Loring Tu gives the following definition of compatible charts in his book An Introduction to Manifolds (Second Edition, page no. 49).
Definition 5.5. Two charts $(U, \phi:U \to \mathbb{R}^n)$, $(V, \psi:V \to \mathbb{R}^n)$ of a topological manifold are $C^{\infty}$-compatible if the to maps $$\phi \circ \psi^{-1}: \psi(U \cap V) \to \phi(U \cap V), \quad \psi \circ \phi^{-1}: \phi(U \cap V) \to \psi(U \cap V) $$ are $C^{\infty}$. These two maps are called the transition functions between the charts. If $U\cap V$ is empty, then the two charts automatically $C^{\infty}$-compatible.
My Question
Which definition of smoothness is used for these transition functions?
I am asking this because smooth functions between two manifolds are defined in terms of infinite differentiability in the latter part of the book.
Why are the two charts automatically $C^{\infty}$-compatible, if $U\cap V$ is empty?