## Summary
Consider single element subscript expr for implicit optional.
On `main`, the cases where there is only a single element in the
subscript
list was giving false positives such as for the following:
```python
typing.Union[None]
typing.Literal[None]
```
## Test Plan
`cargo test`
---------
Co-authored-by: Charlie Marsh <charlie.r.marsh@gmail.com>
## Summary
Check for `Any` in other types for `ANN401`. This reuses the logic from
`implicit-optional` rule to resolve the type to `Any`.
Following types are supported:
* `Union[Any, ...]`
* `Any | ...`
* `Optional[Any]`
* `Annotated[<any of the above variant>, ...]`
* Forward references i.e., `"Any | ..."`
## Test Plan
Added test cases for various combinations.
fixes: #5458
## Summary
Do not raise `EXE001` and `EXE002` if WSL is detected. Uses the
[`wsl`](https://crates.io/crates/wsl) crate.
Closes#5445.
## Test Plan
`cargo test`
I don't use Windows, so was unable to test on a WSL environment. It
would be good if someone who runs Windows could check the functionality.
## Summary
Python doesn't allow `"Foo" | None` if the annotation will be evaluated
at runtime (see the comments in the PR, or the semantic model
documentation for more on what this means and when it is true), but it
_does_ allow it if the annotation is typing-only.
This, for example, is invalid, as Python will evaluate `"Foo" | None` at
runtime in order to
populate the function's `__annotations__`:
```python
def f(x: "Foo" | None): ...
```
This, however, is valid:
```python
def f():
x: "Foo" | None
```
As is this:
```python
from __future__ import annotations
def f(x: "Foo" | None): ...
```
Closes#5706.
## Summary
`StmtAnnAssign` would not insert parentheses when breaking the same way
`StmtAssign` does, causing unstable formatting and likely some syntax
errors.
## Test Plan
I added a regression test.
## Summary
The previous dummy was causing instabilities since it turned a string
into a variable.
E.g.
```python
script_header_dict[
"slurm_partition_line"
] = f"#SBATCH --partition {resources.queue_name}"
```
has an instability as
```python
- script_header_dict["slurm_partition_line"] = (
- NOT_YET_IMPLEMENTED_ExprJoinedStr
- )
+ script_header_dict[
+ "slurm_partition_line"
+ ] = NOT_YET_IMPLEMENTED_ExprJoinedStr
```
## Test Plan
The instability is gone, otherwise it's still a dummy
## Summary
Implement Pylint rule [`consider-using-in`
(`R1714`)](https://pylint.pycqa.org/en/latest/user_guide/messages/refactor/consider-using-in.html)
as `repeated-equality-comparison-target` (`PLR1714`). This rule checks
for expressions that can be re-written as a membership test for better
readability and performance.
For example,
```python
foo == "bar" or foo == "baz" or foo == "qux"
```
should be rewritten as
```python
foo in {"bar", "baz", "qux"}
```
Related to #970. Includes documentation.
### Implementation quirks
The implementation does not work with Yoda conditions (e.g., `"a" ==
foo` instead of `foo == "a"`). The Pylint version does. I couldn't find
a way of supporting Yoda-style conditions without it being inefficient,
so didn't (I don't think people write Yoda conditions any way).
## Test Plan
Added fixture.
`cargo test`
## Summary
We have two `Cursor` implementations. This PR moves the implementation
from the formatter into `ruff_python_whitespace` (kind of a poorly-named
crate now) and uses it for both use-cases.
## Summary
Document all `ruff_dev` subcommands and document the `format_dev` flags
in the formatter readme.
CC @zanieb please flag everything that isn't clear or missing
## Test Plan
n/a
Detects invalid types for tuple, list, bytes, string indices.
For example, the following will raise a `TypeError` at runtime and when
imported Python will display a `SyntaxWarning`
```python
var = [1, 2, 3]["x"]
```
```
example.py:1: SyntaxWarning: list indices must be integers or slices, not str; perhaps you missed a comma?
var = [1, 2, 3]["x"]
Traceback (most recent call last):
File "example.py", line 1, in <module>
var = [1, 2, 3]["x"]
~~~~~~~~~^^^^^
TypeError: list indices must be integers or slices, not str
```
Previously, Ruff would not report the invalid syntax but now a violation
will be reported. This does not apply to cases where a variable, call,
or complex expression is used in the index — detection is roughly
limited to static definitions, which matches Python's warnings.
