Mach speed uncertainty

[presidente20]

This PR adds the application of Mach uncertainty to lateral speeds in a similar manner to the way it is done for CAS.

• Apply speed uncertainty to Mach speeds in cas_and_mach_from_table.
• Add mach_cl/mach_cr/mach_des profile values and uncertainty parameters for all six weight classes in the performance profile and uncertainty data files (this will also be added in the Starling uncertainty file).
• Move the CAS uncertainty application out of the interpolation else branch so it runs in all cases, which is consistent with the way it is done for vertical speeds and now Mach speeds.
• Rework the predictor tests to drop the defunct "full mode"/"simple mode" terminology (now "user-supplied" vs "fallback" performance data).
• Bump the test_springfield::test_landing per-step evolve time (10 → 12) to account for the slightly slower time-to-fix introduced by Mach uncertainty.

I will create a Starling PR with some small follow up changes to match this PR.

[natsmike]

Looks good

[ahdgkn]

Requesting two changes:

1. Assign mach values such that the mach and cas profiles give the same TAS at each FL
2. Calculate the transition altitude from the cas and mach values of the aircraft after uncertainty is applied, rather than the nominal values (is_below_transition() in LinearPredictor) - this avoids need of perfectly aligning the CAS/Mach uncertainty

[ahdgkn]

Happy with changes to remove discontinuity at CAS-mach transition. Profiles for simple mode look realistic. Runs without bugs on a Springfield scenario.