## Summary
This PR updates the entire parser stack in multiple ways:
### Make the lexer lazy
* https://github.com/astral-sh/ruff/pull/11244
* https://github.com/astral-sh/ruff/pull/11473
Previously, Ruff's lexer would act as an iterator. The parser would
collect all the tokens in a vector first and then process the tokens to
create the syntax tree.
The first task in this project is to update the entire parsing flow to
make the lexer lazy. This includes the `Lexer`, `TokenSource`, and
`Parser`. For context, the `TokenSource` is a wrapper around the `Lexer`
to filter out the trivia tokens[^1]. Now, the parser will ask the token
source to get the next token and only then the lexer will continue and
emit the token. This means that the lexer needs to be aware of the
"current" token. When the `next_token` is called, the current token will
be updated with the newly lexed token.
The main motivation to make the lexer lazy is to allow re-lexing a token
in a different context. This is going to be really useful to make the
parser error resilience. For example, currently the emitted tokens
remains the same even if the parser can recover from an unclosed
parenthesis. This is important because the lexer emits a
`NonLogicalNewline` in parenthesized context while a normal `Newline` in
non-parenthesized context. This different kinds of newline is also used
to emit the indentation tokens which is important for the parser as it's
used to determine the start and end of a block.
Additionally, this allows us to implement the following functionalities:
1. Checkpoint - rewind infrastructure: The idea here is to create a
checkpoint and continue lexing. At a later point, this checkpoint can be
used to rewind the lexer back to the provided checkpoint.
2. Remove the `SoftKeywordTransformer` and instead use lookahead or
speculative parsing to determine whether a soft keyword is a keyword or
an identifier
3. Remove the `Tok` enum. The `Tok` enum represents the tokens emitted
by the lexer but it contains owned data which makes it expensive to
clone. The new `TokenKind` enum just represents the type of token which
is very cheap.
This brings up a question as to how will the parser get the owned value
which was stored on `Tok`. This will be solved by introducing a new
`TokenValue` enum which only contains a subset of token kinds which has
the owned value. This is stored on the lexer and is requested by the
parser when it wants to process the data. For example:
8196720f80/crates/ruff_python_parser/src/parser/expression.rs (L1260-L1262)
[^1]: Trivia tokens are `NonLogicalNewline` and `Comment`
### Remove `SoftKeywordTransformer`
* https://github.com/astral-sh/ruff/pull/11441
* https://github.com/astral-sh/ruff/pull/11459
* https://github.com/astral-sh/ruff/pull/11442
* https://github.com/astral-sh/ruff/pull/11443
* https://github.com/astral-sh/ruff/pull/11474
For context,
https://github.com/RustPython/RustPython/pull/4519/files#diff-5de40045e78e794aa5ab0b8aacf531aa477daf826d31ca129467703855408220
added support for soft keywords in the parser which uses infinite
lookahead to classify a soft keyword as a keyword or an identifier. This
is a brilliant idea as it basically wraps the existing Lexer and works
on top of it which means that the logic for lexing and re-lexing a soft
keyword remains separate. The change here is to remove
`SoftKeywordTransformer` and let the parser determine this based on
context, lookahead and speculative parsing.
* **Context:** The transformer needs to know the position of the lexer
between it being at a statement position or a simple statement position.
This is because a `match` token starts a compound statement while a
`type` token starts a simple statement. **The parser already knows
this.**
* **Lookahead:** Now that the parser knows the context it can perform
lookahead of up to two tokens to classify the soft keyword. The logic
for this is mentioned in the PR implementing it for `type` and `match
soft keyword.
* **Speculative parsing:** This is where the checkpoint - rewind
infrastructure helps. For `match` soft keyword, there are certain cases
for which we can't classify based on lookahead. The idea here is to
create a checkpoint and keep parsing. Based on whether the parsing was
successful and what tokens are ahead we can classify the remaining
cases. Refer to #11443 for more details.
If the soft keyword is being parsed in an identifier context, it'll be
converted to an identifier and the emitted token will be updated as
well. Refer
8196720f80/crates/ruff_python_parser/src/parser/expression.rs (L487-L491).
The `case` soft keyword doesn't require any special handling because
it'll be a keyword only in the context of a match statement.
