On Jul 2, 9:11 pm, Steven D'Aprano <steve +comp.lang.pyt...@pearwood.info> wrote: > John Salerno wrote: > > But why does the documentation say "The return value of the decorator > > need not be callable"? > > The thing returned by a decorator does not need to be callable, but if you > want to call it, then it better be! > > This is no different from this: > > my_func = "hello world" > my_func() # fails because strings aren't callable > > So if I do this: > > def decorator(func): > # ignores the function and returns a string > return "hello world" > > @decorator > def my_func(): > x = 1 > y = 2 > return x+y > > print(my_func) # prints "hello world" > my_func() # fails because strings aren't callable > > If that's useful to you, decorator syntax allows it. That is all the > documentation means. > > > And why, if I remove the decorator and just > > leave the two functions as if, does the call to move(roll_die()) work? > > Isn't that what the decorator syntax is essentially doing? > > The call move(roll_die()) is similar to this: > > temp = roll_die() > my_string = move(temp) > > which is perfectly fine, because you never call my_string. If you did, you'd > get the same error, because strings aren't callable. > > The decorator syntax is completely different. It is doing this: > > # replace the function roll_die with the output of move, which is a string > roll_die = move(roll_die) > # now try to call "roll_die", actually a string > roll_die() > > which is more like: > > move(roll_die)() > > See the difference? > > -- > Steven
Eesh, ok, I think I *still* don't quite get decorators, but I get it a little more now. Definitely not what I needed to do here. :) -- http://mail.python.org/mailman/listinfo/python-list
