parameter specification/type variable tuple variance by KotlinIsland · Pull Request #2215 · python/typing

KotlinIsland · 2026-03-09T03:08:20Z

discussion: https://discuss.python.org/t/parameter-specifications-should-have-variance/106452

support in basedpyright variance for parameter specification and type variable tuples DetachHead/basedpyright#1771
support in pycharm: JetBrains/intellij-community@7196a70
cpython tvt variance: gh-148207: add additional keywords to typing.TypeVarTuple cpython#148212
typeshed: typing.TypeVarTuple: add bound/variance properties from 3.15 typeshed#15670
typing_extensions: Type variable tuple variance typing_extensions#741
typing-council: spec change: support variance on parameter specification and type variable tuple typing-council#59

JelleZijlstra

I like adding this if we can get it specified nicely, but this PR is not ready; the proposed test is incorrect.

Also, https://typing.python.org/en/latest/spec/generics.html#paramspec-variables still says variance on ParamSpec is unsupported; this should be updated.

I'd also like to see an implementation in at least one type checker, even if only as a draft PR, so we can be confident this is something that can be feasibly implemented.

davidhalter · 2026-04-02T22:01:54Z

I personally would also like to see tests added for param spec variance inference, since that would probably need to be supported as well.

JelleZijlstra · 2026-04-02T22:04:19Z

The test cases do have variance inference, though as I noted some of the cases are wrong. But it would probably be useful to have a few more cases, and I'd recommend putting the tests for paramspec variance in their own file so we can track type checker support more precisely.

KotlinIsland · 2026-04-03T04:00:17Z

I'd also like to see an implementation in at least one type checker

I have wip support in PyCharm, but I could also add it to basedpyright

it was very straightforward to implement

davidhalter · 2026-04-03T19:15:41Z

The test cases do have variance inference, though as I noted some of the cases are wrong. But it would probably be useful to have a few more cases, and I'd recommend putting the tests for paramspec variance in their own file so we can track type checker support more precisely.

Oh right, sorry I didn't see them, because I thought they would be in a different file. I'm also very much +1 on putting those tests into a different file (for example generics_paramspec_variance.py).

KotlinIsland · 2026-04-23T03:05:17Z

support has landed in pycharm and cpython

JelleZijlstra · 2026-04-23T03:18:25Z

https://typing.python.org/en/latest/spec/generics.html#paramspec-variables stills says that we don't support variance in ParamSpec, that should be fixed in this PR.

JelleZijlstra · 2026-04-23T03:18:43Z

Also can you open an issue on python/typing-council asking for a formal pronouncement?

KotlinIsland · 2026-04-23T03:47:56Z

python/typing-council#59

JelleZijlstra · 2026-04-23T04:06:01Z

Noticed some more issues:

generics_typevartuple_basic.py still has a line Ts1 = TypeVarTuple("Ts1", covariant=True) # E. But of course that's legal now. I'd delete the line and leave all variance testing to the new file generics_typevartuple_variance.py.
The description of the variance inference algorithm says "In the upper specialized class, specialize the target type parameter with an object instance". But that doesn't make sense for ParamSpec and TypeVarTuple. I think TypeVarTuple should instead be specialized with *tuple[object, ...]. For ParamSpec perhaps Callable[..., object] is right but not 100% sure.
The way the tests are written produces some incidental errors, like from ty generics_paramspec_variance.py:28:23: error[empty-body] Function always implicitly returns `None`, which is not assignable to return type `(...) -> Unknown` and mypy generics_typevartuple_variance.py:18: error: Missing return statement [empty-body]. Should add something like raise NotImplementedError.

KotlinIsland · 2026-04-23T05:07:56Z

Callable[..., object] is right but not 100% sure.

not Callable, and not ..., as this is the gradual form. but (*object, **object), which is a non-denotable psuedo representation of the widest value for a parameter specification

JelleZijlstra · 2026-04-23T14:53:28Z

I don't think that's right. The form we want here should be a supertype of every other possible value, and something like a one-parameter callable is not a subtype of (*: object, **: object).

