mirrors/ruff
1
0
Fork 0
mirror of https://github.com/astral-sh/ruff.git synced 2025-09-27 04:19:43 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 3
3951 commits 414 branches 382 tags 224 MiB
Commit graph

3951 commits

This branch
This branch
All branches
Author SHA1 Message Date
Tom Kuson
0e67757edb
[pylint] Implement Pylint typevar-name-mismatch (C0132) (#5501) 
## Summary

Implement Pylint `typevar-name-mismatch` (`C0132`) as
`type-param-name-mismatch` (`PLC0132`). Includes documentation. Related
to #970.

The Pylint implementation checks only `TypeVar`, but this PR checks
`TypeVarTuple`, `ParamSpec`, and `NewType` as well. This seems to better
represent the Pylint rule's [intended
behaviour](https://github.com/pylint-dev/pylint/issues/5224).

Full disclosure: I am not a fan of the translated name and think it
should probably be different.

## Test Plan

`cargo test`
 2023-07-04 18:49:43 +00:00
Charlie Marsh
c395e44bd7
Avoid PERF rules for iteration-dependent assignments (#5508) 
## Summary

We need to avoid raising "rewrite as a comprehension" violations in
cases like:

```python
d = defaultdict(list)

for i in [1, 2, 3]:
    d[i].append(i**2)
```

Closes https://github.com/astral-sh/ruff/issues/5494.

Closes https://github.com/astral-sh/ruff/issues/5500.
 2023-07-04 18:21:05 +00:00
Charlie Marsh
75da72bd7f
Update documentation to list double-quote preference first (#5507) 
Closes https://github.com/astral-sh/ruff/issues/5496.
 2023-07-04 18:06:01 +00:00
Charlie Marsh
521e6de2c8
Fix eval detection for suspicious-eval-usage (#5506) 
Closes https://github.com/astral-sh/ruff/issues/5505.
 2023-07-04 18:01:29 +00:00
Thomas de Zeeuw
0b963ddcfa
Add unreachable code rule (#5384) 
Co-authored-by: Thomas de Zeeuw <thomas@astral.sh>
Co-authored-by: Micha Reiser <micha@reiser.io>
 2023-07-04 14:27:23 +00:00
konsti
937de121f3
check-formatter-stability: Remove newlines and add --error-file (#5491) 
## Summary

This makes the output of `check-formatter-stability` more concise by
removing extraneous newlines. It also adds a `--error-file` option to
that script that allows creating a file with just the errors (without
the status messages) to share with others.

## Test Plan

I ran it over CPython and looked at the output. I then added the
`--error-file` option and looked at the contents of the file
 2023-07-04 07:54:35 +00:00
konsti
787e2fd49d
Format import statements (#5493) 
## Summary

Format import statements in all their variants. Specifically, this
implemented formatting `StmtImport`, `StmtImportFrom` and `Alias`.

## Test Plan

I added some custom snapshots, even though this has been covered well by
black's tests.
 2023-07-04 07:07:20 +00:00
Aarni Koskela
6acc316d19
Turn Linters', etc. implicit into_iter()s into explicit rules() (#5436) 
## Summary

As discussed on ~IRC~ Discord, this will make it easier for e.g. the
docs generation stuff to get all rules for a linter (using
`all_rules()`) instead of just non-nursery ones, and it also makes it
more Explicit Is Better Than Implicit to iterate over linter rules.

Grepping for `Item = Rule` reveals some remaining implicit
`IntoIterator`s that I didn't feel were necessarily in scope for this
(and honestly, iterating over a `RuleSet` makes sense).
 2023-07-03 19:35:16 -04:00
konsti
a647f31600
Don't add a magic trailing comma for a single entry (#5463) 
## Summary

If a comma separated list has only one entry, black will respect the
magic trailing comma, but it will not add a new one.