### Update the parser API
* https://github.com/astral-sh/ruff/pull/11494
* https://github.com/astral-sh/ruff/pull/11505
Now that the lexer is in sync with the parser, and the parser helps to
determine whether a soft keyword is a keyword or an identifier, the
lexer cannot be used on its own. The reason being that it's not
sensitive to the context (which is correct). This means that the parser
API needs to be updated to not allow any access to the lexer.
Previously, there were multiple ways to parse the source code:
1. Passing the source code itself
2. Or, passing the tokens
Now that the lexer and parser are working together, the API
corresponding to (2) cannot exists. The final API is mentioned in this
PR description: https://github.com/astral-sh/ruff/pull/11494.
### Refactor the downstream tools (linter and formatter)
* https://github.com/astral-sh/ruff/pull/11511
* https://github.com/astral-sh/ruff/pull/11515
* https://github.com/astral-sh/ruff/pull/11529
* https://github.com/astral-sh/ruff/pull/11562
* https://github.com/astral-sh/ruff/pull/11592
And, the final set of changes involves updating all references of the
lexer and `Tok` enum. This was done in two-parts:
1. Update all the references in a way that doesn't require any changes
from this PR i.e., it can be done independently
* https://github.com/astral-sh/ruff/pull/11402
* https://github.com/astral-sh/ruff/pull/11406
* https://github.com/astral-sh/ruff/pull/11418
* https://github.com/astral-sh/ruff/pull/11419
* https://github.com/astral-sh/ruff/pull/11420
* https://github.com/astral-sh/ruff/pull/11424
2. Update all the remaining references to use the changes made in this
PR
For (2), there were various strategies used:
1. Introduce a new `Tokens` struct which wraps the token vector and add
methods to query a certain subset of tokens. These includes:
1. `up_to_first_unknown` which replaces the `tokenize` function
2. `in_range` and `after` which replaces the `lex_starts_at` function
where the former returns the tokens within the given range while the
latter returns all the tokens after the given offset
2. Introduce a new `TokenFlags` which is a set of flags to query certain
information from a token. Currently, this information is only limited to
any string type token but can be expanded to include other information
in the future as needed. https://github.com/astral-sh/ruff/pull/11578
3. Move the `CommentRanges` to the parsed output because this
information is common to both the linter and the formatter. This removes
the need for `tokens_and_ranges` function.
## Test Plan
- [x] Update and verify the test snapshots
- [x] Make sure the entire test suite is passing
- [x] Make sure there are no changes in the ecosystem checks
- [x] Run the fuzzer on the parser
- [x] Run this change on dozens of open-source projects
### Running this change on dozens of open-source projects
Refer to the PR description to get the list of open source projects used
for testing.
Now, the following tests were done between `main` and this branch:
1. Compare the output of `--select=E999` (syntax errors)
2. Compare the output of default rule selection
3. Compare the output of `--select=ALL`
**Conclusion: all output were same**
## What's next?
The next step is to introduce re-lexing logic and update the parser to
feed the recovery information to the lexer so that it can emit the
correct token. This moves us one step closer to having error resilience
in the parser and provides Ruff the possibility to lint even if the
source code contains syntax errors.
## Summary
Implement support for RDJson output for `ruff check`, as requested in
#8655.
## Test Plan
Tested using a snapshot test. Same approach as for e.g. the JSON output
formatter.
## Additional info
I tried to keep the implementation close to the JSON implementation.
I had to deviate a bit to make the `suggestions` key work: If there are
no suggestions, then setting `suggestions` to `null` is invalid
according to the JSONSchema. Therefore, I opted for a slightly more
complex implementation, that skips the `suggestions` key entirely if
there are no fixes available for the given diagnostic. Maybe it would
have been easier to set `"suggestions": []`, but I ended up doing it
this way.
I didn't consider notebooks, as I _think_ that RDJson doesn't work with
notebooks. This should be confirmed, and if so, there should be some
form of warning or error emitted when trying to output diagnostics for a
notebook.
I also didn't consider `ruff format`, as this comment:
https://github.com/astral-sh/ruff/issues/8655#issuecomment-1811446160
suggests that that wouldn't be compatible.
I'm new to Rust, any feedback is appreciated. 🙂 I
implemented this in order to have a productive rainy saturday afternoon,
I'm not knowledgeable about RDJson beyond the sources linked in the
issue.
