[ty] remove unnecessary Either

[carljm]

Just a quick review-comment follow-up.

[github-actions]

mypy_primer results

No ecosystem changes detected ✅

[AlexWaygood]

It could be even simpler, FWIW...

--- a/crates/ty_python_semantic/src/types/infer.rs
+++ b/crates/ty_python_semantic/src/types/infer.rs
@@ -9021,9 +9021,9 @@ impl<'db> TypeInferenceBuilder<'db, '_> {
     ) -> Type<'db> {
         let arguments = &*subscript_node.slice;
         let (args, args_number) = if let ast::Expr::Tuple(t) = arguments {
-            (t.iter(), t.len())
+            (&*t.elts, t.len())
         } else {
-            (std::slice::from_ref(arguments).iter(), 1)
+            (std::slice::from_ref(arguments), 1)
         };
         if args_number != expected_arg_count {
             if let Some(builder) = self.context.report_lint(&INVALID_TYPE_FORM, subscript_node) {
@@ -9040,8 +9040,7 @@ impl<'db> TypeInferenceBuilder<'db, '_> {
         }
         let ty = class.to_specialized_instance(
             self.db(),
-            args.into_iter()
-                .map(|node| self.infer_type_expression(node)),
+            args.iter().map(|node| self.infer_type_expression(node)),
         );
         if arguments.is_tuple_expr() {
             self.store_expression_type(arguments, ty);

...but I don't think it's worth a followup PR just for this :P

[MichaReiser]

In which case you could even remove the len because of slice.len ;)

Maybe something for claude

[carljm]

PRs are cheap :) #18526

And this is too simple for Claude -- it'd be way more work to explain it to Claude than to just make the change.