The following code will remain as is:

```python
b1 = [
    aksjdhflsakhdflkjsadlfajkslhfdkjsaldajlahflashdfljahlfksajlhfajfjfsaahflakjslhdfkjalhdskjfa
]
b2 = [
    aksjdhflsakhdflkjsadlfajkslhfdkjsaldajlahflashdfljahlfksajlhfajfjfsaahflakjslhdfkjalhdskjfa,
]
b3 = [
    aksjdhflsakhdflkjsadlfajkslhfdkjsaldajlahflashdfljahlfksajlhfajfjfsaahflakjslhdfkjalhdskjfa,
    aksjdhflsakhdflkjsadlfajkslhfdkjsaldajlahflashdfljahlfksajlhfajfjfsaahflakjslhdfkjalhdskjfa
]
```

## Test Plan

This was first discovered in
7eeadc82c2/django/contrib/admin/checks.py (L674-L681),
which i've minimized into a call test.

I've added tests for the three cases (one entry + no comma, one entry +
comma, more than one entry) to the list tests.

The diffs from the black tests get smaller.
 2023-07-03 21:48:44 +02:00
Charlie Marsh
3992c47c00
Bump version to 0.0.276 (#5488) 2023-07-03 18:02:49 +00:00
Charlie Marsh
8de5a3d29d
Allow Final assignments in stubs (#5490) 
## Summary

This fixes one incompatibility with `flake8-pyi`, and gives us a clean
pass on `typeshed`.
 2023-07-03 17:57:49 +00:00
Charlie Marsh
ed1dd09d02
Refine some perflint rules (#5484) 
## Summary

Removing some false positives based on running over `zulip`.

`PERF401` now also detects cases like:

```py
original = list(range(10000))
filtered = []
for i in original:
    filtered.append(i * i)
```

Previously, these were caught by the list-copy rule, but these too need
comprehensions.
 2023-07-03 13:53:17 -04:00
Charlie Marsh
ca497fabbd
Remove some diagnostics.extend calls (#5483) 
## Summary

It's more efficient (and more idiomatic for us) to pass in the `Checker`
directly.
 2023-07-03 16:47:23 +00:00
Charlie Marsh
00fbbe4223
Remove some additional manual iterator matches (#5482) 
## Summary

I've done a few of these PRs, I thought I'd caught them all, but missed
this pattern.
 2023-07-03 16:29:59 +00:00
Charlie Marsh
dadad0e9ed
Remove some allocations in argument detection (#5481) 
## Summary

Drive-by PR to remove some allocations around argument name matching.
 2023-07-03 12:21:26 -04:00
Charlie Marsh
d2450c25ab
Audit remove_argument usages to use end-of-function (#5480) 
## Summary

This PR applies the fix in #5478 to a variety of other call-sites, and
fixes some other range hygienic stuff in the rules that were modified.
 2023-07-03 12:21:01 -04:00
Harutaka Kawamura
1e4b88969c
Fix unnecessary-encode-utf8 to fix encode on parenthesized strings correctly (#5478) 
## Summary

Fixes #5477

## Test Plan

New test cases.
 2023-07-03 10:11:09 -04:00
Louis Dispa
dc072537e5
Fix python_formatter generate.py with rust path (#5475) 
## Summary

This PR fix an issue with the `generate.py` file of the python
formatter.
Since https://github.com/astral-sh/ruff/pull/5369 the [node.rs
file](f51dc20497/crates/ruff_python_ast/src/node.rs)
used to generate the types now has `ast::` in the enum.

```rust
pub enum AnyNode {
   ModModule(ModModule),
   ModInteractive(ModInteractive),
   ModExpression(ModExpression),
   ModFunctionType(ModFunctionType),
   ...
```

And now:

```rust
pub enum AnyNode {
   ModModule(ast::ModModule),
   ModInteractive(ast::ModInteractive),
   ModExpression(ast::ModExpression),
   ModFunctionType(ast::ModFunctionType),
   ...
```

The python script was not parsing rust paths. This PR adds the
possibility to have it.

## Test Plan

This was tested locally.

