[FFmpeg-devel] [PATCH] rmdec.c: merge old/new packet reading code

[Ronald S. Bultje]

Hi,

On Tue, Mar 10, 2009 at 3:04 PM, Kostya <kostya.shishkov at gmail.com> wrote:
> And from what I remember seq might have been used for marking as keyframe
> only the video packets that contained first slice of video frame.

That looks right given other parts of the code using a variable called
"seq" in a similar way. That logic would never have worked since this
code is only reached if a complete frame is reached, so either this
variable is useless or this is the last frame and therefore it is
false.

I'll try to re-add that logic in another way. For now, I guess we
might end up losing one frame while seeking.

Ronald