By the usage of compose! in main [30 and 34], you should see clearly what it does: it takes as many closures as you want, and combines them into a new closure that runs them one by one. This is really useful for runtime user-driven functionality composition.

The macro is implemented similarly to the standard macro for variable arguments from Chapter 1, Learning the Basics; Accepting a variable number of arguments, with its edge case being a single closure [6]. When encountering more, it will recursively go through them and combine them in pairs by calling the helper function compose_two [14]. Usually, type parameters are written as a single character, but we are using full words for them in this recipe for readability reasons, as there are quite a number of types involved. The type constraints used should illustrate how the types are used pretty well [18 to 20]:

where
    FunOne: Fn(Input) -> Intermediate,
    FunTwo: Fn(Intermediate) -> Output,

FunOne is a closure that takes an Input, turns it into an Intermediate, and passes it to FunTwo, which returns an Output. As you can see from the implementation, the only thing we do is call fun_one on a value, and then call fun_two on its returned value [22]:

move |x| fun_two(fun_one(x))