[red knot] support EllipsisType as a singleton

.cargo/config.toml

@@ -8,7 +8,3 @@ benchmark = "bench -p ruff_benchmark --bench linter --bench formatter --"
 # See: https://github.com/astral-sh/ruff/issues/11503
 [target.'cfg(all(target_env="msvc", target_os = "windows"))']
 rustflags = ["-C", "target-feature=+crt-static"]
-
-[target.'wasm32-unknown-unknown']
-# See https://docs.rs/getrandom/latest/getrandom/#webassembly-support
-rustflags = ["--cfg", 'getrandom_backend="wasm_js"']

.config/nextest.toml

@@ -6,10 +6,3 @@ failure-output = "immediate-final"
 fail-fast = false
 
 status-level = "skip"
-
-# Mark tests that take longer than 1s as slow.
-# Terminate after 60s as a stop-gap measure to terminate on deadlock.
-slow-timeout =  { period = "1s", terminate-after = 60 }
-
-# Show slow jobs in the final summary
-final-status-level = "slow"

.editorconfig

@@ -17,7 +17,4 @@ indent_size = 4
 trim_trailing_whitespace = false
 
 [*.md]
-max_line_length = 100
-
-[*.toml]
-indent_size = 4
+max_line_length = 100

.gitattributes

@@ -12,16 +12,7 @@ crates/ruff_python_parser/resources/invalid/re_lexing/line_continuation_windows_
 crates/ruff_python_parser/resources/invalid/re_lex_logical_token_windows_eol.py text eol=crlf
 crates/ruff_python_parser/resources/invalid/re_lex_logical_token_mac_eol.py text eol=cr
 
-crates/ruff_linter/resources/test/fixtures/ruff/RUF046_CR.py text eol=cr
-crates/ruff_linter/resources/test/fixtures/ruff/RUF046_LF.py text eol=lf
-
-crates/ruff_linter/resources/test/fixtures/pyupgrade/UP018_CR.py text eol=cr
-crates/ruff_linter/resources/test/fixtures/pyupgrade/UP018_LF.py text eol=lf
-
 crates/ruff_python_parser/resources/inline linguist-generated=true
 
-ruff.schema.json -diff linguist-generated=true text=auto eol=lf
-ty.schema.json -diff linguist-generated=true text=auto eol=lf
-crates/ruff_python_ast/src/generated.rs -diff linguist-generated=true text=auto eol=lf
-crates/ruff_python_formatter/src/generated.rs -diff linguist-generated=true text=auto eol=lf
+ruff.schema.json linguist-generated=true text=auto eol=lf
 *.md.snap linguist-language=Markdown

.github/CODEOWNERS

@@ -9,16 +9,13 @@
 /crates/ruff_formatter/ @MichaReiser
 /crates/ruff_python_formatter/ @MichaReiser
 /crates/ruff_python_parser/ @MichaReiser @dhruvmanila
-/crates/ruff_annotate_snippets/ @BurntSushi
 
 # flake8-pyi
 /crates/ruff_linter/src/rules/flake8_pyi/ @AlexWaygood
 
-# Script for fuzzing the parser/ty etc.
-/python/py-fuzzer/ @AlexWaygood
+# Script for fuzzing the parser
+/scripts/fuzz-parser/ @AlexWaygood
 
-# ty
-/crates/ty* @carljm @MichaReiser @AlexWaygood @sharkdp @dcreager
-/crates/ruff_db/ @carljm @MichaReiser @AlexWaygood @sharkdp @dcreager
-/scripts/ty_benchmark/ @carljm @MichaReiser @AlexWaygood @sharkdp @dcreager
-/crates/ty_python_semantic @carljm @AlexWaygood @sharkdp @dcreager
+# red-knot
+/crates/red_knot* @carljm @MichaReiser @AlexWaygood
+/crates/ruff_db/ @carljm @MichaReiser @AlexWaygood

.github/ISSUE_TEMPLATE.md

@@ -0,0 +1,12 @@
+<!--
+Thank you for taking the time to report an issue! We're glad to have you involved with Ruff.
+
+If you're filing a bug report, please consider including the following information:
+
+* List of keywords you searched for before creating this issue. Write them down here so that others can find this issue more easily and help provide feedback.
+  e.g. "RUF001", "unused variable", "Jupyter notebook"
+* A minimal code snippet that reproduces the bug.
+* The command you invoked (e.g., `ruff /path/to/file.py --fix`), ideally including the `--isolated` flag.
+* The current Ruff settings (any relevant sections from your `pyproject.toml`).
+* The current Ruff version (`ruff --version`).
+-->

.github/ISSUE_TEMPLATE/1_bug_report.yaml

@@ -1,31 +0,0 @@
-name: Bug report
-description: Report an error or unexpected behavior
-body:
-  - type: markdown
-    attributes:
-      value: |
-        Thank you for taking the time to report an issue! We're glad to have you involved with Ruff.
-
-        **Before reporting, please make sure to search through [existing issues](https://github.com/astral-sh/ruff/issues?q=is:issue+is:open+label:bug) (including [closed](https://github.com/astral-sh/ruff/issues?q=is:issue%20state:closed%20label:bug)).**
-
-  - type: textarea
-    attributes:
-      label: Summary
-      description: |
-        A clear and concise description of the bug, including a minimal reproducible example.
-
-        Be sure to include the command you invoked (e.g., `ruff check /path/to/file.py --fix`), ideally including the `--isolated` flag and
-        the current Ruff settings (e.g., relevant sections from your `pyproject.toml`).
-
-        If possible, try to include the [playground](https://play.ruff.rs) link that reproduces this issue.
-
-    validations:
-      required: true
-
-  - type: input
-    attributes:
-      label: Version
-      description: What version of ruff are you using? (see `ruff version`)
-      placeholder: e.g., ruff 0.9.3 (90589372d 2025-01-23)
-    validations:
-      required: false

.github/ISSUE_TEMPLATE/2_rule_request.yaml

@@ -1,10 +0,0 @@
-name: Rule request
-description: Anything related to lint rules (proposing new rules, changes to existing rules, auto-fixes, etc.)
-body:
-  - type: textarea
-    attributes:
-      label: Summary
-      description: |
-        A clear and concise description of the relevant request. If applicable, please describe the current behavior as well.
-    validations:
-      required: true

.github/ISSUE_TEMPLATE/3_question.yaml

@@ -1,18 +0,0 @@
-name: Question
-description: Ask a question about Ruff
-labels: ["question"]
-body:
-  - type: textarea
-    attributes:
-      label: Question
-      description: Describe your question in detail.
-    validations:
-      required: true
-
-  - type: input
-    attributes:
-      label: Version
-      description: What version of ruff are you using? (see `ruff version`)
-      placeholder: e.g., ruff 0.9.3 (90589372d 2025-01-23)
-    validations:
-      required: false

.github/ISSUE_TEMPLATE/config.yml

@@ -1,11 +0,0 @@
-blank_issues_enabled: true
-contact_links:
-  - name: Report an issue with ty
-    url: https://github.com/astral-sh/ty/issues/new/choose
-    about: Please report issues for our type checker ty in the ty repository.
-  - name: Documentation
-    url: https://docs.astral.sh/ruff
-    about: Please consult the documentation before creating an issue.
-  - name: Community
-    url: https://discord.com/invite/astral-sh
-    about: Join our Discord community to ask questions and collaborate.

.github/PULL_REQUEST_TEMPLATE.md

@@ -1,9 +1,8 @@
 <!--
-Thank you for contributing to Ruff/ty! To help us out with reviewing, please consider the following:
+Thank you for contributing to Ruff! To help us out with reviewing, please consider the following:
 
 - Does this pull request include a summary of the change? (See below.)
-- Does this pull request include a descriptive title? (Please prefix with `[ty]` for ty pull
-  requests.)
+- Does this pull request include a descriptive title?
 - Does this pull request include references to any relevant issues?
 -->
 

.github/actionlint.yaml

@@ -1,11 +0,0 @@
-# Configuration for the actionlint tool, which we run via pre-commit
-# to verify the correctness of the syntax in our GitHub Actions workflows.
-
-self-hosted-runner:
-  # Various runners we use that aren't recognized out-of-the-box by actionlint:
-  labels:
-    - depot-ubuntu-latest-8
-    - depot-ubuntu-22.04-16
-    - depot-ubuntu-22.04-32
-    - github-windows-2025-x86_64-8
-    - github-windows-2025-x86_64-16

.github/mypy-primer-ty.toml

@@ -1,8 +0,0 @@
-#:schema ../ty.schema.json
-# Configuration overrides for the mypy primer run
-
-# Enable off-by-default rules.
-[rules]
-possibly-unresolved-reference = "warn"
-unused-ignore-comment = "warn"
-division-by-zero = "warn"

.github/renovate.json5

@@ -40,23 +40,12 @@
     enabled: true,
   },
   packageRules: [
-    // Pin GitHub Actions to immutable SHAs.
-    {
-      matchDepTypes: ["action"],
-      pinDigests: true,
-    },
-    // Annotate GitHub Actions SHAs with a SemVer version.
-    {
-      extends: ["helpers:pinGitHubActionDigests"],
-      extractVersion: "^(?<version>v?\\d+\\.\\d+\\.\\d+)$",
-      versioning: "regex:^v?(?<major>\\d+)(\\.(?<minor>\\d+)\\.(?<patch>\\d+))?$",
-    },
     {
       // Group upload/download artifact updates, the versions are dependent
       groupName: "Artifact GitHub Actions dependencies",
       matchManagers: ["github-actions"],
       matchDatasources: ["gitea-tags", "github-tags"],
-      matchPackageNames: ["actions/.*-artifact"],
+      matchPackagePatterns: ["actions/.*-artifact"],
       description: "Weekly update of artifact-related GitHub Actions dependencies",
     },
     {
@@ -72,7 +61,7 @@
     {
       // Disable updates of `zip-rs`; intentionally pinned for now due to ownership change
       // See: https://github.com/astral-sh/uv/issues/3642
-      matchPackageNames: ["zip"],
+      matchPackagePatterns: ["zip"],
       matchManagers: ["cargo"],
       enabled: false,
     },
@@ -81,7 +70,7 @@
       // with `mkdocs-material-insider`.
       // See: https://squidfunk.github.io/mkdocs-material/insiders/upgrade/
       matchManagers: ["pip_requirements"],
-      matchPackageNames: ["mkdocs-material"],
+      matchPackagePatterns: ["mkdocs-material"],
       enabled: false,
     },
     {
@@ -98,15 +87,22 @@
     {
       groupName: "Monaco",
       matchManagers: ["npm"],
-      matchPackageNames: ["monaco"],
+      matchPackagePatterns: ["monaco"],
       description: "Weekly update of the Monaco editor",
     },
     {
       groupName: "strum",
       matchManagers: ["cargo"],
-      matchPackageNames: ["strum"],
+      matchPackagePatterns: ["strum"],
       description: "Weekly update of strum dependencies",
-    }
+    },
+    {
+      groupName: "ESLint",
+      matchManagers: ["npm"],
+      matchPackageNames: ["eslint"],
+      allowedVersions: "<9",
+      description: "Constraint ESLint to version 8 until TypeScript-eslint supports ESLint 9", // https://github.com/typescript-eslint/typescript-eslint/issues/8211
+    },
   ],
   vulnerabilityAlerts: {
     commitMessageSuffix: "",

.github/workflows/build-binaries.yml

@@ -23,12 +23,10 @@ concurrency:
   group: ${{ github.workflow }}-${{ github.ref }}
   cancel-in-progress: true
 
-permissions: {}
-
 env:
   PACKAGE_NAME: ruff
   MODULE_NAME: ruff
-  PYTHON_VERSION: "3.13"
+  PYTHON_VERSION: "3.11"
   CARGO_INCREMENTAL: 0
   CARGO_NET_RETRY: 10
   CARGO_TERM_COLOR: always
@@ -39,52 +37,50 @@ jobs:
     if: ${{ !contains(github.event.pull_request.labels.*.name, 'no-build') }}
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
+      - uses: actions/checkout@v4
         with:
           submodules: recursive
-          persist-credentials: false
-      - uses: actions/setup-python@a26af69be951a213d495a4c3e4e4022e16d87065 # v5.6.0
+      - uses: actions/setup-python@v5
         with:
           python-version: ${{ env.PYTHON_VERSION }}
       - name: "Prep README.md"
         run: python scripts/transform_readme.py --target pypi
       - name: "Build sdist"
-        uses: PyO3/maturin-action@aef21716ff3dcae8a1c301d23ec3e4446972a6e3 # v1.49.1
+        uses: PyO3/maturin-action@v1
         with:
           command: sdist
           args: --out dist
       - name: "Test sdist"
         run: |
-          pip install dist/"${PACKAGE_NAME}"-*.tar.gz --force-reinstall
-          "${MODULE_NAME}" --help
-          python -m "${MODULE_NAME}" --help
+          pip install dist/${{ env.PACKAGE_NAME }}-*.tar.gz --force-reinstall
+          ${{ env.MODULE_NAME }} --help
+          python -m ${{ env.MODULE_NAME }} --help
       - name: "Upload sdist"
-        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
+        uses: actions/upload-artifact@v4
         with:
           name: wheels-sdist
           path: dist
 
   macos-x86_64:
     if: ${{ !contains(github.event.pull_request.labels.*.name, 'no-build') }}
-    runs-on: macos-14
+    runs-on: macos-12
     steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
+      - uses: actions/checkout@v4
         with:
           submodules: recursive
-          persist-credentials: false
-      - uses: actions/setup-python@a26af69be951a213d495a4c3e4e4022e16d87065 # v5.6.0
+      - uses: actions/setup-python@v5
         with:
           python-version: ${{ env.PYTHON_VERSION }}
           architecture: x64
       - name: "Prep README.md"
         run: python scripts/transform_readme.py --target pypi
       - name: "Build wheels - x86_64"
-        uses: PyO3/maturin-action@aef21716ff3dcae8a1c301d23ec3e4446972a6e3 # v1.49.1
+        uses: PyO3/maturin-action@v1
         with:
           target: x86_64
           args: --release --locked --out dist
       - name: "Upload wheels"
-        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
+        uses: actions/upload-artifact@v4
         with:
           name: wheels-macos-x86_64
           path: dist
@@ -99,7 +95,7 @@ jobs:
           tar czvf $ARCHIVE_FILE $ARCHIVE_NAME
           shasum -a 256 $ARCHIVE_FILE > $ARCHIVE_FILE.sha256
       - name: "Upload binary"
-        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
+        uses: actions/upload-artifact@v4
         with:
           name: artifacts-macos-x86_64
           path: |
@@ -110,28 +106,27 @@ jobs:
     if: ${{ !contains(github.event.pull_request.labels.*.name, 'no-build') }}
     runs-on: macos-14
     steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
+      - uses: actions/checkout@v4
         with:
           submodules: recursive
-          persist-credentials: false
-      - uses: actions/setup-python@a26af69be951a213d495a4c3e4e4022e16d87065 # v5.6.0
+      - uses: actions/setup-python@v5
         with:
           python-version: ${{ env.PYTHON_VERSION }}
           architecture: arm64
       - name: "Prep README.md"
         run: python scripts/transform_readme.py --target pypi
       - name: "Build wheels - aarch64"
-        uses: PyO3/maturin-action@aef21716ff3dcae8a1c301d23ec3e4446972a6e3 # v1.49.1
+        uses: PyO3/maturin-action@v1
         with:
           target: aarch64
           args: --release --locked --out dist
       - name: "Test wheel - aarch64"
         run: |
-          pip install dist/"${PACKAGE_NAME}"-*.whl --force-reinstall
+          pip install dist/${{ env.PACKAGE_NAME }}-*.whl --force-reinstall
           ruff --help
           python -m ruff --help
       - name: "Upload wheels"
-        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
+        uses: actions/upload-artifact@v4
         with:
           name: wheels-aarch64-apple-darwin
           path: dist
@@ -146,7 +141,7 @@ jobs:
           tar czvf $ARCHIVE_FILE $ARCHIVE_NAME
           shasum -a 256 $ARCHIVE_FILE > $ARCHIVE_FILE.sha256
       - name: "Upload binary"
-        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
+        uses: actions/upload-artifact@v4
         with:
           name: artifacts-aarch64-apple-darwin
           path: |
@@ -166,18 +161,17 @@ jobs:
           - target: aarch64-pc-windows-msvc
             arch: x64
     steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
+      - uses: actions/checkout@v4
         with:
           submodules: recursive
-          persist-credentials: false
-      - uses: actions/setup-python@a26af69be951a213d495a4c3e4e4022e16d87065 # v5.6.0
+      - uses: actions/setup-python@v5
         with:
           python-version: ${{ env.PYTHON_VERSION }}
           architecture: ${{ matrix.platform.arch }}
       - name: "Prep README.md"
         run: python scripts/transform_readme.py --target pypi
       - name: "Build wheels"
-        uses: PyO3/maturin-action@aef21716ff3dcae8a1c301d23ec3e4446972a6e3 # v1.49.1
+        uses: PyO3/maturin-action@v1
         with:
           target: ${{ matrix.platform.target }}
           args: --release --locked --out dist
@@ -188,11 +182,11 @@ jobs:
         if: ${{ !startsWith(matrix.platform.target, 'aarch64') }}
         shell: bash
         run: |
-          python -m pip install dist/"${PACKAGE_NAME}"-*.whl --force-reinstall
-          "${MODULE_NAME}" --help
-          python -m "${MODULE_NAME}" --help
+          python -m pip install dist/${{ env.PACKAGE_NAME }}-*.whl --force-reinstall
+          ${{ env.MODULE_NAME }} --help
+          python -m ${{ env.MODULE_NAME }} --help
       - name: "Upload wheels"
-        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
+        uses: actions/upload-artifact@v4
         with:
           name: wheels-${{ matrix.platform.target }}
           path: dist
@@ -203,7 +197,7 @@ jobs:
           7z a $ARCHIVE_FILE ./target/${{ matrix.platform.target }}/release/ruff.exe
           sha256sum $ARCHIVE_FILE > $ARCHIVE_FILE.sha256
       - name: "Upload binary"
-        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
+        uses: actions/upload-artifact@v4
         with:
           name: artifacts-${{ matrix.platform.target }}
           path: |
@@ -219,18 +213,17 @@ jobs:
           - x86_64-unknown-linux-gnu
           - i686-unknown-linux-gnu
     steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
+      - uses: actions/checkout@v4
         with:
           submodules: recursive
-          persist-credentials: false
-      - uses: actions/setup-python@a26af69be951a213d495a4c3e4e4022e16d87065 # v5.6.0
+      - uses: actions/setup-python@v5
         with:
           python-version: ${{ env.PYTHON_VERSION }}
           architecture: x64
       - name: "Prep README.md"
         run: python scripts/transform_readme.py --target pypi
       - name: "Build wheels"
-        uses: PyO3/maturin-action@aef21716ff3dcae8a1c301d23ec3e4446972a6e3 # v1.49.1
+        uses: PyO3/maturin-action@v1
         with:
           target: ${{ matrix.target }}
           manylinux: auto
@@ -238,11 +231,11 @@ jobs:
       - name: "Test wheel"
         if: ${{ startsWith(matrix.target, 'x86_64') }}
         run: |
-          pip install dist/"${PACKAGE_NAME}"-*.whl --force-reinstall
-          "${MODULE_NAME}" --help
-          python -m "${MODULE_NAME}" --help
+          pip install dist/${{ env.PACKAGE_NAME }}-*.whl --force-reinstall
+          ${{ env.MODULE_NAME }} --help
+          python -m ${{ env.MODULE_NAME }} --help
       - name: "Upload wheels"
-        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
+        uses: actions/upload-artifact@v4
         with:
           name: wheels-${{ matrix.target }}
           path: dist
@@ -260,7 +253,7 @@ jobs:
           tar czvf $ARCHIVE_FILE $ARCHIVE_NAME
           shasum -a 256 $ARCHIVE_FILE > $ARCHIVE_FILE.sha256
       - name: "Upload binary"
-        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
+        uses: actions/upload-artifact@v4
         with:
           name: artifacts-${{ matrix.target }}
           path: |
@@ -294,24 +287,23 @@ jobs:
             arch: arm
 
     steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
+      - uses: actions/checkout@v4
         with:
           submodules: recursive
-          persist-credentials: false
-      - uses: actions/setup-python@a26af69be951a213d495a4c3e4e4022e16d87065 # v5.6.0
+      - uses: actions/setup-python@v5
         with:
           python-version: ${{ env.PYTHON_VERSION }}
       - name: "Prep README.md"
         run: python scripts/transform_readme.py --target pypi
       - name: "Build wheels"
-        uses: PyO3/maturin-action@aef21716ff3dcae8a1c301d23ec3e4446972a6e3 # v1.49.1
+        uses: PyO3/maturin-action@v1
         with:
           target: ${{ matrix.platform.target }}
           manylinux: auto
           docker-options: ${{ matrix.platform.maturin_docker_options }}
           args: --release --locked --out dist
-      - uses: uraimo/run-on-arch-action@d94c13912ea685de38fccc1109385b83fd79427d # v3.0.1
-        if: ${{ matrix.platform.arch != 'ppc64' && matrix.platform.arch != 'ppc64le'}}
+      - uses: uraimo/run-on-arch-action@v2
+        if: matrix.platform.arch != 'ppc64'
         name: Test wheel
         with:
           arch: ${{ matrix.platform.arch == 'arm' && 'armv6' || matrix.platform.arch }}
@@ -325,7 +317,7 @@ jobs:
             pip3 install ${{ env.PACKAGE_NAME }} --no-index --find-links dist/ --force-reinstall
             ruff --help
       - name: "Upload wheels"
-        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
+        uses: actions/upload-artifact@v4
         with:
           name: wheels-${{ matrix.platform.target }}
           path: dist
@@ -343,7 +335,7 @@ jobs:
           tar czvf $ARCHIVE_FILE $ARCHIVE_NAME
           shasum -a 256 $ARCHIVE_FILE > $ARCHIVE_FILE.sha256
       - name: "Upload binary"
-        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
+        uses: actions/upload-artifact@v4
         with:
           name: artifacts-${{ matrix.platform.target }}
           path: |
@@ -359,25 +351,24 @@ jobs:
           - x86_64-unknown-linux-musl
           - i686-unknown-linux-musl
     steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
+      - uses: actions/checkout@v4
         with:
           submodules: recursive
-          persist-credentials: false
-      - uses: actions/setup-python@a26af69be951a213d495a4c3e4e4022e16d87065 # v5.6.0
+      - uses: actions/setup-python@v5
         with:
           python-version: ${{ env.PYTHON_VERSION }}
           architecture: x64
       - name: "Prep README.md"
         run: python scripts/transform_readme.py --target pypi
       - name: "Build wheels"
-        uses: PyO3/maturin-action@aef21716ff3dcae8a1c301d23ec3e4446972a6e3 # v1.49.1
+        uses: PyO3/maturin-action@v1
         with:
           target: ${{ matrix.target }}
           manylinux: musllinux_1_2
           args: --release --locked --out dist
       - name: "Test wheel"
         if: matrix.target == 'x86_64-unknown-linux-musl'
-        uses: addnab/docker-run-action@4f65fabd2431ebc8d299f8e5a018d79a769ae185 # v3
+        uses: addnab/docker-run-action@v3
         with:
           image: alpine:latest
           options: -v ${{ github.workspace }}:/io -w /io
@@ -387,7 +378,7 @@ jobs:
             .venv/bin/pip3 install ${{ env.PACKAGE_NAME }} --no-index --find-links dist/ --force-reinstall
             .venv/bin/${{ env.MODULE_NAME }} --help
       - name: "Upload wheels"
-        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
+        uses: actions/upload-artifact@v4
         with:
           name: wheels-${{ matrix.target }}
           path: dist
@@ -405,7 +396,7 @@ jobs:
           tar czvf $ARCHIVE_FILE $ARCHIVE_NAME
           shasum -a 256 $ARCHIVE_FILE > $ARCHIVE_FILE.sha256
       - name: "Upload binary"
-        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
+        uses: actions/upload-artifact@v4
         with:
           name: artifacts-${{ matrix.target }}
           path: |
@@ -425,23 +416,22 @@ jobs:
             arch: armv7
 
     steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
+      - uses: actions/checkout@v4
         with:
           submodules: recursive
-          persist-credentials: false
-      - uses: actions/setup-python@a26af69be951a213d495a4c3e4e4022e16d87065 # v5.6.0
+      - uses: actions/setup-python@v5
         with:
           python-version: ${{ env.PYTHON_VERSION }}
       - name: "Prep README.md"
         run: python scripts/transform_readme.py --target pypi
       - name: "Build wheels"
-        uses: PyO3/maturin-action@aef21716ff3dcae8a1c301d23ec3e4446972a6e3 # v1.49.1
+        uses: PyO3/maturin-action@v1
         with:
           target: ${{ matrix.platform.target }}
           manylinux: musllinux_1_2
           args: --release --locked --out dist
           docker-options: ${{ matrix.platform.maturin_docker_options }}
-      - uses: uraimo/run-on-arch-action@d94c13912ea685de38fccc1109385b83fd79427d # v3.0.1
+      - uses: uraimo/run-on-arch-action@v2
         name: Test wheel
         with:
           arch: ${{ matrix.platform.arch }}
@@ -454,7 +444,7 @@ jobs:
             .venv/bin/pip3 install ${{ env.PACKAGE_NAME }} --no-index --find-links dist/ --force-reinstall
             .venv/bin/${{ env.MODULE_NAME }} --help
       - name: "Upload wheels"
-        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
+        uses: actions/upload-artifact@v4
         with:
           name: wheels-${{ matrix.platform.target }}
           path: dist
@@ -472,7 +462,7 @@ jobs:
           tar czvf $ARCHIVE_FILE $ARCHIVE_NAME
           shasum -a 256 $ARCHIVE_FILE > $ARCHIVE_FILE.sha256
       - name: "Upload binary"
-        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
+        uses: actions/upload-artifact@v4
         with:
           name: artifacts-${{ matrix.platform.target }}
           path: |

.github/workflows/build-docker.yml

@@ -17,283 +17,52 @@ on:
     paths:
       - .github/workflows/build-docker.yml
 
-env:
-  RUFF_BASE_IMG: ghcr.io/${{ github.repository_owner }}/ruff
-
 jobs:
-  docker-build:
-    name: Build Docker image (ghcr.io/astral-sh/ruff) for ${{ matrix.platform }}
+  docker-publish:
+    name: Build Docker image (ghcr.io/astral-sh/ruff)
     runs-on: ubuntu-latest
     environment:
       name: release
-    strategy:
-      fail-fast: false
-      matrix:
-        platform:
-          - linux/amd64
-          - linux/arm64
     steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
+      - uses: actions/checkout@v4
         with:
           submodules: recursive
-          persist-credentials: false
 
-      - uses: docker/setup-buildx-action@b5ca514318bd6ebac0fb2aedd5d36ec1b5c232a2 # v3.10.0
+      - uses: docker/setup-buildx-action@v3
 
-      - uses: docker/login-action@74a5d142397b4f367a81961eba4e8cd7edddf772 # v3.4.0
+      - uses: docker/login-action@v3
         with:
           registry: ghcr.io
           username: ${{ github.repository_owner }}
           password: ${{ secrets.GITHUB_TOKEN }}
 
+      - name: Extract metadata (tags, labels) for Docker
+        id: meta
+        uses: docker/metadata-action@v5
+        with:
+          images: ghcr.io/astral-sh/ruff
+
       - name: Check tag consistency
         if: ${{ inputs.plan != '' && !fromJson(inputs.plan).announcement_tag_is_implicit }}
-        env:
-          TAG: ${{ inputs.plan != '' && fromJson(inputs.plan).announcement_tag || 'dry-run' }}
         run: |
-          version=$(grep -m 1 "^version = " pyproject.toml | sed -e 's/version = "\(.*\)"/\1/g')
-          if [ "${TAG}" != "${version}" ]; then
+          version=$(grep "version = " pyproject.toml | sed -e 's/version = "\(.*\)"/\1/g')
+          if [ "${{ fromJson(inputs.plan).announcement_tag }}" != "${version}" ]; then
             echo "The input tag does not match the version from pyproject.toml:" >&2
-            echo "${TAG}" >&2
+            echo "${{ fromJson(inputs.plan).announcement_tag }}" >&2
             echo "${version}" >&2
             exit 1
           else
             echo "Releasing ${version}"
           fi
 
