docs: update Python classes section of the guide
#3914
Conversation
| @@ -1303,7 +1304,7 @@ impl pyo3::impl_::pyclass::PyClassImpl for MyClass { | |||
| static DOC: pyo3::sync::GILOnceCell<::std::borrow::Cow<'static, ::std::ffi::CStr>> = pyo3::sync::GILOnceCell::new(); | |||
| DOC.get_or_try_init(py, || { | |||
| let collector = PyClassImplCollector::<Self>::new(); | |||
| build_pyclass_doc(<MyClass as pyo3::PyTypeInfo>::NAME, "", None.or_else(|| collector.new_text_signature())) | |||
| build_pyclass_doc(<MyClass as pyo3::PyTypeInfo>::NAME, "\0", collector.new_text_signature()) | |||
There was a problem hiding this comment.
Can you explain this bit?
There was a problem hiding this comment.
I just expanded the #[pyclass] macro using rust-analyzer and updated whats changed. 😇
There was a problem hiding this comment.
Thanks very much for kicking off the docs migration!
I think we should probably give existing users of PyCell a nudge to use Bound. Probably worth adding to the migration section as part of this PR?
I wonder if we should also deprecate PyCell completely, to give a stronger push? I suggested it before (I think it's an item on the tracking issue), and I think it'll make sense to do so given where we're heading for the API.
| let dict: &PyDict = unsafe { py.from_borrowed_ptr_or_err(self_.as_ptr())? }; | ||
| fn set(slf: &Bound<'_, Self>, key: String, value: &PyAny) -> PyResult<()> { | ||
| slf.borrow_mut().counter.entry(key.clone()).or_insert(0); | ||
| let dict = slf.downcast::<PyDict>()?; |
There was a problem hiding this comment.
👍 yep I think even though downcast_unchecked would be ok here I think best to recommend the safe option in the guide.
There was a problem hiding this comment.
Yes, that was the intention. The code from the guide is often copied as a first step, so unsafe should be avoided I think.
| @@ -627,6 +627,12 @@ impl IntoPy<Py<PyTuple>> for Bound<'_, PyTuple> { | |||
| } | |||
| } | |||
|
|
|||
| impl IntoPy<Py<PyTuple>> for &'_ Bound<'_, PyTuple> { | |||
There was a problem hiding this comment.
Yep this makes sense to add I think, for simplicity of using as argument tuple.
Icxolu
force-pushed
the
doc-guide-class
branch
from
c5dd57e
to
29070f6
Compare
I think that would make sense. I have worded the entry in the migration guide accordingly and prepared a commit doing the deprecation. I will push that as a followup PR since the diff is also decently large. |
Part of #3684
This adapts the "Python classes" section of the guide to the new
BoundAPI.