Fixes#8721
## Summary
This implements the rule proposed in #8721, as RUF021. `and` always
binds more tightly than `or` when chaining the two together.
(This should definitely be autofixable, but I'm leaving that to a
followup PR for now.)
## Test Plan
`cargo test` / `cargo insta review`
## Summary
On `main`, we flag redefinitions in cases like:
```python
import os
x = 1
if x > 0:
import os
```
That is, we consider these to be in the "same branch", since they're not
in disjoint branches. This matches Flake8's behavior, but it seems to
lead to false positives.
## Summary
Given a docstring like:
```python
def func(x: int, args: tuple[int]):
"""Toggle the gizmo.
Args:
x: Some argument.
args: Some other arguments.
"""
```
We were considering the `args:` descriptor to be an indented docstring
section header (since `Args:`) is a valid header name. This led to very
confusing diagnostics.
This PR makes the parsing a bit more lax in this case, such that if we
see a nested header that's more deeply indented than the preceding
header, and the preceding section allows sub-items (like `Args:`), we
avoid treating the nested item as a section header.
Closes https://github.com/astral-sh/ruff/issues/9426.
I just fixed this false negative in flake8-pyi
(https://github.com/PyCQA/flake8-pyi/pull/460), and then realised ruff
has the exact same bug! Luckily it's a very easy fix.
(The bug is that unused protocols go undetected if they're generic.)
## Summary
Ensures that any lint rules that include line locations render them as
relative to the cell (and include the cell number) when inside a Jupyter
notebook.
Closes https://github.com/astral-sh/ruff/issues/6672.
## Test Plan
`cargo test`
## Summary
This PR adds an autofix for the newly added PYI058 rule (added in
#9313). ~~The PR's current implementation is that the fix is only
available if the fully qualified name of `Generator` or `AsyncGenerator`
is being used:~~
- ~~`-> typing.Generator` is converted to `-> typing.Iterator`;~~
- ~~`-> collections.abc.AsyncGenerator[str, Any]` is converted to `->
collections.abc.AsyncIterator[str]`;~~
- ~~but `-> Generator` is _not_ converted to `-> Iterator`. (It would
require more work to figure out if `Iterator` was already imported or
not. And if it wasn't, where should we import it from? `typing`,
`typing_extensions`, or `collections.abc`? It seems much more
complicated.)~~
The fix is marked as always safe for `__iter__` or `__aiter__` methods
in `.pyi` files, but unsafe for all such methods in `.py` files that
have more than one statement in the method body.
This felt slightly fiddly to accomplish, but I couldn't _see_ any
utilities in
https://github.com/astral-sh/ruff/tree/main/crates/ruff_linter/src/fix
that would have made it simpler to implement. Lmk if I'm missing
something, though -- my first time implementing an autofix! :)
## Test Plan
`cargo test` / `cargo insta review`.
## Summary
We had an early `continue` in this loop, and we weren't setting
`comparator = next;` when continuing... This PR removes the early
continue altogether for clarity.
Closes https://github.com/astral-sh/ruff/issues/9374.
## Test Plan
`cargo test`
## Summary
This PR implements Y058 from flake8-pyi -- this is a new flake8-pyi rule
that was released as part of `flake8-pyi 23.11.0`. I've followed the
Python implementation as closely as possible (see
858c0918a8),
except that for the Ruff version, the rule also applies to `.py` files
as well as for `.pyi` files. (For `.py` files, we only emit the
diagnostic in very specific situations, however, as there's a much
higher likelihood of emitting false positives when applying this rule to
a `.py` file.)
## Test Plan
`cargo test`/`cargo insta review`
## Summary
Historically, we encoded this list by extracting the `__all__`. I went
to update it, but... is there really any value in it? Seems easier to
just treat `typing_extensions` as an alias for `typing`.
Closes https://github.com/astral-sh/ruff/issues/9334.
## Summary
The logic that detects continuations assumed that tokens themselves
cannot span multiple lines. However, strings _can_ -- even single-quoted
strings.
Closes https://github.com/astral-sh/ruff/issues/9323.
## Summary
This PR modifies the semantics of `runtime-evaluated-decorators` to
respect decorators on both classes _and_ functions. Historically, this
only respected classes, since the common use-case is (e.g.)
`pydantic.BaseModel` -- but functions are equally valid.
Closes https://github.com/astral-sh/ruff/issues/9312.