-      - name: Extract metadata (tags, labels) for Docker
-        id: meta
-        uses: docker/metadata-action@902fa8ec7d6ecbf8d84d538b9b233a880e428804 # v5.7.0
-        with:
-          images: ${{ env.RUFF_BASE_IMG }}
-          # Defining this makes sure the org.opencontainers.image.version OCI label becomes the actual release version and not the branch name
-          tags: |
-            type=raw,value=dry-run,enable=${{ inputs.plan == '' || fromJson(inputs.plan).announcement_tag_is_implicit }}
-            type=pep440,pattern={{ version }},value=${{ inputs.plan != '' && fromJson(inputs.plan).announcement_tag || 'dry-run' }},enable=${{ inputs.plan != '' && !fromJson(inputs.plan).announcement_tag_is_implicit }}
-
-      - name: Normalize Platform Pair (replace / with -)
-        run: |
-          platform=${{ matrix.platform }}
-          echo "PLATFORM_TUPLE=${platform//\//-}" >> "$GITHUB_ENV"
-
-      # Adapted from https://docs.docker.com/build/ci/github-actions/multi-platform/
-      - name: Build and push by digest
-        id: build
-        uses: docker/build-push-action@1dc73863535b631f98b2378be8619f83b136f4a0 # v6.17.0
-        with:
-          context: .
-          platforms: ${{ matrix.platform }}
-          cache-from: type=gha,scope=ruff-${{ env.PLATFORM_TUPLE }}
-          cache-to: type=gha,mode=min,scope=ruff-${{ env.PLATFORM_TUPLE }}
-          labels: ${{ steps.meta.outputs.labels }}
-          outputs: type=image,name=${{ env.RUFF_BASE_IMG }},push-by-digest=true,name-canonical=true,push=${{ inputs.plan != '' && !fromJson(inputs.plan).announcement_tag_is_implicit }}
-
-      - name: Export digests
-        env:
-          digest: ${{ steps.build.outputs.digest }}
-        run: |
-          mkdir -p /tmp/digests
-          touch "/tmp/digests/${digest#sha256:}"
-
-      - name: Upload digests
-        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
-        with:
-          name: digests-${{ env.PLATFORM_TUPLE }}
-          path: /tmp/digests/*
-          if-no-files-found: error
-          retention-days: 1
-
-  docker-publish:
-    name: Publish Docker image (ghcr.io/astral-sh/ruff)
-    runs-on: ubuntu-latest
-    environment:
-      name: release
-    needs:
-      - docker-build
-    if: ${{ inputs.plan != '' && !fromJson(inputs.plan).announcement_tag_is_implicit }}
-    steps:
-      - name: Download digests
-        uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093 # v4.3.0
-        with:
-          path: /tmp/digests
-          pattern: digests-*
-          merge-multiple: true
-
-      - uses: docker/setup-buildx-action@b5ca514318bd6ebac0fb2aedd5d36ec1b5c232a2 # v3.10.0
-
-      - name: Extract metadata (tags, labels) for Docker
-        id: meta
-        uses: docker/metadata-action@902fa8ec7d6ecbf8d84d538b9b233a880e428804 # v5.7.0
-        with:
-          images: ${{ env.RUFF_BASE_IMG }}
-          # Order is on purpose such that the label org.opencontainers.image.version has the first pattern with the full version
-          tags: |
-            type=pep440,pattern={{ version }},value=${{ fromJson(inputs.plan).announcement_tag }}
-            type=pep440,pattern={{ major }}.{{ minor }},value=${{ fromJson(inputs.plan).announcement_tag }}
-
-      - uses: docker/login-action@74a5d142397b4f367a81961eba4e8cd7edddf772 # v3.4.0
-        with:
-          registry: ghcr.io
-          username: ${{ github.repository_owner }}
-          password: ${{ secrets.GITHUB_TOKEN }}
-
-      # Adapted from https://docs.docker.com/build/ci/github-actions/multi-platform/
-      - name: Create manifest list and push
-        working-directory: /tmp/digests
-        # The jq command expands the docker/metadata json "tags" array entry to `-t tag1 -t tag2 ...` for each tag in the array
-        # The printf will expand the base image with the `<RUFF_BASE_IMG>@sha256:<sha256> ...` for each sha256 in the directory
-        # The final command becomes `docker buildx imagetools create -t tag1 -t tag2 ... <RUFF_BASE_IMG>@sha256:<sha256_1> <RUFF_BASE_IMG>@sha256:<sha256_2> ...`
-        run: |
-          # shellcheck disable=SC2046
-          docker buildx imagetools create \
-            $(jq -cr '.tags | map("-t " + .) | join(" ")' <<< "$DOCKER_METADATA_OUTPUT_JSON") \
-            $(printf "${RUFF_BASE_IMG}@sha256:%s " *)
-
-  docker-publish-extra:
-    name: Publish additional Docker image based on ${{ matrix.image-mapping }}
-    runs-on: ubuntu-latest
-    environment:
-      name: release
-    needs:
-      - docker-publish
-    if: ${{ inputs.plan != '' && !fromJson(inputs.plan).announcement_tag_is_implicit }}
-    strategy:
-      fail-fast: false
-      matrix:
-        # Mapping of base image followed by a comma followed by one or more base tags (comma separated)
-        # Note, org.opencontainers.image.version label will use the first base tag (use the most specific tag first)
-        image-mapping:
-          - alpine:3.21,alpine3.21,alpine
-          - debian:bookworm-slim,bookworm-slim,debian-slim
-          - buildpack-deps:bookworm,bookworm,debian
-    steps:
-      - uses: docker/setup-buildx-action@b5ca514318bd6ebac0fb2aedd5d36ec1b5c232a2 # v3.10.0
-
-      - uses: docker/login-action@74a5d142397b4f367a81961eba4e8cd7edddf772 # v3.4.0
-        with:
-          registry: ghcr.io
-          username: ${{ github.repository_owner }}
-          password: ${{ secrets.GITHUB_TOKEN }}
-
-      - name: Generate Dynamic Dockerfile Tags
-        shell: bash
-        env:
-          TAG_VALUE: ${{ fromJson(inputs.plan).announcement_tag }}
-        run: |
-          set -euo pipefail
-
-          # Extract the image and tags from the matrix variable
-          IFS=',' read -r BASE_IMAGE BASE_TAGS <<< "${{ matrix.image-mapping }}"
-
-          # Generate Dockerfile content
-          cat <<EOF > Dockerfile
-          FROM ${BASE_IMAGE}
-          COPY --from=${RUFF_BASE_IMG}:latest /ruff /usr/local/bin/ruff
-          ENTRYPOINT []
-          CMD ["/usr/local/bin/ruff"]
-          EOF
-
-          # Initialize a variable to store all tag docker metadata patterns
-          TAG_PATTERNS=""
-
-          # Loop through all base tags and append its docker metadata pattern to the list
-          # Order is on purpose such that the label org.opencontainers.image.version has the first pattern with the full version
-          IFS=','; for TAG in ${BASE_TAGS}; do
-            TAG_PATTERNS="${TAG_PATTERNS}type=pep440,pattern={{ version }},suffix=-${TAG},value=${TAG_VALUE}\n"
-            TAG_PATTERNS="${TAG_PATTERNS}type=pep440,pattern={{ major }}.{{ minor }},suffix=-${TAG},value=${TAG_VALUE}\n"
-            TAG_PATTERNS="${TAG_PATTERNS}type=raw,value=${TAG}\n"
-          done
-
-          # Remove the trailing newline from the pattern list
-          TAG_PATTERNS="${TAG_PATTERNS%\\n}"
-
-          # Export image cache name
-          echo "IMAGE_REF=${BASE_IMAGE//:/-}" >> "$GITHUB_ENV"
-
-          # Export tag patterns using the multiline env var syntax
-          {
-            echo "TAG_PATTERNS<<EOF"
-            echo -e "${TAG_PATTERNS}"
-            echo EOF
-          } >> "$GITHUB_ENV"
-
-      - name: Extract metadata (tags, labels) for Docker
-        id: meta
-        uses: docker/metadata-action@902fa8ec7d6ecbf8d84d538b9b233a880e428804 # v5.7.0
-        # ghcr.io prefers index level annotations
-        env:
-          DOCKER_METADATA_ANNOTATIONS_LEVELS: index
-        with:
-          images: ${{ env.RUFF_BASE_IMG }}
-          flavor: |
-            latest=false
-          tags: |
-            ${{ env.TAG_PATTERNS }}
-
-      - name: Build and push
-        uses: docker/build-push-action@1dc73863535b631f98b2378be8619f83b136f4a0 # v6.17.0
+      - name: "Build and push Docker image"
+        uses: docker/build-push-action@v6
         with:
           context: .
           platforms: linux/amd64,linux/arm64
-          # We do not really need to cache here as the Dockerfile is tiny
-          #cache-from: type=gha,scope=ruff-${{ env.IMAGE_REF }}
-          #cache-to: type=gha,mode=min,scope=ruff-${{ env.IMAGE_REF }}
-          push: true
-          tags: ${{ steps.meta.outputs.tags }}
+          # Reuse the builder
+          cache-from: type=gha
+          cache-to: type=gha,mode=max
+          push: ${{ inputs.plan != '' && !fromJson(inputs.plan).announcement_tag_is_implicit }}
+          tags: ghcr.io/astral-sh/ruff:latest,ghcr.io/astral-sh/ruff:${{ (inputs.plan != '' && fromJson(inputs.plan).announcement_tag) || 'dry-run' }}
           labels: ${{ steps.meta.outputs.labels }}
-          annotations: ${{ steps.meta.outputs.annotations }}
-
-  # This is effectively a duplicate of `docker-publish` to make https://github.com/astral-sh/ruff/pkgs/container/ruff
-  # show the ruff base image first since GitHub always shows the last updated image digests
-  # This works by annotating the original digests (previously non-annotated) which triggers an update to ghcr.io
-  docker-republish:
-    name: Annotate Docker image (ghcr.io/astral-sh/ruff)
-    runs-on: ubuntu-latest
-    environment:
-      name: release
-    needs:
-      - docker-publish-extra
-    if: ${{ inputs.plan != '' && !fromJson(inputs.plan).announcement_tag_is_implicit }}
-    steps:
-      - name: Download digests
-        uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093 # v4.3.0
-        with:
-          path: /tmp/digests
-          pattern: digests-*
-          merge-multiple: true
-
-      - uses: docker/setup-buildx-action@b5ca514318bd6ebac0fb2aedd5d36ec1b5c232a2 # v3.10.0
-
-      - name: Extract metadata (tags, labels) for Docker
-        id: meta
-        uses: docker/metadata-action@902fa8ec7d6ecbf8d84d538b9b233a880e428804 # v5.7.0
-        env:
-          DOCKER_METADATA_ANNOTATIONS_LEVELS: index
-        with:
-          images: ${{ env.RUFF_BASE_IMG }}
-          # Order is on purpose such that the label org.opencontainers.image.version has the first pattern with the full version
-          tags: |
-            type=pep440,pattern={{ version }},value=${{ fromJson(inputs.plan).announcement_tag }}
-            type=pep440,pattern={{ major }}.{{ minor }},value=${{ fromJson(inputs.plan).announcement_tag }}
-
-      - uses: docker/login-action@74a5d142397b4f367a81961eba4e8cd7edddf772 # v3.4.0
-        with:
-          registry: ghcr.io
-          username: ${{ github.repository_owner }}
-          password: ${{ secrets.GITHUB_TOKEN }}
-
-      # Adapted from https://docs.docker.com/build/ci/github-actions/multi-platform/
-      - name: Create manifest list and push
-        working-directory: /tmp/digests
-        # The readarray part is used to make sure the quoting and special characters are preserved on expansion (e.g. spaces)
-        # The jq command expands the docker/metadata json "tags" array entry to `-t tag1 -t tag2 ...` for each tag in the array
-        # The printf will expand the base image with the `<RUFF_BASE_IMG>@sha256:<sha256> ...` for each sha256 in the directory
-        # The final command becomes `docker buildx imagetools create -t tag1 -t tag2 ... <RUFF_BASE_IMG>@sha256:<sha256_1> <RUFF_BASE_IMG>@sha256:<sha256_2> ...`
-        run: |
-          readarray -t lines <<< "$DOCKER_METADATA_OUTPUT_ANNOTATIONS"; annotations=(); for line in "${lines[@]}"; do annotations+=(--annotation "$line"); done
-
-          # shellcheck disable=SC2046
-          docker buildx imagetools create \
-            "${annotations[@]}" \
-            $(jq -cr '.tags | map("-t " + .) | join(" ")' <<< "$DOCKER_METADATA_OUTPUT_JSON") \
-            $(printf "${RUFF_BASE_IMG}@sha256:%s " *)

.github/workflows/daily_fuzz.yaml

@@ -31,31 +31,25 @@ jobs:
     # Don't run the cron job on forks:
     if: ${{ github.repository == 'astral-sh/ruff' || github.event_name != 'schedule' }}
     steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v5
         with:
-          persist-credentials: false
-      - uses: astral-sh/setup-uv@6b9c6063abd6010835644d4c2e1bef4cf5cd0fca # v6.0.1
+          python-version: "3.12"
+      - name: Install uv
+        run: curl -LsSf https://astral.sh/uv/install.sh | sh
+      - name: Install Python requirements
+        run: uv pip install -r scripts/fuzz-parser/requirements.txt --system
       - name: "Install Rust toolchain"
         run: rustup show
       - name: "Install mold"
-        uses: rui314/setup-mold@e16410e7f8d9e167b74ad5697a9089a35126eb50 # v1
-      - uses: Swatinem/rust-cache@9d47c6ad4b02e050fd481d890b2ea34778fd09d6 # v2.7.8
+        uses: rui314/setup-mold@v1
+      - uses: Swatinem/rust-cache@v2
       - name: Build ruff
         # A debug build means the script runs slower once it gets started,
         # but this is outweighed by the fact that a release build takes *much* longer to compile in CI
         run: cargo build --locked
       - name: Fuzz
-        run: |
-          # shellcheck disable=SC2046
-          (
-            uvx \
-            --python=3.12 \
-            --from=./python/py-fuzzer \
-            fuzz \
-            --test-executable=target/debug/ruff \
-            --bin=ruff \
-            $(shuf -i 0-9999999999999999999 -n 1000)
-          )
+        run: python scripts/fuzz-parser/fuzz.py $(shuf -i 0-9999999999999999999 -n 1000) --test-executable target/debug/ruff
 
   create-issue-on-failure:
     name: Create an issue if the daily fuzz surfaced any bugs
@@ -65,7 +59,7 @@ jobs:
     permissions:
       issues: write
     steps:
-      - uses: actions/github-script@60a0d83039c74a4aee543508d2ffcb1c3799cdea # v7.0.1
+      - uses: actions/github-script@v7
         with:
           github-token: ${{ secrets.GITHUB_TOKEN }}
           script: |
@@ -73,6 +67,6 @@ jobs:
               owner: "astral-sh",
               repo: "ruff",
               title: `Daily parser fuzz failed on ${new Date().toDateString()}`,
-              body: "Run listed here: https://github.com/${{ github.repository }}/actions/runs/${{ github.run_id }}",
+              body: "Runs listed here: https://github.com/astral-sh/ruff/actions/workflows/daily_fuzz.yml",
               labels: ["bug", "parser", "fuzzer"],
             })

.github/workflows/mypy_primer.yaml

@@ -1,101 +0,0 @@
-name: Run mypy_primer
-
-permissions: {}
-
-on:
-  pull_request:
-    paths:
-      - "crates/ty*/**"
-      - "crates/ruff_db"
-      - "crates/ruff_python_ast"
-      - "crates/ruff_python_parser"
-      - ".github/workflows/mypy_primer.yaml"
-      - ".github/workflows/mypy_primer_comment.yaml"
-      - "Cargo.lock"
-
-concurrency:
-  group: ${{ github.workflow }}-${{ github.ref_name }}-${{ github.event.pull_request.number || github.sha }}
-  cancel-in-progress: true
-
-env:
-  CARGO_INCREMENTAL: 0
-  CARGO_NET_RETRY: 10
-  CARGO_TERM_COLOR: always
-  RUSTUP_MAX_RETRIES: 10
-  RUST_BACKTRACE: 1
-
-jobs:
-  mypy_primer:
-    name: Run mypy_primer
-    runs-on: depot-ubuntu-22.04-32
-    timeout-minutes: 20
-    steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
-        with:
-          path: ruff
-          fetch-depth: 0
-          persist-credentials: false
-
-      - name: Install the latest version of uv
-        uses: astral-sh/setup-uv@6b9c6063abd6010835644d4c2e1bef4cf5cd0fca # v6.0.1
-
-      - uses: Swatinem/rust-cache@9d47c6ad4b02e050fd481d890b2ea34778fd09d6 # v2.7.8
-        with:
-          workspaces: "ruff"
-
-      - name: Install Rust toolchain
-        run: rustup show
-
-      - name: Run mypy_primer
-        shell: bash
-        run: |
-          cd ruff
-
-          echo "Enabling mypy primer specific configuration overloads (see .github/mypy-primer-ty.toml)"
-          mkdir -p ~/.config/ty
-          cp .github/mypy-primer-ty.toml ~/.config/ty/ty.toml
-
-          PRIMER_SELECTOR="$(paste -s -d'|' crates/ty_python_semantic/resources/primer/good.txt)"
-
-          echo "new commit"
-          git rev-list --format=%s --max-count=1 "$GITHUB_SHA"
-
-          MERGE_BASE="$(git merge-base "$GITHUB_SHA" "origin/$GITHUB_BASE_REF")"
-          git checkout -b base_commit "$MERGE_BASE"
-          echo "base commit"
-          git rev-list --format=%s --max-count=1 base_commit
-
-          cd ..
-
-          echo "Project selector: $PRIMER_SELECTOR"
-          # Allow the exit code to be 0 or 1, only fail for actual mypy_primer crashes/bugs
-          uvx \
-            --from="git+https://github.com/hauntsaninja/mypy_primer@01a7ca325f674433c58e02416a867178d1571128" \
-            mypy_primer \
-            --repo ruff \
-            --type-checker ty \
-            --old base_commit \
-            --new "$GITHUB_SHA" \
-            --project-selector "/($PRIMER_SELECTOR)\$" \
-            --output concise \
-            --debug > mypy_primer.diff || [ $? -eq 1 ]
-
-          # Output diff with ANSI color codes
-          cat mypy_primer.diff
-
-          # Remove ANSI color codes before uploading
-          sed -ie 's/\x1b\[[0-9;]*m//g' mypy_primer.diff
-
-          echo ${{ github.event.number }} > pr-number
-
-      - name: Upload diff
-        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
-        with:
-          name: mypy_primer_diff
-          path: mypy_primer.diff
-
-      - name: Upload pr-number
-        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
-        with:
-          name: pr-number
-          path: pr-number

.github/workflows/mypy_primer_comment.yaml

@@ -1,97 +0,0 @@
-name: PR comment (mypy_primer)
-
-on: # zizmor: ignore[dangerous-triggers]
-  workflow_run:
-    workflows: [Run mypy_primer]
-    types: [completed]
-  workflow_dispatch:
-    inputs:
-      workflow_run_id:
-        description: The mypy_primer workflow that triggers the workflow run
-        required: true
-
-jobs:
-  comment:
-    runs-on: ubuntu-24.04
-    permissions:
-      pull-requests: write
-    steps:
-      - uses: dawidd6/action-download-artifact@20319c5641d495c8a52e688b7dc5fada6c3a9fbc # v8
-        name: Download PR number
-        with:
-          name: pr-number
-          run_id: ${{ github.event.workflow_run.id ||  github.event.inputs.workflow_run_id }}
-          if_no_artifact_found: ignore
-          allow_forks: true
-
-      - name: Parse pull request number
-        id: pr-number
-        run: |
-          if [[ -f pr-number ]]
-          then
-            echo "pr-number=$(<pr-number)" >> "$GITHUB_OUTPUT"
-          fi
-
-      - uses: dawidd6/action-download-artifact@20319c5641d495c8a52e688b7dc5fada6c3a9fbc # v8
-        name: "Download mypy_primer results"
-        id: download-mypy_primer_diff
-        if: steps.pr-number.outputs.pr-number
-        with:
-          name: mypy_primer_diff
-          workflow: mypy_primer.yaml
-          pr: ${{ steps.pr-number.outputs.pr-number }}
-          path: pr/mypy_primer_diff
-          workflow_conclusion: completed
-          if_no_artifact_found: ignore
-          allow_forks: true
-
-      - name: Generate comment content
-        id: generate-comment
-        if: steps.download-mypy_primer_diff.outputs.found_artifact == 'true'
-        run: |
-          # Guard against malicious mypy_primer results that symlink to a secret
-          # file on this runner
-          if [[ -L pr/mypy_primer_diff/mypy_primer.diff ]]
-          then
-              echo "Error: mypy_primer.diff cannot be a symlink"
-              exit 1
-          fi
-
-          # Note this identifier is used to find the comment to update on
-          # subsequent runs
-          echo '<!-- generated-comment mypy_primer -->' >> comment.txt
-
-          echo '## `mypy_primer` results' >> comment.txt
-          if [ -s "pr/mypy_primer_diff/mypy_primer.diff" ]; then
-            echo '<details>' >> comment.txt
-            echo '<summary>Changes were detected when running on open source projects</summary>' >> comment.txt
-            echo '' >> comment.txt
-            echo '```diff' >> comment.txt
-            cat pr/mypy_primer_diff/mypy_primer.diff >> comment.txt
-            echo '```' >> comment.txt
-            echo '</details>' >> comment.txt
-          else
-            echo 'No ecosystem changes detected ✅' >> comment.txt
-          fi
-
-          echo 'comment<<EOF' >> "$GITHUB_OUTPUT"
-          cat comment.txt >> "$GITHUB_OUTPUT"
-          echo 'EOF' >> "$GITHUB_OUTPUT"
-
-      - name: Find existing comment
-        uses: peter-evans/find-comment@3eae4d37986fb5a8592848f6a574fdf654e61f9e # v3.1.0
-        if: steps.generate-comment.outcome == 'success'
-        id: find-comment
-        with:
-          issue-number: ${{ steps.pr-number.outputs.pr-number }}
-          comment-author: "github-actions[bot]"
-          body-includes: "<!-- generated-comment mypy_primer -->"
-
-      - name: Create or update comment
-        if: steps.find-comment.outcome == 'success'
-        uses: peter-evans/create-or-update-comment@71345be0265236311c031f5c7866368bd1eff043 # v4
-        with:
-          comment-id: ${{ steps.find-comment.outputs.comment-id }}
-          issue-number: ${{ steps.pr-number.outputs.pr-number }}
-          body-path: comment.txt
-          edit-mode: replace

.github/workflows/notify-dependents.yml

@@ -17,7 +17,7 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - name: "Update pre-commit mirror"
-        uses: actions/github-script@60a0d83039c74a4aee543508d2ffcb1c3799cdea # v7.0.1
+        uses: actions/github-script@v7
         with:
           github-token: ${{ secrets.RUFF_PRE_COMMIT_PAT }}
           script: |

.github/workflows/pr-comment.yaml

@@ -10,13 +10,14 @@ on:
         description: The ecosystem workflow that triggers the workflow run
         required: true
 
+permissions:
+  pull-requests: write
+
 jobs:
   comment:
     runs-on: ubuntu-latest
-    permissions:
-      pull-requests: write
     steps:
-      - uses: dawidd6/action-download-artifact@20319c5641d495c8a52e688b7dc5fada6c3a9fbc # v8
+      - uses: dawidd6/action-download-artifact@v6
         name: Download pull request number
         with:
           name: pr-number
@@ -29,10 +30,10 @@ jobs:
         run: |
           if [[ -f pr-number ]]
           then
-            echo "pr-number=$(<pr-number)" >> "$GITHUB_OUTPUT"
+            echo "pr-number=$(<pr-number)" >> $GITHUB_OUTPUT
           fi
 
-      - uses: dawidd6/action-download-artifact@20319c5641d495c8a52e688b7dc5fada6c3a9fbc # v8
+      - uses: dawidd6/action-download-artifact@v6
         name: "Download ecosystem results"
         id: download-ecosystem-result
         if: steps.pr-number.outputs.pr-number
@@ -65,12 +66,12 @@ jobs:
           cat pr/ecosystem/ecosystem-result >> comment.txt
           echo "" >> comment.txt
 
-          echo 'comment<<EOF' >> "$GITHUB_OUTPUT"
-          cat comment.txt >> "$GITHUB_OUTPUT"
-          echo 'EOF' >> "$GITHUB_OUTPUT"
+          echo 'comment<<EOF' >> $GITHUB_OUTPUT
+          cat comment.txt >> $GITHUB_OUTPUT
+          echo 'EOF' >> $GITHUB_OUTPUT
 
       - name: Find existing comment
-        uses: peter-evans/find-comment@3eae4d37986fb5a8592848f6a574fdf654e61f9e # v3.1.0
+        uses: peter-evans/find-comment@v3
         if: steps.generate-comment.outcome == 'success'
         id: find-comment
         with:
@@ -80,7 +81,7 @@ jobs:
 
       - name: Create or update comment
         if: steps.find-comment.outcome == 'success'
-        uses: peter-evans/create-or-update-comment@71345be0265236311c031f5c7866368bd1eff043 # v4
+        uses: peter-evans/create-or-update-comment@v4
         with:
           comment-id: ${{ steps.find-comment.outputs.comment-id }}
           issue-number: ${{ steps.pr-number.outputs.pr-number }}

.github/workflows/publish-docs.yml

@@ -23,19 +23,17 @@ jobs:
     env:
       MKDOCS_INSIDERS_SSH_KEY_EXISTS: ${{ secrets.MKDOCS_INSIDERS_SSH_KEY != '' }}
     steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
+      - uses: actions/checkout@v4
         with:
           ref: ${{ inputs.ref }}
-          persist-credentials: true
 
-      - uses: actions/setup-python@a26af69be951a213d495a4c3e4e4022e16d87065 # v5.6.0
+      - uses: actions/setup-python@v5
         with:
           python-version: 3.12
 
       - name: "Set docs version"
-        env:
-          version: ${{ (inputs.plan != '' && fromJson(inputs.plan).announcement_tag) || inputs.ref }}
         run: |
+          version="${{ (inputs.plan != '' && fromJson(inputs.plan).announcement_tag) || inputs.ref }}"
           # if version is missing, use 'latest'
           if [ -z "$version" ]; then
             echo "Using 'latest' as version"
@@ -45,30 +43,32 @@ jobs:
           # Use version as display name for now
           display_name="$version"
 
-          echo "version=$version" >> "$GITHUB_ENV"
-          echo "display_name=$display_name" >> "$GITHUB_ENV"
+          echo "version=$version" >> $GITHUB_ENV
+          echo "display_name=$display_name" >> $GITHUB_ENV
 
       - name: "Set branch name"
         run: |
+          version="${{ env.version }}"
+          display_name="${{ env.display_name }}"
           timestamp="$(date +%s)"
 
           # create branch_display_name from display_name by replacing all
           # characters disallowed in git branch names with hyphens
-          branch_display_name="$(echo "${display_name}" | tr -c '[:alnum:]._' '-' | tr -s '-')"
+          branch_display_name="$(echo "$display_name" | tr -c '[:alnum:]._' '-' | tr -s '-')"
 
-          echo "branch_name=update-docs-$branch_display_name-$timestamp" >> "$GITHUB_ENV"
-          echo "timestamp=$timestamp" >> "$GITHUB_ENV"
+          echo "branch_name=update-docs-$branch_display_name-$timestamp" >> $GITHUB_ENV
+          echo "timestamp=$timestamp" >> $GITHUB_ENV
 
       - name: "Add SSH key"
         if: ${{ env.MKDOCS_INSIDERS_SSH_KEY_EXISTS == 'true' }}
-        uses: webfactory/ssh-agent@a6f90b1f127823b31d4d4a8d96047790581349bd # v0.9.1
+        uses: webfactory/ssh-agent@v0.9.0
         with:
           ssh-private-key: ${{ secrets.MKDOCS_INSIDERS_SSH_KEY }}
 
       - name: "Install Rust toolchain"
         run: rustup show
 
-      - uses: Swatinem/rust-cache@9d47c6ad4b02e050fd481d890b2ea34778fd09d6 # v2.7.8
+      - uses: Swatinem/rust-cache@v2
 
       - name: "Install Insiders dependencies"
         if: ${{ env.MKDOCS_INSIDERS_SSH_KEY_EXISTS == 'true' }}
@@ -92,7 +92,9 @@ jobs:
         run: mkdocs build --strict -f mkdocs.public.yml
 
       - name: "Clone docs repo"
-        run: git clone https://${{ secrets.ASTRAL_DOCS_PAT }}@github.com/astral-sh/docs.git astral-docs
+        run: |
+          version="${{ env.version }}"
+          git clone https://${{ secrets.ASTRAL_DOCS_PAT }}@github.com/astral-sh/docs.git astral-docs
 
       - name: "Copy docs"
         run: rm -rf astral-docs/site/ruff && mkdir -p astral-docs/site && cp -r site/ruff astral-docs/site/
@@ -100,10 +102,12 @@ jobs:
       - name: "Commit docs"
         working-directory: astral-docs
         run: |
+          branch_name="${{ env.branch_name }}"
+
           git config user.name "astral-docs-bot"
           git config user.email "176161322+astral-docs-bot@users.noreply.github.com"
 
-          git checkout -b "${branch_name}"
+          git checkout -b $branch_name
           git add site/ruff
           git commit -m "Update ruff documentation for $version"
 
@@ -112,8 +116,12 @@ jobs:
         env:
           GITHUB_TOKEN: ${{ secrets.ASTRAL_DOCS_PAT }}
         run: |
+          version="${{ env.version }}"
+          display_name="${{ env.display_name }}"
+          branch_name="${{ env.branch_name }}"
+
           # set the PR title
-          pull_request_title="Update ruff documentation for ${display_name}"
+          pull_request_title="Update ruff documentation for $display_name"
 
           # Delete any existing pull requests that are open for this version
           # by checking against pull_request_title because the new PR will
@@ -122,15 +130,13 @@ jobs:
             xargs -I {} gh pr close {}
 
           # push the branch to GitHub
-          git push origin "${branch_name}"
+          git push origin $branch_name
 
           # create the PR
-          gh pr create \
-            --base=main \
-            --head="${branch_name}" \
-            --title="${pull_request_title}" \
-            --body="Automated documentation update for ${display_name}" \
-            --label="documentation"
+          gh pr create --base main --head $branch_name \
+            --title "$pull_request_title" \
+            --body "Automated documentation update for $display_name" \
+            --label "documentation"
 
       - name: "Merge Pull Request"
         if: ${{ inputs.plan != '' && !fromJson(inputs.plan).announcement_tag_is_implicit }}
@@ -138,7 +144,9 @@ jobs:
         env:
           GITHUB_TOKEN: ${{ secrets.ASTRAL_DOCS_PAT }}
         run: |
+          branch_name="${{ env.branch_name }}"
+
           # auto-merge the PR if the build was triggered by a release. Manual builds should be reviewed by a human.
           # give the PR a few seconds to be created before trying to auto-merge it
           sleep 10
-          gh pr merge --squash "${branch_name}"
+          gh pr merge --squash $branch_name

.github/workflows/publish-playground.yml

@@ -24,31 +24,32 @@ jobs:
     env:
       CF_API_TOKEN_EXISTS: ${{ secrets.CF_API_TOKEN != '' }}
     steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
-        with:
-          persist-credentials: false
+      - uses: actions/checkout@v4
       - name: "Install Rust toolchain"
         run: rustup target add wasm32-unknown-unknown
-      - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0
+      - uses: actions/setup-node@v4
         with:
-          node-version: 22
+          node-version: 18
           cache: "npm"
           cache-dependency-path: playground/package-lock.json
-      - uses: jetli/wasm-bindgen-action@20b33e20595891ab1a0ed73145d8a21fc96e7c29 # v0.2.0
+      - uses: jetli/wasm-pack-action@v0.4.0
+      - uses: jetli/wasm-bindgen-action@v0.2.0
+      - name: "Run wasm-pack"
+        run: wasm-pack build --target web --out-dir ../../playground/src/pkg crates/ruff_wasm
       - name: "Install Node dependencies"
         run: npm ci
         working-directory: playground
       - name: "Run TypeScript checks"
         run: npm run check
         working-directory: playground
-      - name: "Build Ruff playground"
-        run: npm run build --workspace ruff-playground
+      - name: "Build JavaScript bundle"
+        run: npm run build
         working-directory: playground
       - name: "Deploy to Cloudflare Pages"
         if: ${{ env.CF_API_TOKEN_EXISTS == 'true' }}
-        uses: cloudflare/wrangler-action@da0e0dfe58b7a431659754fdf3f186c529afbe65 # v3.14.1
+        uses: cloudflare/wrangler-action@v3.7.0
         with:
           apiToken: ${{ secrets.CF_API_TOKEN }}
           accountId: ${{ secrets.CF_ACCOUNT_ID }}
           # `github.head_ref` is only set during pull requests and for manual runs or tags we use `main` to deploy to production
-          command: pages deploy playground/ruff/dist --project-name=ruff-playground --branch ${{ github.head_ref || 'main' }} --commit-hash ${GITHUB_SHA}
+          command: pages deploy playground/dist --project-name=ruff-playground --branch ${{ github.head_ref || 'main' }} --commit-hash ${GITHUB_SHA}

.github/workflows/publish-pypi.yml

@@ -21,12 +21,14 @@ jobs:
       # For PyPI's trusted publishing.
       id-token: write
     steps:
-      - name: "Install uv"
-        uses: astral-sh/setup-uv@6b9c6063abd6010835644d4c2e1bef4cf5cd0fca # v6.0.1
-      - uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093 # v4.3.0
+      - uses: actions/download-artifact@v4
         with:
           pattern: wheels-*
           path: wheels
           merge-multiple: true
       - name: Publish to PyPi
-        run: uv publish -v wheels/*
+        uses: pypa/gh-action-pypi-publish@release/v1
+        with:
+          skip-existing: true
+          packages-dir: wheels
+          verbose: true

.github/workflows/publish-ty-playground.yml

@@ -1,58 +0,0 @@
-# Publish the ty playground.
-name: "[ty Playground] Release"
-
-permissions: {}
-
-on:
-  push:
-    branches: [main]
-    paths:
-      - "crates/ty*/**"
-      - "crates/ruff_db/**"
-      - "crates/ruff_python_ast/**"
-      - "crates/ruff_python_parser/**"
-      - "playground/**"
-      - ".github/workflows/publish-ty-playground.yml"
-
-concurrency:
-  group: ${{ github.workflow }}-${{ github.ref_name }}
-  cancel-in-progress: true
-
-env:
-  CARGO_INCREMENTAL: 0
-  CARGO_NET_RETRY: 10
-  CARGO_TERM_COLOR: always
-  RUSTUP_MAX_RETRIES: 10
-
-jobs:
-  publish:
-    runs-on: ubuntu-latest
-    env:
-      CF_API_TOKEN_EXISTS: ${{ secrets.CF_API_TOKEN != '' }}
-    steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
-        with:
-          persist-credentials: false
-      - name: "Install Rust toolchain"
-        run: rustup target add wasm32-unknown-unknown
-      - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0
-        with:
-          node-version: 22
-      - uses: jetli/wasm-bindgen-action@20b33e20595891ab1a0ed73145d8a21fc96e7c29 # v0.2.0
-      - name: "Install Node dependencies"
-        run: npm ci
-        working-directory: playground
-      - name: "Run TypeScript checks"
-        run: npm run check
-        working-directory: playground
-      - name: "Build ty playground"
-        run: npm run build --workspace ty-playground
-        working-directory: playground
-      - name: "Deploy to Cloudflare Pages"
-        if: ${{ env.CF_API_TOKEN_EXISTS == 'true' }}
-        uses: cloudflare/wrangler-action@da0e0dfe58b7a431659754fdf3f186c529afbe65 # v3.14.1
-        with:
-          apiToken: ${{ secrets.CF_API_TOKEN }}
-          accountId: ${{ secrets.CF_ACCOUNT_ID }}
-          # `github.head_ref` is only set during pull requests and for manual runs or tags we use `main` to deploy to production
-          command: pages deploy playground/ty/dist --project-name=ty-playground --branch ${{ github.head_ref || 'main' }} --commit-hash ${GITHUB_SHA}

