This is the one refactor in the NumPy 2.0 upgrade rule that isn't
compatible with earlier versions of NumPy, so I'm marking it as unsafe
and adding a dedicated message.
## Summary
Currently, `UP032` applied to raw strings results in format strings with
the prefix 'fr'. This gets changed to 'rf' by Ruff format (or Black). In
order to avoid that, this PR uses the prefix 'rf' to begin with.
## Test Plan
Updated the expectation on an existing test.
## Summary
<!-- What's the purpose of the change? What does it do, and why? -->
Hi! Currently NumPy Python API is undergoing a cleanup process that will
be delivered in NumPy 2.0 (release is planned for the end of the year).
Most changes are rather simple (renaming, removing or moving a member of
the main namespace to a new place), and they could be flagged/fixed by
an additional ruff rule for numpy (e.g. changing occurrences of
`np.float_` to `np.float64`).
Would you accept such rule?
I named it `NPY201` in the existing group, so people will receive a
heads-up for changes arriving in 2.0 before actually migrating to it.
~~This is still a draft PR.~~ I'm not an expert in rust so if any part
of code can be done better please share!
NumPy 2.0 migration guide:
https://numpy.org/devdocs/numpy_2_0_migration_guide.html
NEP 52: https://numpy.org/neps/nep-0052-python-api-cleanup.html
NumPy cleanup tracking issue:
https://github.com/numpy/numpy/issues/23999
## Test Plan
A unit test is provided that checks all rule's fix cases.
## Summary
LangChain is attempting to use Ruff over their Jupyter notebooks
(https://github.com/langchain-ai/langchain/pull/12677/files), but
running into a bunch of syntax errors, the majority of which come from
our inability to recognize automagic.
If you run this in a cell:
```jupyter
pip install requests
```
Jupyter will automatically treat that as:
```jupyter
%pip install requests
```
We need to ignore cells that use these automagics, since the parser
doesn't understand them. (I guess we could support it in the parser, but
that seems much harder?). The good news is that AFAICT Jupyter doesn't
let you mix automagics with code, so by skipping these cells, we don't
miss out on analyzing any Python code.
## Test Plan
1. `cargo test`
2. Ran over LangChain and verified that there are no more errors
relating to `pip install` automagics.
## Summary
This PR removes the `unicode` flag from the string literal in
`ComparableExpr`. This flag isn't required as all strings are unicode in
Python 3 so `"foo" == u"foo"`.
## Summary
We typically avoid enforcing exclusions if a file was passed to Ruff
directly on the CLI. However, we also allow `--force-exclude`, which
ignores excluded files _even_ if they're passed to Ruff directly. This
is really important for pre-commit, which always passes changed files --
we need to exclude files passed by pre-commit if they're in the
`exclude` lists.
Turns out the new `lint.exclude` and `format.exclude` settings weren't
respecting `--force-exclude`.
Closes https://github.com/astral-sh/ruff/issues/8391.
## Summary
By using `set`, we were setting the bracket flag to `false` if another
operator was visited.
Closes https://github.com/astral-sh/ruff/issues/8379.
## Test Plan
`cargo test`
## Summary
This PR adds a new `LiteralExpressionRef` which wraps all of the literal
expression nodes in a single enum. This allows for a narrow type when
working exclusively with a literal node. Additionally, it also
implements a `Expr::as_literal_expr` method to return the new enum if
the expression is indeed a literal one.
A few rules have been updated to account for the new enum:
1. `redundant_literal_union`
2. `if_else_block_instead_of_dict_lookup`
3. `magic_value_comparison`
To account for the change in (2), a new `ComparableLiteral` has been
added which can be constructed from the new enum
(`ComparableLiteral::from(<LiteralExpressionRef>)`).
### Open Questions
1. The new `ComparableLiteral` can be exclusively used via the
`LiteralExpressionRef` enum. Should we remove all of the literal
variants from `ComparableExpr` and instead have a single
`ComparableExpr::Literal(ComparableLiteral)` variant instead?
## Test Plan
`cargo test`
## Summary
We were considering the `{` within an f-string to be a left brace, which
caused the "space-after-colon" rule to trigger incorrectly.
Closes https://github.com/astral-sh/ruff/issues/8299.
## Summary
Uses `warn_user_once!` instead of `warn!` to ensure that every warning
is shown exactly once, regardless of whether there are duplicates in the
list, or warnings that are raised by multiple configuration files.
Closes#8271.
## Summary
If the value of `shell` wasn't literally `True`, we now show a message
describing it as truthy, rather than the (misleading) `shell=True`
literal in the diagnostic.
Closes https://github.com/astral-sh/ruff/issues/8310.
**Summary** Previously, own line comment following after a docstring
followed by newline(s) before the first content statement were treated
as trailing on the docstring and we didn't insert a newline after the
docstring as black would.
Before:
```python
class ModuleBrowser:
"""Browse module classes and functions in IDLE."""
# This class is also the base class for pathbrowser.PathBrowser.
def __init__(self, master, path, *, _htest=False, _utest=False):
pass
```
After:
```python
class ModuleBrowser:
"""Browse module classes and functions in IDLE."""
# This class is also the base class for pathbrowser.PathBrowser.
def __init__(self, master, path, *, _htest=False, _utest=False):
pass
```
I'm not entirely happy about hijacking
`handle_own_line_comment_between_statements`, but i don't know a better
spot to put it.
