[mythtv] [patch] bugfix for 166 & 167

[Isaac Richards]

On Tuesday 01 February 2005 05:26 pm, Daniel Thor Kristjansson wrote:
> I decided to go looking for bugs in the database since there is a
> release forthcoming. I found 166 and went about finding a better
> fix than running "xvattr -a XV_COLORKEY -v 0" because I knew that
> solution made the video show through on my konsole...
>
> This turned out to be caused by two problems. First, didn't set the
> frame size properly when painting the second frame. But fixing this
> leaves the line at the top. The only way to fix this is to disable
> the "XV_AUTOPAINT_COLORKEY" attribute, and not paint that line
> at the top to the colorkey color. Since this was being enabled
> before videoout_xv is told if the video is interlaced, I broke
> out the colorkey code from Init() into InitColorKey(). Then I just
> call this again after any call to VideoOutput::SetupDeinterlace().
>
> But this investigation led me to discover that bug 167 wasn't
> actually a duplicate. This is caused by always moving the second
> frame down 1 line. In fact we should be moving it down half a line in
> the source resolution. If the source is 480 lines, and our output is
> 1080 lines, then this is 2 lines in the output resolution.
>
> When I tried to apply these same fixes to videoout_xvmc the fields
> were displayed in the wrong order, so I fixed this as well. But
> otherwise that fix is basically the same.

Doesn't seem like it does the right thing for source = output = 480 lines of 
resolution with no overscan compensation, though.

Isaac