If a builtin is conditionally shadowed by a global, we didn't correctly
fall back to builtins for the not-defined-in-globals path (see added
test for an example.)
List and set comprehensions using `async for` cannot be replaced with
underlying generators; this PR modifies C419 to skip such
comprehensions.
Closes#12891.
## Summary
Occasionally, we receive bug reports that imports in `src` directories
aren't correctly detected. The root of the problem is that we default to
`src = ["."]`, so users have to set `src = ["src"]` explicitly. This PR
extends the default to cover _both_ of them: `src = [".", "src"]`.
Closes https://github.com/astral-sh/ruff/issues/12454.
## Test Plan
I replicated the structure described in
https://github.com/astral-sh/ruff/issues/12453, and verified that the
imports were considered sorted, but that adding `src = ["."]` showed an
error.
## Summary
This PR adds very basic support for using the line / column information
from the diagnostic message. This makes it easier to validate
diagnostics in an editor as oppose to going through the diff one
diagnostic at a time and confirming it at the location.
## Summary
This PR adds a fallback logic for `is_python_notebook` to check the
`kernelspec.language` field.
Reference implementation in VS Code:
1c31e75898/extensions/ipynb/src/deserializers.ts (L20-L22)
It's also required for the kernel to provide the `language` they're
implementing based on
https://jupyter-client.readthedocs.io/en/stable/kernels.html#kernel-specs
reference although that's for the `kernel.json` file but is also
included in the notebook metadata.
Closes: #12281
## Test Plan
Add a test case for `is_python_notebook` and include the test notebook
for round trip validation.
The test notebook contains two cells, one is JavaScript (denoted via the
`vscode.languageId` metadata) and the other is Python (no metadata). The
notebook metadata only contains `kernelspec` and the `language_info` is
absent.
I also verified that this is a valid notebook by opening it in Jupyter
Lab, VS Code and using `nbformat` validator.
## Summary
This PR adds support for VS Code specific cell metadata to consider when
collecting valid code cells.
For context, Ruff only runs on valid code cells. These are the code
cells that doesn't contain cell magics. Previously, Ruff only used the
notebook's metadata to determine whether it's a Python notebook. But, in
VS Code, a notebook's preferred language might be Python but it could
still contain code cells for other languages. This can be determined
with the `metadata.vscode.languageId` field.
### References:
* https://code.visualstudio.com/docs/languages/identifiers
* e6c009a3d4/extensions/ipynb/src/serializers.ts (L104-L107)
*
e6c009a3d4/extensions/ipynb/src/serializers.ts (L117-L122)
This brings us one step closer to fixing #12281.
## Test Plan
Add test cases for `is_valid_python_code_cell` and an integration test
case which showcase running it end to end. The test notebook contains a
JavaScript code cell and a Python code cell.
## Summary
This PR fixes a bug in the semantic model where it would evaluate the
default parameter value in the type parameter scope. For example,
```py
def foo[T1: int](a = T1):
pass
```
Here, the `T1` in `a = T1` is undefined but Ruff doesn't flag it
(https://play.ruff.rs/ba2f7c2f-4da6-417e-aa2a-104aa63e6d5e).
The fix here is to evaluate the default parameter value in the
_enclosing_ scope instead.
## Test Plan
Add a test case which includes the above code under `F821`
(`undefined-name`) and validate the snapshot.
## Summary
See #12703. This only addresses the first bullet point, adding a space
after the comma in the suggested fix from list/tuple to string.
## Test Plan
Updated the snapshots and compared.
## Summary
This PR adds scope and definition for comprehension nodes. This includes
the following nodes:
* List comprehension
* Dictionary comprehension
* Set comprehension
* Generator expression
### Scope
Each expression here adds it's own scope with one caveat - the `iter`
expression of the first generator is part of the parent scope. For
example, in the following code snippet the `iter1` variable is evaluated
in the outer scope.
```py
[x for x in iter1]
```
> The iterable expression in the leftmost for clause is evaluated
directly in the enclosing scope and then passed as an argument to the
implicitly nested scope.
>
> Reference:
https://docs.python.org/3/reference/expressions.html#displays-for-lists-sets-and-dictionaries
There's another special case for assignment expressions:
> There is one special case: an assignment expression occurring in a
list, set or dict comprehension or in a generator expression (below
collectively referred to as “comprehensions”) binds the target in the
containing scope, honoring a nonlocal or global declaration for the
target in that scope, if one exists.
>
> Reference: https://peps.python.org/pep-0572/#scope-of-the-target
For example, in the following code snippet, the variables `a` and `b`
are available after the comprehension while `x` isn't:
```py
[a := 1 for x in range(2) if (b := 2)]
```
### Definition
Each comprehension node adds a single definition, the "target" variable
(`[_ for target in iter]`). This has been accounted for and a new
variant has been added to `DefinitionKind`.
### Type Inference
Currently, type inference is limited to a single scope. It doesn't
_enter_ in another scope to infer the types of the remaining expressions
of a node. To accommodate this, the type inference for a **scope**
requires new methods which _doesn't_ infer the type of the `iter`
expression of the leftmost outer generator (that's defined in the
enclosing scope).
