TypeVarTuple as type arg for subclasses of generic TypedDict (fixes #16975) #16977
Conversation
for more information, see https://pre-commit.ci
| @@ -316,6 +324,7 @@ def analyze_typeddict_classdef_fields( | |||
| analyzed = self.api.anal_type( | |||
| stmt.type, | |||
| allow_required=True, | |||
| allow_unpack=True, | |||
There was a problem hiding this comment.
Leaving allow_unpack out here, or setting it to False, doesn't have an effect on the test suite... Intuitively I would guess that allow_unpack should be True here but would love to understand this a little bit better.
| @@ -520,6 +529,7 @@ def parse_typeddict_fields_with_types( | |||
| analyzed = self.api.anal_type( | |||
| type, | |||
| allow_required=True, | |||
| allow_unpack=True, | |||
There was a problem hiding this comment.
And the same applies here.
|
According to mypy_primer, this change doesn't affect type check results on a corpus of open source code. ✅ |
|
Not wanting to appear impatient, and I appreciate that the reviewers are quite busy -- but is there anything I can do to nudge this along? Perhaps my first sentence didn't inspire confidence:
For what it's worth, though, I took another long look and am now thinking this is in fact the right approach. If it helps makes things more palatable and you're unsure about the two |
(fixes #16975)
I'm not entirely sure about this implementation -- it fixes the issue and doesn't seem to break anything else -- but it feels like I might be doing things in the wrong phase? If so, a little nudge in the right direction would be much appreciated.