[ty] fix comparisons and arithmetic with NewTypes of float#22105
Merged
oconnor663 merged 2 commits intomainfrom Jan 6, 2026
Merged
[ty] fix comparisons and arithmetic with NewTypes of float#22105oconnor663 merged 2 commits intomainfrom
NewTypes of float#22105oconnor663 merged 2 commits intomainfrom
Conversation
oconnor663
force-pushed
the
newtype_float_ops
branch
from
2f3e834 to
0aeee50
Compare
Diagnostic diff on typing conformance testsNo changes detected when running ty on typing conformance tests ✅ |
oconnor663
commented
Dec 20, 2025
oconnor663
commented
Dec 20, 2025
crates/ty_python_semantic/resources/mdtest/annotations/new_types.md
Outdated
Show resolved
Hide resolved
|
AlexWaygood
added
ty
oconnor663
commented
Dec 24, 2025
oconnor663
force-pushed
the
newtype_float_ops
branch
2 times, most recently
from
a2bb4c4 to
9cefff0
Compare
Contributor
Author
|
I've added support/tests for Edit: ChatGPT caught a real mistake below, which I think is responsible for the failing fuzz tests and the excessive new ecosystem errors. Will update. |
oconnor663
requested review from
AlexWaygood,
carljm,
dcreager and
sharkdp
as code owners
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
oconnor663
force-pushed
the
newtype_float_ops
branch
from
9cefff0 to
95cbec2
Compare
|
| Lint rule | Added | Removed | Changed |
|---|---|---|---|
invalid-await |
0 | 2 | 6 |
invalid-return-type |
0 | 1 | 5 |
possibly-missing-attribute |
0 | 3 | 1 |
unresolved-attribute |
0 | 0 | 2 |
invalid-argument-type |
0 | 1 | 0 |
unused-ignore-comment |
1 | 0 | 0 |
| Total | 1 | 7 | 14 |
Contributor
Author
|
Most of the new diagnostics are gone, but there are still a few that look like they're probably my fault: |
Contributor
Author
|
Minimal repro for these failures, new in this PR: from typing import NewType
Key = NewType("Key", object)
class ThingWithKeys:
def __contains__(self, refname: Key) -> bool:
return False
def _(key: Key, keys: ThingWithKeys):
key in keys # error[unsupported-operator]: Unsupported `in` operation |
oconnor663
commented
Dec 31, 2025
Contributor
Author
|
Ok all of the added and removed diagnostics are gone. (Is there really nobody in the ecosystem report trying to do arithmetic on newtypes of float?) What remains is just unions that get printed differently. Some of them look like random ordering flakes, but others are cases where |
oconnor663
force-pushed
the
newtype_float_ops
branch
from
f3d440a to
94d22ea
Compare
Contributor
Author
Gankra
approved these changes
Jan 6, 2026
Contributor
There was a problem hiding this comment.
This all makes sense on paper. This special-case is only triggered by float/complex because NewType otherwise doesn't permit taking a union as an argument, right? Are there other cases where you can smuggle unions that we might not be checking for though? Type aliases (of various flavours)?
Comment on lines
324
to
330
| reveal_type(Foo(3.14) + Bing()) # revealed: Foo | ||
| reveal_type(Bing() + Foo(42)) # revealed: Foo | ||
| reveal_type(Foo(3.14) < Bing()) # revealed: bool | ||
| reveal_type(Bing() < Foo(42)) # revealed: bool | ||
| reveal_type(Foo(3.14) in Bing()) # revealed: bool | ||
| ``` | ||
|
|
Contributor
There was a problem hiding this comment.
Can you include the negative cases like you had for int? Those seemed important/valuable.
