Bump version to 0.0.264 (#4179)

Nice. Thank you

.editorconfig

@@ -12,3 +12,6 @@ indent_size = 2
 
 [*.{rs,py}]
 indent_size = 4
+
+[*.snap]
+trim_trailing_whitespace = false

.github/workflows/ci.yaml

@@ -121,15 +121,6 @@ jobs:
       - run: cargo check
       - run: cargo fmt --all --check
 
-  typos:
-    name: "spell check"
-    runs-on: ubuntu-latest
-    steps:
-      - uses: actions/checkout@v3
-      - uses: crate-ci/typos@master
-        with:
-          files: .
-
   ecosystem:
     name: "ecosystem"
     runs-on: ubuntu-latest
@@ -230,3 +221,21 @@ jobs:
           exit_code=${PIPESTATUS[0]}
           echo '```' >> $GITHUB_STEP_SUMMARY
           exit $exit_code
+
+  docs:
+    name: "mkdocs"
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v3
+      - uses: actions/setup-python@v4
+      - name: "Install Rust toolchain"
+        run: rustup show
+      - uses: Swatinem/rust-cache@v2
+      - name: "Install dependencies"
+        run: pip install -r docs/requirements.txt
+      - name: "Update README File"
+        run: python scripts/transform_readme.py --target mkdocs
+      - name: "Generate docs"
+        run: python scripts/generate_mkdocs.py
+      - name: "Build docs"
+        run: mkdocs build --strict

.gitignore

@@ -3,6 +3,7 @@
 crates/ruff/resources/test/cpython
 mkdocs.yml
 .overrides
+github_search.jsonl
 
 ###
 # Rust.gitignore

.pre-commit-config.yaml

@@ -23,6 +23,11 @@ repos:
           - MD033 # no-inline-html
           - --
 
+  - repo: https://github.com/crate-ci/typos
+    rev: v1.14.8
+    hooks:
+      - id: typos
+
   - repo: local
     hooks:
       - id: cargo-fmt

BREAKING_CHANGES.md

@@ -1,5 +1,40 @@
 # Breaking Changes
 
+## 0.0.260
+
+### Fixes are now represented as a list of edits ([#3709](https://github.com/charliermarsh/ruff/pull/3709))
+
+Previously, Ruff represented each fix as a single edit, which prohibited Ruff from automatically
+fixing violations that required multiple edits across a file. As such, Ruff now represents each
+fix as a list of edits.
+
+This primarily affects the JSON API. Ruff's JSON representation used to represent the `fix` field as
+a single edit, like so:
+
+```json
+{
+    "message": "Remove unused import: `sys`",
+    "content": "",
+    "location": {"row": 1, "column": 0},
+    "end_location": {"row": 2, "column": 0}
+}
+```
+
+The updated representation instead includes a list of edits:
+
+```json
+{
+    "message": "Remove unused import: `sys`",
+    "edits": [
+        {
+            "content": "",
+            "location": {"row": 1, "column": 0},
+            "end_location": {"row": 2, "column": 0},
+        }
+    ]
+}
+```
+
 ## 0.0.246
 
 ### `multiple-statements-on-one-line-def` (`E704`) was removed ([#2773](https://github.com/charliermarsh/ruff/pull/2773))

CONTRIBUTING.md

@@ -116,8 +116,7 @@ At a high level, the steps involved in adding a new lint rule are as follows:
 To define the violation, start by creating a dedicated file for your rule under the appropriate
 rule linter (e.g., `crates/ruff/src/rules/flake8_bugbear/rules/abstract_base_class.rs`). That file should
 contain a struct defined via `#[violation]`, along with a function that creates the violation
-based on any required inputs. (Many of the existing examples live in `crates/ruff/src/violations.rs`,
-but we're looking to place new rules in their own files.)
+based on any required inputs.
 
 To trigger the violation, you'll likely want to augment the logic in `crates/ruff/src/checkers/ast.rs`,
 which defines the Python AST visitor, responsible for iterating over the abstract syntax tree and
@@ -215,6 +214,20 @@ them to [PyPI](https://pypi.org/project/ruff/).
 Ruff follows the [semver](https://semver.org/) versioning standard. However, as pre-1.0 software,
 even patch releases may contain [non-backwards-compatible changes](https://semver.org/#spec-item-4).
 
+## Ecosystem CI
+
+GitHub Actions will run your changes against a number of real-world projects from GitHub and
+report on any diagnostic differences. You can also run those checks locally via:
+
+```shell
+python scripts/check_ecosystem.py path/to/your/ruff path/to/older/ruff
+```
+
+You can also run the Ecosystem CI check in a Docker container across a larger set of projects by
+downloading the [`known-github-tomls.json`](https://github.com/akx/ruff-usage-aggregate/blob/master/data/known-github-tomls.jsonl)
+as `github_search.jsonl` and following the instructions in [scripts/Dockerfile.ecosystem](https://github.com/charliermarsh/ruff/blob/main/scripts/Dockerfile.ecosystem).
+Note that this check will take a while to run.
+
 ## Benchmarks
 
 First, clone [CPython](https://github.com/python/cpython). It's a large and diverse Python codebase,

Cargo.toml

@@ -3,7 +3,7 @@ members = ["crates/*"]
 
 [workspace.package]
 edition = "2021"
-rust-version = "1.67"
+rust-version = "1.69"
 homepage = "https://beta.ruff.rs/docs/"
 documentation = "https://beta.ruff.rs/docs/"
 repository = "https://github.com/charliermarsh/ruff"
@@ -11,7 +11,7 @@ authors = ["Charlie Marsh <charlie.r.marsh@gmail.com>"]
 
 [workspace.dependencies]
 anyhow = { version = "1.0.69" }
-bitflags = { version = "1.3.2" }
+bitflags = { version = "2.1.0" }
 chrono = { version = "0.4.23", default-features = false, features = ["clock"] }
 clap = { version = "4.1.8", features = ["derive"] }
 colored = { version = "2.0.0" }
@@ -30,12 +30,10 @@ path-absolutize = { version = "3.0.14" }
 proc-macro2 = { version = "1.0.51" }
 quote = { version = "1.0.23" }
 regex = { version = "1.7.1" }
+ruff_text_size = { git = "https://github.com/charliermarsh/RustPython.git", rev = "c3147d2c1524ebd0e90cf1c2938d770314fd5a5a" }
 rustc-hash = { version = "1.1.0" }
-rustpython-common = { git = "https://github.com/RustPython/RustPython.git", rev = "c15f670f2c30cfae6b41a1874893590148c74bc4" }
-rustpython-parser = { features = [
-  "lalrpop",
-  "serde",
-], git = "https://github.com/RustPython/RustPython.git", rev = "c15f670f2c30cfae6b41a1874893590148c74bc4" }
+rustpython-common = { git = "https://github.com/charliermarsh/RustPython.git", rev = "c3147d2c1524ebd0e90cf1c2938d770314fd5a5a" }
+rustpython-parser = { git = "https://github.com/charliermarsh/RustPython.git", rev = "c3147d2c1524ebd0e90cf1c2938d770314fd5a5a" }
 schemars = { version = "0.8.12" }
 serde = { version = "1.0.152", features = ["derive"] }
 serde_json = { version = "1.0.93", features = ["preserve_order"] }
@@ -44,7 +42,7 @@ similar = { version = "2.2.1" }
 smallvec = { version = "1.10.0" }
 strum = { version = "0.24.1", features = ["strum_macros"] }
 strum_macros = { version = "0.24.3" }
-syn = { version = "1.0.109" }
+syn = { version = "2.0.15" }
 test-case = { version = "3.0.0" }
 textwrap = { version = "0.16.0" }
 toml = { version = "0.7.2" }

README.md

@@ -47,16 +47,16 @@ all while executing tens or hundreds of times faster than any individual tool.
 
 Ruff is extremely actively developed and used in major open-source projects like:
 
-- [pandas](https://github.com/pandas-dev/pandas)
-- [FastAPI](https://github.com/tiangolo/fastapi)
-- [Transformers (Hugging Face)](https://github.com/huggingface/transformers)
 - [Apache Airflow](https://github.com/apache/airflow)
+- [FastAPI](https://github.com/tiangolo/fastapi)
+- [Hugging Face](https://github.com/huggingface/transformers)
+- [Pandas](https://github.com/pandas-dev/pandas)
 - [SciPy](https://github.com/scipy/scipy)
 
 ...and many more.
 
-Read the [launch blog post](https://notes.crmarsh.com/python-tooling-could-be-much-much-faster) or
-the most recent [project update](https://notes.crmarsh.com/ruff-the-first-200-releases).
+Ruff is backed by [Astral](https://astral.sh). Read the [launch post](https://astral.sh/blog/announcing-astral-the-company-behind-ruff),
+or the original [project announcement](https://notes.crmarsh.com/python-tooling-could-be-much-much-faster).
 
 ## Testimonials
 
@@ -137,7 +137,7 @@ Ruff can also be used as a [pre-commit](https://pre-commit.com) hook:
 ```yaml
 - repo: https://github.com/charliermarsh/ruff-pre-commit
   # Ruff version.
-  rev: 'v0.0.261'
+  rev: 'v0.0.264'
   hooks:
     - id: ruff
 ```
@@ -332,55 +332,68 @@ Ruff is released under the MIT license.
 
 ## Who's Using Ruff?
 
-Ruff is used in a number of major open-source projects, including:
+Ruff is used by a number of major open-source projects and companies, including:
 
-- [pandas](https://github.com/pandas-dev/pandas)
-- [FastAPI](https://github.com/tiangolo/fastapi)
-- [Transformers (Hugging Face)](https://github.com/huggingface/transformers)
-- [Diffusers (Hugging Face)](https://github.com/huggingface/diffusers)
+- Amazon ([AWS SAM](https://github.com/aws/serverless-application-model))
 - [Apache Airflow](https://github.com/apache/airflow)
-- [SciPy](https://github.com/scipy/scipy)
-- [Zulip](https://github.com/zulip/zulip)
-- [Bokeh](https://github.com/bokeh/bokeh)
-- [Pydantic](https://github.com/pydantic/pydantic)
-- [PostHog](https://github.com/PostHog/posthog)
-- [Dagster](https://github.com/dagster-io/dagster)
-- [Dagger](https://github.com/dagger/dagger)
-- [Sphinx](https://github.com/sphinx-doc/sphinx)
-- [Hatch](https://github.com/pypa/hatch)
-- [PDM](https://github.com/pdm-project/pdm)
-- [Jupyter](https://github.com/jupyter-server/jupyter_server)
-- [Great Expectations](https://github.com/great-expectations/great_expectations)
-- [ONNX](https://github.com/onnx/onnx)
-- [Polars](https://github.com/pola-rs/polars)
-- [Ibis](https://github.com/ibis-project/ibis)
-- [Synapse (Matrix)](https://github.com/matrix-org/synapse)
-- [SnowCLI (Snowflake)](https://github.com/Snowflake-Labs/snowcli)
-- [Dispatch (Netflix)](https://github.com/Netflix/dispatch)
-- [Saleor](https://github.com/saleor/saleor)
-- [Pynecone](https://github.com/pynecone-io/pynecone)
-- [OpenBB](https://github.com/OpenBB-finance/OpenBBTerminal)
-- [Home Assistant](https://github.com/home-assistant/core)
-- [Pylint](https://github.com/PyCQA/pylint)
-- [Cryptography (PyCA)](https://github.com/pyca/cryptography)
-- [cibuildwheel (PyPA)](https://github.com/pypa/cibuildwheel)
-- [build (PyPA)](https://github.com/pypa/build)
+- AstraZeneca ([Magnus](https://github.com/AstraZeneca/magnus-core))
+- Benchling ([Refac](https://github.com/benchling/refac))
 - [Babel](https://github.com/python-babel/babel)
-- [featuretools](https://github.com/alteryx/featuretools)
-- [meson-python](https://github.com/mesonbuild/meson-python)
-- [ZenML](https://github.com/zenml-io/zenml)
-- [delta-rs](https://github.com/delta-io/delta-rs)
-- [Starlite](https://github.com/starlite-api/starlite)
-- [telemetry-airflow (Mozilla)](https://github.com/mozilla/telemetry-airflow)
-- [Stable Baselines3](https://github.com/DLR-RM/stable-baselines3)
-- [PaddlePaddle](https://github.com/PaddlePaddle/Paddle)
-- [nox](https://github.com/wntrblm/nox)
-- [Neon](https://github.com/neondatabase/neon)
-- [The Algorithms](https://github.com/TheAlgorithms/Python)
-- [Openverse](https://github.com/WordPress/openverse)
-- [MegaLinter](https://github.com/oxsecurity/megalinter)
+- [Bokeh](https://github.com/bokeh/bokeh)
+- [Cryptography (PyCA)](https://github.com/pyca/cryptography)
+- [Dagger](https://github.com/dagger/dagger)
+- [Dagster](https://github.com/dagster-io/dagster)
+- [DVC](https://github.com/iterative/dvc)
+- [FastAPI](https://github.com/tiangolo/fastapi)
+- [Gradio](https://github.com/gradio-app/gradio)
+- [Great Expectations](https://github.com/great-expectations/great_expectations)
+- Hugging Face ([Transformers](https://github.com/huggingface/transformers), [Datasets](https://github.com/huggingface/datasets), [Diffusers](https://github.com/huggingface/diffusers))
+- [Hatch](https://github.com/pypa/hatch)
+- [Home Assistant](https://github.com/home-assistant/core)
+- [Ibis](https://github.com/ibis-project/ibis)
+- [Jupyter](https://github.com/jupyter-server/jupyter_server)
 - [LangChain](https://github.com/hwchase17/langchain)
 - [LlamaIndex](https://github.com/jerryjliu/llama_index)
+- Matrix ([Synapse](https://github.com/matrix-org/synapse))
+- Meltano ([Meltano CLI](https://github.com/meltano/meltano), [Singer SDK](https://github.com/meltano/sdk))
+- Modern Treasury ([Python SDK](https://github.com/Modern-Treasury/modern-treasury-python-sdk))
+- Mozilla ([Firefox](https://github.com/mozilla/gecko-dev))
+- [MegaLinter](https://github.com/oxsecurity/megalinter)
+- Microsoft ([Semantic Kernel](https://github.com/microsoft/semantic-kernel), [ONNX Runtime](https://github.com/microsoft/onnxruntime))
+- Netflix ([Dispatch](https://github.com/Netflix/dispatch))
+- [Neon](https://github.com/neondatabase/neon)
+- [ONNX](https://github.com/onnx/onnx)
+- [OpenBB](https://github.com/OpenBB-finance/OpenBBTerminal)
+- [PDM](https://github.com/pdm-project/pdm)
+- [PaddlePaddle](https://github.com/PaddlePaddle/Paddle)
+- [Pandas](https://github.com/pandas-dev/pandas)
+- [Poetry](https://github.com/python-poetry/poetry)
+- [Polars](https://github.com/pola-rs/polars)
+- [PostHog](https://github.com/PostHog/posthog)
+- Prefect ([Python SDK](https://github.com/PrefectHQ/prefect), [Marvin](https://github.com/PrefectHQ/marvin))
+- [Pydantic](https://github.com/pydantic/pydantic)
+- [PyInstaller](https://github.com/pyinstaller/pyinstaller)
+- [Pylint](https://github.com/PyCQA/pylint)
+- [Pynecone](https://github.com/pynecone-io/pynecone)
+- [Robyn](https://github.com/sansyrox/robyn)
+- Scale AI ([Launch SDK](https://github.com/scaleapi/launch-python-client))
+- Snowflake ([SnowCLI](https://github.com/Snowflake-Labs/snowcli))
+- [Saleor](https://github.com/saleor/saleor)
+- [SciPy](https://github.com/scipy/scipy)
+- [Sphinx](https://github.com/sphinx-doc/sphinx)
+- [Stable Baselines3](https://github.com/DLR-RM/stable-baselines3)
+- [Starlite](https://github.com/starlite-api/starlite)
+- [The Algorithms](https://github.com/TheAlgorithms/Python)
+- [Vega-Altair](https://github.com/altair-viz/altair)
+- WordPress ([Openverse](https://github.com/WordPress/openverse))
+- [ZenML](https://github.com/zenml-io/zenml)
+- [Zulip](https://github.com/zulip/zulip)
+- [build (PyPA)](https://github.com/pypa/build)
+- [cibuildwheel (PyPA)](https://github.com/pypa/cibuildwheel)
+- [delta-rs](https://github.com/delta-io/delta-rs)
+- [featuretools](https://github.com/alteryx/featuretools)
+- [meson-python](https://github.com/mesonbuild/meson-python)
+- [nox](https://github.com/wntrblm/nox)
 
 ## License
 

_typos.toml

@@ -5,3 +5,6 @@ extend-exclude = ["snapshots", "black"]
 trivias = "trivias"
 hel = "hel"
 whos = "whos"
+spawnve = "spawnve"
+ned = "ned"
+poit = "poit"

crates/flake8_to_ruff/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "flake8-to-ruff"
-version = "0.0.261"
+version = "0.0.264"
 edition = { workspace = true }
 rust-version = { workspace = true }
 

crates/ruff/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "ruff"
-version = "0.0.261"
+version = "0.0.264"
 authors.workspace = true
 edition.workspace = true
 rust-version.workspace = true
@@ -17,17 +17,19 @@ name = "ruff"
 ruff_cache = { path = "../ruff_cache" }
 ruff_diagnostics = { path = "../ruff_diagnostics", features = ["serde"] }
 ruff_macros = { path = "../ruff_macros" }
-ruff_python_ast = { path = "../ruff_python_ast" }
+ruff_python_ast = { path = "../ruff_python_ast", features = ["serde"] }
 ruff_python_semantic = { path = "../ruff_python_semantic" }
 ruff_python_stdlib = { path = "../ruff_python_stdlib" }
 ruff_rustpython = { path = "../ruff_rustpython" }
+ruff_text_size = { workspace = true }
 
+annotate-snippets = { version = "0.9.1", features = ["color"] }
 anyhow = { workspace = true }
 bitflags = { workspace = true }
 chrono = { workspace = true }
 clap = { workspace = true, features = ["derive", "string"], optional = true }
 colored = { workspace = true }
-dirs = { version = "4.0.0" }
+dirs = { version = "5.0.0" }
 fern = { version = "0.6.1" }
 glob = { workspace = true }
 globset = { workspace = true }
@@ -48,6 +50,7 @@ path-absolutize = { workspace = true, features = [
 ] }
 pathdiff = { version = "0.2.1" }
 pep440_rs = { version = "0.3.1", features = ["serde"] }
+quick-junit = { version = "0.3.2" }
 regex = { workspace = true }
 result-like = { version = "0.4.6" }
 rustc-hash = { workspace = true }
@@ -57,6 +60,7 @@ schemars = { workspace = true }
 semver = { version = "1.0.16" }
 serde = { workspace = true }
 serde_json = { workspace = true }
+similar = { workspace = true, features = ["inline"] }
 shellexpand = { workspace = true }
 smallvec = { workspace = true }
 strum = { workspace = true }
@@ -68,9 +72,11 @@ typed-arena = { version = "2.0.2" }
 unicode-width = { version = "0.1.10" }
 
 [dev-dependencies]
-insta = { workspace = true, features = ["yaml", "redactions"] }
+insta = { workspace = true }
 pretty_assertions = "1.3.0"
 test-case = { workspace = true }
+# Disable colored output in tests
+colored = { workspace = true, features = ["no-color"] }
 
 [features]
 default = []

crates/ruff/resources/test/fixtures/eradicate/ERA001.py

@@ -9,6 +9,7 @@ def foo(x, y, z):
     print(x, y, z)
 