Fixes#7948
**Test Plan** Fixtures
We previously incorrectly treated byte strings in docstring position as
docstrings because black does so
(https://github.com/astral-sh/ruff/pull/8283#discussion_r1375682931,
https://github.com/psf/black/issues/4002), even CPython doesn't
recognize them:
```console
$ python3.12
Python 3.12.0 (main, Oct 6 2023, 17:57:44) [GCC 11.4.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> def f():
... b""" a"""
...
>>> print(str(f.__doc__))
None
```
<!--
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
This PR adds `Default` for the following literal nodes:
* `StringLiteral`
* `BytesLiteral`
* `BooleanLiteral`
* `NoneLiteral`
* `EllipsisLiteral`
The implementation creates the zero value of the respective literal
nodes in terms of the Python language.
## Test Plan
`cargo test`
## Summary
This PR inlines the formatting logic for `ExprNumberLiteral` and removes
the need of having dedicated `Format*` struct for each number type.
## Test Plan
`cargo test`
## Summary
This PR splits the `Constant` enum as individual literal nodes. It
introduces the following new nodes for each variant:
* `ExprStringLiteral`
* `ExprBytesLiteral`
* `ExprNumberLiteral`
* `ExprBooleanLiteral`
* `ExprNoneLiteral`
* `ExprEllipsisLiteral`
The main motivation behind this refactor is to introduce the new AST
node for implicit string concatenation in the coming PR. The elements of
that node will be either a string literal, bytes literal or a f-string
which can be implemented using an enum. This means that a string or
bytes literal cannot be represented by `Constant::Str` /
`Constant::Bytes` which creates an inconsistency.
This PR avoids that inconsistency by splitting the constant nodes into
it's own literal nodes, literal being the more appropriate naming
convention from a static analysis tool perspective.
This also makes working with literals in the linter and formatter much
more ergonomic like, for example, if one would want to check if this is
a string literal, it can be done easily using
`Expr::is_string_literal_expr` or matching against `Expr::StringLiteral`
as oppose to matching against the `ExprConstant` and enum `Constant`. A
few AST helper methods can be simplified as well which will be done in a
follow-up PR.
This introduces a new `Expr::is_literal_expr` method which is the same
as `Expr::is_constant_expr`. There are also intermediary changes related
to implicit string concatenation which are quiet less. This is done so
as to avoid having a huge PR which this already is.
## Test Plan
1. Verify and update all of the existing snapshots (parser, visitor)
2. Verify that the ecosystem check output remains **unchanged** for both
the linter and formatter
### Formatter ecosystem check
#### `main`
| project | similarity index | total files | changed files |
|----------------|------------------:|------------------:|------------------:|
| cpython | 0.75803 | 1799 | 1647 |
| django | 0.99983 | 2772 | 34 |
| home-assistant | 0.99953 | 10596 | 186 |
| poetry | 0.99891 | 317 | 17 |
| transformers | 0.99966 | 2657 | 330 |
| twine | 1.00000 | 33 | 0 |
| typeshed | 0.99978 | 3669 | 20 |
| warehouse | 0.99977 | 654 | 13 |
| zulip | 0.99970 | 1459 | 22 |
#### `dhruv/constant-to-literal`
| project | similarity index | total files | changed files |
|----------------|------------------:|------------------:|------------------:|
| cpython | 0.75803 | 1799 | 1647 |
| django | 0.99983 | 2772 | 34 |
| home-assistant | 0.99953 | 10596 | 186 |
| poetry | 0.99891 | 317 | 17 |
| transformers | 0.99966 | 2657 | 330 |
| twine | 1.00000 | 33 | 0 |
| typeshed | 0.99978 | 3669 | 20 |
| warehouse | 0.99977 | 654 | 13 |
| zulip | 0.99970 | 1459 | 22 |
## Summary
This PR adds a new `Singleton` enum for the `PatternMatchSingleton`
node.
Earlier the node was using the `Constant` enum but the value for this
pattern can only be either `None`, `True` or `False`. With the coming PR
to remove the `Constant`, this node required a new type to fill in.
This also has the benefit of narrowing the type down to only the
possible values for the node as evident by the removal of `unreachable`.
## Test Plan
Update the AST snapshots and run `cargo test`.
## Summary
Refactor for isort implementation. Closes#7738.
I introduced a `NatOrdString` and a `NatOrdStr` type to have a naturally
ordered `String` and `&str`, and I pretty much went back to the original
implementation based on `module_key`, `member_key` and
`sorted_by_cached_key` from itertools. I tried my best to avoid
unnecessary allocations but it may have been clumsy in some places, so
feedback is appreciated! I also renamed the `Prefix` enum to
`MemberType` (and made some related adjustments) because I think this
fits more what it is, and it's closer to the wording found in the isort
documentation.
I think the result is nicer to work with, and it should make
implementing #1567 and the like easier :)
Of course, I am very much open to any and all remarks on what I did!
## Test Plan
I didn't add any test, I am relying on the existing tests since this is
just a refactor.
## Summary
Implements pylint C0415 (import-outside-toplevel) — imports should be at
the top level of a file.
The great debate I had on this implementation is whether "top-level" is
one word or two (`toplevel` or `top_level`). I opted for 2 because that
seemed to be how it is used in the codebase but the rule string itself
uses one-word "toplevel." 🤷 I'd be happy to change it as desired.
I suppose this could be auto-fixed by moving the import to the
top-level, but it seems likely that the author's intent was to actually
import this dynamically, so I view the main point of this rule is to
force some sort of explanation, and auto-fixing might be annoying.