### Script output

Before

```
['ast::ModModule),', 'ast::ModInteractive),', 'ast::ModExpression),', 'ast::ModFunctionType),', 'ast::StmtFunctionDef),', 'ast::StmtAsyncFunctionDef),', 'ast::StmtClassDef),', 'ast::StmtReturn),', 'ast::StmtDelete),', 'ast::StmtAssign),', 'ast::StmtAugAssign),', 'ast::StmtAnnAssign),', 'ast::StmtFor),', 'ast::StmtAsyncFor),', 'ast::StmtWhile),', 'ast::StmtIf),', 'ast::StmtWith),', 'ast::StmtAsyncWith),', 'ast::StmtMatch),', 'ast::StmtRaise),', 'ast::StmtTry),', 'ast::StmtTryStar),', 'ast::StmtAssert),', 'ast::StmtImport),', 'ast::StmtImportFrom),', 'ast::StmtGlobal),', 'ast::StmtNonlocal),', 'ast::StmtExpr),', 'ast::StmtPass),', 'ast::StmtBreak),', 'ast::StmtContinue),', 'ast::ExprBoolOp),', 'ast::ExprNamedExpr),', 'ast::ExprBinOp),', 'ast::ExprUnaryOp),', 'ast::ExprLambda),', 'ast::ExprIfExp),', 'ast::ExprDict),', 'ast::ExprSet),', 'ast::ExprListComp),', 'ast::ExprSetComp),', 'ast::ExprDictComp),', 'ast::ExprGeneratorExp),', 'ast::ExprAwait),', 'ast::ExprYield),', 'ast::ExprYieldFrom),', 'ast::ExprCompare),', 'ast::ExprCall),', 'ast::ExprFormattedValue),', 'ast::ExprJoinedStr),', 'ast::ExprConstant),', 'ast::ExprAttribute),', 'ast::ExprSubscript),', 'ast::ExprStarred),', 'ast::ExprName),', 'ast::ExprList),', 'ast::ExprTuple),', 'ast::ExprSlice),', 'ast::ExceptHandlerExceptHandler),', 'ast::PatternMatchValue),', 'ast::PatternMatchSingleton),', 'ast::PatternMatchSequence),', 'ast::PatternMatchMapping),', 'ast::PatternMatchClass),', 'ast::PatternMatchStar),', 'ast::PatternMatchAs),', 'ast::PatternMatchOr),', 'ast::TypeIgnoreTypeIgnore),', 'Comprehension),', 'Arguments),', 'Arg),', 'ArgWithDefault),', 'Keyword),', 'Alias),', 'WithItem),', 'MatchCase),', 'Decorator),']

error: unexpected closing delimiter: `)`
 --> <stdin>:3:55
  |
2 |             use ruff_formatter::{write, Buffer, FormatResult};
  |                                 - this opening brace...     - ...matches this closing brace
3 |             use rustpython_parser::ast::ast::ModModule),;
  |                                                       ^ unexpected closing delimiter

Traceback (most recent call last):
  File "/Users/ldispa/Documents/perso/ruff/crates/ruff_python_formatter/generate.py", line 100, in <module>
    node_path.write_text(rustfmt(code))
                         ^^^^^^^^^^^^^
  File "/Users/ldispa/Documents/perso/ruff/crates/ruff_python_formatter/generate.py", line 12, in rustfmt
    return check_output(["rustfmt", "--emit=stdout"], input=code, text=True)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/opt/homebrew/Cellar/python@3.11/3.11.4_1/Frameworks/Python.framework/Versions/3.11/lib/python3.11/subprocess.py", line 466, in check_output
    return run(*popenargs, stdout=PIPE, timeout=timeout, check=True,
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/opt/homebrew/Cellar/python@3.11/3.11.4_1/Frameworks/Python.framework/Versions/3.11/lib/python3.11/subprocess.py", line 571, in run
    raise CalledProcessError(retcode, process.args,
subprocess.CalledProcessError: Command '['rustfmt', '--emit=stdout']' returned non-zero exit status 1.
```

