HexagonCubic and HexagonAxial inconsistency

[hsseung]

According to the cited reference,
HexagonCubic(q, r, s) == HexagonAxial(q, r), but convert isn't consistent about that convention.

julia> convert(HexagonCubic, HexagonAxial(1, 1))
HexagonCubic(1, 1, -2)

julia> convert(HexagonAxial, HexagonCubic(1, 1, -2))
HexagonAxial(1, -2)

vertices follows the second convention that HexagonCubic(q, r, s) == HexagonAxial(q, s), as shown by the examples below.

julia> collect(vertices(HexagonAxial(1, 1)))
6-element Vector{Any}:
 (3.464101615137755, 2.0)
 (2.598076211353316, 2.5)
 (1.7320508075688776, 2.0000000000000004)
 (1.7320508075688772, 1.0000000000000002)
 (2.598076211353316, 0.5)
 (3.4641016151377544, 0.9999999999999996)

julia> collect(vertices(HexagonCubic(1, 1, -2)))
6-element Vector{Any}:
 (0.8660254037844387, -2.5)
 (6.123233995736766e-17, -2.0)
 (-0.8660254037844385, -2.4999999999999996)
 (-0.8660254037844388, -3.4999999999999996)
 (-1.8369701987210297e-16, -4.0)
 (0.8660254037844384, -3.5000000000000004)

julia> collect(vertices(HexagonCubic(1, -2, 1)))
6-element Vector{Any}:
 (3.464101615137755, 2.0)
 (2.598076211353316, 2.5)
 (1.7320508075688776, 2.0000000000000004)
 (1.7320508075688772, 1.0000000000000002)
 (2.598076211353316, 0.5)
 (3.4641016151377544, 0.9999999999999996)