ruff/crates/ty_python_semantic/resources/mdtest
David Peter 71d711257a
[ty] No union with Unknown for module-global symbols (#20664)
## Summary

Quoting from the newly added comment:

Module-level globals can be mutated externally. A `MY_CONSTANT = 1`
global might be changed to `"some string"` from code outside of the
module that we're looking at, and so from a gradual-guarantee
perspective, it makes sense to infer a type of `Literal[1] | Unknown`
for global symbols. This allows the code that does the mutation to type
check correctly, and for code that uses the global, it accurately
reflects the lack of knowledge about the type.

External modifications (or modifications through `global` statements)
that would require a wider type are relatively rare. From a practical
perspective, we can therefore achieve a better user experience by
trusting the inferred type. Users who need the external mutation to work
can always annotate the global with the wider type. And everyone else
benefits from more precise type inference.

I initially implemented this by applying literal promotion to the type
of the unannotated module globals (as suggested in
https://github.com/astral-sh/ty/issues/1069), but the ecosystem impact
showed a lot of problems (https://github.com/astral-sh/ruff/pull/20643).
I fixed/patched some of these problems, but this PR seems like a good
first step, and it seems sensible to apply the literal promotion change
in a second step that can be evaluated separately.

closes https://github.com/astral-sh/ty/issues/1069

## Ecosystem impact

This seems like an (unexpectedly large) net positive with 650 fewer
diagnostics overall.. even though this change will certainly catch more
true positives.

* There are 666 removed `type-assertion-failure` diagnostics, where we
were previously used the correct type already, but removing the
`Unknown` now leads to an "exact" match.
* 1464 of the 1805 total new diagnostics are `unresolved-attribute`
errors, most (1365) of which were previously
`possibly-missing-attribute` errors. So they could also be counted as
"changed" diagnostics.
* For code that uses constants like
  ```py
  IS_PYTHON_AT_LEAST_3_10 = sys.version_info >= (3, 10)
  ```
where we would have previously inferred a type of `Literal[True/False] |
Unknown`, removing the `Unknown` now allows us to do reachability
analysis on branches that use these constants, and so we get a lot of
favorable ecosystem changes because of that.
* There is code like the following, where we previously emitted
`conflicting-argument-forms` diagnostics on calls to the aliased
`assert_type`, because its type was `Unknown | def …` (and the call to
`Unknown` "used" the type form argument in a non type-form way):
  ```py
  if sys.version_info >= (3, 11):
      import typing
  
      assert_type = typing.assert_type
  else:
      import typing_extensions
  
      assert_type = typing_extensions.assert_type
  ```
* ~100 new `invalid-argument-type` false positives, due to missing
`**kwargs` support (https://github.com/astral-sh/ty/issues/247)

## Typing conformance

```diff
+protocols_modules.py:25:1: error[invalid-assignment] Object of type `<module '_protocols_modules1'>` is not assignable to `Options1`
```

This diagnostic should apparently not be there, but it looks like we
also fail other tests in that file, so it seems to be a limitation that
was previously hidden by `Unknown` somehow.