The type inference for the scope region is split into two parts:
* `infer_generator_expression` (similarly for comprehensions) infers the
type of the `iter` expression of the leftmost outer generator
* `infer_generator_expression_scope` (similarly for comprehension)
infers the type of the remaining expressions except for the one
mentioned in the previous point
The type inference for the **definition** also needs to account for this
special case of leftmost generator. This is done by defining a `first`
boolean parameter which indicates whether this comprehension definition
occurs first in the enclosing expression.
## Test Plan
New test cases were added to validate multiple scenarios. Refer to the
documentation for each test case which explains what is being tested.
Make `cargo doc -p red_knot_python_semantic --document-private-items`
run warning-free. I'd still like to do this for all of ruff and start
enforcing it in CI (https://github.com/astral-sh/ruff/issues/12372) but
haven't gotten to it yet. But in the meantime I'm trying to maintain it
for at least `red_knot_python_semantic`, as it helps to ensure our doc
comments stay up to date.
A few of the comments I just removed or shortened, as their continued
relevance wasn't clear to me; please object in review if you think some
of them are important to keep!
Also remove a no-longer-needed `allow` attribute.
For type narrowing, we'll need intersections (since applying type
narrowing is just a type intersection.)
Add `IntersectionBuilder`, along with some tests for it and
`UnionBuilder` (renamed from `UnionTypeBuilder`).
We use smart builders to ensure that we always keep these types in
disjunctive normal form (DNF). That means that we never have deeply
nested trees of unions and intersections: unions flatten into unions,
intersections flatten into intersections, and intersections distribute
over unions, so the most complex tree we can ever have is a union of
intersections. We also never have a single-element union or a
single-positive-element intersection; these both just simplify to the
contained type.
Maintaining these invariants means that `UnionBuilder` doesn't
necessarily end up building a `Type::Union` (e.g. if you only add a
single type to the union, it'll just return that type instead), and
`IntersectionBuilder` doesn't necessarily build a `Type::Intersection`
(if you add a union to the intersection, we distribute the intersection
over that union, and `IntersectionBuilder` will end up returning a
`Type::Union` of intersections).
We also simplify intersections by ensuring that if a type and its
negation are both in an intersection, they simplify out. (In future this
should also respect subtyping, not just type identity, but we don't have
subtyping yet.) We do implement subtyping of `Never` as a special case
for now.
Most of this PR is unused for now until type narrowing lands; I'm just
breaking it out to reduce the review fatigue of a single massive PR.
## Summary
I'm not sure if this is useful but this is a hacky implementation to add
the filename and row / column numbers to the current Red Knot
diagnostics.
## Summary
Related to https://github.com/astral-sh/ruff-vscode/issues/571, this PR
updates the settings index builder to trace all the errors it
encountered. Without this, there's no way for user to know that
something failed and some of the capability might not work as expected.
For example, in the linked PR, the settings were invalid which means
notebooks weren't included and there were no log messages for it.
## Test Plan
Create an invalid `ruff.toml` file:
```toml
[tool.ruff]
extend-exclude = ["*.ipynb"]
```
Logs:
```
2024-08-12 18:33:09.873 [info] [Trace - 6:33:09 PM] 12.217043000s ERROR ruff:main ruff_server::session::index::ruff_settings: Failed to parse /Users/dhruv/playground/ruff/pyproject.toml
```
Notification Preview:
<img width="483" alt="Screenshot 2024-08-12 at 18 33 20"
src="https://github.com/user-attachments/assets/a4f303e5-f073-454f-bdcd-ba6af511e232">
Another way to trigger is to provide an invalid `cache-dir` value:
```toml
[tool.ruff]
cache-dir = "$UNKNOWN"
```
Same notification preview but different log message:
```
2024-08-12 18:41:37.571 [info] [Trace - 6:41:37 PM] 21.700112208s ERROR ThreadId(30) ruff_server::session::index::ruff_settings: Error while resolving settings from /Users/dhruv/playground/ruff/pyproject.toml: Invalid `cache-dir` value: error looking key 'UNKNOWN' up: environment variable not found
```
With multiple `pyproject.toml` file:
```
2024-08-12 18:41:15.887 [info] [Trace - 6:41:15 PM] 0.016636833s ERROR ThreadId(04) ruff_server::session::index::ruff_settings: Error while resolving settings from /Users/dhruv/playground/ruff/pyproject.toml: Invalid `cache-dir` value: error looking key 'UNKNOWN' up: environment variable not found
2024-08-12 18:41:15.888 [info] [Trace - 6:41:15 PM] 0.017378833s ERROR ThreadId(13) ruff_server::session::index::ruff_settings: Failed to parse /Users/dhruv/playground/ruff/tools/pyproject.toml
```
In most cases we should suggest a ternary operator, but there are three
edge cases where a binary operator is more appropriate.
