[mythtv] conflict resolution

[Derek Atkins]

"Edward Wildgoose" <Edward.Wildgoose at FRMHedge.com> writes:

>> Provided it does a "closest match", sure.  Let's say I have it set to
>> record at 7am.  The actual schedule has a 6:15a, 7:10a, and 7:55a.
>> Which would it record?
>
> Well, under the old system you wouldn't get anything because the
> program wasn't at 7am....

Well, ok, yea...  I concede that; I'm just trying to work wtihin your
strawman.

> ...I'm thinking about a fuzzy +- 1 hour match so in your case you
> are probably going to get a "random one" (as in depends on the
> actual algorithm finally decided on).

And I dont like this idea..  I dont want a random one, I want the one
closest to what I asked for.

> SQL is not great at ordered operations, it's a set based language,
> and trying to order by closest match is going to be slow and
> difficult I think.  I need to have a think about how to make the SQL
> re-usable so we get the same query everywhere in the code otherwise
> it's going to get a little hairy working out what is actually going
> to be recorded...

EHH?  What do you mean?

   SELECT ... FROM ... WHERE ... ORDER BY abs(starttime-now());

This does exactly what I want, and will order the results such that
the closest program (to now) is first.

> It seems you are better off whatever happens though?  

Maybe.  I'd prefer not to choose a random one.

-derek
-- 
       Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory
       Member, MIT Student Information Processing Board  (SIPB)
       URL: http://web.mit.edu/warlord/    PP-ASEL-IA     N1NWH
       warlord at MIT.EDU                        PGP key available