Inject a virtual SDT instead of mapping PID 17 from the first PMT

[Jalle19]

Should fix accidentally overwriting PID 18 (the virtual EIT) on the DDCI adapter with an SDT that has been mapped from PID 17 to PID 18.

Still not runtime tested, but the test cases pass now.

[Jalle19]

Considering rewriting ddci_create_epg to return just the table, then we can use buffer_to_ts just like we do for PAT, SDT and PMTs. I also believe it's not necessary to write an EIT for the virtual CA service ID (MAKE_SID_FOR_CA), so we could skip that. The main point is that the EIT PID should be present, that's all the standard says.

[catalinii]

I added this for my CAM, I can test it next time it needs an update.

[catalinii]

I am fine either way...I will test this later anyway

[Yuri666]

Now my old Viaccess CI+ 1.2 CAM get URI and start decrypting after 3-4 secs instead 10-12 secs when without patch. Now decrypting starts at DDCI as fast as in embedded CI

[Jalle19]

Wow, that's great!

[Jalle19]

Runtime testing this too right now, so far so good

[Jalle19]

Testing out changes to the EIT PID at the same time. Biggest change is that we no longer include an EIT for the dummy SID, only for the SIDs defined in the SDT and that have PMTs. Up until now we've been doing it wrong, the specification says that when we're using:

actual TS, present/following event information = table_id = 0x4E

Then the table should only contain events for services in the current transport stream. The dummy SID is not part of the transport stream.

[Jalle19]

https://github.com/Jalle19/minisatip/commit/5a7c363a934470ce916942751e047ccb1a27bae2

[Jalle19]

Rebased