```
❯ ./target/debug/ruff example.py --select RUF015 --show-source --no-cache
example.py:1:17: RUF015 Indexed access to type `list` uses type `str` instead of an integer or slice.
|
1 | var = [1, 2, 3]["x"]
| ^^^ RUF015
|
```
Closes https://github.com/astral-sh/ruff/issues/5082
xref
ffff1440d1
## Summary
Replaces `DictionaryKey` enum with the more general `ComparableExpr`
when checking for duplicate keys
## Test Plan
Added test fixture from issue. Can potentially be expanded further
depending on what exactly we want to flag (e.g. do we also want to check
for unhashable types?) and which `ComparableExpr::XYZ` types we consider
literals.
## Issue link
Closes: https://github.com/astral-sh/ruff/issues/5691
## Summary
Similar to #5567, we can remove the use of regex, plus simplify the
representation (use `Option`), add snapshot tests, etc.
This is about 100x faster than using a regex for cases that match (2.5ns
vs. 250ns). It's obviously not a hot path, but I prefer the consistency
with other similar comment-parsing. I may DRY these up into some common
functionality later on.
<!--
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 matches Black' behavior where it only omits the optional parentheses if the expression starts or ends with a parenthesized expression:
```python
a + [aaa, bbb, cccc] * c # Don't omit
[aaa, bbb, cccc] + a * c # Split
a + c * [aaa, bbb, ccc] # Split
```
<!-- What's the purpose of the change? What does it do, and why? -->
## Test Plan
This improves the Jaccard index from 0.945 to 0.946
<!--
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 improves the Black compatibility when it comes to breaking comprehensions.
We want to avoid line breaks before the target and `in` whenever possible. Furthermore, `if X is not None` should be grouped together, similar to other binary like expressions
<!-- What's the purpose of the change? What does it do, and why? -->
## Test Plan
`cargo test`
<!-- How was it tested? -->
## Summary
We don't use `ModExpression` anywhere but it's part of the AST, removes
one `not_implemented_yet` and is a trivial 2-liner, so i implemented
formatting for `ModExpression`.
## Test Plan
None, this kind of node does not occur in file input. Otherwise all the
tests for expressions
<!--
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
I started working on this because I assumed that I would need access to options inside of `NeedsParantheses` but it then turned out that I won't.
Anyway, it kind of felt nice to pass fewer arguments. So I'm gonna put this out here to get your feedback if you prefer this over passing individual fiels.
Oh, I sneeked in another change. I renamed `context.contents` to `source`. `contents` is too generic and doesn't tell you anything.
<!-- What's the purpose of the change? What does it do, and why? -->
## Test Plan
It compiles
<!--
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 removes the `mode` field from `BestFitting` because it is no longer used (we now use `conditional_group` and `fits_expanded).
<!-- What's the purpose of the change? What does it do, and why? -->
## Test Plan
`cargo test`
<!-- How was it tested? -->
<!--
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 implements Black's behavior where it first splits off parenthesized expressions before splitting before operands to avoid unnecessary parentheses:
```python
# We want
if a + [
b,
c
]:
pass
# Rather than
if (
a
+ [b, c]
):
pass
```
This is implemented by using the new IR elements introduced in #5596.
* We give the group wrapping the optional parentheses an ID (`parentheses_id`)
* We use `conditional_group` for the lower priority groups (all non-parenthesized expressions) with the condition that the `parentheses_id` group breaks (we want to split before operands only if the parentheses are necessary)
* We use `fits_expanded` to wrap all other parenthesized expressions (lists, dicts, sets), to prevent that expanding e.g. a list expands the `parentheses_id` group. We gate the `fits_expand` to only apply if the `parentheses_id` group fits (because we prefer `a\n+[b, c]` over expanding `[b, c]` if the whole expression gets parenthesized).
We limit using `fits_expanded` and `conditional_group` only to expressions that themselves are not in parentheses (checking the conditions isn't free)
## Test Plan
It increases the Jaccard index for Django from 0.915 to 0.917
## Incompatibilites
There are two incompatibilities left that I'm aware of (there may be more, I didn't go through all snapshot differences).
### Long string literals
I commented on the regression. The issue is that a very long string (or any content without a split point) may not fit when only breaking the right side. The formatter than inserts the optional parentheses. But this is kind of useless because the overlong string will still not fit, because there are no new split points.
I think we should ignore this incompatibility for now
### Expressions on statement level
I don't fully understand the logic behind this yet, but black doesn't break before the operators for the following example even though the expression exceeds the configured line width
```python
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa < bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb > ccccccccccccccccccccccccccccc == ddddddddddddddddddddd
```
But it would if the expression is used inside of a condition.
