[mythtv] slight scheduling bug

[Bruce Markey]

Erik Arendse wrote:
> At 26-2-03 07:19, Cedar McKay wrote:
> 
>> howdy. I'm using cvs from yesterday. The schedule that appears in my 
>> program guide will frequently have a few "unknown" entries, 
>> particularly around midnight. No biggy. However, if I select one to 
>> record (for instance if I know something I want to see falls into that 
>> slot) all "unknown" shows that share stop and start times will be 
>> marked to record, regardless of the channel. So the same slot in 
>> channels 4, 5 and 7 might be marked to record. The conflict resolution 
>> screen shows no conflicts, and it doesn't appear to record any show 
>> whatsoever during that time slot. Doesn't really bother me too much, 
>> but maybe someone cares enough to give it a look.
> 
> 
> The "unknown" programs are not real programs to be recordable. In fact 
> the bug here is that they are markable for record at all.

That's true for time gaps where there is no program info but
sad truth is that all the midnight shows are in the database
("Tonight Show", "Late Show", etc.) but don't have an endtime
so the're marked as unknown so they seem to be the same title
in the same time slot. There's a new workaround that now marks
these as ending at midnight.

The problem is that xmltv isn't getting the endtimes from
zap2it and therefore use the starttime of the next show.
But when grabbing a days worth of data, at midninght there
is no next show. Mythfilldatabase then has no idea when these
shows end. When you look at the EPG it's obvious when the
show ends and it seems that myth could fix these endtimes
once the next day's data is loaded but should myth be
second guessing the data and changing endtimes after the
fact?

> In the future Myth should support real "freeform" recordings, perhaps 
> something like pressing "F" within the EPG on a channel, and entering 
> start/stop times (and perhaps a description) by hand.

Everyone suggests this but no one writes the code ;-). The
adjust timeslot in conflict resolution may be a good starting
point.

--  bjm