After:
```
['ModModule', 'ModInteractive', 'ModExpression', 'ModFunctionType', 'StmtFunctionDef', 'StmtAsyncFunctionDef', 'StmtClassDef', 'StmtReturn', 'StmtDelete', 'StmtAssign', 'StmtAugAssign', 'StmtAnnAssign', 'StmtFor', 'StmtAsyncFor', 'StmtWhile', 'StmtIf', 'StmtWith', 'StmtAsyncWith', 'StmtMatch', 'StmtRaise', 'StmtTry', 'StmtTryStar', 'StmtAssert', 'StmtImport', 'StmtImportFrom', 'StmtGlobal', 'StmtNonlocal', 'StmtExpr', 'StmtPass', 'StmtBreak', 'StmtContinue', 'ExprBoolOp', 'ExprNamedExpr', 'ExprBinOp', 'ExprUnaryOp', 'ExprLambda', 'ExprIfExp', 'ExprDict', 'ExprSet', 'ExprListComp', 'ExprSetComp', 'ExprDictComp', 'ExprGeneratorExp', 'ExprAwait', 'ExprYield', 'ExprYieldFrom', 'ExprCompare', 'ExprCall', 'ExprFormattedValue', 'ExprJoinedStr', 'ExprConstant', 'ExprAttribute', 'ExprSubscript', 'ExprStarred', 'ExprName', 'ExprList', 'ExprTuple', 'ExprSlice', 'ExceptHandlerExceptHandler', 'PatternMatchValue', 'PatternMatchSingleton', 'PatternMatchSequence', 'PatternMatchMapping', 'PatternMatchClass', 'PatternMatchStar', 'PatternMatchAs', 'PatternMatchOr', 'TypeIgnoreTypeIgnore', 'Comprehension', 'Arguments', 'Arg', 'ArgWithDefault', 'Keyword', 'Alias', 'WithItem', 'MatchCase', 'Decorator']
```
 2023-07-03 16:07:57 +02:00
konsti
7ac9e0252e
Document Checking formatter stability and panics (#5415) 
This adds the documentation, but ideally we should add the CI first
 2023-07-03 11:22:19 +02:00
konsti
ca6ff72404
Change generator formatting dummy to include NOT_YET_IMPLEMENTED (#5464) 
## Summary

Change generator formatting dummy to include `NOT_YET_IMPLEMENTED`. This
makes it easier to correctly identify them as dummies

## Test Plan

This is a dummy change
 2023-07-03 09:11:14 +02:00
Charlie Marsh
94ac2c4e1b
Reorganize some flake8-pyi rules (#5472) 2023-07-03 04:39:22 +00:00
qdegraaf
93b2bd7184
[perflint] Add PERF401 and PERF402 rules (#5298) 
## Summary

Adds `PERF401` and `PERF402` mirroring `W8401` and `W8402` from
https://github.com/tonybaloney/perflint

Implementation is not super smart but should be at parity with upstream
implementation judging by:
c07391c176/perflint/comprehension_checker.py (L42-L73)

It essentially checks:

- If the body of a for-loop is just one statement
- If that statement is an `if` and the if-statement contains a call to
`append()` we flag `PERF401` and suggest a list comprehension
- If that statement is a plain call to `append()` or `insert()` we flag
`PERF402` and suggest `list()` or `list.copy()`

I've set the violation to only flag the first append call in a long
`if-else` statement for `PERF401`. Happy to change this to some other
location or make it multiple violations if that makes more sense.

## Test Plan

Fixtures were added with the relevant scenarios for both rules

## Issue Links

Refers: https://github.com/astral-sh/ruff/issues/4789
 2023-07-03 04:03:09 +00:00
Justin Prieto
0bff4ed4d3
[flake8-pyi] Implement PYI002, PYI003, PYI004, PYI005 (#5457) 
## Summary

Implements flake8-pyi checks 002, 003, 004, 005. The logic is a bit
complex, as you can see in the [original
code](57921813c1/pyi.py (L1403C18-L1403C18)).

ref: #848 

## Test Plan

Updated snapshot tests. Ran flake8 to double check lints, and ran ruff
with all PYI lints enabled to check for incorrect overlapping lint
errors.
 2023-07-02 23:52:16 -04:00
Anders Kaseorg
df13e69c3c
Format let-else with rustfmt nightly (#5461) 
Support for `let…else` formatting was just merged to nightly
(rust-lang/rust#113225). Rerun `cargo fmt` with Rust nightly 2023-07-02
to pick this up. Followup to #939.

Signed-off-by: Anders Kaseorg <andersk@mit.edu>
 2023-07-03 02:13:35 +00:00
Charlie Marsh
c8b9a46e2b
[pyupgrade] Restore the keep-runtime-typing setting (#5470) 
## Summary

This PR reverts #4427. See the included documentation for a detailed
explanation.

