[mythtv] Trying to debug seek problems

[Isaac Richards]

On Sunday 29 January 2006 09:08, David Abrahams wrote:
> Isaac Richards <ijr at case.edu> writes:
> > On Saturday 28 January 2006 22:11, David Abrahams wrote:
> >> Isaac Richards <ijr at case.edu> writes:
> >> > The problem is that VideoOutputQuartz::UpdatePauseFrame() isn't
> >> > updating the scratch frame's frame number like it should, since it's
> >> > using that style of pausing.
> >>
> >> Okay, so how should it update the scratch frame's frame number?  And
> >> why, if it's supposed to work that way, is it apparently perfect with
> >> my change?  What is broken that I'm not seeing?
>
> I'd still like to know what's broken by the change I proposed.  It
> would help me to understand the code and not waste as much of your
> time the next time around.

Your fix never updates the scratch frame's frame number - it could get out of 
sync with reality.  Putting it in UpdatePauseFrame copies the video data and 
the frame number at the same time, so that can't possibly happen..

> > Check how it's done in VideoOutputXv::UpdatePauseFrame.  Obviously, it
> > works there.  Simple grepping would have found that.
>
> Simple grepping only works when you know what to look for.  It wasn't
> obvious to me that VideoOutputQuartz was supposed to be following a
> pattern from VideoOutputXv.

<shrug>  Seems fairly obvious if you look for stuff in the output classes 
touching frameNumber and framesPlayed.

Isaac