Given an if-else block of the form
```python
if test:
target_var = body_value
else:
target_var = else_value
```
This PR updates the check for SIM108 to the following:
- If `test == body_value` and preview enabled, suggest to replace with
`target_var = test or else_value`
- If `test == not body_value` and preview enabled, suggest to replace
with `target_var = body_value and else_value`
- If `not test == body_value` and preview enabled, suggest to replace
with `target_var = body_value and else_value`
- Otherwise, suggest to replace with `target_var = body_value if test
else else_value`
Closes#12189.
## Summary
Adding parentheses to a tuple in a subscript with elements that include
slice expressions causes a syntax error. For example, `d[(1,2,:)]` is a
syntax error.
So, when `lint.ruff.parenthesize-tuple-in-subscript = true` and the
tuple includes a slice expression, we skip this check and fix.
Closes#12766.
> ~Builtins are also more efficient than `for` loops.~
Let's not promise performance because this code transformation does not
deliver.
Benchmark written by @dcbaker
> `any()` seems to be about 1/3 as fast (Python 3.11.9, NixOS):
```python
loop = 'abcdef'.split()
found = 'f'
nfound = 'g'
def test1():
for x in loop:
if x == found:
return True
return False
def test2():
return any(x == found for x in loop)
def test3():
for x in loop:
if x == nfound:
return True
return False
def test4():
return any(x == nfound for x in loop)
if __name__ == "__main__":
import timeit
print('for loop (found) :', timeit.timeit(test1))
print('for loop (not found):', timeit.timeit(test3))
print('any() (found) :', timeit.timeit(test2))
print('any() (not found) :', timeit.timeit(test4))
```
```
for loop (found) : 0.051076093994197436
for loop (not found): 0.04388196699437685
any() (found) : 0.15422860698890872
any() (not found) : 0.15568504799739458
```
I have retested with longer lists and on multiple Python versions with
similar results.
Implements the new fixable lint rule `RUF031` which checks for the use or omission of parentheses around tuples in subscripts, depending on the setting `lint.ruff.parenthesize-tuple-in-getitem`. By default, the use of parentheses is considered a violation.
## Summary
Follow-up from https://github.com/astral-sh/ruff/pull/12725, this is
just a small refactor to use a wrapper struct instead of type alias for
workspace settings index. This avoids the need to have the
`register_workspace_settings` as a static method on `Index` and instead
is a method on the new struct itself.
## Summary
This PR updates the server to ignore non-file workspace URL.
This is to avoid crashing the server if the URL scheme is not "file".
We'd still raise an error if the URL to file path conversion fails.
Also, as per the docs of
[`to_file_path`](https://docs.rs/url/2.5.2/url/struct.Url.html#method.to_file_path):
> Note: This does not actually check the URL’s scheme, and may give
nonsensical results for other schemes. It is the user’s responsibility
to check the URL’s scheme before calling this.
resolves: #12660
## Test Plan
I'm not sure how to test this locally but the change is small enough to
validate on its own.
## Summary
This PR updates the `red_knot` CLI to make the subcommand optional.
## Test Plan
Run the following commands:
* `cargo run --bin red_knot --
--current-directory=~/playground/ruff/type_inference` (no subcommand
requirement)
* `cargo run --bin red_knot -- server` (should start the server)
<!--
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
Resolves#12636
Consider docstrings which begin with the word "Returns" as having
satisfactorily documented they're returns. For example
```python
def f():
"""Returns 1."""
return 1
```
is valid.
## Test Plan
Added example to test fixture.
---------
Co-authored-by: Dhruv Manilawala <dhruvmanila@gmail.com>
## Summary
Removes set comprehension as a violation for `sum` when checking `C419`,
because set comprehension may de-duplicate entries in a generator,
thereby modifying the value of the sum.
Closes#12690.
## Summary
Make it a violation of `C409` to call `tuple` with a list or set
comprehension, and
implement the (unsafe) fix of calling the `tuple` with the underlying
generator instead.
Closes#12648.
## Test Plan
Test fixture updated, cargo test, docs checked for updated description.
## Summary
Adds autofix for `RUF007`
## Test Plan
`cargo test`, however I get errors for `test resolver::tests::symlink
... FAILED` which seems to not be my fault
## Summary
Fixes#12630.
DOC501 and DOC502 now understand functions with constructs like this to
be explicitly raising `TypeError` (which should be documented in a
function's docstring):
```py
try:
foo():
except TypeError:
...
raise
```
I made an exception for `Exception` and `BaseException`, however.
Constructs like this are reasonably common, and I don't think anybody
would say that it's worth putting in the docstring that it raises "some
kind of generic exception":
```py
try:
foo()
except BaseException:
do_some_logging()
raise
```
## Test Plan
`cargo test -p ruff_linter --lib`
## Summary
Please see
https://github.com/astral-sh/ruff/pull/12605#discussion_r1699957443 for
a description of the issue.
They way I fixed it is to get the *last* timeout item in the `with`, and
if it's an `async with` and there are items after it, then don't trigger
the lint.
## Test Plan
Updated the fixture with some more cases.
