On Thu, Dec 2, 2021 at 12:42 PM David Lukeš <dafydd.lu...@gmail.com> wrote:
>
> > That's an oversimplification:
> >
> > foo = 42
> > def bar():
> > print(foo)
> > foo = 1
> >
> > This won't print 42.
>
> Oh, right. I've run afoul of this in the past, but clearly not often
> enough to make it second nature :) Thanks for being helpful!
>
> I still think nudging people to only refer to previous arguments in
> their late-bound defaults is a good idea though, for readability and
> intuitiveness.
Oh yes, absolutely. Whether it's legal or not, referring to future
arguments will always be frowned upon.
> Maybe a good way to enforce it naturally (at least w.r.t.
> to arguments without late-bound defaults) would be via a similar syntax
> as keyword-only arguments? Something like:
>
> def foo(bar, *, baz=1, @, qux=[baz]):
> ...
>
That's rather restricting, and it implies that there's a major
difference in the way that functions would be defined ("late-bound
default keyword-only arguments" would become a brand new type of
argument) rather than a small difference in the way a function default
is defined.
> Now that I look at it, I actually like this bikeshed quite a lot. It
> trades some of the flexibility for expliciteness (the @ is very
> conspicuous in this position), consistency and predictability, which
> feels Pythonic. The @-as-barrier-to-eager-evaluation mnemonic still
> works. No potential => footgun.
>
> But it's late at night and I need to go to bed, so I probably messed
> something up again?
Not messed up, other than the fact that it's more change than I think
this needs. :)
ChrisA
_______________________________________________
Python-ideas mailing list -- python-ideas@python.org
To unsubscribe send an email to python-ideas-le...@python.org
https://mail.python.org/mailman3/lists/python-ideas.python.org/
Message archived at
https://mail.python.org/archives/list/python-ideas@python.org/message/SQBJDFUQR3UIKIG6FWWL4FYQT65GYWWT/
Code of Conduct: http://python.org/psf/codeofconduct/