.github/workflows/publish-wasm.yml

@@ -29,15 +29,11 @@ jobs:
         target: [web, bundler, nodejs]
       fail-fast: false
     steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
-        with:
-          persist-credentials: false
+      - uses: actions/checkout@v4
       - name: "Install Rust toolchain"
         run: rustup target add wasm32-unknown-unknown
-      - uses: jetli/wasm-pack-action@0d096b08b4e5a7de8c28de67e11e945404e9eefa # v0.4.0
-        with:
-          version: v0.13.1
-      - uses: jetli/wasm-bindgen-action@20b33e20595891ab1a0ed73145d8a21fc96e7c29 # v0.2.0
+      - uses: jetli/wasm-pack-action@v0.4.0
+      - uses: jetli/wasm-bindgen-action@v0.2.0
       - name: "Run wasm-pack build"
         run: wasm-pack build --target ${{ matrix.target }} crates/ruff_wasm
       - name: "Rename generated package"
@@ -45,9 +41,9 @@ jobs:
           jq '.name="@astral-sh/ruff-wasm-${{ matrix.target }}"' crates/ruff_wasm/pkg/package.json > /tmp/package.json
           mv /tmp/package.json crates/ruff_wasm/pkg
       - run: cp LICENSE crates/ruff_wasm/pkg # wasm-pack does not put the LICENSE file in the pkg
-      - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0
+      - uses: actions/setup-node@v4
         with:
-          node-version: 20
+          node-version: 18
           registry-url: "https://registry.npmjs.org"
       - name: "Publish (dry-run)"
         if: ${{ inputs.plan == '' || fromJson(inputs.plan).announcement_tag_is_implicit }}

.github/workflows/release.yml

@@ -1,13 +1,12 @@
-# This file was autogenerated by dist: https://github.com/astral-sh/cargo-dist
+# This file was autogenerated by cargo-dist: https://opensource.axo.dev/cargo-dist/
 #
 # Copyright 2022-2024, axodotdev
-# Copyright 2025 Astral Software Inc.
 # SPDX-License-Identifier: MIT or Apache-2.0
 #
 # CI that:
 #
 # * checks for a Git Tag that looks like a release
-# * builds artifacts with dist (archives, installers, hashes)
+# * builds artifacts with cargo-dist (archives, installers, hashes)
 # * uploads those artifacts to temporary workflow zip
 # * on success, uploads the artifacts to a GitHub Release
 #
@@ -25,10 +24,10 @@ permissions:
 # must be a Cargo-style SemVer Version (must have at least major.minor.patch).
 #
 # If PACKAGE_NAME is specified, then the announcement will be for that
-# package (erroring out if it doesn't have the given version or isn't dist-able).
+# package (erroring out if it doesn't have the given version or isn't cargo-dist-able).
 #
 # If PACKAGE_NAME isn't specified, then the announcement will be for all
-# (dist-able) packages in the workspace with that version (this mode is
+# (cargo-dist-able) packages in the workspace with that version (this mode is
 # intended for workspaces with only one dist-able package, or with all dist-able
 # packages versioned/released in lockstep).
 #
@@ -40,7 +39,6 @@ permissions:
 # If there's a prerelease-style suffix to the version, then the release(s)
 # will be marked as a prerelease.
 on:
-  pull_request:
   workflow_dispatch:
     inputs:
       tag:
@@ -50,9 +48,9 @@ on:
         type: string
 
 jobs:
-  # Run 'dist plan' (or host) to determine what tasks we need to do
+  # Run 'cargo dist plan' (or host) to determine what tasks we need to do
   plan:
-    runs-on: "depot-ubuntu-latest-4"
+    runs-on: "ubuntu-20.04"
     outputs:
       val: ${{ steps.plan.outputs.manifest }}
       tag: ${{ (inputs.tag != 'dry-run' && inputs.tag) || '' }}
@@ -61,20 +59,19 @@ jobs:
     env:
       GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
     steps:
-      - uses: actions/checkout@85e6279cec87321a52edac9c87bce653a07cf6c2
+      - uses: actions/checkout@v4
         with:
-          persist-credentials: false
           submodules: recursive
-      - name: Install dist
+      - name: Install cargo-dist
         # we specify bash to get pipefail; it guards against the `curl` command
         # failing. otherwise `sh` won't catch that `curl` returned non-0
         shell: bash
-        run: "curl --proto '=https' --tlsv1.2 -LsSf https://github.com/astral-sh/cargo-dist/releases/download/v0.28.5-prerelease.1/cargo-dist-installer.sh | sh"
-      - name: Cache dist
-        uses: actions/upload-artifact@6027e3dd177782cd8ab9af838c04fd81a07f1d47
+        run: "curl --proto '=https' --tlsv1.2 -LsSf https://github.com/axodotdev/cargo-dist/releases/download/v0.22.1/cargo-dist-installer.sh | sh"
+      - name: Cache cargo-dist
+        uses: actions/upload-artifact@v4
         with:
           name: cargo-dist-cache
-          path: ~/.cargo/bin/dist
+          path: ~/.cargo/bin/cargo-dist
       # sure would be cool if github gave us proper conditionals...
       # so here's a doubly-nested ternary-via-truthiness to try to provide the best possible
       # functionality based on whether this is a pull_request, and whether it's from a fork.
@@ -82,12 +79,12 @@ jobs:
       # but also really annoying to build CI around when it needs secrets to work right.)
       - id: plan
         run: |
-          dist ${{ (inputs.tag && inputs.tag != 'dry-run' && format('host --steps=create --tag={0}', inputs.tag)) || 'plan' }} --output-format=json > plan-dist-manifest.json
-          echo "dist ran successfully"
+          cargo dist ${{ (inputs.tag && inputs.tag != 'dry-run' && format('host --steps=create --tag={0}', inputs.tag)) || 'plan' }} --output-format=json > plan-dist-manifest.json
+          echo "cargo dist ran successfully"
           cat plan-dist-manifest.json
           echo "manifest=$(jq -c "." plan-dist-manifest.json)" >> "$GITHUB_OUTPUT"
       - name: "Upload dist-manifest.json"
-        uses: actions/upload-artifact@6027e3dd177782cd8ab9af838c04fd81a07f1d47
+        uses: actions/upload-artifact@v4
         with:
           name: artifacts-plan-dist-manifest
           path: plan-dist-manifest.json
@@ -119,24 +116,23 @@ jobs:
       - plan
       - custom-build-binaries
       - custom-build-docker
-    runs-on: "depot-ubuntu-latest-4"
+    runs-on: "ubuntu-20.04"
     env:
       GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
       BUILD_MANIFEST_NAME: target/distrib/global-dist-manifest.json
     steps:
-      - uses: actions/checkout@85e6279cec87321a52edac9c87bce653a07cf6c2
+      - uses: actions/checkout@v4
         with:
-          persist-credentials: false
           submodules: recursive
-      - name: Install cached dist
-        uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093
+      - name: Install cached cargo-dist
+        uses: actions/download-artifact@v4
         with:
           name: cargo-dist-cache
           path: ~/.cargo/bin/
-      - run: chmod +x ~/.cargo/bin/dist
+      - run: chmod +x ~/.cargo/bin/cargo-dist
       # Get all the local artifacts for the global tasks to use (for e.g. checksums)
       - name: Fetch local artifacts
-        uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093
+        uses: actions/download-artifact@v4
         with:
           pattern: artifacts-*
           path: target/distrib/
@@ -144,8 +140,8 @@ jobs:
       - id: cargo-dist
         shell: bash
         run: |
-          dist build ${{ needs.plan.outputs.tag-flag }} --output-format=json "--artifacts=global" > dist-manifest.json
-          echo "dist ran successfully"
+          cargo dist build ${{ needs.plan.outputs.tag-flag }} --output-format=json "--artifacts=global" > dist-manifest.json
+          echo "cargo dist ran successfully"
 
           # Parse out what we just built and upload it to scratch storage
           echo "paths<<EOF" >> "$GITHUB_OUTPUT"
@@ -154,7 +150,7 @@ jobs:
 
           cp dist-manifest.json "$BUILD_MANIFEST_NAME"
       - name: "Upload artifacts"
-        uses: actions/upload-artifact@6027e3dd177782cd8ab9af838c04fd81a07f1d47
+        uses: actions/upload-artifact@v4
         with:
           name: artifacts-build-global
           path: |
@@ -171,23 +167,22 @@ jobs:
     if: ${{ always() && needs.plan.outputs.publishing == 'true' && (needs.build-global-artifacts.result == 'skipped' || needs.build-global-artifacts.result == 'success') && (needs.custom-build-binaries.result == 'skipped' || needs.custom-build-binaries.result == 'success') && (needs.custom-build-docker.result == 'skipped' || needs.custom-build-docker.result == 'success') }}
     env:
       GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-    runs-on: "depot-ubuntu-latest-4"
+    runs-on: "ubuntu-20.04"
     outputs:
       val: ${{ steps.host.outputs.manifest }}
     steps:
-      - uses: actions/checkout@85e6279cec87321a52edac9c87bce653a07cf6c2
+      - uses: actions/checkout@v4
         with:
-          persist-credentials: false
           submodules: recursive
-      - name: Install cached dist
-        uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093
+      - name: Install cached cargo-dist
+        uses: actions/download-artifact@v4
         with:
           name: cargo-dist-cache
           path: ~/.cargo/bin/
-      - run: chmod +x ~/.cargo/bin/dist
+      - run: chmod +x ~/.cargo/bin/cargo-dist
       # Fetch artifacts from scratch-storage
       - name: Fetch artifacts
-        uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093
+        uses: actions/download-artifact@v4
         with:
           pattern: artifacts-*
           path: target/distrib/
@@ -196,12 +191,12 @@ jobs:
       - id: host
         shell: bash
         run: |
-          dist host ${{ needs.plan.outputs.tag-flag }} --steps=upload --steps=release --output-format=json > dist-manifest.json
+          cargo dist host ${{ needs.plan.outputs.tag-flag }} --steps=upload --steps=release --output-format=json > dist-manifest.json
           echo "artifacts uploaded and released successfully"
           cat dist-manifest.json
           echo "manifest=$(jq -c "." dist-manifest.json)" >> "$GITHUB_OUTPUT"
       - name: "Upload dist-manifest.json"
-        uses: actions/upload-artifact@6027e3dd177782cd8ab9af838c04fd81a07f1d47
+        uses: actions/upload-artifact@v4
         with:
           # Overwrite the previous copy
           name: artifacts-dist-manifest
@@ -247,17 +242,16 @@ jobs:
     # still allowing individual publish jobs to skip themselves (for prereleases).
     # "host" however must run to completion, no skipping allowed!
     if: ${{ always() && needs.host.result == 'success' && (needs.custom-publish-pypi.result == 'skipped' || needs.custom-publish-pypi.result == 'success') && (needs.custom-publish-wasm.result == 'skipped' || needs.custom-publish-wasm.result == 'success') }}
-    runs-on: "depot-ubuntu-latest-4"
+    runs-on: "ubuntu-20.04"
     env:
       GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
     steps:
-      - uses: actions/checkout@85e6279cec87321a52edac9c87bce653a07cf6c2
+      - uses: actions/checkout@v4
         with:
-          persist-credentials: false
           submodules: recursive
       # Create a GitHub Release while uploading all files to it
       - name: "Download GitHub Artifacts"
-        uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093
+        uses: actions/download-artifact@v4
         with:
           pattern: artifacts-*
           path: artifacts

.github/workflows/sync_typeshed.yaml

@@ -21,17 +21,15 @@ jobs:
       contents: write
       pull-requests: write
     steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
+      - uses: actions/checkout@v4
         name: Checkout Ruff
         with:
           path: ruff
-          persist-credentials: true
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
+      - uses: actions/checkout@v4
         name: Checkout typeshed
         with:
           repository: python/typeshed
           path: typeshed
-          persist-credentials: false
       - name: Setup git
         run: |
           git config --global user.name typeshedbot
@@ -39,13 +37,13 @@ jobs:
       - name: Sync typeshed
         id: sync
         run: |
-          rm -rf ruff/crates/ty_vendored/vendor/typeshed
-          mkdir ruff/crates/ty_vendored/vendor/typeshed
-          cp typeshed/README.md ruff/crates/ty_vendored/vendor/typeshed
-          cp typeshed/LICENSE ruff/crates/ty_vendored/vendor/typeshed
-          cp -r typeshed/stdlib ruff/crates/ty_vendored/vendor/typeshed/stdlib
-          rm -rf ruff/crates/ty_vendored/vendor/typeshed/stdlib/@tests
-          git -C typeshed rev-parse HEAD > ruff/crates/ty_vendored/vendor/typeshed/source_commit.txt
+          rm -rf ruff/crates/red_knot_vendored/vendor/typeshed
+          mkdir ruff/crates/red_knot_vendored/vendor/typeshed
+          cp typeshed/README.md ruff/crates/red_knot_vendored/vendor/typeshed
+          cp typeshed/LICENSE ruff/crates/red_knot_vendored/vendor/typeshed
+          cp -r typeshed/stdlib ruff/crates/red_knot_vendored/vendor/typeshed/stdlib
+          rm -rf ruff/crates/red_knot_vendored/vendor/typeshed/stdlib/@tests
+          git -C typeshed rev-parse HEAD > ruff/crates/red_knot_vendored/vendor/typeshed/source_commit.txt
       - name: Commit the changes
         id: commit
         if: ${{ steps.sync.outcome == 'success' }}
@@ -59,7 +57,7 @@ jobs:
         run: |
           cd ruff
           git push --force origin typeshedbot/sync-typeshed
-          gh pr list --repo "$GITHUB_REPOSITORY" --head typeshedbot/sync-typeshed --json id --jq length | grep 1 && exit 0 # exit if there is existing pr
+          gh pr list --repo $GITHUB_REPOSITORY --head typeshedbot/sync-typeshed --json id --jq length | grep 1 && exit 0 # exit if there is existing pr
           gh pr create --title "Sync vendored typeshed stubs" --body "Close and reopen this PR to trigger CI" --label "internal"
 
   create-issue-on-failure:
@@ -70,7 +68,7 @@ jobs:
     permissions:
       issues: write
     steps:
-      - uses: actions/github-script@60a0d83039c74a4aee543508d2ffcb1c3799cdea # v7.0.1
+      - uses: actions/github-script@v7
         with:
           github-token: ${{ secrets.GITHUB_TOKEN }}
           script: |
@@ -78,6 +76,5 @@ jobs:
               owner: "astral-sh",
               repo: "ruff",
               title: `Automated typeshed sync failed on ${new Date().toDateString()}`,
-              body: "Run listed here: https://github.com/${{ github.repository }}/actions/runs/${{ github.run_id }}",
-              labels: ["bug", "ty"],
+              body: "Runs are listed here: https://github.com/astral-sh/ruff/actions/workflows/sync_typeshed.yaml",
             })

.github/zizmor.yml

@@ -1,19 +0,0 @@
-# Configuration for the zizmor static analysis tool, run via pre-commit in CI
-# https://woodruffw.github.io/zizmor/configuration/
-#
-# TODO: can we remove the ignores here so that our workflows are more secure?
-rules:
-  dangerous-triggers:
-    ignore:
-      - pr-comment.yaml
-  cache-poisoning:
-    ignore:
-      - build-docker.yml
-      - publish-playground.yml
-  excessive-permissions:
-    # it's hard to test what the impact of removing these ignores would be
-    # without actually running the release workflow...
-    ignore:
-      - build-docker.yml
-      - publish-playground.yml
-      - publish-docs.yml

.gitignore

@@ -29,10 +29,6 @@ tracing.folded
 tracing-flamechart.svg
 tracing-flamegraph.svg
 
-# insta
-*.rs.pending-snap
-
-
 ###
 # Rust.gitignore
 ###

.ignore

@@ -1 +0,0 @@
-!/.github/

.markdownlint.yaml

@@ -21,15 +21,3 @@ MD014: false
 MD024:
   # Allow when nested under different parents e.g. CHANGELOG.md
   siblings_only: true
-
-# MD046/code-block-style
-#
-# Ignore this because it conflicts with the code block style used in content
-# tabs of mkdocs-material which is to add a blank line after the content title.
-#
-# Ref: https://github.com/astral-sh/ruff/pull/15011#issuecomment-2544790854
-MD046: false
-
-# Link text should be descriptive
-# Disallows link text like *here* which is annoying.
-MD059: false

.pre-commit-config.yaml

@@ -1,12 +1,9 @@
-fail_fast: false
+fail_fast: true
 
 exclude: |
   (?x)^(
-    .github/workflows/release.yml|
-    crates/ty_vendored/vendor/.*|
-    crates/ty_project/resources/.*|
-    crates/ty/docs/(configuration|rules|cli).md|
-    crates/ruff_benchmark/resources/.*|
+    crates/red_knot_vendored/vendor/.*|
+    crates/red_knot_workspace/resources/.*|
     crates/ruff_linter/resources/.*|
     crates/ruff_linter/src/rules/.*/snapshots/.*|
     crates/ruff_notebook/resources/.*|
@@ -19,23 +16,19 @@ exclude: |
   )$
 
 repos:
-  - repo: https://github.com/pre-commit/pre-commit-hooks
-    rev: v5.0.0
-    hooks:
-      - id: check-merge-conflict
-
   - repo: https://github.com/abravalheri/validate-pyproject
-    rev: v0.24.1
+    rev: v0.20.2
     hooks:
       - id: validate-pyproject
 
   - repo: https://github.com/executablebooks/mdformat
-    rev: 0.7.22
+    rev: 0.7.17
     hooks:
       - id: mdformat
         additional_dependencies:
-          - mdformat-mkdocs==4.0.0
-          - mdformat-footnote==0.1.1
+          - mdformat-mkdocs
+          - mdformat-admon
+          - mdformat-footnote
         exclude: |
           (?x)^(
             docs/formatter/black\.md
@@ -43,7 +36,7 @@ repos:
           )$
 
   - repo: https://github.com/igorshubovych/markdownlint-cli
-    rev: v0.45.0
+    rev: v0.42.0
     hooks:
       - id: markdownlint-fix
         exclude: |
@@ -52,21 +45,8 @@ repos:
             | docs/\w+\.md
           )$
 
-  - repo: https://github.com/adamchainz/blacken-docs
-    rev: 1.19.1
-    hooks:
-      - id: blacken-docs
-        args: ["--pyi", "--line-length", "130"]
-        files: '^crates/.*/resources/mdtest/.*\.md'
-        exclude: |
-          (?x)^(
-            .*?invalid(_.+)*_syntax\.md
-          )$
-        additional_dependencies:
-          - black==25.1.0
-
   - repo: https://github.com/crate-ci/typos
-    rev: v1.32.0
+    rev: v1.25.0
     hooks:
       - id: typos
 
@@ -80,7 +60,7 @@ repos:
         pass_filenames: false # This makes it a lot faster
 
   - repo: https://github.com/astral-sh/ruff-pre-commit
-    rev: v0.11.10
+    rev: v0.6.9
     hooks:
       - id: ruff-format
       - id: ruff
@@ -90,42 +70,10 @@ repos:
 
   # Prettier
   - repo: https://github.com/rbubley/mirrors-prettier
-    rev: v3.5.3
+    rev: v3.3.3
     hooks:
       - id: prettier
         types: [yaml]
 
-  # zizmor detects security vulnerabilities in GitHub Actions workflows.
-  # Additional configuration for the tool is found in `.github/zizmor.yml`
-  - repo: https://github.com/woodruffw/zizmor-pre-commit
-    rev: v1.7.0
-    hooks:
-      - id: zizmor
-
-  - repo: https://github.com/python-jsonschema/check-jsonschema
-    rev: 0.33.0
-    hooks:
-      - id: check-github-workflows
-
-  # `actionlint` hook, for verifying correct syntax in GitHub Actions workflows.
-  # Some additional configuration for `actionlint` can be found in `.github/actionlint.yaml`.
-  - repo: https://github.com/rhysd/actionlint
-    rev: v1.7.7
-    hooks:
-      - id: actionlint
-        stages:
-          # This hook is disabled by default, since it's quite slow.
-          # To run all hooks *including* this hook, use `uvx pre-commit run -a --hook-stage=manual`.
-          # To run *just* this hook, use `uvx pre-commit run -a actionlint --hook-stage=manual`.
-          - manual
-        args:
-          - "-ignore=SC2129" # ignorable stylistic lint from shellcheck
-          - "-ignore=SC2016" # another shellcheck lint: seems to have false positives?
-        additional_dependencies:
-          # actionlint has a shellcheck integration which extracts shell scripts in `run:` steps from GitHub Actions
-          # and checks these with shellcheck. This is arguably its most useful feature,
-          # but the integration only works if shellcheck is installed
-          - "github.com/wasilibs/go-shellcheck/cmd/shellcheck@v0.10.0"
-
 ci:
   skip: [cargo-fmt, dev-generate-all]

BREAKING_CHANGES.md

@@ -1,97 +1,5 @@
 # Breaking Changes
 
-## 0.11.0
-
-This is a follow-up to release 0.10.0. Because of a mistake in the release process, the `requires-python` inference changes were not included in that release. Ruff 0.11.0 now includes this change as well as the stabilization of the preview behavior for `PGH004`.
-
-- **Changes to how the Python version is inferred when a `target-version` is not specified** ([#16319](https://github.com/astral-sh/ruff/pull/16319))
-
-    In previous versions of Ruff, you could specify your Python version with:
-
-    - The `target-version` option in a `ruff.toml` file or the `[tool.ruff]` section of a pyproject.toml file.
-    - The `project.requires-python` field in a `pyproject.toml` file with a `[tool.ruff]` section.
-
-    These options worked well in most cases, and are still recommended for fine control of the Python version. However, because of the way Ruff discovers config files, `pyproject.toml` files without a `[tool.ruff]` section would be ignored, including the `requires-python` setting. Ruff would then use the default Python version (3.9 as of this writing) instead, which is surprising when you've attempted to request another version.
-
-    In v0.10, config discovery has been updated to address this issue:
-
-    - If Ruff finds a `ruff.toml` file without a `target-version`, it will check
-        for a `pyproject.toml` file in the same directory and respect its
-        `requires-python` version, even if it does not contain a `[tool.ruff]`
-        section.
-    - If Ruff finds a user-level configuration, the `requires-python` field of the closest `pyproject.toml` in a parent directory will take precedence.
-    - If there is no config file (`ruff.toml`or `pyproject.toml` with a
-        `[tool.ruff]` section) in the directory of the file being checked, Ruff will
-        search for the closest `pyproject.toml` in the parent directories and use its
-        `requires-python` setting.
-
-## 0.10.0
-
-- **Changes to how the Python version is inferred when a `target-version` is not specified** ([#16319](https://github.com/astral-sh/ruff/pull/16319))
-
-    Because of a mistake in the release process, the `requires-python` inference changes are not included in this release and instead shipped as part of 0.11.0.
-    You can find a description of this change in the 0.11.0 section.
-
-- **Updated `TYPE_CHECKING` behavior** ([#16669](https://github.com/astral-sh/ruff/pull/16669))
-
-    Previously, Ruff only recognized typechecking blocks that tested the `typing.TYPE_CHECKING` symbol. Now, Ruff recognizes any local variable named `TYPE_CHECKING`. This release also removes support for the legacy `if 0:` and `if False:` typechecking checks. Use a local `TYPE_CHECKING` variable instead.
-
-- **More robust noqa parsing** ([#16483](https://github.com/astral-sh/ruff/pull/16483))
-
-    The syntax for both file-level and in-line suppression comments has been unified and made more robust to certain errors. In most cases, this will result in more suppression comments being read by Ruff, but there are a few instances where previously read comments will now log an error to the user instead. Please refer to the documentation on [_Error suppression_](https://docs.astral.sh/ruff/linter/#error-suppression) for the full specification.
-
-- **Avoid unnecessary parentheses around with statements with a single context manager and a trailing comment** ([#14005](https://github.com/astral-sh/ruff/pull/14005))
-
-    This change fixes a bug in the formatter where it introduced unnecessary parentheses around with statements with a single context manager and a trailing comment. This change may result in a change in formatting for some users.
-
-- **Bump alpine default tag to 3.21 for derived Docker images** ([#16456](https://github.com/astral-sh/ruff/pull/16456))
-
-    Alpine 3.21 was released in Dec 2024 and is used in the official Alpine-based Python images. Now the ruff:alpine image will use 3.21 instead of 3.20 and ruff:alpine3.20 will no longer be updated.
-
-- **\[`unsafe-markup-use`\]: `RUF035` has been recoded to `S704`** ([#15957](https://github.com/astral-sh/ruff/pull/15957))
-
-## 0.9.0
-
-Ruff now formats your code according to the 2025 style guide. As a result, your code might now get formatted differently. See the [changelog](./CHANGELOG.md#090) for a detailed list of changes.
-
-## 0.8.0
-
-- **Default to Python 3.9**
-
-    Ruff now defaults to Python 3.9 instead of 3.8 if no explicit Python version is configured using [`ruff.target-version`](https://docs.astral.sh/ruff/settings/#target-version) or [`project.requires-python`](https://packaging.python.org/en/latest/guides/writing-pyproject-toml/#python-requires) ([#13896](https://github.com/astral-sh/ruff/pull/13896))
-
-- **Changed location of `pydoclint` diagnostics**
-
-    [`pydoclint`](https://docs.astral.sh/ruff/rules/#pydoclint-doc) diagnostics now point to the first-line of the problematic docstring. Previously, this was not the case.
-
-    If you've opted into these preview rules but have them suppressed using
-    [`noqa`](https://docs.astral.sh/ruff/linter/#error-suppression) comments in
-    some places, this change may mean that you need to move the `noqa` suppression
-    comments. Most users should be unaffected by this change.
-
-- **Use XDG (i.e. `~/.local/bin`) instead of the Cargo home directory in the standalone installer**
-
-    Previously, Ruff's installer used `$CARGO_HOME` or `~/.cargo/bin` for its target install directory. Now, Ruff will be installed into `$XDG_BIN_HOME`, `$XDG_DATA_HOME/../bin`, or `~/.local/bin` (in that order).
-
-    This change is only relevant to users of the standalone Ruff installer (using the shell or PowerShell script). If you installed Ruff using uv or pip, you should be unaffected.
-
-- **Changes to the line width calculation**
-
-    Ruff now uses a new version of the [unicode-width](https://github.com/unicode-rs/unicode-width) Rust crate to calculate the line width. In very rare cases, this may lead to lines containing Unicode characters being reformatted, or being considered too long when they were not before ([`E501`](https://docs.astral.sh/ruff/rules/line-too-long/)).
-
-## 0.7.0
-
-- The pytest rules `PT001` and `PT023` now default to omitting the decorator parentheses when there are no arguments
-    ([#12838](https://github.com/astral-sh/ruff/pull/12838), [#13292](https://github.com/astral-sh/ruff/pull/13292)).
-    This was a change that we attempted to make in Ruff v0.6.0, but only partially made due to an error on our part.
-    See the [blog post](https://astral.sh/blog/ruff-v0.7.0) for more details.
-- The `useless-try-except` rule (in our `tryceratops` category) has been recoded from `TRY302` to
-    `TRY203` ([#13502](https://github.com/astral-sh/ruff/pull/13502)). This ensures Ruff's code is consistent with
-    the same rule in the [`tryceratops`](https://github.com/guilatrova/tryceratops) linter.
-- The `lint.allow-unused-imports` setting has been removed ([#13677](https://github.com/astral-sh/ruff/pull/13677)). Use
-    [`lint.pyflakes.allow-unused-imports`](https://docs.astral.sh/ruff/settings/#lint_pyflakes_allowed-unused-imports)
-    instead.
-
 ## 0.6.0
 
 - Detect imports in `src` layouts by default for `isort` rules ([#12848](https://github.com/astral-sh/ruff/pull/12848))
@@ -246,7 +154,7 @@ flag or `unsafe-fixes` configuration option can be used to enable unsafe fixes.
 
 See the [docs](https://docs.astral.sh/ruff/configuration/#fix-safety) for details.
 
-### Remove formatter-conflicting rules from the default rule set ([#7900](https://github.com/astral-sh/ruff/pull/7900))
+### Remove formatter-conflicting rules from the default rule set  ([#7900](https://github.com/astral-sh/ruff/pull/7900))
 
 Previously, Ruff enabled all implemented rules in Pycodestyle (`E`) by default. Ruff now only includes the
 Pycodestyle prefixes `E4`, `E7`, and `E9` to exclude rules that conflict with automatic formatters. Consequently,
@@ -259,8 +167,8 @@ This change only affects those using Ruff under its default rule set. Users that
 
 ### Remove support for emoji identifiers ([#7212](https://github.com/astral-sh/ruff/pull/7212))
 
-Previously, Ruff supported non-standards-compliant emoji identifiers such as `📦 = 1`.
-We decided to remove this non-standard language extension. Ruff now reports syntax errors for invalid emoji identifiers in your code, the same as CPython.
+Previously, Ruff supported the non-standard compliant emoji identifiers e.g. `📦 = 1`.
+We decided to remove this non-standard language extension, and Ruff now reports syntax errors for emoji identifiers in your code, the same as CPython.
 
 ### Improved GitLab fingerprints ([#7203](https://github.com/astral-sh/ruff/pull/7203))
 

CODE_OF_CONDUCT.md

@@ -71,7 +71,8 @@ representative at an online or offline event.
 ## Enforcement
 
 Instances of abusive, harassing, or otherwise unacceptable behavior may be
-reported to the community leaders responsible for enforcement at <hey@astral.sh>.
+reported to the community leaders responsible for enforcement at
+<charlie.r.marsh@gmail.com>.
 All complaints will be reviewed and investigated promptly and fairly.
 
 All community leaders are obligated to respect the privacy and security of the

CONTRIBUTING.md

@@ -2,11 +2,6 @@
 
 Welcome! We're happy to have you here. Thank you in advance for your contribution to Ruff.
 
-> [!NOTE]
->
-> This guide is for Ruff. If you're looking to contribute to ty, please see [the ty contributing
-> guide](https://github.com/astral-sh/ruff/blob/main/crates/ty/CONTRIBUTING.md).
-
 ## The Basics
 
 Ruff welcomes contributions in the form of pull requests.
@@ -144,7 +139,7 @@ At a high level, the steps involved in adding a new lint rule are as follows:
 1. Create a file for your rule (e.g., `crates/ruff_linter/src/rules/flake8_bugbear/rules/assert_false.rs`).
 
 1. In that file, define a violation struct (e.g., `pub struct AssertFalse`). You can grep for
-    `#[derive(ViolationMetadata)]` to see examples.
+    `#[violation]` to see examples.
 
 1. In that file, define a function that adds the violation to the diagnostic list as appropriate
     (e.g., `pub(crate) fn assert_false`) based on whatever inputs are required for the rule (e.g.,
@@ -371,15 +366,6 @@ uvx --from ./python/ruff-ecosystem ruff-ecosystem format ruff "./target/debug/ru
 
 See the [ruff-ecosystem package](https://github.com/astral-sh/ruff/tree/main/python/ruff-ecosystem) for more details.
 
-## Upgrading Rust
-
-1. Change the `channel` in `./rust-toolchain.toml` to the new Rust version (`<latest>`)
-1. Change the `rust-version` in the `./Cargo.toml` to `<latest> - 2` (e.g. 1.84 if the latest is 1.86)
-1. Run `cargo clippy --fix --allow-dirty --allow-staged` to fix new clippy warnings
-1. Create and merge the PR
-1. Bump the Rust version in Ruff's conda forge recipe. See [this PR](https://github.com/conda-forge/ruff-feedstock/pull/266) for an example.
-1. Enjoy the new Rust version!
-
 ## Benchmarking and Profiling
 
 We have several ways of benchmarking and profiling Ruff:
@@ -411,7 +397,7 @@ cargo install hyperfine
 To benchmark the release build:
 
 ```shell
-cargo build --release --bin ruff && hyperfine --warmup 10 \
+cargo build --release && hyperfine --warmup 10 \
   "./target/release/ruff check ./crates/ruff_linter/resources/test/cpython/ --no-cache -e" \
   "./target/release/ruff check ./crates/ruff_linter/resources/test/cpython/ -e"
 
@@ -481,7 +467,7 @@ cargo build --release && hyperfine --warmup 10 \
   "./target/release/ruff check ./crates/ruff_linter/resources/test/cpython/ --no-cache -e --select W505,E501"
 ```
 
-You can run `uv venv --project ./scripts/benchmarks`, activate the venv and then run `uv sync --project ./scripts/benchmarks` to create a working environment for the
+You can run `poetry install` from `./scripts/benchmarks` to create a working environment for the
 above. All reported benchmarks were computed using the versions specified by
 `./scripts/benchmarks/pyproject.toml` on Python 3.11.
 
@@ -540,7 +526,7 @@ cargo benchmark
 #### Benchmark-driven Development
 
 Ruff uses [Criterion.rs](https://bheisler.github.io/criterion.rs/book/) for benchmarks. You can use
-`--save-baseline=<name>` to store an initial baseline benchmark (e.g., on `main`) and then use
+`--save-baseline=<name>` to store an initial baseline benchmark (e.g. on `main`) and then use
 `--benchmark=<name>` to compare against that benchmark. Criterion will print a message telling you
 if the benchmark improved/regressed compared to that baseline.
 
