Summary
--
This PR tweaks Ruff's internal usage of the new diagnostic model to more
closely
match the intended use, as I understand it. Specifically, it moves the
fix/help
suggestion from the primary annotation's message to a subdiagnostic. In
turn, it
adds the secondary/noqa code as the new primary annotation message. As
shown in
the new `ruff_db` tests, this more closely mirrors Ruff's current
diagnostic
output.
I also added `Severity::Help` to render the fix suggestion with a
`help:` prefix
instead of `info:`.
These changes don't have any external impact now but should help a bit
with #19415.
Test Plan
--
New full output format tests in `ruff_db`
Rendered Diagnostics
--
Full diagnostic output from `annotate-snippets` in this PR:
```
error[unused-import]: `os` imported but unused
--> fib.py:1:8
|
1 | import os
| ^^
|
help: Remove unused import: `os`
```
Current Ruff output for the same code:
```
fib.py:1:8: F401 [*] `os` imported but unused
|
1 | import os
| ^^ F401
|
= help: Remove unused import: `os`
```
Proposed final output after #19415:
```
F401 [*] `os` imported but unused
--> fib.py:1:8
|
1 | import os
| ^^
|
help: Remove unused import: `os`
```
These are slightly updated from
https://github.com/astral-sh/ruff/pull/19464#issuecomment-3097377634
below to remove the extra noqa codes in the primary annotation messages
for the first and third cases.
## Summary
This was originally stacked on #19129, but some of the changes I made
for JSON also impacted the Azure format, so I went ahead and combined
them. The main changes here are:
- Implementing `FileResolver` for Ruff's `EmitterContext`
- Adding `FileResolver::notebook_index` and `FileResolver::is_notebook`
methods
- Adding a `DisplayDiagnostics` (with an "s") type for rendering a group
of diagnostics at once
- Adding `Azure`, `Json`, and `JsonLines` as new `DiagnosticFormat`s
I tried a couple of alternatives to the `FileResolver::notebook` methods
like passing down the `NotebookIndex` separately and trying to reparse a
`Notebook` from Ruff's `SourceFile`. The latter seemed promising, but
the `SourceFile` only stores the concatenated plain text of the
notebook, not the re-parsable JSON. I guess the current version is just
a variation on passing the `NotebookIndex`, but at least we can reuse
the existing `resolver` argument. I think a lot of this can be cleaned
up once Ruff has its own actual file resolver.
As suggested, I also tried deleting the corresponding `Emitter` files in
`ruff_linter`, but it doesn't look like git was able to follow this as a
rename. It did, however, track that the tests were moved, so the
snapshots should be easy to review.
## Test Plan
Existing Ruff tests ported to tests in `ruff_db`. I think some other
existing ruff tests also cover parts of this refactor.
---------
Co-authored-by: Micha Reiser <micha@reiser.io>
## Summary
This PR addresses some additional feedback on #19053:
- Renaming the `syntax_error` methods to `invalid_syntax` to match the
lint id
- Moving the standalone `diagnostic_from_violation` function to
`Violation::into_diagnostic`
- Removing the `Ord` and `PartialOrd` implementations from `Diagnostic`
in favor of `Diagnostic::start_ordering`
## Test Plan
Existing tests
## Additional Follow-ups
Besides these, I also put the following comments on my todo list, but
they seemed like they might be big enough to have their own PRs:
- [Use `LintId::IOError` for IO
errors](https://github.com/astral-sh/ruff/pull/19053#discussion_r2189425922)
- [Move `Fix` and
`Edit`](https://github.com/astral-sh/ruff/pull/19053#discussion_r2189448647)
- [Avoid so many
unwraps](https://github.com/astral-sh/ruff/pull/19053#discussion_r2189465980)
## Summary
This PR is a collaboration with @AlexWaygood from our pairing session
last Friday.
The main goal here is removing `ruff_linter::message::OldDiagnostic` in
favor of
using `ruff_db::diagnostic::Diagnostic` directly. This involved a few
major steps:
- Transferring the fields
- Transferring the methods and trait implementations, where possible
- Converting some constructor methods to free functions
- Moving the `SecondaryCode` struct
- Updating the method names
I'm hoping that some of the methods, especially those in the
`expect_ruff_*`
family, won't be necessary long-term, but I avoided trying to replace
them
entirely for now to keep the already-large diff a bit smaller.
### Related refactors
Alex and I noticed a few refactoring opportunities while looking at the
code,
specifically the very similar implementations for
`create_parse_diagnostic`,
`create_unsupported_syntax_diagnostic`, and
`create_semantic_syntax_diagnostic`.
We combined these into a single generic function, which I then copied
into
`ruff_linter::message` with some small changes and a TODO to combine
them in the
future.
