I see the reason why there was an adjustment. No problem, and with the extra line in the .ini there is a acceptable solution. But from a technical point of view it is interesting what the reason is.
Today I did two additional tests:. One was online flying so the traffic was injected by the network. That network is not as big as the two major ones but the technology is in principle the same as VATSIM and the pilot client is also the same. The behaviour of the AI traffic in my EFB2 application was the same as by the two test former test with external AI sources: Without the fix no en-route traffic visible, with the fix all traffic was visible.
The other test today was flying in a environment where the AI traffic was generated by MSFS2020. In this situation the fix was not necessary: all trafic was visible with the default 7.4.18 version so without the fix.
So it seems that there is a difference in the way AI traffic is injected by external sources (or handled by MSFS) compared to the internal traffic of MSFS2020.
Regards,
Hans