@@ -610,7 +596,8 @@ Then convert the recorded profile
 perf script -F +pid > /tmp/test.perf
 ```
 
-You can now view the converted file with [firefox profiler](https://profiler.firefox.com/). To learn more about Firefox profiler, read the [Firefox profiler profiling-guide](https://profiler.firefox.com/docs/#/./guide-perf-profiling).
+You can now view the converted file with [firefox profiler](https://profiler.firefox.com/), with a
+more in-depth guide [here](https://profiler.firefox.com/docs/#/./guide-perf-profiling)
 
 An alternative is to convert the perf data to `flamegraph.svg` using
 [flamegraph](https://github.com/flamegraph-rs/flamegraph) (`cargo install flamegraph`):
@@ -691,9 +678,9 @@ utils with it:
 23 Newline 24
 ```
 
-- `cargo dev print-cst <file>`: Print the CST of a Python file using
+- `cargo dev print-cst <file>`: Print the CST of a python file using
     [LibCST](https://github.com/Instagram/LibCST), which is used in addition to the RustPython parser
-    in Ruff. For example, for `if True: pass # comment`, everything, including the whitespace, is represented:
+    in Ruff. E.g. for `if True: pass # comment` everything including the whitespace is represented:
 
 ```text
 Module {
@@ -876,7 +863,7 @@ each configuration file.
 
 The package root is used to determine a file's "module path". Consider, again, `baz.py`. In that
 case, `./my_project/src/foo` was identified as the package root, so the module path for `baz.py`
-would resolve to `foo.bar.baz` — as computed by taking the relative path from the package root
+would resolve to  `foo.bar.baz` — as computed by taking the relative path from the package root
 (inclusive of the root itself). The module path can be thought of as "the path you would use to
 import the module" (e.g., `import foo.bar.baz`).
 

Cargo.toml

@@ -3,9 +3,8 @@ members = ["crates/*"]
 resolver = "2"
 
 [workspace.package]
-# Please update rustfmt.toml when bumping the Rust edition
-edition = "2024"
-rust-version = "1.85"
+edition = "2021"
+rust-version = "1.76"
 homepage = "https://docs.astral.sh/ruff"
 documentation = "https://docs.astral.sh/ruff"
 repository = "https://github.com/astral-sh/ruff"
@@ -14,7 +13,6 @@ license = "MIT"
 
 [workspace.dependencies]
 ruff = { path = "crates/ruff" }
-ruff_annotate_snippets = { path = "crates/ruff_annotate_snippets" }
 ruff_cache = { path = "crates/ruff_cache" }
 ruff_db = { path = "crates/ruff_db", default-features = false }
 ruff_diagnostics = { path = "crates/ruff_diagnostics" }
@@ -24,7 +22,6 @@ ruff_index = { path = "crates/ruff_index" }
 ruff_linter = { path = "crates/ruff_linter" }
 ruff_macros = { path = "crates/ruff_macros" }
 ruff_notebook = { path = "crates/ruff_notebook" }
-ruff_options_metadata = { path = "crates/ruff_options_metadata" }
 ruff_python_ast = { path = "crates/ruff_python_ast" }
 ruff_python_codegen = { path = "crates/ruff_python_codegen" }
 ruff_python_formatter = { path = "crates/ruff_python_formatter" }
@@ -37,70 +34,61 @@ ruff_python_trivia = { path = "crates/ruff_python_trivia" }
 ruff_server = { path = "crates/ruff_server" }
 ruff_source_file = { path = "crates/ruff_source_file" }
 ruff_text_size = { path = "crates/ruff_text_size" }
+red_knot_vendored = { path = "crates/red_knot_vendored" }
 ruff_workspace = { path = "crates/ruff_workspace" }
 
-ty = { path = "crates/ty" }
-ty_ide = { path = "crates/ty_ide" }
-ty_project = { path = "crates/ty_project", default-features = false }
-ty_python_semantic = { path = "crates/ty_python_semantic" }
-ty_server = { path = "crates/ty_server" }
-ty_test = { path = "crates/ty_test" }
-ty_vendored = { path = "crates/ty_vendored" }
+red_knot_python_semantic = { path = "crates/red_knot_python_semantic" }
+red_knot_server = { path = "crates/red_knot_server" }
+red_knot_test = { path = "crates/red_knot_test" }
+red_knot_workspace = { path = "crates/red_knot_workspace", default-features = false }
 
 aho-corasick = { version = "1.1.3" }
-anstream = { version = "0.6.18" }
-anstyle = { version = "1.0.10" }
+annotate-snippets = { version = "0.9.2", features = ["color"] }
 anyhow = { version = "1.0.80" }
 assert_fs = { version = "1.1.0" }
 argfile = { version = "0.2.0" }
-bincode = { version = "2.0.0" }
+bincode = { version = "1.3.3" }
 bitflags = { version = "2.5.0" }
 bstr = { version = "1.9.1" }
 cachedir = { version = "0.3.1" }
 camino = { version = "1.1.7" }
+chrono = { version = "0.4.35", default-features = false, features = ["clock"] }
 clap = { version = "4.5.3", features = ["derive"] }
 clap_complete_command = { version = "0.6.0" }
-clearscreen = { version = "4.0.0" }
+clearscreen = { version = "3.0.0" }
 codspeed-criterion-compat = { version = "2.6.0", default-features = false }
-colored = { version = "3.0.0" }
+colored = { version = "2.1.0" }
 console_error_panic_hook = { version = "0.1.7" }
 console_log = { version = "1.0.0" }
 countme = { version = "3.0.1" }
-compact_str = "0.9.0"
-criterion = { version = "0.6.0", default-features = false }
+compact_str = "0.8.0"
+criterion = { version = "0.5.1", default-features = false }
 crossbeam = { version = "0.8.4" }
 dashmap = { version = "6.0.1" }
-dir-test = { version = "0.4.0" }
-dunce = { version = "1.0.5" }
 drop_bomb = { version = "0.1.5" }
 env_logger = { version = "0.11.0" }
-etcetera = { version = "0.10.0" }
-fern = { version = "0.7.0" }
+etcetera = { version = "0.8.0" }
+fern = { version = "0.6.1" }
 filetime = { version = "0.2.23" }
-getrandom = { version = "0.3.1" }
 glob = { version = "0.3.1" }
 globset = { version = "0.4.14" }
 globwalk = { version = "0.9.1" }
 hashbrown = { version = "0.15.0", default-features = false, features = [
     "raw-entry",
-    "equivalent",
     "inline-more",
 ] }
-heck = "0.5.0"
 ignore = { version = "0.4.22" }
 imara-diff = { version = "0.1.5" }
 imperative = { version = "1.0.4" }
-indexmap = { version = "2.6.0" }
 indicatif = { version = "0.17.8" }
 indoc = { version = "2.0.4" }
 insta = { version = "1.35.1" }
 insta-cmd = { version = "0.6.0" }
 is-macro = { version = "0.3.5" }
 is-wsl = { version = "0.4.0" }
-itertools = { version = "0.14.0" }
-jiff = { version = "0.2.0" }
+itertools = { version = "0.13.0" }
 js-sys = { version = "0.3.69" }
-jod-thread = { version = "1.0.0" }
+jod-thread = { version = "0.1.2" }
 libc = { version = "0.2.153" }
 libcst = { version = "1.1.0", default-features = false }
 log = { version = "0.4.17" }
@@ -112,24 +100,24 @@ matchit = { version = "0.8.1" }
 memchr = { version = "2.7.1" }
 mimalloc = { version = "0.1.39" }
 natord = { version = "1.0.9" }
-notify = { version = "8.0.0" }
+notify = { version = "6.1.1" }
+once_cell = { version = "1.19.0" }
 ordermap = { version = "0.5.0" }
 path-absolutize = { version = "3.1.1" }
 path-slash = { version = "0.2.1" }
 pathdiff = { version = "0.2.1" }
-pep440_rs = { version = "0.7.1" }
+pep440_rs = { version = "0.6.0", features = ["serde"] }
 pretty_assertions = "1.3.0"
 proc-macro2 = { version = "1.0.79" }
-pyproject-toml = { version = "0.13.4" }
+pyproject-toml = { version = "0.9.0" }
 quick-junit = { version = "0.5.0" }
 quote = { version = "1.0.23" }
-rand = { version = "0.9.0" }
+rand = { version = "0.8.5" }
 rayon = { version = "1.10.0" }
 regex = { version = "1.10.2" }
+rstest = { version = "0.22.0", default-features = false }
 rustc-hash = { version = "2.0.0" }
-rustc-stable-hash = { version = "0.1.2" }
-# When updating salsa, make sure to also update the revision in `fuzz/Cargo.toml`
-salsa = { git = "https://github.com/salsa-rs/salsa.git", rev = "4818b15f3b7516555d39f5a41cb75970448bee4c" }
+salsa = { git = "https://github.com/salsa-rs/salsa.git", rev = "4a7c955255e707e64e43f3ce5eabb771ae067768" }
 schemars = { version = "0.8.16" }
 seahash = { version = "4.1.0" }
 serde = { version = "1.0.197", features = ["derive"] }
@@ -142,38 +130,30 @@ serde_with = { version = "3.6.0", default-features = false, features = [
 shellexpand = { version = "3.0.0" }
 similar = { version = "2.4.0", features = ["inline"] }
 smallvec = { version = "1.13.2" }
-snapbox = { version = "0.6.0", features = [
-    "diff",
-    "term-svg",
-    "cmd",
-    "examples",
-] }
 static_assertions = "1.1.0"
-strum = { version = "0.27.0", features = ["strum_macros"] }
-strum_macros = { version = "0.27.0" }
+strum = { version = "0.26.0", features = ["strum_macros"] }
+strum_macros = { version = "0.26.0" }
 syn = { version = "2.0.55" }
 tempfile = { version = "3.9.0" }
 test-case = { version = "3.3.1" }
-thiserror = { version = "2.0.0" }
+thiserror = { version = "1.0.58" }
 tikv-jemallocator = { version = "0.6.0" }
 toml = { version = "0.8.11" }
 tracing = { version = "0.1.40" }
 tracing-flame = { version = "0.2.0" }
 tracing-indicatif = { version = "0.3.6" }
-tracing-log = { version = "0.2.0" }
 tracing-subscriber = { version = "0.3.18", default-features = false, features = [
     "env-filter",
     "fmt",
-    "ansi",
-    "smallvec"
 ] }
-tryfn = { version = "0.2.1" }
+tracing-tree = { version = "0.4.0" }
 typed-arena = { version = "2.0.2" }
 unic-ucd-category = { version = "0.9" }
 unicode-ident = { version = "1.0.12" }
-unicode-width = { version = "0.2.0" }
+unicode-width = { version = "0.1.11" }
 unicode_names2 = { version = "1.2.2" }
 unicode-normalization = { version = "0.1.23" }
+ureq = { version = "2.9.6" }
 url = { version = "2.5.0" }
 uuid = { version = "1.6.1", features = [
     "v4",
@@ -187,10 +167,6 @@ wasm-bindgen-test = { version = "0.3.42" }
 wild = { version = "2" }
 zip = { version = "0.6.6", default-features = false }
 
-[workspace.metadata.cargo-shear]
-ignored = ["getrandom", "ruff_options_metadata"]
-
-
 [workspace.lints.rust]
 unsafe_code = "warn"
 unreachable_pub = "warn"
@@ -213,11 +189,8 @@ missing_panics_doc = "allow"
 module_name_repetitions = "allow"
 must_use_candidate = "allow"
 similar_names = "allow"
-single_match_else = "allow"
 too_many_lines = "allow"
-needless_continue = "allow"       # An explicit continue can be more readable, especially if the alternative is an empty block.
-unnecessary_debug_formatting = "allow"  # too many instances, the display also doesn't quote the path which is often desired in logs where we use them the most often.
-# Without the hashes we run into a `rustfmt` bug in some snapshot tests, see #13250
+# To allow `#[allow(clippy::all)]` in `crates/ruff_python_parser/src/python.rs`.
 needless_raw_string_hashes = "allow"
 # Disallowed restriction lints
 print_stdout = "warn"
@@ -230,13 +203,6 @@ get_unwrap = "warn"
 rc_buffer = "warn"
 rc_mutex = "warn"
 rest_pat_in_fully_bound_structs = "warn"
-# nursery rules
-redundant_clone = "warn"
-debug_assert_with_mut_call = "warn"
-unused_peekable = "warn"
-
-# Diagnostics are not actionable: Enable once https://github.com/rust-lang/rust-clippy/issues/13774 is resolved.
-large_stack_arrays = "allow"
 
 [profile.release]
 # Note that we set these explicitly, and these values
@@ -261,9 +227,6 @@ opt-level = 3
 [profile.dev.package.similar]
 opt-level = 3
 
-[profile.dev.package.salsa]
-opt-level = 3
-
 # Reduce complexity of a parser function that would trigger a locals limit in a wasm tool.
 # https://github.com/bytecodealliance/wasm-tools/blob/b5c3d98e40590512a3b12470ef358d5c7b983b15/crates/wasmparser/src/limits.rs#L29
 [profile.dev.package.ruff_python_parser]
@@ -278,3 +241,64 @@ debug = 1
 # The profile that 'cargo dist' will build with.
 [profile.dist]
 inherits = "release"
+
+# Config for 'cargo dist'
+[workspace.metadata.dist]
+# The preferred cargo-dist version to use in CI (Cargo.toml SemVer syntax)
+cargo-dist-version = "0.22.1"
+# CI backends to support
+ci = "github"
+# The installers to generate for each app
+installers = ["shell", "powershell"]
+# The archive format to use for windows builds (defaults .zip)
+windows-archive = ".zip"
+# The archive format to use for non-windows builds (defaults .tar.xz)
+unix-archive = ".tar.gz"
+# Target platforms to build apps for (Rust target-triple syntax)
+targets = [
+    "aarch64-apple-darwin",
+    "aarch64-pc-windows-msvc",
+    "aarch64-unknown-linux-gnu",
+    "aarch64-unknown-linux-musl",
+    "arm-unknown-linux-musleabihf",
+    "armv7-unknown-linux-gnueabihf",
+    "armv7-unknown-linux-musleabihf",
+    "i686-pc-windows-msvc",
+    "i686-unknown-linux-gnu",
+    "i686-unknown-linux-musl",
+    "powerpc64-unknown-linux-gnu",
+    "powerpc64le-unknown-linux-gnu",
+    "s390x-unknown-linux-gnu",
+    "x86_64-apple-darwin",
+    "x86_64-pc-windows-msvc",
+    "x86_64-unknown-linux-gnu",
+    "x86_64-unknown-linux-musl",
+]
+# Whether to auto-include files like READMEs, LICENSEs, and CHANGELOGs (default true)
+auto-includes = false
+# Whether cargo-dist should create a GitHub Release or use an existing draft
+create-release = true
+# Which actions to run on pull requests
+pr-run-mode = "skip"
+# Whether CI should trigger releases with dispatches instead of tag pushes
+dispatch-releases = true
+# Which phase cargo-dist should use to create the GitHub release
+github-release = "announce"
+# Whether CI should include auto-generated code to build local artifacts
+build-local-artifacts = false
+# Local artifacts jobs to run in CI
+local-artifacts-jobs = ["./build-binaries", "./build-docker"]
+# Publish jobs to run in CI
+publish-jobs = ["./publish-pypi", "./publish-wasm"]
+# Post-announce jobs to run in CI
+post-announce-jobs = [
+    "./notify-dependents",
+    "./publish-docs",
+    "./publish-playground",
+]
+# Custom permissions for GitHub Jobs
+github-custom-job-permissions = { "build-docker" = { packages = "write", contents = "read" }, "publish-wasm" = { contents = "read", id-token = "write", packages = "write" } }
+# Whether to install an updater program
+install-updater = false
+# Path that installers should place binaries in
+install-path = "CARGO_HOME"

Dockerfile

@@ -1,4 +1,4 @@
-FROM --platform=$BUILDPLATFORM ubuntu AS build
+FROM --platform=$BUILDPLATFORM ubuntu as build
 ENV HOME="/root"
 WORKDIR $HOME
 

README.md

@@ -116,22 +116,9 @@ For more, see the [documentation](https://docs.astral.sh/ruff/).
 
 ### Installation
 
-Ruff is available as [`ruff`](https://pypi.org/project/ruff/) on PyPI.
-
-Invoke Ruff directly with [`uvx`](https://docs.astral.sh/uv/):
+Ruff is available as [`ruff`](https://pypi.org/project/ruff/) on PyPI:
 
 ```shell
-uvx ruff check   # Lint all files in the current directory.
-uvx ruff format  # Format all files in the current directory.
-```
-
-Or install Ruff with `uv` (recommended), `pip`, or `pipx`:
-
-```shell
-# With uv.
-uv tool install ruff@latest  # Install Ruff globally.
-uv add --dev ruff            # Or add Ruff to your project.
-
 # With pip.
 pip install ruff
 
@@ -149,8 +136,8 @@ curl -LsSf https://astral.sh/ruff/install.sh | sh
 powershell -c "irm https://astral.sh/ruff/install.ps1 | iex"
 
 # For a specific version.
-curl -LsSf https://astral.sh/ruff/0.11.11/install.sh | sh
-powershell -c "irm https://astral.sh/ruff/0.11.11/install.ps1 | iex"
+curl -LsSf https://astral.sh/ruff/0.6.9/install.sh | sh
+powershell -c "irm https://astral.sh/ruff/0.6.9/install.ps1 | iex"
 ```
 
 You can also install Ruff via [Homebrew](https://formulae.brew.sh/formula/ruff), [Conda](https://anaconda.org/conda-forge/ruff),
@@ -183,7 +170,7 @@ Ruff can also be used as a [pre-commit](https://pre-commit.com/) hook via [`ruff
 ```yaml
 - repo: https://github.com/astral-sh/ruff-pre-commit
   # Ruff version.
-  rev: v0.11.11
+  rev: v0.6.9
   hooks:
     # Run the linter.
     - id: ruff
@@ -205,7 +192,7 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - uses: actions/checkout@v4
-      - uses: astral-sh/ruff-action@v3
+      - uses: astral-sh/ruff-action@v1
 ```
 
 ### Configuration<a id="configuration"></a>
@@ -251,11 +238,11 @@ exclude = [
 line-length = 88
 indent-width = 4
 
-# Assume Python 3.9
-target-version = "py39"
+# Assume Python 3.8
+target-version = "py38"
 
 [lint]
-# Enable Pyflakes (`F`) and a subset of the pycodestyle (`E`) codes by default.
+# Enable Pyflakes (`F`) and a subset of the pycodestyle (`E`)  codes by default.
 select = ["E4", "E7", "E9", "F"]
 ignore = []
 
@@ -430,7 +417,6 @@ Ruff is used by a number of major open-source projects and companies, including:
 - [Babel](https://github.com/python-babel/babel)
 - Benchling ([Refac](https://github.com/benchling/refac))
 - [Bokeh](https://github.com/bokeh/bokeh)
-- CrowdCent ([NumerBlox](https://github.com/crowdcent/numerblox)) <!-- typos: ignore -->
 - [Cryptography (PyCA)](https://github.com/pyca/cryptography)
 - CERN ([Indico](https://getindico.io/))
 - [DVC](https://github.com/iterative/dvc)
@@ -452,7 +438,6 @@ Ruff is used by a number of major open-source projects and companies, including:
 - ING Bank ([popmon](https://github.com/ing-bank/popmon), [probatus](https://github.com/ing-bank/probatus))
 - [Ibis](https://github.com/ibis-project/ibis)
 - [ivy](https://github.com/unifyai/ivy)
-- [JAX](https://github.com/jax-ml/jax)
 - [Jupyter](https://github.com/jupyter-server/jupyter_server)
 - [Kraken Tech](https://kraken.tech/)
 - [LangChain](https://github.com/hwchase17/langchain)

SECURITY.md

@@ -1,15 +0,0 @@
-# Security policy
-
-## Reporting a vulnerability
-
-If you have found a possible vulnerability, please email `security at astral dot sh`.
-
-## Bug bounties
-
-While we sincerely appreciate and encourage reports of suspected security problems, please note that
-Astral does not currently run any bug bounty programs.
-
-## Vulnerability disclosures
-
-Critical vulnerabilities will be disclosed via GitHub's
-[security advisory](https://github.com/astral-sh/ruff/security) system.

_typos.toml

@@ -1,10 +1,6 @@
 [files]
 # https://github.com/crate-ci/typos/issues/868
-extend-exclude = [
-    "crates/ty_vendored/vendor/**/*",
-    "**/resources/**/*",
-    "**/snapshots/**/*",
-]
+extend-exclude = ["crates/red_knot_vendored/vendor/**/*", "**/resources/**/*", "**/snapshots/**/*"]
 
 [default.extend-words]
 "arange" = "arange"  # e.g. `numpy.arange`
@@ -16,18 +12,10 @@ pn = "pn"  # `import panel as pn` is a thing
 poit = "poit"
 BA = "BA" # acronym for "Bad Allowed", used in testing.
 jod = "jod" # e.g., `jod-thread`
-Numer = "Numer" # Library name 'NumerBlox' in "Who's Using Ruff?"
 
 [default]
 extend-ignore-re = [
-    # Line ignore with trailing "spellchecker:disable-line"
-    "(?Rm)^.*#\\s*spellchecker:disable-line$",
-    "LICENSEs",
-    # Various third party dependencies uses `typ` as struct field names (e.g., lsp_types::LogMessageParams)
-    "typ",
-    # TODO: Remove this once the `TYP` redirects are removed from `rule_redirects.rs`
-    "TYP",
+  # Line ignore with trailing "spellchecker:disable-line"
+  "(?Rm)^.*#\\s*spellchecker:disable-line$",
+  "LICENSEs",
 ]
-
-[default.extend-identifiers]
-"FrIeNdLy" = "FrIeNdLy"

clippy.toml

@@ -1,26 +1,21 @@
 doc-valid-idents = [
-    "..",
-    "CodeQL",
-    "CPython",
-    "FastAPI",
-    "IPython",
-    "LangChain",
-    "LibCST",
-    "McCabe",
-    "NumPy",
-    "SCREAMING_SNAKE_CASE",
-    "SQLAlchemy",
-    "StackOverflow",
-    "PyCharm",
-    "SNMPv1",
-    "SNMPv2",
-    "SNMPv3",
-    "PyFlakes",
+  "..",
+  "CodeQL",
+  "FastAPI",
+  "IPython",
+  "LangChain",
+  "LibCST",
+  "McCabe",
+  "NumPy",
+  "SCREAMING_SNAKE_CASE",
+  "SQLAlchemy",
+  "StackOverflow",
+  "PyCharm",
 ]
 
 ignore-interior-mutability = [
-    # Interned is read-only. The wrapped `Rc` never gets updated.
-    "ruff_formatter::format_element::Interned",
-    # The expression is read-only.
-    "ruff_python_ast::hashable::HashableExpr",
+  # Interned is read-only. The wrapped `Rc` never gets updated.
+  "ruff_formatter::format_element::Interned",
+  # The expression is read-only.
+  "ruff_python_ast::hashable::HashableExpr",
 ]

crates/red_knot/Cargo.toml

@@ -0,0 +1,39 @@
+[package]
+name = "red_knot"
+version = "0.0.0"
+edition.workspace = true
+rust-version.workspace = true
+homepage.workspace = true
+documentation.workspace = true
+repository.workspace = true
+authors.workspace = true
+license.workspace = true
+
+# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
+
+[dependencies]
+red_knot_python_semantic = { workspace = true }
+red_knot_workspace = { workspace = true, features = ["zstd"] }
+red_knot_server = { workspace = true }
+ruff_db = { workspace = true, features = ["os", "cache"] }
+
+anyhow = { workspace = true }
+chrono = { workspace = true }
+clap = { workspace = true, features = ["wrap_help"] }
+colored = { workspace = true }
+countme = { workspace = true, features = ["enable"] }
+crossbeam = { workspace = true }
+ctrlc = { version = "3.4.4" }
+rayon = { workspace = true }
+salsa = { workspace = true }
+tracing = { workspace = true, features = ["release_max_level_debug"] }
+tracing-subscriber = { workspace = true, features = ["env-filter", "fmt"] }
+tracing-flame = { workspace = true }
+tracing-tree = { workspace = true }
+
+[dev-dependencies]
+filetime = { workspace = true }
+tempfile = { workspace = true }
+
+[lints]
+workspace = true

crates/red_knot/docs/tracing.md

@@ -1,7 +1,6 @@
 # Tracing
 
-Traces are a useful tool to narrow down the location of a bug or, at least, to understand why the compiler is doing a
-particular thing.
+Traces are a useful tool to narrow down the location of a bug or, at least, to understand why the compiler is doing a particular thing.
 Note, tracing messages with severity `debug` or greater are user-facing. They should be phrased accordingly.
 Tracing spans are only shown when using `-vvv`.
 
@@ -10,28 +9,20 @@ Tracing spans are only shown when using `-vvv`.
 The CLI supports different verbosity levels.
 
 - default: Only show errors and warnings.
-- `-v` activates `info!`: Show generally useful information such as paths of configuration files, detected platform,
-    etc., but it's not a lot of messages, it's something you'll activate in CI by default. cargo build e.g. shows you
-    which packages are fresh.
-- `-vv` activates `debug!` and timestamps: This should be enough information to get to the bottom of bug reports. When
-    you're processing many packages or files, you'll get pages and pages of output, but each line is link to a specific
-    action or state change.
-- `-vvv` activates `trace!` (only in debug builds) and shows tracing-spans: At this level, you're logging everything.
-    Most of this is wasted, it's really slow, we dump e.g. the entire resolution graph. Only useful to developers, and you
-    almost certainly want to use `TY_LOG` to filter it down to the area your investigating.
+- `-v` activates `info!`: Show generally useful information such as paths of configuration files, detected platform, etc., but it's not a lot of messages, it's something you'll activate in CI by default. cargo build e.g. shows you which packages are fresh.
+- `-vv` activates `debug!` and timestamps: This should be enough information to get to the bottom of bug reports. When you're processing many packages or files, you'll get pages and pages of output, but each line is link to a specific action or state change.
+- `-vvv` activates `trace!` (only in debug builds) and shows tracing-spans: At this level, you're logging everything. Most of this is wasted, it's really slow, we dump e.g. the entire resolution graph. Only useful to developers, and you almost certainly want to use `RED_KNOT_LOG` to filter it down to the area your investigating.
 
-## Better logging with `TY_LOG` and `TY_MAX_PARALLELISM`
+## Better logging with `RED_KNOT_LOG` and `RAYON_NUM_THREADS`
 
-By default, the CLI shows messages from the `ruff` and `ty` crates. Tracing messages from other crates are not shown.
-The `TY_LOG` environment variable allows you to customize which messages are shown by specifying one
-or
-more [filter directives](https://docs.rs/tracing-subscriber/latest/tracing_subscriber/filter/struct.EnvFilter.html#directives).
+By default, the CLI shows messages from the `ruff` and `red_knot` crates. Tracing messages from other crates are not shown.
+The `RED_KNOT_LOG` environment variable allows you to customize which messages are shown by specifying one
+or more [filter directives](https://docs.rs/tracing-subscriber/latest/tracing_subscriber/filter/struct.EnvFilter.html#directives).
 
-The `TY_MAX_PARALLELISM` environment variable, meanwhile, can be used to control the level of parallelism ty uses.
-By default, ty will attempt to parallelize its work so that multiple files are checked simultaneously,
-but this can result in a confused logging output where messages from different threads are intertwined and non
-determinism.
-To switch off parallelism entirely and have more readable logs, use `TY_MAX_PARALLELISM=1` (or `RAYON_NUM_THREADS=1`).
+The `RAYON_NUM_THREADS` environment variable, meanwhile, can be used to control the level of concurrency red-knot uses.
+By default, red-knot will attempt to parallelize its work so that multiple files are checked simultaneously,
+but this can result in a confused logging output where messages from different threads are intertwined.
+To switch off concurrency entirely and have more readable logs, use `RAYON_NUM_THREADS=1`.
 
 ### Examples
 
@@ -40,23 +31,23 @@ To switch off parallelism entirely and have more readable logs, use `TY_MAX_PARA
 Shows debug messages from all crates.
 
 ```bash
-TY_LOG=debug
+RED_KNOT_LOG=debug
 ```
 
 #### Show salsa query execution messages
 
-Show the salsa `execute: my_query` messages in addition to all ty messages.
+Show the salsa `execute: my_query` messages in addition to all red knot messages.
 
 ```bash
-TY_LOG=ruff=trace,ty=trace,salsa=info
+RED_KNOT_LOG=ruff=trace,red_knot=trace,salsa=info
 ```
 
 #### Show typing traces
 
-Only show traces for the `ty_python_semantic::types` module.
+Only show traces for the `red_knot_python_semantic::types` module.
 
 ```bash
-TY_LOG="ty_python_semantic::types"
+RED_KNOT_LOG="red_knot_python_semantic::types"
 ```
 
 Note: Ensure that you use `-vvv` to see tracing spans.
@@ -66,7 +57,7 @@ Note: Ensure that you use `-vvv` to see tracing spans.
 Shows all messages that are inside of a span for a specific file.
 
 ```bash
-TY_LOG=ty[{file=/home/micha/astral/test/x.py}]=trace
+RED_KNOT_LOG=red_knot[{file=/home/micha/astral/test/x.py}]=trace
 ```
 
 **Note**: Tracing still shows all spans because tracing can't know at the time of entering the span
@@ -88,24 +79,22 @@ query to return the failure as part of the query's result or use a Salsa accumul
 
 ## Tracing in tests
 
-You can use `ruff_db::testing::setup_logging` or `ruff_db::testing::setup_logging_with_filter` to set up logging in
-tests.
+You can use `ruff_db::testing::setup_logging` or `ruff_db::testing::setup_logging_with_filter` to set up logging in tests.
 
 ```rust
 use ruff_db::testing::setup_logging;
 
 #[test]
 fn test() {
-    let _logging = setup_logging();
+  let _logging = setup_logging();
 
-    tracing::info!("This message will be printed to stderr");
+  tracing::info!("This message will be printed to stderr");
 }
 ```
 
 Note: Most test runners capture stderr and only show its output when a test fails.
 
-Note also that `setup_logging` only sets up logging for the current thread because
-[`set_global_default`](https://docs.rs/tracing/latest/tracing/subscriber/fn.set_global_default.html) can only be
+Note also that `setup_logging` only sets up logging for the current thread because [`set_global_default`](https://docs.rs/tracing/latest/tracing/subscriber/fn.set_global_default.html) can only be
 called **once**.
 
 ## Release builds
@@ -114,11 +103,10 @@ called **once**.
 
 ## Profiling
 
-ty generates a folded stack trace to the current directory named `tracing.folded` when setting the environment variable
-`TY_LOG_PROFILE` to `1` or `true`.
+Red Knot generates a folded stack trace  to the current directory named `tracing.folded` when setting the environment variable `RED_KNOT_LOG_PROFILE` to `1` or `true`.
 
