## Summary
This PR introduces a new semantic model flag `DOCSTRING` which suggests
that the model is currently in a module / class / function docstring.
This is the first step in eliminating the docstring detection state
machine which is prone to bugs as stated in #7595.
## Test Plan
~TODO: Is there a way to add a test case for this?~
I tested this using the following code snippet and adding a print
statement in the `string_like` analyzer to print if we're currently in a
docstring or not.
<details><summary>Test code snippet:</summary>
<p>
```python
"Docstring" ", still a docstring"
"Not a docstring"
def foo():
"Docstring"
"Not a docstring"
if foo:
"Not a docstring"
pass
class Foo:
"Docstring"
"Not a docstring"
foo: int
"Unofficial variable docstring"
def method():
"Docstring"
"Not a docstring"
pass
def bar():
"Not a docstring".strip()
def baz():
_something_else = 1
"""Not a docstring"""
```
</p>
</details>
## Summary
Implement [implicit readlines
(FURB129)](https://github.com/dosisod/refurb/blob/master/refurb/checks/iterable/implicit_readlines.py)
lint.
## Notes
I need a help/an opinion about suggested implementations.
This implementation differs from the original one from `refurb` in the
following way. This implementation checks syntactically the call of the
method with the name `readlines()` inside `for` {loop|generator
expression}. The implementation from refurb also
[checks](https://github.com/dosisod/refurb/blob/master/refurb/checks/iterable/implicit_readlines.py#L43)
that callee is a variable with a type `io.TextIOWrapper` or
`io.BufferedReader`.
- I do not see a simple way to implement the same logic.
- The best I can have is something like
```rust
checker.semantic().binding(checker.semantic().resolve_name(attr_expr.value.as_name_expr()?)?).statement(checker.semantic())
```
and analyze cases. But this will be not about types, but about guessing
the type by assignment (or with) expression.
- Also this logic has several false negatives, when the callee is not a
variable, but the result of function call (e.g. `open(...)`).
- On the other side, maybe it is good to lint this on other things,
where this suggestion is not safe, and push the developers to change
their interfaces to be less surprising, comparing with the standard
library.
- Anyway while the current implementation has false-positives (I
mentioned some of them in the test) I marked the fixes to be unsafe.
## Summary
Accept 0.0 and 1.0 as common magic values. This is in line with the
pylint behaviour, and I think makes sense conceptually.
## Test Plan
Test cases were added to
`crates/ruff_linter/resources/test/fixtures/pylint/magic_value_comparison.py`
## Summary
I was surprised to learn that we treat `x` in `[_ for x in y]` as an
"assignment" binding kind, rather than a dedicated comprehension
variable.
The docs previously mentioned an irrelevant config option, but were
missing a link to the relevant `ignore-init-module-imports` config
option which _is_ actually used.
Additionally, this commit adds a link to the documentation to explain
the conventions around a module interface which includes using a
redundant import alias to preserve an unused import.
(noticed this while filing #9962)
## Summary
This PR renames the semantic model flag `MODULE_DOCSTRING` to
`MODULE_DOCSTRING_BOUNDARY`. The main reason is for readability and for
the new semantic model flag `DOCSTRING` which tracks that the model is
in a module / class / function docstring.
I got confused earlier with the name until I looked at the use case and
it seems that the `_BOUNDARY` prefix is more appropriate for the
use-case and is consistent with other flags.
## Summary
This PR ensures that if a list `x` is modified within a `for` loop, we
avoid flagging `list(x)` as unnecessary. Previously, we only detected
calls to exactly `.append`, and they couldn't be nested within other
statements.
Closes https://github.com/astral-sh/ruff/issues/9925.
## Summary
If these are defined within class scopes, they're actually attributes of
the class, and can be accessed through the class itself.
(We preserve our existing behavior for `.pyi` files.)
Closes https://github.com/astral-sh/ruff/issues/9948.