     # This is a real comment.
+    # # This is a (nested) comment.
     #return True
     return False
 

crates/ruff/resources/test/fixtures/flake8_bandit/S101.py

@@ -1,11 +1,13 @@
-# Error
-assert True
+assert True  # S101
+
 
 def fn():
     x = 1
+    assert x == 1  # S101
+    assert x == 2  # S101
 
-    # Error
-    assert x == 1
 
-    # Error
-    assert x == 2
+from typing import TYPE_CHECKING
+
+if TYPE_CHECKING:
+    assert True  # OK

crates/ruff/resources/test/fixtures/flake8_bandit/S602.py

@@ -0,0 +1,20 @@
+from subprocess import Popen, call, check_call, check_output, run
+
+# Check different Popen wrappers are checked.
+Popen("true", shell=True)
+call("true", shell=True)
+check_call("true", shell=True)
+check_output("true", shell=True)
+run("true", shell=True)
+
+# Check values that truthy values are treated as true.
+Popen("true", shell=1)
+Popen("true", shell=[1])
+Popen("true", shell={1: 1})
+Popen("true", shell=(1,))
+
+# Check command argument looks unsafe.
+var_string = "true"
+Popen(var_string, shell=True)
+Popen([var_string], shell=True)
+Popen([var_string, ""], shell=True)

crates/ruff/resources/test/fixtures/flake8_bandit/S603.py

@@ -0,0 +1,20 @@
+from subprocess import Popen, call, check_call, check_output, run
+
+# Different Popen wrappers are checked.
+Popen("true", shell=False)
+call("true", shell=False)
+check_call("true", shell=False)
+check_output("true", shell=False)
+run("true", shell=False)
+
+# Values that falsey values are treated as false.
+Popen("true", shell=0)
+Popen("true", shell=[])
+Popen("true", shell={})
+Popen("true", shell=None)
+
+# Unknown values are treated as falsey.
+Popen("true", shell=True if True else False)
+
+# No value is also caught.
+Popen("true")

crates/ruff/resources/test/fixtures/flake8_bandit/S604.py

@@ -0,0 +1,5 @@
+def foo(shell):
+    pass
+
+
+foo(shell=True)

crates/ruff/resources/test/fixtures/flake8_bandit/S605.py

@@ -0,0 +1,25 @@
+import os
+
+import commands
+import popen2
+
+# Check all shell functions.
+os.system("true")
+os.popen("true")
+os.popen2("true")
+os.popen3("true")
+os.popen4("true")
+popen2.popen2("true")
+popen2.popen3("true")
+popen2.popen4("true")
+popen2.Popen3("true")
+popen2.Popen4("true")
+commands.getoutput("true")
+commands.getstatusoutput("true")
+
+
+# Check command argument looks unsafe.
+var_string = "true"
+os.system(var_string)
+os.system([var_string])
+os.system([var_string, ""])

crates/ruff/resources/test/fixtures/flake8_bandit/S606.py

@@ -0,0 +1,20 @@
+import os
+
+# Check all shell functions.
+os.execl("true")
+os.execle("true")
+os.execlp("true")
+os.execlpe("true")
+os.execv("true")
+os.execve("true")
+os.execvp("true")
+os.execvpe("true")
+os.spawnl("true")
+os.spawnle("true")
+os.spawnlp("true")
+os.spawnlpe("true")
+os.spawnv("true")
+os.spawnve("true")
+os.spawnvp("true")
+os.spawnvpe("true")
+os.startfile("true")

crates/ruff/resources/test/fixtures/flake8_bandit/S607.py

@@ -0,0 +1,44 @@
+import os
+
+# Check all functions.
+subprocess.Popen("true")
+subprocess.call("true")
+subprocess.check_call("true")
+subprocess.check_output("true")
+subprocess.run("true")
+os.system("true")
+os.popen("true")
+os.popen2("true")
+os.popen3("true")
+os.popen4("true")
+popen2.popen2("true")
+popen2.popen3("true")
+popen2.popen4("true")
+popen2.Popen3("true")
+popen2.Popen4("true")
+commands.getoutput("true")
+commands.getstatusoutput("true")
+os.execl("true")
+os.execle("true")
+os.execlp("true")
+os.execlpe("true")
+os.execv("true")
+os.execve("true")
+os.execvp("true")
+os.execvpe("true")
+os.spawnl("true")
+os.spawnle("true")
+os.spawnlp("true")
+os.spawnlpe("true")
+os.spawnv("true")
+os.spawnve("true")
+os.spawnvp("true")
+os.spawnvpe("true")
+os.startfile("true")
+
+# Check it does not fail for full paths.
+os.system("/bin/ls")
+os.system("./bin/ls")
+os.system(["/bin/ls"])
+os.system(["/bin/ls", "/tmp"])
+os.system(r"C:\\bin\ls")

crates/ruff/resources/test/fixtures/flake8_bugbear/B017.py

@@ -1,9 +1,10 @@
 """
 Should emit:
-B017 - on lines 20
+B017 - on lines 23 and 41
 """
 import asyncio
 import unittest
+import pytest
 
 CONSTANT = True
 
@@ -34,3 +35,14 @@ class Foobar(unittest.TestCase):
     def raises_with_absolute_reference(self):
         with self.assertRaises(asyncio.CancelledError):
             Foo()
+
+
+def test_pytest_raises():
+    with pytest.raises(Exception):
+        raise ValueError("Hello")
+
+    with pytest.raises(Exception, "hello"):
+        raise ValueError("This is fine")
+
+    with pytest.raises(Exception, match="hello"):
+        raise ValueError("This is also fine")

crates/ruff/resources/test/fixtures/flake8_bugbear/B023.py

@@ -172,3 +172,14 @@ def iter_f(names):
 
         if False:
             return [lambda: i for i in range(3)]  # error
+
+
+for val in range(3):
+    def make_func(val=val):
+        def tmp():
+            return print(val)
+
+        return tmp
+
+
+    funcs.append(make_func())

crates/ruff/resources/test/fixtures/flake8_bugbear/B031.py

@@ -78,6 +78,21 @@ for _section, section_items in itertools.groupby(items, key=lambda p: p[1]):
     for shopper in shoppers:
         collect_shop_items(shopper, section_items)  # B031
 
+for _section, section_items in itertools.groupby(items, key=lambda p: p[1]):
+    _ = [collect_shop_items(shopper, section_items) for shopper in shoppers]  # B031
+
+for _section, section_items in itertools.groupby(items, key=lambda p: p[1]):
+    # The variable is overridden, skip checking.
+    _ = [_ for section_items in range(3)]
+    _ = [collect_shop_items(shopper, section_items) for shopper in shoppers]
+
+for _section, section_items in itertools.groupby(items, key=lambda p: p[1]):
+    _ = [item for item in section_items]
+
+for _section, section_items in itertools.groupby(items, key=lambda p: p[1]):
+    # The iterator is being used for the second time.
+    _ = [(item1, item2) for item1 in section_items for item2 in section_items]  # B031
+
 for _section, section_items in itertools.groupby(items, key=lambda p: p[1]):
     if _section == "greens":
         collect_shop_items(shopper, section_items)
@@ -134,6 +149,16 @@ for group in groupby(items, key=lambda p: p[1]):
     collect_shop_items("Joe", group[1])
 
 
+# https://github.com/charliermarsh/ruff/issues/4050
+for _section, section_items in itertools.groupby(items, key=lambda p: p[1]):
+    if _section == "greens":
+        for item in section_items:
+            collect_shop_items(shopper, item)
+    elif _section == "frozen items":
+        _ = [item for item in section_items]
+    else:
+        collect_shop_items(shopper, section_items)
+
 #  Make sure we ignore - but don't fail on more complicated invocations
 for _key, (_value1, _value2) in groupby(
     [("a", (1, 2)), ("b", (3, 4)), ("a", (5, 6))], key=lambda p: p[1]

crates/ruff/resources/test/fixtures/flake8_comprehensions/C414.py

@@ -7,11 +7,14 @@ set(set(x))
 set(list(x))
 set(tuple(x))
 set(sorted(x))
+set(sorted(x, key=lambda y: y))
 set(reversed(x))
 sorted(list(x))
 sorted(tuple(x))
 sorted(sorted(x))
+sorted(sorted(x, key=lambda y: y))
 sorted(reversed(x))
+sorted(list(x), key=lambda y: y)
 tuple(
     list(
         [x, 3, "hell"\

crates/ruff/resources/test/fixtures/flake8_comprehensions/C418.py

@@ -0,0 +1,10 @@
+dict({})
+dict({'a': 1})
+dict({'x': 1 for x in range(10)})
+dict(
+    {'x': 1 for x in range(10)}
+)
+
+dict({}, a=1)
+dict({x: 1 for x in range(1)}, a=1)
+

crates/ruff/resources/test/fixtures/flake8_comprehensions/C419.py

@@ -1,4 +1,3 @@
-# PIE802
 any([x.id for x in bar])
 all([x.id for x in bar])
 any(  # first comment
@@ -15,5 +14,6 @@ all(x.id for x in bar)
 any(x.id for x in bar)
 all((x.id for x in bar))
 
+
 async def f() -> bool:
     return all([await use_greeting(greeting) for greeting in await greetings()])

crates/ruff/resources/test/fixtures/flake8_errmsg/EM.py

@@ -21,3 +21,36 @@ def f_c():
 def f_ok():
     msg = "hello"
     raise RuntimeError(msg)
+
+
+def f_unfixable():
+    msg = "hello"
+    raise RuntimeError("This is an example exception")
+
+
+def f_msg_in_nested_scope():
+    def nested():
+        msg = "hello"
+
+    raise RuntimeError("This is an example exception")
+
+
+def f_msg_in_parent_scope():
+    msg = "hello"
+
+    def nested():
+        raise RuntimeError("This is an example exception")
+
+
+def f_fix_indentation_check(foo):
+    if foo:
+        raise RuntimeError("This is an example exception")
+    else:
+        if foo == "foo":
+            raise RuntimeError(f"This is an exception: {foo}")
+    raise RuntimeError("This is an exception: {}".format(foo))
+
+
+# Report these, but don't fix them
+if foo: raise RuntimeError("This is an example exception")
+if foo: x = 1; raise RuntimeError("This is an example exception")

crates/ruff/resources/test/fixtures/flake8_import_conventions/custom_banned.py

@@ -0,0 +1,16 @@
+import typing as t  # banned
+import typing as ty  # banned
+
+import numpy as nmp  # banned
+import numpy as npy  # banned
+import tensorflow.keras.backend as K  # banned
+import torch.nn.functional as F  # banned
+from tensorflow.keras import backend as K  # banned
+from torch.nn import functional as F  # banned
+
+from typing import Any  # ok
+
+import numpy as np  # ok
+import tensorflow as tf  # ok
+import torch.nn as nn  # ok
+from tensorflow.keras import backend  # ok

crates/ruff/resources/test/fixtures/flake8_import_conventions/custom_banned_from.py

@@ -0,0 +1,10 @@
+from logging.config import BaseConfigurator  # banned
+from typing import Any, Dict  # banned
+from typing import *  # banned
+
+from pandas import DataFrame  # banned
+from pandas import *  # banned
+
+import logging.config  # ok
+import typing  # ok
+import pandas  # ok

crates/ruff/resources/test/fixtures/flake8_pyi/PYI001.py

@@ -11,3 +11,7 @@ _T = TypeVar("_T")  # OK
 _TTuple = TypeVarTuple("_TTuple")  # OK
 
 _P = ParamSpec("_P")  # OK
+
+
+def f():
+    T = TypeVar("T")  # OK

crates/ruff/resources/test/fixtures/flake8_pyi/PYI001.pyi

@@ -11,3 +11,6 @@ _T = TypeVar("_T")  # OK
 _TTuple = TypeVarTuple("_TTuple")  # OK
 
 _P = ParamSpec("_P")  # OK
+
+def f():
+    T = TypeVar("T")  # OK

crates/ruff/resources/test/fixtures/flake8_pyi/PYI015.py

@@ -46,3 +46,48 @@ field229: dict[int, int] = {1: 2, **{3: 4}}  # Y015 Only simple default values a
 field23 = "foo" + "bar"  # Y015 Only simple default values are allowed for assignments
 field24 = b"foo" + b"bar"  # Y015 Only simple default values are allowed for assignments
 field25 = 5 * 5  # Y015 Only simple default values are allowed for assignments
+
+# We shouldn't emit Y015 within functions
+def f():
+  field26: list[int] = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
+
+
+# We shouldn't emit Y015 for __slots__ or __match_args__
+class Class1:
+  __slots__ = (
+    '_one',
+    '_two',
+    '_three',
+    '_four',
+    '_five',
+    '_six',
+    '_seven',
+    '_eight',
+    '_nine',
+    '_ten',
+    '_eleven',
+  )
+
+  __match_args__ = (
+    'one',
+    'two',
+    'three',
+    'four',
+    'five',
+    'six',
+    'seven',
+    'eight',
+    'nine',
+    'ten',
+    'eleven',
+  )
+
+# We shouldn't emit Y015 for __all__
+__all__ = ["Class1"]
+
+# Ignore the following for PYI015
+field26 = typing.Sequence[int]
+field27 = list[str]
+field28 = builtins.str
+field29 = str
+field30 = str | bytes | None

crates/ruff/resources/test/fixtures/flake8_pyi/PYI015.pyi

@@ -53,3 +53,48 @@ field229: dict[int, int] = {1: 2, **{3: 4}}  # Y015 Only simple default values a
 field23 = "foo" + "bar"  # Y015 Only simple default values are allowed for assignments
 field24 = b"foo" + b"bar"  # Y015 Only simple default values are allowed for assignments
 field25 = 5 * 5  # Y015 Only simple default values are allowed for assignments
+
+# We shouldn't emit Y015 within functions
+def f():
+  field26: list[int] = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
+
+
+# We shouldn't emit Y015 for __slots__ or __match_args__
+class Class1:
+  __slots__ = (
+    '_one',
+    '_two',
+    '_three',
+    '_four',
+    '_five',
+    '_six',
+    '_seven',
+    '_eight',
+    '_nine',
+    '_ten',
+    '_eleven',
+  )
+
+  __match_args__ = (
+    'one',
+    'two',
+    'three',
+    'four',
+    'five',
+    'six',
+    'seven',
+    'eight',
+    'nine',
+    'ten',
+    'eleven',
+  )
+
+# We shouldn't emit Y015 for __all__
+__all__ = ["Class1"]
+
+# Ignore the following for PYI015
+field26 = typing.Sequence[int]
+field27 = list[str]
+field28 = builtins.str
+field29 = str
+field30 = str | bytes | None

crates/ruff/resources/test/fixtures/flake8_pyi/PYI016.py

@@ -0,0 +1,35 @@
+# Shouldn't affect non-union field types.
+field1: str
+
+# Should emit for duplicate field types.
+field2: str | str  # PYI016: Duplicate union member `str`
+
+
+# Should emit for union types in arguments.
+def func1(arg1: int | int):  # PYI016: Duplicate union member `int`
+    print(arg1)
+
+
+# Should emit for unions in return types.
+def func2() -> str | str:  # PYI016: Duplicate union member `str`
+    return "my string"
+
+
+# Should emit in longer unions, even if not directly adjacent.
+field3: str | str | int  # PYI016: Duplicate union member `str`
+field4: int | int | str  # PYI016: Duplicate union member `int`
+field5: str | int | str  # PYI016: Duplicate union member `str`
+field6: int | bool | str | int  # PYI016: Duplicate union member `int`
+
+# Shouldn't emit for non-type unions.
+field7 = str | str
+
+# Should emit for strangely-bracketed unions.
+field8: int | (str | int)  # PYI016: Duplicate union member `int`
+
+# Should handle user brackets when fixing.
+field9: int | (int | str)  # PYI016: Duplicate union member `int`
+field10: (str | int) | str  # PYI016: Duplicate union member `str`
+
+# Should emit for nested unions.
+field11: dict[int | int, str]

crates/ruff/resources/test/fixtures/flake8_pyi/PYI016.pyi

@@ -0,0 +1,32 @@
+# Shouldn't affect non-union field types.
+field1: str
+
+# Should emit for duplicate field types.
+field2: str | str  # PYI016: Duplicate union member `str`
+
+# Should emit for union types in arguments.
+def func1(arg1: int | int):  # PYI016: Duplicate union member `int`
+    print(arg1)
+
+# Should emit for unions in return types.
+def func2() -> str | str:  # PYI016: Duplicate union member `str`
+    return "my string"
+
+# Should emit in longer unions, even if not directly adjacent.
+field3: str | str | int  # PYI016: Duplicate union member `str`
+field4: int | int | str  # PYI016: Duplicate union member `int`
+field5: str | int | str  # PYI016: Duplicate union member `str`
+field6: int | bool | str | int  # PYI016: Duplicate union member `int`
+
+# Shouldn't emit for non-type unions.
+field7 = str | str
+
+# Should emit for strangely-bracketed unions.
+field8: int | (str | int)  # PYI016: Duplicate union member `int`
+
+# Should handle user brackets when fixing.
+field9: int | (int | str)  # PYI016: Duplicate union member `int`
+field10: (str | int) | str  # PYI016: Duplicate union member `str`
+
+# Should emit for nested unions.
+field11: dict[int | int, str]

crates/ruff/resources/test/fixtures/flake8_pytest_style/PT006.py

@@ -49,3 +49,18 @@ def test_list_expressions(param1, param2):
 @pytest.mark.parametrize([some_expr, "param2"], [1, 2, 3])
 def test_list_mixed_expr_literal(param1, param2):
     ...
+
+
+@pytest.mark.parametrize(("param1, " "param2, " "param3"), [(1, 2, 3), (4, 5, 6)])
+def test_implicit_str_concat_with_parens(param1, param2, param3):
+    ...
+
+
+@pytest.mark.parametrize("param1, " "param2, " "param3", [(1, 2, 3), (4, 5, 6)])
+def test_implicit_str_concat_no_parens(param1, param2, param3):
+    ...
+
+
+@pytest.mark.parametrize((("param1, " "param2, " "param3")), [(1, 2, 3), (4, 5, 6)])
+def test_implicit_str_concat_with_multi_parens(param1, param2, param3):
+    ...

crates/ruff/resources/test/fixtures/flake8_return/RET504.py

@@ -3,7 +3,7 @@
 ###
 def x():
     a = 1
-    return a  # error
+    return a  # RET504
 
 
 # Can be refactored false positives
@@ -211,10 +211,10 @@ def nonlocal_assignment():
 def decorator() -> Flask:
     app = Flask(__name__)
 
-    @app.route('/hello')
+    @app.route("/hello")
     def hello() -> str:
         """Hello endpoint."""
-        return 'Hello, World!'
+        return "Hello, World!"
 
     return app
 
@@ -222,12 +222,13 @@ def decorator() -> Flask:
 def default():
     y = 1
 
-    def f(x = y) -> X:
+    def f(x=y) -> X:
         return x
 
     return y
 
 
+# Multiple assignment
 def get_queryset(option_1, option_2):
     queryset: Any = None
     queryset = queryset.filter(a=1)
@@ -246,4 +247,28 @@ def get_queryset():
 
 def get_queryset():
     queryset = Model.filter(a=1)
-    return queryset  # error
+    return queryset  # RET504
+
+
+# Function arguments
+def str_to_bool(val):
+    if isinstance(val, bool):
+        return val
+    val = val.strip().lower()
+    if val in ("1", "true", "yes"):
+        return True
+
+    return False
+
+
+def str_to_bool(val):
+    if isinstance(val, bool):
+        return val
+    val = 1
+    return val  # RET504
+
+
+def str_to_bool(val):
+    if isinstance(val, bool):
+        return some_obj
+    return val

crates/ruff/resources/test/fixtures/flake8_simplify/SIM105.py

@@ -59,3 +59,15 @@ def bar():
         return foo()
     except ValueError:
         pass
+
+def with_ellipsis():
+    try:
+        foo()
+    except ValueError:
+        ...
+
+def with_ellipsis_and_return():
+    try:
+        return foo()
+    except ValueError:
+        ...

