[ty] Infer lambda return types from body expressions#23328
Closed
alex wants to merge 1 commit intoastral-sh:mainfrom
Closed
[ty] Infer lambda return types from body expressions#23328alex wants to merge 1 commit intoastral-sh:mainfrom
alex wants to merge 1 commit intoastral-sh:mainfrom
Conversation
Infer the return type of lambda expressions from their body expression, rather than always using `Unknown`. This can introduce new Salsa cycles when a lambda body references a name involved in a circular definition. Added cycle recovery guards to prevent panics from non-converging cycles, though I'm not confident this is the correct solution. https://claude.ai/code/session_014XYWYmMYnqZdVjBQCs5XCd
Typing conformance resultsNo changes detected ✅ |
|
Memory usage reportSummary
Significant changesClick to expand detailed breakdownprefect
sphinx
trio
flake8
|
alex
requested review from
AlexWaygood,
carljm,
dcreager,
ibraheemdev and
sharkdp
as code owners
|
| Lint rule | Added | Removed | Changed |
|---|---|---|---|
invalid-argument-type |
17 | 16 | 118 |
invalid-assignment |
10 | 0 | 42 |
unresolved-attribute |
7 | 0 | 23 |
unused-type-ignore-comment |
1 | 19 | 0 |
unsupported-operator |
7 | 0 | 4 |
type-assertion-failure |
0 | 5 | 1 |
no-matching-overload |
4 | 0 | 0 |
not-subscriptable |
4 | 0 | 0 |
invalid-return-type |
2 | 0 | 1 |
invalid-await |
1 | 0 | 0 |
not-iterable |
0 | 0 | 1 |
| Total | 53 | 40 | 190 |
Contributor
Author
|
Important note, as indicated in a comment, this is kind of the bare minimum to infer types for lambdas where the return type isn't specific to the callsite/argument types. I'm hoping that since the volume of code is so minimal that it's an ok stepping stone. |
Member
|
you may have seen the previous attempt in #22633 -- that one was attempting something slightly more ambitious with full type context, but it might be helpful for you |
Contributor
Author
|
Yeah, the test case from that is what led to discovering the cyclic issues. Given my lack of certainty about the correctness of those fixes, I figured I'd start with the most basic implementation. |
Member
|
I'm not sure this usefully narrows scope compared to #22633. The benefit of inferring with type context is that we avoid false positives on e.g., I also don't think introducing cycle fallback based on an iteration threshold is a good idea. Based on the dicussion in #22818, we have a good idea of what the issue is, and I would rather delay this feature and work on a concrete solution than introduce fallbacks in arbitrary queries. Our recursive type inference should be robust to all of these issues. |
Contributor
Author
|
Ok! More than happy to defer to that.
All that is necessary for evil to succeed is for good people to do nothing.
…On Thu, Feb 19, 2026, 8:23 AM Ibraheem Ahmed ***@***.***> wrote:
*ibraheemdev* left a comment (astral-sh/ruff#23328)
<#23328 (comment)>
I'm not sure this usefully narrows scope compared to #22633
<#22633>. The benefit of inferring
with type context is that we avoid false positives on e.g., _:
Callable[[], MyTypedDict] = lambda: { "foo": 1 }, and extracting the type
context from the callable annotation is relatively trivial.
I also don't think introducing cycle fallback based on an iteration
threshold is a good idea. Based on the dicussion in #22818
<#22818>, we have a good idea of
what the issue is, and I would rather delay this feature and work on a
concrete solution than introduce fallbacks in arbitrary queries. Our
recursive type inference should be robust to all of these issues.
—
Reply to this email directly, view it on GitHub
<#23328 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAAAGBFE2N3QJ2Y2DEVUGE34MW2OXAVCNFSM6AAAAACVHLGOYCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTSMRXGI4TIMZSGU>
.
You are receiving this because you authored the thread.Message ID:
***@***.***>
|
Member
|
Closing this for now. |
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
4 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.
Infer the return type of lambda expressions from their body expression, rather than always using
Unknown.This can introduce new Salsa cycles when a lambda body references a name involved in a circular definition. Added cycle recovery guards to prevent panics from non-converging cycles, though I'm not confident this is the correct solution.