Fixes#8368
Fixes https://github.com/astral-sh/ruff/issues/9186
## Summary
Arbitrary TOML strings can be provided via the command-line to override
configuration options in `pyproject.toml` or `ruff.toml`. As an example:
to run over typeshed and respect typeshed's `pyproject.toml`, but
override a specific isort setting and enable an additional pep8-naming
setting:
```
cargo run -- check ../typeshed --no-cache --config ../typeshed/pyproject.toml --config "lint.isort.combine-as-imports=false" --config "lint.extend-select=['N801']"
```
---------
Co-authored-by: Micha Reiser <micha@reiser.io>
Co-authored-by: Zanie Blue <contact@zanie.dev>
## Summary
Currently these rules apply the heuristic that if the original sequence
doesn't have a newline in between the final sequence item and the
closing parenthesis, the autofix won't add one for you. The feedback
from @ThiefMaster, however, was that this was producing slightly unusual
formatting -- things like this:
```py
__all__ = [
"b", "c",
"a", "d"]
```
were being autofixed to this:
```py
__all__ = [
"a",
"b",
"c",
"d"]
```
When, if it was _going_ to be exploded anyway, they'd prefer something
like this (with the closing parenthesis on its own line, and a trailing comma added):
```py
__all__ = [
"a",
"b",
"c",
"d",
]
```
I'm still pretty skeptical that we'll be able to please everybody here
with the formatting choices we make; _but_, on the other hand, this
_specific_ change is pretty easy to make.
## Test Plan
`cargo test`. I also ran the autofixes for RUF022 and RUF023 on CPython
to check how they looked; they looked fine to me.
## Summary
If a generic appears multiple times on the right-hand side, we should
only include it once on the left-hand side when rewriting.
Closes https://github.com/astral-sh/ruff/issues/9904.
## Summary
This review contains a fix for
[D405](https://docs.astral.sh/ruff/rules/capitalize-section-name/)
(capitalize-section-name)
The problem is that Ruff considers the sub-section header as a normal
section if it has the same name as some section name. For instance, a
function/method has an argument named "parameters". This only applies if
you use Numpy style docstring.
See: [ISSUE](https://github.com/astral-sh/ruff/issues/9806)
The following will not raise D405 after the fix:
```python
def some_function(parameters: list[str]):
"""A function with a parameters parameter
Parameters
----------
parameters:
A list of string parameters
"""
...
```
## Test Plan
```bash
cargo test
```
---------
Co-authored-by: Mikko Leppänen <mikko.leppanen@vaisala.com>
Co-authored-by: Charlie Marsh <charlie.r.marsh@gmail.com>
## Summary
This PR reduces the size of `Expr` from 80 to 64 bytes, by reducing the
sizes of...
- `ExprCall` from 72 to 56 bytes, by using boxed slices for `Arguments`.
- `ExprCompare` from 64 to 48 bytes, by using boxed slices for its
various vectors.
In testing, the parser gets a bit faster, and the linter benchmarks
improve quite a bit.
## Summary
Corrects mentions of `Path.is_link` to `Path.is_symlink` (the former
doesn't exist).
## Test Plan
```sh
python scripts/generate_mkdocs.py && mkdocs serve -f mkdocs.public.yml
```
Fixes#9857.
## Summary
Statements like `logging.info("Today it is: {day}")` will no longer be
ignored by RUF027. As before, statements like `"Today it is:
{day}".format(day="Tuesday")` will continue to be ignored.
## Test Plan
The snapshot tests were expanded to include new cases. Additionally, the
snapshot tests have been split in two to separate positive cases from
negative cases.
## Summary
Django's `mark_safe` can also be used as a decorator, so we should
detect usages of `@mark_safe` for the purpose of the relevant Bandit
rule.
Closes https://github.com/astral-sh/ruff/issues/9780.
## Summary
Given:
```python
"""Make a summary line.
Note:
----
Per the code comment the next two lines are blank. "// The first blank line is the line containing the closing
triple quotes, so we need at least two."
"""
```
It turns out we excluded the line ending in `"""`, because it's empty
(unlike for functions, where it consists of the indent). This PR changes
the `following_lines` iterator to always include the trailing newline,
which gives us correct and consistent handling between function and
module-level docstrings.