Changes the red-knot benchmark to run on the stdlib "tomllib" library
(which is self-contained, four files, uses type annotations) instead of
on very small bits of handwritten code.
Also remove the `without_parse` benchmark: now that we are running on
real code that uses typeshed, we'd either have to pre-parse all of
typeshed (slow) or find some way to determine which typeshed modules
will be used by the benchmark (not feasible with reasonable complexity.)
## Test Plan
`cargo bench -p ruff_benchmark --bench red_knot`
## Summary
This PR separates the current `red_knot` crate into two crates:
1. `red_knot` - This will be similar to the `ruff` crate, it'll act as
the CLI crate
2. `red_knot_workspace` - This includes everything except for the CLI
functionality from the existing `red_knot` crate
Note that the code related to the file watcher is in
`red_knot_workspace` for now but might be required to extract it out in
the future.
The main motivation for this change is so that we can have a `red_knot
server` command. This makes it easier to test the server out without
making any changes in the VS Code extension. All we need is to specify
the `red_knot` executable path in `ruff.path` extension setting.
## Test Plan
- `cargo build`
- `cargo clippy --workspace --all-targets --all-features`
- `cargo shear --fix`
## Summary
There's still a problem here. Given:
```python
class Class():
pass
# comment
# another comment
a = 1
```
We only add one newline before `a = 1` on the first pass, because
`max_precedling_blank_lines` is 1... We then add the second newline on
the second pass, so it ends up in the right state, but the logic is
clearly wonky.
Closes https://github.com/astral-sh/ruff/issues/11508.
I hit this `todo!` trying to run type inference over some real modules.
Since it's a one-liner to implement it, I just did that rather than
changing to `Type::Unknown`.
## Summary
@zanieb noticed while we were discussing #12595 that this flag is now
unnecessary, so remove it and the flags which reference it.
## Test Plan
Question for maintainers: is there a test to add *or* remove here? (I’ve
opened this as a draft PR with that in view!)
## Summary
This pull request adds support for logging via `$/logTrace` RPC
messages. It also enables that code path for when a client is Zed editor
or VS Code (as there's no way for us to generically tell whether a client prefers
`$/logTrace` over stderr.
Related to: #12523
## Test Plan
I've built Ruff from this branch and tested it manually with Zed.
---------
Co-authored-by: Dhruv Manilawala <dhruvmanila@gmail.com>
## Summary
<!-- What's the purpose of the change? What does it do, and why? -->
Extend `flake8-builtins` to imports, lambda-arguments, and modules to be
consistent with original checker
[flake8_builtins](https://github.com/gforcada/flake8-builtins/blob/main/flake8_builtins.py).
closes#12540
## Details
- Implement builtin-import-shadowing (A004)
- Stop tracking imports shadowing in builtin-variable-shadowing (A001)
in preview mode.
- Implement builtin-lambda-argument-shadowing (A005)
- Implement builtin-module-shadowing (A006)
- Add new option `linter.flake8_builtins.builtins_allowed_modules`
## Test Plan
cargo test
## Summary
If an import is marked as "required", we should never flag it as unused.
In practice, this is rare, since required imports are typically used for
`__future__` annotations, which are always considered "used".
Closes https://github.com/astral-sh/ruff/issues/12458.
Now that we have builtins available, resolve some simple cases to the
right builtin type.
We should also adjust the display for types to include their module
name; that's not done yet here.
## Summary
This PR adds support for untitled files in the Red Knot project.
Refer to the [design
discussion](https://github.com/astral-sh/ruff/discussions/12336) for
more details.
### Changes
* The `parsed_module` always assumes that the `SystemVirtual` path is of
`PySourceType::Python`.
* For the module resolver, as suggested, I went ahead by adding a new
`SystemOrVendoredPath` enum and renamed `FilePathRef` to
`SystemOrVendoredPathRef` (happy to consider better names here).
* The `file_to_module` query would return if it's a
`FilePath::SystemVirtual` variant because a virtual file doesn't belong
to any module.
* The sync implementation for the system virtual path is basically the
same as that of system path except that it uses the
`virtual_path_metadata`. The reason for this is that the system
(language server) would provide the metadata on whether it still exists
or not and if it exists, the corresponding metadata.
For point (1), VS Code would use `Untitled-1` for Python files and
`Untitled-1.ipynb` for Jupyter Notebooks. We could use this distinction
to determine whether the source type is `Python` or `Ipynb`.
## Test Plan
Added test cases in #12526
Extend red-knot type inference to cover all syntax, so that inferring
types for a scope gives all expressions a type. This means we can run
the red-knot semantic lint on all Python code without panics. It also
means we can infer types for `builtins.pyi` without panics.
To keep things simple, this PR intentionally doesn't add any new type
inference capabilities: the expanded coverage is all achieved with
`Type::Unknown`. But this puts the skeleton in place for adding better
inference of all these language features.
I also had to add basic Salsa cycle recovery (with just `Type::Unknown`
for now), because some `builtins.pyi` definitions are cyclic.