crates/ruff/resources/test/fixtures/flake8_simplify/SIM222.py

@@ -42,3 +42,113 @@ if False and f() and a and g() and b:  # OK
 
 if a and False and f() and b and g():  # OK
     pass
+
+
+a or "" or True  # SIM222
+
+a or "foo" or True or "bar"  # SIM222
+
+a or 0 or True  # SIM222
+
+a or 1 or True or 2  # SIM222
+
+a or 0.0 or True  # SIM222
+
+a or 0.1 or True or 0.2  # SIM222
+
+a or [] or True  # SIM222
+
+a or list([]) or True  # SIM222
+
+a or [1] or True or [2]  # SIM222
+
+a or list([1]) or True or list([2])  # SIM222
+
+a or {} or True  # SIM222
+
+a or dict() or True  # SIM222
+
+a or {1: 1} or True or {2: 2}  # SIM222
+
+a or dict({1: 1}) or True or dict({2: 2})  # SIM222
+
+a or set() or True  # SIM222
+
+a or set(set()) or True  # SIM222
+
+a or {1} or True or {2}  # SIM222
+
+a or set({1}) or True or set({2})  # SIM222
+
+a or () or True  # SIM222
+
+a or tuple(()) or True  # SIM222
+
+a or (1,) or True or (2,)  # SIM222
+
+a or tuple((1,)) or True or tuple((2,))  # SIM222
+
+a or frozenset() or True  # SIM222
+
+a or frozenset(frozenset()) or True  # SIM222
+
+a or frozenset({1}) or True or frozenset({2})  # SIM222
+
+a or frozenset(frozenset({1})) or True or frozenset(frozenset({2}))  # SIM222
+
+
+# Inside test `a` is simplified.
+
+bool(a or [1] or True or [2])  # SIM222
+
+assert a or [1] or True or [2]  # SIM222
+
+if (a or [1] or True or [2]) and (a or [1] or True or [2]):  # SIM222
+    pass
+
+0 if a or [1] or True or [2] else 1  # SIM222
+
+while a or [1] or True or [2]:  # SIM222
+    pass
+
+[
+    0
+    for a in range(10)
+    for b in range(10)
+    if a or [1] or True or [2]  # SIM222
+    if b or [1] or True or [2]  # SIM222
+]
+
+{
+    0
+    for a in range(10)
+    for b in range(10)
+    if a or [1] or True or [2]  # SIM222
+    if b or [1] or True or [2]  # SIM222
+}
+
+{
+    0: 0
+    for a in range(10)
+    for b in range(10)
+    if a or [1] or True or [2]  # SIM222
+    if b or [1] or True or [2]  # SIM222
+}
+
+(
+    0
+    for a in range(10)
+    for b in range(10)
+    if a or [1] or True or [2]  # SIM222
+    if b or [1] or True or [2]  # SIM222
+)
+
+# Outside test `a` is not simplified.
+
+a or [1] or True or [2]  # SIM222
+
+if (a or [1] or True or [2]) == (a or [1]):  # SIM222
+    pass
+
+if f(a or [1] or True or [2]):  # SIM222
+    pass

crates/ruff/resources/test/fixtures/flake8_simplify/SIM223.py

@@ -37,3 +37,113 @@ if True or f() or a or g() or b:  # OK
 
 if a or True or f() or b or g():  # OK
     pass
+
+
+a and "" and False  # SIM223
+
+a and "foo" and False and "bar"  # SIM223
+
+a and 0 and False  # SIM223
+
+a and 1 and False and 2  # SIM223
+
+a and 0.0 and False  # SIM223
+
+a and 0.1 and False and 0.2  # SIM223
+
+a and [] and False  # SIM223
+
+a and list([]) and False  # SIM223
+
+a and [1] and False and [2]  # SIM223
+
+a and list([1]) and False and list([2])  # SIM223
+
+a and {} and False  # SIM223
+
+a and dict() and False  # SIM223
+
+a and {1: 1} and False and {2: 2}  # SIM223
+
+a and dict({1: 1}) and False and dict({2: 2})  # SIM223
+
+a and set() and False  # SIM223
+
+a and set(set()) and False  # SIM223
+
+a and {1} and False and {2}  # SIM223
+
+a and set({1}) and False and set({2})  # SIM223
+
+a and () and False  # SIM222
+
+a and tuple(()) and False  # SIM222
+
+a and (1,) and False and (2,)  # SIM222
+
+a and tuple((1,)) and False and tuple((2,))  # SIM222
+
+a and frozenset() and False  # SIM222
+
+a and frozenset(frozenset()) and False  # SIM222
+
+a and frozenset({1}) and False and frozenset({2})  # SIM222
+
+a and frozenset(frozenset({1})) and False and frozenset(frozenset({2}))  # SIM222
+
+
+# Inside test `a` is simplified.
+
+bool(a and [] and False and [])  # SIM223
+
+assert a and [] and False and []  # SIM223
+
+if (a and [] and False and []) or (a and [] and False and []):  # SIM223
+    pass
+
+0 if a and [] and False and [] else 1  # SIM222
+
+while a and [] and False and []:  # SIM223
+    pass
+
+[
+    0
+    for a in range(10)
+    for b in range(10)
+    if a and [] and False and []  # SIM223
+    if b and [] and False and []  # SIM223
+]
+
+{
+    0
+    for a in range(10)
+    for b in range(10)
+    if a and [] and False and []  # SIM223
+    if b and [] and False and []  # SIM223
+}
+
+{
+    0: 0
+    for a in range(10)
+    for b in range(10)
+    if a and [] and False and []  # SIM223
+    if b and [] and False and []  # SIM223
+}
+
+(
+    0
+    for a in range(10)
+    for b in range(10)
+    if a and [] and False and []  # SIM223
+    if b and [] and False and []  # SIM223
+)
+
+# Outside test `a` is not simplified.
+
+a and [] and False and []  # SIM223
+
+if (a and [] and False and []) == (a and []):  # SIM223
+    pass
+
+if f(a and [] and False and []):  # SIM223
+    pass

crates/ruff/resources/test/fixtures/flake8_tidy_imports/TID251.py

@@ -31,3 +31,6 @@ typing.TypedDict.anything()
 # import aliases are resolved
 import typing as totally_not_typing
 totally_not_typing.TypedDict
+
+# relative imports are respected
+from .typing import TypedDict

crates/ruff/resources/test/fixtures/flake8_tidy_imports/TID251_false_positives.py

@@ -1,11 +0,0 @@
-# module members cannot be imported with that syntax
-import typing.TypedDict
-
-# we don't track reassignments
-import typing, other
-typing = other
-typing.TypedDict()
-
-# yet another false positive
-def foo(typing):
-    typing.TypedDict()

crates/ruff/resources/test/fixtures/flake8_type_checking/strict.py

@@ -1,3 +1,6 @@
+from __future__ import annotations
+
+
 def f():
     # Even in strict mode, this shouldn't rase an error, since `pkg` is used at runtime,
     # and implicitly imports `pkg.bar`.

crates/ruff/resources/test/fixtures/isort/force_single_line.py

@@ -22,3 +22,6 @@ from bar import (
      a, # comment 7
      b, # comment 8
 )
+
+# comment 9
+from baz import *  # comment 10

crates/ruff/resources/test/fixtures/isort/propagate_inline_comments.py

@@ -0,0 +1,4 @@
+from mypackage.subpackage import (  # long comment that seems to be a problem
+    a_long_variable_name_that_causes_problems,
+    items,
+)

crates/ruff/resources/test/fixtures/isort/required_imports/docstring.pyi

@@ -0,0 +1,3 @@
+"""Hello, world!"""
+
+x = 1

crates/ruff/resources/test/fixtures/isort/sections.py

@@ -0,0 +1,7 @@
+from __future__ import annotations
+import os
+import sys
+import pytz
+import django.settings
+from library import foo
+from . import local

crates/ruff/resources/test/fixtures/pep8_naming/N802.py

@@ -39,3 +39,11 @@ class Test(unittest.TestCase):
 
     def testTest(self):
         assert True
+
+
+from typing import override
+
+
+@override
+def BAD_FUNC():
+    pass

crates/ruff/resources/test/fixtures/pep8_naming/N815.py

@@ -13,3 +13,11 @@ class C:
     myObj2 = namedtuple("MyObj2", ["a", "b"])
     Employee = NamedTuple('Employee', [('name', str), ('id', int)])
     Point2D = TypedDict('Point2D', {'in': int, 'x-y': int})
+
+
+class D(TypedDict):
+    lower: int
+    CONSTANT: str
+    mixedCase: bool
+    _mixedCase: list
+    mixed_Case: set

crates/ruff/resources/test/fixtures/pycodestyle/E714.py

@@ -4,11 +4,11 @@ if not X is Y:
 #: E714
 if not X.B is Y:
     pass
-#: E714
+
+#: Okay
 if not X is Y is not Z:
     pass
 
-#: Okay
 if not X is not Y:
     pass
 

crates/ruff/resources/test/fixtures/pycodestyle/E731.py

@@ -13,6 +13,7 @@ f = lambda: (yield from g())
 class F:
     f = lambda x: 2 * x
 
+
 f = object()
 f.method = lambda: "Method"
 f = {}
@@ -21,3 +22,30 @@ f = []
 f.append(lambda x: x**2)
 f = g = lambda x: x**2
 lambda: "no-op"
+
+# Annotated
+from typing import Callable, ParamSpec
+
+P = ParamSpec("P")
+
+# ParamSpec cannot be used in this context, so do not preserve the annotation.
+f: Callable[P, int] = lambda *args: len(args)
+f: Callable[[], None] = lambda: None
+f: Callable[..., None] = lambda a, b: None
+f: Callable[[int], int] = lambda x: 2 * x
+
+# Let's use the `Callable` type from `collections.abc` instead.
+from collections.abc import Callable
+
+f: Callable[[str, int], str] = lambda a, b: a * b
+f: Callable[[str, int], tuple[str, int]] = lambda a, b: (a, b)
+f: Callable[[str, int, list[str]], list[str]] = lambda a, b, /, c: [*c, a * b]
+
+
+# Override `Callable`
+class Callable:
+    pass
+
+
+# Do not copy the annotation from here on out.
+f: Callable[[str, int], str] = lambda a, b: a * b

crates/ruff/resources/test/fixtures/pycodestyle/W19.py

@@ -97,10 +97,10 @@ if length > options.max_line_length:
 if os.path.exists(os.path.join(path, PEP8_BIN)):
 	cmd = ([os.path.join(path, PEP8_BIN)] +
 	       self._pep8_options(targetfile))
-#: W191
+#: W191 - okay
 '''
 	multiline string with tab in it'''
-#: E101 W191
+#: E101 (W191 okay)
 '''multiline string
 	with tabs
    and spaces
@@ -142,4 +142,10 @@ def test_keys(self):
 x = [
 	'abc'
 ]
-#:
+#: W191 - okay
+'''	multiline string with tab in it, same lines'''
+"""	here we're using '''different delimiters'''"""
+'''
+	multiline string with tab in it, different lines
+'''
+"	single line string with tab in it"

crates/ruff/resources/test/fixtures/pycodestyle/W505.py

@@ -9,6 +9,9 @@ def f():
 
     # Here's a standalone comment that's over the limit.
 
+    x = 2
+    # Another standalone that is preceded by a newline and indent toke and is over the limit.
+
     print("Here's a string that's over the limit, but it's not a docstring.")
 
 

crates/ruff/resources/test/fixtures/pydocstyle/D417.py

@@ -115,6 +115,20 @@ def f(x, *args, **kwargs):
     return x
 
 
+def f(x, *, y, z):
+    """Do something.
+
+    Args:
+        x: some first value
+
+    Keyword Args:
+        y (int): the other value
+        z (int): the last value
+
+    """
+    return x, y, z
+
+
 class Test:
     def f(self, /, arg1: int) -> None:
         """

crates/ruff/resources/test/fixtures/pyflakes/F523.py

@@ -11,3 +11,9 @@
 "{}".format(1, 2, 3)  # F523
 "{:{}}".format(1, 2)  # No issues
 "{:{}}".format(1, 2, 3)  # F523
+
+# With *args
+"{0}{1}".format(*args)  # No issues
+"{0}{1}".format(1, *args)  # No issues
+"{0}{1}".format(1, 2, *args)  # No issues
+"{0}{1}".format(1, 2, 3, *args)  # F523

crates/ruff/resources/test/fixtures/pyflakes/F811_22.py

@@ -0,0 +1,13 @@
+def redef(value):
+    match value:
+        case True:
+
+            def fun(x, y):
+                return x
+
+        case False:
+
+            def fun(x, y):
+                return y
+
+    return fun

crates/ruff/resources/test/fixtures/pyflakes/F821_0.py

@@ -132,3 +132,8 @@ def in_ipython_notebook() -> bool:
     except NameError:
         return False  # not in notebook
     return True
+
+
+def named_expr():
+    if any((key := (value := x)) for x in ["ok"]):
+        print(key)

crates/ruff/resources/test/fixtures/pyflakes/F841_0.py

@@ -121,3 +121,8 @@ def f(x: int):
             print("A")
         case y:
             pass
+
+
+def f():
+    if any((key := (value := x)) for x in ["ok"]):
+        print(key)

crates/ruff/resources/test/fixtures/pygrep-hooks/PGH003_0.py

@@ -1,11 +1,16 @@
 x = 1  # type: ignore
-x = 1  # type ignore
 x = 1  # type:ignore
+x = 1  # type: ignore[attr-defined]  # type: ignore
 
 x = 1
+x = 1  # type ignore
 x = 1  # type ignore  # noqa
 x = 1  # type: ignore[attr-defined]
 x = 1  # type: ignore[attr-defined, name-defined]
+x = 1  # type: ignore[attr-defined]  # type: ignore[type-mismatch]
 x = 1  # type: ignore[type-mismatch]  # noqa
+x = 1  # type: ignore [attr-defined]
+x = 1  # type: ignore [attr-defined, name-defined]
+x = 1  # type: ignore [type-mismatch]  # noqa
 x = 1  # type: Union[int, str]
 x = 1  # type: ignoreme

crates/ruff/resources/test/fixtures/pygrep-hooks/PGH003_1.py

@@ -0,0 +1,16 @@
+x = 1  # pyright: ignore
+x = 1  # pyright:ignore
+x = 1  # pyright: ignore[attr-defined]  # pyright: ignore
+
+x = 1
+x = 1  # pyright ignore
+x = 1  # pyright ignore  # noqa
+x = 1  # pyright: ignore[attr-defined]
+x = 1  # pyright: ignore[attr-defined, name-defined]
+x = 1  # pyright: ignore[attr-defined]  # pyright: ignore[type-mismatch]
+x = 1  # pyright: ignore[type-mismatch]  # noqa
+x = 1  # pyright: ignore [attr-defined]
+x = 1  # pyright: ignore [attr-defined, name-defined]
+x = 1  # pyright: ignore [type-mismatch]  # noqa
+x = 1  # pyright: Union[int, str]
+x = 1  # pyright: ignoreme

crates/ruff/resources/test/fixtures/pylint/redefined_loop_name.py

@@ -1,3 +1,6 @@
+import typing
+from typing import cast
+
 # For -> for, variable reused
 for i in []:
     for i in []:  # error
@@ -43,6 +46,9 @@ for i in []:
 
 # For -> assignment
 for i in []:
+    # ignore typing cast
+    i = cast(int, i)
+    i = typing.cast(int, i)
     i = 5  # error
 
 # For -> augmented assignment
@@ -53,6 +59,10 @@ for i in []:
 for i in []:
     i: int = 5  # error
 
+# For -> annotated assignment without value
+for i in []:
+    i: int  # no error
+
 # Async for -> for, variable reused
 async for i in []:
     for i in []:  # error

crates/ruff/resources/test/fixtures/pylint/unexpected_special_method_signature.py

@@ -0,0 +1,51 @@
+class TestClass:
+    def __bool__(self):
+        ...
+    
+    def __bool__(self, x):  # too many mandatory args
+        ...
+    
+    def __bool__(self, x=1):  # additional optional args OK
+        ...
+    
+    def __bool__(self, *args):  # varargs OK
+        ...
+    
+    def __bool__():  # ignored; should be caughty by E0211/N805
+        ...
+    
+    @staticmethod
+    def __bool__():
+        ...
+
+    @staticmethod
+    def __bool__(x):  # too many mandatory args
+        ...
+    
+    @staticmethod
+    def __bool__(x=1):  # additional optional args OK
+        ...
+    
+    def __eq__(self, other):  # multiple args
+        ...
+    
+    def __eq__(self, other=1):  # expected arg is optional
+        ...
+    
+    def __eq__(self):  # too few mandatory args
+        ...
+    
+    def __eq__(self, other, other_other):  # too many mandatory args
+        ...
+    
+    def __round__(self):  # allow zero additional args.
+        ...
+    
+    def __round__(self, x):  # allow one additional args.
+        ...
+    
+    def __round__(self, x, y):  # disallow 2 args
+        ...
+    
+    def __round__(self, x, y, z=2):  # disallow 3 args even when one is optional
+        ...

crates/ruff/resources/test/fixtures/pyupgrade/UP031_0.py

@@ -83,3 +83,26 @@ print('Hello %s (%s)' % bar['bop'])
 print('Hello %(arg)s' % bar)
 print('Hello %(arg)s' % bar.baz)
 print('Hello %(arg)s' % bar['bop'])
+
+# Hanging modulos
+(
+    "foo %s "
+    "bar %s"
+) % (x, y)
+
+(
+    "foo %(foo)s "
+    "bar %(bar)s"
+) % {"foo": x, "bar": y}
+
+(
+    """foo %s"""
+    % (x,)
+)
+
+(
+    """
+    foo %s
+    """
+    % (x,)
+)

crates/ruff/resources/test/fixtures/pyupgrade/UP031_1.py

@@ -34,28 +34,6 @@ pytest.param('"%8s" % (None,)', id="unsafe width-string conversion"),
 "%(and)s" % {"and": 2}
 
 # OK (arguably false negatives)
-(
-    "foo %s "
-    "bar %s"
-) % (x, y)
-
-(
-    "foo %(foo)s "
-    "bar %(bar)s"
-) % {"foo": x, "bar": y}
-
-(
-    """foo %s"""
-    % (x,)
-)
-
-(
-    """
-    foo %s
-    """
-    % (x,)
-)
-
 'Hello %s' % bar
 
 'Hello %s' % bar.baz

crates/ruff/resources/test/fixtures/pyupgrade/UP032.py

@@ -46,6 +46,14 @@ print("foo {} ".format(x))
 
 '({}={{0!e}})'.format(a)
 
+"{[b]}".format(a)
+
+'{[b]}'.format(a)
+
+"""{[b]}""".format(a)
+
+'''{[b]}'''.format(a)
+
 ###
 # Non-errors
 ###

crates/ruff/resources/test/fixtures/ruff/RUF008.py

@@ -0,0 +1,27 @@
+import typing
+from dataclasses import dataclass, field
+from typing import ClassVar, Sequence
+
+KNOWINGLY_MUTABLE_DEFAULT = []
+
+
+@dataclass()
+class A:
+    mutable_default: list[int] = []
+    immutable_annotation: typing.Sequence[int] = []
+    without_annotation = []
+    ignored_via_comment: list[int] = []  # noqa: RUF008
+    correct_code: list[int] = KNOWINGLY_MUTABLE_DEFAULT
+    perfectly_fine: list[int] = field(default_factory=list)
+    class_variable: typing.ClassVar[list[int]] = []
+
+
+@dataclass
+class B:
+    mutable_default: list[int] = []
+    immutable_annotation: Sequence[int] = []
+    without_annotation = []
+    ignored_via_comment: list[int] = []  # noqa: RUF008
+    correct_code: list[int] = KNOWINGLY_MUTABLE_DEFAULT
+    perfectly_fine: list[int] = field(default_factory=list)
+    class_variable: ClassVar[list[int]] = []

crates/ruff/resources/test/fixtures/ruff/RUF009.py

@@ -0,0 +1,42 @@
+import datetime
+import re
+import typing
+from dataclasses import dataclass, field
+from pathlib import Path
+from typing import ClassVar, NamedTuple
+
+
+def default_function() -> list[int]:
+    return []
+
+
+class ImmutableType(NamedTuple):
+    something: int = 8
+
+
+@dataclass()
+class A:
+    hidden_mutable_default: list[int] = default_function()
+    class_variable: typing.ClassVar[list[int]] = default_function()
+    another_class_var: ClassVar[list[int]] = default_function()
+
+    fine_path: Path = Path()
+    fine_date: datetime.date = datetime.date(2042, 1, 1)
+    fine_timedelta: datetime.timedelta = datetime.timedelta(hours=7)
+    fine_tuple: tuple[int] = tuple([1])
+    fine_regex: re.Pattern = re.compile(r".*")
+
+
+DEFAULT_IMMUTABLETYPE_FOR_ALL_DATACLASSES = ImmutableType(40)
+DEFAULT_A_FOR_ALL_DATACLASSES = A([1, 2, 3])
+
+
+@dataclass
+class B:
+    hidden_mutable_default: list[int] = default_function()
+    another_dataclass: A = A()
+    not_optimal: ImmutableType = ImmutableType(20)
+    good_variant: ImmutableType = DEFAULT_IMMUTABLETYPE_FOR_ALL_DATACLASSES
+    okay_variant: A = DEFAULT_A_FOR_ALL_DATACLASSES
+
+    fine_dataclass_function: list[int] = field(default_factory=list)

crates/ruff/src/autofix/actions.rs

@@ -4,12 +4,12 @@ use itertools::Itertools;
 use libcst_native::{
     Codegen, CodegenState, ImportNames, ParenthesizableWhitespace, SmallStatement, Statement,
 };
-use rustpython_parser::ast::{ExcepthandlerKind, Expr, Keyword, Location, Stmt, StmtKind};
+use ruff_text_size::{TextLen, TextRange, TextSize};
+use rustpython_parser::ast::{ExcepthandlerKind, Expr, Keyword, Stmt, StmtKind};
 use rustpython_parser::{lexer, Mode, Tok};
 
 use ruff_diagnostics::Edit;
 use ruff_python_ast::helpers;
-use ruff_python_ast::helpers::to_absolute;
 use ruff_python_ast::imports::{AnyImport, Import};
 use ruff_python_ast::newlines::NewlineWithTrailingNewline;
 use ruff_python_ast::source_code::{Indexer, Locator, Stylist};
@@ -102,20 +102,17 @@ fn is_lone_child(child: &Stmt, parent: &Stmt, deleted: &[&Stmt]) -> Result<bool>
 
 /// Return the location of a trailing semicolon following a `Stmt`, if it's part
 /// of a multi-statement line.
-fn trailing_semicolon(stmt: &Stmt, locator: &Locator) -> Option<Location> {
-    let contents = locator.skip(stmt.end_location.unwrap());
-    for (row, line) in NewlineWithTrailingNewline::from(contents).enumerate() {
-        let trimmed = line.trim();
+fn trailing_semicolon(stmt: &Stmt, locator: &Locator) -> Option<TextSize> {
+    let contents = locator.after(stmt.end());
+
+    for line in NewlineWithTrailingNewline::from(contents) {
+        let trimmed = line.trim_start();
+
         if trimmed.starts_with(';') {
-            let column = line
-                .char_indices()
-                .find_map(|(column, char)| if char == ';' { Some(column) } else { None })
-                .unwrap();
-            return Some(to_absolute(
-                Location::new(row + 1, column),
-                stmt.end_location.unwrap(),
-            ));
+            let colon_offset = line.text_len() - trimmed.text_len();
+            return Some(stmt.end() + line.start() + colon_offset);
         }
+
         if !trimmed.starts_with('\\') {
             break;
         }
@@ -124,42 +121,36 @@ fn trailing_semicolon(stmt: &Stmt, locator: &Locator) -> Option<Location> {
 }
 
