[pylint] Implement invalid-length-returned (E0303)
#10963
Conversation
|
Try bumping the value in |
CodSpeed Performance ReportMerging #10963 will not alter performanceComparing Summary
|
tibor-reiss
force-pushed
the
add-E0303-invalid-length-returned
branch
2 times, most recently
from
ee331b8
to
57fab81
Compare
|
| code | total | + violation | - violation | + fix | - fix |
|---|---|---|---|---|---|
| PYI001 | 1 | 1 | 0 | 0 | 0 |
Linter (preview)
ℹ️ ecosystem check detected linter changes. (+4 -0 violations, +0 -0 fixes in 3 projects; 41 projects unchanged)
ibis-project/ibis (+1 -0 violations, +0 -0 fixes)
ruff check --no-cache --exit-zero --ignore RUF9 --output-format concise --preview
+ ibis/expr/types/relations.py:831:9: PLE0303 `__len__` does not return a non-negative integer
pandas-dev/pandas (+2 -0 violations, +0 -0 fixes)
ruff check --no-cache --exit-zero --ignore RUF9 --output-format concise --preview
+ pandas/core/arrays/base.py:483:9: PLE0303 `__len__` does not return a non-negative integer + pandas/core/base.py:349:9: PLE0303 `__len__` does not return a non-negative integer
python/typeshed (+1 -0 violations, +0 -0 fixes)
ruff check --no-cache --exit-zero --ignore RUF9 --output-format concise --preview --select E,F,FA,I,PYI,RUF,UP,W
+ stdlib/typing.pyi:330:10: PYI001 Name of private `TypeVar` must start with `_`
Changes by rule (2 rules affected)
| code | total | + violation | - violation | + fix | - fix |
|---|---|---|---|---|---|
| PLE0303 | 3 | 3 | 0 | 0 | 0 |
| PYI001 | 1 | 1 | 0 | 0 | 0 |
@charliermarsh Is the performance drop of 4% also related to this please? The newly introduced rule is quite lightweight. Solved, thanks @MichaReiser |
| @@ -97,6 +97,9 @@ pub(crate) fn statement(stmt: &Stmt, checker: &mut Checker) { | |||
| if checker.enabled(Rule::InvalidBoolReturnType) { | |||
| pylint::rules::invalid_bool_return(checker, name, body); | |||
| } | |||
| if checker.enabled(Rule::InvalidLengthReturnType) { | |||
| pylint::rules::invalid_length_return(checker, name, body); | |||
There was a problem hiding this comment.
Nit: You can pass the entire function definition instead of just passing the name and body
| pylint::rules::invalid_length_return(checker, name, body); | |
| pylint::rules::invalid_length_return(checker, function_def); |
There was a problem hiding this comment.
I took the current bool and str cases as a blueprint, which passed also name and body. What would be the advantage of passing the function_def please? I would still need to reference name and body.
| let Expr::UnaryOp(ast::ExprUnaryOp { op, .. }) = value else { | ||
| return false; | ||
| }; | ||
| matches!(op, ast::UnaryOp::USub) |
There was a problem hiding this comment.
Nit:
| let Expr::UnaryOp(ast::ExprUnaryOp { op, .. }) = value else { | |
| return false; | |
| }; | |
| matches!(op, ast::UnaryOp::USub) | |
| matches!( | |
| value, | |
| Expr::UnaryOp(ast::ExprUnaryOp { | |
| op: ast::UnaryOp::USub, | |
| .. | |
| }) | |
| ) |
| if !matches!( | ||
| ResolvedPythonType::from(value), | ||
| ResolvedPythonType::Unknown | ||
| | ResolvedPythonType::Atom(PythonType::Number(NumberLike::Integer)) | ||
| ) || is_negative_integer(value) | ||
| { |
There was a problem hiding this comment.
It could make sense to change the order of the checks because is_negative_integer is relatively simple compared to ResolvedPythonType::from
| if !matches!( | |
| ResolvedPythonType::from(value), | |
| ResolvedPythonType::Unknown | |
| | ResolvedPythonType::Atom(PythonType::Number(NumberLike::Integer)) | |
| ) || is_negative_integer(value) | |
| { | |
| if is_negative_integer(value) || | |
| !matches!( | |
| ResolvedPythonType::from(value), | |
| ResolvedPythonType::Unknown | |
| | ResolvedPythonType::Atom(PythonType::Number(NumberLike::Integer)) | |
| ) | |
| { |
There was a problem hiding this comment.
TLDR: changed.
I was not sure about which should be first because I would expect that there are more cases where the return type is wrong, compared to where the type is ok (int) but the sign not.
|
@tibor-reiss I bumped the index (and fixed the perf regression). You can rebase and we should then see the impact of your rule. |
tibor-reiss
force-pushed
the
add-E0303-invalid-length-returned
branch
from
57fab81
to
2b50cd0
Compare
charliermarsh
changed the title
pylint] Implement invalid-length-returned (E0303)
Add pylint rule invalid-length-returned (PLE0303)
See #970 for rules
Test Plan:
cargo testTBD: from the description: "Strictly speaking
boolis a subclass ofint, thus returningTrue/Falseis valid. To be consistent with other rules (e.g. PLE0305 invalid-index-returned), ruff will raise, compared to pylint which will not raise."