For reference, this is what "pylint" reports:
```
> pylint crates/ruff/resources/test/fixtures/pylint/import_outside_top_level.py
************* Module import_outside_top_level
...
crates/ruff/resources/test/fixtures/pylint/import_outside_top_level.py:4:4: C0415: Import outside toplevel (string) (import-outside-toplevel)
```
ruff would now report:
```
import_outside_top_level.py:4:5: PLC0415 `import` should be used only at the top level of a file
|
3 | def import_outside_top_level():
4 | import string # [import-outside-toplevel]
| ^^^^^^^^^^^^^ PLC0415
|
```
Contributes to https://github.com/astral-sh/ruff/issues/970.
## Test Plan
Snapshot test.
## Summary
If a file has no diagnostics, then we can read and write that
information from and to the cache, even if the fix mode is `--fix` or
`--diff`. (Typically, we can't read or write such results from or to the
cache, because `--fix` and `--diff` have side effects that take place
during diagnostic analysis (writing to disk or outputting the diff).)
This greatly improves performance when running `--fix` on a codebase in
the common case (few diagnostics).
Closes#8311.
Closes https://github.com/astral-sh/ruff/issues/8315.
## Summary
This makes use of memchr and other methods to parse the strings
(hopefully) faster. It might also be worth converting the
`parse_fstring_middle` helper to use similar techniques, but I did not
implement it in this PR.
## Test Plan
This was tested using the existing tests and passed all of them.
## Summary
Implement
[`no-isinstance-type-none`](https://github.com/dosisod/refurb/blob/master/refurb/checks/builtin/no_isinstance_type_none.py)
as `isinstance-type-none` (`FURB168`).
Auto-fixes calls to `isinstance` to check if an object is `None` to a
`None` identity check. For example,
```python
isinstance(foo, type(None))
```
becomes
```python
foo is None
```
Related to #1348.
## Test Plan
`cargo test`
---------
Co-authored-by: Charlie Marsh <charlie.r.marsh@gmail.com>
## Summary
Remove wrong note on `tool.ruff.format` `exclude` option from
documentation which is referencing `extend-exclude` even if it's not
relevant for the formatter options (`exclude` is additive). See #8301
## Test Plan
N/A (Docs change)
## Summary
Add missing "is":
```diff
- The 🧪 emoji indicates that a rule in "preview".
+ The 🧪 emoji indicates that a rule is in "preview".
```
## Test Plan
N/A
## Summary
Since `--format` was changed to `--output-format` for `check`, it feels
like it makes sense for the same to work for the auxiliary commands.
This
* adds the same deprecation warning that used to be a thing in #7514
(and un-became a thing in #7984)
Fixes#7990.
## Test Plan
* `cargo run --bin=ruff -- rule --all --output-format=json` works
* `cargo run --bin=ruff -- rule --format=json` works with warnings
Change
```python
"""Test docstring"""
a = 1
```
to
```python
"""Test docstring"""
a = 1
```
in preview style, but don't touch the docstring otherwise.
Do we want to ask black to also format the content of module level
docstrings? Seems inconsistent to me that we change function and class
docstring indentation/contents but not module docstrings.
Fixes https://github.com/astral-sh/ruff/issues/7995
## Summary
This PR removes the `debug_assertion` in the `Indexer` to allow
unterminated f-strings. This is mainly a fix in the development build
which now matches the release build.
The fix is simple: remove the `debug_assertion` which means that the
there could be `FStringStart` and possibly `FStringMiddle` tokens
without a corresponding f-string range in the `Indexer`. This means that
the code requesting for the f-string index need to account for the
`None` case, making the code safer.
This also updates the code which queries the `FStringRanges` to account
for the `None` case. This will happen when the `FStringStart` /
`FStringMiddle` tokens are present but the `FStringEnd` token isn't
which means that the `Indexer` won't contain the range for that
f-string.
## Test Plan
`cargo test`
Taking the following code as an example:
```python
f"{123}
```
This only emits a `FStringStart` token, but no `FStringMiddle` or
`FStringEnd` tokens.
And,
```python
f"\.png${
```
This emits a `FStringStart` and `FStringMiddle` token, but no
`FStringEnd` token.
fixes: #8065
## Summary
Explain the meaning of the icon for screen readers (and mouse over).
Hide "inactive" (low opacity) icons from screen readers.
Remove opacity: 1 styling, it's the default opacity.
Without this change a screen reader will just read "Hammer and spanner
test tube" for the last column in each row.
## Summary
Based on [this
feedback](https://github.com/astral-sh/ruff/issues/8185#issuecomment-1780092525).
Avoid warning about `force-wrap-aliases` and `split-on-trailing-comma`
if `force-single-line` is true (which creates a dedicated import for
each imported member).
## Test Plan
Ran `ruff format . --no-cache` and verified that the warning show up
when `force-single-line=false` and aren't shown when
`force-single-line=true`
## Summary
This PR fixes the `W605` rule implementation to provide the quickfix
message as
per the fix provided.
## Test Plan
Update snapshots.
fixes: #8155
## Summary
Avoid warning about incompatible rules except if their configuration
directly conflicts with the formatter. This should reduce the noise and
potentially the need for https://github.com/astral-sh/ruff/issues/8175
and https://github.com/astral-sh/ruff/issues/8185
I also extended the rule and option documentation to mention any
potential formatter incompatibilities or whether they're redundant when
using the formatter.