## Test Plan
`cargo test`
## Summary
Given:
```python
from somewhere import get_cfg
def lookup_cfg(cfg_description):
cfg = get_cfg(cfg_description)
if cfg is not None:
return cfg
raise AttributeError(f"No cfg found matching {cfg_description}")
```
We were analyzing the method from last-to-first statement. So we saw the
`raise`, then assumed the method _always_ raised. In reality, though, it
_might_ return. This PR improves the branch analysis to respect these
mixed cases.
Closes https://github.com/astral-sh/ruff/issues/9269.
Closes https://github.com/astral-sh/ruff/issues/9304.
## Summary
Part of #970.
This adds Pylint's [R0244
empty_comment](https://pylint.pycqa.org/en/latest/user_guide/messages/refactor/empty-comment.html)
lint as well as an always-safe fix.
## Test Plan
The included snapshot verifies the following:
- A line containing only a non-empty comment is not changed
- A line containing leading whitespace before a non-empty comment is not
changed
- A line containing only an empty comment has its content deleted
- A line containing only leading whitespace before an empty comment has
its content deleted
- A line containing only leading and trailing whitespace on an empty
comment has its content deleted
- A line containing trailing whitespace after a non-empty comment is not
changed
- A line containing only a single newline character (i.e. a blank line)
is not changed
- A line containing code followed by a non-empty comment is not changed
- A line containing code followed by an empty comment has its content
deleted after the last non-whitespace character
- Lines containing code and no comments are not changed
- Empty comment lines within block comments are ignored
- Empty comments within triple-quoted sections are ignored
## Comparison to Pylint
Running Ruff and Pylint 3.0.3 with Python 3.12.0 against the
`empty_comment.py` file added in this PR, we see the following:
* Identical behavior:
* empty_comment.py:3:0: R2044: Line with empty comment (empty-comment)
* empty_comment.py:4:0: R2044: Line with empty comment (empty-comment)
* empty_comment.py:5:0: R2044: Line with empty comment (empty-comment)
* empty_comment.py:18:0: R2044: Line with empty comment (empty-comment)
* Differing behavior:
* Pylint doesn't ignore empty comments in block comments commonly used
for visual spacing; I decided these were fine in this implementation
since many projects use these and likely do not want them removed.
* empty_comment.py:28:0: R2044: Line with empty comment (empty-comment)
* Pylint detects "empty comments" within the triple-quoted section at
the bottom of the file, which is arguably a bug in the Pylint
implementation since these are not truly comments. These are ignored by
this implementation.
* empty_comment.py:37:0: R2044: Line with empty comment (empty-comment)
* empty_comment.py:38:0: R2044: Line with empty comment (empty-comment)
* empty_comment.py:39:0: R2044: Line with empty comment (empty-comment)
## Summary
Hey there 👋 thanks for this great project!
On python code looking like the following
```
import yaml
from yaml.loader import SafeLoader
with MY_FILE_PATH.open("r") as my_file:
my_data = yaml.load(my_file, Loader=SafeLoader)
```
ruff reports this error:
```
S506 Probable use of unsafe loader `SafeLoader` with `yaml.load`. Allows instantiation of arbitrary objects. Consider `yaml.safe_load`.
```
This PR is an attempt to support SafeLoader being imported for either
`yaml` or `yaml.loader`
Disclaimer:
I am not familiar with Rust so this is likely not the better way of
doing it. Interested in hearing how to adapt this PR to provide similar
behavior in a better way
## Test Plan
The S506.py file was updated accordingly to cover the use cases and test
were confirmed to pass with this change.
## Summary
If `RUF100` is ignored via `per-file-ignores`, we need to avoid raising
it. `RUF100` has special "self-ignore" logic, since the rule itself
deals with `# noqa` directives. This PR wires up `per-file-ignores` to
that "self-ignore" logic.
Closes https://github.com/astral-sh/ruff/issues/9297.
We should avoid adding `-> None` to stubs in `.pyi` files, along with a
few other cases. (We already ignore abstract methods.)
Closes https://github.com/astral-sh/ruff/issues/9270.
## Summary
Adds a rule to detect unions that include `typing.NoReturn` or
`typing.Never`. In such cases, the use of the bottom type is redundant.
Closes https://github.com/astral-sh/ruff/issues/9113.
