On Feb 15, 4:27 am, Hrvoje Niksic <hnik...@xemacs.org> wrote: > Steven D'Aprano <st...@pearwood.info> writes: > > Suppose I have a function f() which I know has been decorated, but I don't > > have access to the original undecorated function any longer: > > > def reverse(func): > > def f(*args): > > args = list(args) > > args.reverse() > > return func(*args) > > return f > > > def say(*args): > > print args > > > rsay = reverse(say) > > del say > > > Is there any way to peek inside the decorated function rsay() to get access > > to the undecorated function say()? > > This works in Python 2.5.2: > > >>> rsay.func_closure[0].cell_contents > > <function say at 0xb7e67224> > > Of course, this applies only if you know there's only one free > variable, and you know that the decorator is in fact implemented with > a closure, and so on. > > >>>> dir(rsay.func_closure[0]) > > ['__class__', '__cmp__', '__delattr__', '__doc__', > > '__getattribute__', '__hash__', '__init__', '__new__', '__reduce__', > > '__reduce_ex__', '__repr__', '__setattr__', '__str__'] > > I got 'cell_contents' as well when I did the 'dir', at least under > Python 2.5.2.
Assume the decorator preserves the function at all; that is, it doesn't discard it. Then you can scan func_closure. In poor cases, you might have to search through a nested decoration as well. In others, such as when the decorator is an object, not a function, you might have to search an instance dictionary, also possibly nested. The leading assumption also requires that the function is stored in a Python object; that is, not in a C structure. In Python 3.0.1: >>> def f( fun ): ... def g( *ar, **kw ): ... return fun( *ar, **kw ) ... return g ... >>> def h( x ): ... print( x ) ... >>> i= f( h ) >>> i.__closure__[0].cell_contents <function h at 0x00B59810> -- http://mail.python.org/mailman/listinfo/python-list