Closes #5434.
 2023-07-03 02:11:31 +00:00
Charlie Marsh
6cc04d64e4
[flake8-django] Skip duplicate violations in DJ012 (#5469) 
## Summary

This PR reduces the noise from `DJ012` by emitting a single violation
when you have multiple consecutive violations of the same "type".

For example, given:

```py
class MultipleConsecutiveFields(models.Model):
    """Model that contains multiple out-of-order field definitions in a row."""


    class Meta:
        verbose_name = "test"

    first_name = models.CharField(max_length=32)
    last_name = models.CharField(max_length=32)
```

It's convenient to only error on `first_name`, and not `last_name`,
since we're really flagging that the _section_ is out-of-order.

Closes #5465.
 2023-07-02 21:09:49 -04:00
Charlie Marsh
d0b2fffb87
[numpy] Add numpy-deprecated-function (NPY003) (#5468) 
## Summary

Closes #5456.
 2023-07-02 20:50:14 -04:00
Charlie Marsh
b32d1e8d78
Detect consecutive, non-newline-delimited NumPy sections (#5467) 
## Summary

Given a docstring like:

```py
def f(a: int, b: int) -> int:
    """Showcase function.
    Parameters
    ----------
    a : int
        _description_
    b : int
        _description_
    Returns
    -------
    int
        _description
    """
```

We were failing to identify `Returns` as a section, because the previous
line was neither empty nor ended with punctuation. This was causing a
false negative, where by we weren't flagging a missing line before
`Returns`. So, the very reason for the rule (no blank line) was causing
us to fail to catch it.

Note that, we did have a test case for this, which was working properly:

```py
def f() -> int:
    """Showcase function.
    Parameters
    ----------
    Returns
    -------
    """
```

...because the line before `Returns` "ends in a punctuation mark" (`-`).

Closes #5442.
 2023-07-02 20:29:45 -04:00
Charlie Marsh
af7051b976
Include BaseException in B017 rule (#5466) 
Closes #5462.
 2023-07-02 20:18:33 -04:00
Micha Reiser
f0ec9ecd67
Show BestFitting mode if it isn't FirstLine (#5452) 2023-06-30 09:49:00 +00:00
Micha Reiser
f9129e435a
Normalize '\r' in string literals to '\n' 
<!--
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 normalizes line endings inside of strings to `\n` as required by the printer.

<!-- What's the purpose of the change? What does it do, and why? -->

## Test Plan

I added a new test using `\r\n` and ran the ecosystem check. There are no remaining end of line panics. 


https://gist.github.com/MichaReiser/8f36b1391ca7b48475b3a4f592d74ff4

<!-- How was it tested? -->
 2023-06-30 10:13:23 +02:00
Micha Reiser
dc65007fe9
Use rayon to parallelize the stability 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

This PR uses rayon to parallelize the stability check by scheduling each project as its own task.

<!-- What's the purpose of the change? What does it do, and why? -->

## Test Plan

I ran the ecosystem check. It now makes use of all cores (except at the end, there are some large projects). 

## Performance

The check now completes in minutes where it took about 30 minutes before.

<!-- How was it tested? -->
 2023-06-30 10:05:25 +02:00
Micha Reiser
9c2a75284b
Preserve parentheses around left side of binary expression 
<!--
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 fixes an issue where the binary expression formatting removed parentheses around the left hand side of an expression.

<!-- What's the purpose of the change? What does it do, and why? -->

## Test Plan

I added a new regression test and re-ran the ecosystem check. It brings down the `check-formatter-stability` output from a 3.4MB file down to 900KB. 

