[mythtv-users] Re: Re: Fixing my "no data" problem in EPG
mythtv at braddolman.com
Wed Jul 27 12:05:19 EDT 2005
Thanks Chris & Michael,
I don't think I'm suffering from the 11 vs 14 day problem you
describe -- although I've hardly ever had a full set of EPG data, so
I may be having that problem too and just haven't seen noticed it yet.
The threshold that you both mention seems like the culprit to me. I
checked my current listings and I have 35 channels reporting "no
data" and 26 with data. So, even though there are a lot of missing
channels, I guess that it is returning a high number of programs for
after 6pm, so the listing isn't updated.
Unfortunately, I'm not compiling from source. So I don't think I can
fix this easily. I guess I'll just keep flushing my database and
restore every few weeks when it starts bothering me again. Perhaps
Michael's change should be merged into the release in an effort to
improve this area? It seems to me that the threshold is set
unnecessarily high currently.
Thanks again and regards,
> > Other solutions would be to find a way to adjust the threshold
> > percentage at which a day's schedule is considered complete, or
> to find
> > a way to make a certain range of future dates mandatory updates (so
> > that the percentage of filled time slots is not considered beyond
> > 10, for example, meaning that days 2, 11, 12, 13 and 14 are all
> > downloaded every night). Those probably require hacking the sources.
> This is what I have had to recently do. mythfilldatabase was checking
> that there where at least 4 times as many programs starting after 6pm
> for a given source as there are channels. This is in
> programs/mythfilldatabase/filldata.cpp . After changing the code to
> require only twice as many programs starting after 10pm as there are
> channels, my problem was fixed. (My data was only missing for cable
> channels after 10pm).
> If you search for "d 18" in that file, that is the sql line that gets
> the number of programs starting after 6pm (18:00), and about twelve
> lines down is a line that ends in "chancnt * 4", that decides if there
> are enough programs for that source. I ran the query manually against
> filled schedules, and broken schedules, to come up with the timespan
> and multiplier that fixed my issue.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the mythtv-users