I also deleted the `DisplayParseErrorType` and `TruncateAtNewline` types
for
reporting parse errors. These were added in #4124, I believe to work
around the
error messages from LALRPOP. Removing these didn't affect any tests, so
I think
they were unnecessary now that we fully control the error messages from
the
parser.
On a more minor note, I factored out some calls to the
`OldDiagnostic::filename`
(now `Diagnostic::expect_ruff_filename`) function to avoid repeatedly
allocating
`String`s in some places.
### Snapshot changes
The `show_statistics_syntax_errors` integration test changed because the
`OldDiagnostic::name` method used `syntax-error` instead of
`invalid-syntax`
like in ty. I think this (`--statistics`) is one of the only places we
actually
use this name for syntax errors, so I hope this is okay. An alternative
is to
use `syntax-error` in ty too.
The other snapshot changes are from removing this code, as discussed on
[Discord](1388252408):
34052a1185/crates/ruff_linter/src/message/mod.rs (L128-L135)
I think both of these are technically breaking changes, but they only
affect
syntax errors and are very narrow in scope, while also pretty
substantially
simplifying the refactor, so I hope they're okay to include in a patch
release.
## Test plan
Existing tests, with the adjustments mentioned above
---------
Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com>
## Summary
I think this should be the last step before combining `OldDiagnostic`
and `ruff_db::Diagnostic`. We can't store a `NoqaCode` on
`ruff_db::Diagnostic`, so I converted the `noqa_code` field to an
`Option<String>` and then propagated this change to all of the callers.
I tried to use `&str` everywhere it was possible, so I think the
remaining `to_string` calls are necessary. I spent some time trying to
convert _everything_ to `&str` but ran into lifetime issues, especially
in the `FixTable`. Maybe we can take another look at that if it causes a
performance regression, but hopefully these paths aren't too hot. We
also avoid some `to_string` calls, so it might even out a bit too.
## Test Plan
Existing tests
---------
Co-authored-by: Micha Reiser <micha@reiser.io>
## Summary
Ref:
https://github.com/astral-sh/ruff/issues/14820#issuecomment-2996690681
This PR fixes a bug where virtual paths or any paths that doesn't exists
on the file system weren't being considered for checking inclusion /
exclusion. This was because the logic used `file_path` which returns
`None` for those path. This PR fixes that by using the
`virtual_file_path` method that returns a `Path` corresponding to the
actual file on disk or any kind of virtual path.
This should ideally just fix the above linked issue by way of excluding
the documents representing the interactive window because they aren't in
the inclusion set. It failed only on Windows previously because the file
path construction would fail and then Ruff would default to including
all the files.
## Test Plan
On my machine, the `.interactive` paths are always excluded so I'm using
the inclusion set instead:
```json
{
"ruff.nativeServer": "on",
"ruff.path": ["/Users/dhruv/work/astral/ruff/target/debug/ruff"],
"ruff.configuration": {
"extend-include": ["*.interactive"]
}
}
```
The diagnostics are shown for both the file paths and the interactive
window:
<img width="1727" alt="Screenshot 2025-06-24 at 14 56 40"
src="https://github.com/user-attachments/assets/d36af96a-777e-4367-8acf-4d9c9014d025"
/>
And, the logs:
```
2025-06-24 14:56:26.478275000 DEBUG notification{method="notebookDocument/didChange"}: Included path via `extend-include`: /Interactive-1.interactive
```
And, when using `ruff.exclude` via:
```json
{
"ruff.exclude": ["*.interactive"]
}
```
With logs:
```
2025-06-24 14:58:41.117743000 DEBUG notification{method="notebookDocument/didChange"}: Ignored path via `exclude`: /Interactive-1.interactive
```
Summary
--
This PR unifies the remaining differences between `OldDiagnostic` and
`Message` (`OldDiagnostic` was only missing an optional `noqa_offset`
field) and
replaces `Message` with `OldDiagnostic`.
The biggest functional difference is that the combined `OldDiagnostic`
kind no
longer implements `AsRule` for an infallible conversion to `Rule`. This
was
pretty easy to work around with `is_some_and` and `is_none_or` in the
few places
it was needed. In `LintContext::report_diagnostic_if_enabled` we can
just use
the new `Violation::rule` method, which takes care of most cases.
Most of the interesting changes are in [this
range](8156992540)
before I started renaming.
Test Plan
--
Existing tests
Future Work
--
I think it's time to start shifting some of these fields to the new
`Diagnostic`
kind. I believe we want `Fix` for sure, but I'm less sure about the
others. We
may want to keep a thin wrapper type here anyway to implement a `rule`
method,
so we could leave some of these fields on that too.