What I understand so far is that Black doesn't insert optional parentheses on the expression statement level (and a few other places) and, therefore, only breaks after opening parentheses. I propose to keep this deviation for now to avoid overlong-lines and use the compatibility report to make a decision if we should implement the same behavior.
## Summary
The similarity index, the fraction of unchanged lines, is easier to
understand than the jaccard index, the fraction between intersection and
union.
## Test Plan
I ran this on django and git a 0.945 index, meaning 5.5% of lines are
currently reformatted when compared to black
## Summary
Format statements such as `tree_depth += 1`. This is a statement that
does not allow any line breaks, the only thing to be mindful of is to
parenthesize the assigned expression
Jaccard index on django: 0.915 -> 0.918
## Test Plan
black tests, and two new tests, a basic one and one that ensures that
the child gets parentheses. I ran the django stability check.
## Summary
This is the result of running `cargo +nightly clippy --workspace
--all-targets --all-features -- -D warnings` and fixing all violations.
Just wanted to see if there were any interesting new checks on nightly
👀
## Summary
This PR implements the formatting of `raise` statements. I haven't
looked at the black implementation, this is inspired from from the
`return` statements formatting.
## Test Plan
The black differences with insta.
I also compared manually some edge cases with very long string and call
chaining and it seems to do the same formatting as black.
There is one issue:
```python
# input
raise OsError(
"aksjdhflsakhdflkjsadlfajkslhfdkjsaldajlahflashdfljahlfksajlhfajfjfsaahflakjslhdfkjalhdskjfa"
) from a.aaaaa(aksjdhflsakhdflkjsadlfajkslhfdkjsaldajlahflashdfljahlfksajlhfajfjfsaahflakjslhdfkjalhdskjfa).a(aaaa)
# black
raise OsError(
"aksjdhflsakhdflkjsadlfajkslhfdkjsaldajlahflashdfljahlfksajlhfajfjfsaahflakjslhdfkjalhdskjfa"
) from a.aaaaa(
aksjdhflsakhdflkjsadlfajkslhfdkjsaldajlahflashdfljahlfksajlhfajfjfsaahflakjslhdfkjalhdskjfa
).a(
aaaa
)
# ruff
raise OsError(
"aksjdhflsakhdflkjsadlfajkslhfdkjsaldajlahflashdfljahlfksajlhfajfjfsaahflakjslhdfkjalhdskjfa"
) from a.aaaaa(
aksjdhflsakhdflkjsadlfajkslhfdkjsaldajlahflashdfljahlfksajlhfajfjfsaahflakjslhdfkjalhdskjfa
).a(aaaa)
```
But I'm not sure this diff is the raise formatting implementation.
---------
Co-authored-by: Louis Dispa <ldispa@deezer.com>
## Summary
#5658 didn't actually ignore bivariate types in some all cases (sorry
about that). This PR fixes that and adds bivariate types to the test
fixture.
## Test Plan
`cargo test`
## Summary
Change the `type-name-incorrect-variance` diagnostic message to include
the detected variance and a name change recommendation. For example,
```
`TypeVar` name "T_co" does not reflect its contravariance; consider renaming it to "T_contra"
```
Related to #5651.
## Test Plan
`cargo test`
## Summary
Non-behavior-changing refactors to delay some `.is_builtin` calls in a
few older rules. Cheaper pre-conditions should always be checked first.
## Summary
Fixes#5503. Ready for final review as the `mkdocs` issue involving SSH
keys is fixed.
Note that this will only throw on a `Name` - it will be refactorable
once we have a type-checker. This means that this is the only sort of
input that will throw.
```python
x = range(10)
list(x)[0]
```
I thought it'd be confusing if we supported direct function results.
Consider this example, assuming we support direct results:
```python
# throws
list(range(10))[0]
def createRange(bound):
return range(bound)
# "why doesn't this throw, but a direct `range(10)` call does?"
list(createRange(10))[0]
```
If it's necessary, I can go through the list of built-ins and find those
which produce iterables, then add them to the throwing list.
## Test Plan
Added a new fixture, then ran `cargo t`
## Summary
Implement Pylint `typevar-name-incorrect-variance` (`C0105`) as
`type-name-incorrect-variance` (`PLC0105`). Includes documentation.
Related to #970.
The Pylint implementation checks only `TypeVar`, but this PR checks
`ParamSpec` as well.
## Test Plan
Added test fixture.
`cargo test`
## Summary
Completes all the documentation for the `pandas-vet` rules, except for
`pandas-use-of-dot-read-table` as I am unclear of the rule's motivation
(see #5628).