 /// Find the next valid break for a `Stmt` after a semicolon.
-fn next_stmt_break(semicolon: Location, locator: &Locator) -> Location {
-    let start_location = Location::new(semicolon.row(), semicolon.column() + 1);
-    let contents = locator.skip(start_location);
-    for (row, line) in NewlineWithTrailingNewline::from(contents).enumerate() {
+fn next_stmt_break(semicolon: TextSize, locator: &Locator) -> TextSize {
+    let start_location = semicolon + TextSize::from(1);
+
+    let contents = &locator.contents()[usize::from(start_location)..];
+    for line in NewlineWithTrailingNewline::from(contents) {
         let trimmed = line.trim();
         // Skip past any continuations.
         if trimmed.starts_with('\\') {
             continue;
         }
-        return if trimmed.is_empty() {
-            // If the line is empty, then despite the previous statement ending in a
-            // semicolon, we know that it's not a multi-statement line.
-            to_absolute(Location::new(row + 1, 0), start_location)
-        } else {
-            // Otherwise, find the start of the next statement. (Or, anything that isn't
-            // whitespace.)
-            let column = line
-                .char_indices()
-                .find_map(|(column, char)| {
-                    if char.is_whitespace() {
-                        None
-                    } else {
-                        Some(column)
-                    }
-                })
-                .unwrap();
-            to_absolute(Location::new(row + 1, column), start_location)
-        };
+
+        return start_location
+            + if trimmed.is_empty() {
+                // If the line is empty, then despite the previous statement ending in a
+                // semicolon, we know that it's not a multi-statement line.
+                line.start()
+            } else {
+                // Otherwise, find the start of the next statement. (Or, anything that isn't
+                // whitespace.)
+                let relative_offset = line.find(|c: char| !c.is_whitespace()).unwrap();
+                line.start() + TextSize::try_from(relative_offset).unwrap()
+            };
     }
-    Location::new(start_location.row() + 1, 0)
+
+    locator.line_end(start_location)
 }
 
 /// Return `true` if a `Stmt` occurs at the end of a file.
 fn is_end_of_file(stmt: &Stmt, locator: &Locator) -> bool {
-    let contents = locator.skip(stmt.end_location.unwrap());
-    contents.is_empty()
+    stmt.end() == locator.contents().text_len()
 }
 
 /// Return the `Fix` to use when deleting a `Stmt`.
@@ -190,33 +181,23 @@ pub fn delete_stmt(
     {
         // If removing this node would lead to an invalid syntax tree, replace
         // it with a `pass`.
-        Ok(Edit::replacement(
-            "pass".to_string(),
-            stmt.location,
-            stmt.end_location.unwrap(),
-        ))
+        Ok(Edit::range_replacement("pass".to_string(), stmt.range()))
     } else {
         Ok(if let Some(semicolon) = trailing_semicolon(stmt, locator) {
             let next = next_stmt_break(semicolon, locator);
-            Edit::deletion(stmt.location, next)
-        } else if helpers::match_leading_content(stmt, locator) {
-            Edit::deletion(stmt.location, stmt.end_location.unwrap())
-        } else if helpers::preceded_by_continuation(stmt, indexer) {
-            if is_end_of_file(stmt, locator) && stmt.location.column() == 0 {
+            Edit::deletion(stmt.start(), next)
+        } else if helpers::has_leading_content(stmt, locator) {
+            Edit::range_deletion(stmt.range())
+        } else if helpers::preceded_by_continuation(stmt, indexer, locator) {
+            if is_end_of_file(stmt, locator) && locator.is_at_start_of_line(stmt.start()) {
                 // Special-case: a file can't end in a continuation.
-                Edit::replacement(
-                    stylist.line_ending().to_string(),
-                    stmt.location,
-                    stmt.end_location.unwrap(),
-                )
+                Edit::range_replacement(stylist.line_ending().to_string(), stmt.range())
             } else {
-                Edit::deletion(stmt.location, stmt.end_location.unwrap())
+                Edit::range_deletion(stmt.range())
             }
         } else {
-            Edit::deletion(
-                Location::new(stmt.location.row(), 0),
-                Location::new(stmt.end_location.unwrap().row() + 1, 0),
-            )
+            let range = locator.full_lines_range(stmt.range());
+            Edit::range_deletion(range)
         })
     }
 }
@@ -231,7 +212,7 @@ pub fn remove_unused_imports<'a>(
     indexer: &Indexer,
     stylist: &Stylist,
 ) -> Result<Edit> {
-    let module_text = locator.slice(stmt);
+    let module_text = locator.slice(stmt.range());
     let mut tree = match_module(module_text)?;
 
     let Some(Statement::Simple(body)) = tree.body.first_mut() else {
@@ -337,11 +318,7 @@ pub fn remove_unused_imports<'a>(
         };
         tree.codegen(&mut state);
 
-        Ok(Edit::replacement(
-            state.to_string(),
-            stmt.location,
-            stmt.end_location.unwrap(),
-        ))
+        Ok(Edit::range_replacement(state.to_string(), stmt.range()))
     }
 }
 
@@ -353,15 +330,14 @@ pub fn remove_unused_imports<'a>(
 /// For this behavior, set `remove_parentheses` to `true`.
 pub fn remove_argument(
     locator: &Locator,
-    call_at: Location,
-    expr_at: Location,
-    expr_end: Location,
+    call_at: TextSize,
+    expr_range: TextRange,
     args: &[Expr],
     keywords: &[Keyword],
     remove_parentheses: bool,
 ) -> Result<Edit> {
     // TODO(sbrugman): Preserve trailing comments.
-    let contents = locator.skip(call_at);
+    let contents = locator.after(call_at);
 
     let mut fix_start = None;
     let mut fix_end = None;
@@ -374,13 +350,13 @@ pub fn remove_argument(
     if n_arguments == 1 {
         // Case 1: there is only one argument.
         let mut count: usize = 0;
-        for (start, tok, end) in lexer::lex_located(contents, Mode::Module, call_at).flatten() {
+        for (tok, range) in lexer::lex_located(contents, Mode::Module, call_at).flatten() {
             if matches!(tok, Tok::Lpar) {
                 if count == 0 {
                     fix_start = Some(if remove_parentheses {
-                        start
+                        range.start()
                     } else {
-                        Location::new(start.row(), start.column() + 1)
+                        range.start() + TextSize::from(1)
                     });
                 }
                 count += 1;
@@ -390,9 +366,9 @@ pub fn remove_argument(
                 count -= 1;
                 if count == 0 {
                     fix_end = Some(if remove_parentheses {
-                        end
+                        range.end()
                     } else {
-                        Location::new(end.row(), end.column() - 1)
+                        range.end() - TextSize::from(1)
                     });
                     break;
                 }
@@ -400,27 +376,27 @@ pub fn remove_argument(
         }
     } else if args
         .iter()
-        .map(|node| node.location)
-        .chain(keywords.iter().map(|node| node.location))
-        .any(|location| location > expr_at)
+        .map(Expr::start)
+        .chain(keywords.iter().map(Keyword::start))
+        .any(|location| location > expr_range.start())
     {
         // Case 2: argument or keyword is _not_ the last node.
         let mut seen_comma = false;
-        for (start, tok, end) in lexer::lex_located(contents, Mode::Module, call_at).flatten() {
+        for (tok, range) in lexer::lex_located(contents, Mode::Module, call_at).flatten() {
             if seen_comma {
                 if matches!(tok, Tok::NonLogicalNewline) {
                     // Also delete any non-logical newlines after the comma.
                     continue;
                 }
                 fix_end = Some(if matches!(tok, Tok::Newline) {
-                    end
+                    range.end()
                 } else {
-                    start
+                    range.start()
                 });
                 break;
             }
-            if start == expr_at {
-                fix_start = Some(start);
+            if range.start() == expr_range.start() {
+                fix_start = Some(range.start());
             }
             if fix_start.is_some() && matches!(tok, Tok::Comma) {
                 seen_comma = true;
@@ -429,13 +405,13 @@ pub fn remove_argument(
     } else {
         // Case 3: argument or keyword is the last node, so we have to find the last
         // comma in the stmt.
-        for (start, tok, _) in lexer::lex_located(contents, Mode::Module, call_at).flatten() {
-            if start == expr_at {
-                fix_end = Some(expr_end);
+        for (tok, range) in lexer::lex_located(contents, Mode::Module, call_at).flatten() {
+            if range.start() == expr_range.start() {
+                fix_end = Some(expr_range.end());
                 break;
             }
             if matches!(tok, Tok::Comma) {
-                fix_start = Some(start);
+                fix_start = Some(range.start());
             }
         }
     }
@@ -482,11 +458,8 @@ pub fn get_or_import_symbol(
         //
         // By adding this no-op edit, we force the `unused-imports` fix to conflict with the
         // `sys-exit-alias` fix, and thus will avoid applying both fixes in the same pass.
-        let import_edit = Edit::replacement(
-            locator.slice(source).to_string(),
-            source.location,
-            source.end_location.unwrap(),
-        );
+        let import_edit =
+            Edit::range_replacement(locator.slice(source.range()).to_string(), source.range());
         Ok((import_edit, binding))
     } else {
         if let Some(stmt) = importer.get_import_from(module) {
@@ -527,8 +500,8 @@ pub fn get_or_import_symbol(
 #[cfg(test)]
 mod tests {
     use anyhow::Result;
+    use ruff_text_size::TextSize;
     use rustpython_parser as parser;
-    use rustpython_parser::ast::Location;
 
     use ruff_python_ast::source_code::Locator;
 
@@ -546,19 +519,13 @@ mod tests {
         let program = parser::parse_program(contents, "<filename>")?;
         let stmt = program.first().unwrap();
         let locator = Locator::new(contents);
-        assert_eq!(
-            trailing_semicolon(stmt, &locator),
-            Some(Location::new(1, 5))
-        );
+        assert_eq!(trailing_semicolon(stmt, &locator), Some(TextSize::from(5)));
 
         let contents = "x = 1 ; y = 1";
         let program = parser::parse_program(contents, "<filename>")?;
         let stmt = program.first().unwrap();
         let locator = Locator::new(contents);
-        assert_eq!(
-            trailing_semicolon(stmt, &locator),
-            Some(Location::new(1, 6))
-        );
+        assert_eq!(trailing_semicolon(stmt, &locator), Some(TextSize::from(6)));
 
         let contents = r#"
 x = 1 \
@@ -568,10 +535,7 @@ x = 1 \
         let program = parser::parse_program(contents, "<filename>")?;
         let stmt = program.first().unwrap();
         let locator = Locator::new(contents);
-        assert_eq!(
-            trailing_semicolon(stmt, &locator),
-            Some(Location::new(2, 2))
-        );
+        assert_eq!(trailing_semicolon(stmt, &locator), Some(TextSize::from(10)));
 
         Ok(())
     }
@@ -581,15 +545,15 @@ x = 1 \
         let contents = "x = 1; y = 1";
         let locator = Locator::new(contents);
         assert_eq!(
-            next_stmt_break(Location::new(1, 4), &locator),
-            Location::new(1, 5)
+            next_stmt_break(TextSize::from(4), &locator),
+            TextSize::from(5)
         );
 
         let contents = "x = 1 ; y = 1";
         let locator = Locator::new(contents);
         assert_eq!(
-            next_stmt_break(Location::new(1, 5), &locator),
-            Location::new(1, 6)
+            next_stmt_break(TextSize::from(5), &locator),
+            TextSize::from(6)
         );
 
         let contents = r#"
@@ -599,8 +563,8 @@ x = 1 \
         .trim();
         let locator = Locator::new(contents);
         assert_eq!(
-            next_stmt_break(Location::new(2, 2), &locator),
-            Location::new(2, 4)
+            next_stmt_break(TextSize::from(10), &locator),
+            TextSize::from(12)
         );
     }
 }

crates/ruff/src/autofix/mod.rs

@@ -1,12 +1,11 @@
 use std::collections::BTreeSet;
 
 use itertools::Itertools;
+use ruff_text_size::{TextRange, TextSize};
 use rustc_hash::FxHashMap;
-use rustpython_parser::ast::Location;
 
 use ruff_diagnostics::{Diagnostic, Edit, Fix};
 use ruff_python_ast::source_code::Locator;
-use ruff_python_ast::types::Range;
 
 use crate::linter::FixTable;
 use crate::registry::{AsRule, Rule};
@@ -15,10 +14,15 @@ pub mod actions;
 
 /// Auto-fix errors in a file, and write the fixed source code to disk.
 pub fn fix_file(diagnostics: &[Diagnostic], locator: &Locator) -> Option<(String, FixTable)> {
-    if diagnostics.iter().all(|check| check.fix.is_empty()) {
+    let mut with_fixes = diagnostics
+        .iter()
+        .filter(|diag| !diag.fix.is_empty())
+        .peekable();
+
+    if with_fixes.peek().is_none() {
         None
     } else {
-        Some(apply_fixes(diagnostics.iter(), locator))
+        Some(apply_fixes(with_fixes, locator))
     }
 }
 
@@ -28,7 +32,7 @@ fn apply_fixes<'a>(
     locator: &'a Locator<'a>,
 ) -> (String, FixTable) {
     let mut output = String::with_capacity(locator.len());
-    let mut last_pos: Option<Location> = None;
+    let mut last_pos: Option<TextSize> = None;
     let mut applied: BTreeSet<&Edit> = BTreeSet::default();
     let mut fixed = FxHashMap::default();
 
@@ -52,7 +56,7 @@ fn apply_fixes<'a>(
         // Best-effort approach: if this fix overlaps with a fix we've already applied,
         // skip it.
         if last_pos.map_or(false, |last_pos| {
-            fix.location()
+            fix.min_start()
                 .map_or(false, |fix_location| last_pos >= fix_location)
         }) {
             continue;
@@ -60,14 +64,14 @@ fn apply_fixes<'a>(
 
         for edit in fix.edits() {
             // Add all contents from `last_pos` to `fix.location`.
-            let slice = locator.slice(Range::new(last_pos.unwrap_or_default(), edit.location));
+            let slice = locator.slice(TextRange::new(last_pos.unwrap_or_default(), edit.start()));
             output.push_str(slice);
 
             // Add the patch itself.
-            output.push_str(&edit.content);
+            output.push_str(edit.content().unwrap_or_default());
 
             // Track that the edit was applied.
-            last_pos = Some(edit.end_location);
+            last_pos = Some(edit.end());
             applied.insert(edit);
         }
 
@@ -75,7 +79,7 @@ fn apply_fixes<'a>(
     }
 
     // Add the remaining content.
-    let slice = locator.skip(last_pos.unwrap_or_default());
+    let slice = locator.after(last_pos.unwrap_or_default());
     output.push_str(slice);
 
     (output, fixed)
@@ -83,8 +87,8 @@ fn apply_fixes<'a>(
 
 /// Compare two fixes.
 fn cmp_fix(rule1: Rule, rule2: Rule, fix1: &Fix, fix2: &Fix) -> std::cmp::Ordering {
-    fix1.location()
-        .cmp(&fix2.location())
+    fix1.min_start()
+        .cmp(&fix2.min_start())
         .then_with(|| match (&rule1, &rule2) {
             // Apply `EndsInPeriod` fixes before `NewLineAfterLastParagraph` fixes.
             (Rule::EndsInPeriod, Rule::NewLineAfterLastParagraph) => std::cmp::Ordering::Less,
@@ -95,7 +99,7 @@ fn cmp_fix(rule1: Rule, rule2: Rule, fix1: &Fix, fix2: &Fix) -> std::cmp::Orderi
 
 #[cfg(test)]
 mod tests {
-    use rustpython_parser::ast::Location;
+    use ruff_text_size::TextSize;
 
     use ruff_diagnostics::Diagnostic;
     use ruff_diagnostics::Edit;
@@ -109,8 +113,7 @@ mod tests {
             .map(|edit| Diagnostic {
                 // The choice of rule here is arbitrary.
                 kind: MissingNewlineAtEndOfFile.into(),
-                location: edit.location,
-                end_location: edit.end_location,
+                range: edit.range(),
                 fix: edit.into(),
                 parent: None,
             })
@@ -135,11 +138,11 @@ class A(object):
 "#
             .trim(),
         );
-        let diagnostics = create_diagnostics([Edit {
-            content: "Bar".to_string(),
-            location: Location::new(1, 8),
-            end_location: Location::new(1, 14),
-        }]);
+        let diagnostics = create_diagnostics([Edit::replacement(
+            "Bar".to_string(),
+            TextSize::new(8),
+            TextSize::new(14),
+        )]);
         let (contents, fixed) = apply_fixes(diagnostics.iter(), &locator);
         assert_eq!(
             contents,
@@ -161,11 +164,7 @@ class A(object):
 "#
             .trim(),
         );
-        let diagnostics = create_diagnostics([Edit {
-            content: String::new(),
-            location: Location::new(1, 7),
-            end_location: Location::new(1, 15),
-        }]);
+        let diagnostics = create_diagnostics([Edit::deletion(TextSize::new(7), TextSize::new(15))]);
         let (contents, fixed) = apply_fixes(diagnostics.iter(), &locator);
         assert_eq!(
             contents,
@@ -188,16 +187,8 @@ class A(object, object, object):
             .trim(),
         );
         let diagnostics = create_diagnostics([
-            Edit {
-                content: String::new(),
-                location: Location::new(1, 8),
-                end_location: Location::new(1, 16),
-            },
-            Edit {
-                content: String::new(),
-                location: Location::new(1, 22),
-                end_location: Location::new(1, 30),
-            },
+            Edit::deletion(TextSize::from(8), TextSize::from(16)),
+            Edit::deletion(TextSize::from(22), TextSize::from(30)),
         ]);
         let (contents, fixed) = apply_fixes(diagnostics.iter(), &locator);
 
@@ -222,16 +213,8 @@ class A(object):
             .trim(),
         );
         let diagnostics = create_diagnostics([
-            Edit {
-                content: String::new(),
-                location: Location::new(1, 7),
-                end_location: Location::new(1, 15),
-            },
-            Edit {
-                content: "ignored".to_string(),
-                location: Location::new(1, 9),
-                end_location: Location::new(1, 11),
-            },
+            Edit::deletion(TextSize::from(7), TextSize::from(15)),
+            Edit::replacement("ignored".to_string(), TextSize::from(9), TextSize::from(11)),
         ]);
         let (contents, fixed) = apply_fixes(diagnostics.iter(), &locator);
         assert_eq!(

crates/ruff/src/checkers/ast/deferred.rs

@@ -1,14 +1,14 @@
-use ruff_python_semantic::scope::ScopeStack;
+use ruff_text_size::TextRange;
 use rustpython_parser::ast::{Expr, Stmt};
 
-use ruff_python_ast::types::Range;
 use ruff_python_ast::types::RefEquality;
 use ruff_python_semantic::analyze::visibility::{Visibility, VisibleScope};
+use ruff_python_semantic::scope::ScopeId;
 
 use crate::checkers::ast::AnnotationContext;
 use crate::docstrings::definition::Definition;
 
-type Context<'a> = (ScopeStack, Vec<RefEquality<'a, Stmt>>);
+type Context<'a> = (ScopeId, Vec<RefEquality<'a, Stmt>>);
 
 /// A collection of AST nodes that are deferred for later analysis.
 /// Used to, e.g., store functions, whose bodies shouldn't be analyzed until all
@@ -16,7 +16,7 @@ type Context<'a> = (ScopeStack, Vec<RefEquality<'a, Stmt>>);
 #[derive(Default)]
 pub struct Deferred<'a> {
     pub definitions: Vec<(Definition<'a>, Visibility, Context<'a>)>,
-    pub string_type_definitions: Vec<(Range, &'a str, AnnotationContext, Context<'a>)>,
+    pub string_type_definitions: Vec<(TextRange, &'a str, AnnotationContext, Context<'a>)>,
     pub type_definitions: Vec<(&'a Expr, AnnotationContext, Context<'a>)>,
     pub functions: Vec<(&'a Stmt, Context<'a>, VisibleScope)>,
     pub lambdas: Vec<(&'a Expr, Context<'a>)>,

crates/ruff/src/checkers/imports.rs

@@ -9,6 +9,7 @@ use ruff_python_ast::helpers::to_module_path;
 use ruff_python_ast::imports::{ImportMap, ModuleImport};
 use ruff_python_ast::source_code::{Indexer, Locator, Stylist};
 use ruff_python_ast::visitor::Visitor;
+use ruff_python_stdlib::path::is_python_stub_file;
 
 use crate::directives::IsortDirectives;
 use crate::registry::Rule;
@@ -29,13 +30,11 @@ fn extract_import_map(path: &Path, package: Option<&Path>, blocks: &[&Block]) ->
     for stmt in blocks.iter().flat_map(|block| &block.imports) {
         match &stmt.node {
             StmtKind::Import { names } => {
-                module_imports.extend(names.iter().map(|name| {
-                    ModuleImport::new(
-                        name.node.name.clone(),
-                        stmt.location,
-                        stmt.end_location.unwrap(),
-                    )
-                }));
+                module_imports.extend(
+                    names
+                        .iter()
+                        .map(|name| ModuleImport::new(name.node.name.clone(), stmt.range())),
+                );
             }
             StmtKind::ImportFrom {
                 module,
@@ -60,11 +59,7 @@ fn extract_import_map(path: &Path, package: Option<&Path>, blocks: &[&Block]) ->
                     Cow::Owned(module_path[..module_path.len() - level].join("."))
                 };
                 module_imports.extend(names.iter().map(|name| {
-                    ModuleImport::new(
-                        format!("{}.{}", module, name.node.name),
-                        name.location,
-                        name.end_location.unwrap(),
-                    )
+                    ModuleImport::new(format!("{}.{}", module, name.node.name), name.range())
                 }));
             }
             _ => panic!("Expected StmtKind::Import | StmtKind::ImportFrom"),
@@ -88,9 +83,11 @@ pub fn check_imports(
     path: &Path,
     package: Option<&Path>,
 ) -> (Vec<Diagnostic>, Option<ImportMap>) {
+    let is_stub = is_python_stub_file(path);
+
     // Extract all imports from the AST.
     let tracker = {
-        let mut tracker = ImportTracker::new(locator, directives, path);
+        let mut tracker = ImportTracker::new(locator, directives, is_stub);
         tracker.visit_body(python_ast);
         tracker
     };
@@ -111,7 +108,7 @@ pub fn check_imports(
     }
     if settings.rules.enabled(Rule::MissingRequiredImport) {
         diagnostics.extend(isort::rules::add_required_imports(
-            &blocks, python_ast, locator, stylist, settings, autofix,
+            &blocks, python_ast, locator, stylist, settings, autofix, is_stub,
         ));
     }
 

crates/ruff/src/checkers/logical_lines.rs

@@ -1,9 +1,9 @@
-use rustpython_parser::ast::Location;
+use ruff_text_size::TextRange;
 use rustpython_parser::lexer::LexResult;
 