* `LineTooLong`: This is a use case we explicitly want to support. Don't
warn about it
* `TabIndentation`, `IndentWithSpaces`: Only warn if
`indent-style="tab"`
* `IndentationWithInvalidMultiple`,
`IndentationWithInvalidMultipleComment`: Only warn if `indent-width !=
4`
* `OverIndented`: Don't warn, but mention that the rule is redundant
* `BadQuotesInlineString`: Warn if quote setting is different from
`format.quote-style`
* `BadQuotesMultilineString`, `BadQuotesDocstring`: Warn if `quote !=
"double"`
## Test Plan
I added a new integration test for the default configuration with `ALL`.
`ruff format` now only shows two incompatible rules, which feels more
reasonable.
## Summary
This rule is now unsafe if we can't verify that the `obj` in `raise
obj()` is a class or builtin. (If we verify that it's a function, we
don't raise at all, as before.)
See the documentation change for motivation behind the unsafe edit.
Closes https://github.com/astral-sh/ruff/issues/8228.
**Summary** Prepare for the black preview style becoming the black
stable style at the end of the year.
This adds a new test file to compare stable and preview on some relevant
preview options in black, and makes `format_dev` understand the black
preview flag. I've added poetry as a project that uses preview.
I've implemented one specific deviation (collapsing of stub
implementation in non-stub files) which showed up in poetry for testing.
This also improves poetry compatibility from 0.99891 to 0.99919.
Fixes#7440
New compatibility stats:
| project | similarity index | total files | changed files |
|----------------|------------------:|------------------:|------------------:|
| cpython | 0.75803 | 1799 | 1647 |
| django | 0.99983 | 2772 | 35 |
| home-assistant | 0.99953 | 10596 | 189 |
| poetry | 0.99919 | 317 | 12 |
| transformers | 0.99963 | 2657 | 332 |
| twine | 1.00000 | 33 | 0 |
| typeshed | 0.99978 | 3669 | 20 |
| warehouse | 0.99969 | 654 | 15 |
| zulip | 0.99970 | 1459 | 22 |
## Summary
This PR refactors the formatter diff code to reuse the
`SourceKind::diff` logic. This has the benefit that the Notebook diff
now includes the cell numbers which was not present before.
## Test Plan
Update the snapshots and verified the cell numbers.
## Summary
Given:
```python
# comment
class A:
def foo(self):
pass
```
We need to insert an additional newline between `# comment` and `class
A`. We were missing this handling for the case in which `# comment` is a
leading comment on `class A`, as opposed to a trailing comment of some
preceding statement.
In practice, I think this only applies to the specific case in which a
class or function is the first statement in a module, and there's a
single empty line between a leading comment and that class or function.
If there are no empty lines, then the comment "sticks" to the
definition; if there are two or more, then `leading_comments` will
truncate appropriately. If the class or function is nested, then we only
need one empty line anyway.
Closes https://github.com/astral-sh/ruff/issues/8215.
## Test Plan
No change in similarity.
Before:
| project | similarity index | total files | changed files |
|----------------|------------------:|------------------:|------------------:|
| cpython | 0.75803 | 1799 | 1647 |
| django | 0.99983 | 2772 | 34 |
| home-assistant | 0.99953 | 10596 | 186 |
| poetry | 0.99891 | 317 | 17 |
| transformers | 0.99966 | 2657 | 330 |
| twine | 1.00000 | 33 | 0 |
| typeshed | 0.99978 | 3669 | 20 |
| warehouse | 0.99977 | 654 | 13 |
| zulip | 0.99970 | 1459 | 22 |
After:
| project | similarity index | total files | changed files |
|----------------|------------------:|------------------:|------------------:|
| cpython | 0.75803 | 1799 | 1648 |
| django | 0.99983 | 2772 | 34 |
| home-assistant | 0.99953 | 10596 | 186 |
| poetry | 0.99891 | 317 | 17 |
| transformers | 0.99966 | 2657 | 330 |
| twine | 1.00000 | 33 | 0 |
| typeshed | 0.99978 | 3669 | 20 |
| warehouse | 0.99977 | 654 | 13 |
| zulip | 0.99970 | 1459 | 22 |
## Summary
This PR removes the `todo!()` around `IpyEscapeCommand` in the
formatter.
The `NeedsParentheses` trait needs to be implemented which always return
`Never`. The reason being that if an escape command is parenthesized,
then that's not parsed as an escape command. IOW, the parentheses
shouldn't be present around an escape command.
In the similar way, the `CanSkipOptionalParenthesesVisitor` will skip
this node.
## Test Plan
Updated the `unformatted.ipynb` fixture with new cells containing
IPython escape commands and the corresponding snapshot was verified.
Also, tested it out in a few open source repositories containing
notebooks (`openai/openai-cookbook`, `huggingface/notebooks`).
#### New cells in `unformatted.ipynb`
**Cell 2**
```markdown
A markdown cell
```
**Cell 3**
```python
def some_function(foo, bar):
pass
%matplotlib inline
```
**Cell 4**
```python
foo = %pwd
def some_function(foo,bar,):
foo = %pwd
print(foo
)
```
fixes: #8204
## Summary
This PR fixes the bug where if a Notebook contained IPython syntax, then
the format command would fail. This was because the correct mode was not
being used while parsing through the formatter code path.
## Test Plan
This PR isn't the only requirement for Notebook formatting to start
working with IPython escape commands. The following PR in the stack is
required as well.
## Summary
Python 3.12 added the `__buffer__()`/`__release_buffer_()` special
methods, which are incorrectly flagged as invalid dunder methods by
`PLW3201`.
## Test Plan
Added definitions to the test suite, and confirmed they failed without
the fix and are ignored after the fix was done.