Closes https://github.com/astral-sh/ruff/issues/9877.
## Summary
The benchmarks show a pretty consistent 1% speedup here for all-rules,
though not enough to trigger our threshold of course:
#2977 added the `allow-dict-calls-with-keyword-arguments` configuration
option for the `unnecessary-collection-call (C408)` rule, but it did not
update the rule description.
## Summary
When we fall through to parsing, the comment-detection rule is a
significant portion of lint time. This PR adds an additional fast
heuristic whereby we abort if a comment contains two consecutive name
tokens (via the zero-allocation lexer). For the `ctypeslib.py`, which
has a few cases that are now caught by this, it's a 2.5x speedup for the
rule (and a 20% speedup for token-based rules).
These are for descriptors which affects the behavior of the object _as a
property_; I do not think they should be called directly but there is no
alternative when working with the object directly.
Closes https://github.com/astral-sh/ruff/issues/9789
## Summary
These run over nearly every identifier. It's rare to override them, so
when not provided, we can just use a match against the hardcoded default
set.
<!--
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
Fixes#8151
This PR implements a new rule, `RUF027`.
## What it does
Checks for strings that contain f-string syntax but are not f-strings.
### Why is this bad?
An f-string missing an `f` at the beginning won't format anything, and
instead treat the interpolation syntax as literal.
### Example
```python
name = "Sarah"
dayofweek = "Tuesday"
msg = "Hello {name}! It is {dayofweek} today!"
```
It should instead be:
```python
name = "Sarah"
dayofweek = "Tuesday"
msg = f"Hello {name}! It is {dayofweek} today!"
```
## Heuristics
Since there are many possible string literals which contain syntax
similar to f-strings yet are not intended to be,
this lint will disqualify any literal that satisfies any of the
following conditions:
1. The string literal is a standalone expression. For example, a
docstring.
2. The literal is part of a function call with keyword arguments that
match at least one variable (for example: `format("Message: {value}",
value = "Hello World")`)
3. The literal (or a parent expression of the literal) has a direct
method call on it (for example: `"{value}".format(...)`)
4. The string has no `{...}` expression sections, or uses invalid
f-string syntax.
5. The string references variables that are not in scope, or it doesn't
capture variables at all.
6. Any format specifiers in the potential f-string are invalid.
## Test Plan
I created a new test file, `RUF027.py`, which is both an example of what
the lint should catch and a way to test edge cases that may trigger
false positives.
## Summary
It turns out we saw a panic in cases when dedenting blocks like the `def
wrapper` here:
```python
def instrument_url(f: UrlFuncT) -> UrlFuncT:
# TODO: Type this with ParamSpec to preserve the function signature.
if not INSTRUMENTING: # nocoverage -- option is always enabled; should we remove?
return f
else:
def wrapper(
self: "ZulipTestCase", url: str, info: object = {}, **kwargs: Union[bool, str]
) -> HttpResponseBase:
```
Since we relied on the first line to determine the indentation, instead
of the first non-empty line.
## Test Plan
`cargo test`
## Summary
This is a simple idea to avoid unnecessary work in the linter,
especially for rules that run on all name and/or all attribute nodes.
Imagine a rule like the NumPy deprecation check. If the user never
imported `numpy`, we should be able to skip that rule entirely --
whereas today, we do a `resolve_call_path` check on _every_ name in the
file. It turns out that there's basically a finite set of modules that
we care about, so we now track imports on those modules as explicit
flags on the semantic model. In rules that can _only_ ever trigger if
those modules were imported, we add a dedicated and extremely cheap
check to the top of the rule.
We could consider generalizing this to all modules, but I would expect
that not to be much faster than `resolve_call_path`, which is just a
hash map lookup on `TextSize` anyway.
It would also be nice to make this declarative, such that rules could
declare the modules they care about, the analyzers could call the rules
as appropriate. But, I don't think such a design should block merging
this.