## Summary
As the title says, this PR removes the `Message::to_rule` method by
replacing related uses of `Rule` with `NoqaCode` (or the rule's name in
the case of the cache). Where it seemed a `Rule` was really needed, we
convert back to the `Rule` by parsing either the rule name (with
`str::parse`) or the `NoqaCode` (with `Rule::from_code`).
I thought this was kind of like cheating and that it might not resolve
this part of Micha's
[comment](https://github.com/astral-sh/ruff/pull/18391#issuecomment-2933764275):
> because we can't add Rule to Diagnostic or **have it anywhere in our
shared rendering logic**
but after looking again, the only remaining `Rule` conversion in
rendering code is for the SARIF output format. The other two non-test
`Rule` conversions are for caching and writing a fix summary, which I
don't think fall into the shared rendering logic. That leaves the SARIF
format as the only real problem, but maybe we can delay that for now.
The motivation here is that we won't be able to store a `Rule` on the
new `Diagnostic` type, but we should be able to store a `NoqaCode`,
likely as a string.
## Test Plan
Existing tests
##
[Benchmarks](https://codspeed.io/astral-sh/ruff/branches/brent%2Fremove-to-rule)
Almost no perf regression, only -1% on
`linter/default-rules[large/dataset.py]`.
---------
Co-authored-by: Micha Reiser <micha@reiser.io>
## Summary
This is a spin-off from
https://github.com/astral-sh/ruff/pull/18447#discussion_r2125844669 to
avoid using `Message::noqa_code` to differentiate between lints and
syntax errors. I went through all of the calls on `main` and on the
branch from #18447, and the instance in `ruff_server` noted in the
linked comment was actually the primary place where this was being done.
Other calls to `noqa_code` are typically some variation of
`message.noqa_code().map_or(String::new, format!(...))`, with the major
exception of the gitlab output format:
a120610b5b/crates/ruff_linter/src/message/gitlab.rs (L93-L105)
which obviously assumes that `None` means syntax error. A simple fix
here would be to use `message.name()` for `check_name` instead of the
noqa code, but I'm not sure how breaking that would be. This could just
be:
```rust
let description = message.body();
let description = description.strip_prefix("SyntaxError: ").unwrap_or(description).to_string();
let check_name = message.name();
```
In that case. This sounds reasonable based on the [Code Quality report
format](https://docs.gitlab.com/ci/testing/code_quality/#code-quality-report-format)
docs:
> | Name | Type | Description|
> |-----|-----|----|
> |`check_name` | String | A unique name representing the check, or
rule, associated with this violation. |
## Test Plan
Existing tests
## Summary
This PR unifies the ruff `Message` enum variants for syntax errors and
rule violations into a single `Message` struct consisting of a shared
`db::Diagnostic` and some additional, optional fields used for some rule
violations.
This version of `Message` is nearly a drop-in replacement for
`ruff_diagnostics::Diagnostic`, which is the next step I have in mind
for the refactor.
I think this is also a useful checkpoint because we could possibly add
some of these optional fields to the new `Diagnostic` type. I think
we've previously discussed wanting support for `Fix`es, but the other
fields seem less relevant, so we may just need to preserve the `Message`
wrapper for a bit longer.
## Test plan
Existing tests
---------
Co-authored-by: Micha Reiser <micha@reiser.io>
## Summary
This PR deletes the `DiagnosticKind` type by inlining its three fields
(`name`, `body`, and `suggestion`) into three other diagnostic types:
`Diagnostic`, `DiagnosticMessage`, and `CacheMessage`.
Instead of deferring to an internal `DiagnosticKind`, both `Diagnostic`
and `DiagnosticMessage` now have their own macro-generated `AsRule`
implementations.
This should make both https://github.com/astral-sh/ruff/pull/18051 and
another follow-up PR changing the type of `name` on `CacheMessage`
easier since its type will be able to change separately from
`Diagnostic` and `DiagnosticMessage`.
## Test Plan
Existing tests
## Summary
This PR is a first step toward integration of the new `Diagnostic` type
into ruff. There are two main changes:
- A new `UnifiedFile` enum wrapping `File` for red-knot and a
`SourceFile` for ruff
- ruff's `Message::SyntaxError` variant is now a `Diagnostic` instead of
a `SyntaxErrorMessage`
The second of these changes was mostly just a proof of concept for the
first, and it went pretty smoothly. Converting `DiagnosticMessage`s will
be most of the work in replacing `Message` entirely.
## Test Plan
Existing tests, which show no changes.
