Mailing List archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[linux-dvb] Re: VDR 0.92 : No more EIT on Canal Satellite




----- Original Message -----
From: "Klaus Schmidinger" <Klaus.Schmidinger@cadsoft.de>
To: "Rolf Hakenes" <hakenes@hippomi.de>
Cc: <jrepetto@mxmlab.com>
Sent: Tuesday, August 21, 2001 12:25 AM
Subject: Re: [linux-dvb] VDR 0.92 : No more EIT on Canal Satellite


> Rolf Hakenes wrote:
> >
> > > -----Ursprüngliche Nachricht-----
> > > Von: Jean-Claude REPETTO [mailto:jrepetto@francenet.fr]
> > > Gesendet: Sonntag, 19. August 2001 21:53
> > > An: linux-dvb@linuxtv.org
> > > Betreff: [linux-dvb] VDR 0.92 : No more EIT on Canal Satellite
> > >
> > >
> > > I have noticed a small problem on the french channels of
> > > Canal Satellite on
> > > Astra. The current and next program aren't displayed anymore.
> > > I suppose it
> > > is related to the changes in the EIT processing. I haven't
> > > them with DTV
> > > either.
> > >
> > > Jean-Claude
> > >
> > Hi again,
> >
> > unfortunately the problem is not that small. It is caused by wrong
time-stamps
> > for the EPG data the CANALSATTELITE channels deliver. I have checked
this here,
> > the information is existent, but always in the future, therefore it is
not
> > possible to get a 'present' (and no 'following') broadcasting by
analysing the
> > time, as Klaus has to do since he uses my library. There are two
possibilities
> > to fix this,
>
> Well, actually there's a third one: they could adhere to the standard!!
> Gee, I hate to fiddle around to fix stuff that was intentionally broken.
>
> > either to re-introduce the 'present-following' flag to the
> > EPG-data, as I supposed to Klaus some mails earlier (with all the
drawbacks that
> > means), or to correct the time-stamp of the EPG. I would prefer the
second
> > choice, as only this enables using the recording feature in the correct
way. I
> > remember some discussions a long time ago concerning wrong TDT
time-stamps of
> > some channels (maybe Klaus knows more exactly about that, I am not sure
whether
> > these are the same channels), possibly a correction of the stamps is
possible
> > while using these TDT messages...
>
> This is getting more and more annoying. First Pro7 with their wrong date
on events
> between 00:00 and 06:00, now some French and Spanish channels who are
unable to
> send the timestamps in UTC - what's next?
>
> If it just were that easy to identify a specific channel, but as has been
posted
> in several messages here, the service ID is by no means unique, and using
the
> channel name is also pretty unreliable, because not everybody wants to
name the channels
> the same way.
>
> Isn't there _some_ bit in the EIT data stream indicating that the
timestamps are
> in local time instead of UTC? Maybe that's what Rolf was suggesting above:
if
> libdtv realizes that an event which is marked as "present" has a timestamp
that
> is more than, say, 30 minutes in the future, it could assume that all
events
> from this service ID have wrong timestamps and shift them by the
difference
> between the "present" event and the current time?
>
> Of course this would only be curing the symptoms. The real solution would
be
> if those braindead idiots would send the correct timestamp in the first
place...
>
> Klaus
> --

I have asked on french sat forums, and it appears that commercial receivers
(Nokia 9800S, XSAT CDTV310VM , ASTON 1600 ...) display the wrong time on
Canal Satellite channels.
I'll have a look at the EIT data stream to see if there is a bit indicating
that it is UTC or local time, but it won't help, since the local time is
unknown.
So I would suggest to add an optional time zone field in channels.conf, only
for the channels that are not in UTC, then to correct the time according to
the time zone.
For example, all Canal Satellite channels in local time would have a
"Europe/Paris" time zone.

What do you think about that ?

Jean-Claude







-- 
Info:
To unsubscribe send a mail to listar@linuxtv.org with "unsubscribe linux-dvb" as subject.


Home | Main Index | Thread Index