cleanup

crates/ty_python_semantic/resources/mdtest/classes.md

@@ -25,3 +25,22 @@ B = bytes
 
 reveal_mro(C)  # revealed: (<class 'C'>, <class 'int'>, <class 'G[bytes]'>, typing.Generic, <class 'object'>)
 ```
+
+## Starred bases
+
+These are currently not supported, but ideally we would support them in some limited situations.
+
+```py
+from ty_extensions import reveal_mro
+
+class A: ...
+class B: ...
+class C: ...
+
+bases = (A, B, C)
+
+class Foo(*bases): ...
+
+# revealed: (<class 'Foo'>, @Todo(Starred expressions in class bases), <class 'object'>)
+reveal_mro(Foo)
+```

crates/ty_python_semantic/src/types/infer.rs

@@ -386,6 +386,8 @@ impl<'db> TypeContext<'db> {
             .is_some_and(|ty| ty.is_typealias_special_form())
     }
 
+    // TODO: we could just always use `Iterable[<expected_element_type>]`
+    // as the type context once <https://github.com/astral-sh/ty/issues/1576> is fixed.
     pub(crate) fn for_starred_expression(
         db: &'db dyn Db,
         expected_element_type: Type<'db>,
@@ -401,8 +403,6 @@ impl<'db> TypeContext<'db> {
             ast::Expr::Tuple(_) => {
                 Self::new(Some(Type::homogeneous_tuple(db, expected_element_type)))
             }
-            // `Iterable[<expected_element_type>]` would work well for an arbitrary other node
-            // if <https://github.com/astral-sh/ty/issues/1576> is implemented.
             _ => Self::default(),
         }
     }