[mythtv] Is there anyone still on list who can give me some help with the channelscan_sm code?

[roger]

Hi,

I am in the process of rewriting the DVB SI table handling code in 
dvbstreamdata. The objective is to maintain robust in memory caches of 
some of the SI tables. At the moment these are the Service Description 
Table (SDT) and the Event Information Table (EIT). The code uses and 
combination of timestamps and cryptographic hashes to detect version 
wrap around.

The objective of this is that higher level users of dvbstreamdata should 
be able to rely on being passed complete tables not just random 
segments  and only when a new version of a table has been received. I am 
trying to avoid the multiple caches of table sections that are currently 
held in various places in the dvb/mpeg code.

The SDT is used in the channel scanning code in channelscan_sm.cpp. Is 
there anyone familiar with this code who can give me a heads up on the 
strategy being used to ensure a that a reaonably complete service (a.k.a 
channel) list is retrieved. Is seems to me at the moment that little 
distinction is being made between "actual" and "other" SDTs. I would 
assume a scanner would use the NIT "other" tables to discover other 
multiplexes and then attempt to grab "actual" SDTs from the  new 
multiplex before falling back to using "other" SDTs.

All help gratefully received.

Roger