On Tue, May 26, 2020 at 8:30 PM Greg Ewing <[email protected]>
wrote:
> "Did the user supply a value for this optional argument?" is
> a simple and reasonable question to ask. It deserves to have a
> simple and direct way of answering it that always works.
>
It does so very little that I would definitely not want syntax ONLY for
that. The example of lru_cache() is god here... None is a sentinel, but it
is NOT the default value. So there is a value of 128 supplied, even if the
user doesn't supply one. Changing syntax for something that doesn't even
actually save a line, isn't worth it. I'm assuming you mean something
along the lines of:
def foo(a, b, mode=):
if undef mode:
mode = whatever
It's really just a sentinel, no better than None. If we want to save more,
we're back to PEP 505 and None-coalescing. I'm feeling more sympathetic to
that than I did in earlier discussion. But I'm still a bit horrified by
some of the use cases that were presented at that time (not directly in the
PEP). E.g.:
data = json.load(fh)
needle ?= data?.child?.grandchild?[key]?[index]?.attr
When it looks like that, I get visions of APL or Perl unreadability. So I
guess what I'm sympathetic with is about 1/3 of PEP 505. Like only:
stuff = this ?? that
Yes, there are still rare non-None sentinels to deal with differently, but
that is 90% of the issue being discussed.
--
The dead increasingly dominate and strangle both the living and the
not-yet born. Vampiric capital and undead corporate persons abuse
the lives and control the thoughts of homo faber. Ideas, once born,
become abortifacients against new conceptions.
_______________________________________________
Python-ideas mailing list -- [email protected]
To unsubscribe send an email to [email protected]
https://mail.python.org/mailman3/lists/python-ideas.python.org/
Message archived at
https://mail.python.org/archives/list/[email protected]/message/XBA4VCCRP4LLVZ3S7JYBZMDADF3C4ZPS/
Code of Conduct: http://python.org/psf/codeofconduct/