 ```bash
-TY_LOG_PROFILE=1 ty -- --current-directory=../test -vvv
+RED_KNOT_LOG_PROFILE=1 red_knot -- --current-directory=../test -vvv
 ```
 
 You can convert the textual representation into a visual one using `inferno`.

crates/red_knot/src/logging.rs

@@ -1,17 +1,16 @@
-//! Sets up logging for ty
+//! Sets up logging for Red Knot
 
-use crate::args::TerminalColor;
 use anyhow::Context;
 use colored::Colorize;
 use std::fmt;
 use std::fs::File;
-use std::io::{BufWriter, IsTerminal};
+use std::io::BufWriter;
 use tracing::{Event, Subscriber};
-use tracing_subscriber::EnvFilter;
 use tracing_subscriber::filter::LevelFilter;
 use tracing_subscriber::fmt::format::Writer;
 use tracing_subscriber::fmt::{FmtContext, FormatEvent, FormatFields};
 use tracing_subscriber::registry::LookupSpan;
+use tracing_subscriber::EnvFilter;
 
 /// Logging flags to `#[command(flatten)]` into your CLI
 #[derive(clap::Args, Debug, Clone, Default)]
@@ -43,14 +42,14 @@ impl Verbosity {
 
 #[derive(Debug, Copy, Clone, Eq, PartialEq, Ord, PartialOrd)]
 pub(crate) enum VerbosityLevel {
-    /// Default output level. Only shows Ruff and ty events up to the [`WARN`](tracing::Level::WARN).
+    /// Default output level. Only shows Ruff and Red Knot events up to the [`WARN`](tracing::Level::WARN).
     Default,
 
-    /// Enables verbose output. Emits Ruff and ty events up to the [`INFO`](tracing::Level::INFO).
+    /// Enables verbose output. Emits Ruff and Red Knot events up to the [`INFO`](tracing::Level::INFO).
     /// Corresponds to `-v`.
     Verbose,
 
-    /// Enables a more verbose tracing format and emits Ruff and ty events up to [`DEBUG`](tracing::Level::DEBUG).
+    /// Enables a more verbose tracing format and emits Ruff and Red Knot events up to [`DEBUG`](tracing::Level::DEBUG).
     /// Corresponds to `-vv`
     ExtraVerbose,
 
@@ -77,17 +76,14 @@ impl VerbosityLevel {
     }
 }
 
-pub(crate) fn setup_tracing(
-    level: VerbosityLevel,
-    color: TerminalColor,
-) -> anyhow::Result<TracingGuard> {
+pub(crate) fn setup_tracing(level: VerbosityLevel) -> anyhow::Result<TracingGuard> {
     use tracing_subscriber::prelude::*;
 
-    // The `TY_LOG` environment variable overrides the default log level.
-    let filter = if let Ok(log_env_variable) = std::env::var("TY_LOG") {
+    // The `RED_KNOT_LOG` environment variable overrides the default log level.
+    let filter = if let Ok(log_env_variable) = std::env::var("RED_KNOT_LOG") {
         EnvFilter::builder()
             .parse(log_env_variable)
-            .context("Failed to parse directives specified in TY_LOG environment variable.")?
+            .context("Failed to parse directives specified in RED_KNOT_LOG environment variable.")?
     } else {
         match level {
             VerbosityLevel::Default => {
@@ -97,9 +93,9 @@ pub(crate) fn setup_tracing(
             level => {
                 let level_filter = level.level_filter();
 
-                // Show info|debug|trace events, but allow `TY_LOG` to override
+                // Show info|debug|trace events, but allow `RED_KNOT_LOG` to override
                 let filter = EnvFilter::default().add_directive(
-                    format!("ty={level_filter}")
+                    format!("red_knot={level_filter}")
                         .parse()
                         .expect("Hardcoded directive to be valid"),
                 );
@@ -119,33 +115,27 @@ pub(crate) fn setup_tracing(
         .with(filter)
         .with(profiling_layer);
 
-    let ansi = match color {
-        TerminalColor::Auto => {
-            colored::control::SHOULD_COLORIZE.should_colorize() && std::io::stderr().is_terminal()
-        }
-        TerminalColor::Always => true,
-        TerminalColor::Never => false,
-    };
-
     if level.is_trace() {
         let subscriber = registry.with(
-            tracing_subscriber::fmt::layer()
-                .event_format(tracing_subscriber::fmt::format().pretty())
+            tracing_tree::HierarchicalLayer::default()
+                .with_indent_lines(true)
+                .with_indent_amount(2)
+                .with_bracketed_fields(true)
                 .with_thread_ids(true)
-                .with_ansi(ansi)
-                .with_writer(std::io::stderr),
+                .with_targets(true)
+                .with_writer(std::io::stderr)
+                .with_timer(tracing_tree::time::Uptime::default()),
         );
 
         subscriber.init();
     } else {
         let subscriber = registry.with(
             tracing_subscriber::fmt::layer()
-                .event_format(TyFormat {
+                .event_format(RedKnotFormat {
                     display_level: true,
                     display_timestamp: level.is_extra_verbose(),
                     show_spans: false,
                 })
-                .with_ansi(ansi)
                 .with_writer(std::io::stderr),
         );
 
@@ -157,7 +147,7 @@ pub(crate) fn setup_tracing(
     })
 }
 
-#[expect(clippy::type_complexity)]
+#[allow(clippy::type_complexity)]
 fn setup_profile<S>() -> (
     Option<tracing_flame::FlameLayer<S, BufWriter<File>>>,
     Option<tracing_flame::FlushGuard<BufWriter<File>>>,
@@ -165,7 +155,7 @@ fn setup_profile<S>() -> (
 where
     S: Subscriber + for<'span> LookupSpan<'span>,
 {
-    if let Ok("1" | "true") = std::env::var("TY_LOG_PROFILE").as_deref() {
+    if let Ok("1" | "true") = std::env::var("RED_KNOT_LOG_PROFILE").as_deref() {
         let (layer, guard) = tracing_flame::FlameLayer::with_file("tracing.folded")
             .expect("Flame layer to be created");
         (Some(layer), Some(guard))
@@ -178,14 +168,14 @@ pub(crate) struct TracingGuard {
     _flame_guard: Option<tracing_flame::FlushGuard<BufWriter<File>>>,
 }
 
-struct TyFormat {
+struct RedKnotFormat {
     display_timestamp: bool,
     display_level: bool,
     show_spans: bool,
 }
 
 /// See <https://docs.rs/tracing-subscriber/0.3.18/src/tracing_subscriber/fmt/format/mod.rs.html#1026-1156>
-impl<S, N> FormatEvent<S, N> for TyFormat
+impl<S, N> FormatEvent<S, N> for RedKnotFormat
 where
     S: Subscriber + for<'a> LookupSpan<'a>,
     N: for<'a> FormatFields<'a> + 'static,
@@ -200,8 +190,8 @@ where
         let ansi = writer.has_ansi_escapes();
 
         if self.display_timestamp {
-            let timestamp = jiff::Zoned::now()
-                .strftime("%Y-%m-%d %H:%M:%S.%f")
+            let timestamp = chrono::Local::now()
+                .format("%Y-%m-%d %H:%M:%S.%f")
                 .to_string();
             if ansi {
                 write!(writer, "{} ", timestamp.dimmed())?;
@@ -209,7 +199,7 @@ where
                 write!(
                     writer,
                     "{} ",
-                    jiff::Zoned::now().strftime("%Y-%m-%d %H:%M:%S.%f")
+                    chrono::Local::now().format("%Y-%m-%d %H:%M:%S.%f")
                 )?;
             }
         }

crates/red_knot/src/main.rs

@@ -0,0 +1,384 @@
+use std::process::{ExitCode, Termination};
+use std::sync::Mutex;
+
+use anyhow::{anyhow, Context};
+use clap::Parser;
+use colored::Colorize;
+use crossbeam::channel as crossbeam_channel;
+use salsa::plumbing::ZalsaDatabase;
+
+use red_knot_python_semantic::SitePackages;
+use red_knot_server::run_server;
+use red_knot_workspace::db::RootDatabase;
+use red_knot_workspace::watch;
+use red_knot_workspace::watch::WorkspaceWatcher;
+use red_knot_workspace::workspace::settings::Configuration;
+use red_knot_workspace::workspace::WorkspaceMetadata;
+use ruff_db::system::{OsSystem, System, SystemPath, SystemPathBuf};
+use target_version::TargetVersion;
+
+use crate::logging::{setup_tracing, Verbosity};
+
+mod logging;
+mod target_version;
+mod verbosity;
+
+#[derive(Debug, Parser)]
+#[command(
+    author,
+    name = "red-knot",
+    about = "An extremely fast Python type checker."
+)]
+#[command(version)]
+struct Args {
+    #[command(subcommand)]
+    pub(crate) command: Option<Command>,
+
+    #[arg(
+        long,
+        help = "Changes the current working directory.",
+        long_help = "Changes the current working directory before any specified operations. This affects the workspace and configuration discovery.",
+        value_name = "PATH"
+    )]
+    current_directory: Option<SystemPathBuf>,
+
+    #[arg(
+        long,
+        help = "Path to the virtual environment the project uses",
+        long_help = "\
+Path to the virtual environment the project uses. \
+If provided, red-knot will use the `site-packages` directory of this virtual environment \
+to resolve type information for the project's third-party dependencies.",
+        value_name = "PATH"
+    )]
+    venv_path: Option<SystemPathBuf>,
+
+    #[arg(
+        long,
+        value_name = "DIRECTORY",
+        help = "Custom directory to use for stdlib typeshed stubs"
+    )]
+    custom_typeshed_dir: Option<SystemPathBuf>,
+
+    #[arg(
+        long,
+        value_name = "PATH",
+        help = "Additional path to use as a module-resolution source (can be passed multiple times)"
+    )]
+    extra_search_path: Option<Vec<SystemPathBuf>>,
+
+    #[arg(
+        long,
+        help = "Python version to assume when resolving types",
+        value_name = "VERSION"
+    )]
+    target_version: Option<TargetVersion>,
+
+    #[clap(flatten)]
+    verbosity: Verbosity,
+
+    #[arg(
+        long,
+        help = "Run in watch mode by re-running whenever files change",
+        short = 'W'
+    )]
+    watch: bool,
+}
+
+impl Args {
+    fn to_configuration(&self, cli_cwd: &SystemPath) -> Configuration {
+        let mut configuration = Configuration::default();
+
+        if let Some(target_version) = self.target_version {
+            configuration.target_version = Some(target_version.into());
+        }
+
+        if let Some(venv_path) = &self.venv_path {
+            configuration.search_paths.site_packages = Some(SitePackages::Derived {
+                venv_path: SystemPath::absolute(venv_path, cli_cwd),
+            });
+        }
+
+        if let Some(custom_typeshed_dir) = &self.custom_typeshed_dir {
+            configuration.search_paths.custom_typeshed =
+                Some(SystemPath::absolute(custom_typeshed_dir, cli_cwd));
+        }
+
+        if let Some(extra_search_paths) = &self.extra_search_path {
+            configuration.search_paths.extra_paths = extra_search_paths
+                .iter()
+                .map(|path| Some(SystemPath::absolute(path, cli_cwd)))
+                .collect();
+        }
+
+        configuration
+    }
+}
+
+#[derive(Debug, clap::Subcommand)]
+pub enum Command {
+    /// Start the language server
+    Server,
+}
+
+#[allow(clippy::print_stdout, clippy::unnecessary_wraps, clippy::print_stderr)]
+pub fn main() -> ExitStatus {
+    run().unwrap_or_else(|error| {
+        use std::io::Write;
+
+        // Use `writeln` instead of `eprintln` to avoid panicking when the stderr pipe is broken.
+        let mut stderr = std::io::stderr().lock();
+
+        // This communicates that this isn't a linter error but Red Knot itself hard-errored for
+        // some reason (e.g. failed to resolve the configuration)
+        writeln!(stderr, "{}", "Red Knot failed".red().bold()).ok();
+        // Currently we generally only see one error, but e.g. with io errors when resolving
+        // the configuration it is help to chain errors ("resolving configuration failed" ->
+        // "failed to read file: subdir/pyproject.toml")
+        for cause in error.chain() {
+            writeln!(stderr, "  {} {cause}", "Cause:".bold()).ok();
+        }
+
+        ExitStatus::Error
+    })
+}
+
+fn run() -> anyhow::Result<ExitStatus> {
+    let args = Args::parse_from(std::env::args().collect::<Vec<_>>());
+
+    if matches!(args.command, Some(Command::Server)) {
+        return run_server().map(|()| ExitStatus::Success);
+    }
+
+    let verbosity = args.verbosity.level();
+    countme::enable(verbosity.is_trace());
+    let _guard = setup_tracing(verbosity)?;
+
+    // The base path to which all CLI arguments are relative to.
+    let cli_base_path = {
+        let cwd = std::env::current_dir().context("Failed to get the current working directory")?;
+        SystemPathBuf::from_path_buf(cwd)
+            .map_err(|path| {
+                anyhow!(
+                    "The current working directory `{}` contains non-Unicode characters. Red Knot only supports Unicode paths.",
+                    path.display()
+                )
+            })?
+    };
+
+    let cwd = args
+        .current_directory
+        .as_ref()
+        .map(|cwd| {
+            if cwd.as_std_path().is_dir() {
+                Ok(SystemPath::absolute(cwd, &cli_base_path))
+            } else {
+                Err(anyhow!(
+                    "Provided current-directory path `{cwd}` is not a directory"
+                ))
+            }
+        })
+        .transpose()?
+        .unwrap_or_else(|| cli_base_path.clone());
+
+    let system = OsSystem::new(cwd.clone());
+    let cli_configuration = args.to_configuration(&cwd);
+    let workspace_metadata = WorkspaceMetadata::from_path(
+        system.current_directory(),
+        &system,
+        Some(cli_configuration.clone()),
+    )?;
+
+    // TODO: Use the `program_settings` to compute the key for the database's persistent
+    //   cache and load the cache if it exists.
+    let mut db = RootDatabase::new(workspace_metadata, system)?;
+
+    let (main_loop, main_loop_cancellation_token) = MainLoop::new(cli_configuration);
+
+    // Listen to Ctrl+C and abort the watch mode.
+    let main_loop_cancellation_token = Mutex::new(Some(main_loop_cancellation_token));
+    ctrlc::set_handler(move || {
+        let mut lock = main_loop_cancellation_token.lock().unwrap();
+
+        if let Some(token) = lock.take() {
+            token.stop();
+        }
+    })?;
+
+    let exit_status = if args.watch {
+        main_loop.watch(&mut db)?
+    } else {
+        main_loop.run(&mut db)
+    };
+
+    tracing::trace!("Counts for entire CLI run:\n{}", countme::get_all());
+
+    std::mem::forget(db);
+
+    Ok(exit_status)
+}
+
+#[derive(Copy, Clone)]
+pub enum ExitStatus {
+    /// Checking was successful and there were no errors.
+    Success = 0,
+
+    /// Checking was successful but there were errors.
+    Failure = 1,
+
+    /// Checking failed.
+    Error = 2,
+}
+
+impl Termination for ExitStatus {
+    fn report(self) -> ExitCode {
+        ExitCode::from(self as u8)
+    }
+}
+
+struct MainLoop {
+    /// Sender that can be used to send messages to the main loop.
+    sender: crossbeam_channel::Sender<MainLoopMessage>,
+
+    /// Receiver for the messages sent **to** the main loop.
+    receiver: crossbeam_channel::Receiver<MainLoopMessage>,
+
+    /// The file system watcher, if running in watch mode.
+    watcher: Option<WorkspaceWatcher>,
+
+    cli_configuration: Configuration,
+}
+
+impl MainLoop {
+    fn new(cli_configuration: Configuration) -> (Self, MainLoopCancellationToken) {
+        let (sender, receiver) = crossbeam_channel::bounded(10);
+
+        (
+            Self {
+                sender: sender.clone(),
+                receiver,
+                watcher: None,
+                cli_configuration,
+            },
+            MainLoopCancellationToken { sender },
+        )
+    }
+
+    fn watch(mut self, db: &mut RootDatabase) -> anyhow::Result<ExitStatus> {
+        tracing::debug!("Starting watch mode");
+        let sender = self.sender.clone();
+        let watcher = watch::directory_watcher(move |event| {
+            sender.send(MainLoopMessage::ApplyChanges(event)).unwrap();
+        })?;
+
+        self.watcher = Some(WorkspaceWatcher::new(watcher, db));
+
+        self.run(db);
+
+        Ok(ExitStatus::Success)
+    }
+
+    fn run(mut self, db: &mut RootDatabase) -> ExitStatus {
+        self.sender.send(MainLoopMessage::CheckWorkspace).unwrap();
+
+        let result = self.main_loop(db);
+
+        tracing::debug!("Exiting main loop");
+
+        result
+    }
+
+    fn main_loop(&mut self, db: &mut RootDatabase) -> ExitStatus {
+        // Schedule the first check.
+        tracing::debug!("Starting main loop");
+
+        let mut revision = 0u64;
+
+        while let Ok(message) = self.receiver.recv() {
+            match message {
+                MainLoopMessage::CheckWorkspace => {
+                    let db = db.snapshot();
+                    let sender = self.sender.clone();
+
+                    // Spawn a new task that checks the workspace. This needs to be done in a separate thread
+                    // to prevent blocking the main loop here.
+                    rayon::spawn(move || {
+                        if let Ok(result) = db.check() {
+                            // Send the result back to the main loop for printing.
+                            sender
+                                .send(MainLoopMessage::CheckCompleted { result, revision })
+                                .unwrap();
+                        }
+                    });
+                }
+
+                MainLoopMessage::CheckCompleted {
+                    result,
+                    revision: check_revision,
+                } => {
+                    let has_diagnostics = !result.is_empty();
+                    if check_revision == revision {
+                        for diagnostic in result {
+                            tracing::error!("{}", diagnostic);
+                        }
+                    } else {
+                        tracing::debug!(
+                            "Discarding check result for outdated revision: current: {revision}, result revision: {check_revision}"
+                        );
+                    }
+
+                    if self.watcher.is_none() {
+                        return if has_diagnostics {
+                            ExitStatus::Failure
+                        } else {
+                            ExitStatus::Success
+                        };
+                    }
+
+                    tracing::trace!("Counts after last check:\n{}", countme::get_all());
+                }
+
+                MainLoopMessage::ApplyChanges(changes) => {
+                    revision += 1;
+                    // Automatically cancels any pending queries and waits for them to complete.
+                    db.apply_changes(changes, Some(&self.cli_configuration));
+                    if let Some(watcher) = self.watcher.as_mut() {
+                        watcher.update(db);
+                    }
+                    self.sender.send(MainLoopMessage::CheckWorkspace).unwrap();
+                }
+                MainLoopMessage::Exit => {
+                    // Cancel any pending queries and wait for them to complete.
+                    // TODO: Don't use Salsa internal APIs
+                    //  [Zulip-Thread](https://salsa.zulipchat.com/#narrow/stream/333573-salsa-3.2E0/topic/Expose.20an.20API.20to.20cancel.20other.20queries)
+                    let _ = db.zalsa_mut();
+                    return ExitStatus::Success;
+                }
+            }
+
+            tracing::debug!("Waiting for next main loop message.");
+        }
+
+        ExitStatus::Success
+    }
+}
+
+#[derive(Debug)]
+struct MainLoopCancellationToken {
+    sender: crossbeam_channel::Sender<MainLoopMessage>,
+}
+
+impl MainLoopCancellationToken {
+    fn stop(self) {
+        self.sender.send(MainLoopMessage::Exit).unwrap();
+    }
+}
+
+/// Message sent from the orchestrator to the main loop.
+#[derive(Debug)]
+enum MainLoopMessage {
+    CheckWorkspace,
+    CheckCompleted { result: Vec<String>, revision: u64 },
+    ApplyChanges(Vec<watch::ChangeEvent>),
+    Exit,
+}

crates/red_knot/src/target_version.rs

@@ -0,0 +1,48 @@
+/// Enumeration of all supported Python versions
+///
+/// TODO: unify with the `PythonVersion` enum in the linter/formatter crates?
+#[derive(Copy, Clone, Hash, Debug, PartialEq, Eq, PartialOrd, Ord, Default, clap::ValueEnum)]
+pub enum TargetVersion {
+    Py37,
+    #[default]
+    Py38,
+    Py39,
+    Py310,
+    Py311,
+    Py312,
+    Py313,
+}
+
+impl TargetVersion {
+    const fn as_str(self) -> &'static str {
+        match self {
+            Self::Py37 => "py37",
+            Self::Py38 => "py38",
+            Self::Py39 => "py39",
+            Self::Py310 => "py310",
+            Self::Py311 => "py311",
+            Self::Py312 => "py312",
+            Self::Py313 => "py313",
+        }
+    }
+}
+
+impl std::fmt::Display for TargetVersion {
+    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        f.write_str(self.as_str())
+    }
+}
+
+impl From<TargetVersion> for red_knot_python_semantic::PythonVersion {
+    fn from(value: TargetVersion) -> Self {
+        match value {
+            TargetVersion::Py37 => Self::PY37,
+            TargetVersion::Py38 => Self::PY38,
+            TargetVersion::Py39 => Self::PY39,
+            TargetVersion::Py310 => Self::PY310,
+            TargetVersion::Py311 => Self::PY311,
+            TargetVersion::Py312 => Self::PY312,
+            TargetVersion::Py313 => Self::PY313,
+        }
+    }
+}

crates/red_knot/src/verbosity.rs

@@ -0,0 +1 @@
+

crates/red_knot_python_semantic/Cargo.toml

@@ -1,5 +1,5 @@
 [package]
-name = "ty_python_semantic"
+name = "red_knot_python_semantic"
 version = "0.0.0"
 publish = false
 authors = { workspace = true }
@@ -12,54 +12,41 @@ license = { workspace = true }
 
 [dependencies]
 ruff_db = { workspace = true }
-ruff_index = { workspace = true, features = ["salsa"] }
-ruff_macros = { workspace = true }
-ruff_python_ast = { workspace = true, features = ["salsa"] }
-ruff_python_parser = { workspace = true }
+ruff_index = { workspace = true }
+ruff_python_ast = { workspace = true }
 ruff_python_stdlib = { workspace = true }
 ruff_source_file = { workspace = true }
 ruff_text_size = { workspace = true }
 ruff_python_literal = { workspace = true }
-ruff_python_trivia = { workspace = true }
 
 anyhow = { workspace = true }
 bitflags = { workspace = true }
 camino = { workspace = true }
 compact_str = { workspace = true }
 countme = { workspace = true }
-drop_bomb = { workspace = true }
-indexmap = { workspace = true }
-itertools = { workspace = true }
+itertools = { workspace = true}
 ordermap = { workspace = true }
-salsa = { workspace = true, features = ["compact_str"] }
+salsa = { workspace = true }
 thiserror = { workspace = true }
 tracing = { workspace = true }
 rustc-hash = { workspace = true }
 hashbrown = { workspace = true }
-schemars = { workspace = true, optional = true }
-serde = { workspace = true, optional = true }
 smallvec = { workspace = true }
 static_assertions = { workspace = true }
 test-case = { workspace = true }
 memchr = { workspace = true }
-strum = { workspace = true }
-strum_macros = { workspace = true }
 
 [dev-dependencies]
-ruff_db = { workspace = true, features = ["testing", "os"] }
+ruff_db = { workspace = true, features = ["os", "testing"] }
 ruff_python_parser = { workspace = true }
-ty_test = { workspace = true }
-ty_vendored = { workspace = true }
+red_knot_test = { workspace = true }
+red_knot_vendored = { workspace = true }
 
 anyhow = { workspace = true }
-dir-test = { workspace = true }
 insta = { workspace = true }
+rstest = { workspace = true }
 tempfile = { workspace = true }
-quickcheck = { version = "1.0.3", default-features = false }
-quickcheck_macros = { version = "1.0.0" }
-
-[features]
-serde = ["ruff_db/serde", "dep:serde", "ruff_python_ast/serde"]
 
 [lints]
 workspace = true
+

crates/red_knot_python_semantic/build.rs

@@ -1,4 +1,4 @@
 /// Rebuild the crate if a test file is added or removed from
 pub fn main() {
-    println!("cargo::rerun-if-changed=resources/mdtest");
+    println!("cargo:rerun-if-changed=resources/mdtest");
 }

crates/red_knot_python_semantic/resources/README.md

@@ -1,4 +1,4 @@
 Markdown files within the `mdtest/` subdirectory are tests of type inference and type checking;
 executed by the `tests/mdtest.rs` integration test.
 
-See `crates/ty_test/README.md` for documentation of this test format.
+See `crates/red_knot_test/README.md` for documentation of this test format.

crates/red_knot_python_semantic/resources/mdtest/comparison/byte_literals.md

@@ -0,0 +1,43 @@
+### Comparison: Byte literals
+
+These tests assert that we infer precise `Literal` types for comparisons between objects
+inferred as having `Literal` bytes types:
+
+```py
+reveal_type(b"abc" == b"abc")  # revealed: Literal[True]
+reveal_type(b"abc" == b"ab")   # revealed: Literal[False]
+
+reveal_type(b"abc" != b"abc")  # revealed: Literal[False]
+reveal_type(b"abc" != b"ab")   # revealed: Literal[True]
+
+reveal_type(b"abc" < b"abd")   # revealed: Literal[True]
+reveal_type(b"abc" < b"abb")   # revealed: Literal[False]
+
+reveal_type(b"abc" <= b"abc")  # revealed: Literal[True]
+reveal_type(b"abc" <= b"abb")  # revealed: Literal[False]
+
+reveal_type(b"abc" > b"abd")   # revealed: Literal[False]
+reveal_type(b"abc" > b"abb")   # revealed: Literal[True]
+
+reveal_type(b"abc" >= b"abc")  # revealed: Literal[True]
+reveal_type(b"abc" >= b"abd")  # revealed: Literal[False]
+
+reveal_type(b"" in b"")                      # revealed: Literal[True]
+reveal_type(b"" in b"abc")                   # revealed: Literal[True]
+reveal_type(b"abc" in b"")                   # revealed: Literal[False]
+reveal_type(b"ab" in b"abc")                 # revealed: Literal[True]
+reveal_type(b"abc" in b"abc")                # revealed: Literal[True]
+reveal_type(b"d" in b"abc")                  # revealed: Literal[False]
+reveal_type(b"ac" in b"abc")                 # revealed: Literal[False]
+reveal_type(b"\x81\x82" in b"\x80\x81\x82")  # revealed: Literal[True]
+reveal_type(b"\x82\x83" in b"\x80\x81\x82")  # revealed: Literal[False]
+
+reveal_type(b"ab" not in b"abc")             # revealed: Literal[False]
+reveal_type(b"ac" not in b"abc")             # revealed: Literal[True]
+
+reveal_type(b"abc" is b"abc")      # revealed: bool
+reveal_type(b"abc" is b"ab")       # revealed: Literal[False]
+
+reveal_type(b"abc" is not b"abc")  # revealed: bool
+reveal_type(b"abc" is not b"ab")   # revealed: Literal[True]
+```

crates/red_knot_python_semantic/resources/mdtest/narrow/conditionals_is.md

@@ -0,0 +1,29 @@
+# Narrowing for `is` conditionals
+
+## `is None`
+
+```py
+x = None if flag else 1
+
+if x is None:
+    # TODO the following should be simplified to 'None'
+    reveal_type(x)  # revealed: None | Literal[1] & None
+
+reveal_type(x)  # revealed: None | Literal[1]
+```
+
+## `is` for other types
+
+```py
+class A:
+    ...
+
+x = A()
+y = x if flag else None
+
+if y is x:
+    # TODO the following should be simplified to 'A'
+    reveal_type(y)  # revealed: A | None & A
+
+reveal_type(y)  # revealed: A | None
+```

crates/red_knot_python_semantic/resources/mdtest/narrow/conditionals_is_not.md

@@ -0,0 +1,40 @@
+# Narrowing for `is not` conditionals
+
+## `is not None`
+
+The type guard removes `None` from the union type:
+
+```py
+x = None if flag else 1
+
+if x is not None:
+    reveal_type(x)  # revealed: Literal[1]
+
+reveal_type(x)  # revealed: None | Literal[1]
+```
+
+## `is not` for other singleton types
+
+```py
+x = True if flag else False
+reveal_type(x)  # revealed: bool
+
+if x is not False:
+    # TODO the following should be `Literal[True]`
+    reveal_type(x)  # revealed: bool & ~Literal[False]
+```
+
+## `is not` for non-singleton types
+
+Non-singleton types should *not* narrow the type: two instances of a
+non-singleton class may occupy different addresses in memory even if
+they compare equal.
+
+```py
+x = [1]
+y = [1]
+
+if x is not y:
+    # TODO: should include type parameter: list[int]
+    reveal_type(x)  # revealed: list
+```

crates/red_knot_python_semantic/resources/mdtest/narrow/match.md

@@ -0,0 +1,17 @@
+# Narrowing for `match` statements
+
+## Single `match` pattern
+
+```py
+x = None if flag else 1
+reveal_type(x)  # revealed: None | Literal[1]
+
+y = 0
+
+match x:
+    case None:
+        y = x
+
+# TODO intersection simplification: should be just Literal[0] | None
+reveal_type(y)  # revealed: Literal[0] | None | Literal[1] & None
+```

crates/red_knot_python_semantic/resources/mdtest/numbers.md

@@ -0,0 +1,35 @@
+# Numbers
+
+## Integers
+
+### Literals
+
+We can infer an integer literal type:
+
+```py
+reveal_type(1)  # revealed: Literal[1]
+```
+
+### Overflow
+
+We only track integer literals within the range of an i64:
+
+```py
+reveal_type(9223372036854775808)  # revealed: int
+```
+
+## Floats
+
+There aren't literal float types, but we infer the general float type:
+
+```py
+reveal_type(1.0)  # revealed: float
+```
+
+## Complex
+
+Same for complex:
+
+```py
+reveal_type(2j)  # revealed: complex
+```

crates/red_knot_python_semantic/src/ast_node_ref.rs

@@ -12,40 +12,22 @@ use ruff_db::parsed::ParsedModule;
 /// Holding on to any [`AstNodeRef`] prevents the [`ParsedModule`] from being released.
 ///
 /// ## Equality
-/// Two `AstNodeRef` are considered equal if their pointer addresses are equal.
-///
-/// ## Usage in salsa tracked structs
-/// It's important that [`AstNodeRef`] fields in salsa tracked structs are tracked fields
-/// (attributed with `#[tracked`]). It prevents that the tracked struct gets a new ID
-/// every time the AST changes, which in turn, invalidates the result of any query
-/// that takes said tracked struct as a query argument or returns the tracked struct as part of its result.
-///
-/// For example, marking the [`AstNodeRef`] as tracked on `Expression`
-/// has the effect that salsa will consider the expression as "unchanged" for as long as it:
-///
-/// * belongs to the same file
-/// * belongs to the same scope
-/// * has the same kind
-/// * was created in the same order
-///
-/// This means that changes to expressions in other scopes don't invalidate the expression's id, giving
-/// us some form of scope-stable identity for expressions. Only queries accessing the node field
-/// run on every AST change. All other queries only run when the expression's identity changes.
+/// Two `AstNodeRef` are considered equal if their wrapped nodes are equal.
 #[derive(Clone)]
 pub struct AstNodeRef<T> {
     /// Owned reference to the node's [`ParsedModule`].
     ///
     /// The node's reference is guaranteed to remain valid as long as it's enclosing
     /// [`ParsedModule`] is alive.
-    parsed: ParsedModule,
+    _parsed: ParsedModule,
 
     /// Pointer to the referenced node.
     node: std::ptr::NonNull<T>,
 }
 
-#[expect(unsafe_code)]
+#[allow(unsafe_code)]
 impl<T> AstNodeRef<T> {
-    /// Creates a new `AstNodeRef` that references `node`. The `parsed` is the [`ParsedModule`] to
+    /// Creates a new `AstNodeRef` that reference `node`. The `parsed` is the [`ParsedModule`] to
     /// which the `AstNodeRef` belongs.
     ///
     /// ## Safety
@@ -55,13 +37,13 @@ impl<T> AstNodeRef<T> {
     /// the invariant `node belongs to parsed` is upheld.
     pub(super) unsafe fn new(parsed: ParsedModule, node: &T) -> Self {
         Self {
-            parsed,