<!-- How was it tested? -->
 2023-06-30 09:52:14 +02:00
Micha Reiser
ae25638b0b
Update Black tests (#5438) 2023-06-30 06:32:50 +00:00
Micha Reiser
f7969cf23c
ecosystem: Run git command with no human interaction flag (#5435) 2023-06-29 09:19:11 +02:00
Micha Reiser
955e9ef821
Fix invalid syntax for binary expression in unary op (#5370) 2023-06-29 08:09:26 +02:00
Micha Reiser
38189ed913
Fix invalid printer IR error (#5422) 2023-06-29 08:09:13 +02:00
David Szotten
ca5e10b5ea
format StmtTryStar (#5418) 2023-06-29 08:07:33 +02:00
Charlie Marsh
a973019358
Rewrite a variety of .contains() calls as matches! statements (#5432) 
## Summary

These have the potential to be much more efficient, as we've seen in the
past.
 2023-06-28 22:42:27 -04:00
Charlie Marsh
aa887d5a1d
Use "manual" fixability for E731 in shadowed context (#5430) 
## Summary

This PR makes E731 a "manual" fix in one other context: when the lambda
is shadowing another variable in the scope. Function declarations (with
shadowing) cause issues for type checkers, and so rewriting an
annotation, e.g., in branches of an `if` statement can lead to failures.

Closes https://github.com/astral-sh/ruff/issues/5421.
 2023-06-28 22:00:06 -04:00
Charlie Marsh
72f7f11bac
Use matches! for reserved attribute lookup (#5431) 2023-06-29 01:52:11 +00:00
Tom Kuson
5aa2a90e17
Add documentation to flake8-logging-format rules (#5417) 
## Summary

Completes the documentation for the `flake8-logging-format` rules.
Related to #2646.

I included both the `flake8-logging-format` recommendation to use the
`extra` keyword and the Pylint recommendation to pass format values as
parameters so that formatting is done lazily, as #970 suggests the
Pylint logging rules are covered by this ruleset. Using lazy formatting
via parameters is probably more common than avoiding formatting entirely
in favour of the `extra` argument, regardless.

## Test Plan

`python scripts/check_docs_formatted.py`
 2023-06-29 01:30:11 +00:00
Charlie Marsh
0e89c94947
Run shadowed-variable analyses in deferred handlers (#5181) 
## Summary

This PR extracts a bunch of complex logic from `add_binding`, instead
running the the shadowing rules in the deferred handler, thereby
decoupling the binding phase (during which we build up the semantic
model) from the analysis phase, and generally making `add_binding` much
more focused.

This was made possible by improving the semantic model to better handle
deletions -- previously, we'd "lose track" of bindings if they were
deleted, which made this kind of refactor impossible.

## Test Plan

We have good automated coverage for this, but I want to benchmark it
separately.
 2023-06-29 00:08:18 +00:00
Eric H
139a9f757b
Update default configuration.md to mention C901 rule (#5397) 2023-06-28 21:22:16 +00:00
Charlie Marsh
c5e20505f8
Remove an unsafe access in the resolver (#5428) 2023-06-28 19:08:10 +00:00
Charlie Marsh
69c4b7fa11
Add dedicated struct for implicit imports (#5427) 
## Summary

This was some feedback on a prior PR that I decided to act on
separately.
 2023-06-28 18:55:43 +00:00
Charlie Marsh
0e12eb3071
Add a snapshot test for native module resolution (#5423) 2023-06-28 18:16:39 +00:00
Charlie Marsh
864f50a3a4
Remove all unwrap calls from the resolver (#5426) 2023-06-28 18:06:17 +00:00
Charlie Marsh
4d90a5a9bc
Move resolver tests out to top-level (#5424) 
## Summary

These are really tests for the entire crate.
 2023-06-28 13:25:37 -04:00
Charlie Marsh
1d2d015bc5
Make standard input detection robust to invalid arguments (#5393) 
## Summary

This PR fixes a silent failure that manifested itself in
https://github.com/astral-sh/ruff-vscode/issues/238. In short, if the
user provided invalid arguments to Ruff in the VS Code extension (like
`"ruff.args": ["a"]`), then we generated something like the following
command:

```console
/path/to/ruff --force-exclude --no-cache --no-fix --format json - --fix a --stdin-filename /path/to/file.py
```

Since this contains both `-` and `a` as the "input files", Ruff would
treat this as if we're linting the files names `-` and `a`, rather than
linting standard input.

This PR modifies out standard input detection to force standard input
when `--stdin-filename` is present, or at least one file is `-`. (We
then warn and ignore the others.)
 2023-06-28 14:52:23 +00:00