## Summary
This PR implements the rule B901, which is part of the opinionated rules
of `flake8-bugbear`.
This rule seems to be desired in `ruff` as per
https://github.com/astral-sh/ruff/issues/3758 and
https://github.com/astral-sh/ruff/issues/2954#issuecomment-1441162976.
## Test Plan
As this PR was made closely following the
[CONTRIBUTING.md](8a25531a71/CONTRIBUTING.md),
it tests using the snapshot approach, that is described there.
## Sources
The implementation is inspired by [the original implementation in the
`flake8-bugbear`
repository](d1aec4cbef/bugbear.py (L1092)).
The error message and [test
file](d1aec4cbef/tests/b901.py)
where also copied from there.
The documentation I came up with on my own and needs improvement. Maybe
the example given in
https://github.com/astral-sh/ruff/issues/2954#issuecomment-1441162976
could be used, but maybe they are too complex, I'm not sure.
## Open Questions
- [ ] Documentation. (See above.)
- [x] Can I access the parent in a visitor?
The [original
implementation](d1aec4cbef/bugbear.py (L1100))
references the `yield` statement's parent to check if it is an
expression statement. I didn't find a way to do this in `ruff` and used
the `is_expresssion_statement` field on the visitor instead. What are
your thoughts on this? Is it possible and / or desired to access the
parent node here?
- [x] Is `Option::is_some(...)` -> `...unwrap()` the right thing to do?
Referring to [this piece of
code](9d5a280f71/crates/ruff_linter/src/rules/flake8_bugbear/rules/return_x_in_generator.rs (L91-L96)).
From my understanding, the `.unwrap()` is safe, because it is checked
that `return_` is not `None`. However, I feel like I missed a more
elegant solution that does both in one.
## Other
I don't know a lot about this rule, I just implemented it because I
found it in a
https://github.com/astral-sh/ruff/labels/good%20first%20issue.
I'm new to Rust, so any constructive critisism is appreciated.
---------
Co-authored-by: Charlie Marsh <charlie.r.marsh@gmail.com>
* Potentially resolves#11619 (nondeterministic hashmap order across
different architectures) in F401 by replacing a hashmap with
nondeterministic traversal order with an ordered mapping.
I'm not sure how to test this with our CI/CD. I don't have an s390x
machine at home. Should I try it in Qemu?
## Summary
In an `__init__.py` file, it's not uncommon to lack a logical indent
(since it may just contain imports). In such cases, we were always
falling back to four-space indent. This PR adds detection for indents
within import groups.
Closes https://github.com/astral-sh/ruff/issues/11606.
## Summary
This PR aims to close#10095 by adding an option
`init-allow-undef-export` to the `pyflakes` settings. This option is
currently set to `true` such that behavior is kept identical.
But setting this option to `false` will lead to `F822` warnings to be
shown in all files, **including** `__init__.py` files.
As I've mentioned on #10095, I think `init-allow-undef-export=false`
would be the more user-friendly default option, as it creates fewer
surprises. @charliermarsh what do you think about making that the
default?
With this option in place, it's a single line fix for people that rely
on the old behavior.
And thinking longer term, for future major releases, one could probably
consider deprecating the option and eventually having people just `noqa`
these warnings if they are not wanted.
## Test Plan
I've added a `test_init_f822_enabled` test which repeats the test that
is done in the `init` test but this time with
`init-allow-undef-export=false` and the snap file correctly shows that
ruff will then trigger the otherwise suppressed F822 warning.
closes#10095
## Summary
Removed stray space in sample code snippet that is against ruff's own
default formatting rules.
This documentation appears on
https://docs.astral.sh/ruff/rules/unused-import/
## Test Plan
This is a trivially obvious change, verifiable with `ruff format
--check`
## Summary
- Implements `Y066` from `flake8-pyi` as `PYI066`
- Fixes `PYI006` not being raised for `elif` clauses. This would have
conflicted with PYI006's implementation, so decided to do it in the same
PR.
## Test Plan
`cargo test` / `cargo insta review`
## Summary
This PR ensures that if a variable is bound via `global`, and then the
`global` is read, the originating variable is also marked as read. It's
not perfect, in that it won't detect _rebindings_, like:
```python
from app import redis_connection
def func():
global redis_connection
redis_connection = 1
redis_connection()
```
So, above, `redis_connection` is still marked as unused.