To test this, I added a comprehensive corpus of test snippets sourced
from Cinder under [MIT
license](https://github.com/facebookincubator/cinder/blob/cinder/3.10/cinderx/LICENSE),
which matches Ruff's license. I also added to this corpus some
additional snippets for newer language features: all the
`27_func_generic_*` and `73_class_generic_*` files, as well as
`20_lambda_default_arg.py`, and added a test which runs semantic-lint
over all these files. (The test doesn't assert the test-corpus files are
lint-free; just that they are able to lint without a panic.)
## Summary
Right now, in the isort comment model, there's nowhere for trailing
comments on the _statement_ to go, as in:
```python
from mylib import (
MyClient,
MyMgmtClient,
) # some comment
```
If the comment is on the _alias_, we do preserve it, because we attach
it to the alias, as in:
```python
from mylib import (
MyClient,
MyMgmtClient, # some comment
)
```
Similarly, if the comment is trailing on an import statement
(non-`from`), we again attach it to the alias, because it can't be
parenthesized, as in:
```python
import foo # some comment
```
This PR adds logic to track and preserve those trailing comments.
We also no longer drop several other comments, like:
```python
from mylib import (
# some comment
MyClient
)
```
Closes https://github.com/astral-sh/ruff/issues/12487.
## Summary
When working on improving Ruff integration with Zed I noticed that it
errors out when we try to resolve a code action of a `QUICKFIX` kind;
apparently, per @dhruvmanila we shouldn't need to resolve it, as the
edit is provided in the initial response for the code action. However,
it's possible for the `resolve` call to fill out other fields (such as
`command`).
AFAICT Helix also tries to resolve the code actions unconditionally (as
in, when either `edit` or `command` is absent); so does VSC. They can
still apply the quickfixes though, as they do not error out on a failed
call to resolve code actions - Zed does. Following suit on Zed's side
does not cut it though, as we still get a log request from Ruff for that
failure (which is surfaced in the UI).
There are also other language servers (such as
[rust-analyzer](c1c9e10f72/crates/rust-analyzer/src/handlers/request.rs (L1257)))
that fill out both `command` and `edit` fields as a part of code action
resolution.
This PR makes the resolve calls for quickfix actions return the input
value.
## Test Plan
N/A
Add support for while-loop control flow.
This doesn't yet include general support for terminals and reachability;
that is wider than just while loops and belongs in its own PR.
This also doesn't yet add support for cyclic definitions in loops; that
comes with enough of its own complexity in Salsa that I want to handle
it separately.
Add a lint rule to detect if a name is definitely or possibly undefined
at a given usage.
If I create the file `undef/main.py` with contents:
```python
x = int
def foo():
z
return x
if flag:
y = x
y
```
And then run `cargo run --bin red_knot -- --current-directory
../ruff-examples/undef`, I get the output:
```
Name 'z' used when not defined.
Name 'flag' used when not defined.
Name 'y' used when possibly not defined.
```
If I modify the file to add `y = 0` at the top, red-knot re-checks it
and I get the new output:
```
Name 'z' used when not defined.
Name 'flag' used when not defined.
```
Note that `int` is not flagged, since it's a builtin, and `return x` in
the function scope is not flagged, since it refers to the global `x`.
## Summary
This PR fixes a bug to raise a syntax error when an unparenthesized
generator expression is used as an argument to a call when there are
more than one argument.
For reference, the grammar is:
```
primary:
| ...
| primary genexp
| primary '(' [arguments] ')'
| ...
genexp:
| '(' ( assignment_expression | expression !':=') for_if_clauses ')'
```
The `genexp` requires the parenthesis as mentioned in the grammar. So,
the grammar for a call expression is either a name followed by a
generator expression or a name followed by a list of argument. In the
former case, the parenthesis are excluded because the generator
expression provides them while in the later case, the parenthesis are
explicitly provided for a list of arguments which means that the
generator expression requires it's own parenthesis.
This was discovered in https://github.com/astral-sh/ruff/issues/12420.
## Test Plan
Add test cases for valid and invalid syntax.
Make sure that the parser from CPython also raises this at the parsing
step:
```console
$ python3.13 -m ast parser/_.py
File "parser/_.py", line 1
total(1, 2, x for x in range(5), 6)
^^^^^^^^^^^^^^^^^^^
SyntaxError: Generator expression must be parenthesized
$ python3.13 -m ast parser/_.py
File "parser/_.py", line 1
sum(x for x in range(10), 10)
^^^^^^^^^^^^^^^^^^^^
SyntaxError: Generator expression must be parenthesized
```
## Summary
Fix panic reported in #12428. Where a string would sometimes get split
within a character boundary. This bypasses the need to split the string.
This does not guarantee the correct formatting of the docstring, but
neither did the previous implementation.
Resolves#12428
## Test Plan
Test case added to fixture
## Summary
These are the first rules implemented as part of #458, but I plan to
implement more.
Specifically, this implements `docstring-missing-exception` which checks
for raised exceptions not documented in the docstring, and
`docstring-extraneous-exception` which checks for exceptions in the
docstring not present in the body.
