Moving PyObjectInit into impl_ and sealing the trait
#4611
Conversation
src/impl_/pyclass_init.rs
Outdated
| #[doc(hidden)] | ||
| fn can_be_subclassed(&self) -> bool; | ||
|
|
||
| private_decl! {} |
There was a problem hiding this comment.
This is a different method of sealing, it can be removed now, together with private_impl!() below.
There was a problem hiding this comment.
It would actually be a nice change in this PR (or a follow up) to replace all uses of private_decl! / private_impl! with sealing, and remove that older method.
There was a problem hiding this comment.
Sure, that means some of the traits in pyclass need to be Sealed, which is what I plan to do next. I think I will do it in another PR
src/sealed.rs
Outdated
| @@ -46,3 +49,7 @@ impl<T> Sealed for AddTypeToModule<T> {} | |||
| impl<T> Sealed for AddClassToModule<T> {} | |||
| impl Sealed for PyMethodDef {} | |||
| impl Sealed for ModuleDef {} | |||
|
|
|||
| impl<T: PyObjectInit<T>> Sealed for T {} | |||
There was a problem hiding this comment.
| impl<T: PyObjectInit<T>> Sealed for T {} |
I don't think we should add a blanket implementation for Sealed. If we implement it for types that we don't control we have to be careful to not open the seal. In this case it seems we don't need that impl at all, so we should just remove it.
There was a problem hiding this comment.
Thanks for catching that, I have no idea why I added that. Me just being silly I guess
| /// Initializer for Python native types, like `PyDict`. | ||
| pub struct PyNativeTypeInitializer<T: PyTypeInfo>(pub PhantomData<T>); | ||
|
|
||
| impl<T: PyTypeInfo> PyObjectInit<T> for PyNativeTypeInitializer<T> { |
There was a problem hiding this comment.
I'm happy with moving this to impl_ for now; I think there is a good chance we need to come up with a new way of initializing native super-types, but it won't stick around in this current form 👍
There was a problem hiding this comment.
Hope this got merged before further development 😅
github-merge-queue
bot
removed this pull request from the merge queue due to failed status checks
|
Looks like there's still some freethreaded flake to investigate another time |
Potentially closes #4552
PyNativeTypeInitializerandPyObjectInitare moved intoimpl_but notPyClassInitializeras it is used in the prelude and used in the lib so I assume that is available to the public.PyObjectInitis now a Sealed trait