Diagnostic diff on typing conformance tests

No changes detected when running ty on typing conformance tests ✅

mypy_primer results

No ecosystem changes detected ✅
No memory usage changes detected ✅

ruff-ecosystem results

Linter (stable)

✅ ecosystem check detected no linter changes.

Linter (preview)

✅ ecosystem check detected no linter changes.

Formatter (stable)

✅ ecosystem check detected no format changes.

Formatter (preview)

✅ ecosystem check detected no format changes.

@BurntSushi any suggestions on what we should change on the ruff/ty side as part of this update? I see from the commits that there's a new global cwd option. Is this something we have to or should set now?

Yeah, the WalkBuilder::current_dir method lets you set the directory used for matching gitignore patterns from global gitignore exclusions or files added via WalkBuilder::add_ignore. If you don't set it, then it will automatically be pulled from std::env::current_dir for every use of WalkBuilder::build{_parallel}.

The main reasons to use WalkBuilder::current_dir explicitly are:

1. You have a slightly different way of discovering the current working directory.
2. You use many WalkBuilder::build{_parallel} calls but only want to query std::env::current_dir once.
3. You don't want your libraries inspecting process global state for $reasons.

If those don't apply to ruff, then I think you can just leave it.

Thanks, that's helpful

2. applies for both ty and ruff (although I consider this more a nice to have rather than being very performance critical)
   1 and 3: Apply for ty where we want to sue OsSystem::current_directory over std::env::current_dir (they should be the same but for consistency's sake)