[mythtv] Broken DEC2000-T filtering. (was generic digest subject)

[Taylor Jacob]

Quoting Nathaniel Daw <daw at gatsby.ucl.ac.uk>:

> On Wed, 20 Apr 2005 mythtv-dev-request at mythtv.org wrote:
>
> > It's not as trivial as you make it sound.. There are reasons most of the
> > masks were chosen the way they are.. There are some networks that misuse
> > the 0x46 and 0x42 tables and have them inverted.. EIT also seems to be
> > barely within spec as i've seen it most of the time and its easier to
> > just read whatever is being sent on that PID instead of opening up a
> > pile of section filters..
> Fair enough, but I am suggesting leaving the permissive mask unchanged and
> just changing the masked out part of the filter ID underneath it to align
> with the table you are most likely want. So for the channel tables that
> would be 0x42 instead of 0x46, and for EIT that would be 0x50 instead of
> 0x00. For drivers that properly implement masking, this shouldn't change
> anything.  If it doesn't, as on the dec-2000t, it should fall back to
> acceptable behavior, at least picking up the most important packet type.

Well if I request the table I most likely want, and filter out the one that I
end up wanting what good does that do?

> I agree it probably doesn't make sense to open up a whole pile of filters
> for the EIT -- I'm sorry I suggested that -- but I don't see how the
> narrower change would hurt anything. And it would make hardware work out
> of the box that right now mythtv can't even configure.

There are plenty of networks that send tables other than 0x50-0x5F, so this
really isn't an option, and besides how would fitering on 0x50 / 0xF0 help me
any if the masks/filters on 0x42/0x46 don't work as they are now?

Taylor