## Test Plan
Test fixtures added for both google and numpy style.
When poring over traces, the ones that just include a definition or
symbol or expression ID aren't very useful, because you don't know which
file it comes from. This adds that information to the trace.
I guess the downside here is that if calling `.file(db)` on a
scope/definition/expression would execute other traced code, it would be
marked as outside the span? I don't think that's a concern, because I
don't think a simple field access on a tracked struct should ever
execute our code. If I'm wrong and this is a problem, it seems like the
tracing crate has this feature where you can record a field as
`tracing::field::Empty` and then fill in its value later with
`span.record(...)`, but when I tried this it wasn't working for me, not
sure why.
I think there's a lot more we can do to make our tracing output more
useful for debugging (e.g. record an event whenever a
definition/symbol/expression/use id is created with the details of that
definition/symbol/expression/use), this is just dipping my toes in the
water.
<!--
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? -->
This PR updates D301 rule to allow inclduing escaped docstring, e.g.
`\"""Foo.\"""` or `\"\"\"Bar.\"\"\"`, within a docstring.
Related issue: #12152
## Test Plan
Add more test cases to D301.py and update the snapshot file.
<!-- How was it tested? -->
In preparation for supporting resolving builtins, simplify the benchmark
so it doesn't look up `str`, which is actually a complex builtin to deal
with because it inherits `Sequence[str]`.
Co-authored-by: Alex Waygood <alex.waygood@gmail.com>
Per comments in https://github.com/astral-sh/ruff/pull/12269, "module
global" is kind of long, and arguably redundant.
I tried just using "module" but there were too many cases where I felt
this was ambiguous. I like the way "global" works out better, though it
does require an understanding that in Python "global" generally means
"module global" not "globally global" (though in a sense module globals
are also globally global since modules are singletons).
Support falling back to a global name lookup if a name isn't defined in
the local scope, in the cases where that is correct according to Python
semantics.
In class scopes, a name lookup checks the local namespace first, and if
the name isn't found there, looks it up in globals.
In function scopes (and type parameter scopes, which are function-like),
if a name has any definitions in the local scope, it is a local, and
accessing it when none of those definitions have executed yet just
results in an `UnboundLocalError`, it does not fall back to a global. If
the name does not have any definitions in the local scope, then it is an
implicit global.
Public symbol type lookups never include such a fall back. For example,
if a name is not defined in a class scope, it is not available as a
member on that class, even if a name lookup within the class scope would
have fallen back to a global lookup.
This PR makes the `@override` lint rule work again.
Not yet included/supported in this PR:
* Support for free variables / closures: a free symbol in a nested
function-like scope referring to a symbol in an outer function-like
scope.
* Support for `global` and `nonlocal` statements, which force a symbol
to be treated as global or nonlocal even if it has definitions in the
local scope.
* Module-global lookups should fall back to builtins if the name isn't
found in the module scope.
I would like to expose nicer APIs for the various kinds of symbols
(explicit global, implicit global, free, etc), but this will also wait
for a later PR, when more kinds of symbols are supported.
Adds inference tests sufficient to give full test coverage of the
`UseDefMapBuilder::merge` method.
In the process I realized that we could implement visiting of if
statements in `SemanticBuilder` with fewer `snapshot`, `restore`, and
`merge` operations, so I restructured that visit a bit.
I also found one correctness bug in the `merge` method (it failed to
extend the given snapshot with "unbound" for any missing symbols,
meaning we would just lose the fact that the symbol could be unbound in
the merged-in path), and two efficiency bugs (if one of the ranges to
merge is empty, we can just use the other one, no need for copies, and
if the ranges are overlapping -- which can occur with nested branches --
we can still just merge them with no copies), and fixed all three.
## Summary
This PR allows us to fix both expressions in `foo == "a" or foo == "b"
or ("c" != bar and "d" != bar)`, but limits the rule to consecutive
comparisons, following https://github.com/astral-sh/ruff/issues/7797.
I think this logic was _probably_ added because of
https://github.com/astral-sh/ruff/pull/12368 -- the intent being that
we'd replace the _entire_ expression.
## Summary
This PR adds documentation for the Ruff language server.
It mainly does the following:
1. Combines various READMEs containing instructions for different editor
setup in their respective section on the online docs
2. Provide an enumerated list of server settings. Additionally, it also
provides a section for VS Code specific options.
3. Adds a "Features" section which enumerates all the current
capabilities of the native server
For (2), the settings documentation is done manually but a future
improvement (easier after `ruff-lsp` is deprecated) is to move the docs
in to Rust struct and generate the documentation from the code itself.
And, the VS Code extension specific options can be generated by diffing
against the `package.json` in `ruff-vscode` repository.
