On 4/20/06, Greg Ewing <[EMAIL PROTECTED]> wrote:
> Some wilder ideas for keyword-only arguments:
> def spam(a, b, c, {d, e, f=x}):
> # d, e are mandatory keyword-only
> # f is optional keyword-only
I see several sources of complexity.
Long sequences where order matters are bad, but probably can't be
eliminated. We have this today.
A rest-of-the-arguments collector is a useful special case. We have
this today with *.
Arguments which can be either positional or keywords are tricky, but
also useful. We have this today.
A distinction between expected and unexpected keywords is awkward, but
sometimes useful. We have this today with **.
Today, we do not have any way to say that a parameter is keyword only
without also marking it unexpected. This motivates the addition of
def(pos1, pos2, *args, key1=XXX):
I also see the value of keyword-only arguments without unlimited
positional arguments. Whether it deserves syntax and what that syntax
should be (such as * or *None) may need a pronouncement eventually,
but it doesn't have to be complex.
The *new* complexity seems to come almost entirely from trying to
support keyword-only arguments without a default. I don't think
(mandatory) + (keyword only) + (no default) + (no sentinel)
is important enough to justify that complexity.
(That said, I would like a better way to tell whether the keyword was
passed explicitly; creating a sentinel outside the function always
seems awkward.)
-jJ
_______________________________________________
Python-3000 mailing list
[email protected]
http://mail.python.org/mailman/listinfo/python-3000
Unsubscribe:
http://mail.python.org/mailman/options/python-3000/archive%40mail-archive.com
