fix(simulator): incorrect initialisation of preflight checks

[philmoz]

When creating a new model for the T15 in the simulator the pre-flight checks state for the switches is incorrect.
SA is set to down instead of up.

Happens on MacOS, not tested on Linux or Windows.

[pfeerick]

Colorlcd specific issue - can reproduce on simulator with TX16S and EL18 in addition to T15 on Linux. PR appears to resolve this, no impact on hardware (since this is changing firmware, not just simulator). Seems somewhat hit and miss when it will occur... TX16S doesn't always have this issue, but T15 is pretty consistent.

[philmoz]

It depends on the number of switches - as soon as it goes over 11 it can happen.
On TX16S you would need to have SI, SJ and some EXT switches configured.

[pfeerick]

It's more than just that. 2.10.1 simulator does not do this with TX16S with default settings (but does for T15 - but only for SA), but 2.11 main does (and complains about all the switches)... even with the 2.10.1 ETX, which has no SI/SJ/ETX/FLx switches configured.

[philmoz]

It depends on how many switches pass the SWITCH_EXISTS() check.

The issue is the way different compilers handle the code - for some reason when building the simulator it is not correctly expanding the shifts to 64 bits.

[pfeerick]

Yeah, there have been quite a few occasions where firmware code acts differently for simulator than when built for radio... 😢

[pfeerick]

There is some sort of inconsistently / hard to reproduce occurrence whereby on with TX16S, if SI/SJ are configured 2POS, the simulator may sometimes still show the warning just for those switches if even if in the correct position, but it will not consistencly reproduce yet... i.e. it happened first run of simulator after restarting Companion with blank ETX + SI/SJ enabled, and then wouldn't do it for subsequent runs, and on a repeat run it wouldn't do it then either. However, the default settings for both TX16S and T15 are good now. 🤷