detail and irrelevant for the
>> user. Given that it is also set for demuxers without this flag actually
>> means that libavformat lies when using a demuxer without this flag.
> 
> This is moot since I'm switching to a new field but would either of
> these satisfy you?
> 
> s/Demuxer/Format/
> 
> or
> 
> /**< start_time_realtime is populated with conveyed origin wallclock
> timestamp */
> 

Of course not. My point was not about the semantics of the field being
populated, but about the flag and the fact that the field was populated
for demuxers that don't have this flag set. To remedy this, you should
stop setting start_time_realtime for demuxers that don't have this flag set.

- Andreas
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".