## Test Plan
`cargo test`
## Summary
fixes#6956
details in issue
Following an advice in
https://github.com/astral-sh/ruff/issues/6956#issuecomment-1817672585,
this change separates expressions to 3 levels of "constant likelihood":
* literals, empty dict and tuples... (definitely constant, level 2)
* CONSTANT_CASE vars (probably constant, 1)
* all other expressions (0)
a comparison is marked yoda if the level is strictly higher on its left
hand side
following
https://github.com/astral-sh/ruff/issues/6956#issuecomment-1697107822
marking compound expressions of literals (e.g. `60 * 60` ) as constants
this change current behaviour on
`SomeClass().settings.SOME_CONSTANT_VALUE > (60 * 60)` in the fixture
from error to ok
## Summary
Given a function like:
```python
def func(x: int):
if not x:
raise ValueError
else:
raise TypeError
```
We now correctly use `NoReturn` as the return type, rather than `None`.
Closes https://github.com/astral-sh/ruff/issues/9201.
## Summary
Part of #8771. flake8-pyi will emit a Y018 error for unused TypeVars,
ParamSpecs or TypeVarTuples; Ruff currently only emits PYI018 for unused
TypeVars.
This is my first "proper" Ruff PR -- let me know if there's a better way
of doing this! Not sure if the repeated calls to `match_typing_expr()`
are ideal.
## Test Plan
I manually updated the fixtures to add some unused ParamSpecs and
TypeVarTuples, and then updated the snapshots using `cargo insta
review`. All tests then passed when run using `cargo test`.
## Summary
Given `Callable[[Callable[_P, _R]], Callable[_P, _R]]` from the
originating issue, when quoting `Callable`, we quoted the inner
`[Callable[_P, _R]]`, and then created a separate edit for the outer
`Callable`. Since there's an extra level of nesting in the subscript,
the edit for `[Callable[_P, _R]]` correctly did _not_ expand to the
entire expression. However, in this case, we should discard the inner
edit, since the expression is getting quoted by the outer edit anyway.
Closes https://github.com/astral-sh/ruff/issues/9162.
## Summary
A common mistake is to add quotes around one member in an `X | Y`-style
type union, as in:
```python
contract_versions_list: list[ContractVersion] | 'QuerySet[ContractVersion]' | None = None
```
However, doing so will lead to a runtime error if the annotation is
runtime-evaluated. This PR lints against such patterns.
Closes https://github.com/astral-sh/ruff/issues/9139.
## Summary
Fix dropped union expressions for piped non-types in `PYI055` autofix
If you had `type[int] | type[str] | str`, it would have dropped the
`str`, which breaks the type!
Closes#9156
## Test Plan
`cargo test`
Fix#9080
Example, where `[]` is a 2 byte non-breaking space:
```
def f():
""" Docstring header
^^^^ Real indentation is 4 chars
docstring body, over-indented
^^^^^^ Over-indentation is 6 - 4 = 2 chars due to this line
[] [] docstring body 2, further indented
^^^^^ We take these 4 chars/5 bytes to match the docstring ...
^^^ ... and these 2 chars/3 bytes to remove the `over_indented_size` ...
^^ ... but preserve this real indent
```
Given:
```python
x: DataFrame[
int
] = 1
```
We currently wrap the annotation in single quotes, which leads to a
syntax error:
```python
x: "DataFrame[
int
]" = 1
```
There are a few options for what to suggest for users here... Use triple
quotes:
```python
x: """DataFrame[
int
]""" = 1
```
Or, use an implicit string concatenation (which may require
parentheses):
```python
x: ("DataFrame["
"int"
"]") = 1
```
The solution I settled on here is to use the `Generator`, which
effectively means we write it out on a single line, like:
```python
x: "DataFrame[int]" = 1
```
It's kind of the "least opinionated" solution, but it does mean we'll
expand to a very long line in some cases.
Closes https://github.com/astral-sh/ruff/issues/9136.
## Summary
This allows us to fix usages like:
```python
from pandas import DataFrame
def baz() -> DataFrame:
...
```
By quoting the `DataFrame` in `-> DataFrame`. Without quotes, moving
`from pandas import DataFrame` into an `if TYPE_CHECKING:` block will
fail at runtime, since Python tries to evaluate the annotation to add it
to the function's `__annotations__`.
