PEP 572: Add another UAEAP bullet point #729
Conversation
(This section would gzip very efficiently. "Unparenthesized assignment expressions are prohibited...")
| Example:: | ||
|
|
||
| (lambda: x := 1) # INVALID | ||
| lambda: (x := 1) # Valid, but unlikely to be useful |
There was a problem hiding this comment.
Hm. I'd much rather show an example here that is actually useful (and not obviously terrible style). E.g.
lambda line: (m := re.match(pattern, line)) and m.group(1)
Such a lambda might be useful (or be the starting point for something useful) with sorted(lines, key=...).
There was a problem hiding this comment.
That one wouldn't work in Python 3 because None and str are not comparable (None < str throws an error). Something like lambda line: m.group(1) if m := re.match(pattern, line) else '' would work I think.
There was a problem hiding this comment.
I'll keep the super-simple one, but also add the more useful one.
It won't really work as a sort key function, since None isn't comparable. Unfortunately, the if/else variant reads somewhat badly:
lambda line: m.group(1) if m := re.match(pattern, line) else ''
(it's confusingly similar to a comparison), and I'm not even sure how I'd write a bool-tuple version as a clean one-liner.
There was a problem hiding this comment.
This would work:
lambda line: (m := re.match(pattern, line)) and m.group(1) or ''but disqualifies itself by relying on the questionable A and B or C idiom. So let's stick to what Chris just added.
There was a problem hiding this comment.
I didn't ignore this. In the back of my head, I've been trying all day to dream up a realistic example of using an assignment expression in a lambda that didn't plain suck. The problem: if a lambda is fancy enough to benefit from one, it's already so fancy that I'd almost certainly write it as a def instead. This is the closest I got to a legitimate use, and it's still on the "use a def" side to me:
lowest_terms = lambda n, d: (n // (g := gcd(n, d)), d // g)There was a problem hiding this comment.
Agreed. I think here we're just looking to define the edge cases of the grammar precisely, and future generations may find a use for what the grammar allows. (As with so many other grammar edge cases.)
pep-0572.rst
Outdated
|
|
||
| This allows ``lambda`` to always bind less tightly than ``:=``; creating | ||
| a name binding inside the lambda function is unlikely to be of value, as | ||
| there is no way to make use of it. In cases where the name will be used |
There was a problem hiding this comment.
"creating a name binding inside the lambda function" ->
"having a name binding at the top level inside a lambda function"
| Example:: | ||
|
|
||
| (lambda: x := 1) # INVALID | ||
| lambda: (x := 1) # Valid, but unlikely to be useful |
There was a problem hiding this comment.
This would work:
lambda line: (m := re.match(pattern, line)) and m.group(1) or ''but disqualifies itself by relying on the questionable A and B or C idiom. So let's stick to what Chris just added.
|
AFUAEAPBP? :-) |
(This section would gzip very efficiently. "Unparenthesized assignment
expressions are prohibited...")