AlexWaygood · 2026-04-23T15:44:46Z

In ty we have a type that we spell as Top[Callable[..., object]] in ty's type display (a fully static type describing the infinite union of all possible Callable types). It's not representable using standard Python notation currently.

JelleZijlstra · 2026-04-23T15:46:24Z

We only need assignability not subtyping here, so I think ... is enough.

carljm

Thanks! Definitely support this change.

carljm · 2026-04-26T20:25:24Z

-# > TypeVarTuple does not yet support specification of variance, bounds, constraints.
+# > TypeVarTuple does not yet support specification of bounds, constraints.

-Ts1 = TypeVarTuple("Ts1", covariant=True)  # E


This is is used by Array3 below, so line 107 is now failing for the wrong reason (name error) in all type checkers.

It might be better for the Array3 test to define its own TypeVarTuples right above that test, instead of relying on these TypeVarTuples doing double duty.

carljm · 2026-04-26T20:27:25Z

+arguments in the declaration just as ``typing.TypeVar`` does, but for ``bound`` we
+will defer the standardization of the semantics of this option to a later PEP.


Suggested change

arguments in the declaration just as ``typing.TypeVar`` does, but for ``bound`` we

will defer the standardization of the semantics of this option to a later PEP.

arguments in the declaration just as ``typing.TypeVar`` does.

We defer the standardization of the semantics of the ``bound`` option to a later PEP.

carljm · 2026-04-26T20:28:31Z

+``TypeVar``/``TypeVarTuple``/``ParamSpec`` declaration and is not specified
+as ``infer_variance`` (see below), its variance is specified by the


Suggested change

``TypeVar``/``TypeVarTuple``/``ParamSpec`` declaration and is not specified

as ``infer_variance`` (see below), its variance is specified by the

``TypeVar``/``TypeVarTuple``/``ParamSpec`` declaration and is not constructed

with ``infer_variance=True`` (see below), its variance is specified by the

carljm · 2026-04-26T20:35:54Z

+
+ an ``object`` instance for a type variable.
+ a ``*tuple[object, ...]`` value for a type variable tuple.
+ a ``...`` value for a parameter specification.


I don't think using ... here is sound. Because it is both assignable-to and assignable-from any paramspec, it means that a case like the ContravariantParamSpec test will wrongly infer the paramspec as covariant (because lower is assignable to upper) before it even checks the contravariant direction.

I don't think there is any way to make the variance-inference algorithm described here correct and support inference of ParamSpec variance, without introducing the new concept of a "top signature".

(I'm also not sure how much value there is in including this particular variance-inference algorithm in the spec. I think it describes pyright's algorithm. I know that ty and pyrefly do not use this algorithm. I don't know about mypy or zuban.)

carljm · 2026-04-26T20:42:52Z

@@ -0,0 +1,50 @@
+"""
+Tests variance of ParamSpec.


I think we should also have tests for invariant paramspecs (would have to be used more than once in the class)

carljm · 2026-04-26T20:44:17Z

@@ -0,0 +1,53 @@
+"""
+Tests variance of TypeVarTuple.


Here too, I think we should have tests for invariant TypeVarTuple.

carljm · 2026-04-26T20:45:48Z

We only need assignability not subtyping here, so I think ... is enough.

But we also need assignability to fail in the "wrong" direction, and this workaround fails that requirement.

