[mythtv] Multiple tuner conflict resolution bug or feature?

Steve Davies steve at one47.co.uk
Fri Jan 16 13:02:19 EST 2004


Mark Chou said:
> I have a peculiar issue with conflict scheduling with 2 tuners (each on
> a different backend).  The output of "mythbackend --printsched" is
> attached below (with two cuts from brevity). Each of the shows last one
> hour; only a global pre/postroll of 60 seconds is specified; and each of
> the shows is a "channelrecord," i.e. record whenever it comes on this
> channel. Mythtv is from cvs dated 1/12.
>
> I'm wondering why the first Nova program (below cut1) gets overridden,
> and does not get assigned to "tuner 2," when it's clearly available.
> Instead the "overridden" (as shown by the 'O' in column N.  Nova gets
> recorded at a later time, 1/18 (it just happens to be a repeat).

I've recently been playing with the scheduler code - The algorithm will
generally try and resolve conflicts by leaving a program on the card/input
it was originally requested to be recorded on, if a resolution is
available that does not involve a card move, it is used.

> This was not the behavior in myth .12.  In myth .12 the first Nova would
> have been recorded.

Yup, same behaviour. There was probably never a repeat in your schedule
that could satisfy the scheduler, so in the past it had to use the second
card.

> How do I go about changing the behavior so it's similar to myth .12?  I
> really don't want recordings deferred to the last (repeat) showing,
> because if that recording messes up for whatever reason, it would mean
> that I would have completely missed that program :).
>
I do not promise rapid results, but I will investigate changing this
behavior along with the other "tweaks" I am doing. If you are feeling
brave, you could try moving the conditional call to "DoMultiCards()" above
the call to "PruneList()" in programs/mythbackend/sceduler.cpp - Might not
be eactly that, but from memory it is near line 220 in that file.

> Thanks,
> Mark
>
[snip]

All IIRC, IMHO etc etc...

Steve


More information about the mythtv-dev mailing list