Unfortunately, this does require us to split our "annotation kind" flags
into three categories, rather than two:
- `typing-only`: The annotation is only evaluated at type-checking-time.
- `runtime-evaluated`: Python will evaluate the annotation at runtime
(like above) -- but we're willing to quote it.
- `runtime-required`: Python will evaluate the annotation at runtime
(like above), and some library (like Pydantic) needs it to be available
at runtime, so we _can't_ quote it.
This functionality is gated behind a setting
(`flake8-type-checking.quote-annotations`).
Closes https://github.com/astral-sh/ruff/issues/5559.
## Summary
Adds `find_assigned_value` a function which gets the `&Expr` assigned to
a given `id` if one exists in the semantic model.
Open TODOs:
- [ ] Handle `binding.kind.is_unpacked_assignment()`: I am bit confused
by this one. The snippet from its documentation does not appear to be
counted as an unpacked assignment and the only ones I could find for
which that was true were invalid Python like:
```python
x, y = 1
```
- [ ] How to handle AugAssign. Can we combine statements like:
```python
(a, b) = [(1, 2, 3), (4,)]
a += (6, 7)
```
to get the full value for a? Code currently just returns `None` for
these assign types
- [ ] Multi target assigns
```python
m_c = (m_d, m_e) = (0, 0)
trio.sleep(m_c) # OK
trio.sleep(m_d) # TRIO115
trio.sleep(m_e) # TRIO115
```
## Test Plan
Used the function in two rules:
- `TRIO115`
- `PERF101`
Expanded both their fixtures for explicit multi target check
<!--
Thank you for contributing to Ruff! To help us out with reviewing,
please consider the following:
- Does this pull request include a summary of the change? (See below.)
- Does this pull request include a descriptive title?
- Does this pull request include references to any relevant issues?
-->
## Summary
<!-- What's the purpose of the change? What does it do, and why? -->
Check floating-point numbers similarly to integers in FURB163. For
example, both `math.log(x, 10)` and `math.log(x, 10.0)` should be
changed to `math.log10(x)`.
## Test Plan
<!-- How was it tested? -->
Added couple of test cases.
## Summary
E274 currently flags any keyword at the start of a line indented with
tabs. This turns out to be due to a bug in `Whitespace::trailing` that
never considers any whitespace containing a tab as indentation.
## Test Plan
Added a simple test case.
This PR allows `matplotlib.use` calls to intersperse imports without
triggering `E402`. This is a pragmatic choice as it's common to require
`matplotlib.use` calls prior to importing from within `matplotlib`
itself.
Closes https://github.com/astral-sh/ruff/issues/9091.
## Summary
Fixes#8863 : Detect asyncio-dangling-task (RUF006) when discarding
return value
## Test Plan
added new two testcases, changed result of an old one that was made more
specific
## Summary
It's common to interleave a `sys.path` modification between imports at
the top of a file. This is a frequent cause of `# noqa: E402` false
positives, as seen in the ecosystem checks. This PR modifies E402 to
omit such modifications when determining the "import boundary".
(We could consider linting against `sys.path` modifications, but that
should be a separate rule.)
Closes: https://github.com/astral-sh/ruff/issues/5557.
Rebase of #6365 authored by @davidszotten.
## Summary
This PR updates the AST structure for an f-string elements.
The main **motivation** behind this change is to have a dedicated node
for the string part of an f-string. Previously, the existing
`ExprStringLiteral` node was used for this purpose which isn't exactly
correct. The `ExprStringLiteral` node should include the quotes as well
in the range but the f-string literal element doesn't include the quote
as it's a specific part within an f-string. For example,
```python
f"foo {x}"
# ^^^^
# This is the literal part of an f-string
```
The introduction of `FStringElement` enum is helpful which represent
either the literal part or the expression part of an f-string.
### Rule Updates
This means that there'll be two nodes representing a string depending on
the context. One for a normal string literal while the other is a string
literal within an f-string. The AST checker is updated to accommodate
this change. The rules which work on string literal are updated to check
on the literal part of f-string as well.
#### Notes
1. The `Expr::is_literal_expr` method would check for
`ExprStringLiteral` and return true if so. But now that we don't
represent the literal part of an f-string using that node, this improves
the method's behavior and confines to the actual expression. We do have
the `FStringElement::is_literal` method.
2. We avoid checking if we're in a f-string context before adding to
`string_type_definitions` because the f-string literal is now a
dedicated node and not part of `Expr`.