+            _parsed: parsed,
             node: std::ptr::NonNull::from(node),
         }
     }
 
     /// Returns a reference to the wrapped node.
-    pub const fn node(&self) -> &T {
+    pub fn node(&self) -> &T {
         // SAFETY: Holding on to `parsed` ensures that the AST to which `node` belongs is still
         // alive and not moved.
         unsafe { self.node.as_ref() }
@@ -90,14 +72,7 @@ where
     T: PartialEq,
 {
     fn eq(&self, other: &Self) -> bool {
-        if self.parsed == other.parsed {
-            // Comparing the pointer addresses is sufficient to determine equality
-            // if the parsed are the same.
-            self.node.eq(&other.node)
-        } else {
-            // Otherwise perform a deep comparison.
-            self.node().eq(other.node())
-        }
+        self.node().eq(other.node())
     }
 }
 
@@ -112,23 +87,9 @@ where
     }
 }
 
-#[expect(unsafe_code)]
-unsafe impl<T> salsa::Update for AstNodeRef<T> {
-    unsafe fn maybe_update(old_pointer: *mut Self, new_value: Self) -> bool {
-        let old_ref = unsafe { &mut (*old_pointer) };
-
-        if old_ref.parsed == new_value.parsed && old_ref.node.eq(&new_value.node) {
-            false
-        } else {
-            *old_ref = new_value;
-            true
-        }
-    }
-}
-
-#[expect(unsafe_code)]
+#[allow(unsafe_code)]
 unsafe impl<T> Send for AstNodeRef<T> where T: Send {}
-#[expect(unsafe_code)]
+#[allow(unsafe_code)]
 unsafe impl<T> Sync for AstNodeRef<T> where T: Sync {}
 
 #[cfg(test)]
@@ -139,7 +100,7 @@ mod tests {
     use ruff_python_parser::parse_unchecked_source;
 
     #[test]
-    #[expect(unsafe_code)]
+    #[allow(unsafe_code)]
     fn equality() {
         let parsed_raw = parse_unchecked_source("1 + 2", PySourceType::Python);
         let parsed = ParsedModule::new(parsed_raw.clone());
@@ -167,11 +128,11 @@ mod tests {
         assert_ne!(node1, other_node);
     }
 
-    #[expect(unsafe_code)]
+    #[allow(unsafe_code)]
     #[test]
     fn inequality() {
         let parsed_raw = parse_unchecked_source("1 + 2", PySourceType::Python);
-        let parsed = ParsedModule::new(parsed_raw);
+        let parsed = ParsedModule::new(parsed_raw.clone());
 
         let stmt = &parsed.syntax().body[0];
         let node = unsafe { AstNodeRef::new(parsed.clone(), stmt) };
@@ -186,10 +147,10 @@ mod tests {
     }
 
     #[test]
-    #[expect(unsafe_code)]
+    #[allow(unsafe_code)]
     fn debug() {
         let parsed_raw = parse_unchecked_source("1 + 2", PySourceType::Python);
-        let parsed = ParsedModule::new(parsed_raw);
+        let parsed = ParsedModule::new(parsed_raw.clone());
 
         let stmt = &parsed.syntax().body[0];
 

crates/red_knot_python_semantic/src/db.rs

@@ -1,59 +1,51 @@
+use ruff_db::files::File;
 use ruff_db::{Db as SourceDb, Upcast};
-use ty_python_semantic::Db as SemanticDb;
 
+/// Database giving access to semantic information about a Python program.
 #[salsa::db]
-pub trait Db: SemanticDb + Upcast<dyn SemanticDb> + Upcast<dyn SourceDb> {}
+pub trait Db: SourceDb + Upcast<dyn SourceDb> {
+    fn is_file_open(&self, file: File) -> bool;
+}
 
 #[cfg(test)]
 pub(crate) mod tests {
-    use std::sync::{Arc, Mutex};
+    use std::sync::Arc;
 
-    use super::Db;
     use ruff_db::files::{File, Files};
     use ruff_db::system::{DbWithTestSystem, System, TestSystem};
     use ruff_db::vendored::VendoredFileSystem;
     use ruff_db::{Db as SourceDb, Upcast};
-    use ty_python_semantic::lint::{LintRegistry, RuleSelection};
-    use ty_python_semantic::{Db as SemanticDb, Program, default_lint_registry};
 
-    type Events = Arc<Mutex<Vec<salsa::Event>>>;
+    use super::Db;
 
     #[salsa::db]
-    #[derive(Clone)]
     pub(crate) struct TestDb {
         storage: salsa::Storage<Self>,
         files: Files,
         system: TestSystem,
         vendored: VendoredFileSystem,
-        events: Events,
-        rule_selection: Arc<RuleSelection>,
+        events: std::sync::Arc<std::sync::Mutex<Vec<salsa::Event>>>,
     }
 
-    #[expect(dead_code)]
     impl TestDb {
         pub(crate) fn new() -> Self {
-            let events = Events::default();
             Self {
-                storage: salsa::Storage::new(Some(Box::new({
-                    let events = events.clone();
-                    move |event| {
-                        tracing::trace!("event: {event:?}");
-                        let mut events = events.lock().unwrap();
-                        events.push(event);
-                    }
-                }))),
+                storage: salsa::Storage::default(),
                 system: TestSystem::default(),
-                vendored: ty_vendored::file_system().clone(),
-                events,
+                vendored: red_knot_vendored::file_system().clone(),
+                events: std::sync::Arc::default(),
                 files: Files::default(),
-                rule_selection: Arc::new(RuleSelection::from_registry(default_lint_registry())),
             }
         }
 
         /// Takes the salsa events.
+        ///
+        /// ## Panics
+        /// If there are any pending salsa snapshots.
         pub(crate) fn take_salsa_events(&mut self) -> Vec<salsa::Event> {
-            let mut events = self.events.lock().unwrap();
+            let inner = Arc::get_mut(&mut self.events).expect("no pending salsa snapshots");
 
+            let events = inner.get_mut().unwrap();
             std::mem::take(&mut *events)
         }
 
@@ -89,10 +81,6 @@ pub(crate) mod tests {
         fn files(&self) -> &Files {
             &self.files
         }
-
-        fn python_version(&self) -> ruff_python_ast::PythonVersion {
-            Program::get(self).python_version(self)
-        }
     }
 
     impl Upcast<dyn SourceDb> for TestDb {
@@ -104,34 +92,20 @@ pub(crate) mod tests {
         }
     }
 
-    impl Upcast<dyn SemanticDb> for TestDb {
-        fn upcast(&self) -> &(dyn SemanticDb + 'static) {
-            self
-        }
-
-        fn upcast_mut(&mut self) -> &mut dyn SemanticDb {
-            self
-        }
-    }
-
     #[salsa::db]
-    impl SemanticDb for TestDb {
+    impl Db for TestDb {
         fn is_file_open(&self, file: File) -> bool {
             !file.path(self).is_vendored_path()
         }
-
-        fn rule_selection(&self) -> &RuleSelection {
-            &self.rule_selection
-        }
-
-        fn lint_registry(&self) -> &LintRegistry {
-            default_lint_registry()
-        }
     }
 
     #[salsa::db]
-    impl Db for TestDb {}
-
-    #[salsa::db]
-    impl salsa::Database for TestDb {}
+    impl salsa::Database for TestDb {
+        fn salsa_event(&self, event: &dyn Fn() -> salsa::Event) {
+            let event = event();
+            tracing::trace!("event: {event:?}");
+            let mut events = self.events.lock().unwrap();
+            events.push(event);
+        }
+    }
 }

crates/red_knot_python_semantic/src/lib.rs

@@ -0,0 +1,25 @@
+use std::hash::BuildHasherDefault;
+
+use rustc_hash::FxHasher;
+
+pub use db::Db;
+pub use module_name::ModuleName;
+pub use module_resolver::{resolve_module, system_module_search_paths, Module};
+pub use program::{Program, ProgramSettings, SearchPathSettings, SitePackages};
+pub use python_version::PythonVersion;
+pub use semantic_model::{HasTy, SemanticModel};
+
+pub mod ast_node_ref;
+mod db;
+mod module_name;
+mod module_resolver;
+mod node_key;
+mod program;
+mod python_version;
+pub mod semantic_index;
+mod semantic_model;
+pub(crate) mod site_packages;
+mod stdlib;
+pub mod types;
+
+type FxOrderSet<V> = ordermap::set::OrderSet<V, BuildHasherDefault<FxHasher>>;

crates/red_knot_python_semantic/src/module_name.rs

@@ -1,15 +1,10 @@
 use std::fmt;
-use std::num::NonZeroU32;
 use std::ops::Deref;
 
 use compact_str::{CompactString, ToCompactString};
 
-use ruff_db::files::File;
-use ruff_python_ast as ast;
 use ruff_python_stdlib::identifiers::is_identifier;
 
-use crate::{db::Db, module_resolver::file_to_module};
-
 /// A module name, e.g. `foo.bar`.
 ///
 /// Always normalized to the absolute form (never a relative module name, i.e., never `.foo`).
@@ -47,7 +42,7 @@ impl ModuleName {
     /// ## Examples
     ///
     /// ```
-    /// use ty_python_semantic::ModuleName;
+    /// use red_knot_python_semantic::ModuleName;
     ///
     /// assert_eq!(ModuleName::new_static("foo.bar").as_deref(), Some("foo.bar"));
     /// assert_eq!(ModuleName::new_static(""), None);
@@ -73,7 +68,7 @@ impl ModuleName {
     /// # Examples
     ///
     /// ```
-    /// use ty_python_semantic::ModuleName;
+    /// use red_knot_python_semantic::ModuleName;
     ///
     /// assert_eq!(ModuleName::new_static("foo.bar.baz").unwrap().components().collect::<Vec<_>>(), vec!["foo", "bar", "baz"]);
     /// ```
@@ -87,7 +82,7 @@ impl ModuleName {
     /// # Examples
     ///
     /// ```
-    /// use ty_python_semantic::ModuleName;
+    /// use red_knot_python_semantic::ModuleName;
     ///
     /// assert_eq!(ModuleName::new_static("foo.bar").unwrap().parent(), Some(ModuleName::new_static("foo").unwrap()));
     /// assert_eq!(ModuleName::new_static("foo.bar.baz").unwrap().parent(), Some(ModuleName::new_static("foo.bar").unwrap()));
@@ -106,7 +101,7 @@ impl ModuleName {
     /// # Examples
     ///
     /// ```
-    /// use ty_python_semantic::ModuleName;
+    /// use red_knot_python_semantic::ModuleName;
     ///
     /// assert!(ModuleName::new_static("foo.bar").unwrap().starts_with(&ModuleName::new_static("foo").unwrap()));
     ///
@@ -138,7 +133,7 @@ impl ModuleName {
     /// # Examples
     ///
     /// ```
-    /// use ty_python_semantic::ModuleName;
+    /// use red_knot_python_semantic::ModuleName;
     ///
     /// assert_eq!(&*ModuleName::from_components(["a"]).unwrap(), "a");
     /// assert_eq!(&*ModuleName::from_components(["a", "b"]).unwrap(), "a.b");
@@ -179,7 +174,7 @@ impl ModuleName {
     /// # Examples
     ///
     /// ```
-    /// use ty_python_semantic::ModuleName;
+    /// use red_knot_python_semantic::ModuleName;
     ///
     /// let mut module_name = ModuleName::new_static("foo").unwrap();
     /// module_name.extend(&ModuleName::new_static("bar").unwrap());
@@ -191,49 +186,6 @@ impl ModuleName {
         self.0.push('.');
         self.0.push_str(other);
     }
-
-    /// Returns an iterator of this module name and all of its parent modules.
-    ///
-    /// # Examples
-    ///
-    /// ```
-    /// use ty_python_semantic::ModuleName;
-    ///
-    /// assert_eq!(
-    ///     ModuleName::new_static("foo.bar.baz").unwrap().ancestors().collect::<Vec<_>>(),
-    ///     vec![
-    ///         ModuleName::new_static("foo.bar.baz").unwrap(),
-    ///         ModuleName::new_static("foo.bar").unwrap(),
-    ///         ModuleName::new_static("foo").unwrap(),
-    ///     ],
-    /// );
-    /// ```
-    pub fn ancestors(&self) -> impl Iterator<Item = Self> {
-        std::iter::successors(Some(self.clone()), Self::parent)
-    }
-
-    pub(crate) fn from_import_statement<'db>(
-        db: &'db dyn Db,
-        importing_file: File,
-        node: &'db ast::StmtImportFrom,
-    ) -> Result<Self, ModuleNameResolutionError> {
-        let ast::StmtImportFrom {
-            module,
-            level,
-            names: _,
-            range: _,
-        } = node;
-
-        let module = module.as_deref();
-
-        if let Some(level) = NonZeroU32::new(*level) {
-            relative_module_name(db, importing_file, module, level)
-        } else {
-            module
-                .and_then(Self::new)
-                .ok_or(ModuleNameResolutionError::InvalidSyntax)
-        }
-    }
 }
 
 impl Deref for ModuleName {
@@ -262,58 +214,3 @@ impl std::fmt::Display for ModuleName {
         f.write_str(&self.0)
     }
 }
-
-/// Given a `from .foo import bar` relative import, resolve the relative module
-/// we're importing `bar` from into an absolute [`ModuleName`]
-/// using the name of the module we're currently analyzing.
-///
-/// - `level` is the number of dots at the beginning of the relative module name:
-///   - `from .foo.bar import baz` => `level == 1`
-///   - `from ...foo.bar import baz` => `level == 3`
-/// - `tail` is the relative module name stripped of all leading dots:
-///   - `from .foo import bar` => `tail == "foo"`
-///   - `from ..foo.bar import baz` => `tail == "foo.bar"`
-fn relative_module_name(
-    db: &dyn Db,
-    importing_file: File,
-    tail: Option<&str>,
-    level: NonZeroU32,
-) -> Result<ModuleName, ModuleNameResolutionError> {
-    let module = file_to_module(db, importing_file)
-        .ok_or(ModuleNameResolutionError::UnknownCurrentModule)?;
-    let mut level = level.get();
-
-    if module.kind().is_package() {
-        level = level.saturating_sub(1);
-    }
-
-    let mut module_name = module
-        .name()
-        .ancestors()
-        .nth(level as usize)
-        .ok_or(ModuleNameResolutionError::TooManyDots)?;
-
-    if let Some(tail) = tail {
-        let tail = ModuleName::new(tail).ok_or(ModuleNameResolutionError::InvalidSyntax)?;
-        module_name.extend(&tail);
-    }
-
-    Ok(module_name)
-}
-
-/// Various ways in which resolving a [`ModuleName`]
-/// from an [`ast::StmtImport`] or [`ast::StmtImportFrom`] node might fail
-#[derive(Debug, Copy, Clone, PartialEq, Eq)]
-pub(crate) enum ModuleNameResolutionError {
-    /// The import statement has invalid syntax
-    InvalidSyntax,
-
-    /// We couldn't resolve the file we're currently analyzing back to a module
-    /// (Only necessary for relative import statements)
-    UnknownCurrentModule,
-
-    /// The relative import statement seems to take us outside of the module search path
-    /// (e.g. our current module is `foo.bar`, and the relative import statement in `foo.bar`
-    /// is `from ....baz import spam`)
-    TooManyDots,
-}

crates/red_knot_python_semantic/src/module_resolver/mod.rs

@@ -1,12 +1,12 @@
 use std::iter::FusedIterator;
 
-pub use module::{KnownModule, Module};
+pub use module::Module;
 pub use resolver::resolve_module;
-pub(crate) use resolver::{SearchPaths, file_to_module};
+pub(crate) use resolver::{file_to_module, SearchPaths};
 use ruff_db::system::SystemPath;
 
-use crate::Db;
 use crate::module_resolver::resolver::search_paths;
+use crate::Db;
 use resolver::SearchPathIterator;
 
 mod module;

crates/red_knot_python_semantic/src/module_resolver/module.rs

@@ -0,0 +1,85 @@
+use std::fmt::Formatter;
+use std::sync::Arc;
+
+use ruff_db::files::File;
+
+use super::path::SearchPath;
+use crate::module_name::ModuleName;
+
+/// Representation of a Python module.
+#[derive(Clone, PartialEq, Eq)]
+pub struct Module {
+    inner: Arc<ModuleInner>,
+}
+
+impl Module {
+    pub(crate) fn new(
+        name: ModuleName,
+        kind: ModuleKind,
+        search_path: SearchPath,
+        file: File,
+    ) -> Self {
+        Self {
+            inner: Arc::new(ModuleInner {
+                name,
+                kind,
+                search_path,
+                file,
+            }),
+        }
+    }
+
+    /// The absolute name of the module (e.g. `foo.bar`)
+    pub fn name(&self) -> &ModuleName {
+        &self.inner.name
+    }
+
+    /// The file to the source code that defines this module
+    pub fn file(&self) -> File {
+        self.inner.file
+    }
+
+    /// The search path from which the module was resolved.
+    pub(crate) fn search_path(&self) -> &SearchPath {
+        &self.inner.search_path
+    }
+
+    /// Determine whether this module is a single-file module or a package
+    pub fn kind(&self) -> ModuleKind {
+        self.inner.kind
+    }
+}
+
+impl std::fmt::Debug for Module {
+    fn fmt(&self, f: &mut Formatter<'_>) -> std::fmt::Result {
+        f.debug_struct("Module")
+            .field("name", &self.name())
+            .field("kind", &self.kind())
+            .field("file", &self.file())
+            .field("search_path", &self.search_path())
+            .finish()
+    }
+}
+
+#[derive(PartialEq, Eq)]
+struct ModuleInner {
+    name: ModuleName,
+    kind: ModuleKind,
+    search_path: SearchPath,
+    file: File,
+}
+
+#[derive(Copy, Clone, Debug, Eq, PartialEq, Hash)]
+pub enum ModuleKind {
+    /// A single-file module (e.g. `foo.py` or `foo.pyi`)
+    Module,
+
+    /// A python package (`foo/__init__.py` or `foo/__init__.pyi`)
+    Package,
+}
+
+impl ModuleKind {
+    pub const fn is_package(self) -> bool {
+        matches!(self, ModuleKind::Package)
+    }
+}

crates/red_knot_python_semantic/src/module_resolver/path.rs

@@ -4,11 +4,12 @@ use std::fmt;
 use std::sync::Arc;
 
 use camino::{Utf8Path, Utf8PathBuf};
-use ruff_db::files::{File, FileError, system_path_to_file, vendored_path_to_file};
+
+use ruff_db::files::{system_path_to_file, vendored_path_to_file, File, FileError};
 use ruff_db::system::{System, SystemPath, SystemPathBuf};
 use ruff_db::vendored::{VendoredPath, VendoredPathBuf};
 
-use super::typeshed::{TypeshedVersionsParseError, TypeshedVersionsQueryResult, typeshed_versions};
+use super::typeshed::{typeshed_versions, TypeshedVersionsParseError, TypeshedVersionsQueryResult};
 use crate::db::Db;
 use crate::module_name::ModuleName;
 use crate::module_resolver::resolver::ResolverContext;
@@ -62,10 +63,6 @@ impl ModulePath {
         self.relative_path.pop()
     }
 
-    pub(super) fn search_path(&self) -> &SearchPath {
-        &self.search_path
-    }
-
     #[must_use]
     pub(super) fn is_directory(&self, resolver: &ResolverContext) -> bool {
         let ModulePath {
@@ -115,9 +112,8 @@ impl ModulePath {
             | SearchPathInner::SitePackages(search_path)
             | SearchPathInner::Editable(search_path) => {
                 let absolute_path = search_path.join(relative_path);
-
                 system_path_to_file(resolver.db.upcast(), absolute_path.join("__init__.py")).is_ok()
-                    || system_path_to_file(resolver.db.upcast(), absolute_path.join("__init__.pyi"))
+                    || system_path_to_file(resolver.db.upcast(), absolute_path.join("__init__.py"))
                         .is_ok()
             }
             SearchPathInner::StandardLibraryCustom(search_path) => {
@@ -143,23 +139,6 @@ impl ModulePath {
         }
     }
 
-    pub(super) fn to_system_path(&self) -> Option<SystemPathBuf> {
-        let ModulePath {
-            search_path,
-            relative_path,
-        } = self;
-        match &*search_path.0 {
-            SearchPathInner::Extra(search_path)
-            | SearchPathInner::FirstParty(search_path)
-            | SearchPathInner::SitePackages(search_path)
-            | SearchPathInner::Editable(search_path) => Some(search_path.join(relative_path)),
-            SearchPathInner::StandardLibraryCustom(stdlib_root) => {
-                Some(stdlib_root.join(relative_path))
-            }
-            SearchPathInner::StandardLibraryVendored(_) => None,
-        }
-    }
-
     #[must_use]
     pub(super) fn to_file(&self, resolver: &ResolverContext) -> Option<File> {
         let db = resolver.db.upcast();
@@ -205,18 +184,7 @@ impl ModulePath {
             stdlib_path_to_module_name(relative_path)
         } else {
             let parent = relative_path.parent()?;
-            let parent_components = parent.components().enumerate().map(|(index, component)| {
-                let component = component.as_str();
-
-                // For stub packages, strip the `-stubs` suffix from the first component
-                // because it isn't a valid module name part AND the module name is the name without the `-stubs`.
-                if index == 0 {
-                    component.strip_suffix("-stubs").unwrap_or(component)
-                } else {
-                    component
-                }
-            });
-
+            let parent_components = parent.components().map(|component| component.as_str());
             let skip_final_part =
                 relative_path.ends_with("__init__.py") || relative_path.ends_with("__init__.pyi");
             if skip_final_part {
@@ -315,9 +283,9 @@ fn query_stdlib_version(
     let Some(module_name) = stdlib_path_to_module_name(relative_path) else {
         return TypeshedVersionsQueryResult::DoesNotExist;
     };
-    let ResolverContext { db, python_version } = context;
+    let ResolverContext { db, target_version } = context;
 
-    typeshed_versions(*db).query_module(&module_name, *python_version)
+    typeshed_versions(*db).query_module(&module_name, *target_version)
 }
 
 /// Enumeration describing the various ways in which validation of a search path might fail.
@@ -660,26 +628,13 @@ impl PartialEq<SearchPath> for VendoredPathBuf {
     }
 }
 
-impl fmt::Display for SearchPath {
-    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
-        match &*self.0 {
-            SearchPathInner::Extra(system_path_buf)
-            | SearchPathInner::FirstParty(system_path_buf)
-            | SearchPathInner::SitePackages(system_path_buf)
-            | SearchPathInner::Editable(system_path_buf)
-            | SearchPathInner::StandardLibraryCustom(system_path_buf) => system_path_buf.fmt(f),
-            SearchPathInner::StandardLibraryVendored(vendored_path_buf) => vendored_path_buf.fmt(f),
-        }
-    }
-}
-
 #[cfg(test)]
 mod tests {
     use ruff_db::Db;
-    use ruff_python_ast::PythonVersion;
 
     use crate::db::tests::TestDb;
     use crate::module_resolver::testing::{FileSpec, MockedTypeshed, TestCase, TestCaseBuilder};
+    use crate::python_version::PythonVersion;
 
     use super::*;
 
@@ -703,7 +658,7 @@ mod tests {
         let TestCase {
             db, src, stdlib, ..
         } = TestCaseBuilder::new()
-            .with_mocked_typeshed(MockedTypeshed::default())
+            .with_custom_typeshed(MockedTypeshed::default())
             .build();
 
         assert_eq!(
@@ -824,7 +779,7 @@ mod tests {
     #[should_panic(expected = "Extension must be `pyi`; got `py`")]
     fn stdlib_path_invalid_join_py() {
         let TestCase { db, stdlib, .. } = TestCaseBuilder::new()
-            .with_mocked_typeshed(MockedTypeshed::default())
+            .with_custom_typeshed(MockedTypeshed::default())
             .build();
         SearchPath::custom_stdlib(&db, stdlib.parent().unwrap())
             .unwrap()
@@ -836,7 +791,7 @@ mod tests {
     #[should_panic(expected = "Extension must be `pyi`; got `rs`")]
     fn stdlib_path_invalid_join_rs() {
         let TestCase { db, stdlib, .. } = TestCaseBuilder::new()
-            .with_mocked_typeshed(MockedTypeshed::default())
+            .with_custom_typeshed(MockedTypeshed::default())
             .build();
         SearchPath::custom_stdlib(&db, stdlib.parent().unwrap())
             .unwrap()
@@ -867,7 +822,7 @@ mod tests {
     #[test]
     fn relativize_stdlib_path_errors() {
         let TestCase { db, stdlib, .. } = TestCaseBuilder::new()
-            .with_mocked_typeshed(MockedTypeshed::default())
+            .with_custom_typeshed(MockedTypeshed::default())
             .build();
 
         let root = SearchPath::custom_stdlib(&db, stdlib.parent().unwrap()).unwrap();
@@ -912,11 +867,11 @@ mod tests {
 
     fn typeshed_test_case(
         typeshed: MockedTypeshed,
-        python_version: PythonVersion,
+        target_version: PythonVersion,
     ) -> (TestDb, SearchPath) {
         let TestCase { db, stdlib, .. } = TestCaseBuilder::new()
-            .with_mocked_typeshed(typeshed)
-            .with_python_version(python_version)
+            .with_custom_typeshed(typeshed)
+            .with_target_version(target_version)
             .build();
         let stdlib = SearchPath::custom_stdlib(&db, stdlib.parent().unwrap()).unwrap();
         (db, stdlib)
@@ -950,12 +905,10 @@ mod tests {
         assert!(asyncio_regular_package.is_regular_package(&resolver));
         // Paths to directories don't resolve to VfsFiles
         assert_eq!(asyncio_regular_package.to_file(&resolver), None);
-        assert!(
-            asyncio_regular_package
-                .join("__init__.pyi")
-                .to_file(&resolver)
-                .is_some()
-        );
+        assert!(asyncio_regular_package
+            .join("__init__.pyi")
+            .to_file(&resolver)
+            .is_some());
 
         // The `asyncio` package exists on Python 3.8, but the `asyncio.tasks` submodule does not,
         // according to the `VERSIONS` file in our typeshed mock:
@@ -1085,12 +1038,10 @@ mod tests {
         assert!(collections_regular_package.is_regular_package(&resolver));
         // (This is still `None`, as directories don't resolve to `Vfs` files)
         assert_eq!(collections_regular_package.to_file(&resolver), None);
-        assert!(
-            collections_regular_package
-                .join("__init__.pyi")
-                .to_file(&resolver)
-                .is_some()
-        );
+        assert!(collections_regular_package
+            .join("__init__.pyi")
+            .to_file(&resolver)
+            .is_some());
 
         // ...and so should the `asyncio.tasks` submodule (though it's still not a directory):
         let asyncio_tasks_module = stdlib_path.join("asyncio/tasks.pyi");

crates/red_knot_python_semantic/src/module_resolver/testing.rs

@@ -1,14 +1,10 @@
-use ruff_db::system::{
-    DbWithTestSystem as _, DbWithWritableSystem as _, SystemPath, SystemPathBuf,
-};
+use ruff_db::system::{DbWithTestSystem, SystemPath, SystemPathBuf};
 use ruff_db::vendored::VendoredPathBuf;
-use ruff_python_ast::PythonVersion;
 
 use crate::db::tests::TestDb;
 use crate::program::{Program, SearchPathSettings};
-use crate::{
-    ProgramSettings, PythonPath, PythonPlatform, PythonVersionSource, PythonVersionWithSource,
-};
+use crate::python_version::PythonVersion;
+use crate::{ProgramSettings, SitePackages};
 
 /// A test case for the module resolver.
 ///
@@ -22,7 +18,7 @@ pub(crate) struct TestCase<T> {
     // so this is a single directory instead of a `Vec` of directories,
     // like it is in `ruff_db::Program`.
     pub(crate) site_packages: SystemPathBuf,
-    pub(crate) python_version: PythonVersion,
+    pub(crate) target_version: PythonVersion,
 }
 
 /// A `(file_name, file_contents)` tuple
@@ -71,13 +67,13 @@ pub(crate) struct UnspecifiedTypeshed;
 /// ```rs
 /// let test_case = TestCaseBuilder::new()
 ///     .with_src_files(...)
-///     .with_python_version(...)
+///     .with_target_version(...)
 ///     .build();
 /// ```
 ///
 /// For tests checking that standard-library module resolution is working
 /// correctly, you should usually create a [`MockedTypeshed`] instance
-/// and pass it to the [`TestCaseBuilder::with_mocked_typeshed`] method.
+/// and pass it to the [`TestCaseBuilder::with_custom_typeshed`] method.
 /// If you need to check something that involves the vendored typeshed stubs
 /// we include as part of the binary, you can instead use the
 /// [`TestCaseBuilder::with_vendored_typeshed`] method.
@@ -89,13 +85,13 @@ pub(crate) struct UnspecifiedTypeshed;
 /// const TYPESHED = MockedTypeshed { ... };
 ///
 /// let test_case = resolver_test_case()
-///     .with_mocked_typeshed(TYPESHED)
-///     .with_python_version(...)
+///     .with_custom_typeshed(TYPESHED)
+///     .with_target_version(...)
 ///     .build();
 ///
 /// let test_case2 = resolver_test_case()
 ///     .with_vendored_typeshed()
-///     .with_python_version(...)
+///     .with_target_version(...)
 ///     .build();
 /// ```
 ///
@@ -104,8 +100,7 @@ pub(crate) struct UnspecifiedTypeshed;
 /// to `()`.
 pub(crate) struct TestCaseBuilder<T> {
     typeshed_option: T,
-    python_version: PythonVersion,
-    python_platform: PythonPlatform,
+    target_version: PythonVersion,
     first_party_files: Vec<FileSpec>,
     site_packages_files: Vec<FileSpec>,
 }
@@ -123,9 +118,9 @@ impl<T> TestCaseBuilder<T> {
         self
     }
 
-    /// Specify the Python version the module resolver should assume
-    pub(crate) fn with_python_version(mut self, python_version: PythonVersion) -> Self {
-        self.python_version = python_version;
+    /// Specify the target Python version the module resolver should assume
+    pub(crate) fn with_target_version(mut self, target_version: PythonVersion) -> Self {
+        self.target_version = target_version;
         self
     }
 
@@ -151,8 +146,7 @@ impl TestCaseBuilder<UnspecifiedTypeshed> {
     pub(crate) fn new() -> TestCaseBuilder<UnspecifiedTypeshed> {
         Self {
             typeshed_option: UnspecifiedTypeshed,
-            python_version: PythonVersion::default(),
-            python_platform: PythonPlatform::default(),
+            target_version: PythonVersion::default(),
             first_party_files: vec![],
             site_packages_files: vec![],
         }
@@ -162,37 +156,33 @@ impl TestCaseBuilder<UnspecifiedTypeshed> {
     pub(crate) fn with_vendored_typeshed(self) -> TestCaseBuilder<VendoredTypeshed> {
         let TestCaseBuilder {
             typeshed_option: _,
-            python_version,
-            python_platform,
+            target_version,
             first_party_files,
             site_packages_files,
         } = self;
         TestCaseBuilder {
             typeshed_option: VendoredTypeshed,
-            python_version,
-            python_platform,
+            target_version,
             first_party_files,
             site_packages_files,
         }
     }
 
     /// Use a mock typeshed directory for this test case
-    pub(crate) fn with_mocked_typeshed(
+    pub(crate) fn with_custom_typeshed(
         self,
         typeshed: MockedTypeshed,
     ) -> TestCaseBuilder<MockedTypeshed> {
         let TestCaseBuilder {
             typeshed_option: _,
-            python_version,
-            python_platform,
+            target_version,
             first_party_files,
             site_packages_files,
         } = self;
 
         TestCaseBuilder {
             typeshed_option: typeshed,
-            python_version,
-            python_platform,
+            target_version,
             first_party_files,
             site_packages_files,
         }
@@ -204,15 +194,15 @@ impl TestCaseBuilder<UnspecifiedTypeshed> {
             src,
             stdlib: _,
             site_packages,
-            python_version,
-        } = self.with_mocked_typeshed(MockedTypeshed::default()).build();
+            target_version,
+        } = self.with_custom_typeshed(MockedTypeshed::default()).build();
 
         TestCase {
             db,
             src,
             stdlib: (),
             site_packages,
-            python_version,
+            target_version,
         }
     }
 }
@@ -221,8 +211,7 @@ impl TestCaseBuilder<MockedTypeshed> {
     pub(crate) fn build(self) -> TestCase<SystemPathBuf> {
         let TestCaseBuilder {
             typeshed_option,
-            python_version,
-            python_platform,
+            target_version,
             first_party_files,
             site_packages_files,
         } = self;
@@ -236,17 +225,13 @@ impl TestCaseBuilder<MockedTypeshed> {
 
         Program::from_settings(
             &db,
-            ProgramSettings {