Related to #2646.
## Test Plan
`python scripts/check_docs_formatted.py && mkdocs serve`
## Summary
Fixes#5246. We generate a hash set of all exception IDs caught by the
`try` statement, then check that the inner `raise` actually raises a
caught exception.
## Test Plan
Added a new test, `cargo t`.
## Summary
Fix an oversight in `find_only_token_in_range` where the following code
would panic due do the closing and opening parentheses being in the
range we scan:
```python
d1 = [
("a") if # 1
("b") else # 2
("c")
]
```
Closing and opening parentheses respectively are now correctly skipped.
## Test Plan
I added a regression test
## Summary
This PR reworks the `upstream_categories` mechanism that is only used
for documentation purposes to make it easier to generate docs using
`all_rules()`. The new implementation also relies on "tribal knowledge"
about rule codes, so it's not the best implementation, but gets us
forward.
Another option would be to change the rule-defining proc macros to allow
configuring an optional `RuleCategory`, but that seems more heavy-handed
and possibly unnecessary in the long run...
Draft since this builds on #5439.
cc @charliermarsh :)
## Summary
Format named expressions (walrus operator) such a `value := f()`.
Unlike tuples, named expression parentheses are not part of the range
even when mandatory, so mapping optional parentheses to always gives us
decent formatting without implementing all [PEP
572](https://peps.python.org/pep-0572/) rules on when we need
parentheses where other expressions wouldn't. We might want to revisit
this decision later and implement special cases, but for now this gives
us what we need.
## Test Plan
black fixtures, i added some fixtures and checked django and cpython for
stability.
Closes#5613
## Summary
This changes the docs to show a nursery icon (🌅) for rules in the
nursery.
It currently doesn't do that for the rules that are in sub-categories
(Pylint, Pycodestyle) because there is no `all_rules()` for the
`RuleCodePrefix` that's returned by `UpstreamCategory` iteration (and as
mentioned on Discord, I think `UpstreamCategory` maybe shouldn't be a
thing). (That would be enabled by #5591.)
## Test Plan
Generated docs to see new icons (with the caveat above).
## Summary
It turns out that just doing this match directly without `AhoCorasick`
is much faster, like 2x (and removes one dependency, though we likely
already rely on this transitively).
## Summary
Adds `import tkinter as tk` to the list of default import conventions.
Closes#5620.
## Test Plan
Added `tkinter` to test fixture.
`cargo test`
## Summary
We're doing some unsafe accesses to advance these iterators. It's easier
to model these as actual iterators to ensure safety everywhere. Also
added some additional test cases.
Closes#5621.
## Summary
We now treat `# flake8: noqa: F401` as turning off F401 for the entire
file. (Flake8 treats this as turning off _all rules_ for the entire
file).
This deviates from Flake8, but I think it's a much more user-friendly
deviation than what I introduced in #5571. See
https://github.com/astral-sh/ruff/issues/5617 for an explanation.
Closes https://github.com/astral-sh/ruff/issues/5617.
## Summary
This PR adds a `ParseError` type to the `noqa` parsing system to enable
us to render useful warnings instead of silently failing when parsing
`noqa` codes.
For example, given `foo.py`:
```python
# ruff: noqa: x
# ruff: noqa foo
# flake8: noqa: F401
import os # noqa: foo-bar
```
We would now output:
```console
warning: Invalid `# noqa` directive on line 2: expected a comma-separated list of codes (e.g., `# noqa: F401, F841`).
warning: Invalid `# noqa` directive on line 4: expected `:` followed by a comma-separated list of codes (e.g., `# noqa: F401, F841`).
warning: Invalid `# noqa` directive on line 6: Flake8's blanket exemption does not support exempting specific codes. To exempt specific codes, use, e.g., `# ruff: noqa: F401, F841` instead.
warning: Invalid `# noqa` directive on line 7: expected a comma-separated list of codes (e.g., `# noqa: F401, F841`).
```
There's one important behavior change here too. Right now, with Flake8,
if you do `# flake8: noqa: F401`, Flake8 treats that as equivalent to `#
flake8: noqa` -- it turns off _all_ diagnostics in the file, not just
`F401`. Historically, we respected this... but, I think it's confusing.
So we now raise a warning, and don't respect it at all. This will lead
to errors in some projects, but I'd argue that right now, those
directives are almost certainly behaving in an unintended way for users
anyway.
Closes https://github.com/astral-sh/ruff/issues/3339.
## Summary
I was testing some changes on Airflow, and I realized that we _always_
run the `pyproject.toml` validation rules, even if they're not enabled.
