[mythtv] Re: [mythtv-commits] mythtv commits

[Isaac Richards]

On Friday 27 August 2004 12:15 am, mythtv at cvs.mythtv.org wrote:
> ---------------------------------------------------------------------------
>- Changes committed by cpinkham on Fri Aug 27 04:14:45 2004

>   The changes have been tested with mpeg2 files from ivtv-compatible cards
>   and with mpeg4/nuppel files from a bttv card.  The two things that need
>   heavy testing to make sure things weren't broken in the consolidation
> are:
>
>     1) the ivtv decoder output for people using a PVR-350 for output
>     2) the keyFrameAjustMap code for people who have transcoded recordings
>        and used the cutlist to cut out portions.
>
>   We need to make sure that skipping forward/backward in files works
>   correctly for the above two cases.  We wanted to get the code put into   
>       CVS to get it exposed to a much wider test base.  This is a much
> better solution for the short-term fix I put in place a couple weeks ago to
> fix the seeking problems people were experiencing especially when watching
> an in-progress recording.

The keyFrameAdjust code looks slightly wrong - GetKeyFrameAdjust is supposed 
to be equivalent to the old code?  Shouldn't it only be doing the adjustment 
if ki.key() == desiredKey?  Not sure, since I'm just going over it quickly...

Also, I think exact forward seeks are going to go (leftoverframes * 2) frames 
forware instead of (leftoverframes), right?  And shouldn't both calls to 
SeekReset() be using normalframes so the exactseeks setting is used?

Isaac