Recursive function to develop permutations

[Steven Bethard]

Steve Goldman <steve_g <at> ix.netcom.com> writes:

> But now I have another question.  In the elegant recursive generator, below,
> the yield expression for the terminating condition is followed by a return
> statement.  Why is that necessary?
[snip]
> > def permute(Xs, N):
> >     if N <= 0:
> >         yield []
> >         return
> >     for x in Xs:
> >         for sub in permute(Xs, N-1):
> >             yield [x]+sub

The 'return' statement tells the generator to exit the function (without 
saving the state.  So,

>>> def gen():
... 	yield 1
... 	return
... 	yield 2
... 	
>>> list(gen())
[1]

Note that the second yield is never reached.  So the source of your example 
above chose to use the 'return' statement instead of the perhaps more 
intuitive (but slightly more indented):

def permute(Xs, N):
    if N <= 0:
        yield []
    else:
        for x in Xs:
            for sub in permute(Xs, N-1):
                yield [x]+sub

Steve