But it does avoid flagging `redis_connection` as unused in:
```python
from app import redis_connection
def func():
global redis_connection
redis_connection()
```
Closes https://github.com/astral-sh/ruff/issues/11518.
## Summary
Follow up to https://github.com/astral-sh/ruff/pull/11521
Removes the extra added complexity for catch all match cases. This
matches the implementation of plain `else` statements.
## Test Plan
Added new test cases.
---------
Co-authored-by: Dhruv Manilawala <dhruvmanila@gmail.com>
## Summary
This PR brings back the functionality to remove empty strings when
converting to an f-string in `UP032`.
For context, https://github.com/astral-sh/ruff/pull/8712 added this
functionality to remove _trailing_ empty strings but it got removed in
https://github.com/astral-sh/ruff/pull/8697 possibly unexpectedly so.
There's one difference which is that this PR will remove _any_ empty
strings and not just trailing ones. For example,
```diff
--- /Users/dhruv/playground/ruff/src/UP032.py
+++ /Users/dhruv/playground/ruff/src/UP032.py
@@ -1,7 +1,5 @@
(
- "{a}"
- ""
- "{b}"
- ""
-).format(a=1, b=1)
+ f"{1}"
+ f"{1}"
+)
```
## Test Plan
Run `cargo insta test` and update the snapshots.
## Summary
This PR updates the sequence sorting (`RUF022` and `RUF023`) to avoid
using the owned data from the string token. Instead, we will directly
use the reference to the data on the AST. This does introduce a lot of
lifetimes but that's required.
The main motivation for this is to allow removing the `lex_starts_at`
usage easily.
### Alternatives
1. Extract the raw string content (stripping the prefix and quotes)
using the `Locator` and use that for comparison
2. Build up an
[`IndexVec`](3e30962077/crates/ruff_index/src/vec.rs)
and use the newtype index in place of the string value itself. This also
does require lifetimes so we might as well just use the method in this
PR.
## Test Plan
`cargo insta test` and no ecosystem changes
## Summary
Concurrent GitLab runners clone projects into separate directories, e.g.
`{builds_dir}/$RUNNER_TOKEN_KEY/$CONCURRENT_ID/$NAMESPACE/$PROJECT_NAME`.
Since the fingerprint uses the full path to the file, the fingerprints
calculated by Ruff are different depending on which concurrent runner it
executes on, so often an MR will appear to remove all existing issues
and add them with new fingerprints.
I've adjusted the fingerprint function to use the project relative path,
which fixes this. Unfortunately this will have a breaking change for any
current users of this output - the fingerprints will change and appear
in GitLab as all linting messages having been fixed and then created.
## Test Plan
`cargo nextest run`
Running `ruff check --output-format gitlab` in a git repo, moving the
repo and running again, verifying no diffs between the outputs
Hi!
I left out some of the functions in the migration rule which became
removed in NumPy 2.0:
- `np.alltrue`
- `np.anytrue`
- `np.cumproduct`
- `np.product`
Addressing: https://github.com/numpy/numpy/issues/26493
## Summary
<!-- What's the purpose of the change? What does it do, and why? -->
Current doc says `sys.version[0]` will select the first digit of a major
version number (correct) then as an example says
> e.g., `"3.10"` would evaluate to `"1"`
(would actually evaluate to `"3"`). Changed the example version to a
two-digit number to make the problem more clear.
## Test Plan
<!-- How was it tested? -->
ran the following:
- `cargo run -p ruff -- check
crates/ruff_linter/resources/test/fixtures/flake8_2020/YTT301.py
--no-cache`
- `cargo insta review`
- `cargo test`
which all passed.
## Summary
Rule `logging-warn` (`G010`) prescribes a change from `warn` to
`warning` and has a corresponding autofix, but the autofix is mistakenly
titled ```"Convert to `warn`"``` instead of ```"Convert to `warning`"```
(the latter is what the autofix actually does). Seems to be a plain
typo.
## Summary
It turns out that `singledispatch` does end up evaluating all arguments,
even though only the first is used to dispatch.
Closes https://github.com/astral-sh/ruff/issues/11520.