This PR gates them behind the appropriate enablement flags.
## Test Plan
- Ran: `cargo run -p ruff_cli -- check ../airflow -n`. Verified that no
RUF200 violations were raised.
- Run: `cargo run -p ruff_cli -- check ../airflow -n --select RUF200`.
Verified that two RUF200 violations were raised.
Since the (implicit) update to cargo-insta 1.30, CI would pass even when
the tests failed. This downgrades to cargo insta 1.29.0 and CI fails
again when it should (which i can't show here, because CI needs to pass
to merge this PR). I've improved the unreferenced snapshot handling in
the process
See https://github.com/mitsuhiko/insta/issues/392
<!--
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
Fix typos found by
[codespell](https://github.com/codespell-project/codespell).
I have left out `memoize` for now (see #5606).
<!-- What's the purpose of the change? What does it do, and why? -->
## Test Plan
CI tests.
<!-- How was it tested? -->
## Summary
Format `ExprIfExp`, also known as the ternary operator or inline `if`.
It can look like
```python
a1 = 1 if True else 2
```
but also
```python
b1 = (
# We return "a" ...
"a" # that's our True value
# ... if this condition matches ...
if True # that's our test
# ... otherwise we return "b§
else "b" # that's our False value
)
```
This also fixes a visitor order bug.
The jaccard index on django goes from 0.911 to 0.915.
## Test Plan
I added fixtures without and with comments in strange places.
Implements PYI030 as part of
https://github.com/astral-sh/ruff/issues/848
> Union expressions should never have more than one Literal member, as
Literal[1] | Literal[2] is semantically identical to Literal[1, 2].
Note we differ slightly from the flake8-pyi implementation:
- We detect cases where there are parentheses or nested unions
- We detect cases with mixed `Union` and `|` syntax
- We use the same error message for all violations; flake8-pyi has two
different messages
- We retain the user's quoting style when displaying string literals;
flake8-pyi uses single quotes
- We warn on duplicates of the same literal `Literal[1] | Literal[1]`
## Summary
In addition to `# noqa` codes, we also support file-level exemptions,
which look like:
- `# flake8: noqa` (ignore all rules in the file, for compatibility)
- `# ruff: noqa` (all rules in the file)
- `# ruff: noqa: F401` (ignore `F401` in the file, Flake8 doesn't
support this)
This PR moves that logic to something that looks a lot more like our `#
noqa` parser. Performance is actually quite a bit _worse_ than the
previous approach (lexing `# flake8: noqa` goes from 2ns to 11ns; lexing
`# ruff: noqa: F401, F841` is about the same`; lexing `# type: ignore #
noqa: E501` fgoes from 4ns to 6ns), but the numbers are very small so
it's... maybe worth it?
The primary benefit here is that we now properly support flexible
whitespace, like: `#flake8:noqa`. Previously, we required exact string
matching, and we also didn't support all case-insensitive variants of
`noqa`.
## Summary
There are two pypi links in the documentation that link to specific
version numbers of other packages. Removing these versioned links allows
users to immediately view the latest version of the package and
maintains consistency with the other links.
## Test Plan
N/A
## Summary
This extends the `ruff_dev` formatter script util. Instead of only doing
stability checks, you can now choose different compatible options on the
CLI and get statistics.
* It adds an option the formats all files that ruff would check to allow
looking at an entire black-formatted repository with `git diff`
* It computes the [Jaccard
index](https://en.wikipedia.org/wiki/Jaccard_index) as a measure of
deviation between input and output, which is useful as single number
metric for assessing our current deviations from black.
* It adds progress bars to both the single projects as well as the
multi-project mode.
* It adds an option to write the multi-project output to a file
Sample usage:
```
$ cargo run --bin ruff_dev -- format-dev --stability-check crates/ruff/resources/test/cpython
$ cargo run --bin ruff_dev -- format-dev --stability-check /home/konsti/projects/django
Syntax error in /home/konsti/projects/django/tests/test_runner_apps/tagged/tests_syntax_error.py: source contains syntax errors (parser error): BaseError { error: UnrecognizedToken(Name { name: "syntax_error" }, None), offset: 131, source_path: "<filename>" }
Found 0 stability errors in 2755 files (jaccard index 0.911) in 9.75s
$ cargo run --bin ruff_dev -- format-dev --write /home/konsti/projects/django
```
Options:
```
Several utils related to the formatter which can be run on one or more repositories. The selected set of files in a repository is the same as for `ruff check`.
