The Scalar::from_bits function allows constructing unreduced scalars with
exactly the given bit pattern, for compatibility with X/Ed2559. Prior to
1.2.3, this API had a sharp edge where it was possible to construct unreduced
scalars, then perform certain operations on them, and compute incorrect
results. This was fixed in 1.2.3; I don't believe it affected any actually
existing code.
The Scalar::from_bits function allows constructing unreduced scalars with exactly the given bit pattern, for compatibility with X/Ed2559. Prior to 1.2.3, this API had a sharp edge where it was possible to construct unreduced scalars, then perform certain operations on them, and compute incorrect results. This was fixed in 1.2.3; I don't believe it affected any actually existing code.