PSI: parse multiple sections correctly

[tmm1]

cc #25

[coveralls]

Coverage decreased (-76.6%) to 0.0% when pulling d1f8a96 on tmm1:psi-parse-multiple into 077af8e on asticode:master.

[barbashov]

It seems that &3 is not correct here, it should be 0xf, since the specification says that section length is 12 bits long.

[barbashov]

FYI: I always use this copy of spec: http://www.itu.int/rec/T-REC-H.222.0-201410-S/en

[tmm1]

On page 50 it says the section_length is only 10 bits:

section_length – This is a 12-bit field, the first two bits of which shall be '00'. The remaining 10 bits specify the number of bytes of the section, starting immediately following the section_length field, and including the CRC. The value in this field shall not exceed 1021 (0x3FD).

[barbashov]

Fair enough, I missed that one. Can you please fix this one in muxer as well?

[asticode]

To be honest I don't quite get why removing || tableID.isUnknown() would fix your issue here 🤔 Can you explain it a little bit?

[asticode]

Can you just replace this by an else if?

[tmm1]

To be honest I don't quite get why removing || tableID.isUnknown() would fix your issue here 🤔 Can you explain it a little bit?

The issue was that when an unknown table was seen previously, parsing stopped altogether. This means that a known table later in the same packet (such as the program_map) was ignored altogether.

[asticode]

OK, got it. In that case, I feel like doing the following would better fix the problem and clean everything behind:

1. Remove the shouldStopPSIParsing function
2. Make sure the parsePSISectionHeader function returns stop and upon calling it, add an else if stop { return } (and hence remove the if shouldStopPSIParsing(s.Header.TableID) below)
3. Replace this if with if tableID == PSITableIDNull directly
4. Remove this function

What do you think?