I have seen that floating point methods such as is_nan or is_finite, whose implementation is trivial, are implemented in libcore in the unstable Float trait and exposed in libstd as stable methods. This means that you can't use these methods with #![no_std]. Is there any reason it is done this way? Why not allow to use these methods without libstd?
I have also seen that to_degrees and to_radians methods in libcore were stabilized in rust 1.11 (according to the release notes). However, it turns out they are in the same situation described before.