Maximilian Hils
Maximilian Hils
> I don't think it is correct to link #748 to this. The reported issue was specifically about particular members of Pydantic classes. Eh, you are correct. I did not...
Thanks for posting a potential solution! This is out of scope for pdoc itself for now, the workaround you posted seems to be a good approach.
I'd be open to a well-written PR if the PEP gets accepted, but to be frank I'm not a fan of the syntax at all. So unless this becomes widely...
We probably need to get https://github.com/python/typeshed/tree/main/stdlib/_typeshed into pdoc for this somehow.
We already try to evaluate the TYPE_CHECKING blocks, this is where the error is coming from. We're missing the _typeshed package import in those sections
pdoc does do that, in a way we already try to set `TYPE_SETTING = True`. But the `_typeshed` module is not in your typical Python install, which is why the...
More specifically: I think we need to vendor the _typeshed from the link above, and make it available as _typeshed during evaluation.
It turns out that vendoring isn't easy, `_typeshed/__init__.pyi` also bails on import [here](https://github.com/python/typeshed/blob/57c524d8225faae98611e6b643bb29769afdb3fa/stdlib/_typeshed/__init__.pyi#L343).
Unclear if this is a bug or if this is technically correct. At least theoretically there is a `Dog` member in the local namespace. Do the Python typing specs say...
We added TLS 1.3 Post Handshake Authentication in https://github.com/mitmproxy/mitmproxy/pull/7576. Stuff for TLS 1.2 and below is missing, contributions welcome!