Rebecca Chen

Move tests to end of solver.rs Summary: Moves a `mod test {...}` that was sandwiched in the middle of a file. No other changes. Reviewed By: grievejia Differential Revision: D115827083 fbshipit-source-id: e1307a251debc80a27043ba6cbc7523d4faf6645

Stop nested classes from using an outer class's scoped type parameters Summary: PEP 695-style type parameters declared by a class are not available inside a nested class, but Pyrefly allowed them in bases, defaults, and annotations. Record which names are class type parameters and stop lookup when it reaches one from an outer class, so the invalid use is reported. Reviewed By: grievejia Differential Revision: D115705194 fbshipit-source-id: f2657c711e7e9b25493d779954e9084c94f16172

Keep inferred types from the selected overload Summary: When a higher-order call selected one overload, type information learned while checking that overload was discarded. This could leave internal placeholder types in results such as `list(map(list, rows))`. Recheck the selected overload after choosing it so the inferred nested types are kept, while multiple possible overloads remain separate. Reviewed By: grievejia Differential Revision: D115799014 fbshipit-source-id: 6e5b6de2a1cef832d866ea88f25d2437a69bfb19

Update conformance sources Summary: Pyrefly's score goes down a bit because variance for TypeVarTuple and ParamSpec landed, which pyrefly does not yet support. Reviewed By: samwgoldman Differential Revision: D115950269 fbshipit-source-id: 4751524f7d637b01419226c773b81ac019289674

Add Type::Overloaded Summary: See https://github.com/facebook/pyrefly/discussions/4655 for context. This diff adds `Type::Overloaded`, which will be used to store the result of a higher-order function call when we need to preserve the structure of an overloaded function passed as an argument. Semantics: * `Overloaded` is assignable to a type if any of its alternatives is. This is consistent with overloaded functions. (Note that this diff does not add the `is_subset_eq` rule that implements this. It's pretty complicated, so it comes in a later diff so it can be reviewed on its own.) * Attribute access looks up the attribute on all alternatives and re-combines the lookup results. * Calls follow overload semantics. * Other operations (e.g., subscription) approximate overload semantics: alternatives with incompatible input signatures are dropped, and the results of the operation on the remaining alternatives are re-combined. `Overloaded` is not constructed anywhere yet; that'll come later. Reviewed By: stroxler Differential Revision: D117467797 fbshipit-source-id: 86e1b93fd394127b0dafaa19c77068dd7c1ec063

Prefer to insert `from` imports Summary: The previous diffs in this stack made it so that when we double-click on an inlay hint to import it, we also insert any required imports. However, the implementation defaulted to inserting `import foo` and `foo.X`, even though the hint shows `X` unless the module name is needed for disambiguation. This diff makes it so that we prefer to insert `from foo import X` and `X`. We still reuse `import foo` if it already exists in the file. Reviewed By: yangdanny97 Differential Revision: D114664767 fbshipit-source-id: 2c3ba897507b95750fb6fb7e31d5438a8ad16393

Split pyrefly_types/src/callable.rs into two files Summary: `callable.rs` was getting really long. This diff splits out a new `function.rs` (named so because it contains `FuncMetadata`, `FuncFacts`, etc.) and updates import paths accordingly, with no other changes. Reviewed By: maggiemoss Differential Revision: D114987156 fbshipit-source-id: 5dfec425605fbf2d7807f045bc50839c9446275e

Rename residual capture terminology Summary: See https://github.com/facebook/pyrefly/discussions/4655 for context. This diff is a pure rename that gets rid of most references to residuals, in preparation for moving away from them to a correlation table design. I did the rename first because the new logic is easier to understand if appropriate names are already in place. ___ Differential Revision: D117934032 fbshipit-source-id: 8c7c7658a10663851b9ce361348f88691f0c3f03

Make fields of `Tuple::Unpacked` private Summary: This is the second time (that I remember) that I've had to fix a subtle bug caused by direct construction of `Tuple::Unpacked`. Let's just make it impossible to construct directly. Note for reviewing: the only interesting parts of this change are in `crates/pyrefly_types/src/tuple.rs`, where I've added an `UnpackedTupleParts` struct with a few accessors. The rest is just mechanically rewriting code to use the accessors. Reviewed By: yangdanny97 Differential Revision: D115503925 fbshipit-source-id: c5a013b7f802d8cd2351df9200151c5ee9900086

Thread overload solutions through call inference Summary: See https://github.com/facebook/pyrefly/discussions/4655 for context. This diff adds an (empty) `OverloadTable` struct and threads it through call inference. This is a no-op change that adds the mechanical plumbing for `OverloadTable` so that it's not mixed with complicated solver logic in subsequent diffs, which require more careful review. ___ Differential Revision: D118248180 fbshipit-source-id: c6e8cd48462272beec139384a62bb119f1596551