[red-knot] Initial tests for protocols #17436
Conversation
|
|
we really need to figure out why we aren't getting any red-knot tests run on PRs that only touch MarkDown files right now 😄 but I ran the new tests locally and I promise that they all pass! |
| satisfy the interface declared by `X`. But if `Y` is not `@final`, then this does not hold true, since a subclass of `Y` | ||
| could always provide additional methods or attributes that would lead to it satisfying `X`'s interface: |
There was a problem hiding this comment.
If Y does provide methods or attributes that are part of Xs interface, and the types of those are such that no subtype of Y could (respecting Liskov) possibly satisfy Xs interface, we can still consider them disjoint.
There was a problem hiding this comment.
Indeed! This is something I wanted to discuss; thanks for bringing it up. I seem to remember us having this conversation about nominal types on one of @sharkdp's PRs (I can't easily find the reference), and IIRC you were opposed to applying the same principle there. E.g. we could consider A and B here disjoint, because they both have a covariant (read-only) x attribute, but the type of A.x is disjoint from B.x. There can be no possible Liskov-compliant subclass of both A and B here; they are definitely disjoint:
class A:
@property
def x(self) -> bool:
return True
class B:
@property
def x(self) -> str:
return "x"Is it justifiable to apply this principle to structural types but not nominal types? It feels inconsistent to me
There was a problem hiding this comment.
(Going to land the PR for now as I don't think my current language precludes the idea that we might infer other pairs of protocols as disjoint in the future. Happy to revisit it later!)
There was a problem hiding this comment.
I do think there might be reasons to consider it for structural types and not nominal types, but I also think it's not high priority and fine to defer it either way, so let's just not worry about it for now.
AlexWaygood
force-pushed
the
alex/protocol-tests
branch
from
683c2d6 to
5e059ee
Compare
* main: [red-knot] Detect version-related syntax errors (#16379) [`pyflakes`] Add fix safety section (`F841`) (#17410) [red-knot] Add `KnownFunction` variants for `is_protocol`, `get_protocol_members` and `runtime_checkable` (#17450) Bump 0.11.6 (#17449) Auto generate `visit_source_order` (#17180) [red-knot] Initial tests for protocols (#17436) [red-knot] Dataclasses: synthesize `__init__` with proper signature (#17428) [red-knot] Dataclasses: support `order=True` (#17406)
* main: (123 commits) [red-knot] Handle explicit class specialization in type expressions (#17434) [red-knot] allow assignment expression in call compare narrowing (#17461) [red-knot] fix building unions with literals and AlwaysTruthy/AlwaysFalsy (#17451) [red-knot] Type narrowing for assertions (take 2) (#17345) [red-knot] class bases are not affected by __future__.annotations (#17456) [red-knot] Add support for overloaded functions (#17366) [`pyupgrade`] Add fix safety section to docs (`UP036`) (#17444) [red-knot] more type-narrowing in match statements (#17302) [red-knot] Add some narrowing for assignment expressions (#17448) [red-knot] Understand `typing.Protocol` and `typing_extensions.Protocol` as equivalent (#17446) Server: Use `min` instead of `max` to limit the number of threads (#17421) [red-knot] Detect version-related syntax errors (#16379) [`pyflakes`] Add fix safety section (`F841`) (#17410) [red-knot] Add `KnownFunction` variants for `is_protocol`, `get_protocol_members` and `runtime_checkable` (#17450) Bump 0.11.6 (#17449) Auto generate `visit_source_order` (#17180) [red-knot] Initial tests for protocols (#17436) [red-knot] Dataclasses: synthesize `__init__` with proper signature (#17428) [red-knot] Dataclasses: support `order=True` (#17406) [red-knot] Super-basic generic inference at call sites (#17301) ...
Summary
There are still many tests missing from this test suite (see the long TODO list at the bottom of this file). But this PR already adds around 1,000 lines of tests, which feels like enough for now: I'd like to actually start implementing some of this stuff 😄
TODO: figure out how to make pre-commit happy...
Test Plan
this pr is all tests