---------
Co-authored-by: Carl Meyer <carl@astral.sh>
Co-authored-by: Micha Reiser <micha@reiser.io>
## Summary
This PR partially addresses #16418 via the following:
- `LinterSettings::unresolved_python_version` is now a `TargetVersion`,
which is a thin wrapper around an `Option<PythonVersion>`
- `Checker::target_version` now calls `TargetVersion::linter_version`
internally, which in turn uses `unwrap_or_default` to preserve the
current default behavior
- Calls to the parser now call `TargetVersion::parser_version`, which
calls `unwrap_or_else(PythonVersion::latest)`
- The `Checker`'s implementation of
`SemanticSyntaxContext::python_version` also uses
`TargetVersion::parser_version` to use `PythonVersion::latest` for
semantic errors
In short, all lint rule behavior should be unchanged, but we default to
the latest Python version for the new syntax errors, which should
minimize confusing version-related syntax errors for users without a
version configured.
## Test Plan
Existing tests, which showed no changes (except for printing default
settings).
Summary
--
This PR updates `check_path` in the `ruff_linter` crate to return a
`Vec<Message>` instead of a `Vec<Diagnostic>`. The main motivation for
this is to make it easier to convert semantic syntax errors directly
into `Message`s rather than `Diagnostic`s in #16106. However, this also
has the benefit of keeping the preview check on unsupported syntax
errors in `check_path`, as suggested in
https://github.com/astral-sh/ruff/pull/16429#discussion_r1974748024.
All of the interesting changes are in the first commit. The second
commit just renames variables like `diagnostics` to `messages`, and the
third commit is a tiny import fix.
I also updated the `ExpandedMessage::location` field name, which caused
a few extra commits tidying up the playground code. I thought it was
nicely symmetric with `end_location`, but I'm happy to revert that too.
Test Plan
--
Existing tests. I also tested the playground and server manually.
Split from F841 following discussion in #8884.
Fixes#8884.
<!--
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? -->
Add a new rule for unused assignments in tuples. Remove similar behavior
from F841.
## Test Plan
Adapt F841 tests and move them over to the new rule.
<!-- How was it tested? -->
---------
Co-authored-by: Micha Reiser <micha@reiser.io>
## Summary
This PR builds on the changes in #16220 to pass a target Python version
to the parser. It also adds the `Parser::unsupported_syntax_errors` field, which
collects version-related syntax errors while parsing. These syntax
errors are then turned into `Message`s in ruff (in preview mode).
This PR only detects one syntax error (`match` statement before Python
3.10), but it has been pretty quick to extend to several other simple
errors (see #16308 for example).
## Test Plan
The current tests are CLI tests in the linter crate, but these could be
supplemented with inline parser tests after #16357.
I also tested the display of these syntax errors in VS Code:
---------
Co-authored-by: Alex Waygood <alex.waygood@gmail.com>
## Summary
This PR refactors the `RuffSettings` struct to directly include the
resolved `Settings` instead of including the specific fields from it.
The server utilizes a lot of it already, so it makes sense to just
include the entire struct for simplicity.
### `Deref`
I implemented `Deref` on `RuffSettings` to return the `Settings` because
`RuffSettings` is now basically a wrapper around it with the config path
as the other field. This path field is only used for debugging
("printDebugInformation" command).
## Summary
This PR creates separate functions to check whether the document path is
excluded for linting or formatting. The main motivation is to avoid the
double `Option` for the call sites and makes passing the correct
settings simpler.
## Summary
The implicit namespace package rule currently fails to detect cases like
the following:
```text
foo/
├── __init__.py
└── bar/
└── baz/
└── __init__.py
```
The problem is that we detect a root at `foo`, and then an independent
root at `baz`. We _would_ detect that `bar` is an implicit namespace
package, but it doesn't contain any files! So we never check it, and
have no place to raise the diagnostic.
This PR adds detection for these kinds of nested packages, and augments
the `INP` rule to flag the `__init__.py` file above with a specialized
message. As a side effect, I've introduced a dedicated `PackageRoot`
struct which we can pass around in lieu of Yet Another `Path`.
For now, I'm only enabling this in preview (and the approach doesn't
affect any other rules). It's a bug fix, but it may end up expanding the
rule.
Closes https://github.com/astral-sh/ruff/issues/13519.
## Summary
This PR updates the native server to consider the `include` and
`extend-include` file resolver settings.
fixes: #12242
## Test Plan
Note: Settings reloading doesn't work for nested configs which is fixed
in #12253 so the preview here only showcases root level config.
e8969128-c175-4f98-8114-0d692b906cc8
## Summary
This PR removes the need to check for `E999` code to find the diagnostic
severity in the server.
**Note:** This is just removing a redundant check because all
`ParseErrors` are converted to `Diagnostic` with default `Error`
severity by
63c92586a1/crates/ruff_server/src/lint.rs (L309-L346)
## Test Plan
Verify that syntax errors are still shown with error severity as it did
before:
<img width="1313" alt="Screenshot 2024-06-28 at 09 30 20"
src="75e389a7-01ea-461c-86a2-0dfc244e515d">
## Summary
Follow-up from #11901
This PR adds a new server setting to show / hide syntax errors.