* Check formatter stability: Format a repository twice and ensure that it looks that the first and second formatting look the same. * Format: Format the files in a repository to be able to check them with `git diff` * Statistics: The subcommand the Jaccard index between the (assumed to be black formatted) input and the ruff formatted output
Usage: ruff_dev format-dev [OPTIONS] [FILES]...
Arguments:
[FILES]...
Like `ruff check`'s files. See `--multi-project` if you want to format an ecosystem checkout
Options:
--stability-check
Check stability
We want to ensure that once formatted content stays the same when formatted again, which is known as formatter stability or formatter idempotency, and that the formatter prints syntactically valid code. As our test cases cover only a limited amount of code, this allows checking entire repositories.
--write
Format the files. Without this flag, the python files are not modified
--format <FORMAT>
Control the verbosity of the output
[default: default]
Possible values:
- minimal: Filenames only
- default: Filenames and reduced diff
- full: Full diff and invalid code
-x, --exit-first-error
Print only the first error and exit, `-x` is same as pytest
--multi-project
Checks each project inside a directory, useful e.g. if you want to check all of the ecosystem checkouts
--error-file <ERROR_FILE>
Write all errors to this file in addition to stdout. Only used in multi-project mode
```
## Test Plan
I ran this on django (2755 files, jaccard index 0.911) and discovered a
magic trailing comma problem and that we really needed to implement
import formatting. I ran the script on cpython to identify
https://github.com/astral-sh/ruff/pull/5558.
## Summary
In Python, the annotations on `x` and `y` here have very different
treatment:
```python
def foo(x: int):
y: int
```
The `int` in `x: int` is a runtime-required annotation, because `x` gets
added to the function's `__annotations__`. You'll notice, for example,
that this fails:
```python
from typing import TYPE_CHECKING
if TYPE_CHECKING:
from foo import Bar
def f(x: Bar):
...
```
Because `Bar` is required to be available at runtime, not just at typing
time. Meanwhile, this succeeds:
```python
from typing import TYPE_CHECKING
if TYPE_CHECKING:
from foo import Bar
def f():
x: Bar = 1
f()
```
(Both cases are fine if you use `from __future__ import annotations`.)
Historically, we've tracked those annotations that are _not_
runtime-required via the semantic model's `ANNOTATION` flag. But
annotations that _are_ runtime-required have been treated as "type
definitions" that aren't annotations.
This causes problems for the flake8-future-annotations rules, which try
to detect whether adding `from __future__ import annotations` would
_allow_ you to rewrite a type annotation. We need to know whether we're
in _any_ type annotation, runtime-required or not, since adding `from
__future__ import annotations` will convert any runtime-required
annotation to a typing-only annotation.
This PR adds separate state to track these runtime-required annotations.
The changes in the test fixtures are correct -- these were false
negatives before.
Closes https://github.com/astral-sh/ruff/issues/5574.
## Summary
I'll write up a more detailed description tomorrow, but in short, this
PR removes our regex-based implementation in favor of "manual" parsing.
I tried a couple different implementations. In the benchmarks below:
- `Directive/Regex` is our implementation on `main`.
- `Directive/Find` just uses `text.find("noqa")`, which is insufficient,
since it doesn't cover case-insensitive variants like `NOQA`, and
doesn't handle multiple `noqa` matches in a single like, like ` # Here's
a noqa comment # noqa: F401`. But it's kind of a baseline.
- `Directive/Memchr` uses three `memchr` iterative finders (one for
`noqa`, `NOQA`, and `NoQA`).
- `Directive/AhoCorasick` is roughly the variant checked-in here.
The raw results:
```
Directive/Regex/# noqa: F401
time: [273.69 ns 274.71 ns 276.03 ns]
change: [+1.4467% +1.8979% +2.4243%] (p = 0.00 < 0.05)
Performance has regressed.
Found 15 outliers among 100 measurements (15.00%)
3 (3.00%) low mild
8 (8.00%) high mild
4 (4.00%) high severe
Directive/Find/# noqa: F401
time: [66.972 ns 67.048 ns 67.132 ns]
change: [+2.8292% +2.9377% +3.0540%] (p = 0.00 < 0.05)
Performance has regressed.
Found 15 outliers among 100 measurements (15.00%)
1 (1.00%) low severe
3 (3.00%) low mild
8 (8.00%) high mild
3 (3.00%) high severe
Directive/AhoCorasick/# noqa: F401
time: [76.922 ns 77.189 ns 77.536 ns]
change: [+0.4265% +0.6862% +0.9871%] (p = 0.00 < 0.05)
Change within noise threshold.