## Summary
Addresses #8451 by implementing rule 116 to add an unsafe fix when sleep
is used with a >24 hour interval to instead consider sleeping forever.
This rule is added as async instead as I my understanding was that these
trio rules would be moved to async anyway.
There are a couple of TODOs, which address further extending the rule by
adding support for lookups and evaluations, and also supporting `anyio`.
## Summary
This PR updates the `FA102` rule logic to use the `Importer` which is
available on the `Checker`.
The main motivation is that this would make updating the `Importer` to
use the `Tokens` struct which will be required to remove the
`lex_starts_at` usage in `Insertion::start_of_block` method.
## Test Plan
`cargo insta test`
## Summary
Similar to #11414, this PR extends `UP037` to flag quoted annotations
that are located in positions that won't be evaluated at runtime.
For example, the quotes on `Tuple` are unnecessary in:
```python
from typing import TYPE_CHECKING
if TYPE_CHECKING:
from typing import Tuple
def foo():
x: "Tuple[int, int]" = (0, 0)
foo()
```
## Summary
This PR moves the `has_comments` function from `Indexer` to
`CommentRanges`. The main motivation is that the `CommentRanges` will
now be built by the parser which is shared between the linter and the
formatter. Thus, the `CommentRanges` will be removed from the `Indexer`.
## Test Plan
`cargo test`
## Summary
Matching Pylint, we now omit the `try` body itself from branch counting.
Each `except` counts as a branch, as does the `else` and the `finally`.
Closes https://github.com/astral-sh/ruff/issues/11205.
The wording 'negative comparison' is a rather vague description of the
'is not' operation and does not describe what the 'not in' operation
does (potentially copied from 'is not'). This was replaced with more
precise language to describe the operators taken from the official
python docs[1].
Both rules didn't have a strong reasoning besides 'it's bad, use the
other'. The origin of these rules seems to be PEP8[2] which prefers 'is
not' over 'not ... is' for readability. This is now reflected in the
description.
[1]:
https://docs.python.org/3/reference/expressions.html#membership-test-operations
[2]: https://peps.python.org/pep-0008/#programming-recommendations
## Summary
If an annotation won't be evaluated at runtime, we don't need to flag
`from __future__ import annotations` as required. This applies both to
quoted annotations and annotations outside of runtime-evaluated
positions, like:
```python
def main() -> None:
a_list: list[str] | None = []
a_list.append("hello")
```
Closes https://github.com/astral-sh/ruff/issues/11397.
## Summary
* Update documentation for F401 following recent PRs
* #11168
* #11314
* Deprecate `ignore_init_module_imports`
* Add a deprecation pragma to the option and a "warn user once" message
when the option is used.
* Restore the old behavior for stable (non-preview) mode:
* When `ignore_init_module_imports` is set to `true` (default) there are
no `__init_.py` fixes (but we get nice fix titles!).
* When `ignore_init_module_imports` is set to `false` there are unsafe
`__init__.py` fixes to remove unused imports.
* When preview mode is enabled, it overrides
`ignore_init_module_imports`.
* Fixed a bug in fix titles where `import foo as bar` would recommend
reexporting `bar as bar`. It now says to reexport `foo as foo`. (In this
case we don't issue a fix, fwiw; it was just a fix title bug.)
## Test plan
Added new fixture tests that reuse the existing fixtures for
`__init__.py` files. Each of the three situations listed above has
fixture tests. The F401 "stable" tests cover:
> * When `ignore_init_module_imports` is set to `true` (default) there
are no `__init_.py` fixes (but we get nice fix titles!).
The F401 "deprecated option" tests cover:
> * When `ignore_init_module_imports` is set to `false` there are unsafe
`__init__.py` fixes to remove unused imports.
These complement existing "preview" tests that show the new behavior
which recommends fixes in `__init__.py` according to whether the import
is 1st party and other circumstances (for more on that behavior see:
#11314).
## Summary
This is a follow-up PR to #11445 update the `E27` rules to consider soft
keywords as well.
## Test Plan
Add test cases consisting of soft keywords and update the snapshot.
## Summary
We weren't treating the escaped newline as a valid condition to trigger
the safer fix (add an extra backslash before each invalid escape
sequence).
Closes https://github.com/astral-sh/ruff/issues/11461.