### Structure
1. Setup: This section contains the configuration for setting up the
language server for different editors
2. Features: This section contains a list of capabilities provided by
the server along with short GIF to showcase it
3. Settings: This section contains an enumerated list of settings in a
similar format to the one for the linter / formatter
4. Migrating from `ruff-lsp`
> [!NOTE]
>
> The settings page is manually written but could possibly be
auto-generated via a macro similar to `OptionsMetadata` on the
`ClientSettings` struct
resolves: #11217
## Test Plan
Generate and open the documentation locally using:
1. `python scripts/generate_mkdocs.py`
2. `mkdocs serve -f mkdocs.insiders.yml`
## Summary
This PR removes the requirement of `--preview` flag to run the `ruff
server` and instead considers it to be an indicator to turn on preview
mode for the linter and the formatter.
resolves: #12161
## Test Plan
Add test cases to assert the `preview` value is updated accordingly.
In an editor context, I used the local `ruff` executable in Neovim with
the `--preview` flag and verified that the preview-only violations are
being highlighted.
Running with:
```lua
require('lspconfig').ruff.setup({
cmd = {
'/Users/dhruv/work/astral/ruff/target/debug/ruff',
'server',
'--preview',
},
})
```
The screenshot shows that `E502` is highlighted with the below config in
`pyproject.toml`:
<img width="877" alt="Screenshot 2024-07-17 at 16 43 09"
src="https://github.com/user-attachments/assets/c7016ef3-55b1-4a14-bbd3-a07b1bcdd323">
## Summary
This PR updates the settings index building logic in the language server
to consider the fallback settings for applying ignore filters in
`WalkBuilder` and the exclusion via `exclude` / `extend-exclude`.
This flow matches the one in the `ruff` CLI where the root settings is
built by (1) finding the workspace setting in the ancestor directory (2)
finding the user configuration if that's missing and (3) fallback to
using the default configuration.
Previously, the index building logic was being executed before (2) and
(3). This PR reverses the logic so that the exclusion /
`respect_gitignore` is being considered from the default settings if
there's no workspace / user settings. This has the benefit that the
server no longer enters the `.git` directory or any other excluded
directory when a user opens a file in the home directory.
Related to #11366
## Test plan
Opened a test file from the home directory and confirmed with the debug
trace (removed in #12360) that the server excludes the `.git` directory
when indexing.
## Summary
Add new rule and implement for `unnecessary default type arguments`
under the `UP` category (`UP043`).
```py
// < py313
Generator[int, None, None]
// >= py313
Generator[int]
```
I think that as Python 3.13 develops, there might be more default type
arguments added besides `Generator` and `AsyncGenerator`. So, I made
this more flexible to accommodate future changes.
related issue: #12286
## Test Plan
snapshot included..!
## Summary
Pretty sure this should still be an error, but also, I think I added
this because of ecosystem CI? So want to see what pops up.
Closes https://github.com/astral-sh/ruff/issues/12164.
## Summary
This is the _intended_ default that PEP 597 _wants_, but it's not
backwards compatible. The fix is already unsafe, so it's better for us
to recommend the desired and expected behavior.
Closes https://github.com/astral-sh/ruff/issues/12069.
Improve semantic index tests with better assertions than just `.len()`,
and re-add use-definition test that was commented out in the switch to
Salsa initially.
Implements definition-level type inference, with basic control flow
(only if statements and if expressions so far) in Salsa.
There are a couple key ideas here:
1) We can do type inference queries at any of three region
granularities: an entire scope, a single definition, or a single
expression. These are represented by the `InferenceRegion` enum, and the
entry points are the salsa queries `infer_scope_types`,
`infer_definition_types`, and `infer_expression_types`. Generally
per-scope will be used for scopes that we are directly checking and
per-definition will be used anytime we are looking up symbol types from
another module/scope. Per-expression should be uncommon: used only for
the RHS of an unpacking or multi-target assignment (to avoid
re-inferring the RHS once per symbol defined in the assignment) and for
test nodes in type narrowing (e.g. the `test` of an `If` node). All
three queries return a `TypeInference` with a map of types for all
definitions and expressions within their region. If you do e.g.
scope-level inference, when it hits a definition, or an
independently-inferable expression, it should use the relevant query
(which may already be cached) to get all types within the smaller
region. This avoids double-inferring smaller regions, even though larger
regions encompass smaller ones.
2) Instead of building a control-flow graph and lazily traversing it to
find definitions which reach a use of a name (which is O(n^2) in the
worst case), instead semantic indexing builds a use-def map, where every
use of a name knows which definitions can reach that use. We also no
longer track all definitions of a symbol in the symbol itself; instead
the use-def map also records which defs remain visible at the end of the
scope, and considers these the publicly-visible definitions of the
symbol (see below).
Major items left as TODOs in this PR, to be done in follow-up PRs:
1) Free/global references aren't supported yet (only lookup based on
definitions in current scope), which means the override-check example
doesn't currently work. This is the first thing I'll fix as follow-up to
this PR.
2) Control flow outside of if statements and expressions.
3) Type narrowing.
There are also some smaller relevant changes here:
1) Eliminate `Option` in the return type of member lookups; instead
always return `Type::Unbound` for a name we can't find. Also use
`Type::Unbound` for modules we can't resolve (not 100% sure about this
one yet.)