-use ruff_diagnostics::{Diagnostic, Fix};
+use ruff_diagnostics::{Diagnostic, DiagnosticKind, Fix};
 use ruff_python_ast::source_code::{Locator, Stylist};
-use ruff_python_ast::types::Range;
+use ruff_python_ast::token_kind::TokenKind;
 
 use crate::registry::{AsRule, Rule};
 use crate::rules::pycodestyle::rules::logical_lines::{
@@ -37,7 +37,7 @@ pub fn check_logical_lines(
     settings: &Settings,
     autofix: flags::Autofix,
 ) -> Vec<Diagnostic> {
-    let mut diagnostics = vec![];
+    let mut context = LogicalLinesContext::new(settings);
 
     #[cfg(feature = "logical_lines")]
     let should_fix_missing_whitespace =
@@ -59,120 +59,51 @@ pub fn check_logical_lines(
 
     for line in &LogicalLines::from_tokens(tokens, locator) {
         if line.flags().contains(TokenFlags::OPERATOR) {
-            for (location, kind) in space_around_operator(&line) {
-                if settings.rules.enabled(kind.rule()) {
-                    diagnostics.push(Diagnostic {
-                        kind,
-                        location,
-                        end_location: location,
-                        fix: Fix::empty(),
-                        parent: None,
-                    });
-                }
-            }
-
-            for (location, kind) in whitespace_around_named_parameter_equals(&line.tokens()) {
-                if settings.rules.enabled(kind.rule()) {
-                    diagnostics.push(Diagnostic {
-                        kind,
-                        location,
-                        end_location: location,
-                        fix: Fix::empty(),
-                        parent: None,
-                    });
-                }
-            }
-            for (location, kind) in missing_whitespace_around_operator(&line.tokens()) {
-                if settings.rules.enabled(kind.rule()) {
-                    diagnostics.push(Diagnostic {
-                        kind,
-                        location,
-                        end_location: location,
-                        fix: Fix::empty(),
-                        parent: None,
-                    });
-                }
-            }
-
-            for diagnostic in missing_whitespace(&line, should_fix_missing_whitespace) {
-                if settings.rules.enabled(diagnostic.kind.rule()) {
-                    diagnostics.push(diagnostic);
-                }
-            }
+            space_around_operator(&line, &mut context);
+            whitespace_around_named_parameter_equals(&line, &mut context);
+            missing_whitespace_around_operator(&line, &mut context);
+            missing_whitespace(&line, should_fix_missing_whitespace, &mut context);
         }
+
         if line
             .flags()
             .contains(TokenFlags::OPERATOR | TokenFlags::PUNCTUATION)
         {
-            for (location, kind) in extraneous_whitespace(&line) {
-                if settings.rules.enabled(kind.rule()) {
-                    diagnostics.push(Diagnostic {
-                        kind,
-                        location,
-                        end_location: location,
-                        fix: Fix::empty(),
-                        parent: None,
-                    });
-                }
-            }
+            extraneous_whitespace(&line, &mut context);
         }
         if line.flags().contains(TokenFlags::KEYWORD) {
-            for (location, kind) in whitespace_around_keywords(&line) {
-                if settings.rules.enabled(kind.rule()) {
-                    diagnostics.push(Diagnostic {
-                        kind,
-                        location,
-                        end_location: location,
-                        fix: Fix::empty(),
-                        parent: None,
-                    });
-                }
-            }
-
-            for (location, kind) in missing_whitespace_after_keyword(&line.tokens()) {
-                if settings.rules.enabled(kind.rule()) {
-                    diagnostics.push(Diagnostic {
-                        kind,
-                        location,
-                        end_location: location,
-                        fix: Fix::empty(),
-                        parent: None,
-                    });
-                }
-            }
+            whitespace_around_keywords(&line, &mut context);
+            missing_whitespace_after_keyword(&line, &mut context);
         }
+
         if line.flags().contains(TokenFlags::COMMENT) {
-            for (range, kind) in whitespace_before_comment(&line.tokens(), locator) {
-                if settings.rules.enabled(kind.rule()) {
-                    diagnostics.push(Diagnostic {
-                        kind,
-                        location: range.location,
-                        end_location: range.end_location,
-                        fix: Fix::empty(),
-                        parent: None,
-                    });
-                }
-            }
+            whitespace_before_comment(&line, locator, prev_line.is_none(), &mut context);
         }
 
         if line.flags().contains(TokenFlags::BRACKET) {
-            for diagnostic in whitespace_before_parameters(
-                &line.tokens(),
+            whitespace_before_parameters(
+                &line,
                 should_fix_whitespace_before_parameters,
-            ) {
-                if settings.rules.enabled(diagnostic.kind.rule()) {
-                    diagnostics.push(diagnostic);
-                }
-            }
+                &mut context,
+            );
         }
 
         // Extract the indentation level.
-        let Some(start_loc) = line.first_token_location() else { continue; };
-        let start_line = locator.slice(Range::new(Location::new(start_loc.row(), 0), start_loc));
-        let indent_level = expand_indent(start_line);
+        let Some(first_token) = line.first_token() else {
+            continue;
+        };
+
+        let range = if first_token.kind() == TokenKind::Indent {
+            first_token.range()
+        } else {
+            TextRange::new(locator.line_start(first_token.start()), first_token.start())
+        };
+
+        let indent_level = expand_indent(locator.slice(range));
+
         let indent_size = 4;
 
-        for (location, kind) in indentation(
+        for kind in indentation(
             &line,
             prev_line.as_ref(),
             indent_char,
@@ -181,13 +112,7 @@ pub fn check_logical_lines(
             indent_size,
         ) {
             if settings.rules.enabled(kind.rule()) {
-                diagnostics.push(Diagnostic {
-                    kind,
-                    location,
-                    end_location: location,
-                    fix: Fix::empty(),
-                    parent: None,
-                });
+                context.push(kind, range);
             }
         }
 
@@ -196,7 +121,40 @@ pub fn check_logical_lines(
             prev_indent_level = Some(indent_level);
         }
     }
-    diagnostics
+    context.diagnostics
+}
+
+#[derive(Debug, Clone)]
+pub(crate) struct LogicalLinesContext<'a> {
+    settings: &'a Settings,
+    diagnostics: Vec<Diagnostic>,
+}
+
+impl<'a> LogicalLinesContext<'a> {
+    fn new(settings: &'a Settings) -> Self {
+        Self {
+            settings,
+            diagnostics: Vec::new(),
+        }
+    }
+
+    pub fn push<K: Into<DiagnosticKind>>(&mut self, kind: K, range: TextRange) {
+        let kind = kind.into();
+        if self.settings.rules.enabled(kind.rule()) {
+            self.diagnostics.push(Diagnostic {
+                kind,
+                range,
+                fix: Fix::empty(),
+                parent: None,
+            });
+        }
+    }
+
+    pub fn push_diagnostic(&mut self, diagnostic: Diagnostic) {
+        if self.settings.rules.enabled(diagnostic.kind.rule()) {
+            self.diagnostics.push(diagnostic);
+        }
+    }
 }
 
 #[cfg(test)]

crates/ruff/src/checkers/noqa.rs

@@ -1,15 +1,13 @@
 //! `NoQA` enforcement and validation.
 
-use nohash_hasher::IntMap;
-use rustpython_parser::ast::Location;
+use itertools::Itertools;
+use ruff_text_size::{TextLen, TextRange, TextSize};
 
 use ruff_diagnostics::{Diagnostic, Edit};
-use ruff_python_ast::newlines::StrExt;
-use ruff_python_ast::types::Range;
+use ruff_python_ast::source_code::Locator;
 
-use crate::codes::NoqaCode;
 use crate::noqa;
-use crate::noqa::{Directive, FileExemption};
+use crate::noqa::{Directive, FileExemption, NoqaDirectives, NoqaMapping};
 use crate::registry::{AsRule, Rule};
 use crate::rule_redirects::get_redirect_target;
 use crate::rules::ruff::rules::{UnusedCodes, UnusedNOQA};
@@ -17,37 +15,25 @@ use crate::settings::{flags, Settings};
 
 pub fn check_noqa(
     diagnostics: &mut Vec<Diagnostic>,
-    contents: &str,
-    commented_lines: &[usize],
-    noqa_line_for: &IntMap<usize, usize>,
+    locator: &Locator,
+    comment_ranges: &[TextRange],
+    noqa_line_for: &NoqaMapping,
     settings: &Settings,
     autofix: flags::Autofix,
 ) -> Vec<usize> {
     let enforce_noqa = settings.rules.enabled(Rule::UnusedNOQA);
 
-    let lines: Vec<&str> = contents.universal_newlines().collect();
-
     // Identify any codes that are globally exempted (within the current file).
-    let exemption = noqa::file_exemption(&lines, commented_lines);
-
-    // Map from line number to `noqa` directive on that line, along with any codes
-    // that were matched by the directive.
-    let mut noqa_directives: IntMap<usize, (Directive, Vec<NoqaCode>)> = IntMap::default();
+    let exemption = noqa::file_exemption(locator.contents(), comment_ranges);
 
     // Extract all `noqa` directives.
-    if enforce_noqa {
-        for lineno in commented_lines {
-            noqa_directives
-                .entry(lineno - 1)
-                .or_insert_with(|| (noqa::extract_noqa_directive(lines[lineno - 1]), vec![]));
-        }
-    }
+    let mut noqa_directives = NoqaDirectives::from_commented_ranges(comment_ranges, locator);
 
     // Indices of diagnostics that were ignored by a `noqa` directive.
     let mut ignored_diagnostics = vec![];
 
     // Remove any ignored diagnostics.
-    for (index, diagnostic) in diagnostics.iter().enumerate() {
+    'outer: for (index, diagnostic) in diagnostics.iter().enumerate() {
         if matches!(diagnostic.kind.rule(), Rule::BlanketNOQA) {
             continue;
         }
@@ -68,92 +54,65 @@ pub fn check_noqa(
             FileExemption::None => {}
         }
 
-        let diagnostic_lineno = diagnostic.location.row();
+        let noqa_offsets = diagnostic
+            .parent
+            .into_iter()
+            .chain(std::iter::once(diagnostic.start()))
+            .map(|position| noqa_line_for.resolve(position))
+            .unique();
 
-        // Is the violation ignored by a `noqa` directive on the parent line?
-        if let Some(parent_lineno) = diagnostic.parent.map(|location| location.row()) {
-            if parent_lineno != diagnostic_lineno {
-                let noqa_lineno = noqa_line_for.get(&parent_lineno).unwrap_or(&parent_lineno);
-                if commented_lines.contains(noqa_lineno) {
-                    let noqa = noqa_directives.entry(noqa_lineno - 1).or_insert_with(|| {
-                        (noqa::extract_noqa_directive(lines[noqa_lineno - 1]), vec![])
-                    });
-                    match noqa {
-                        (Directive::All(..), matches) => {
-                            matches.push(diagnostic.kind.rule().noqa_code());
-                            ignored_diagnostics.push(index);
-                            continue;
-                        }
-                        (Directive::Codes(.., codes, _), matches) => {
-                            if noqa::includes(diagnostic.kind.rule(), codes) {
-                                matches.push(diagnostic.kind.rule().noqa_code());
-                                ignored_diagnostics.push(index);
-                                continue;
-                            }
-                        }
-                        (Directive::None, ..) => {}
-                    }
-                }
-            }
-        }
-
-        // Is the diagnostic ignored by a `noqa` directive on the same line?
-        let noqa_lineno = noqa_line_for
-            .get(&diagnostic_lineno)
-            .unwrap_or(&diagnostic_lineno);
-        if commented_lines.contains(noqa_lineno) {
-            let noqa = noqa_directives
-                .entry(noqa_lineno - 1)
-                .or_insert_with(|| (noqa::extract_noqa_directive(lines[noqa_lineno - 1]), vec![]));
-            match noqa {
-                (Directive::All(..), matches) => {
-                    matches.push(diagnostic.kind.rule().noqa_code());
-                    ignored_diagnostics.push(index);
-                    continue;
-                }
-                (Directive::Codes(.., codes, _), matches) => {
-                    if noqa::includes(diagnostic.kind.rule(), codes) {
-                        matches.push(diagnostic.kind.rule().noqa_code());
+        for noqa_offset in noqa_offsets {
+            if let Some(directive_line) = noqa_directives.find_line_with_directive_mut(noqa_offset)
+            {
+                let suppressed = match &directive_line.directive {
+                    Directive::All(..) => {
+                        directive_line
+                            .matches
+                            .push(diagnostic.kind.rule().noqa_code());
                         ignored_diagnostics.push(index);
-                        continue;
+                        true
                     }
+                    Directive::Codes(.., codes, _) => {
+                        if noqa::includes(diagnostic.kind.rule(), codes) {
+                            directive_line
+                                .matches
+                                .push(diagnostic.kind.rule().noqa_code());
+                            ignored_diagnostics.push(index);
+                            true
+                        } else {
+                            false
+                        }
+                    }
+                    Directive::None => unreachable!(),
+                };
+
+                if suppressed {
+                    continue 'outer;
                 }
-                (Directive::None, ..) => {}
             }
         }
     }
 
     // Enforce that the noqa directive was actually used (RUF100).
     if enforce_noqa {
-        for (row, (directive, matches)) in noqa_directives {
-            match directive {
-                Directive::All(leading_spaces, start_byte, end_byte, trailing_spaces) => {
-                    if matches.is_empty() {
-                        let start_char = lines[row][..start_byte].chars().count();
-                        let end_char =
-                            start_char + lines[row][start_byte..end_byte].chars().count();
-
-                        let mut diagnostic = Diagnostic::new(
-                            UnusedNOQA { codes: None },
-                            Range::new(
-                                Location::new(row + 1, start_char),
-                                Location::new(row + 1, end_char),
-                            ),
-                        );
+        for line in noqa_directives.lines() {
+            match &line.directive {
+                Directive::All(leading_spaces, noqa_range, trailing_spaces) => {
+                    if line.matches.is_empty() {
+                        let mut diagnostic =
+                            Diagnostic::new(UnusedNOQA { codes: None }, *noqa_range);
                         if autofix.into() && settings.rules.should_fix(diagnostic.kind.rule()) {
                             diagnostic.set_fix(delete_noqa(
-                                row,
-                                lines[row],
-                                leading_spaces,
-                                start_byte,
-                                end_byte,
-                                trailing_spaces,
+                                *leading_spaces,
+                                *noqa_range,
+                                *trailing_spaces,
+                                locator,
                             ));
                         }
                         diagnostics.push(diagnostic);
                     }
                 }
-                Directive::Codes(leading_spaces, start_byte, end_byte, codes, trailing_spaces) => {
+                Directive::Codes(leading_spaces, range, codes, trailing_spaces) => {
                     let mut disabled_codes = vec![];
                     let mut unknown_codes = vec![];
                     let mut unmatched_codes = vec![];
@@ -166,7 +125,9 @@ pub fn check_noqa(
                             break;
                         }
 
-                        if matches.iter().any(|m| *m == code) || settings.external.contains(code) {
+                        if line.matches.iter().any(|m| *m == code)
+                            || settings.external.contains(code)
+                        {
                             valid_codes.push(code);
                         } else {
                             if let Ok(rule) = Rule::from_code(code) {
@@ -189,10 +150,6 @@ pub fn check_noqa(
                         && unknown_codes.is_empty()
                         && unmatched_codes.is_empty())
                     {
-                        let start_char = lines[row][..start_byte].chars().count();
-                        let end_char =
-                            start_char + lines[row][start_byte..end_byte].chars().count();
-
                         let mut diagnostic = Diagnostic::new(
                             UnusedNOQA {
                                 codes: Some(UnusedCodes {
@@ -210,26 +167,20 @@ pub fn check_noqa(
                                         .collect(),
                                 }),
                             },
-                            Range::new(
-                                Location::new(row + 1, start_char),
-                                Location::new(row + 1, end_char),
-                            ),
+                            *range,
                         );
                         if autofix.into() && settings.rules.should_fix(diagnostic.kind.rule()) {
                             if valid_codes.is_empty() {
                                 diagnostic.set_fix(delete_noqa(
-                                    row,
-                                    lines[row],
-                                    leading_spaces,
-                                    start_byte,
-                                    end_byte,
-                                    trailing_spaces,
+                                    *leading_spaces,
+                                    *range,
+                                    *trailing_spaces,
+                                    locator,
                                 ));
                             } else {
-                                diagnostic.set_fix(Edit::replacement(
+                                diagnostic.set_fix(Edit::range_replacement(
                                     format!("# noqa: {}", valid_codes.join(", ")),
-                                    Location::new(row + 1, start_char),
-                                    Location::new(row + 1, end_char),
+                                    *range,
                                 ));
                             }
                         }
@@ -247,39 +198,37 @@ pub fn check_noqa(
 
 /// Generate a [`Edit`] to delete a `noqa` directive.
 fn delete_noqa(
-    row: usize,
-    line: &str,
-    leading_spaces: usize,
-    start_byte: usize,
-    end_byte: usize,
-    trailing_spaces: usize,
+    leading_spaces: TextSize,
+    noqa_range: TextRange,
+    trailing_spaces: TextSize,
+    locator: &Locator,
 ) -> Edit {
-    if start_byte - leading_spaces == 0 && end_byte == line.len() {
-        // Ex) `# noqa`
-        Edit::deletion(Location::new(row + 1, 0), Location::new(row + 2, 0))
-    } else if end_byte == line.len() {
-        // Ex) `x = 1  # noqa`
-        let start_char = line[..start_byte].chars().count();
-        let end_char = start_char + line[start_byte..end_byte].chars().count();
-        Edit::deletion(
-            Location::new(row + 1, start_char - leading_spaces),
-            Location::new(row + 1, end_char + trailing_spaces),
+    let line_range = locator.line_range(noqa_range.start());
+
+    // Ex) `# noqa`
+    if line_range
+        == TextRange::new(
+            noqa_range.start() - leading_spaces,
+            noqa_range.end() + trailing_spaces,
         )
-    } else if line[end_byte..].trim_start().starts_with('#') {
-        // Ex) `x = 1  # noqa  # type: ignore`
-        let start_char = line[..start_byte].chars().count();
-        let end_char = start_char + line[start_byte..end_byte].chars().count();
+    {
+        let full_line_end = locator.full_line_end(line_range.end());
+        Edit::deletion(line_range.start(), full_line_end)
+    }
+    // Ex) `x = 1  # noqa`
+    else if noqa_range.end() + trailing_spaces == line_range.end() {
+        Edit::deletion(noqa_range.start() - leading_spaces, line_range.end())
+    }
+    // Ex) `x = 1  # noqa  # type: ignore`
+    else if locator.contents()[usize::from(noqa_range.end() + trailing_spaces)..].starts_with('#')
+    {
+        Edit::deletion(noqa_range.start(), noqa_range.end() + trailing_spaces)
+    }
+    // Ex) `x = 1  # noqa here`
+    else {
         Edit::deletion(
-            Location::new(row + 1, start_char),
-            Location::new(row + 1, end_char + trailing_spaces),
-        )
-    } else {
-        // Ex) `x = 1  # noqa here`
-        let start_char = line[..start_byte].chars().count();
-        let end_char = start_char + line[start_byte..end_byte].chars().count();
-        Edit::deletion(
-            Location::new(row + 1, start_char + 1 + 1),
-            Location::new(row + 1, end_char + trailing_spaces),
+            noqa_range.start() + "# ".text_len(),
+            noqa_range.end() + trailing_spaces,
         )
     }
 }

crates/ruff/src/checkers/physical_lines.rs

@@ -1,10 +1,11 @@
 //! Lint rules based on checking physical lines.
 