## Test Plan
### VS Code
Using https://github.com/astral-sh/ruff-vscode/pull/504 with the
following config:
```json
{
"ruff.nativeServer": true,
"ruff.path": ["/Users/dhruv/work/astral/ruff/target/debug/ruff"],
"ruff.showSyntaxErrors": true
}
```
First, set `ruff.showSyntaxErrors` to `true`:
<img width="1177" alt="Screenshot 2024-06-27 at 08 34 58"
src="5d77547a-a908-4a00-8714-7c00784e8679">
And then set it to `false`:
<img width="1185" alt="Screenshot 2024-06-27 at 08 35 19"
src="9720f089-f10c-420b-a2c1-2bbb2245be35">
### Neovim
Using the following Ruff server config:
```lua
require('lspconfig').ruff.setup {
init_options = {
settings = {
showSyntaxErrors = false,
},
},
}
```
First, set `showSyntaxErrors` to `true`:
<img width="1279" alt="Screenshot 2024-06-27 at 08 28 03"
src="e694e231-91ba-47f8-8e8a-ad2e82b85a45">
And then set it to `false`:
<img width="1284" alt="Screenshot 2024-06-27 at 08 28 20"
src="25b86a57-02b1-44f7-9f65-cf5fdde93b0c">
## Summary
Follow-up to #11902
This PR simplifies the `LinterResult` struct by avoiding the generic and
not store the `ParseError`.
This is possible because the callers already have access to the
`ParseError` via the `Parsed` output. This also means that we can
simplify the return type of `check_path` and avoid the generic `T` on
`LinterResult`.
## Test Plan
`cargo insta test`
## Summary
This PR updates the way syntax errors are handled throughout the linter.
The main change is that it's now not considered as a rule which involves
the following changes:
* Update `Message` to be an enum with two variants - one for diagnostic
message and the other for syntax error message
* Provide methods on the new message enum to query information required
by downstream usages
This means that the syntax errors cannot be hidden / disabled via any
disablement methods. These are:
1. Configuration via `select`, `ignore`, `per-file-ignores`, and their
`extend-*` variants
```console
$ cargo run -- check ~/playground/ruff/src/lsp.py --extend-select=E999
--no-preview --no-cache
Finished `dev` profile [unoptimized + debuginfo] target(s) in 0.10s
Running `target/debug/ruff check /Users/dhruv/playground/ruff/src/lsp.py
--extend-select=E999 --no-preview --no-cache`
warning: Rule `E999` is deprecated and will be removed in a future
release. Syntax errors will always be shown regardless of whether this
rule is selected or not.
/Users/dhruv/playground/ruff/src/lsp.py:1:8: F401 [*] `abc` imported but
unused
|
1 | import abc
| ^^^ F401
2 | from pathlib import Path
3 | import os
|
= help: Remove unused import: `abc`
```
3. Command-line flags via `--select`, `--ignore`, `--per-file-ignores`,
and their `--extend-*` variants
```console
$ cargo run -- check ~/playground/ruff/src/lsp.py --no-cache
--config=~/playground/ruff/pyproject.toml
Finished `dev` profile [unoptimized + debuginfo] target(s) in 0.11s
Running `target/debug/ruff check /Users/dhruv/playground/ruff/src/lsp.py
--no-cache --config=/Users/dhruv/playground/ruff/pyproject.toml`
warning: Rule `E999` is deprecated and will be removed in a future
release. Syntax errors will always be shown regardless of whether this
rule is selected or not.
/Users/dhruv/playground/ruff/src/lsp.py:1:8: F401 [*] `abc` imported but
unused
|
1 | import abc
| ^^^ F401
2 | from pathlib import Path
3 | import os
|
= help: Remove unused import: `abc`
```
This also means that the **output format** needs to be updated:
1. The `code`, `noqa_row`, `url` fields in the JSON output is optional
(`null` for syntax errors)
2. Other formats are changed accordingly
For each format, a new test case specific to syntax errors have been
added. Please refer to the snapshot output for the exact format for
syntax error message.
The output of the `--statistics` flag will have a blank entry for syntax
errors:
```
315 F821 [ ] undefined-name
119 [ ] syntax-error
103 F811 [ ] redefined-while-unused
```
The **language server** is updated to consider the syntax errors by
convert them into LSP diagnostic format separately.
### Preview
There are no quick fixes provided to disable syntax errors. This will
automatically work for `ruff-lsp` because the `noqa_row` field will be
`null` in that case.
