[mythtv] hdtv reorg, crc checking + new gui patch

[Daniel Thor Kristjansson]

I haven't had time to generate a proper set of patches with my latest
changes, but I put up a big unified cvs diff here:
  http://www.mrl.nyu.edu/~danielk/mythtv/cvsdiff.patch3

I've adopted the MythTV brace style, and deleted any style changes I
could find. It's biggest advantage over earlier versions is that it has
B-Frame skipping capability. NuppelVideoPlayer turns on the avcodec's
hurry_up flag when it starts falling behind. Turning it back off isn't
properly tuned yet..

But in any case the problem I noticed doing this is that a lot of things
depend on the framecount reflecting the fps so I added a little hack to
avformatdecoder to progress the frames more quickly when frame skipping
is on (this is not exact, but can't really be without changes to the
whole mpeg decoding library). I think the real answer here is to use
audio/wall time as a reference instead of frames, since the frames shown
per second can vary especially if we support things like video
telephone in the future. But I welcome comments on this.

This patch should not be applied to CVS, though I'd love it if the
use-mouse and hdtv-videodev patches were applied...

Note: I broke the XvMC deinterlacing this time around, I don't know
exactly what went wrong, but I lost some changes in the C++ style
cleanup :/ But the frame skipping really makes up for it, I have no
performance problems with OSD anymore. I also played it for four hours
last night without a problem, but the sound can get out of sync
initially if frame skipping kicks in too quickly, skip ahead and back to
resync. This is because the timing hack expects the B-Frame skipping
to kick in immediately, and it doesn't.

-- Daniel
  << When truth is outlawed; only outlaws will tell the truth. >> - RLiegh

On Tue, 10 Feb 2004, Doug Larrick wrote:

]I am finally getting around to applying your patches.  I doubt I'll
]have time to do more than compile them tonight, though... my SO is
]using the TV.
]
]One quick comment... the prevalent brace style in MythTV is for an open
]brace to be on its own new line, even with the preceding statement.
]You will find your patches much more likely to be accepted w/o grumbles
]if you follow this convention... in fact, there are a few places
](hdtvrecorder.cpp in particular) where you have changed the style of
]existing code as you modified it.  This introduces unnecessary
]differences in your patches and makes them harder to follow.
]
]-Doug