LaxZonedDateTime() <= LaxZonedDateTime()

[mattBrzezinski]

An unrepresentable LaxZonedDateTime is supposed to be treated like a NaN. However currently:

julia> using LaxZonedDateTimes

julia> LaxZonedDateTime() <= LaxZonedDateTime()
true

julia> NaN <= NaN
false

This problem can be corrected but doing so makes it impossible to use an ambiguous or non-existent LZDT as an AnchoredInterval anchor. This can also be corrected by making _isfinite(x::LaxZonedDateTime) = isrepresentable(x) but this implies that ambiguous/non-existent values are finite which doesn't seem right to me.

I suspect we may want to have non-representable values with ordering.