<img width="772" alt="Screenshot 2024-06-26 at 14 57 08"
src="aaac827e-4777-4ac8-8c68-eaf9f2c36774">
Even with `noqa` comment, the syntax error is displayed:
<img width="763" alt="Screenshot 2024-06-26 at 14 59 51"
src="ba1afb68-7eaf-4b44-91af-6d93246475e2">
Rule documentation page:
<img width="1371" alt="Screenshot 2024-06-26 at 16 48 07"
src="524f01df-d91f-4ac0-86cc-40e76b318b24">
## Test Plan
- [x] Disablement methods via config shows a warning
- [x] `select`, `extend-select`
- [ ] ~`ignore`~ _doesn't show any message_
- [ ] ~`per-file-ignores`, `extend-per-file-ignores`~ _doesn't show any
message_
- [x] Disablement methods via command-line flag shows a warning
- [x] `--select`, `--extend-select`
- [ ] ~`--ignore`~ _doesn't show any message_
- [ ] ~`--per-file-ignores`, `--extend-per-file-ignores`~ _doesn't show
any message_
- [x] File with syntax errors should exit with code 1
- [x] Language server
- [x] Should show diagnostics for syntax errors
- [x] Should not recommend a quick fix edit for adding `noqa` comment
- [x] Same for `ruff-lsp`
resolves: #8447
## Summary
This PR updates the parser to remove building the `CommentRanges` and
instead it'll be built by the linter and the formatter when it's
required.
For the linter, it'll be built and owned by the `Indexer` while for the
formatter it'll be built from the `Tokens` struct and passed as an
argument.
## Test Plan
`cargo insta test`
## Summary
As-is, we're using the URL path for all files, leading us to use paths
like:
```
/c%3A/Users/crmar/workspace/fastapi/tests/main.py
```
This doesn't match against per-file ignores and other patterns in Ruff
configuration.
This PR modifies the LSP to use the real file path if available, and the
virtual file path if not.
Closes https://github.com/astral-sh/ruff/issues/11751.
## Test Plan
Ran the LSP on Windows. In the FastAPI repo, added:
```toml
[tool.ruff.lint.per-file-ignores]
"tests/**/*.py" = ["F401"]
```
And verified that an unused import was ignored in `tests` after this
change, but not before.
## 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
Closes https://github.com/astral-sh/ruff/issues/11587.
## Test Plan
- Added a lint error to `test_server.py` in `vscode-ruff`.
- Validated that, prior to this change, diagnostics appeared in the
file.
- Validated that, with this change, no diagnostics were shown.
- Validated that, with this change, no diagnostics were fixed on-save.
## Summary
Fixes https://github.com/astral-sh/ruff/issues/11236.
This PR fixes several issues, most of which relate to non-VS Code
editors (Helix and Neovim).
1. Global-only initialization options are now correctly deserialized
from Neovim and Helix
2. Empty diagnostics are now published correctly for Neovim and Helix.
3. A workspace folder is created at the current working directory if the
initialization parameters send an empty list of workspace folders.
4. The server now gracefully handles opening files outside of any known
workspace, and will use global fallback settings taken from client
editor settings and a user settings TOML, if it exists.
## Test Plan
I've tested to confirm that each issue has been fixed.
* Global-only initialization options are now correctly deserialized from
Neovim and Helix + the server gracefully handles opening files outside
of any known workspace
4f33477f-20c8-4e50-8214-6608b1a1ea6b
* Empty diagnostics are now published correctly for Neovim and Helix
c93f56a0-f75d-466f-9f40-d77f99cf0637
* A workspace folder is created at the current working directory if the
initialization parameters send an empty list of workspace folders.
b4b2e818-4b0d-40ce-961d-5831478cc726
## Summary
Closes https://github.com/astral-sh/ruff/issues/10858.
`ruff server` now supports `*.ipynb` (aka Jupyter Notebook) files.
Extensive internal changes have been made to facilitate this, which I've
done some work to contextualize with documentation and an pre-review
that highlights notable sections of the code.
`*.ipynb` cells should behave similarly to `*.py` documents, with one
major exception. The format command `ruff.applyFormat` will only apply
to the currently selected notebook cell - if you want to format an
entire notebook document, use `Format Notebook` from the VS Code context
menu.
## Test Plan
The VS Code extension does not yet have Jupyter Notebook support
enabled, so you'll first need to enable it manually. To do this,
checkout the `pre-release` branch and modify `src/common/server.ts` as
follows:
Before:
After:
I recommend testing this PR with large, complicated notebook files. I
used notebook files from [this popular
repository](https://github.com/jakevdp/PythonDataScienceHandbook/tree/master/notebooks)
in my preliminary testing.
The main thing to test is ensuring that notebook cells behave the same
as Python documents, besides the aforementioned issue with
`ruff.applyFormat`. You should also test adding and deleting cells (in
particular, deleting all the code cells and ensure that doesn't break
anything), changing the kind of a cell (i.e. from markup -> code or vice
versa), and creating a new notebook file from scratch. Finally, you
should also test that source actions work as expected (and across the
entire notebook).