2) Eliminate the use of the terms "public" and "root" to refer to
module-global scope or symbols. Instead consistently use the term
"module-global". It's longer, but it's the clearest, and the most
consistent with typical Python terminology. In particular I don't like
"public" for this use because it has other implications around author
intent (is an underscore-prefixed module-global symbol "public"?). And
"root" is just not commonly used for this in Python.
3) Eliminate the `PublicSymbol` Salsa ingredient. Many non-module-global
symbols can also be seen from other scopes (e.g. by a free var in a
nested scope, or by class attribute access), and thus need to have a
"public type" (that is, the type not as seen from a particular use in
the control flow of the same scope, but the type as seen from some other
scope.) So all symbols need to have a "public type" (here I want to keep
the use of the term "public", unless someone has a better term to
suggest -- since it's "public type of a symbol" and not "public symbol"
the confusion with e.g. initial underscores is less of an issue.) At
least initially, I would like to try not having special handling for
module-global symbols vs other symbols.
4) Switch to using "definitions that reach end of scope" rather than
"all definitions" in determining the public type of a symbol. I'm
convinced that in general this is the right way to go. We may want to
refine this further in future for some free-variable cases, but it can
be changed purely by making changes to the building of the use-def map
(the `public_definitions` index in it), without affecting any other
code. One consequence of combining this with no control-flow support
(just last-definition-wins) is that some inference tests now give more
wrong-looking results; I left TODO comments on these tests to fix them
when control flow is added.
And some potential areas for consideration in the future:
1) Should `symbol_ty` be a Salsa query? This would require making all
symbols a Salsa ingredient, and tracking even more dependencies. But it
would save some repeated reconstruction of unions, for symbols with
multiple public definitions. For now I'm not making it a query, but open
to changing this in future with actual perf evidence that it's better.
## Summary
I believe these should always bind more tightly -- e.g., in:
```python
for _ in bar(baz for foo in [1]):
pass
```
The inner `baz` and `foo` should be considered comprehension variables,
not for loop bindings.
We need to revisit this more holistically. In some of these cases,
`BindingKind` should probably be a flag, not an enum, since the values
aren't mutually exclusive. Separately, we should probably be more
precise in how we set it (e.g., by passing down from the parent rather
than sniffing in `handle_node_store`).
Closes https://github.com/astral-sh/ruff/issues/12339
When there is a function or class definition at the end of a suite
followed by the beginning of an alternative block, we have to insert a
single empty line between them.
In the if-else-statement example below, we insert an empty line after
the `foo` in the if-block, but none after the else-block `foo`, since in
the latter case the enclosing suite already adds empty lines.
```python
if sys.version_info >= (3, 10):
def foo():
return "new"
else:
def foo():
return "old"
class Bar:
pass
```
To do so, we track whether the current suite is the last one in the
current statement with a new option on the suite kind.
Fixes#12199
---------
Co-authored-by: Micha Reiser <micha@reiser.io>
## Summary
This PR updates the server to build the settings index in parallel using
similar logic as `python_files_in_path`.
This should help with https://github.com/astral-sh/ruff/issues/11366 but
ideally we would want to build it lazily.
## Test Plan
`cargo insta test`
## Summary
I don't know that there's more to do here. We could consider not raising
the violation at all for arguments, but that would have some false
negatives and could also be surprising to users.
Closes https://github.com/astral-sh/ruff/issues/12267.
## Summary
Ensures that, e.g., the following is not considered a
redefinition-without-use:
```python
import contextlib
foo = None
with contextlib.suppress(ImportError):
from some_module import foo
```
Closes https://github.com/astral-sh/ruff/issues/12309.
## Summary
Closes https://github.com/astral-sh/ruff/issues/12291.
## Test Plan
```shell
❯ cargo run check ../uv/foo --select INP
/Users/crmarsh/workspace/uv/foo/bar/baz.py:1:1: INP001 File `/Users/crmarsh/workspace/uv/foo/bar/baz.py` is part of an implicit namespace package. Add an `__init__.py`.
Found 1 error.
```
## Summary
I don't fully understand the purpose of this. In #7905, it was just
copied over from the previous non-preview implementation. But it means
that (e.g.) we don't treat `type(self.foo)` as a type -- which is wrong.
Closes https://github.com/astral-sh/ruff/issues/12290.
Carl Meyer
Jonathan Plasse
Alex Waygood
Micha Reiser
github-actions[bot]
Dylan
Dhruv Manilawala
Charlie Marsh
edhinard
Sid
Tzu-ping Chung
Yury Fedotov
renovate[bot]
Ryan Hoban
Christian Clauss
Steve C
Auguste Lalande
epenet
Ran Benita
Chris Krycho
Eero Vaher
Piotr Osiewicz
Aleksei Latyshev
Sigurd Spieckermann
Uriya Harpeness
Mateusz Sokół
TomerBin
Sashko
ukyen
cake-monotone
Matthew Runyon
konsti
Tim Chan
Matthias
Victorien