KotlinIsland force-pushed the paramspec-variance branch 3 times, most recently from 5f8e0cc to d9d82c9 Compare March 9, 2026 03:13

srittau added the topic: typing spec For improving the typing spec label Mar 9, 2026

KotlinIsland force-pushed the paramspec-variance branch 2 times, most recently from 81ba400 to 54ad5e1 Compare April 1, 2026 07:59

JelleZijlstra requested changes Apr 2, 2026

View reviewed changes

Comment thread conformance/tests/generics_paramspec_semantics.py Outdated

Comment thread conformance/tests/generics_paramspec_semantics.py Outdated

KotlinIsland force-pushed the paramspec-variance branch 2 times, most recently from 8f30004 to 117964e Compare April 7, 2026 01:29

KotlinIsland requested a review from JelleZijlstra April 7, 2026 04:53

KotlinIsland mentioned this pull request Apr 7, 2026

add variance support to TypeVarTuple python/cpython#148207

Closed

KotlinIsland changed the title ~~parameter specification variance~~ parameter specification/type variable tuple variance Apr 7, 2026

KotlinIsland force-pushed the paramspec-variance branch from 7a21592 to 5788125 Compare April 8, 2026 00:48

KotlinIsland mentioned this pull request Apr 23, 2026

spec change: support variance on parameter specification and type variable tuple python/typing-council#59

Open

5 tasks

KotlinIsland force-pushed the paramspec-variance branch from 5788125 to f23d8ae Compare April 23, 2026 03:47

specify that parameter specification should have variance

495b5c9

KotlinIsland force-pushed the paramspec-variance branch from f23d8ae to 414c697 Compare April 23, 2026 05:12

specify that type variable tuple should have variance

02c4628

KotlinIsland force-pushed the paramspec-variance branch from 414c697 to 02c4628 Compare April 23, 2026 05:17

KotlinIsland mentioned this pull request Apr 23, 2026

variance for parameter specification and type variable tuples DetachHead/basedpyright#1771

Draft

carljm self-requested a review April 23, 2026 23:35

fixup! specify that parameter specification should have variance

8ecdde0

carljm requested changes Apr 26, 2026

View reviewed changes

		arguments in the declaration just as ``typing.TypeVar`` does, but for ``bound`` we
		will defer the standardization of the semantics of this option to a later PEP.

		``TypeVar``/``TypeVarTuple``/``ParamSpec`` declaration and is not specified
		as ``infer_variance`` (see below), its variance is specified by the

Conversation

KotlinIsland commented Mar 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

JelleZijlstra left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

davidhalter commented Apr 2, 2026

Uh oh!

JelleZijlstra commented Apr 2, 2026

Uh oh!

KotlinIsland commented Apr 3, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

davidhalter commented Apr 3, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

KotlinIsland commented Apr 23, 2026

Uh oh!

JelleZijlstra commented Apr 23, 2026

Uh oh!

JelleZijlstra commented Apr 23, 2026

Uh oh!

KotlinIsland commented Apr 23, 2026

Uh oh!

JelleZijlstra commented Apr 23, 2026

Uh oh!

KotlinIsland commented Apr 23, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

JelleZijlstra commented Apr 23, 2026

Uh oh!

AlexWaygood commented Apr 23, 2026

Uh oh!

JelleZijlstra commented Apr 23, 2026

Uh oh!

carljm left a comment

Choose a reason for hiding this comment

Uh oh!

carljm Apr 26, 2026

Choose a reason for hiding this comment

Uh oh!

carljm Apr 26, 2026

Choose a reason for hiding this comment

Uh oh!

carljm Apr 26, 2026

Choose a reason for hiding this comment

Uh oh!

carljm Apr 26, 2026

Choose a reason for hiding this comment

Uh oh!

carljm Apr 26, 2026

Choose a reason for hiding this comment

Uh oh!

carljm Apr 26, 2026

Choose a reason for hiding this comment

Uh oh!

carljm commented Apr 26, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

KotlinIsland commented Mar 9, 2026 •

edited

Loading

KotlinIsland commented Apr 3, 2026 •

edited

Loading

davidhalter commented Apr 3, 2026 •

edited

Loading

KotlinIsland commented Apr 23, 2026 •

edited

Loading