Repository navigation
Conversation
|
There were various test failures because of a bad annotation in |
|
Now we have a problem with We could fix typeshed -- maybe the return type should be In any case these issues are very useful findings. Mypyc is a good at finding typeshed bugs! |
|
The match group thing keeps coming up, I'm in favour of a Union Any return (with maybe some overloads for when we don't need it): python/typeshed#3902 (comment) |
|
What's the status quo of the match group thing? Is it still a blocker of this? |
|
I think that it's still an issue. I kind of forgot about this PR. The typeshed fix should be fairly straightforward (but it needs to be synced back to mypy). |
|
|
|
Hi! @hauntsaninja, I'm working on bytes primitives, shall we sync the typeshed to fix current match.group issue? |
|
Sure, I opened #10862 |
|
Closing in favor of #10881. |
This adds minimal support for the
bytesprimitive type. This won'timprove performance until we implement some more primitives.
Actually, this will likely degrade performance a bit, since we now add
type checks for
bytesvalues.Unfortunately, most of the
bytesoperations we'd need are private inCPython, so we'll need to either depend on implementation details or
reimplement (or copy+paste) a bunch of code.
Work on mypyc/mypyc#644.