scoder
I noticed that the separate storage of Unicode and bytes strings was actually unnecessary. We can merge bytes and Unicode strings into a single byte sequence, with the first byte...
I agree with your analysis. There will usually be a mix of unique and non-unique signatures, but I don't think it's worth distinguishing the two. We can always move the...
This code will be removed by https://github.com/cython/cython/pull/6684
> doesn't cover everything that could potentially be removed so I don't think it closes this. I'd say that what remains is mostly the positional argument unpacking code. That's really...
The link is the ability to retrieve [non-volatile UTF-8 buffer pointers](https://docs.python.org/3/c-api/unicode.html#c.PyUnicode_AsUTF8AndSize) of Python Unicode strings in Py3. `default` is a synonym of UTF-8 now. The remaining question probably is whether...
Reopening since we still need to decide about the auto-conversion of UTF-8 strings.
Can we just disable this warning in 3.0.x, maybe? The mistake that it reports doesn't seem something that many users would do.
> Not huge problem since `Optional/Union` is mainly for type annotation. I'd say "only" rather than "mainly". It feels wrong to allow them as C style type declarations at all.
Regarding the general approach, `Union[…, None]` is really redundant with `Optional[…]`. If we wanted to avoid implementing both, we could apply the following two rules: - transform `Union[…, None]` into...
Thanks, yes, that makes the implementation much shorter.