+use ruff_text_size::TextSize;
 use std::path::Path;
 
 use ruff_diagnostics::Diagnostic;
 use ruff_python_ast::newlines::StrExt;
-use ruff_python_ast::source_code::{Locator, Stylist};
+use ruff_python_ast::source_code::{Indexer, Locator, Stylist};
 
 use crate::registry::Rule;
 use crate::rules::flake8_executable::helpers::{extract_shebang, ShebangDirective};
@@ -24,8 +25,8 @@ pub fn check_physical_lines(
     path: &Path,
     locator: &Locator,
     stylist: &Stylist,
-    commented_lines: &[usize],
-    doc_lines: &[usize],
+    indexer: &Indexer,
+    doc_lines: &[TextSize],
     settings: &Settings,
     autofix: flags::Autofix,
 ) -> Vec<Diagnostic> {
@@ -55,17 +56,19 @@ pub fn check_physical_lines(
     let fix_shebang_whitespace =
         autofix.into() && settings.rules.should_fix(Rule::ShebangLeadingWhitespace);
 
-    let mut commented_lines_iter = commented_lines.iter().peekable();
+    let mut commented_lines_iter = indexer.comment_ranges().iter().peekable();
     let mut doc_lines_iter = doc_lines.iter().peekable();
+    let string_lines = indexer.triple_quoted_string_ranges();
+
     for (index, line) in locator.contents().universal_newlines().enumerate() {
         while commented_lines_iter
-            .next_if(|lineno| &(index + 1) == *lineno)
+            .next_if(|comment_range| line.range().contains_range(**comment_range))
             .is_some()
         {
             if enforce_unnecessary_coding_comment {
                 if index < 2 {
                     if let Some(diagnostic) =
-                        unnecessary_coding_comment(index, line, fix_unnecessary_coding_comment)
+                        unnecessary_coding_comment(&line, fix_unnecessary_coding_comment)
                     {
                         diagnostics.push(diagnostic);
                     }
@@ -73,15 +76,11 @@ pub fn check_physical_lines(
             }
 
             if enforce_blanket_type_ignore {
-                if let Some(diagnostic) = blanket_type_ignore(index, line) {
-                    diagnostics.push(diagnostic);
-                }
+                blanket_type_ignore(&mut diagnostics, &line);
             }
 
             if enforce_blanket_noqa {
-                if let Some(diagnostic) = blanket_noqa(index, line) {
-                    diagnostics.push(diagnostic);
-                }
+                blanket_noqa(&mut diagnostics, &line);
             }
 
             if enforce_shebang_missing
@@ -90,31 +89,31 @@ pub fn check_physical_lines(
                 || enforce_shebang_newline
                 || enforce_shebang_python
             {
-                let shebang = extract_shebang(line);
+                let shebang = extract_shebang(&line);
                 if enforce_shebang_not_executable {
-                    if let Some(diagnostic) = shebang_not_executable(path, index, &shebang) {
+                    if let Some(diagnostic) = shebang_not_executable(path, line.range(), &shebang) {
                         diagnostics.push(diagnostic);
                     }
                 }
                 if enforce_shebang_missing {
-                    if !has_any_shebang && matches!(shebang, ShebangDirective::Match(_, _, _, _)) {
+                    if !has_any_shebang && matches!(shebang, ShebangDirective::Match(..)) {
                         has_any_shebang = true;
                     }
                 }
                 if enforce_shebang_whitespace {
                     if let Some(diagnostic) =
-                        shebang_whitespace(index, &shebang, fix_shebang_whitespace)
+                        shebang_whitespace(line.range(), &shebang, fix_shebang_whitespace)
                     {
                         diagnostics.push(diagnostic);
                     }
                 }
                 if enforce_shebang_newline {
-                    if let Some(diagnostic) = shebang_newline(index, &shebang) {
+                    if let Some(diagnostic) = shebang_newline(line.range(), &shebang, index == 0) {
                         diagnostics.push(diagnostic);
                     }
                 }
                 if enforce_shebang_python {
-                    if let Some(diagnostic) = shebang_python(index, &shebang) {
+                    if let Some(diagnostic) = shebang_python(line.range(), &shebang) {
                         diagnostics.push(diagnostic);
                     }
                 }
@@ -122,40 +121,40 @@ pub fn check_physical_lines(
         }
 
         while doc_lines_iter
-            .next_if(|lineno| &(index + 1) == *lineno)
+            .next_if(|doc_line_start| line.range().contains(**doc_line_start))
             .is_some()
         {
             if enforce_doc_line_too_long {
-                if let Some(diagnostic) = doc_line_too_long(index, line, settings) {
+                if let Some(diagnostic) = doc_line_too_long(&line, settings) {
                     diagnostics.push(diagnostic);
                 }
             }
         }
 
         if enforce_mixed_spaces_and_tabs {
-            if let Some(diagnostic) = mixed_spaces_and_tabs(index, line) {
+            if let Some(diagnostic) = mixed_spaces_and_tabs(&line) {
                 diagnostics.push(diagnostic);
             }
         }
 
         if enforce_line_too_long {
-            if let Some(diagnostic) = line_too_long(index, line, settings) {
+            if let Some(diagnostic) = line_too_long(&line, settings) {
                 diagnostics.push(diagnostic);
             }
         }
 
         if enforce_bidirectional_unicode {
-            diagnostics.extend(pylint::rules::bidirectional_unicode(index, line));
+            diagnostics.extend(pylint::rules::bidirectional_unicode(&line));
         }
 
         if enforce_trailing_whitespace || enforce_blank_line_contains_whitespace {
-            if let Some(diagnostic) = trailing_whitespace(index, line, settings, autofix) {
+            if let Some(diagnostic) = trailing_whitespace(&line, settings, autofix) {
                 diagnostics.push(diagnostic);
             }
         }
 
         if enforce_tab_indentation {
-            if let Some(diagnostic) = tab_indentation(index, line) {
+            if let Some(diagnostic) = tab_indentation(&line, string_lines) {
                 diagnostics.push(diagnostic);
             }
         }
@@ -186,7 +185,7 @@ mod tests {
     use rustpython_parser::Mode;
     use std::path::Path;
 
-    use ruff_python_ast::source_code::{Locator, Stylist};
+    use ruff_python_ast::source_code::{Indexer, Locator, Stylist};
 
     use crate::registry::Rule;
     use crate::settings::{flags, Settings};
@@ -198,6 +197,7 @@ mod tests {
         let line = "'\u{4e9c}' * 2"; // 7 in UTF-32, 9 in UTF-8.
         let locator = Locator::new(line);
         let tokens: Vec<_> = lex(line, Mode::Module).collect();
+        let indexer = Indexer::from_tokens(&tokens, &locator);
         let stylist = Stylist::from_tokens(&tokens, &locator);
 
         let check_with_max_line_length = |line_length: usize| {
@@ -205,7 +205,7 @@ mod tests {
                 Path::new("foo.py"),
                 &locator,
                 &stylist,
-                &[],
+                &indexer,
                 &[],
                 &Settings {
                     line_length,

crates/ruff/src/checkers/tokens.rs

@@ -64,7 +64,7 @@ pub fn check_tokens(
     // RUF001, RUF002, RUF003
     if enforce_ambiguous_unicode_character {
         let mut state_machine = StateMachine::default();
-        for &(start, ref tok, end) in tokens.iter().flatten() {
+        for &(ref tok, range) in tokens.iter().flatten() {
             let is_docstring = if enforce_ambiguous_unicode_character {
                 state_machine.consume(tok)
             } else {
@@ -74,8 +74,7 @@ pub fn check_tokens(
             if matches!(tok, Tok::String { .. } | Tok::Comment(_)) {
                 diagnostics.extend(ruff::rules::ambiguous_unicode_character(
                     locator,
-                    start,
-                    end,
+                    range,
                     if matches!(tok, Tok::String { .. }) {
                         if is_docstring {
                             Context::Docstring
@@ -94,10 +93,10 @@ pub fn check_tokens(
 
     // ERA001
     if enforce_commented_out_code {
-        for (start, tok, end) in tokens.iter().flatten() {
+        for (tok, range) in tokens.iter().flatten() {
             if matches!(tok, Tok::Comment(_)) {
                 if let Some(diagnostic) =
-                    eradicate::rules::commented_out_code(locator, *start, *end, settings, autofix)
+                    eradicate::rules::commented_out_code(locator, *range, settings, autofix)
                 {
                     diagnostics.push(diagnostic);
                 }
@@ -107,12 +106,11 @@ pub fn check_tokens(
 
     // W605
     if enforce_invalid_escape_sequence {
-        for (start, tok, end) in tokens.iter().flatten() {
+        for (tok, range) in tokens.iter().flatten() {
             if matches!(tok, Tok::String { .. }) {
                 diagnostics.extend(pycodestyle::rules::invalid_escape_sequence(
                     locator,
-                    *start,
-                    *end,
+                    *range,
                     autofix.into() && settings.rules.should_fix(Rule::InvalidEscapeSequence),
                 ));
             }
@@ -120,10 +118,10 @@ pub fn check_tokens(
     }
     // PLE2510, PLE2512, PLE2513
     if enforce_invalid_string_character {
-        for (start, tok, end) in tokens.iter().flatten() {
+        for (tok, range) in tokens.iter().flatten() {
             if matches!(tok, Tok::String { .. }) {
                 diagnostics.extend(
-                    pylint::rules::invalid_string_characters(locator, *start, *end, autofix.into())
+                    pylint::rules::invalid_string_characters(locator, *range, autofix.into())
                         .into_iter()
                         .filter(|diagnostic| settings.rules.enabled(diagnostic.kind.rule())),
                 );
@@ -155,6 +153,7 @@ pub fn check_tokens(
             flake8_implicit_str_concat::rules::implicit(
                 tokens,
                 &settings.flake8_implicit_str_concat,
+                locator,
             )
             .into_iter()
             .filter(|diagnostic| settings.rules.enabled(diagnostic.kind.rule())),

crates/ruff/src/codes.rs

@@ -1,8 +1,15 @@
 use crate::registry::{Linter, Rule};
+use std::fmt::Formatter;
 
 #[derive(PartialEq, Eq, PartialOrd, Ord)]
 pub struct NoqaCode(&'static str, &'static str);
 
+impl std::fmt::Debug for NoqaCode {
+    fn fmt(&self, f: &mut Formatter<'_>) -> std::fmt::Result {
+        std::fmt::Display::fmt(self, f)
+    }
+}
+
 impl std::fmt::Display for NoqaCode {
     fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> Result<(), std::fmt::Error> {
         write!(f, "{}{}", self.0, self.1)
@@ -202,6 +209,7 @@ pub fn code_to_rule(linter: Linter, code: &str) -> Option<Rule> {
         (Pylint, "W0711") => Rule::BinaryOpException,
         (Pylint, "W1508") => Rule::InvalidEnvvarDefault,
         (Pylint, "W2901") => Rule::RedefinedLoopName,
+        (Pylint, "E0302") => Rule::UnexpectedSpecialMethodSignature,
 
         // flake8-builtins
         (Flake8Builtins, "001") => Rule::BuiltinVariableShadowing,
@@ -263,6 +271,8 @@ pub fn code_to_rule(linter: Linter, code: &str) -> Option<Rule> {
         (Flake8Comprehensions, "15") => Rule::UnnecessarySubscriptReversal,
         (Flake8Comprehensions, "16") => Rule::UnnecessaryComprehension,
         (Flake8Comprehensions, "17") => Rule::UnnecessaryMap,
+        (Flake8Comprehensions, "18") => Rule::UnnecessaryLiteralWithinDictCall,
+        (Flake8Comprehensions, "19") => Rule::UnnecessaryComprehensionAnyAll,
 
         // flake8-debugger
         (Flake8Debugger, "0") => Rule::Debugger,
@@ -507,6 +517,12 @@ pub fn code_to_rule(linter: Linter, code: &str) -> Option<Rule> {
         (Flake8Bandit, "506") => Rule::UnsafeYAMLLoad,
         (Flake8Bandit, "508") => Rule::SnmpInsecureVersion,
         (Flake8Bandit, "509") => Rule::SnmpWeakCryptography,
+        (Flake8Bandit, "602") => Rule::SubprocessPopenWithShellEqualsTrue,
+        (Flake8Bandit, "603") => Rule::SubprocessWithoutShellEqualsTrue,
+        (Flake8Bandit, "604") => Rule::CallWithShellEqualsTrue,
+        (Flake8Bandit, "605") => Rule::StartProcessWithAShell,
+        (Flake8Bandit, "606") => Rule::StartProcessWithNoShell,
+        (Flake8Bandit, "607") => Rule::StartProcessWithPartialPath,
         (Flake8Bandit, "608") => Rule::HardcodedSQLExpression,
         (Flake8Bandit, "612") => Rule::LoggingConfigInsecureListen,
         (Flake8Bandit, "701") => Rule::Jinja2AutoescapeFalse,
@@ -525,6 +541,8 @@ pub fn code_to_rule(linter: Linter, code: &str) -> Option<Rule> {
 
         // flake8-import-conventions
         (Flake8ImportConventions, "001") => Rule::UnconventionalImportAlias,
+        (Flake8ImportConventions, "002") => Rule::BannedImportAlias,
+        (Flake8ImportConventions, "003") => Rule::BannedImportFrom,
 
         // flake8-datetimez
         (Flake8Datetimez, "001") => Rule::CallDatetimeWithoutTzinfo,
@@ -573,6 +591,7 @@ pub fn code_to_rule(linter: Linter, code: &str) -> Option<Rule> {
         (Flake8Pyi, "012") => Rule::PassInClassBody,
         (Flake8Pyi, "014") => Rule::ArgumentDefaultInStub,
         (Flake8Pyi, "015") => Rule::AssignmentDefaultInStub,
+        (Flake8Pyi, "016") => Rule::DuplicateUnionMember,
         (Flake8Pyi, "021") => Rule::DocstringInStub,
         (Flake8Pyi, "033") => Rule::TypeCommentInStub,
 
@@ -608,7 +627,6 @@ pub fn code_to_rule(linter: Linter, code: &str) -> Option<Rule> {
         (Flake8Pie, "794") => Rule::DuplicateClassFieldDefinition,
         (Flake8Pie, "796") => Rule::NonUniqueEnums,
         (Flake8Pie, "800") => Rule::UnnecessarySpread,
-        (Flake8Pie, "802") => Rule::UnnecessaryComprehensionAnyAll,
         (Flake8Pie, "804") => Rule::UnnecessaryDictKwargs,
         (Flake8Pie, "807") => Rule::ReimplementedListBuiltin,
         (Flake8Pie, "810") => Rule::MultipleStartsEndsWith,
@@ -700,6 +718,8 @@ pub fn code_to_rule(linter: Linter, code: &str) -> Option<Rule> {
         (Ruff, "005") => Rule::CollectionLiteralConcatenation,
         (Ruff, "006") => Rule::AsyncioDanglingTask,
         (Ruff, "007") => Rule::PairwiseOverZipped,
+        (Ruff, "008") => Rule::MutableDataclassDefault,
+        (Ruff, "009") => Rule::FunctionCallInDataclassDefaultArgument,
         (Ruff, "100") => Rule::UnusedNOQA,
 
         // flake8-django

crates/ruff/src/directives.rs

@@ -1,16 +1,18 @@
 //! Extract `# noqa` and `# isort: skip` directives from tokenized source.
 
+use crate::noqa::NoqaMapping;
 use bitflags::bitflags;
-use nohash_hasher::{IntMap, IntSet};
-use rustpython_parser::ast::Location;
+use ruff_python_ast::source_code::{Indexer, Locator};
+use ruff_text_size::{TextLen, TextRange, TextSize};
 use rustpython_parser::lexer::LexResult;
 use rustpython_parser::Tok;
 
 use crate::settings::Settings;
 
 bitflags! {
-    pub struct Flags: u32 {
-        const NOQA = 0b0000_0001;
+    #[derive(Debug, Copy, Clone)]
+    pub struct Flags: u8 {
+        const NOQA  = 0b0000_0001;
         const ISORT = 0b0000_0010;
     }
 }
@@ -29,27 +31,50 @@ impl Flags {
     }
 }
 
-#[derive(Default)]
+#[derive(Default, Debug)]
 pub struct IsortDirectives {
-    pub exclusions: IntSet<usize>,
-    pub splits: Vec<usize>,
+    /// Ranges for which sorting is disabled
+    pub exclusions: Vec<TextRange>,
+    /// Text positions at which splits should be inserted
+    pub splits: Vec<TextSize>,
     pub skip_file: bool,
 }
 
+impl IsortDirectives {
+    pub fn is_excluded(&self, offset: TextSize) -> bool {
+        for range in &self.exclusions {
+            if range.contains(offset) {
+                return true;
+            }
+
+            if range.start() > offset {
+                break;
+            }
+        }
+
+        false
+    }
+}
+
 pub struct Directives {
-    pub noqa_line_for: IntMap<usize, usize>,
+    pub noqa_line_for: NoqaMapping,
     pub isort: IsortDirectives,
 }
 
-pub fn extract_directives(lxr: &[LexResult], flags: Flags) -> Directives {
+pub fn extract_directives(
+    lxr: &[LexResult],
+    flags: Flags,
+    locator: &Locator,
+    indexer: &Indexer,
+) -> Directives {
     Directives {
         noqa_line_for: if flags.contains(Flags::NOQA) {
-            extract_noqa_line_for(lxr)
+            extract_noqa_line_for(lxr, locator, indexer)
         } else {
-            IntMap::default()
+            NoqaMapping::default()
         },
         isort: if flags.contains(Flags::ISORT) {
-            extract_isort_directives(lxr)
+            extract_isort_directives(lxr, locator)
         } else {
             IsortDirectives::default()
         },
@@ -57,48 +82,92 @@ pub fn extract_directives(lxr: &[LexResult], flags: Flags) -> Directives {
 }
 
 /// Extract a mapping from logical line to noqa line.
-pub fn extract_noqa_line_for(lxr: &[LexResult]) -> IntMap<usize, usize> {
-    let mut noqa_line_for: IntMap<usize, usize> = IntMap::default();
-    let mut prev_non_newline: Option<(&Location, &Tok, &Location)> = None;
-    for (start, tok, end) in lxr.iter().flatten() {
-        if matches!(tok, Tok::EndOfFile) {
-            break;
-        }
-        // For multi-line strings, we expect `noqa` directives on the last line of the
-        // string.
-        if matches!(tok, Tok::String { .. }) && end.row() > start.row() {
-            for i in start.row()..end.row() {
-                noqa_line_for.insert(i, end.row());
+pub fn extract_noqa_line_for(
+    lxr: &[LexResult],
+    locator: &Locator,
+    indexer: &Indexer,
+) -> NoqaMapping {
+    let mut string_mappings = Vec::new();
+
+    for (tok, range) in lxr.iter().flatten() {
+        match tok {
+            Tok::EndOfFile => {
+                break;
             }
-        }
-        // For continuations, we expect `noqa` directives on the last line of the
-        // continuation.
-        if matches!(
-            tok,
-            Tok::Newline | Tok::NonLogicalNewline | Tok::Comment(..)
-        ) {
-            if let Some((.., end)) = prev_non_newline {
-                for i in end.row()..start.row() {
-                    noqa_line_for.insert(i, start.row());
+
+            // For multi-line strings, we expect `noqa` directives on the last line of the
+            // string.
+            Tok::String {
+                triple_quoted: true,
+                ..
+            } => {
+                if locator.contains_line_break(*range) {
+                    string_mappings.push(*range);
                 }
             }
-            prev_non_newline = None;
-        } else if prev_non_newline.is_none() {
-            prev_non_newline = Some((start, tok, end));
+
+            _ => {}
         }
     }
-    noqa_line_for
+
+    let mut continuation_mappings = Vec::new();
+
+    // For continuations, we expect `noqa` directives on the last line of the
+    // continuation.
+    let mut last: Option<TextRange> = None;
+    for continuation_line in indexer.continuation_line_starts() {
+        let line_end = locator.full_line_end(*continuation_line);
+        if let Some(last_range) = last.take() {
+            if last_range.end() == *continuation_line {
+                last = Some(TextRange::new(last_range.start(), line_end));
+                continue;
+            }
+            // new continuation
+            continuation_mappings.push(last_range);
+        }
+
+        last = Some(TextRange::new(*continuation_line, line_end));
+    }
+
+    if let Some(last_range) = last.take() {
+        continuation_mappings.push(last_range);
+    }
+
+    // Merge the mappings in sorted order
+    let mut mappings =
+        NoqaMapping::with_capacity(continuation_mappings.len() + string_mappings.len());
+
+    let mut continuation_mappings = continuation_mappings.into_iter().peekable();
+    let mut string_mappings = string_mappings.into_iter().peekable();
+
+    while let (Some(continuation), Some(string)) =
+        (continuation_mappings.peek(), string_mappings.peek())
+    {
+        if continuation.start() <= string.start() {
+            mappings.push_mapping(continuation_mappings.next().unwrap());
+        } else {
+            mappings.push_mapping(string_mappings.next().unwrap());
+        }
+    }
+
+    for mapping in continuation_mappings {
+        mappings.push_mapping(mapping);
+    }
+
+    for mapping in string_mappings {
+        mappings.push_mapping(mapping);
+    }
+
+    mappings
 }
 
-/// Extract a set of lines over which to disable isort.
-pub fn extract_isort_directives(lxr: &[LexResult]) -> IsortDirectives {
-    let mut exclusions: IntSet<usize> = IntSet::default();
-    let mut splits: Vec<usize> = Vec::default();
-    let mut off: Option<Location> = None;
-    let mut last: Option<Location> = None;
-    for &(start, ref tok, end) in lxr.iter().flatten() {
-        last = Some(end);
+/// Extract a set of ranges over which to disable isort.
+pub fn extract_isort_directives(lxr: &[LexResult], locator: &Locator) -> IsortDirectives {
+    let mut exclusions: Vec<TextRange> = Vec::default();
+    let mut splits: Vec<TextSize> = Vec::default();
+    let mut off: Option<TextSize> = None;
 