Found 8 outliers among 100 measurements (8.00%)
1 (1.00%) low mild
3 (3.00%) high mild
4 (4.00%) high severe
Directive/Memchr/# noqa: F401
time: [62.627 ns 62.654 ns 62.679 ns]
change: [-0.1780% -0.0887% -0.0120%] (p = 0.03 < 0.05)
Change within noise threshold.
Found 11 outliers among 100 measurements (11.00%)
1 (1.00%) low severe
5 (5.00%) low mild
3 (3.00%) high mild
2 (2.00%) high severe
Directive/Regex/# noqa: F401, F841
time: [321.83 ns 322.39 ns 322.93 ns]
change: [+8602.4% +8623.5% +8644.5%] (p = 0.00 < 0.05)
Performance has regressed.
Found 5 outliers among 100 measurements (5.00%)
1 (1.00%) low severe
2 (2.00%) low mild
1 (1.00%) high mild
1 (1.00%) high severe
Directive/Find/# noqa: F401, F841
time: [78.618 ns 78.758 ns 78.896 ns]
change: [+1.6909% +1.8771% +2.0628%] (p = 0.00 < 0.05)
Performance has regressed.
Found 3 outliers among 100 measurements (3.00%)
3 (3.00%) high mild
Directive/AhoCorasick/# noqa: F401, F841
time: [87.739 ns 88.057 ns 88.468 ns]
change: [+0.1843% +0.4685% +0.7854%] (p = 0.00 < 0.05)
Change within noise threshold.
Found 11 outliers among 100 measurements (11.00%)
5 (5.00%) low mild
3 (3.00%) high mild
3 (3.00%) high severe
Directive/Memchr/# noqa: F401, F841
time: [80.674 ns 80.774 ns 80.860 ns]
change: [-0.7343% -0.5633% -0.4031%] (p = 0.00 < 0.05)
Change within noise threshold.
Found 14 outliers among 100 measurements (14.00%)
4 (4.00%) low severe
9 (9.00%) low mild
1 (1.00%) high mild
Directive/Regex/# noqa time: [194.86 ns 195.93 ns 196.97 ns]
change: [+11973% +12039% +12103%] (p = 0.00 < 0.05)
Performance has regressed.
Found 6 outliers among 100 measurements (6.00%)
5 (5.00%) low mild
1 (1.00%) high mild
Directive/Find/# noqa time: [25.327 ns 25.354 ns 25.383 ns]
change: [+3.8524% +4.0267% +4.1845%] (p = 0.00 < 0.05)
Performance has regressed.
Found 9 outliers among 100 measurements (9.00%)
6 (6.00%) high mild
3 (3.00%) high severe
Directive/AhoCorasick/# noqa
time: [34.267 ns 34.368 ns 34.481 ns]
change: [+0.5646% +0.8505% +1.1281%] (p = 0.00 < 0.05)
Change within noise threshold.
Found 5 outliers among 100 measurements (5.00%)
5 (5.00%) high mild
Directive/Memchr/# noqa time: [21.770 ns 21.818 ns 21.874 ns]
change: [-0.0990% +0.1464% +0.4046%] (p = 0.26 > 0.05)
No change in performance detected.
Found 10 outliers among 100 measurements (10.00%)
4 (4.00%) low mild
4 (4.00%) high mild
2 (2.00%) high severe
Directive/Regex/# type: ignore # noqa: E501
time: [278.76 ns 279.69 ns 280.72 ns]
change: [+7449.4% +7469.8% +7490.5%] (p = 0.00 < 0.05)
Performance has regressed.
Found 3 outliers among 100 measurements (3.00%)
1 (1.00%) low mild
1 (1.00%) high mild
1 (1.00%) high severe
Directive/Find/# type: ignore # noqa: E501
time: [67.791 ns 67.976 ns 68.184 ns]
change: [+2.8321% +3.1735% +3.5418%] (p = 0.00 < 0.05)
Performance has regressed.
Found 6 outliers among 100 measurements (6.00%)
5 (5.00%) high mild
1 (1.00%) high severe
Directive/AhoCorasick/# type: ignore # noqa: E501
time: [75.908 ns 76.055 ns 76.210 ns]
change: [+0.9269% +1.1427% +1.3955%] (p = 0.00 < 0.05)
Change within noise threshold.
Found 1 outliers among 100 measurements (1.00%)
1 (1.00%) high severe
Directive/Memchr/# type: ignore # noqa: E501
time: [72.549 ns 72.723 ns 72.957 ns]
change: [+1.5881% +1.9660% +2.3974%] (p = 0.00 < 0.05)
Performance has regressed.