## Test Plan

Updated tests and relatively thorough ecosystem analysis.
2025-10-01 16:40:30 +02:00
..
annotations [ty] Use typing.Self for the first parameter of instance methods (#20517) 2025-09-29 21:08:08 +02:00
assignment [ty] More precise type inference for dictionary literals (#20523) 2025-09-24 18:12:00 -04:00
binary [ty] Remove special casing for string-literal-in-tuple __contains__ (#19642) 2025-07-31 11:28:03 +01:00
boolean
boundness_declaredness [ty] Reformulation of public symbol inference test suite (#20667) 2025-10-01 14:26:17 +02:00
call [ty] Sync vendored typeshed stubs (#20658) 2025-10-01 10:11:48 +02:00
class [ty] Rename "possibly unbound" diagnostics to "possibly missing" (#20492) 2025-09-23 14:26:55 +00:00
comparison [ty] detect cycles in binary comparison inference (#20446) 2025-09-17 09:45:25 +02:00
comprehensions [ty] Async for loops and async iterables (#19634) 2025-07-30 17:40:24 +02:00
conditional [ty] Support as-patterns in reachability analysis (#19728) 2025-08-04 20:13:50 +02:00
dataclasses [ty] initial support for slots=True in dataclasses (#20278) 2025-09-07 18:25:35 +01:00
declaration
diagnostics [ty] Use fully qualified names to distinguish ambiguous protocols in diagnostics (#20627) 2025-09-29 12:02:07 +00:00
directives [ty] Infer slightly more precise types for comprehensions (#20111) 2025-08-27 13:21:47 +01:00
doc
exception [ty] Use separate Rust types for bound and unbound type variables (#19796) 2025-08-11 15:29:58 -04:00
expression [ty] Rename "possibly unbound" diagnostics to "possibly missing" (#20492) 2025-09-23 14:26:55 +00:00
function [ty] Improve disambiguation of types via fully qualified names (#20141) 2025-08-29 08:44:18 +00:00
generics [ty] Add tests for nested generic functions (#20631) 2025-09-30 08:44:18 +02:00
ide_support [ty] Include NamedTupleFallback members in NamedTuple instance completions (#20356) 2025-09-15 11:00:03 +02:00
import [ty] No union with Unknown for module-global symbols (#20664) 2025-10-01 16:40:30 +02:00
literal [ty] More precise type inference for dictionary literals (#20523) 2025-09-24 18:12:00 -04:00
loops [ty] Add more tests for subtyping/assignability between two protocol types (#20573) 2025-09-26 12:07:57 +01:00
narrow [ty] No union with Unknown for module-global symbols (#20664) 2025-10-01 16:40:30 +02:00
regression
scopes [ty] No union with Unknown for module-global symbols (#20664) 2025-10-01 16:40:30 +02:00
shadowing
snapshots [ty] Rename "possibly unbound" diagnostics to "possibly missing" (#20492) 2025-09-23 14:26:55 +00:00
stubs
subscript [ty] Rename "possibly unbound" diagnostics to "possibly missing" (#20492) 2025-09-23 14:26:55 +00:00
suppressions [ty] Consistent use of American english (in rules) (#19488) 2025-07-22 16:10:38 +02:00
type_compendium [ty] Infer more precise types for collection literals (#20360) 2025-09-17 18:51:50 -04:00
type_of [ty] Improve the Display for generic type[] types (#19667) 2025-07-31 19:45:01 +01:00
type_properties [ty] Fix subtyping of invariant generics specialized with Any (#20650) 2025-10-01 10:05:54 +00:00
type_qualifiers [ty] Allow annotation expressions to be ast::Attribute nodes (#20413) 2025-09-15 12:06:48 +01:00
unary
with [ty] Use typing.Self for the first parameter of instance methods (#20517) 2025-09-29 21:08:08 +02:00
.mdformat.toml
async.md [ty] Support async/await, async with and yield from (#19595) 2025-07-30 11:51:21 +02:00
attributes.md [ty] Rename "possibly unbound" diagnostics to "possibly missing" (#20492) 2025-09-23 14:26:55 +00:00
classes.md [ty] don't assume that deferred type inference means deferred name resolution (#20160) 2025-08-29 16:19:45 -07:00
cycle.md
decorators.md
del.md [ty] No union with Unknown for module-global symbols (#20664) 2025-10-01 16:40:30 +02:00
deprecated.md [ty] Consistent use of American english (in rules) (#19488) 2025-07-22 16:10:38 +02:00
descriptor_protocol.md [ty] Use typing.Self for the first parameter of instance methods (#20517) 2025-09-29 21:08:08 +02:00
enums.md [ty] Enums: allow multiple aliases to point to the same member (#20669) 2025-10-01 15:51:53 +02:00
exhaustiveness_checking.md [ty] Fix subtyping for dynamic specializations (#20592) 2025-09-26 15:05:03 +02:00
final.md
implicit_type_aliases.md [ty] no more diverging query cycles in type expressions (#20359) 2025-09-16 16:44:11 -07:00
instance_layout_conflict.md [ty] initial support for slots=True in dataclasses (#20278) 2025-09-07 18:25:35 +01:00
intersection_types.md [ty] Expansion of enums into unions of literals (#19382) 2025-07-21 19:37:55 +02:00
invalid_syntax.md
known_constants.md
mdtest_config.md
mdtest_custom_typeshed.md [ty] Remove Type::Tuple (#19669) 2025-08-11 22:03:32 +01:00
metaclass.md
mro.md [ty] Treat Hashable, and similar protocols, equivalently to object for subtyping/assignability (#20284) 2025-09-10 11:38:58 +01:00
named_tuple.md [ty] Use typing.Self for the first parameter of instance methods (#20517) 2025-09-29 21:08:08 +02:00
overloads.md [ty] Use C[T] instead of C[Unknown] for the upper bound of Self (#20479) 2025-09-23 14:02:25 +02:00
pep695_type_aliases.md [ty] Use typing.Self for the first parameter of instance methods (#20517) 2025-09-29 21:08:08 +02:00
properties.md [ty] "foo".startswith is not an instance of types.MethodWrapperType (#20317) 2025-09-10 11:14:26 +00:00
protocols.md [ty] Literal promotion refactor (#20646) 2025-09-30 14:22:36 +02:00
public_types.md [ty] No union with Unknown for module-global symbols (#20664) 2025-10-01 16:40:30 +02:00
statically_known_branches.md [ty] Rename "possibly unbound" diagnostics to "possibly missing" (#20492) 2025-09-23 14:26:55 +00:00
sys_platform.md
sys_version_info.md
t_strings.md [ty] Add support for PEP 750 t-strings (#20085) 2025-08-25 18:49:49 +00:00
terminal_statements.md [ty] improve lazy scope place lookup (#19321) 2025-07-25 07:11:11 +00:00
ty_extensions.md [ty] Fix CallableTypeOf[…] for classmethods (#20345) 2025-09-11 10:14:38 +02:00
typed_dict.md [ty] Use typing.Self for the first parameter of instance methods (#20517) 2025-09-29 21:08:08 +02:00
union_types.md [ty] Simplify unions of enum literals and subtypes thereof (#20324) 2025-09-10 15:54:06 +02:00
unpacking.md [ty] Infer more precise types for collection literals (#20360) 2025-09-17 18:51:50 -04:00
unreachable.md [ty] Rename "possibly unbound" diagnostics to "possibly missing" (#20492) 2025-09-23 14:26:55 +00:00