Note: `ruff.applyAutofix` and `ruff.applyOrganizeImports` are currently
broken for notebook files, and I suspect it has something to do with
https://github.com/astral-sh/ruff/issues/11248. Once this is fixed, I
will update the test plan accordingly.
---------
Co-authored-by: nolan <nolan.king90@gmail.com>
## Summary
Alternative to #11237
This PR adds a new `Tokens` struct which is a newtype wrapper around a
vector of lexer output. This allows us to add a `kinds` method which
returns an iterator over the corresponding `TokenKind`. This iterator is
implemented as a separate `TokenKindIter` struct to allow using the type
and provide additional methods like `peek` directly on the iterator.
This exposes the linter to access the stream of `TokenKind` instead of
`Tok`.
Edit: I've made the necessary downstream changes and plan to merge the
entire stack at once.
## Summary
Fixes https://github.com/astral-sh/ruff/issues/10594.
Code actions to disable a diagnostic via `noqa` comment are now
available.
6d3bcf11-a9d9-499b-8c7f-a10cd39cfbba
`DiagnosticFix` has been changed so that `noqa` code actions appear even
for diagnostics with no available quick fix. It can contain quick fix
edits, `noqa` comment edits, or both.
## Test Plan
The scenarios that need to be tested are as follows:
* A code action to disable a diagnostic should be available for every
diagnostic.
* Using this code action should append to the appropriate line with the
diagnostic, or modify an existing `noqa` comment.
* Adding a `noqa` comment manually should make a diagnostic disappear
* `Fix all auto-fixable problems` should not add `noqa` comments
* Removing a code from a `noqa` comment should make the diagnostic
re-appear
## Summary
This PR removes the `ImportMap` implementation and all its routing
through ruff.
The import map was added in https://github.com/astral-sh/ruff/pull/3243
but we then never ended up using it to do cross file analysis.
We are now working on adding multifile analysis to ruff, and revisit
import resolution as part of it.
```
hyperfine --warmup 10 --runs 20 --setup "./target/release/ruff clean" \
"./target/release/ruff check crates/ruff_linter/resources/test/cpython -e -s --extend-select=I" \
"./target/release/ruff-import check crates/ruff_linter/resources/test/cpython -e -s --extend-select=I"
Benchmark 1: ./target/release/ruff check crates/ruff_linter/resources/test/cpython -e -s --extend-select=I
Time (mean ± σ): 37.6 ms ± 0.9 ms [User: 52.2 ms, System: 63.7 ms]
Range (min … max): 35.8 ms … 39.8 ms 20 runs
Benchmark 2: ./target/release/ruff-import check crates/ruff_linter/resources/test/cpython -e -s --extend-select=I
Time (mean ± σ): 36.0 ms ± 0.7 ms [User: 50.3 ms, System: 58.4 ms]
Range (min … max): 34.5 ms … 37.6 ms 20 runs
Summary
./target/release/ruff-import check crates/ruff_linter/resources/test/cpython -e -s --extend-select=I ran
1.04 ± 0.03 times faster than ./target/release/ruff check crates/ruff_linter/resources/test/cpython -e -s --extend-select=I
```
I suspect that the performance improvement should even be more
significant for users that otherwise don't have any diagnostics.
```
hyperfine --warmup 10 --runs 20 --setup "cd ../ecosystem/airflow && ../../ruff/target/release/ruff clean" \
"./target/release/ruff check ../ecosystem/airflow -e -s --extend-select=I" \
"./target/release/ruff-import check ../ecosystem/airflow -e -s --extend-select=I"
Benchmark 1: ./target/release/ruff check ../ecosystem/airflow -e -s --extend-select=I
Time (mean ± σ): 53.7 ms ± 1.8 ms [User: 68.4 ms, System: 63.0 ms]
Range (min … max): 51.1 ms … 58.7 ms 20 runs
Benchmark 2: ./target/release/ruff-import check ../ecosystem/airflow -e -s --extend-select=I
Time (mean ± σ): 50.8 ms ± 1.4 ms [User: 50.7 ms, System: 60.9 ms]
Range (min … max): 48.5 ms … 55.3 ms 20 runs
Summary
./target/release/ruff-import check ../ecosystem/airflow -e -s --extend-select=I ran
1.06 ± 0.05 times faster than ./target/release/ruff check ../ecosystem/airflow -e -s --extend-select=I
```
## Test Plan
`cargo test`
## Summary
Fixes#11185Fixes#11214
Document path and package information is now forwarded to the Ruff
linter, which allows `per-file-ignores` to correctly match against the
file name. This also fixes an issue where the import sorting rule didn't
distinguish between third-party and first-party packages since we didn't
pass in the package root.
