Categorical features for relief based feature selection

I am exploring Feature Selection methods and was wondering if the Relief based feature selection methods can be used for categorical or mixed-type features. From what I understand there is,

• relief: to be used for a binary target/response variable
• reliefF: to be used for multiclass target/response variable
• RreliefF: to be used for continuous target/response variable

But what about the features/explanatory variables? Can they be categorical or mixed type? Hopefully, used through one-hot-encoding?

It seems that relief based measures use a distance measure to compute similarities based on the target/response variable. In which case does it only make sense to use continuous features?