Found 15 outliers among 100 measurements (15.00%)
10 (10.00%) high mild
5 (5.00%) high severe
Directive/Regex/# type: ignore # nosec
time: [66.967 ns 67.075 ns 67.207 ns]
change: [+1713.0% +1715.8% +1718.9%] (p = 0.00 < 0.05)
Performance has regressed.
Found 10 outliers among 100 measurements (10.00%)
1 (1.00%) low severe
3 (3.00%) low mild
2 (2.00%) high mild
4 (4.00%) high severe
Directive/Find/# type: ignore # nosec
time: [18.505 ns 18.548 ns 18.597 ns]
change: [+1.3520% +1.6976% +2.0333%] (p = 0.00 < 0.05)
Performance has regressed.
Found 4 outliers among 100 measurements (4.00%)
4 (4.00%) high mild
Directive/AhoCorasick/# type: ignore # nosec
time: [16.162 ns 16.206 ns 16.252 ns]
change: [+1.2919% +1.5587% +1.8430%] (p = 0.00 < 0.05)
Performance has regressed.
Found 4 outliers among 100 measurements (4.00%)
3 (3.00%) high mild
1 (1.00%) high severe
Directive/Memchr/# type: ignore # nosec
time: [39.192 ns 39.233 ns 39.276 ns]
change: [+0.5164% +0.7456% +0.9790%] (p = 0.00 < 0.05)
Change within noise threshold.
Found 13 outliers among 100 measurements (13.00%)
2 (2.00%) low severe
4 (4.00%) low mild
3 (3.00%) high mild
4 (4.00%) high severe
Directive/Regex/# some very long comment that # is interspersed with characters but # no directive
time: [81.460 ns 81.578 ns 81.703 ns]
change: [+2093.3% +2098.8% +2104.2%] (p = 0.00 < 0.05)
Performance has regressed.
Found 4 outliers among 100 measurements (4.00%)
2 (2.00%) low mild
2 (2.00%) high mild
Directive/Find/# some very long comment that # is interspersed with characters but # no directive
time: [26.284 ns 26.331 ns 26.387 ns]
change: [+0.7554% +1.1027% +1.3832%] (p = 0.00 < 0.05)
Change within noise threshold.
Found 6 outliers among 100 measurements (6.00%)
5 (5.00%) high mild
1 (1.00%) high severe
Directive/AhoCorasick/# some very long comment that # is interspersed with characters but # no direc...
time: [28.643 ns 28.714 ns 28.787 ns]
change: [+1.3774% +1.6780% +2.0028%] (p = 0.00 < 0.05)
Performance has regressed.
Found 2 outliers among 100 measurements (2.00%)
2 (2.00%) high mild
Directive/Memchr/# some very long comment that # is interspersed with characters but # no directive
time: [55.766 ns 55.831 ns 55.897 ns]
change: [+1.5802% +1.7476% +1.9021%] (p = 0.00 < 0.05)
Performance has regressed.
Found 2 outliers among 100 measurements (2.00%)
2 (2.00%) low mild
```
While memchr is faster than aho-corasick in some of the common cases
(like `# noqa: F401`), the latter is way, way faster when there _isn't_
a match (like 2x faster -- see the last two cases). Since most comments
_aren't_ `noqa` comments, this felt like the right tradeoff. Note that
all implementations are significantly faster than the regex version.
(I know I originally reported a 10x speedup, but I ended up improving
the regex version a bit in some prior PRs, so it got unintentionally
faster via some refactors.)
There's also one behavior change in here, which is that we now allow
variable spaces, e.g., `#noqa` or `# noqa`. Previously, we required
exactly one space. This thus closes#5177.
## Summary
The following code was previously leading to unstable formatting:
```python
try:
try:
pass
finally:
print(1) # issue7208
except A:
pass
```
The comment would be formatted as a trailing comment of `try` which is
unstable as an end-of-line comment gets two extra whitespaces.
This was originally found in
99b00efd5e/Lib/getpass.py (L68-L91)
## Test Plan
I added a regression test
## Summary
It's a bit simpler to let the API just take the text itself, plus an
offset (to make the returned `TextRange` absolute, rather than
relative).
Charlie Marsh
skykasko
eggplants
Justin Prieto
Micha Reiser
Dhruv Manilawala
Tom Kuson
konsti
Zanie
qdegraaf
David Szotten
Chris Pryer
Louis Dispa
monosans
Evan Rittenhouse
Harutaka Kawamura
Aarni Koskela
Dimitri Papadopoulos Orfanos
Peter Attia