+    for &(ref tok, range) in lxr.iter().flatten() {
         let Tok::Comment(comment_text) = tok else {
             continue;
         };
@@ -108,7 +177,7 @@ pub fn extract_isort_directives(lxr: &[LexResult]) -> IsortDirectives {
         // required to include the space, and must appear on their own lines.
         let comment_text = comment_text.trim_end();
         if matches!(comment_text, "# isort: split" | "# ruff: isort: split") {
-            splits.push(start.row());
+            splits.push(range.start());
         } else if matches!(
             comment_text,
             "# isort: skip_file"
@@ -122,30 +191,25 @@ pub fn extract_isort_directives(lxr: &[LexResult]) -> IsortDirectives {
             };
         } else if off.is_some() {
             if comment_text == "# isort: on" || comment_text == "# ruff: isort: on" {
-                if let Some(start) = off {
-                    for row in start.row() + 1..=end.row() {
-                        exclusions.insert(row);
-                    }
+                if let Some(exclusion_start) = off {
+                    exclusions.push(TextRange::new(exclusion_start, range.start()));
                 }
                 off = None;
             }
         } else {
             if comment_text.contains("isort: skip") || comment_text.contains("isort:skip") {
-                exclusions.insert(start.row());
+                exclusions.push(locator.line_range(range.start()));
             } else if comment_text == "# isort: off" || comment_text == "# ruff: isort: off" {
-                off = Some(start);
+                off = Some(range.start());
             }
         }
     }
 
     if let Some(start) = off {
         // Enforce unterminated `isort: off`.
-        if let Some(end) = last {
-            for row in start.row() + 1..=end.row() {
-                exclusions.insert(row);
-            }
-        }
+        exclusions.push(TextRange::new(start, locator.contents().text_len()));
     }
+
     IsortDirectives {
         exclusions,
         splits,
@@ -155,120 +219,98 @@ pub fn extract_isort_directives(lxr: &[LexResult]) -> IsortDirectives {
 
 #[cfg(test)]
 mod tests {
-    use nohash_hasher::{IntMap, IntSet};
+    use ruff_python_ast::source_code::{Indexer, Locator};
+    use ruff_text_size::{TextLen, TextRange, TextSize};
     use rustpython_parser::lexer::LexResult;
     use rustpython_parser::{lexer, Mode};
 
     use crate::directives::{extract_isort_directives, extract_noqa_line_for};
+    use crate::noqa::NoqaMapping;
+
+    fn noqa_mappings(contents: &str) -> NoqaMapping {
+        let lxr: Vec<LexResult> = lexer::lex(contents, Mode::Module).collect();
+        let locator = Locator::new(contents);
+        let indexer = Indexer::from_tokens(&lxr, &locator);
+
+        extract_noqa_line_for(&lxr, &locator, &indexer)
+    }
 
     #[test]
     fn noqa_extraction() {
-        let lxr: Vec<LexResult> = lexer::lex(
-            "x = 1
-y = 2
-z = x + 1",
-            Mode::Module,
-        )
-        .collect();
-        assert_eq!(extract_noqa_line_for(&lxr), IntMap::default());
+        let contents = "x = 1
+y = 2 \
+    + 1
+z = x + 1";
 
-        let lxr: Vec<LexResult> = lexer::lex(
-            "
+        assert_eq!(noqa_mappings(contents), NoqaMapping::default());
+
+        let contents = "
 x = 1
 y = 2
-z = x + 1",
-            Mode::Module,
-        )
-        .collect();
-        assert_eq!(extract_noqa_line_for(&lxr), IntMap::default());
+z = x + 1";
+        assert_eq!(noqa_mappings(contents), NoqaMapping::default());
 
-        let lxr: Vec<LexResult> = lexer::lex(
-            "x = 1
+        let contents = "x = 1
 y = 2
 z = x + 1
-        ",
-            Mode::Module,
-        )
-        .collect();
-        assert_eq!(extract_noqa_line_for(&lxr), IntMap::default());
+        ";
+        assert_eq!(noqa_mappings(contents), NoqaMapping::default());
 
-        let lxr: Vec<LexResult> = lexer::lex(
-            "x = 1
+        let contents = "x = 1
 
 y = 2
 z = x + 1
-        ",
-            Mode::Module,
-        )
-        .collect();
-        assert_eq!(extract_noqa_line_for(&lxr), IntMap::default());
+        ";
+        assert_eq!(noqa_mappings(contents), NoqaMapping::default());
 
-        let lxr: Vec<LexResult> = lexer::lex(
-            "x = '''abc
+        let contents = "x = '''abc
 def
 ghi
 '''
 y = 2
-z = x + 1",
-            Mode::Module,
-        )
-        .collect();
+z = x + 1";
         assert_eq!(
-            extract_noqa_line_for(&lxr),
-            IntMap::from_iter([(1, 4), (2, 4), (3, 4)])
+            noqa_mappings(contents),
+            NoqaMapping::from_iter([TextRange::new(TextSize::from(4), TextSize::from(22)),])
         );
 
-        let lxr: Vec<LexResult> = lexer::lex(
-            "x = 1
+        let contents = "x = 1
 y = '''abc
 def
 ghi
 '''
-z = 2",
-            Mode::Module,
-        )
-        .collect();
+z = 2";
         assert_eq!(
-            extract_noqa_line_for(&lxr),
-            IntMap::from_iter([(2, 5), (3, 5), (4, 5)])
+            noqa_mappings(contents),
+            NoqaMapping::from_iter([TextRange::new(TextSize::from(10), TextSize::from(28))])
         );
 
-        let lxr: Vec<LexResult> = lexer::lex(
-            "x = 1
+        let contents = "x = 1
 y = '''abc
 def
 ghi
-'''",
-            Mode::Module,
-        )
-        .collect();
+'''";
         assert_eq!(
-            extract_noqa_line_for(&lxr),
-            IntMap::from_iter([(2, 5), (3, 5), (4, 5)])
+            noqa_mappings(contents),
+            NoqaMapping::from_iter([TextRange::new(TextSize::from(10), TextSize::from(28))])
         );
 
-        let lxr: Vec<LexResult> = lexer::lex(
-            r#"x = \
-    1"#,
-            Mode::Module,
-        )
-        .collect();
-        assert_eq!(extract_noqa_line_for(&lxr), IntMap::from_iter([(1, 2)]));
+        let contents = r#"x = \
+    1"#;
+        assert_eq!(
+            noqa_mappings(contents),
+            NoqaMapping::from_iter([TextRange::new(TextSize::from(0), TextSize::from(6))])
+        );
 
-        let lxr: Vec<LexResult> = lexer::lex(
-            r#"from foo import \
+        let contents = r#"from foo import \
     bar as baz, \
-    qux as quux"#,
-            Mode::Module,
-        )
-        .collect();
+    qux as quux"#;
         assert_eq!(
-            extract_noqa_line_for(&lxr),
-            IntMap::from_iter([(1, 3), (2, 3)])
+            noqa_mappings(contents),
+            NoqaMapping::from_iter([TextRange::new(TextSize::from(0), TextSize::from(36))])
         );
 
-        let lxr: Vec<LexResult> = lexer::lex(
-            r#"
+        let contents = r#"
 # Foo
 from foo import \
     bar as baz, \
@@ -276,13 +318,14 @@ from foo import \
 x = \
     1
 y = \
-    2"#,
-            Mode::Module,
-        )
-        .collect();
+    2"#;
         assert_eq!(
-            extract_noqa_line_for(&lxr),
-            IntMap::from_iter([(3, 5), (4, 5), (6, 7), (8, 9)])
+            noqa_mappings(contents),
+            NoqaMapping::from_iter([
+                TextRange::new(TextSize::from(7), TextSize::from(43)),
+                TextRange::new(TextSize::from(65), TextSize::from(71)),
+                TextRange::new(TextSize::from(77), TextSize::from(83)),
+            ])
         );
     }
 
@@ -292,7 +335,10 @@ y = \
 y = 2
 z = x + 1";
         let lxr: Vec<LexResult> = lexer::lex(contents, Mode::Module).collect();
-        assert_eq!(extract_isort_directives(&lxr).exclusions, IntSet::default());
+        assert_eq!(
+            extract_isort_directives(&lxr, &Locator::new(contents)).exclusions,
+            Vec::default()
+        );
 
         let contents = "# isort: off
 x = 1
@@ -301,8 +347,8 @@ y = 2
 z = x + 1";
         let lxr: Vec<LexResult> = lexer::lex(contents, Mode::Module).collect();
         assert_eq!(
-            extract_isort_directives(&lxr).exclusions,
-            IntSet::from_iter([2, 3, 4])
+            extract_isort_directives(&lxr, &Locator::new(contents)).exclusions,
+            Vec::from_iter([TextRange::new(TextSize::from(0), TextSize::from(25))])
         );
 
         let contents = "# isort: off
@@ -314,8 +360,8 @@ z = x + 1
 # isort: on";
         let lxr: Vec<LexResult> = lexer::lex(contents, Mode::Module).collect();
         assert_eq!(
-            extract_isort_directives(&lxr).exclusions,
-            IntSet::from_iter([2, 3, 4, 5])
+            extract_isort_directives(&lxr, &Locator::new(contents)).exclusions,
+            Vec::from_iter([TextRange::new(TextSize::from(0), TextSize::from(38))])
         );
 
         let contents = "# isort: off
@@ -324,8 +370,8 @@ y = 2
 z = x + 1";
         let lxr: Vec<LexResult> = lexer::lex(contents, Mode::Module).collect();
         assert_eq!(
-            extract_isort_directives(&lxr).exclusions,
-            IntSet::from_iter([2, 3, 4])
+            extract_isort_directives(&lxr, &Locator::new(contents)).exclusions,
+            Vec::from_iter([TextRange::at(TextSize::from(0), contents.text_len())])
         );
 
         let contents = "# isort: skip_file
@@ -333,7 +379,10 @@ x = 1
 y = 2
 z = x + 1";
         let lxr: Vec<LexResult> = lexer::lex(contents, Mode::Module).collect();
-        assert_eq!(extract_isort_directives(&lxr).exclusions, IntSet::default());
+        assert_eq!(
+            extract_isort_directives(&lxr, &Locator::new(contents)).exclusions,
+            Vec::default()
+        );
 
         let contents = "# isort: off
 x = 1
@@ -342,7 +391,10 @@ y = 2
 # isort: skip_file
 z = x + 1";
         let lxr: Vec<LexResult> = lexer::lex(contents, Mode::Module).collect();
-        assert_eq!(extract_isort_directives(&lxr).exclusions, IntSet::default());
+        assert_eq!(
+            extract_isort_directives(&lxr, &Locator::new(contents)).exclusions,
+            Vec::default()
+        );
     }
 
     #[test]
@@ -351,19 +403,28 @@ z = x + 1";
 y = 2
 z = x + 1";
         let lxr: Vec<LexResult> = lexer::lex(contents, Mode::Module).collect();
-        assert_eq!(extract_isort_directives(&lxr).splits, Vec::<usize>::new());
+        assert_eq!(
+            extract_isort_directives(&lxr, &Locator::new(contents)).splits,
+            Vec::new()
+        );
 
         let contents = "x = 1
 y = 2
 # isort: split
 z = x + 1";
         let lxr: Vec<LexResult> = lexer::lex(contents, Mode::Module).collect();
-        assert_eq!(extract_isort_directives(&lxr).splits, vec![3]);
+        assert_eq!(
+            extract_isort_directives(&lxr, &Locator::new(contents)).splits,
+            vec![TextSize::from(12)]
+        );
 
         let contents = "x = 1
 y = 2  # isort: split
 z = x + 1";
         let lxr: Vec<LexResult> = lexer::lex(contents, Mode::Module).collect();
-        assert_eq!(extract_isort_directives(&lxr).splits, vec![2]);
+        assert_eq!(
+            extract_isort_directives(&lxr, &Locator::new(contents)).splits,
+            vec![TextSize::from(13)]
+        );
     }
 }

crates/ruff/src/doc_lines.rs

@@ -1,8 +1,10 @@
 //! Doc line extraction. In this context, a doc line is a line consisting of a
 //! standalone comment or a constant string statement.
 
+use ruff_text_size::{TextRange, TextSize};
 use std::iter::FusedIterator;
 
+use ruff_python_ast::source_code::Locator;
 use rustpython_parser::ast::{Constant, ExprKind, Stmt, StmtKind, Suite};
 use rustpython_parser::lexer::LexResult;
 use rustpython_parser::Tok;
@@ -11,46 +13,56 @@ use ruff_python_ast::visitor;
 use ruff_python_ast::visitor::Visitor;
 
 /// Extract doc lines (standalone comments) from a token sequence.
-pub fn doc_lines_from_tokens(lxr: &[LexResult]) -> DocLines {
-    DocLines::new(lxr)
+pub fn doc_lines_from_tokens<'a>(lxr: &'a [LexResult], locator: &'a Locator<'a>) -> DocLines<'a> {
+    DocLines::new(lxr, locator)
 }
 
 pub struct DocLines<'a> {
     inner: std::iter::Flatten<core::slice::Iter<'a, LexResult>>,
-    prev: Option<usize>,
+    locator: &'a Locator<'a>,
+    prev: TextSize,
 }
 
 impl<'a> DocLines<'a> {
-    fn new(lxr: &'a [LexResult]) -> Self {
+    fn new(lxr: &'a [LexResult], locator: &'a Locator) -> Self {
         Self {
             inner: lxr.iter().flatten(),
-            prev: None,
+            locator,
+            prev: TextSize::default(),
         }
     }
 }
 
 impl Iterator for DocLines<'_> {
-    type Item = usize;
+    type Item = TextSize;
 
     fn next(&mut self) -> Option<Self::Item> {
+        let mut at_start_of_line = true;
         loop {
-            let (start, tok, end) = self.inner.next()?;
+            let (tok, range) = self.inner.next()?;
 
             match tok {
-                Tok::Indent | Tok::Dedent | Tok::Newline => continue,
                 Tok::Comment(..) => {
-                    if let Some(prev) = self.prev {
-                        if start.row() > prev {
-                            break Some(start.row());
-                        }
-                    } else {
-                        break Some(start.row());
+                    if at_start_of_line
+                        || self
+                            .locator
+                            .contains_line_break(TextRange::new(self.prev, range.start()))
+                    {
+                        break Some(range.start());
                     }
                 }
-                _ => {}
+                Tok::Newline => {
+                    at_start_of_line = true;
+                }
+                Tok::Indent | Tok::Dedent => {
+                    // ignore
+                }
+                _ => {
+                    at_start_of_line = false;
+                }
             }
 
-            self.prev = Some(end.row());
+            self.prev = range.end();
         }
     }
 }
@@ -59,7 +71,7 @@ impl FusedIterator for DocLines<'_> {}
 
 #[derive(Default)]
 struct StringLinesVisitor {
-    string_lines: Vec<usize>,
+    string_lines: Vec<TextSize>,
 }
 
 impl Visitor<'_> for StringLinesVisitor {
@@ -70,16 +82,15 @@ impl Visitor<'_> for StringLinesVisitor {
                 ..
             } = &value.node
             {
-                self.string_lines
-                    .extend(value.location.row()..=value.end_location.unwrap().row());
+                self.string_lines.push(value.start());
             }
         }
         visitor::walk_stmt(self, stmt);
     }
 }
 
-/// Extract doc lines (standalone strings) from an AST.
-pub fn doc_lines_from_ast(python_ast: &Suite) -> Vec<usize> {
+/// Extract doc lines (standalone strings) start positions from an AST.
+pub fn doc_lines_from_ast(python_ast: &Suite) -> Vec<TextSize> {
     let mut visitor = StringLinesVisitor::default();
     visitor.visit_body(python_ast);
     visitor.string_lines

crates/ruff/src/docstrings/definition.rs

@@ -1,4 +1,7 @@
+use ruff_text_size::{TextRange, TextSize};
 use rustpython_parser::ast::{Expr, Stmt};
+use std::fmt::{Debug, Formatter};
+use std::ops::Deref;
 
 use ruff_python_semantic::analyze::visibility::{
     class_visibility, function_visibility, method_visibility, Modifier, Visibility, VisibleScope,
@@ -25,11 +28,78 @@ pub struct Definition<'a> {
 pub struct Docstring<'a> {
     pub kind: DefinitionKind<'a>,
     pub expr: &'a Expr,
+    /// The content of the docstring, including the leading and trailing quotes.
     pub contents: &'a str,
-    pub body: &'a str,
+
+    /// The range of the docstring body (without the quotes). The range is relative to [`Self::contents`].
+    pub body_range: TextRange,
     pub indentation: &'a str,
 }
 
+impl<'a> Docstring<'a> {
+    pub fn body(&self) -> DocstringBody {
+        DocstringBody { docstring: self }
+    }
+
+    pub const fn start(&self) -> TextSize {
+        self.expr.start()
+    }
+
+    pub const fn end(&self) -> TextSize {
+        self.expr.end()
+    }
+
+    pub const fn range(&self) -> TextRange {
+        self.expr.range()
+    }
+
+    pub fn leading_quote(&self) -> &'a str {
+        &self.contents[TextRange::up_to(self.body_range.start())]
+    }
+}
+
+#[derive(Copy, Clone)]
+pub struct DocstringBody<'a> {
+    docstring: &'a Docstring<'a>,
+}
+
+impl<'a> DocstringBody<'a> {
+    #[inline]
+    pub fn start(self) -> TextSize {
+        self.range().start()
+    }
+
+    #[inline]
+    pub fn end(self) -> TextSize {
+        self.range().end()
+    }
+
+    pub fn range(self) -> TextRange {
+        self.docstring.body_range + self.docstring.start()
+    }
+
+    pub fn as_str(self) -> &'a str {
+        &self.docstring.contents[self.docstring.body_range]
+    }
+}
+
+impl Deref for DocstringBody<'_> {
+    type Target = str;
+
+    fn deref(&self) -> &Self::Target {
+        self.as_str()
+    }
+}
+
+impl Debug for DocstringBody<'_> {
+    fn fmt(&self, f: &mut Formatter<'_>) -> std::fmt::Result {
+        f.debug_struct("DocstringBody")
+            .field("text", &self.as_str())
+            .field("range", &self.range())
+            .finish()
+    }
+}
+
 #[derive(Copy, Clone)]
 pub enum Documentable {
     Class,

crates/ruff/src/docstrings/google.rs

@@ -26,6 +26,8 @@ pub(crate) static GOOGLE_SECTIONS: &[SectionKind] = &[
     SectionKind::KeywordArguments,
     SectionKind::Note,
     SectionKind::Notes,
+    SectionKind::OtherArgs,
+    SectionKind::OtherArguments,
     SectionKind::Return,
     SectionKind::Tip,
     SectionKind::Todo,

crates/ruff/src/docstrings/numpy.rs

@@ -14,6 +14,7 @@ pub(crate) static NUMPY_SECTIONS: &[SectionKind] = &[
     SectionKind::Yields,
     // NumPy-only
     SectionKind::ExtendedSummary,
+    SectionKind::OtherParams,
     SectionKind::OtherParameters,
     SectionKind::Parameters,
     SectionKind::ShortSummary,

crates/ruff/src/docstrings/sections.rs

@@ -1,5 +1,10 @@
+use ruff_python_ast::newlines::{StrExt, UniversalNewlineIterator};
+use ruff_text_size::{TextLen, TextRange, TextSize};
+use std::fmt::{Debug, Formatter};
+use std::iter::FusedIterator;
 use strum_macros::EnumIter;
 
+use crate::docstrings::definition::{Docstring, DocstringBody};
 use crate::docstrings::styles::SectionStyle;
 use ruff_python_ast::whitespace;
 
@@ -22,6 +27,9 @@ pub enum SectionKind {
     Methods,
     Note,
     Notes,
+    OtherArgs,
+    OtherArguments,
+    OtherParams,
     OtherParameters,
     Parameters,
     Raises,
@@ -59,6 +67,9 @@ impl SectionKind {
             "methods" => Some(Self::Methods),
             "note" => Some(Self::Note),
             "notes" => Some(Self::Notes),
+            "other args" => Some(Self::OtherArgs),
+            "other arguments" => Some(Self::OtherArguments),
+            "other params" => Some(Self::OtherParams),
             "other parameters" => Some(Self::OtherParameters),
             "parameters" => Some(Self::Parameters),
             "raises" => Some(Self::Raises),
@@ -97,6 +108,9 @@ impl SectionKind {
             Self::Methods => "Methods",
             Self::Note => "Note",
             Self::Notes => "Notes",
+            Self::OtherArgs => "Other Args",
+            Self::OtherArguments => "Other Arguments",
+            Self::OtherParams => "Other Params",
             Self::OtherParameters => "Other Parameters",
             Self::Parameters => "Parameters",
             Self::Raises => "Raises",
@@ -116,17 +130,259 @@ impl SectionKind {
     }
 }
 
