CoordinatesSpec sporadic failure

[swalker2m]

[info] CoordinatesSpec
[info] 
[info] Coordinates Offsetting should
[info]   + be consistent with diff  + flip the RA when Dec overflowsCoordinates Diff should
[info]   + have identity (1)  + have identity (2)Coordinates Serialization should
[info]   + Support Java BinaryCoordinates Angular Offset should
[info]   + result in an angular separation equal to distance, regardless of bearingCoordinates Angular Separation should
[info]   + be in [0, 180]  + be 0 for the same Coordinates
[info[   + be consistent with RA offsetting at the equator
[info]   + be consistent with Dec offsetting (includes polar discontinuity)Coordinates Interpolation should
[error]   x be consistent with fractional angular separation
[error]    Falsified after 70 passed tests.
[error]    > ARG_0: Coordinates(RA(Angle(36.69999999999982°)),Dec(Angle(0.10000000000002274°)))
[error]    > ARG_1: Coordinates(RA(Angle(216.70000000000005°)),Dec(Angle(359.9°)))

[tpolecat]

That's a good edge case. When you go over the poles you have to flip the RA by 180° so if the rounding error smears the result across the pole you get wildly different answers.