[mythtv] [experimental patch] pchdtv bttv support

Isaac Richards ijr at case.edu
Mon Jan 17 21:54:27 EST 2005


On Monday 17 January 2005 09:25 pm, Daniel Thor Kristjansson wrote:
> On Mon, 17 Jan 2005, Isaac Richards wrote:
> ]On Monday 17 January 2005 06:25 pm, Daniel Thor Kristjansson wrote:
> ]>    http://www.mrl.nyu.edu/~danielk/mythtv/daughtercard-v4.tbz
> ]> Also this patch modifies the myth protocol so it updates the protocol
> ]> version. So you will need to relink any plugins.
> ]
> ]I don't really like this approach.  Why can't the TVRec object simply
> switch ]between the two types of recorder objects as needed?  I think that
> would be ]much, much simpler.
>
> That's what I thought too. This was my first implementation, and it
> didn't work very well. The scheduler gave me troubles and I had to
> create a completely new channel object that contained both regular and
> hdtv channel and implemented virtualized inputs on top of these. (The
> current channel object is very tied to a single device file).

Why not just switch channel objects, too?  I don't think it'd be difficult.  
What troubles were there from the scheduler?  I don't see how this is any 
different than a normal card with two different channel lineups attached to 
two different inputs.

> It 
> surprisingly turned out more complex. I also needed to keep two
> different ring buffers in TVRec or do some fancy footwork.

And again, why? =)  I don't see the need for two ringbuffers, as long as the 
player's modified to switch appropriately (which would be easy)...

> The beauty of this is that it isn't a special case for the pcHDTV card,
> it will work equally well for dealing with the radio tuner on other
> cards.

You're breaking the entire design of tvrec.cpp & the database layout though.  
I don't like the concept at all.

Isaac


More information about the mythtv-dev mailing list