+pub(crate) struct SectionContexts<'a> {
+    contexts: Vec<SectionContextData>,
+    docstring: &'a Docstring<'a>,
+}
+
+impl<'a> SectionContexts<'a> {
+    /// Extract all `SectionContext` values from a docstring.
+    pub fn from_docstring(docstring: &'a Docstring<'a>, style: SectionStyle) -> Self {
+        let contents = docstring.body();
+
+        let mut contexts = Vec::new();
+        let mut last: Option<SectionContextData> = None;
+        let mut previous_line = None;
+
+        for line in contents.universal_newlines() {
+            if previous_line.is_none() {
+                // skip the first line
+                previous_line = Some(line.as_str());
+                continue;
+            }
+
+            if let Some(section_kind) = suspected_as_section(&line, style) {
+                let indent = whitespace::leading_space(&line);
+                let section_name = whitespace::leading_words(&line);
+
+                let section_name_range = TextRange::at(indent.text_len(), section_name.text_len());
+
+                if is_docstring_section(
+                    &line,
+                    section_name_range,
+                    previous_line.unwrap_or_default(),
+                ) {
+                    if let Some(mut last) = last.take() {
+                        last.range = TextRange::new(last.range.start(), line.start());
+                        contexts.push(last);
+                    }
+
+                    last = Some(SectionContextData {
+                        kind: section_kind,
+                        name_range: section_name_range + line.start(),
+                        range: TextRange::empty(line.start()),
+                        summary_full_end: line.full_end(),
+                    });
+                }
+            }
+
+            previous_line = Some(line.as_str());
+        }
+
+        if let Some(mut last) = last.take() {
+            last.range = TextRange::new(last.range.start(), contents.text_len());
+            contexts.push(last);
+        }
+
+        Self {
+            contexts,
+            docstring,
+        }
+    }
+
+    pub fn len(&self) -> usize {
+        self.contexts.len()
+    }
+
+    pub fn iter(&self) -> SectionContextsIter {
+        SectionContextsIter {
+            docstring_body: self.docstring.body(),
+            inner: self.contexts.iter(),
+        }
+    }
+}
+
+impl<'a> IntoIterator for &'a SectionContexts<'a> {
+    type Item = SectionContext<'a>;
+    type IntoIter = SectionContextsIter<'a>;
+
+    fn into_iter(self) -> Self::IntoIter {
+        self.iter()
+    }
+}
+
+impl Debug for SectionContexts<'_> {
+    fn fmt(&self, f: &mut Formatter<'_>) -> std::fmt::Result {
+        f.debug_list().entries(self.iter()).finish()
+    }
+}
+
+pub struct SectionContextsIter<'a> {
+    docstring_body: DocstringBody<'a>,
+    inner: std::slice::Iter<'a, SectionContextData>,
+}
+
+impl<'a> Iterator for SectionContextsIter<'a> {
+    type Item = SectionContext<'a>;
+
+    fn next(&mut self) -> Option<Self::Item> {
+        let next = self.inner.next()?;
+
+        Some(SectionContext {
+            data: next,
+            docstring_body: self.docstring_body,
+        })
+    }
+
+    fn size_hint(&self) -> (usize, Option<usize>) {
+        self.inner.size_hint()
+    }
+}
+
+impl<'a> DoubleEndedIterator for SectionContextsIter<'a> {
+    fn next_back(&mut self) -> Option<Self::Item> {
+        let back = self.inner.next_back()?;
+        Some(SectionContext {
+            data: back,
+            docstring_body: self.docstring_body,
+        })
+    }
+}
+
+impl FusedIterator for SectionContextsIter<'_> {}
+impl ExactSizeIterator for SectionContextsIter<'_> {}
+
 #[derive(Debug)]
-pub(crate) struct SectionContext<'a> {
-    /// The "kind" of the section, e.g. "SectionKind::Args" or "SectionKind::Returns".
-    pub(crate) kind: SectionKind,
+struct SectionContextData {
+    kind: SectionKind,
+
+    /// Range of the section name, relative to the [`Docstring::body`]
+    name_range: TextRange,
+
+    /// Range from the start to the end of the section, relative to the [`Docstring::body`]
+    range: TextRange,
+
+    /// End of the summary, relative to the [`Docstring::body`]
+    summary_full_end: TextSize,
+}
+
+pub struct SectionContext<'a> {
+    data: &'a SectionContextData,
+    docstring_body: DocstringBody<'a>,
+}
+
+impl<'a> SectionContext<'a> {
+    pub fn is_last(&self) -> bool {
+        self.range().end() == self.docstring_body.end()
+    }
+
+    /// The `kind` of the section, e.g. [`SectionKind::Args`] or [`SectionKind::Returns`].
+    pub const fn kind(&self) -> SectionKind {
+        self.data.kind
+    }
+
     /// The name of the section as it appears in the docstring, e.g. "Args" or "Returns".
-    pub(crate) section_name: &'a str,
-    pub(crate) previous_line: &'a str,
-    pub(crate) line: &'a str,
-    pub(crate) following_lines: &'a [&'a str],
-    pub(crate) is_last_section: bool,
-    pub(crate) original_index: usize,
+    pub fn section_name(&self) -> &'a str {
+        &self.docstring_body.as_str()[self.data.name_range]
+    }
+
+    /// Returns the rest of the summary line after the section name.
+    pub fn summary_after_section_name(&self) -> &'a str {
+        &self.summary_line()[usize::from(self.data.name_range.end() - self.data.range.start())..]
+    }
+
+    fn offset(&self) -> TextSize {
+        self.docstring_body.start()
+    }
+
+    /// The absolute range of the section name
+    pub fn section_name_range(&self) -> TextRange {
+        self.data.name_range + self.offset()
+    }
+
+    /// Summary range relative to the start of the document. Includes the trailing newline.
+    pub fn summary_full_range(&self) -> TextRange {
+        self.summary_full_range_relative() + self.offset()
+    }
+
+    /// The absolute range of the summary line, excluding any trailing newline character.
+    pub fn summary_range(&self) -> TextRange {
+        TextRange::at(self.range().start(), self.summary_line().text_len())
+    }
+
+    /// Range of the summary line relative to [`Docstring::body`], including the trailing newline character.
+    fn summary_full_range_relative(&self) -> TextRange {
+        TextRange::new(self.range_relative().start(), self.data.summary_full_end)
+    }
+
+    /// Returns the range of this section relative to [`Docstring::body`]
+    const fn range_relative(&self) -> TextRange {
+        self.data.range
+    }
+
+    /// The absolute range of the full-section.
+    pub fn range(&self) -> TextRange {
+        self.range_relative() + self.offset()
+    }
+
+    /// Summary line without the trailing newline characters
+    pub fn summary_line(&self) -> &'a str {
+        let full_summary = &self.docstring_body.as_str()[self.summary_full_range_relative()];
+
+        let mut bytes = full_summary.bytes().rev();
+
+        let newline_width = match bytes.next() {
+            Some(b'\n') => {
+                if bytes.next() == Some(b'\r') {
+                    2
+                } else {
+                    1
+                }
+            }
+            Some(b'\r') => 1,
+            _ => 0,
+        };
+
+        &full_summary[..full_summary.len() - newline_width]
+    }
+
+    /// Returns the text of the last line of the previous section or an empty string if it is the first section.
+    pub fn previous_line(&self) -> Option<&'a str> {
+        let previous =
+            &self.docstring_body.as_str()[TextRange::up_to(self.range_relative().start())];
+        previous.universal_newlines().last().map(|l| l.as_str())
+    }
+
+    /// Returns the lines belonging to this section after the summary line.
+    pub fn following_lines(&self) -> UniversalNewlineIterator<'a> {
+        let lines = self.following_lines_str();
+        UniversalNewlineIterator::with_offset(lines, self.offset() + self.data.summary_full_end)
+    }
+
+    fn following_lines_str(&self) -> &'a str {
+        &self.docstring_body.as_str()[self.following_range_relative()]
+    }
+
+    /// Returns the range to the following lines relative to [`Docstring::body`].
+    const fn following_range_relative(&self) -> TextRange {
+        TextRange::new(self.data.summary_full_end, self.range_relative().end())
+    }
+
+    /// Returns the absolute range of the following lines.
+    pub fn following_range(&self) -> TextRange {
+        self.following_range_relative() + self.offset()
+    }
+}
+
+impl Debug for SectionContext<'_> {
+    fn fmt(&self, f: &mut Formatter<'_>) -> std::fmt::Result {
+        f.debug_struct("SectionContext")
+            .field("kind", &self.kind())
+            .field("section_name", &self.section_name())
+            .field("summary_line", &self.summary_line())
+            .field("following_lines", &&self.following_lines_str())
+            .finish()
+    }
 }
 
 fn suspected_as_section(line: &str, style: SectionStyle) -> Option<SectionKind> {
@@ -139,20 +395,15 @@ fn suspected_as_section(line: &str, style: SectionStyle) -> Option<SectionKind>
 }
 
 /// Check if the suspected context is really a section header.
-fn is_docstring_section(context: &SectionContext) -> bool {
-    let section_name_suffix = context
-        .line
-        .trim()
-        .strip_prefix(context.section_name)
-        .unwrap()
-        .trim();
+fn is_docstring_section(line: &str, section_name_range: TextRange, previous_lines: &str) -> bool {
+    let section_name_suffix = line[usize::from(section_name_range.end())..].trim();
     let this_looks_like_a_section_name =
         section_name_suffix == ":" || section_name_suffix.is_empty();
     if !this_looks_like_a_section_name {
         return false;
     }
 
-    let prev_line = context.previous_line.trim();
+    let prev_line = previous_lines.trim();
     let prev_line_ends_with_punctuation = [',', ';', '.', '-', '\\', '/', ']', '}', ')']
         .into_iter()
         .any(|char| prev_line.ends_with(char));
@@ -164,50 +415,3 @@ fn is_docstring_section(context: &SectionContext) -> bool {
 
     true
 }
-
-/// Extract all `SectionContext` values from a docstring.
-pub(crate) fn section_contexts<'a>(
-    lines: &'a [&'a str],
-    style: SectionStyle,
-) -> Vec<SectionContext<'a>> {
-    let mut contexts = vec![];
-    for (kind, lineno) in lines
-        .iter()
-        .enumerate()
-        .skip(1)
-        .filter_map(|(lineno, line)| suspected_as_section(line, style).map(|kind| (kind, lineno)))
-    {
-        let context = SectionContext {
-            kind,
-            section_name: whitespace::leading_words(lines[lineno]),
-            previous_line: lines[lineno - 1],
-            line: lines[lineno],
-            following_lines: &lines[lineno + 1..],
-            original_index: lineno,
-            is_last_section: false,
-        };
-        if is_docstring_section(&context) {
-            contexts.push(context);
-        }
-    }
-
-    let mut truncated_contexts = Vec::with_capacity(contexts.len());
-    let mut end: Option<usize> = None;
-    for context in contexts.into_iter().rev() {
-        let next_end = context.original_index;
-        truncated_contexts.push(SectionContext {
-            kind: context.kind,
-            section_name: context.section_name,
-            previous_line: context.previous_line,
-            line: context.line,
-            following_lines: end.map_or(context.following_lines, |end| {
-                &lines[context.original_index + 1..end]
-            }),
-            original_index: context.original_index,
-            is_last_section: end.is_none(),
-        });
-        end = Some(next_end);
-    }
-    truncated_contexts.reverse();
-    truncated_contexts
-}

crates/ruff/src/fs.rs

@@ -1,36 +1,21 @@
 use std::path::{Path, PathBuf};
 
-use anyhow::{anyhow, Result};
 use globset::GlobMatcher;
 use log::debug;
 use path_absolutize::{path_dedot, Absolutize};
 
 use crate::registry::RuleSet;
 
-/// Extract the absolute path and basename (as strings) from a Path.
-pub fn extract_path_names(path: &Path) -> Result<(&str, &str)> {
-    let file_path = path
-        .to_str()
-        .ok_or_else(|| anyhow!("Unable to parse filename: {:?}", path))?;
-    let file_basename = path
-        .file_name()
-        .ok_or_else(|| anyhow!("Unable to parse filename: {:?}", path))?
-        .to_str()
-        .ok_or_else(|| anyhow!("Unable to parse filename: {:?}", path))?;
-    Ok((file_path, file_basename))
-}
-
 /// Create a set with codes matching the pattern/code pairs.
 pub(crate) fn ignores_from_path(
     path: &Path,
     pattern_code_pairs: &[(GlobMatcher, GlobMatcher, RuleSet)],
 ) -> RuleSet {
-    let (file_path, file_basename) = extract_path_names(path).expect("Unable to parse filename");
-
+    let file_name = path.file_name().expect("Unable to parse filename");
     pattern_code_pairs
         .iter()
         .filter_map(|(absolute, basename, rules)| {
-            if basename.is_match(file_basename) {
+            if basename.is_match(file_name) {
                 debug!(
                     "Adding per-file ignores for {:?} due to basename match on {:?}: {:?}",
                     path,
@@ -38,7 +23,7 @@ pub(crate) fn ignores_from_path(
                     rules
                 );
                 Some(rules)
-            } else if absolute.is_match(file_path) {
+            } else if absolute.is_match(path) {
                 debug!(
                     "Adding per-file ignores for {:?} due to absolute match on {:?}: {:?}",
                     path,

crates/ruff/src/importer.rs

@@ -2,8 +2,9 @@
 
 use anyhow::Result;
 use libcst_native::{Codegen, CodegenState, ImportAlias, Name, NameOrAttribute};
+use ruff_text_size::TextSize;
 use rustc_hash::FxHashMap;
-use rustpython_parser::ast::{Location, Stmt, StmtKind, Suite};
+use rustpython_parser::ast::{Stmt, StmtKind, Suite};
 use rustpython_parser::{lexer, Mode, Tok};
 
 use ruff_diagnostics::Edit;
@@ -95,7 +96,7 @@ impl<'a> Importer<'a> {
 
     /// Add the given member to an existing `StmtKind::ImportFrom` statement.
     pub fn add_member(&self, stmt: &Stmt, member: &str) -> Result<Edit> {
-        let mut tree = match_module(self.locator.slice(stmt))?;
+        let mut tree = match_module(self.locator.slice(stmt.range()))?;
         let import_from = match_import_from(&mut tree)?;
         let aliases = match_aliases(import_from)?;
         aliases.push(ImportAlias {
@@ -113,11 +114,7 @@ impl<'a> Importer<'a> {
             ..CodegenState::default()
         };
         tree.codegen(&mut state);
-        Ok(Edit::replacement(
-            state.to_string(),
-            stmt.location,
-            stmt.end_location.unwrap(),
-        ))
+        Ok(Edit::range_replacement(state.to_string(), stmt.range()))
     }
 }
 
@@ -126,13 +123,13 @@ struct Insertion {
     /// The content to add before the insertion.
     prefix: &'static str,
     /// The location at which to insert.
-    location: Location,
+    location: TextSize,
     /// The content to add after the insertion.
     suffix: &'static str,
 }
 
 impl Insertion {
-    fn new(prefix: &'static str, location: Location, suffix: &'static str) -> Self {
+    fn new(prefix: &'static str, location: TextSize, suffix: &'static str) -> Self {
         Self {
             prefix,
             location,
@@ -142,7 +139,7 @@ impl Insertion {
 }
 
 /// Find the end of the last docstring.
-fn match_docstring_end(body: &[Stmt]) -> Option<Location> {
+fn match_docstring_end(body: &[Stmt]) -> Option<TextSize> {
     let mut iter = body.iter();
     let Some(mut stmt) = iter.next() else {
         return None;
@@ -156,7 +153,7 @@ fn match_docstring_end(body: &[Stmt]) -> Option<Location> {
         }
         stmt = next;
     }
-    Some(stmt.end_location.unwrap())
+    Some(stmt.end())
 }
 
 /// Find the location at which a "top-of-file" import should be inserted,
@@ -173,17 +170,17 @@ fn match_docstring_end(body: &[Stmt]) -> Option<Location> {
 /// The location returned will be the start of the `import os` statement,
 /// along with a trailing newline suffix.
 fn end_of_statement_insertion(stmt: &Stmt, locator: &Locator, stylist: &Stylist) -> Insertion {
-    let location = stmt.end_location.unwrap();
-    let mut tokens = lexer::lex_located(locator.skip(location), Mode::Module, location).flatten();
-    if let Some((.., Tok::Semi, end)) = tokens.next() {
+    let location = stmt.end();
+    let mut tokens = lexer::lex_located(locator.after(location), Mode::Module, location).flatten();
+    if let Some((Tok::Semi, range)) = tokens.next() {
         // If the first token after the docstring is a semicolon, insert after the semicolon as an
         // inline statement;
-        Insertion::new(" ", end, ";")
+        Insertion::new(" ", range.end(), ";")
     } else {
         // Otherwise, insert on the next line.
         Insertion::new(
             "",
-            Location::new(location.row() + 1, 0),
+            locator.line_end(location),
             stylist.line_ending().as_str(),
         )
     }
@@ -207,25 +204,25 @@ fn top_of_file_insertion(body: &[Stmt], locator: &Locator, stylist: &Stylist) ->
     let mut location = if let Some(location) = match_docstring_end(body) {
         // If the first token after the docstring is a semicolon, insert after the semicolon as an
         // inline statement;
-        let first_token = lexer::lex_located(locator.skip(location), Mode::Module, location)
+        let first_token = lexer::lex_located(locator.after(location), Mode::Module, location)
             .flatten()
             .next();
-        if let Some((.., Tok::Semi, end)) = first_token {
-            return Insertion::new(" ", end, ";");
+        if let Some((Tok::Semi, range)) = first_token {
+            return Insertion::new(" ", range.end(), ";");
         }
 
         // Otherwise, advance to the next row.
-        Location::new(location.row() + 1, 0)
+        locator.full_line_end(location)
     } else {
-        Location::default()
+        TextSize::default()
     };
 
     // Skip over any comments and empty lines.
-    for (.., tok, end) in
-        lexer::lex_located(locator.skip(location), Mode::Module, location).flatten()
+    for (tok, range) in
+        lexer::lex_located(locator.after(location), Mode::Module, location).flatten()
     {
         if matches!(tok, Tok::Comment(..) | Tok::Newline) {
-            location = Location::new(end.row() + 1, 0);
+            location = locator.full_line_end(range.end());
         } else {
             break;
         }
@@ -237,11 +234,12 @@ fn top_of_file_insertion(body: &[Stmt], locator: &Locator, stylist: &Stylist) ->
 #[cfg(test)]
 mod tests {
     use anyhow::Result;
+    use ruff_python_ast::newlines::LineEnding;
+    use ruff_text_size::TextSize;
     use rustpython_parser as parser;
-    use rustpython_parser::ast::Location;
     use rustpython_parser::lexer::LexResult;
 
-    use ruff_python_ast::source_code::{LineEnding, Locator, Stylist};
+    use ruff_python_ast::source_code::{Locator, Stylist};
 
     use crate::importer::{top_of_file_insertion, Insertion};
 
@@ -258,7 +256,7 @@ mod tests {
         let contents = "";
         assert_eq!(
             insert(contents)?,
-            Insertion::new("", Location::new(1, 0), LineEnding::default().as_str())
+            Insertion::new("", TextSize::from(0), LineEnding::default().as_str())
         );
 
         let contents = r#"
@@ -266,7 +264,7 @@ mod tests {
             .trim_start();
         assert_eq!(
             insert(contents)?,
-            Insertion::new("", Location::new(2, 0), LineEnding::default().as_str())
+            Insertion::new("", TextSize::from(19), LineEnding::default().as_str())
         );
 
         let contents = r#"
@@ -275,7 +273,7 @@ mod tests {
         .trim_start();
         assert_eq!(
             insert(contents)?,
-            Insertion::new("", Location::new(2, 0), "\n")
+            Insertion::new("", TextSize::from(20), "\n")
         );
 
         let contents = r#"
@@ -285,7 +283,7 @@ mod tests {
         .trim_start();
         assert_eq!(
             insert(contents)?,
-            Insertion::new("", Location::new(3, 0), "\n")
+            Insertion::new("", TextSize::from(40), "\n")
         );
 
         let contents = r#"
@@ -294,7 +292,7 @@ x = 1
         .trim_start();
         assert_eq!(
             insert(contents)?,
-            Insertion::new("", Location::new(1, 0), "\n")
+            Insertion::new("", TextSize::from(0), "\n")
         );
 
         let contents = r#"
@@ -303,7 +301,7 @@ x = 1
         .trim_start();
         assert_eq!(
             insert(contents)?,
-            Insertion::new("", Location::new(2, 0), "\n")
+            Insertion::new("", TextSize::from(23), "\n")
         );
 
         let contents = r#"
@@ -313,7 +311,7 @@ x = 1
         .trim_start();
         assert_eq!(
             insert(contents)?,
-            Insertion::new("", Location::new(3, 0), "\n")
+            Insertion::new("", TextSize::from(43), "\n")
         );
 
         let contents = r#"
@@ -323,7 +321,7 @@ x = 1
         .trim_start();
         assert_eq!(
             insert(contents)?,
-            Insertion::new("", Location::new(3, 0), "\n")
+            Insertion::new("", TextSize::from(43), "\n")
         );
 
         let contents = r#"
@@ -332,7 +330,7 @@ x = 1
         .trim_start();
         assert_eq!(
             insert(contents)?,
-            Insertion::new("", Location::new(1, 0), "\n")
+            Insertion::new("", TextSize::from(0), "\n")
         );
 
         let contents = r#"
@@ -341,7 +339,7 @@ x = 1
         .trim_start();
         assert_eq!(
             insert(contents)?,
-            Insertion::new(" ", Location::new(1, 20), ";")
+            Insertion::new(" ", TextSize::from(20), ";")
         );
 
         let contents = r#"
@@ -351,7 +349,7 @@ x = 1
         .trim_start();
         assert_eq!(
             insert(contents)?,
-            Insertion::new(" ", Location::new(1, 20), ";")
+            Insertion::new(" ", TextSize::from(20), ";")
         );
 
         Ok(())