Contributor
Author
Right. That restriction is implemented here. Apart from the |
MichaReiser
pushed a commit
that referenced
this pull request
Jan 6, 2026
carljm
added a commit
that referenced
this pull request
Jan 13, 2026
…king When calling a method on an instance of a generic class with bounded type parameters (e.g., `C[T: K]` where `K` is a NewType), ty was incorrectly reporting: "Argument type `C[K]` does not satisfy upper bound `C[T@C]` of type variable `Self`" The issue was introduced by PR #22105, which added a catch-all case for NewType assignments that falls back to the concrete base type. This case was placed before the TypeVar handling cases, so when checking `K <: T@C` (where K is a NewType and T@C is a TypeVar with upper bound K): 1. The NewType fallback matched first 2. It delegated to `int` (K's concrete base type) 3. Then checked `int <: T@C`, which checks if `int` satisfies bound `K` 4. But `int` is not assignable to `K` (NewTypes are distinct from their bases) The fix moves the NewType fallback case after the TypeVar cases, so TypeVar handling takes precedence. Now when checking `K <: T@C`, we use the TypeVar case at line 828 which returns `false` for non-inferable typevars - but this is correct because the *other* direction (`T@C <: K`) passes, and for the overall specialization comparison both directions are checked. Fixes astral-sh/ty#2467 Co-Authored-By: Carl Meyer <[email protected]>
carljm
added a commit
that referenced
this pull request
Jan 13, 2026
…king When calling a method on an instance of a generic class with bounded type parameters (e.g., `C[T: K]` where `K` is a NewType), ty was incorrectly reporting: "Argument type `C[K]` does not satisfy upper bound `C[T@C]` of type variable `Self`" The issue was introduced by PR #22105, which added a catch-all case for NewType assignments that falls back to the concrete base type. This case was placed before the TypeVar handling cases, so when checking `K <: T@C` (where K is a NewType and T@C is a TypeVar with upper bound K): 1. The NewType fallback matched first 2. It delegated to `int` (K's concrete base type) 3. Then checked `int <: T@C`, which checks if `int` satisfies bound `K` 4. But `int` is not assignable to `K` (NewTypes are distinct from their bases) The fix moves the NewType fallback case after the TypeVar cases, so TypeVar handling takes precedence. Now when checking `K <: T@C`, we use the TypeVar case at line 828 which returns `false` for non-inferable typevars - but this is correct because the *other* direction (`T@C <: K`) passes, and for the overall specialization comparison both directions are checked. Fixes astral-sh/ty#2467 Co-Authored-By: Carl Meyer <[email protected]>
carljm
added a commit
that referenced
this pull request
Jan 13, 2026
…king When calling a method on an instance of a generic class with bounded type parameters (e.g., `C[T: K]` where `K` is a NewType), ty was incorrectly reporting: "Argument type `C[K]` does not satisfy upper bound `C[T@C]` of type variable `Self`" The issue was introduced by PR #22105, which moved the catch-all case for NewType assignments that falls back to the concrete base type. This case was placed before the TypeVar handling cases, so when checking `K <: T@C` (where K is a NewType and T@C is a TypeVar with upper bound K): 1. The NewType fallback matched first 2. It delegated to `int` (K's concrete base type) 3. Then checked `int <: T@C`, which checks if `int` satisfies bound `K` 4. But `int` is not assignable to `K` (NewTypes are distinct from their bases) The fix moves the NewType fallback case after the TypeVar cases, so TypeVar handling takes precedence. Fixes astral-sh/ty#2467
carljm
added a commit
that referenced
this pull request
Jan 13, 2026
) Fixes astral-sh/ty#2467 When calling a method on an instance of a generic class with bounded type parameters (e.g., `C[T: K]` where `K` is a NewType), ty was incorrectly reporting: "Argument type `C[K]` does not satisfy upper bound `C[T@C]` of type variable `Self`" The issue was introduced by PR #22105, which moved the catch-all case for NewType assignments that falls back to the concrete base type. This case was moved before the TypeVar handling cases, so when checking `K <: T@C` (where K is a NewType and T@C is a TypeVar with upper bound K): 1. The NewType fallback matched first 2. It delegated to `int` (K's concrete base type) 3. Then checked `int <: T@C`, which checks if `int` satisfies bound `K` 4. But `int` is not assignable to `K` (NewTypes are distinct from their bases) The fix moves the NewType fallback case after the TypeVar cases, so TypeVar handling takes precedence. Now when checking `K <: T@C`, we use the TypeVar case at line 828 which returns `false` for non-inferable typevars - but this is correct because the *other* direction (`T@C <: K`) passes, and for the overall specialization comparison both directions are checked.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Fixes astral-sh/ty#2077.