-
-
Notifications
You must be signed in to change notification settings - Fork 2.8k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Disallow no-args generic aliases when using PEP 613 explicit aliases #18173
base: master
Are you sure you want to change the base?
Disallow no-args generic aliases when using PEP 613 explicit aliases #18173
Conversation
Conformance results diff for this PR ( diff --git a/conformance/results/mypy/aliases_explicit.toml b/conformance/results/mypy/aliases_explicit.toml
index ee05762..f07e7ad 100644
--- a/conformance/results/mypy/aliases_explicit.toml
+++ b/conformance/results/mypy/aliases_explicit.toml
@@ -24,11 +24,11 @@ aliases_explicit.py:89: error: Invalid type: try using Literal[1] instead? [val
aliases_explicit.py:90: error: Invalid type alias: expression is not a valid type [valid-type]
aliases_explicit.py:90: error: Function "list" could always be true in boolean context [truthy-function]
aliases_explicit.py:91: error: Invalid type alias: expression is not a valid type [valid-type]
+aliases_explicit.py:100: error: Bad number of arguments for type alias, expected 0, given 1 [type-arg]
aliases_explicit.py:101: error: "UnionType[list[Any], set[Any]]" not callable [operator]
aliases_explicit.py:102: error: Bad number of arguments for type alias, expected 0, given 1 [type-arg]
"""
-conformance_automated = "Fail"
+conformance_automated = "Pass"
errors_diff = """
-Line 100: Expected 1 errors
"""
ignore_errors = ["Function \"list\" could always be true in boolean context"]
diff --git a/conformance/results/mypy/version.toml b/conformance/results/mypy/version.toml
index 699a899..1d9b711 100644
--- a/conformance/results/mypy/version.toml
+++ b/conformance/results/mypy/version.toml
@@ -1,2 +1,2 @@
-version = "mypy 1.14.0+dev.21587f01045246a9ecb54a054a5ba03e20cbbd19"
-test_duration = 4.5
+version = "mypy 1.14.0+dev.6563c362fd5ee8e1fbf17775eab5dcad5a0a7459"
+test_duration = 5.1 |
This comment has been minimized.
This comment has been minimized.
This is expected, now that the below branch is no longer taken for PEP 613 aliases. Implicit aliases are now the only case with eager expansion. https://github.com/python/mypy/blob/21587f01045246a9ecb54a054a5ba03e20cbbd19/mypy/typeanal.py#L2147-L2151
This comment has been minimized.
This comment has been minimized.
The psycopg hits are true positives from aliases of the form |
from typing_extensions import TypeAlias | ||
|
||
Implicit = List | ||
Explicit: TypeAlias = List |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Also test with non-alias generic class (e.g. <alias> = list
)? You could also test with tuple
, since it's a bit special.
Diff from mypy_primer, showing the effect of this PR on open source code: psycopg (https://github.com/psycopg/psycopg)
+ psycopg/psycopg/_acompat.py:43: error: Bad number of arguments for type alias, expected 0, given 1 [type-arg]
+ psycopg/psycopg/_acompat.py:52: error: Returning Any from function declared to return "T" [no-any-return]
+ psycopg/psycopg/_acompat.py:55: error: Bad number of arguments for type alias, expected 0, given 1 [type-arg]
+ psycopg_pool/psycopg_pool/_acompat.py:63: error: Bad number of arguments for type alias, expected 0, given 1 [type-arg]
+ psycopg_pool/psycopg_pool/_acompat.py:72: error: Returning Any from function declared to return "T" [no-any-return]
+ psycopg_pool/psycopg_pool/_acompat.py:105: error: Bad number of arguments for type alias, expected 0, given 1 [type-arg]
|
Per the type system conformance tests, this is ok:
While this is not:
Mypy currently permits both. This PR makes mypy reject the latter case, improving conformance.
As part of this, no-args PEP 613 explicit aliases are no longer eagerly expanded.
(Also removed a stale comment referencing
TypeAliasExpr.no_args
, which was removed in #15924)