## Summary
Related to https://github.com/astral-sh/ruff/issues/8135.
If we're not printing a `--diff`, or a summary of `--check` changes, we
can avoid sorting the list of results. Further, when sorting, we only
need to sort a small subset of the entries, in the common case (i.e., in
general, it's much more likely that a file is formatted than not).
## Test Plan
Local benchmarks suggest a 5-10% speedup on the cached behavior:
```
❯ hyperfine --warmup 3 "./target/release/ruff format ../airflow" "./target/release/sort format ../airflow"
Benchmark 1: ./target/release/ruff format ../airflow
Time (mean ± σ): 70.3 ms ± 5.2 ms [User: 52.1 ms, System: 59.0 ms]
Range (min … max): 68.3 ms … 101.7 ms 42 runs
Warning: Statistical outliers were detected. Consider re-running this benchmark on a quiet PC without any interferences from other programs. It might help to use the '--warmup' or '--prepare' options.
Benchmark 2: ./target/release/sort format ../airflow
Time (mean ± σ): 66.0 ms ± 1.4 ms [User: 48.3 ms, System: 58.4 ms]
Range (min … max): 64.7 ms … 71.8 ms 44 runs
Warning: Statistical outliers were detected. Consider re-running this benchmark on a quiet PC without any interferences from other programs. It might help to use the '--warmup' or '--prepare' options.
Summary
'./target/release/sort format ../airflow' ran
1.07 ± 0.08 times faster than './target/release/ruff format ../airflow'
```
## Summary
This PR renames the `tab-size` configuration option to `indent-width` to
express that the formatter uses the option to determine the indentation
width AND as tab width.
I first preferred naming the option `tab-width` but then decided to go
with `indent-width` because:
* It aligns with the `indent-style` option
* It would allow us to write a lint rule that asserts that each
indentation uses `indent-width` spaces.
Closes#7643
## Test Plan
Added integration test
## Summary
This PR introduces a new `pycodestyl.max-line-length` option that allows overriding the global `line-length` option for `E501` only.
This is useful when using the formatter and `E501` together, where the formatter uses a lower limit and `E501` is only used to catch extra-long lines.
Closes#7644
## Considerations
~~Our fix infrastructure asserts in some places that the fix doesn't exceed the configured `line-width`. With this change, the question is whether it should use the `pycodestyle.max-line-width` or `line-width` option to make that decision.
I opted for the global `line-width` for now, considering that it should be the lower limit. However, this constraint isn't enforced and users not using the formatter may only specify `pycodestyle.max-line-width` because they're unaware of the global option (and it solves their need).~~
~~I'm interested to hear your thoughts on whether we should use `pycodestyle.max-line-width` or `line-width` to decide on whether to emit a fix or not.~~
Edit: The linter users `pycodestyle.max-line-width`. The `line-width` option has been removed from the `LinterSettings`
## Test Plan
Added integration test. Built the documentation and verified that the links are correct.
## Summary
First time contribute to `ruff`, so If there are low-level errors,
please forgive me. 🙇
Introduce auto fix for `E275`, this partially address #8121.
## Test Plan
Already coverd.
<!--
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? -->
Close#8123
## Test Plan
<!-- How was it tested? -->
New test cases
---------
Signed-off-by: harupy <hkawamura0130@gmail.com>
## Summary
This was just a bug in the parser ranges, probably since it was
initially implemented. Given `match n % 3, n % 5: ...`, the "subject"
(i.e., the tuple of two binary operators) was using the entire range of
the `match` statement.
Closes https://github.com/astral-sh/ruff/issues/8091.
## Test Plan
`cargo test`
## Summary
This PR updates our E721 implementation and semantics to match the
updated `pycodestyle` logic, which I think is an improvement.
Specifically, we now allow `type(obj) is int` for exact type
comparisons, which were previously impossible. So now, we're largely
just linting against code like `type(obj) == int`.
This change is gated to preview mode.
Closes https://github.com/astral-sh/ruff/issues/7904.
## Test Plan
Updated the test fixture and ensured parity with latest Flake8.
## Summary
This PR updates our documentation for the upcoming formatter release.
Broadly, the documentation is now structured as follows:
- Overview
- Tutorial
- Installing Ruff
- The Ruff Linter
- Overview
- `ruff check`
- Rule selection
- Error suppression
- Exit codes
- The Ruff Formatter
- Overview
- `ruff format`
- Philosophy
- Configuration
- Format suppression
- Exit codes
- Black compatibility
- Known deviations
- Configuring Ruff
- pyproject.toml
- File discovery
- Configuration discovery
- CLI
- Shell autocompletion
- Preview
- Rules
- Settings
- Integrations
- `pre-commit`
- VS Code
- LSP
- PyCharm
- GitHub Actions
- FAQ
- Contributing
The major changes include:
- Removing the "Usage" section from the docs, and instead folding that
information into "Integrations" and the new Linter and Formatter
sections.
- Breaking up "Configuration" into "Configuring Ruff" (for generic
configuration), and new Linter- and Formatter-specific sections.
- Updating all example configurations to use `[tool.ruff.lint]` and
`[tool.ruff.format]`.
My suggestion is to pull and build the docs locally, and review by
reading them in the browser rather than trying to parse all the code
changes.
Closes https://github.com/astral-sh/ruff/issues/7235.
Closes https://github.com/astral-sh/ruff/issues/7647.
Adds a new `ruff version` sub-command which displays long version
information in the style of `cargo` and `rustc`. We include the number
of commits since the last release tag if its a development build, in the
style of Python's versioneer.