-                python_version: PythonVersionWithSource {
-                    version: python_version,
-                    source: PythonVersionSource::default(),
-                },
-                python_platform,
+            &ProgramSettings {
+                target_version,
                 search_paths: SearchPathSettings {
                     extra_paths: vec![],
-                    src_roots: vec![src.clone()],
+                    src_root: src.clone(),
                     custom_typeshed: Some(typeshed.clone()),
-                    python_path: PythonPath::KnownSitePackages(vec![site_packages.clone()]),
+                    site_packages: SitePackages::Known(vec![site_packages.clone()]),
                 },
             },
         )
@@ -257,7 +242,7 @@ impl TestCaseBuilder<MockedTypeshed> {
             src,
             stdlib: typeshed.join("stdlib"),
             site_packages,
-            python_version,
+            target_version,
         }
     }
 
@@ -283,8 +268,7 @@ impl TestCaseBuilder<VendoredTypeshed> {
     pub(crate) fn build(self) -> TestCase<VendoredPathBuf> {
         let TestCaseBuilder {
             typeshed_option: VendoredTypeshed,
-            python_version,
-            python_platform,
+            target_version,
             first_party_files,
             site_packages_files,
         } = self;
@@ -297,15 +281,11 @@ impl TestCaseBuilder<VendoredTypeshed> {
 
         Program::from_settings(
             &db,
-            ProgramSettings {
-                python_version: PythonVersionWithSource {
-                    version: python_version,
-                    source: PythonVersionSource::default(),
-                },
-                python_platform,
+            &ProgramSettings {
+                target_version,
                 search_paths: SearchPathSettings {
-                    python_path: PythonPath::KnownSitePackages(vec![site_packages.clone()]),
-                    ..SearchPathSettings::new(vec![src.clone()])
+                    site_packages: SitePackages::Known(vec![site_packages.clone()]),
+                    ..SearchPathSettings::new(src.clone())
                 },
             },
         )
@@ -316,7 +296,7 @@ impl TestCaseBuilder<VendoredTypeshed> {
             src,
             stdlib: VendoredPathBuf::from("stdlib"),
             site_packages,
-            python_version,
+            target_version,
         }
     }
 }

crates/red_knot_python_semantic/src/module_resolver/typeshed.rs

@@ -4,12 +4,11 @@ use std::num::{NonZeroU16, NonZeroUsize};
 use std::ops::{RangeFrom, RangeInclusive};
 use std::str::FromStr;
 
-use ruff_python_ast::PythonVersion;
 use rustc_hash::FxHashMap;
 
-use crate::Program;
 use crate::db::Db;
 use crate::module_name::ModuleName;
+use crate::{Program, PythonVersion};
 
 pub(in crate::module_resolver) fn vendored_typeshed_versions(db: &dyn Db) -> TypeshedVersions {
     TypeshedVersions::from_str(
@@ -113,10 +112,10 @@ impl TypeshedVersions {
     pub(in crate::module_resolver) fn query_module(
         &self,
         module: &ModuleName,
-        python_version: PythonVersion,
+        target_version: PythonVersion,
     ) -> TypeshedVersionsQueryResult {
         if let Some(range) = self.exact(module) {
-            if range.contains(python_version) {
+            if range.contains(target_version) {
                 TypeshedVersionsQueryResult::Exists
             } else {
                 TypeshedVersionsQueryResult::DoesNotExist
@@ -126,7 +125,7 @@ impl TypeshedVersions {
             while let Some(module_to_try) = module {
                 if let Some(range) = self.exact(&module_to_try) {
                     return {
-                        if range.contains(python_version) {
+                        if range.contains(target_version) {
                             TypeshedVersionsQueryResult::MaybeExists
                         } else {
                             TypeshedVersionsQueryResult::DoesNotExist
@@ -237,7 +236,7 @@ impl FromStr for TypeshedVersions {
                     return Err(TypeshedVersionsParseError {
                         line_number: Some(line_number),
                         reason,
-                    });
+                    })
                 }
             };
         }
@@ -279,12 +278,12 @@ impl FromStr for PyVersionRange {
         let mut parts = s.split('-').map(str::trim);
         match (parts.next(), parts.next(), parts.next()) {
             (Some(lower), Some(""), None) => {
-                let lower = python_version_from_versions_file_string(lower)?;
+                let lower = PythonVersion::from_versions_file_string(lower)?;
                 Ok(Self::AvailableFrom(lower..))
             }
             (Some(lower), Some(upper), None) => {
-                let lower = python_version_from_versions_file_string(lower)?;
-                let upper = python_version_from_versions_file_string(upper)?;
+                let lower = PythonVersion::from_versions_file_string(lower)?;
+                let upper = PythonVersion::from_versions_file_string(upper)?;
                 Ok(Self::AvailableWithin(lower..=upper))
             }
             _ => Err(TypeshedVersionsParseErrorKind::UnexpectedNumberOfHyphens),
@@ -303,26 +302,25 @@ impl fmt::Display for PyVersionRange {
     }
 }
 
-fn python_version_from_versions_file_string(
-    s: &str,
-) -> Result<PythonVersion, TypeshedVersionsParseErrorKind> {
-    let mut parts = s.split('.').map(str::trim);
-    let (Some(major), Some(minor), None) = (parts.next(), parts.next(), parts.next()) else {
-        return Err(TypeshedVersionsParseErrorKind::UnexpectedNumberOfPeriods(
-            s.to_string(),
-        ));
-    };
-    PythonVersion::try_from((major, minor)).map_err(|int_parse_error| {
-        TypeshedVersionsParseErrorKind::IntegerParsingFailure {
-            version: s.to_string(),
-            err: int_parse_error,
-        }
-    })
+impl PythonVersion {
+    fn from_versions_file_string(s: &str) -> Result<Self, TypeshedVersionsParseErrorKind> {
+        let mut parts = s.split('.').map(str::trim);
+        let (Some(major), Some(minor), None) = (parts.next(), parts.next(), parts.next()) else {
+            return Err(TypeshedVersionsParseErrorKind::UnexpectedNumberOfPeriods(
+                s.to_string(),
+            ));
+        };
+        PythonVersion::try_from((major, minor)).map_err(|int_parse_error| {
+            TypeshedVersionsParseErrorKind::IntegerParsingFailure {
+                version: s.to_string(),
+                err: int_parse_error,
+            }
+        })
+    }
 }
 
 #[cfg(test)]
 mod tests {
-    use std::fmt::Write as _;
     use std::num::{IntErrorKind, NonZeroU16};
     use std::path::Path;
 
@@ -334,7 +332,8 @@ mod tests {
 
     const TYPESHED_STDLIB_DIR: &str = "stdlib";
 
-    const ONE: Option<NonZeroU16> = Some(NonZeroU16::new(1).unwrap());
+    #[allow(unsafe_code)]
+    const ONE: Option<NonZeroU16> = Some(unsafe { NonZeroU16::new_unchecked(1) });
 
     impl TypeshedVersions {
         #[must_use]
@@ -392,7 +391,7 @@ mod tests {
         let db = TestDb::new();
         let vendored_typeshed_versions = vendored_typeshed_versions(&db);
         let vendored_typeshed_dir =
-            Path::new(env!("CARGO_MANIFEST_DIR")).join("../ty_vendored/vendor/typeshed");
+            Path::new(env!("CARGO_MANIFEST_DIR")).join("../red_knot_vendored/vendor/typeshed");
 
         let mut empty_iterator = true;
 
@@ -460,11 +459,11 @@ foo: 3.8-   # trailing comment
 ";
         let parsed_versions = TypeshedVersions::from_str(VERSIONS).unwrap();
         assert_eq!(parsed_versions.len(), 3);
-        assert_snapshot!(parsed_versions.to_string(), @r"
+        assert_snapshot!(parsed_versions.to_string(), @r###"
         bar: 2.7-3.10
         bar.baz: 3.1-3.9
         foo: 3.8-
-        "
+        "###
         );
     }
 
@@ -571,7 +570,7 @@ foo: 3.8-   # trailing comment
 
         let mut massive_versions_file = String::new();
         for i in 0..too_many {
-            let _ = writeln!(&mut massive_versions_file, "x{i}: 3.8-");
+            massive_versions_file.push_str(&format!("x{i}: 3.8-\n"));
         }
 
         assert_eq!(

crates/red_knot_python_semantic/src/node_key.rs

@@ -0,0 +1,24 @@
+use ruff_python_ast::{AnyNodeRef, NodeKind};
+use ruff_text_size::{Ranged, TextRange};
+
+/// Compact key for a node for use in a hash map.
+///
+/// Compares two nodes by their kind and text range.
+#[derive(Copy, Clone, Debug, Eq, PartialEq, Hash)]
+pub(super) struct NodeKey {
+    kind: NodeKind,
+    range: TextRange,
+}
+
+impl NodeKey {
+    pub(super) fn from_node<'a, N>(node: N) -> Self
+    where
+        N: Into<AnyNodeRef<'a>>,
+    {
+        let node = node.into();
+        NodeKey {
+            kind: node.kind(),
+            range: node.range(),
+        }
+    }
+}

crates/red_knot_python_semantic/src/program.rs

@@ -0,0 +1,100 @@
+use crate::python_version::PythonVersion;
+use anyhow::Context;
+use salsa::Durability;
+use salsa::Setter;
+
+use ruff_db::system::{SystemPath, SystemPathBuf};
+
+use crate::module_resolver::SearchPaths;
+use crate::Db;
+
+#[salsa::input(singleton)]
+pub struct Program {
+    pub target_version: PythonVersion,
+
+    #[return_ref]
+    pub(crate) search_paths: SearchPaths,
+}
+
+impl Program {
+    pub fn from_settings(db: &dyn Db, settings: &ProgramSettings) -> anyhow::Result<Self> {
+        let ProgramSettings {
+            target_version,
+            search_paths,
+        } = settings;
+
+        tracing::info!("Target version: Python {target_version}");
+
+        let search_paths = SearchPaths::from_settings(db, search_paths)
+            .with_context(|| "Invalid search path settings")?;
+
+        Ok(Program::builder(settings.target_version, search_paths)
+            .durability(Durability::HIGH)
+            .new(db))
+    }
+
+    pub fn update_search_paths(
+        self,
+        db: &mut dyn Db,
+        search_path_settings: &SearchPathSettings,
+    ) -> anyhow::Result<()> {
+        let search_paths = SearchPaths::from_settings(db, search_path_settings)?;
+
+        if self.search_paths(db) != &search_paths {
+            tracing::debug!("Update search paths");
+            self.set_search_paths(db).to(search_paths);
+        }
+
+        Ok(())
+    }
+
+    pub fn custom_stdlib_search_path(self, db: &dyn Db) -> Option<&SystemPath> {
+        self.search_paths(db).custom_stdlib()
+    }
+}
+
+#[derive(Clone, Debug, Eq, PartialEq)]
+pub struct ProgramSettings {
+    pub target_version: PythonVersion,
+    pub search_paths: SearchPathSettings,
+}
+
+/// Configures the search paths for module resolution.
+#[derive(Eq, PartialEq, Debug, Clone)]
+pub struct SearchPathSettings {
+    /// List of user-provided paths that should take first priority in the module resolution.
+    /// Examples in other type checkers are mypy's MYPYPATH environment variable,
+    /// or pyright's stubPath configuration setting.
+    pub extra_paths: Vec<SystemPathBuf>,
+
+    /// The root of the workspace, used for finding first-party modules.
+    pub src_root: SystemPathBuf,
+
+    /// Optional path to a "custom typeshed" directory on disk for us to use for standard-library types.
+    /// If this is not provided, we will fallback to our vendored typeshed stubs for the stdlib,
+    /// bundled as a zip file in the binary
+    pub custom_typeshed: Option<SystemPathBuf>,
+
+    /// The path to the user's `site-packages` directory, where third-party packages from ``PyPI`` are installed.
+    pub site_packages: SitePackages,
+}
+
+impl SearchPathSettings {
+    pub fn new(src_root: SystemPathBuf) -> Self {
+        Self {
+            src_root,
+            extra_paths: vec![],
+            custom_typeshed: None,
+            site_packages: SitePackages::Known(vec![]),
+        }
+    }
+}
+
+#[derive(Debug, Clone, Eq, PartialEq)]
+pub enum SitePackages {
+    Derived {
+        venv_path: SystemPathBuf,
+    },
+    /// Resolved site packages paths
+    Known(Vec<SystemPathBuf>),
+}

crates/red_knot_python_semantic/src/python_version.rs

@@ -0,0 +1,69 @@
+use std::fmt;
+
+/// Representation of a Python version.
+///
+/// Unlike the `TargetVersion` enums in the CLI crates,
+/// this does not necessarily represent a Python version that we actually support.
+#[derive(Debug, Clone, Copy, Eq, PartialEq, Ord, PartialOrd, Hash)]
+pub struct PythonVersion {
+    pub major: u8,
+    pub minor: u8,
+}
+
+impl PythonVersion {
+    pub const PY37: PythonVersion = PythonVersion { major: 3, minor: 7 };
+    pub const PY38: PythonVersion = PythonVersion { major: 3, minor: 8 };
+    pub const PY39: PythonVersion = PythonVersion { major: 3, minor: 9 };
+    pub const PY310: PythonVersion = PythonVersion {
+        major: 3,
+        minor: 10,
+    };
+    pub const PY311: PythonVersion = PythonVersion {
+        major: 3,
+        minor: 11,
+    };
+    pub const PY312: PythonVersion = PythonVersion {
+        major: 3,
+        minor: 12,
+    };
+    pub const PY313: PythonVersion = PythonVersion {
+        major: 3,
+        minor: 13,
+    };
+
+    pub fn free_threaded_build_available(self) -> bool {
+        self >= PythonVersion::PY313
+    }
+}
+
+impl Default for PythonVersion {
+    fn default() -> Self {
+        Self::PY38
+    }
+}
+
+impl TryFrom<(&str, &str)> for PythonVersion {
+    type Error = std::num::ParseIntError;
+
+    fn try_from(value: (&str, &str)) -> Result<Self, Self::Error> {
+        let (major, minor) = value;
+        Ok(Self {
+            major: major.parse()?,
+            minor: minor.parse()?,
+        })
+    }
+}
+
+impl From<(u8, u8)> for PythonVersion {
+    fn from(value: (u8, u8)) -> Self {
+        let (major, minor) = value;
+        Self { major, minor }
+    }
+}
+
+impl fmt::Display for PythonVersion {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        let PythonVersion { major, minor } = self;
+        write!(f, "{major}.{minor}")
+    }
+}

crates/red_knot_python_semantic/src/semantic_index.rs

@@ -1,44 +1,34 @@
 use std::iter::FusedIterator;
 use std::sync::Arc;
 
+use rustc_hash::{FxBuildHasher, FxHashMap};
+use salsa::plumbing::AsId;
+
 use ruff_db::files::File;
 use ruff_db::parsed::parsed_module;
 use ruff_index::{IndexSlice, IndexVec};
 
-use ruff_python_parser::semantic_errors::SemanticSyntaxError;
-use rustc_hash::{FxBuildHasher, FxHashMap, FxHashSet};
-use salsa::Update;
-use salsa::plumbing::AsId;
-
-use crate::Db;
-use crate::module_name::ModuleName;
-use crate::node_key::NodeKey;
-use crate::semantic_index::ast_ids::AstIds;
 use crate::semantic_index::ast_ids::node_key::ExpressionNodeKey;
+use crate::semantic_index::ast_ids::AstIds;
 use crate::semantic_index::builder::SemanticIndexBuilder;
-use crate::semantic_index::definition::{Definition, DefinitionNodeKey, Definitions};
+use crate::semantic_index::definition::{Definition, DefinitionNodeKey};
 use crate::semantic_index::expression::Expression;
-use crate::semantic_index::narrowing_constraints::ScopedNarrowingConstraint;
 use crate::semantic_index::symbol::{
-    FileScopeId, NodeWithScopeKey, NodeWithScopeRef, Scope, ScopeId, ScopeKind, ScopedSymbolId,
-    SymbolTable,
+    FileScopeId, NodeWithScopeKey, NodeWithScopeRef, Scope, ScopeId, ScopedSymbolId, SymbolTable,
 };
-use crate::semantic_index::use_def::{EagerSnapshotKey, ScopedEagerSnapshotId, UseDefMap};
+use crate::semantic_index::use_def::UseDefMap;
+use crate::Db;
 
 pub mod ast_ids;
 mod builder;
+pub(crate) mod constraint;
 pub mod definition;
 pub mod expression;
-pub(crate) mod narrowing_constraints;
-pub(crate) mod predicate;
-mod re_exports;
 pub mod symbol;
 mod use_def;
-mod visibility_constraints;
 
 pub(crate) use self::use_def::{
-    BindingWithConstraints, BindingWithConstraintsIterator, DeclarationWithConstraint,
-    DeclarationsIterator,
+    BindingWithConstraints, BindingWithConstraintsIterator, DeclarationsIterator,
 };
 
 type SymbolMap = hashbrown::HashMap<ScopedSymbolId, (), FxBuildHasher>;
@@ -46,9 +36,9 @@ type SymbolMap = hashbrown::HashMap<ScopedSymbolId, (), FxBuildHasher>;
 /// Returns the semantic index for `file`.
 ///
 /// Prefer using [`symbol_table`] when working with symbols from a single scope.
-#[salsa::tracked(returns(ref), no_eq)]
+#[salsa::tracked(return_ref, no_eq)]
 pub(crate) fn semantic_index(db: &dyn Db, file: File) -> SemanticIndex<'_> {
-    let _span = tracing::trace_span!("semantic_index", ?file).entered();
+    let _span = tracing::trace_span!("semantic_index", file = %file.path(db)).entered();
 
     let parsed = parsed_module(db.upcast(), file);
 
@@ -60,104 +50,45 @@ pub(crate) fn semantic_index(db: &dyn Db, file: File) -> SemanticIndex<'_> {
 /// Using [`symbol_table`] over [`semantic_index`] has the advantage that
 /// Salsa can avoid invalidating dependent queries if this scope's symbol table
 /// is unchanged.
-#[salsa::tracked(returns(deref))]
+#[salsa::tracked]
 pub(crate) fn symbol_table<'db>(db: &'db dyn Db, scope: ScopeId<'db>) -> Arc<SymbolTable> {
     let file = scope.file(db);
-    let _span = tracing::trace_span!("symbol_table", scope=?scope.as_id(), ?file).entered();
+    let _span =
+        tracing::trace_span!("symbol_table", scope=?scope.as_id(), file=%file.path(db)).entered();
     let index = semantic_index(db, file);
 
     index.symbol_table(scope.file_scope_id(db))
 }
 
-/// Returns the set of modules that are imported anywhere in `file`.
-///
-/// This set only considers `import` statements, not `from...import` statements, because:
-///
-///   - In `from foo import bar`, we cannot determine whether `foo.bar` is a submodule (and is
-///     therefore imported) without looking outside the content of this file.  (We could turn this
-///     into a _potentially_ imported modules set, but that would change how it's used in our type
-///     inference logic.)
-///
-///   - We cannot resolve relative imports (which aren't allowed in `import` statements) without
-///     knowing the name of the current module, and whether it's a package.
-#[salsa::tracked(returns(deref))]
-pub(crate) fn imported_modules<'db>(db: &'db dyn Db, file: File) -> Arc<FxHashSet<ModuleName>> {
-    semantic_index(db, file).imported_modules.clone()
-}
-
 /// Returns the use-def map for a specific `scope`.
 ///
 /// Using [`use_def_map`] over [`semantic_index`] has the advantage that
 /// Salsa can avoid invalidating dependent queries if this scope's use-def map
 /// is unchanged.
-#[salsa::tracked(returns(deref))]
+#[salsa::tracked]
 pub(crate) fn use_def_map<'db>(db: &'db dyn Db, scope: ScopeId<'db>) -> Arc<UseDefMap<'db>> {
     let file = scope.file(db);
-    let _span = tracing::trace_span!("use_def_map", scope=?scope.as_id(), ?file).entered();
+    let _span =
+        tracing::trace_span!("use_def_map", scope=?scope.as_id(), file=%file.path(db)).entered();
     let index = semantic_index(db, file);
 
     index.use_def_map(scope.file_scope_id(db))
 }
 
-/// Returns all attribute assignments (and their method scope IDs) for a specific class body scope.
-/// Only call this when doing type inference on the same file as `class_body_scope`, otherwise it
-/// introduces a direct dependency on that file's AST.
-pub(crate) fn attribute_assignments<'db, 's>(
-    db: &'db dyn Db,
-    class_body_scope: ScopeId<'db>,
-    name: &'s str,
-) -> impl Iterator<Item = (BindingWithConstraintsIterator<'db, 'db>, FileScopeId)> + use<'s, 'db> {
-    let file = class_body_scope.file(db);
-    let index = semantic_index(db, file);
-    let class_scope_id = class_body_scope.file_scope_id(db);
-
-    ChildrenIter::new(index, class_scope_id).filter_map(|(child_scope_id, scope)| {
-        let (function_scope_id, function_scope) =
-            if scope.node().scope_kind() == ScopeKind::Annotation {
-                // This could be a generic method with a type-params scope.
-                // Go one level deeper to find the function scope. The first
-                // descendant is the (potential) function scope.
-                let function_scope_id = scope.descendants().start;
-                (function_scope_id, index.scope(function_scope_id))
-            } else {
-                (child_scope_id, scope)
-            };
-
-        function_scope.node().as_function()?;
-        let attribute_table = index.instance_attribute_table(function_scope_id);
-        let symbol = attribute_table.symbol_id_by_name(name)?;
-        let use_def = &index.use_def_maps[function_scope_id];
-        Some((
-            use_def.instance_attribute_bindings(symbol),
-            function_scope_id,
-        ))
-    })
-}
-
 /// Returns the module global scope of `file`.
 #[salsa::tracked]
 pub(crate) fn global_scope(db: &dyn Db, file: File) -> ScopeId<'_> {
-    let _span = tracing::trace_span!("global_scope", ?file).entered();
+    let _span = tracing::trace_span!("global_scope", file = %file.path(db)).entered();
 
     FileScopeId::global().to_scope_id(db, file)
 }
 
-pub(crate) enum EagerSnapshotResult<'map, 'db> {
-    FoundConstraint(ScopedNarrowingConstraint),
-    FoundBindings(BindingWithConstraintsIterator<'map, 'db>),
-    NotFound,
-    NoLongerInEagerContext,
-}
-
 /// The symbol tables and use-def maps for all scopes in a file.
-#[derive(Debug, Update)]
+#[derive(Debug)]
 pub(crate) struct SemanticIndex<'db> {
     /// List of all symbol tables in this file, indexed by scope.
     symbol_tables: IndexVec<FileScopeId, Arc<SymbolTable>>,
 
-    /// List of all instance attribute tables in this file, indexed by scope.
-    instance_attribute_tables: IndexVec<FileScopeId, SymbolTable>,
-
     /// List of all scopes in this file.
     scopes: IndexVec<FileScopeId, Scope>,
 
@@ -165,7 +96,7 @@ pub(crate) struct SemanticIndex<'db> {
     scopes_by_expression: FxHashMap<ExpressionNodeKey, FileScopeId>,
 
     /// Map from a node creating a definition to its definition.
-    definitions_by_node: FxHashMap<DefinitionNodeKey, Definitions<'db>>,
+    definitions_by_node: FxHashMap<DefinitionNodeKey, Definition<'db>>,
 
     /// Map from a standalone expression to its [`Expression`] ingredient.
     expressions_by_node: FxHashMap<ExpressionNodeKey, Expression<'db>>,
@@ -176,9 +107,6 @@ pub(crate) struct SemanticIndex<'db> {
     /// Map from the file-local [`FileScopeId`] to the salsa-ingredient [`ScopeId`].
     scope_ids_by_scope: IndexVec<FileScopeId, ScopeId<'db>>,
 
-    /// Map from the file-local [`FileScopeId`] to the set of explicit-global symbols it contains.
-    globals_by_scope: FxHashMap<FileScopeId, FxHashSet<ScopedSymbolId>>,
-
     /// Use-def map for each scope in this file.
     use_def_maps: IndexVec<FileScopeId, Arc<UseDefMap<'db>>>,
 
@@ -188,20 +116,8 @@ pub(crate) struct SemanticIndex<'db> {
     /// changing a file invalidates all dependents.
     ast_ids: IndexVec<FileScopeId, AstIds>,
 
-    /// The set of modules that are imported anywhere within this file.
-    imported_modules: Arc<FxHashSet<ModuleName>>,
-
     /// Flags about the global scope (code usage impacting inference)
     has_future_annotations: bool,
-
-    /// Map of all of the eager snapshots that appear in this file.
-    eager_snapshots: FxHashMap<EagerSnapshotKey, ScopedEagerSnapshotId>,
-
-    /// List of all semantic syntax errors in this file.
-    semantic_syntax_errors: Vec<SemanticSyntaxError>,
-
-    /// Set of all generator functions in this file.
-    generator_functions: FxHashSet<FileScopeId>,
 }
 
 impl<'db> SemanticIndex<'db> {
@@ -209,31 +125,23 @@ impl<'db> SemanticIndex<'db> {
     ///
     /// Use the Salsa cached [`symbol_table()`] query if you only need the
     /// symbol table for a single scope.
-    #[track_caller]
     pub(super) fn symbol_table(&self, scope_id: FileScopeId) -> Arc<SymbolTable> {
         self.symbol_tables[scope_id].clone()
     }
 
-    pub(super) fn instance_attribute_table(&self, scope_id: FileScopeId) -> &SymbolTable {
-        &self.instance_attribute_tables[scope_id]
-    }
-
     /// Returns the use-def map for a specific scope.
     ///
     /// Use the Salsa cached [`use_def_map()`] query if you only need the
     /// use-def map for a single scope.
-    #[track_caller]
     pub(super) fn use_def_map(&self, scope_id: FileScopeId) -> Arc<UseDefMap> {
         self.use_def_maps[scope_id].clone()
     }
 
-    #[track_caller]
     pub(crate) fn ast_ids(&self, scope_id: FileScopeId) -> &AstIds {
         &self.ast_ids[scope_id]
     }
 
     /// Returns the ID of the `expression`'s enclosing scope.
-    #[track_caller]
     pub(crate) fn expression_scope_id(
         &self,
         expression: impl Into<ExpressionNodeKey>,
@@ -243,13 +151,11 @@ impl<'db> SemanticIndex<'db> {
 
     /// Returns the [`Scope`] of the `expression`'s enclosing scope.
     #[allow(unused)]
-    #[track_caller]
     pub(crate) fn expression_scope(&self, expression: impl Into<ExpressionNodeKey>) -> &Scope {
         &self.scopes[self.expression_scope_id(expression)]
     }
 
     /// Returns the [`Scope`] with the given id.
-    #[track_caller]
     pub(crate) fn scope(&self, id: FileScopeId) -> &Scope {
         &self.scopes[id]
     }
@@ -258,68 +164,22 @@ impl<'db> SemanticIndex<'db> {
         self.scope_ids_by_scope.iter().copied()
     }
 
-    pub(crate) fn symbol_is_global_in_scope(
-        &self,
-        symbol: ScopedSymbolId,
-        scope: FileScopeId,
-    ) -> bool {
-        self.globals_by_scope
-            .get(&scope)
-            .is_some_and(|globals| globals.contains(&symbol))
-    }
-
     /// Returns the id of the parent scope.
     pub(crate) fn parent_scope_id(&self, scope_id: FileScopeId) -> Option<FileScopeId> {
         let scope = self.scope(scope_id);
-        scope.parent()
+        scope.parent
     }
 
     /// Returns the parent scope of `scope_id`.
-    #[expect(unused)]
-    #[track_caller]
+    #[allow(unused)]
     pub(crate) fn parent_scope(&self, scope_id: FileScopeId) -> Option<&Scope> {
         Some(&self.scopes[self.parent_scope_id(scope_id)?])
     }
 
-    fn is_scope_reachable(&self, db: &'db dyn Db, scope_id: FileScopeId) -> bool {
-        self.parent_scope_id(scope_id)
-            .is_none_or(|parent_scope_id| {
-                if !self.is_scope_reachable(db, parent_scope_id) {
-                    return false;
-                }
-
-                let parent_use_def = self.use_def_map(parent_scope_id);
-                let reachability = self.scope(scope_id).reachability();
-
-                parent_use_def.is_reachable(db, reachability)
-            })
-    }
-
-    /// Returns true if a given AST node is reachable from the start of the scope. For example,
-    /// in the following code, expression `2` is reachable, but expressions `1` and `3` are not:
-    /// ```py
-    /// def f():
-    ///     x = 1
-    ///     if False:
-    ///         x  # 1
-    ///     x  # 2
-    ///     return
-    ///     x  # 3
-    /// ```
-    pub(crate) fn is_node_reachable(
-        &self,
-        db: &'db dyn crate::Db,
-        scope_id: FileScopeId,
-        node_key: NodeKey,
-    ) -> bool {
-        self.is_scope_reachable(db, scope_id)
-            && self.use_def_map(scope_id).is_node_reachable(db, node_key)
-    }
-
     /// Returns an iterator over the descendent scopes of `scope`.
     #[allow(unused)]
-    pub(crate) fn descendent_scopes(&self, scope: FileScopeId) -> DescendantsIter {
-        DescendantsIter::new(self, scope)
+    pub(crate) fn descendent_scopes(&self, scope: FileScopeId) -> DescendentsIter {
+        DescendentsIter::new(self, scope)
     }
 
     /// Returns an iterator over the direct child scopes of `scope`.
@@ -329,48 +189,23 @@ impl<'db> SemanticIndex<'db> {
     }
 
     /// Returns an iterator over all ancestors of `scope`, starting with `scope` itself.
+    #[allow(unused)]
     pub(crate) fn ancestor_scopes(&self, scope: FileScopeId) -> AncestorsIter {
         AncestorsIter::new(self, scope)
     }
 
-    /// Returns the [`definition::Definition`] salsa ingredient(s) for `definition_key`.
-    ///
-    /// There will only ever be >1 `Definition` associated with a `definition_key`
-    /// if the definition is created by a wildcard (`*`) import.
-    #[track_caller]
-    pub(crate) fn definitions(
+    /// Returns the [`Definition`] salsa ingredient for `definition_key`.
+    pub(crate) fn definition(
         &self,
         definition_key: impl Into<DefinitionNodeKey>,
-    ) -> &Definitions<'db> {
-        &self.definitions_by_node[&definition_key.into()]
-    }
-
-    /// Returns the [`definition::Definition`] salsa ingredient for `definition_key`.
-    ///
-    /// ## Panics
-    ///
-    /// If the number of definitions associated with the key is not exactly 1 and
-    /// the `debug_assertions` feature is enabled, this method will panic.
-    #[track_caller]
-    pub(crate) fn expect_single_definition(
-        &self,
-        definition_key: impl Into<DefinitionNodeKey> + std::fmt::Debug + Copy,
     ) -> Definition<'db> {
-        let definitions = self.definitions(definition_key);
-        debug_assert_eq!(
-            definitions.len(),
-            1,
-            "Expected exactly one definition to be associated with AST node {definition_key:?} but found {}",
-            definitions.len()
-        );
-        definitions[0]
+        self.definitions_by_node[&definition_key.into()]
     }
 
     /// Returns the [`Expression`] ingredient for an expression node.
     /// Panics if we have no expression ingredient for that node. We can only call this method for
     /// standalone-inferable expressions, which we call `add_standalone_expression` for in
     /// [`SemanticIndexBuilder`].
-    #[track_caller]
     pub(crate) fn expression(
         &self,
         expression_key: impl Into<ExpressionNodeKey>,
@@ -378,27 +213,8 @@ impl<'db> SemanticIndex<'db> {
         self.expressions_by_node[&expression_key.into()]
     }
 
-    pub(crate) fn try_expression(
-        &self,
-        expression_key: impl Into<ExpressionNodeKey>,
-    ) -> Option<Expression<'db>> {
-        self.expressions_by_node
-            .get(&expression_key.into())
-            .copied()
-    }
-
-    pub(crate) fn is_standalone_expression(
-        &self,
-        expression_key: impl Into<ExpressionNodeKey>,
-    ) -> bool {
-        self.expressions_by_node
-            .contains_key(&expression_key.into())
-    }
-
-    /// Returns the id of the scope that `node` creates.
-    /// This is different from [`definition::Definition::scope`] which
+    /// Returns the id of the scope that `node` creates. This is different from [`Definition::scope`] which
     /// returns the scope in which that definition is defined in.
-    #[track_caller]
     pub(crate) fn node_scope(&self, node: NodeWithScopeRef) -> FileScopeId {
         self.scopes_by_node[&node.node_key()]
     }
@@ -408,44 +224,6 @@ impl<'db> SemanticIndex<'db> {
     pub(super) fn has_future_annotations(&self) -> bool {
         self.has_future_annotations
     }
-
-    /// Returns
-    /// * `NoLongerInEagerContext` if the nested scope is no longer in an eager context
