python / python/typing

Spec/Conformance: subtyping callables with non-constant parameter mapping.

Open
#2,224 1 comment 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

topic: typing spec
Dominant language
Python
Stars
1.8k
Forks
302
Avg merge
23h
Merged PRs (30d)
8

Description

Consider the example below, which errors with all tested type checkers (mypy/pyright/ty/pyrefly):

from typing import Protocol

class Interval: ...

class Make(Protocol):
    def __call__(self, /, lower: float, upper: float) -> Interval: ...

def make_impl(
    string_or_lower: str | float | None = None,
    /,
    lower: float | None = None,
    upper: float | None = None,
) -> Interval: ...

def test() -> None:
    _f: Make = make_impl  # ❌️ type checkers error here.

I believe from a pure type theory POV, this assignment should be legal, because all legal arguments to Make are also legal arguments to make_impl. The spec phrases it in the same spirit:

A callable type B is assignable to a callable type A if the return type of B is assignable to the return type of A and the input signature of B accepts all possible combinations of arguments that the input signature of A accepts. All of the specific assignability rules described below derive from this general rule.

And I couldn't find anything else in https://typing.python.org/en/latest/spec/callables.html#assignability-rules-for-callables that would disallow this assignment.

It seems the type-checkers try to match the KEYWORD_OR_POSITIONAL parameters by name, which is incorrect. Make has 3 legal call signatures:

  1. Make(float, float)
  2. Make(float, upper=float)
  3. Make(lower=float, upper=float)

and all these 3 call signatures are supported by make_impl, but the parameter mapping is not constant:

  1. make_impl(float, float) -> {lower:string_or_lower, upper:lower}
  2. make_impl(float, upper=float) -> {lower:string_or_lower, upper:upper}
  3. make_impl(lower=float, upper=float) -> {lower:lower, upper:upper}

So either the spec should demand a constant parameter mapping, or this example should be added to the conformance tests.

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start with the callable assignability rules linked in the issue and reproduce the example with mypy, pyright, ty, and pyrefly. The issue names no repository files or conformance-test path; completion depends on deciding whether to clarify constant parameter mapping in the spec or add this case to conformance tests.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
documentation, testing-qa
Issue type
Documentation
Difficulty
5/5
Estimated time
Over a week
Activity status
Quiet
Clarity
Mostly clear
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.