```
❯ ruff version
ruff 0.1.0+14 (947940e91 2023-10-18)
```
```
❯ ruff version --output-format json
{
"version": "0.1.0",
"commit_info": {
"short_commit_hash": "947940e91",
"commit_hash": "947940e91269f20f6b3f8f8c7c63f8e914680e80",
"commit_date": "2023-10-18",
"last_tag": "v0.1.0",
"commits_since_last_tag": 14
}
}%
```
```
❯ cargo version
cargo 1.72.1 (103a7ff2e 2023-08-15)
```
## Test plan
I've tested this manually locally, but want to at least add unit tests
for the message formatting. We'd also want to check the next release to
ensure the information is correct.
I checked build behavior with a detached head and branches.
## Future work
We could include rustc and cargo versions from the build, the current
Python version, and other diagnostic information for bug reports.
The `--version` and `-V` output is unchanged. However, we could update
it to display the long ruff version without the rust and cargo versions
(this is what cargo does). We'll need to be careful to ensure this does
not break downstream packages which parse our version string.
```
❯ ruff --version
ruff 0.1.0
```
The LSP should be updated to use `ruff version --output-format json`
instead of parsing `ruff --version`.
This is my first PR and I'm new at rust, so feel free to ask me to
rewrite everything if needed ;)
The rule must be called after deferred lambas have been visited because
of the last check (whether the lambda parameters are used in the body of
the function that's being called). I didn't know where to do it, so I
did what I could to be able to work on the rule itself:
- added a `ruff_linter::checkers::ast::analyze::lambda` module
- build a vec of visited lambdas in `visit_deferred_lambdas`
- call `analyze::lambda` on the vec after they all have been visited
Building that vec of visited lambdas was necessary so that bindings
could be properly resolved in the case of nested lambdas.
Note that there is an open issue in pylint for some false positives, do
we need to fix that before merging the rule?
https://github.com/pylint-dev/pylint/issues/8192
Also, I did not provide any fixes (yet), maybe we want do avoid merging
new rules without fixes?
## Summary
Checks for lambdas whose body is a function call on the same arguments
as the lambda itself.
### Bad
```python
df.apply(lambda x: str(x))
```
### Good
```python
df.apply(str)
```
## Test Plan
Added unit test and snapshot.
Manually compared pylint and ruff output on pylint's test cases.
## References
- [pylint
documentation](https://pylint.readthedocs.io/en/latest/user_guide/messages/warning/unnecessary-lambda.html)
- [pylint
implementation](https://github.com/pylint-dev/pylint/blob/main/pylint/checkers/base/basic_checker.py#L521-L587)
- https://github.com/astral-sh/ruff/issues/970
## Summary
The lint checks for number of arguments in a function *definition*, but
the message says “function *call*”
## Test Plan
See what breaks and change the tests
Given `print(*a_list_with_elements, sep="\n")`, we can't remove the
separator (unlike in `print(a, sep="\n")`), since we don't know how many
arguments were provided.
Closes https://github.com/astral-sh/ruff/issues/8078.
- Add changelog entry for 0.1.1
- Bump version to 0.1.1
- Require preview for fix added in #7967
- Allow duplicate headings in changelog (markdownlint setting)
<!--
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 https://github.com/astral-sh/ruff/issues/7448
Fixes https://github.com/astral-sh/ruff/issues/7892
I've removed automatic dangling comment formatting, we're doing manual
dangling comment formatting everywhere anyway (the
assert-all-comments-formatted ensures this) and dangling comments would
break the formatting there.
## Test Plan
New test file.
---------
Co-authored-by: Micha Reiser <micha@reiser.io>
Split out of #8044: In preview style, ellipsis are also collapsed in
non-stub files. This should only affect function/class contexts since
for other statements stub are generally not used. I've updated our tests
to use `pass` instead to reflect this, which makes tracking the preview
style changes much easier.
## Summary
Given an expression like `[x for (x) in y]`, we weren't skipping over
parentheses when searching for the `in` between `(x)` and `y`.
Closes https://github.com/astral-sh/ruff/issues/8053.
## Summary
In #6157 a warning was introduced when users use `ruff: noqa`
suppression in-line instead of at the file-level. I had this trigger
today after forgetting about it, and the warning is an excellent
improvement.
I knew immediately what the issue was because I raised it previously,
but on reading the warning I'm not sure it would be so obvious to all
users. This PR extends the error with a short sentence explaining that
line-level suppression should omit the `ruff:` prefix.
## Test Plan
Not sure it's necessary for such a trivial change :)
**Summary** `ruff format --diff` is similar to `ruff format --check`,
but we don't only error with the list of file that would be formatted,
but also show a diff between the unformatted input and the formatted
output.
```console
$ ruff format --diff scratch.py scratch.pyi scratch.ipynb
warning: `ruff format` is not yet stable, and subject to change in future versions.
--- scratch.ipynb
+++ scratch.ipynb
@@ -1,3 +1,4 @@
import numpy
-maths = (numpy.arange(100)**2).sum()
-stats= numpy.asarray([1,2,3,4]).median()
+
+maths = (numpy.arange(100) ** 2).sum()
+stats = numpy.asarray([1, 2, 3, 4]).median()
--- scratch.py
+++ scratch.py
@@ -1,3 +1,3 @@
x = 1
-y=2
+y = 2
z = 3
2 files would be reformatted, 1 file left unchanged
```
With `--diff`, the summary message gets printed to stderr to allow e.g.
`ruff format --diff . > format.patch`.
