[mythtv] EIT broken in SVN ?

[Dave]

Stuart wrote:
>  >
>  > Not all channels on all muxes supply EIT data. This is especially true
>  > with DVB-S setups, where it has been known to be confined to a single
>  > EIT data channel.
>  >
> 
> Ah, this is interesting. I definitely have 'useonairguide' switched off 
> of all the dvb-s transponder channels except the ones i watch. So maybe 
> the EIT is on a channel not being scanned. However, how does this 
> explain that when i am watching the channels concerned, epgsnoop will 
> retrieve the EPG data correctly. Seem like the active and passive 
> scanners are using different mechanisms to gather the data. What 
> algorithm does mythtv use to determine what channel EPG data is on ? How 
> does epgsnoop get it right. Can we not get mythtv to do the right thing 
> like epgsnoop ?
> 
 >
 > the active and passive scanners use the same mechanism, so that isn't
 > the problem.
 >
 > useonairguide on a channel is used to indicate if the EIT data should
 > be used for the channel. it is also set by the scanning if the
 > signalling tables indicate that there is EIT available for that
 > channel
 >
 > interesting that epgsnoop finds eit data. it indicates that the eit
 > data could well be present but just in a non standard place. we'll
 > need to do a bit of dvbsnooping to find out just where they might be.
 >
 > Please join us in #mythtv on irc.freenode.net and we can go through
 > it.

Ok, i've done some more snooping and i think i know what's going on. At 
some point in time on some channels (even on HB13E) we insert a blank '' 
program title spanning the *whole* 24hr period which wipes out existing 
valid program data.

Note, 'listingsource' is 1 meaning this is coming from EIT data and not 
me or a script inserting bogus data.

mysql> select chanid,starttime,endtime,listingsource FROM program where 
title='';
+--------+---------------------+---------------------+---------------+
| chanid | starttime           | endtime             | listingsource |
+--------+---------------------+---------------------+---------------+
|   2027 | 2008-02-28 02:00:00 | 2008-02-29 01:59:59 |             1 |
|   2027 | 2008-02-29 02:00:00 | 2008-03-01 01:59:59 |             1 |
|   4314 | 2008-02-28 04:00:00 | 2008-02-29 03:59:59 |             1 |
+--------+---------------------+---------------------+---------------+
3 rows in set (0.00 sec)

Always the start and end time is spanning a whole day (or probably is a 
negative number in the start/end time calculation)

As i can't ask for the satellite provider to fix this, we need to handle 
this in mythtv or just drop empty title EPG data ?

Stuart comments ?
Cheers
D