[mythtv] Ticket #1790: pes packet assembly in mythtv-eit

[Stuart Auchterlonie]

On Wed, May 17, 2006 at 05:30:33PM +0200, Janne Grunau wrote:
> On Wednesday 17 May 2006 16:18, Daniel Kristjansson wrote:
> >
> > Ok, then that sounds fine as a work-around for this broken SDT
> > problem for now. I'll sync the EIT tree to svn head then and so we
> > can start applying those outstanding patches to the EIT tree.
> 
> Im not 100% sure that the assembly is without bugs but it is working 
> Stuart Auchterlonie, me and another guy from irc. But if there is a 
> problem we can sort it out later.

One thing, the "Writing PAT & PMT" message should not be VB_IMPORTANT
I have it at VB_RECORD in my local tree.

> Another minor problem is that consecutive recordings fail. Stuart has 
> discovered it and I'm able to reproduce it. The problem seems to be 
> that we finished not only the first recording but also the second. When 
> TVRec is ready to start the second tuning fails sice the tuningrequest 
> is empty. I haven't tested trunk or anything other than DVB recordings 
> yet. relevant log excerpt: http://pastebin.ca/56336

As well as that, I've found another cause of some zero byte records.
Sometimes it ends up tuned to the wrong transport and so it
can't find the PMT it's looking for in the PAT.

since this has only happened to me since I've added the bbc hd
test mux, I'm going to have a careful look at the mux frequencies
and how close they are. see if I can work out why.

(and it seems to always end up on the new mux...)

> 
> I'm not too happy with the previous global eitscanner patch. I've 
> started to redoing it and I will probably move all EIT processing to 
> that single thread. This will make it easier to limit the load of EIT 
> processing.

This is where it should be. The scanner should control the cards
to retrieve the info for it, and then it can do the processing.

This allows us to have a place to implement more intelligent
scanning methods.


Stuart