3. Annotations cannot use f-string so we avoid changing any rules which
work on annotation and checks for `ExprStringLiteral`.
## Test Plan
- All references of `Expr::StringLiteral` were checked to see if any of
the rules require updating to account for the f-string literal element
node.
- New test cases are added for rules which check against the literal
part of an f-string.
- Check the ecosystem results and ensure it remains unchanged.
## Performance
There's a performance penalty in the parser. The reason for this remains
unknown as it seems that the generated assembly code is now different
for the `__reduce154` function. The reduce function body is just popping
the `ParenthesizedExpr` on top of the stack and pushing it with the new
location.
- The size of `FStringElement` enum is the same as `Expr` which is what
it replaces in `FString::format_spec`
- The size of `FStringExpressionElement` is the same as
`ExprFormattedValue` which is what it replaces
I tried reducing the `Expr` enum from 80 bytes to 72 bytes but it hardly
resulted in any performance gain. The difference can be seen here:
- Original profile: https://share.firefox.dev/3Taa7ES
- Profile after boxing some node fields:
https://share.firefox.dev/3GsNXpD
### Backtracking
I tried backtracking the changes to see if any of the isolated change
produced this regression. The problem here is that the overall change is
so small that there's only a single checkpoint where I can backtrack and
that checkpoint results in the same regression. This checkpoint is to
revert using `Expr` to the `FString::format_spec` field. After this
point, the change would revert back to the original implementation.
## Review process
The review process is similar to #7927. The first set of commits update
the node structure, parser, and related AST files. Then, further commits
update the linter and formatter part to account for the AST change.
---------
Co-authored-by: David Szotten <davidszotten@gmail.com>
## Summary
This PR updates the `ANN201`, `ANN202`, `ANN205`, and `ANN206` rules to
not create a fix for the return type when it's an abstract method and
the function body is empty i.e., it only contains either a pass
statement, docstring or an ellipsis literal.
fixes: #9004
## Test Plan
Add the following test cases:
- Abstract method with pass statement
- Abstract method with docstring
- Abstract method with ellipsis literal
- Abstract method with possible return type
## Summary
If a string has a Unicode prefix, we can't add the `r` prefix on top of
that -- we need to remove and replace it. (The Unicode prefix is
redundant anyway in Python 3.)
Closes https://github.com/astral-sh/ruff/issues/8967.
## Summary
Check PEP 695 type alias definitions for `snake-case-type-alias`
(`PYI042`) and `t-suffixed-type-alias` (`PYI043`)
Related to #8771.
## Test Plan
`cargo test`
## Summary
- Adds `add_argument` similar to existing `remove_argument` utility to
safely add arguments to functions.
- Adds autofix for `PLW1514` as per specs requested in
https://github.com/astral-sh/ruff/issues/8883 as a test
## Test Plan
Checks on existing fixtures as well as additional test and fixture for
Python 3.9 and lower fix
## Issue Link
Closes: https://github.com/astral-sh/ruff/issues/8883
## Summary
Adds detection for branches without a `return` or `raise`, so that we
can properly `Optional` the return types. I'd like to remove this and
replace it with our code graph analysis from the `unreachable.rs` rule,
but it at least fixes the worst offenders.
Closes#8942.
## Summary
Triggers `no-return-argument-annotation-in-stub` (`PYI050`) for vararg
and kwarg `NoReturn` type annotations.
Related to #8771.
## Test Plan
`cargo test`
## Summary
When a function uses `@functools.singledispatch`, we need to treat the
first argument of any implementations as runtime-required.
Closes https://github.com/astral-sh/ruff/issues/6849.
## Summary
Part 2 of implementing the reverted autofix for `PYI030`
Also handles `typing.Union` and `typing_extensions.Literal` etc, uses
the first subscript name it finds for each offensive line.
## Test Plan
<!-- How was it tested? -->
`cargo test` and manually
---------
Co-authored-by: Zanie Blue <contact@zanie.dev>
Charlie Marsh
Alex Waygood
qdegraaf
Steve C
Dimitri Papadopoulos Orfanos
Mike Bernard
Mikael Arguedas
Shantanu
asafamr-mm
Tuomas Siipola
konsti
T-256
Samuel Cormier-Iijima
Simon Brugman
Dhruv Manilawala
Philipp A
Tom Kuson