## Test Plan
`per-file-ignores` should ignore files as expected. One quick way to
check is by adding this to your `pyproject.toml`:
```toml
[tool.ruff.lint.per-file-ignores]
"__init__.py" = ["ALL"]
```
Then, confirm that no diagnostics appear when you add code to an
`__init__.py` file (besides syntax errors).
The import sorting fix can be verified by failing to reproduce the
original issue - an `I001` diagnostic should not appear in
`other_module.py`.
## Summary
This builds off of the work in
https://github.com/astral-sh/ruff/pull/10652 to implement a command
executor, backwards compatible with the commands from the previous LSP
(`ruff.applyAutofix`, `ruff.applyFormat` and
`ruff.applyOrganizeImports`).
This involved a lot of refactoring and tweaks to the code action
resolution code - the most notable change is that workspace edits are
specified in a slightly different way, using the more general `changes`
field instead of the `document_changes` field (which isn't supported on
all LSP clients). Additionally, the API for synchronous request handlers
has been updated to include access to the `Requester`, which we use to
send a `workspace/applyEdit` request to the client.
## Test Plan
7932e30f-d944-4e35-b828-1d81aa56c087
## Summary
`ruff server` now has source action `source.fixAll` as an available code
action.
This also fixes https://github.com/astral-sh/ruff/issues/10593 in the
process of revising the code for quick fix code actions.
## Test Plan
f4c07425-e68a-445f-a4ed-949c9197a6be
## Summary
Fixes#10589.
Code that violates `F401` or `F841` (in other words, unused variables or
imports) should now appear greyed out or 'unused' in an editor.
## Test Plan
Put the following test code in a new file within the extension
development host window:
```python
import math
def func():
if False:
unused = "<- this should be greyed out"
```
The following test code should have greyed out/unused import and
variable names, like so:
<img width="294" alt="Screenshot 2024-03-28 at 4 23 18 AM"
src="e84a6e7a-49e2-4fed-9624-f8f9559e0837">
## Summary
Fixes#10588.
Most diagnostics from `ruff server` now appear as a much less alarming
warning instead of an error. Three diagnostics still appear as errors:
`F821` (`undefined name <name>`), `E902` (`IOError`) and `E999`
(`SyntaxError`).
## Test Plan
With an extension using the path to a locally-built executable, open a
file with multiple highlighted problems. Toggle the `Experimental
Server` setting on and off. The highlights should stay as warnings.
Then, modify the file to have a syntactically incorrect element. The
start of the invalid syntax should now have a red highlight.
<!--
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 introduces the `ruff_server` crate and a new `ruff server`
command. `ruff_server` is a re-implementation of
[`ruff-lsp`](https://github.com/astral-sh/ruff-lsp), written entirely in
Rust. It brings significant performance improvements, much tighter
integration with Ruff, a foundation for supporting entirely new language
server features, and more!
This PR is an early version of `ruff_lsp` that we're calling the
**pre-release** version. Anyone is more than welcome to use it and
submit bug reports for any issues they encounter - we'll have some
documentation on how to set it up with a few common editors, and we'll
also provide a pre-release VSCode extension for those interested.
This pre-release version supports:
- **Diagnostics for `.py` files**
- **Quick fixes**
- **Full-file formatting**
- **Range formatting**
- **Multiple workspace folders**
- **Automatic linter/formatter configuration** - taken from any
`pyproject.toml` files in the workspace.
Many thanks to @MichaReiser for his [proof-of-concept
work](https://github.com/astral-sh/ruff/pull/7262), which was important
groundwork for making this PR possible.
## Architectural Decisions
I've made an executive choice to go with `lsp-server` as a base
framework for the LSP, in favor of `tower-lsp`. There were several
reasons for this:
1. I would like to avoid `async` in our implementation. LSPs are mostly
computationally bound rather than I/O bound, and `async` adds a lot of
complexity to the API, while also making harder to reason about
execution order. This leads into the second reason, which is...
2. Any handlers that mutate state should be blocking and run in the
event loop, and the state should be lock-free. This is the approach that
`rust-analyzer` uses (also with the `lsp-server`/`lsp-types` crates as a
framework), and it gives us assurances about data mutation and execution
order. `tower-lsp` doesn't support this, which has caused some
[issues](https://github.com/ebkalderon/tower-lsp/issues/284) around data
races and out-of-order handler execution.
3. In general, I think it makes sense to have tight control over
scheduling and the specifics of our implementation, in exchange for a
slightly higher up-front cost of writing it ourselves. We'll be able to
fine-tune it to our needs and support future LSP features without
depending on an upstream maintainer.
## Test Plan
The pre-release of `ruff_server` will have snapshot tests for common
document editing scenarios. An expanded test suite is on the roadmap for
future version of `ruff_server`.
