This PR aligns the evaluation of default arguments with other kinds of arguments by extracting the expressions within them as argument values to be evaluated.