Do the to_gregorian functions still lose nanoseconds?

[iamthad]

Epoch::to_gregorian_utc and Epoch::to_gregorian_tai include a warning in their doc comment that nanoseconds are lost.

There is a u32 value for nanoseconds in the returned tuple, though, and I do see that field checked in the appropriate tests in tests/epoch.rs. At least one of the tests has a nonzero value there, too.

Is this warning still applicable, or is the comment stale?

[ChristopherRabotin]

Good find! That comment is stale: these function both rely internally on compute_gregorian, which does indeed return the nanoseconds.