David Hewitt
David Hewitt
Thanks, yes we kept the `PyCell` concept around in documentation because we still have runtime borrow checking, however: - I can agree it's probably confusing and we probably haven't cleaned...
Having recently seen the blog after reading https://github.com/PyO3/pyo3/issues/416#issuecomment-2557013743, I wonder, would it be useful to folks if PyOxidizer was housed under the PyO3 GitHub organization? IMO it's an important part...
Sure thing @indygreg . Perhaps let's follow up via either email or the PyO3 discord to figure out the logistics of the transfer and what permissions need to be reassigned...
`#[pyo3(text_signature = "(index=1)"]` should go on your `#[new]` function since PyO3 0.19.
Ah, thanks for flagging - docs need fixing!
The text signature ends up just below the class header, rather than on the `__new__` help, e.g.: ``` Help on class Foo in module pyo3_scratch: class Foo(builtins.object) | Foo(index, /)
I opened #305 specifically for the case of `-C instrument-coverage`.
What you are doing is inherently `unsafe` given PyO3's lack of support for subinterpreters, but it's possible I can help work out a subset of operations which you _can_ do...
I would suggest looking at the problem like this: - sharing almost all objects between subinterpreters is unsound; PyO3's `Py` does nothing to prevent this so if you have these...
> Unfortunately, we haven’t found any public interfaces in the latest version of PyO3 that allow us to manipulate the internal counter directly. As a result, this issue cannot be...