[FFmpeg-devel] [PATCH v2 4/6] lavc/qsvdec: remove orignal parser code since not needed now

[Li, Zhong]

> From: ffmpeg-devel [mailto:ffmpeg-devel-bounces at ffmpeg.org] On Behalf
> Of Mark Thompson
> Sent: Thursday, February 21, 2019 5:23 AM
> To: ffmpeg-devel at ffmpeg.org
> Subject: Re: [FFmpeg-devel] [PATCH v2 4/6] lavc/qsvdec: remove orignal
> parser code since not needed now
> 
> On 20/02/2019 02:58, Zhong Li wrote:
> > Signed-off-by: Zhong Li <zhong.li at intel.com>
> > ---
> >  configure           | 10 +++++-----
> >  libavcodec/qsvdec.c | 16 +---------------  libavcodec/qsvdec.h |  2
> > --
> >  3 files changed, 6 insertions(+), 22 deletions(-)
> 
> You can't remove this, it's still needed - the stream properties must be
> determined before the get_format() callback.
> 
> Similarly, you will need to extend the VP9 parser to return the relevant
> information for the following patch so that it works in general rather than
> only in cases where the user can supply it externally.  It should be quite
> straightforward; see 182cf170a544bce069c8690c90b49381150a1f10.
> 
> - Mark

There are something different from vp8 for vp9. 
VP9 frame resolution probably can't be got from bitstream but may be referred from reference frames, see: https://github.com/FFmpeg/FFmpeg/blob/master/libavcodec/vp9.c#L565 
If parsing header is separated form decoding process, it will be a problem how to get the reference list information.