At the moment, jupyter notebooks are formatted as code diffs, while
everything else is a real diff that could be applied. This means that
the diffs containing jupyter notebooks are not real diffs and can't be
applied. We could change this to json diffs, but they are hard to read.
We could also split the diff option into a human diff option, where we
deviate from the machine readable diff constraints, and a proper machine
readable, appliable diff output that you can pipe into other tools.
To make the tests work, the results (and errors, if any) are sorted
before printing them. Previously, the print order was random, i.e. two
identical runs could have different output.
Open question: Should this go into the markdown docs? Or will this be
subsumed by the integration of the formatter into `ruff check`?
**Test plan** Fixtures for the change and no change cases, including a
jupyter notebook and for file input and stdin.
Fixes#7231
---------
Co-authored-by: Micha Reiser <micha@reiser.io>
**Summary** Insert a newline after nested function and class
definitions, unless there is a trailing own line comment.
We need to e.g. format
```python
if platform.system() == "Linux":
if sys.version > (3, 10):
def f():
print("old")
else:
def f():
print("new")
f()
```
as
```python
if platform.system() == "Linux":
if sys.version > (3, 10):
def f():
print("old")
else:
def f():
print("new")
f()
```
even though `f()` is directly preceded by an if statement, not a
function or class definition. See the comments and fixtures for trailing
own line comment handling.
**Test Plan** I checked that the new content of `newlines.py` matches
black's formatting.
---------
Co-authored-by: Charlie Marsh <charlie.r.marsh@gmail.com>
## Summary
When linting, we store a map from file path to fixes, which we then use
to show a fix summary in the printer.
In the printer, we assume that if the map is non-empty, then we have at
least one fix. But this isn't enforced by the fix struct, since you can
have an entry from (file path) to (empty fix table). In practice, this
only bites us when linting from `stdin`, since when linting across
multiple files, we have an `AddAssign` on `Diagnostics` that avoids
adding empty entries to the map. When linting from `stdin`, we create
the map directly, and so it _is_ possible to have a non-empty map that
doesn't contain any fixes, leading to a panic.
This PR introduces a dedicated struct to make these constraints part of
the formal interface.
Closes https://github.com/astral-sh/ruff/issues/8027.
## Test Plan
`cargo test` (notice two failures are removed)
<!--
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
In https://github.com/astral-sh/ruff/pull/7968, I introduced a
regression whereby we started to treat imports used _only_ in type
annotation bounds (with `__future__` annotations) as unused.
The root of the issue is that I started using `visit_annotation` for
these bounds. So we'd queue up the bound in the list of deferred type
parameters, then when visiting, we'd further queue it up in the list of
deferred type annotations... Which we'd then never visit, since deferred
type annotations are visited _before_ deferred type parameters.
Anyway, the better solution here is to use a dedicated flag for these,
since they have slightly different behavior than type annotations.
I've also fixed what I _think_ is a bug whereby we previously failed to
resolve `Callable` in:
```python
type RecordCallback[R: Record] = Callable[[R], None]
from collections.abc import Callable
```
IIUC, the values in type aliases should be evaluated lazily, like type
parameters.
Closes https://github.com/astral-sh/ruff/issues/8017.
## Test Plan
`cargo test`
## Summary
Rule B005 of flake8-bugbear docs has a typo in one of the examples that
leads to a confusion in the correctness of `.strip()` method
```python
# Wrong output (used in docs)
"text.txt".strip(".txt") # "ex"
# Correct output
"text.txt".strip(".txt") # "e"
```
## Summary
Fix a typo in the docs for quote style.
> a = "a string without any quotes"
> b = "It's monday morning"
> Ruff will change a to use single quotes when using quote-style =
"single". However, a will be unchanged, as converting to single quotes
would require the inner ' to be escaped, which leads to less readable
code: 'It\'s monday morning'.
It should read "However, **b** will be unchanged".
## Test Plan
N/A.
## Summary
### What it does
This rule triggers an error when a bare raise statement is not in an
except or finally block.
### Why is this bad?
If raise statement is not in an except or finally block, there is no
active exception to
re-raise, so it will fail with a `RuntimeError` exception.
### Example
```python
def validate_positive(x):
if x <= 0:
raise
```
Use instead:
```python
def validate_positive(x):
if x <= 0:
raise ValueError(f"{x} is not positive")
```
## Test Plan
Added unit test and snapshot.
Manually compared ruff and pylint outputs on pylint's tests.
## References
- [pylint
documentation](https://pylint.pycqa.org/en/stable/user_guide/messages/error/misplaced-bare-raise.html)
- [pylint
implementation](https://github.com/pylint-dev/pylint/blob/main/pylint/checkers/exceptions.py#L339)
See the provided breaking changes note for details.
Removes support for the deprecated `--format`option in the `ruff check`
CLI, `format` inference as `output-format` in the configuration file,
and the `RUFF_FORMAT` environment variable.
The error message for use of `format` in the configuration file could be
better, but would require some awkward serde wrappers and it seems hard
to present the correct schema to the user still.
## Summary
Given `type RecordOrThings = Record | int | str`, the right-hand side
won't be evaluated at runtime. Same goes for `Record` in `type
RecordCallback[R: Record] = Callable[[R], None]`. This PR modifies the
visitation logic to treat them as typing-only.
Closes https://github.com/astral-sh/ruff/issues/7966.
## Summary
Unlike other filepath-based settings, the `cache-dir` wasn't being
resolved relative to the project root, when specified as an absolute
path.
Closes https://github.com/astral-sh/ruff/issues/7958.
## Summary
This PR adds a new `cell` field to the JSON output format which
indicates the Notebook cell this diagnostic (and fix) belongs to. It
also updates the location for the diagnostic and fixes as per the
`NotebookIndex`. It will be used in the VSCode extension to display the
diagnostic in the correct cell.
The diagnostic and edit start and end source locations are translated
for the notebook as per the `NotebookIndex`. The end source location for
an edit needs some special handling.
### Edit end location
To understand this, the following context is required:
1. Visible lines in Jupyter Notebook vs JSON array strings: The newline
is part of the string in the JSON format. This means that if there are 3
visible lines in a cell where the last line is empty then the JSON would
contain 2 strings in the source array, both ending with a newline:
**JSON format:**
```json
[
"# first line\n",
"# second line\n",
]
```
**Notebook view:**
```python
1 # first line
2 # second line
3
```
2. If an edit needs to remove an entire line including the newline, then
the end location would be the start of the next row.
To remove a statement in the following code:
```python
import os
```
The edit would be:
```
start: row 1, col 1
end: row 2, col 1
```
Now, here's where the problem lies. The notebook index doesn't have any
information for row 2 because it doesn't exists in the actual notebook.
The newline was added by Ruff to concatenate the source code and it's
removed before writing back. But, the edit is computed looking at that
newline.
This means that while translating the end location for an edit belong to
a Notebook, we need to check if both the start and end location belongs
to the same cell. If not, then the end location should be the first
character of the next row and if so, translate that back to the last
character of the previous row. Taking the above example, the translated
location for Notebook would be:
```
start: row 1, col 1
end: row 1, col 10
```
## Test Plan
Add test cases for notebook output in the JSON format and update
existing snapshots.
## Summary
This PR refactors the `NotebookIndex` struct to use `OneIndexed` to make
the
intent of the code clearer.
## Test Plan
Update the existing test case and run `cargo test` to verify the change.
- [x] Verify `--diff` output
- [x] Verify the diagnostics output
- [x] Verify `--show-source` output
**Summary** Handle comment before the default values of function
parameters correctly by inserting a line break instead of space after
the equals sign where required.
```python
def f(
a = # parameter trailing comment; needs line break
1,
b =
# default leading comment; needs line break
2,
c = ( # the default leading can only be end-of-line with parentheses; no line break
3
),
d = (
# own line leading comment with parentheses; no line break
4
)
)
```
Fixes#7603
**Test Plan** Added the different cases and one more complex case as
fixtures.
## Summary
This PR fixes the bug where the rule `E251` was being triggered on a equal token
inside a f-string which was used in the context of debug expressions.
For example, the following was being flagged before the fix:
```python
print(f"{foo = }")
```
But, now it is not. This leads to false negatives such as:
```python
print(f"{foo(a = 1)}")
```
One solution would be to know if the opened parentheses was inside a f-string or
not. If it was then we can continue flagging until it's closed. If not, then we
should not flag it.
## Test Plan
Add new test cases and check that they don't raise any false positives.
fixes: #7882
## Summary
`foo(**{})` was an overlooked edge case for `PIE804` which introduced a
crash within the Fix, introduced in #7884.
I've made it so that `foo(**{})` turns into `foo()` when applied with
`--fix`, but is that desired/expected? 🤔 Should we just ignore instead?
## Test Plan
`cargo test`
Closes https://github.com/astral-sh/ruff/issues/7572
Drops formatting specific rules from the default rule set as they
conflict with formatters in general (and in particular, conflict with
our formatter). Most of these rules are in preview, but the removal of
`line-too-long` and `mixed-spaces-and-tabs` is a change to the stable
rule set.
## Example
The following no longer raises `E501`
```
echo "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx = 1" | ruff check -
```
## Summary
Throughout the codebase, we have this pattern:
```rust
let mut diagnostic = ...
if checker.patch(Rule::UnusedVariable) {
// Do the fix.
}
diagnostics.push(diagnostic)
```
This was helpful when we computed fixes lazily; however, we now compute
fixes eagerly, and this is _only_ used to ensure that we don't generate
fixes for rules marked as unfixable.
We often forget to add this, and it leads to bugs in enforcing
`--unfixable`.
This PR instead removes all of these checks, moving the responsibility
of enforcing `--unfixable` up to `check_path`. This is similar to how
@zanieb handled the `--extend-unsafe` logic: we post-process the
diagnostics to remove any fixes that should be ignored.
## Summary
Add autofix for `PLR1714` using tuples.
If added complexity is desired, we can lean into the `set` part by doing
some kind of builtin check on all of the comparator elements for
starters, since we otherwise don't know if something's hashable.
## Test Plan
`cargo test`, and manually.
Charlie Marsh
Deepyaman Datta
Micha Reiser
Christopher Covington
Mateusz Sokół
Kar Petrosyan
Zanie Blue
Dhruv Manilawala
konsti
dependabot[bot]
bluthej
Joshua Bronson
Andrew Shannon Brown
T-256
Harutaka Kawamura
Mathieu Kniewallner
Carter Snook
Tom Kuson
Lukas Burgholzer
Giulio Mazzanti
Harmon
Aarni Koskela
Jaap Roes
Spencer Brown
Weijie Guo
gouzil
Claudio Jolowicz
Flowrey
Steve C
Clément Schreiner
Valeriy Savchenko
Philipp A
Tony Lykke
Ahmed Ashraf
Eiko Wagenknecht
Sven Hager
Victor Hugo Gomes
Jake Park