-    ///   (that is, not every scope that will be traversed is eager).
-    /// *  an iterator of bindings for a particular nested eager scope reference if the bindings exist.
-    /// *  a narrowing constraint if there are no bindings, but there is a narrowing constraint for an outer scope symbol.
-    /// * `NotFound` if the narrowing constraint / bindings do not exist in the nested eager scope.
-    pub(crate) fn eager_snapshot(
-        &self,
-        enclosing_scope: FileScopeId,
-        symbol: &str,
-        nested_scope: FileScopeId,
-    ) -> EagerSnapshotResult<'_, 'db> {
-        for (ancestor_scope_id, ancestor_scope) in self.ancestor_scopes(nested_scope) {
-            if ancestor_scope_id == enclosing_scope {
-                break;
-            }
-            if !ancestor_scope.is_eager() {
-                return EagerSnapshotResult::NoLongerInEagerContext;
-            }
-        }
-        let Some(symbol_id) = self.symbol_tables[enclosing_scope].symbol_id_by_name(symbol) else {
-            return EagerSnapshotResult::NotFound;
-        };
-        let key = EagerSnapshotKey {
-            enclosing_scope,
-            enclosing_symbol: symbol_id,
-            nested_scope,
-        };
-        let Some(id) = self.eager_snapshots.get(&key) else {
-            return EagerSnapshotResult::NotFound;
-        };
-        self.use_def_maps[enclosing_scope].eager_snapshot(*id)
-    }
-
-    pub(crate) fn semantic_syntax_errors(&self) -> &[SemanticSyntaxError] {
-        &self.semantic_syntax_errors
-    }
 }
 
 pub struct AncestorsIter<'a> {
@@ -468,7 +246,7 @@ impl<'a> Iterator for AncestorsIter<'a> {
     fn next(&mut self) -> Option<Self::Item> {
         let current_id = self.next_id?;
         let current = &self.scopes[current_id];
-        self.next_id = current.parent();
+        self.next_id = current.parent;
 
         Some((current_id, current))
     }
@@ -476,55 +254,55 @@ impl<'a> Iterator for AncestorsIter<'a> {
 
 impl FusedIterator for AncestorsIter<'_> {}
 
-pub struct DescendantsIter<'a> {
+pub struct DescendentsIter<'a> {
     next_id: FileScopeId,
-    descendants: std::slice::Iter<'a, Scope>,
+    descendents: std::slice::Iter<'a, Scope>,
 }
 
-impl<'a> DescendantsIter<'a> {
+impl<'a> DescendentsIter<'a> {
     fn new(symbol_table: &'a SemanticIndex, scope_id: FileScopeId) -> Self {
         let scope = &symbol_table.scopes[scope_id];
-        let scopes = &symbol_table.scopes[scope.descendants()];
+        let scopes = &symbol_table.scopes[scope.descendents.clone()];
 
         Self {
             next_id: scope_id + 1,
-            descendants: scopes.iter(),
+            descendents: scopes.iter(),
         }
     }
 }
 
-impl<'a> Iterator for DescendantsIter<'a> {
+impl<'a> Iterator for DescendentsIter<'a> {
     type Item = (FileScopeId, &'a Scope);
 
     fn next(&mut self) -> Option<Self::Item> {
-        let descendant = self.descendants.next()?;
+        let descendent = self.descendents.next()?;
         let id = self.next_id;
         self.next_id = self.next_id + 1;
 
-        Some((id, descendant))
+        Some((id, descendent))
     }
 
     fn size_hint(&self) -> (usize, Option<usize>) {
-        self.descendants.size_hint()
+        self.descendents.size_hint()
     }
 }
 
-impl FusedIterator for DescendantsIter<'_> {}
+impl FusedIterator for DescendentsIter<'_> {}
 
-impl ExactSizeIterator for DescendantsIter<'_> {}
+impl ExactSizeIterator for DescendentsIter<'_> {}
 
 pub struct ChildrenIter<'a> {
     parent: FileScopeId,
-    descendants: DescendantsIter<'a>,
+    descendents: DescendentsIter<'a>,
 }
 
 impl<'a> ChildrenIter<'a> {
     fn new(module_symbol_table: &'a SemanticIndex, parent: FileScopeId) -> Self {
-        let descendants = DescendantsIter::new(module_symbol_table, parent);
+        let descendents = DescendentsIter::new(module_symbol_table, parent);
 
         Self {
             parent,
-            descendants,
+            descendents,
         }
     }
 }
@@ -533,8 +311,8 @@ impl<'a> Iterator for ChildrenIter<'a> {
     type Item = (FileScopeId, &'a Scope);
 
     fn next(&mut self) -> Option<Self::Item> {
-        self.descendants
-            .find(|(_, scope)| scope.parent() == Some(self.parent))
+        self.descendents
+            .find(|(_, scope)| scope.parent == Some(self.parent))
     }
 }
 
@@ -542,13 +320,13 @@ impl FusedIterator for ChildrenIter<'_> {}
 
 #[cfg(test)]
 mod tests {
-    use ruff_db::files::{File, system_path_to_file};
+    use ruff_db::files::{system_path_to_file, File};
     use ruff_db::parsed::parsed_module;
-    use ruff_python_ast::{self as ast};
+    use ruff_db::system::DbWithTestSystem;
+    use ruff_python_ast as ast;
     use ruff_text_size::{Ranged, TextRange};
 
-    use crate::Db;
-    use crate::db::tests::{TestDb, TestDbBuilder};
+    use crate::db::tests::TestDb;
     use crate::semantic_index::ast_ids::{HasScopedUseId, ScopedUseId};
     use crate::semantic_index::definition::{Definition, DefinitionKind};
     use crate::semantic_index::symbol::{
@@ -556,16 +334,19 @@ mod tests {
     };
     use crate::semantic_index::use_def::UseDefMap;
     use crate::semantic_index::{global_scope, semantic_index, symbol_table, use_def_map};
+    use crate::Db;
 
     impl UseDefMap<'_> {
         fn first_public_binding(&self, symbol: ScopedSymbolId) -> Option<Definition<'_>> {
             self.public_bindings(symbol)
-                .find_map(|constrained_binding| constrained_binding.binding)
+                .next()
+                .map(|constrained_binding| constrained_binding.binding)
         }
 
         fn first_binding_at_use(&self, use_id: ScopedUseId) -> Option<Definition<'_>> {
             self.bindings_at_use(use_id)
-                .find_map(|constrained_binding| constrained_binding.binding)
+                .next()
+                .map(|constrained_binding| constrained_binding.binding)
         }
     }
 
@@ -574,15 +355,11 @@ mod tests {
         file: File,
     }
 
-    fn test_case(content: &str) -> TestCase {
-        const FILENAME: &str = "test.py";
+    fn test_case(content: impl ToString) -> TestCase {
+        let mut db = TestDb::new();
+        db.write_file("test.py", content).unwrap();
 
-        let db = TestDbBuilder::new()
-            .with_file(FILENAME, content)
-            .build()
-            .unwrap();
-
-        let file = system_path_to_file(&db, FILENAME).unwrap();
+        let file = system_path_to_file(&db, "test.py").unwrap();
 
         TestCase { db, file }
     }
@@ -599,7 +376,7 @@ mod tests {
         let TestCase { db, file } = test_case("");
         let global_table = symbol_table(&db, global_scope(&db, file));
 
-        let global_names = names(global_table);
+        let global_names = names(&global_table);
 
         assert_eq!(global_names, Vec::<&str>::new());
     }
@@ -609,7 +386,7 @@ mod tests {
         let TestCase { db, file } = test_case("x");
         let global_table = symbol_table(&db, global_scope(&db, file));
 
-        assert_eq!(names(global_table), vec!["x"]);
+        assert_eq!(names(&global_table), vec!["x"]);
     }
 
     #[test]
@@ -617,7 +394,7 @@ mod tests {
         let TestCase { db, file } = test_case("x: int");
         let global_table = symbol_table(&db, global_scope(&db, file));
 
-        assert_eq!(names(global_table), vec!["int", "x"]);
+        assert_eq!(names(&global_table), vec!["int", "x"]);
         // TODO record definition
     }
 
@@ -627,7 +404,7 @@ mod tests {
         let scope = global_scope(&db, file);
         let global_table = symbol_table(&db, scope);
 
-        assert_eq!(names(global_table), vec!["foo"]);
+        assert_eq!(names(&global_table), vec!["foo"]);
         let foo = global_table.symbol_id_by_name("foo").unwrap();
 
         let use_def = use_def_map(&db, scope);
@@ -640,7 +417,7 @@ mod tests {
         let TestCase { db, file } = test_case("import foo.bar");
         let global_table = symbol_table(&db, global_scope(&db, file));
 
-        assert_eq!(names(global_table), vec!["foo"]);
+        assert_eq!(names(&global_table), vec!["foo"]);
     }
 
     #[test]
@@ -648,7 +425,7 @@ mod tests {
         let TestCase { db, file } = test_case("import foo.bar as baz");
         let global_table = symbol_table(&db, global_scope(&db, file));
 
-        assert_eq!(names(global_table), vec!["baz"]);
+        assert_eq!(names(&global_table), vec!["baz"]);
     }
 
     #[test]
@@ -657,7 +434,7 @@ mod tests {
         let scope = global_scope(&db, file);
         let global_table = symbol_table(&db, scope);
 
-        assert_eq!(names(global_table), vec!["foo"]);
+        assert_eq!(names(&global_table), vec!["foo"]);
         assert!(
             global_table
                 .symbol_by_name("foo")
@@ -682,7 +459,7 @@ mod tests {
         let scope = global_scope(&db, file);
         let global_table = symbol_table(&db, scope);
 
-        assert_eq!(names(global_table), vec!["foo", "x"]);
+        assert_eq!(names(&global_table), vec!["foo", "x"]);
         assert!(
             global_table
                 .symbol_by_name("foo")
@@ -702,7 +479,7 @@ mod tests {
         let scope = global_scope(&db, file);
         let global_table = symbol_table(&db, scope);
 
-        assert_eq!(names(global_table), vec!["x"]);
+        assert_eq!(names(&global_table), vec!["x"]);
 
         let use_def = use_def_map(&db, scope);
         let binding = use_def
@@ -726,7 +503,7 @@ y = 2
         );
         let global_table = symbol_table(&db, global_scope(&db, file));
 
-        assert_eq!(names(global_table), vec!["C", "y"]);
+        assert_eq!(names(&global_table), vec!["C", "y"]);
 
         let index = semantic_index(&db, file);
 
@@ -798,7 +575,7 @@ def f(a: str, /, b: str, c: int = 1, *args, d: int = 2, **kwargs):
         let index = semantic_index(&db, file);
         let global_table = symbol_table(&db, global_scope(&db, file));
 
-        assert_eq!(names(global_table), vec!["str", "int", "f"]);
+        assert_eq!(names(&global_table), vec!["str", "int", "f"]);
 
         let [(function_scope_id, _function_scope)] = index
             .child_scopes(FileScopeId::global())
@@ -810,11 +587,24 @@ def f(a: str, /, b: str, c: int = 1, *args, d: int = 2, **kwargs):
         let function_table = index.symbol_table(function_scope_id);
         assert_eq!(
             names(&function_table),
-            vec!["a", "b", "c", "d", "args", "kwargs"],
+            vec!["a", "b", "c", "args", "d", "kwargs"],
         );
 
         let use_def = index.use_def_map(function_scope_id);
         for name in ["a", "b", "c", "d"] {
+            let binding = use_def
+                .first_public_binding(
+                    function_table
+                        .symbol_id_by_name(name)
+                        .expect("symbol exists"),
+                )
+                .unwrap();
+            assert!(matches!(
+                binding.kind(&db),
+                DefinitionKind::ParameterWithDefault(_)
+            ));
+        }
+        for name in ["args", "kwargs"] {
             let binding = use_def
                 .first_public_binding(
                     function_table
@@ -824,28 +614,6 @@ def f(a: str, /, b: str, c: int = 1, *args, d: int = 2, **kwargs):
                 .unwrap();
             assert!(matches!(binding.kind(&db), DefinitionKind::Parameter(_)));
         }
-        let args_binding = use_def
-            .first_public_binding(
-                function_table
-                    .symbol_id_by_name("args")
-                    .expect("symbol exists"),
-            )
-            .unwrap();
-        assert!(matches!(
-            args_binding.kind(&db),
-            DefinitionKind::VariadicPositionalParameter(_)
-        ));
-        let kwargs_binding = use_def
-            .first_public_binding(
-                function_table
-                    .symbol_id_by_name("kwargs")
-                    .expect("symbol exists"),
-            )
-            .unwrap();
-        assert!(matches!(
-            kwargs_binding.kind(&db),
-            DefinitionKind::VariadicKeywordParameter(_)
-        ));
     }
 
     #[test]
@@ -855,7 +623,7 @@ def f(a: str, /, b: str, c: int = 1, *args, d: int = 2, **kwargs):
         let index = semantic_index(&db, file);
         let global_table = symbol_table(&db, global_scope(&db, file));
 
-        assert!(names(global_table).is_empty());
+        assert!(names(&global_table).is_empty());
 
         let [(lambda_scope_id, _lambda_scope)] = index
             .child_scopes(FileScopeId::global())
@@ -867,38 +635,25 @@ def f(a: str, /, b: str, c: int = 1, *args, d: int = 2, **kwargs):
         let lambda_table = index.symbol_table(lambda_scope_id);
         assert_eq!(
             names(&lambda_table),
-            vec!["a", "b", "c", "d", "args", "kwargs"],
+            vec!["a", "b", "c", "args", "d", "kwargs"],
         );
 
         let use_def = index.use_def_map(lambda_scope_id);
         for name in ["a", "b", "c", "d"] {
+            let binding = use_def
+                .first_public_binding(lambda_table.symbol_id_by_name(name).expect("symbol exists"))
+                .unwrap();
+            assert!(matches!(
+                binding.kind(&db),
+                DefinitionKind::ParameterWithDefault(_)
+            ));
+        }
+        for name in ["args", "kwargs"] {
             let binding = use_def
                 .first_public_binding(lambda_table.symbol_id_by_name(name).expect("symbol exists"))
                 .unwrap();
             assert!(matches!(binding.kind(&db), DefinitionKind::Parameter(_)));
         }
-        let args_binding = use_def
-            .first_public_binding(
-                lambda_table
-                    .symbol_id_by_name("args")
-                    .expect("symbol exists"),
-            )
-            .unwrap();
-        assert!(matches!(
-            args_binding.kind(&db),
-            DefinitionKind::VariadicPositionalParameter(_)
-        ));
-        let kwargs_binding = use_def
-            .first_public_binding(
-                lambda_table
-                    .symbol_id_by_name("kwargs")
-                    .expect("symbol exists"),
-            )
-            .unwrap();
-        assert!(matches!(
-            kwargs_binding.kind(&db),
-            DefinitionKind::VariadicKeywordParameter(_)
-        ));
     }
 
     /// Test case to validate that the comprehension scope is correctly identified and that the target
@@ -987,7 +742,7 @@ def f(a: str, /, b: str, c: int = 1, *args, d: int = 2, **kwargs):
             panic!("expected generator definition")
         };
         let target = comprehension.target();
-        let name = target.as_name_expr().unwrap().id().as_str();
+        let name = target.id().as_str();
 
         assert_eq!(name, "x");
         assert_eq!(target.range(), TextRange::new(23.into(), 24.into()));
@@ -1106,10 +861,7 @@ def func():
         let global_table = index.symbol_table(FileScopeId::global());
 
         assert_eq!(names(&global_table), vec!["func"]);
-        let [
-            (func_scope1_id, func_scope_1),
-            (func_scope2_id, func_scope_2),
-        ] = index
+        let [(func_scope1_id, func_scope_1), (func_scope2_id, func_scope_2)] = index
             .child_scopes(FileScopeId::global())
             .collect::<Vec<_>>()[..]
         else {
@@ -1242,7 +994,7 @@ class C[T]:
         let ast::Expr::NumberLiteral(ast::ExprNumberLiteral {
             value: ast::Number::Int(num),
             ..
-        }) = assignment.value()
+        }) = &*assignment.assignment().value
         else {
             panic!("should be a number literal")
         };
@@ -1298,9 +1050,9 @@ def x():
 
         let index = semantic_index(&db, file);
 
-        let descendants = index.descendent_scopes(FileScopeId::global());
+        let descendents = index.descendent_scopes(FileScopeId::global());
         assert_eq!(
-            scope_names(descendants, &db, file),
+            scope_names(descendents, &db, file),
             vec!["Test", "foo", "bar", "baz", "x"]
         );
 
@@ -1347,10 +1099,8 @@ match subject:
 
         assert!(global_table.symbol_by_name("Foo").unwrap().is_used());
         assert_eq!(
-            names(global_table),
-            vec![
-                "subject", "a", "b", "c", "d", "e", "f", "g", "h", "Foo", "i", "j", "k", "l"
-            ]
+            names(&global_table),
+            vec!["subject", "a", "b", "c", "d", "e", "f", "g", "h", "Foo", "i", "j", "k", "l"]
         );
 
         let use_def = use_def_map(&db, global_scope_id);
@@ -1394,7 +1144,7 @@ match 1:
         let global_scope_id = global_scope(&db, file);
         let global_table = symbol_table(&db, global_scope_id);
 
-        assert_eq!(names(global_table), vec!["first", "second"]);
+        assert_eq!(names(&global_table), vec!["first", "second"]);
 
         let use_def = use_def_map(&db, global_scope_id);
         for (name, expected_index) in [("first", 0), ("second", 0)] {
@@ -1415,7 +1165,7 @@ match 1:
         let scope = global_scope(&db, file);
         let global_table = symbol_table(&db, scope);
 
-        assert_eq!(&names(global_table), &["a", "x"]);
+        assert_eq!(&names(&global_table), &["a", "x"]);
 
         let use_def = use_def_map(&db, scope);
         let binding = use_def
@@ -1431,7 +1181,7 @@ match 1:
         let scope = global_scope(&db, file);
         let global_table = symbol_table(&db, scope);
 
-        assert_eq!(&names(global_table), &["a", "x", "y"]);
+        assert_eq!(&names(&global_table), &["a", "x", "y"]);
 
         let use_def = use_def_map(&db, scope);
         let x_binding = use_def
@@ -1451,7 +1201,7 @@ match 1:
         let scope = global_scope(&db, file);
         let global_table = symbol_table(&db, scope);
 
-        assert_eq!(&names(global_table), &["e", "a", "b", "c", "d"]);
+        assert_eq!(&names(&global_table), &["e", "a", "b", "c", "d"]);
 
         let use_def = use_def_map(&db, scope);
         let binding = use_def

crates/red_knot_python_semantic/src/semantic_index/ast_ids.rs

@@ -2,12 +2,12 @@ use rustc_hash::FxHashMap;
 
 use ruff_index::newtype_index;
 use ruff_python_ast as ast;
-use ruff_python_ast::ExprRef;
+use ruff_python_ast::ExpressionRef;
 
-use crate::Db;
 use crate::semantic_index::ast_ids::node_key::ExpressionNodeKey;
 use crate::semantic_index::semantic_index;
 use crate::semantic_index::symbol::ScopeId;
+use crate::Db;
 
 /// AST ids for a single scope.
 ///
@@ -24,7 +24,7 @@ use crate::semantic_index::symbol::ScopeId;
 ///
 /// x = foo()
 /// ```
-#[derive(Debug, salsa::Update)]
+#[derive(Debug)]
 pub(crate) struct AstIds {
     /// Maps expressions to their expression id.
     expressions_map: FxHashMap<ExpressionNodeKey, ScopedExpressionId>,
@@ -34,10 +34,7 @@ pub(crate) struct AstIds {
 
 impl AstIds {
     fn expression_id(&self, key: impl Into<ExpressionNodeKey>) -> ScopedExpressionId {
-        let key = &key.into();
-        *self.expressions_map.get(key).unwrap_or_else(|| {
-            panic!("Could not find expression ID for {key:?}");
-        })
+        self.expressions_map[&key.into()]
     }
 
     fn use_id(&self, key: impl Into<ExpressionNodeKey>) -> ScopedUseId {
@@ -49,58 +46,56 @@ fn ast_ids<'db>(db: &'db dyn Db, scope: ScopeId) -> &'db AstIds {
     semantic_index(db, scope.file(db)).ast_ids(scope.file_scope_id(db))
 }
 
+pub trait HasScopedUseId {
+    /// The type of the ID uniquely identifying the use.
+    type Id: Copy;
+
+    /// Returns the ID that uniquely identifies the use in `scope`.
+    fn scoped_use_id(&self, db: &dyn Db, scope: ScopeId) -> Self::Id;
+}
+
 /// Uniquely identifies a use of a name in a [`crate::semantic_index::symbol::FileScopeId`].
 #[newtype_index]
 pub struct ScopedUseId;
 
-pub trait HasScopedUseId {
-    /// Returns the ID that uniquely identifies the use in `scope`.
-    fn scoped_use_id(&self, db: &dyn Db, scope: ScopeId) -> ScopedUseId;
-}
-
-impl HasScopedUseId for ast::Identifier {
-    fn scoped_use_id(&self, db: &dyn Db, scope: ScopeId) -> ScopedUseId {
-        let ast_ids = ast_ids(db, scope);
-        ast_ids.use_id(self)
-    }
-}
-
 impl HasScopedUseId for ast::ExprName {
-    fn scoped_use_id(&self, db: &dyn Db, scope: ScopeId) -> ScopedUseId {
-        let expression_ref = ExprRef::from(self);
+    type Id = ScopedUseId;
+
+    fn scoped_use_id(&self, db: &dyn Db, scope: ScopeId) -> Self::Id {
+        let expression_ref = ExpressionRef::from(self);
         expression_ref.scoped_use_id(db, scope)
     }
 }
 
-impl HasScopedUseId for ast::ExprRef<'_> {
-    fn scoped_use_id(&self, db: &dyn Db, scope: ScopeId) -> ScopedUseId {
+impl HasScopedUseId for ast::ExpressionRef<'_> {
+    type Id = ScopedUseId;
+
+    fn scoped_use_id(&self, db: &dyn Db, scope: ScopeId) -> Self::Id {
         let ast_ids = ast_ids(db, scope);
         ast_ids.use_id(*self)
     }
 }
 
+pub trait HasScopedAstId {
+    /// The type of the ID uniquely identifying the node.
+    type Id: Copy;
+
+    /// Returns the ID that uniquely identifies the node in `scope`.
+    fn scoped_ast_id(&self, db: &dyn Db, scope: ScopeId) -> Self::Id;
+}
+
 /// Uniquely identifies an [`ast::Expr`] in a [`crate::semantic_index::symbol::FileScopeId`].
 #[newtype_index]
-#[derive(salsa::Update)]
 pub struct ScopedExpressionId;
 
-pub trait HasScopedExpressionId {
-    /// Returns the ID that uniquely identifies the node in `scope`.
-    fn scoped_expression_id(&self, db: &dyn Db, scope: ScopeId) -> ScopedExpressionId;
-}
-
-impl<T: HasScopedExpressionId> HasScopedExpressionId for Box<T> {
-    fn scoped_expression_id(&self, db: &dyn Db, scope: ScopeId) -> ScopedExpressionId {
-        self.as_ref().scoped_expression_id(db, scope)
-    }
-}
-
 macro_rules! impl_has_scoped_expression_id {
     ($ty: ty) => {
-        impl HasScopedExpressionId for $ty {
-            fn scoped_expression_id(&self, db: &dyn Db, scope: ScopeId) -> ScopedExpressionId {
-                let expression_ref = ExprRef::from(self);
-                expression_ref.scoped_expression_id(db, scope)
+        impl HasScopedAstId for $ty {
+            type Id = ScopedExpressionId;
+
+            fn scoped_ast_id(&self, db: &dyn Db, scope: ScopeId) -> Self::Id {
+                let expression_ref = ExpressionRef::from(self);
+                expression_ref.scoped_ast_id(db, scope)
             }
         }
     };
@@ -140,20 +135,29 @@ impl_has_scoped_expression_id!(ast::ExprSlice);
 impl_has_scoped_expression_id!(ast::ExprIpyEscapeCommand);
 impl_has_scoped_expression_id!(ast::Expr);
 
-impl HasScopedExpressionId for ast::ExprRef<'_> {
-    fn scoped_expression_id(&self, db: &dyn Db, scope: ScopeId) -> ScopedExpressionId {
+impl HasScopedAstId for ast::ExpressionRef<'_> {
+    type Id = ScopedExpressionId;
+
+    fn scoped_ast_id(&self, db: &dyn Db, scope: ScopeId) -> Self::Id {
         let ast_ids = ast_ids(db, scope);
         ast_ids.expression_id(*self)
     }
 }
 
-#[derive(Debug, Default)]
+#[derive(Debug)]
 pub(super) struct AstIdsBuilder {
     expressions_map: FxHashMap<ExpressionNodeKey, ScopedExpressionId>,
     uses_map: FxHashMap<ExpressionNodeKey, ScopedUseId>,
 }
 
 impl AstIdsBuilder {
+    pub(super) fn new() -> Self {
+        Self {
+            expressions_map: FxHashMap::default(),
+            uses_map: FxHashMap::default(),
+        }
+    }
+
     /// Adds `expr` to the expression ids map and returns its id.
     pub(super) fn record_expression(&mut self, expr: &ast::Expr) -> ScopedExpressionId {
         let expression_id = self.expressions_map.len().into();
@@ -164,7 +168,7 @@ impl AstIdsBuilder {
     }
 
     /// Adds `expr` to the use ids map and returns its id.
-    pub(super) fn record_use(&mut self, expr: impl Into<ExpressionNodeKey>) -> ScopedUseId {
+    pub(super) fn record_use(&mut self, expr: &ast::Expr) -> ScopedUseId {
         let use_id = self.uses_map.len().into();
 
         self.uses_map.insert(expr.into(), use_id);
@@ -189,11 +193,11 @@ pub(crate) mod node_key {
 
     use crate::node_key::NodeKey;
 
-    #[derive(Copy, Clone, Eq, PartialEq, Hash, Debug, salsa::Update)]
+    #[derive(Copy, Clone, Eq, PartialEq, Hash, Debug)]
     pub(crate) struct ExpressionNodeKey(NodeKey);
 
-    impl From<ast::ExprRef<'_>> for ExpressionNodeKey {
-        fn from(value: ast::ExprRef<'_>) -> Self {
+    impl From<ast::ExpressionRef<'_>> for ExpressionNodeKey {
+        fn from(value: ast::ExpressionRef<'_>) -> Self {
             Self(NodeKey::from_node(value))
         }
     }
@@ -203,10 +207,4 @@ pub(crate) mod node_key {
             Self(NodeKey::from_node(value))
         }
     }
-
-    impl From<&ast::Identifier> for ExpressionNodeKey {
-        fn from(value: &ast::Identifier) -> Self {
-            Self(NodeKey::from_node(value))
-        }
-    }
 }

crates/red_knot_python_semantic/src/semantic_index/constraint.rs

@@ -0,0 +1,39 @@
+use ruff_db::files::File;
+use ruff_python_ast as ast;
+
+use crate::ast_node_ref::AstNodeRef;
+use crate::db::Db;
+use crate::semantic_index::expression::Expression;
+use crate::semantic_index::symbol::{FileScopeId, ScopeId};
+
+#[derive(Clone, Copy, Debug, PartialEq, Eq)]
+pub(crate) enum Constraint<'db> {
+    Expression(Expression<'db>),
+    Pattern(PatternConstraint<'db>),
+}
+
+#[salsa::tracked]
+pub(crate) struct PatternConstraint<'db> {
+    #[id]
+    pub(crate) file: File,
+
+    #[id]
+    pub(crate) file_scope: FileScopeId,
+
+    #[no_eq]
+    #[return_ref]
+    pub(crate) subject: AstNodeRef<ast::Expr>,
+
+    #[no_eq]
+    #[return_ref]
+    pub(crate) pattern: AstNodeRef<ast::Pattern>,
+
+    #[no_eq]
+    count: countme::Count<PatternConstraint<'static>>,
+}
+
+impl<'db> PatternConstraint<'db> {
+    pub(crate) fn scope(self, db: &'db dyn Db) -> ScopeId<'db> {
+        self.file_scope(db).to_scope_id(db, self.file(db))
+    }
+}

crates/red_knot_python_semantic/src/semantic_index/definition.rs

@@ -0,0 +1,631 @@
+use ruff_db::files::File;
+use ruff_db::parsed::ParsedModule;
+use ruff_python_ast as ast;
+
+use crate::ast_node_ref::AstNodeRef;
+use crate::module_resolver::file_to_module;
+use crate::node_key::NodeKey;
+use crate::semantic_index::symbol::{FileScopeId, ScopeId, ScopedSymbolId};
+use crate::Db;
+
+#[salsa::tracked]
+pub struct Definition<'db> {
+    /// The file in which the definition occurs.
+    #[id]
+    pub(crate) file: File,
+
+    /// The scope in which the definition occurs.
+    #[id]
+    pub(crate) file_scope: FileScopeId,
+
+    /// The symbol defined.
+    #[id]
+    pub(crate) symbol: ScopedSymbolId,
+
+    #[no_eq]
+    #[return_ref]
+    pub(crate) kind: DefinitionKind,
+
+    #[no_eq]
+    count: countme::Count<Definition<'static>>,
+}
+
+impl<'db> Definition<'db> {
+    pub(crate) fn scope(self, db: &'db dyn Db) -> ScopeId<'db> {
+        self.file_scope(db).to_scope_id(db, self.file(db))
+    }
+
+    pub(crate) fn category(self, db: &'db dyn Db) -> DefinitionCategory {
+        self.kind(db).category()
+    }
+
+    pub(crate) fn is_declaration(self, db: &'db dyn Db) -> bool {
+        self.kind(db).category().is_declaration()
+    }
+
+    pub(crate) fn is_binding(self, db: &'db dyn Db) -> bool {
+        self.kind(db).category().is_binding()
+    }
+
+    /// Return true if this is a symbol was defined in the `typing` or `typing_extensions` modules
+    pub(crate) fn is_typing_definition(self, db: &'db dyn Db) -> bool {
+        file_to_module(db, self.file(db)).is_some_and(|module| {
+            module.search_path().is_standard_library()
+                && matches!(&**module.name(), "typing" | "typing_extensions")
+        })
+    }
+}
+
+#[derive(Copy, Clone, Debug)]
+pub(crate) enum DefinitionNodeRef<'a> {
+    Import(&'a ast::Alias),
+    ImportFrom(ImportFromDefinitionNodeRef<'a>),
+    For(ForStmtDefinitionNodeRef<'a>),
+    Function(&'a ast::StmtFunctionDef),
+    Class(&'a ast::StmtClassDef),
+    NamedExpression(&'a ast::ExprNamed),
+    Assignment(AssignmentDefinitionNodeRef<'a>),
+    AnnotatedAssignment(&'a ast::StmtAnnAssign),
+    AugmentedAssignment(&'a ast::StmtAugAssign),
+    Comprehension(ComprehensionDefinitionNodeRef<'a>),
+    Parameter(ast::AnyParameterRef<'a>),
+    WithItem(WithItemDefinitionNodeRef<'a>),
+    MatchPattern(MatchPatternDefinitionNodeRef<'a>),
+    ExceptHandler(ExceptHandlerDefinitionNodeRef<'a>),
+}
+
+impl<'a> From<&'a ast::StmtFunctionDef> for DefinitionNodeRef<'a> {
+    fn from(node: &'a ast::StmtFunctionDef) -> Self {
+        Self::Function(node)
+    }
+}
+
+impl<'a> From<&'a ast::StmtClassDef> for DefinitionNodeRef<'a> {
+    fn from(node: &'a ast::StmtClassDef) -> Self {
+        Self::Class(node)
+    }
+}
+
+impl<'a> From<&'a ast::ExprNamed> for DefinitionNodeRef<'a> {
+    fn from(node: &'a ast::ExprNamed) -> Self {
+        Self::NamedExpression(node)
+    }
+}
+
+impl<'a> From<&'a ast::StmtAnnAssign> for DefinitionNodeRef<'a> {
+    fn from(node: &'a ast::StmtAnnAssign) -> Self {
+        Self::AnnotatedAssignment(node)
+    }
+}
+
+impl<'a> From<&'a ast::StmtAugAssign> for DefinitionNodeRef<'a> {
+    fn from(node: &'a ast::StmtAugAssign) -> Self {
+        Self::AugmentedAssignment(node)
+    }
+}
+
+impl<'a> From<&'a ast::Alias> for DefinitionNodeRef<'a> {
+    fn from(node_ref: &'a ast::Alias) -> Self {
+        Self::Import(node_ref)
+    }
+}
+
+impl<'a> From<ImportFromDefinitionNodeRef<'a>> for DefinitionNodeRef<'a> {
+    fn from(node_ref: ImportFromDefinitionNodeRef<'a>) -> Self {
+        Self::ImportFrom(node_ref)
+    }
+}
+
+impl<'a> From<ForStmtDefinitionNodeRef<'a>> for DefinitionNodeRef<'a> {
+    fn from(value: ForStmtDefinitionNodeRef<'a>) -> Self {
+        Self::For(value)
+    }
+}
+
+impl<'a> From<AssignmentDefinitionNodeRef<'a>> for DefinitionNodeRef<'a> {
+    fn from(node_ref: AssignmentDefinitionNodeRef<'a>) -> Self {
+        Self::Assignment(node_ref)
+    }
+}
+
+impl<'a> From<WithItemDefinitionNodeRef<'a>> for DefinitionNodeRef<'a> {
+    fn from(node_ref: WithItemDefinitionNodeRef<'a>) -> Self {
+        Self::WithItem(node_ref)
+    }
+}
+
+impl<'a> From<ComprehensionDefinitionNodeRef<'a>> for DefinitionNodeRef<'a> {
+    fn from(node: ComprehensionDefinitionNodeRef<'a>) -> Self {
+        Self::Comprehension(node)
+    }
+}
+
+impl<'a> From<ast::AnyParameterRef<'a>> for DefinitionNodeRef<'a> {
+    fn from(node: ast::AnyParameterRef<'a>) -> Self {
+        Self::Parameter(node)
+    }
+}
+
+impl<'a> From<MatchPatternDefinitionNodeRef<'a>> for DefinitionNodeRef<'a> {
+    fn from(node: MatchPatternDefinitionNodeRef<'a>) -> Self {
+        Self::MatchPattern(node)
+    }
+}
+
+#[derive(Copy, Clone, Debug)]
+pub(crate) struct ImportFromDefinitionNodeRef<'a> {
+    pub(crate) node: &'a ast::StmtImportFrom,
+    pub(crate) alias_index: usize,
+}
+
+#[derive(Copy, Clone, Debug)]
+pub(crate) struct AssignmentDefinitionNodeRef<'a> {
+    pub(crate) assignment: &'a ast::StmtAssign,
+    pub(crate) target: &'a ast::ExprName,
+}
+
+#[derive(Copy, Clone, Debug)]
+pub(crate) struct WithItemDefinitionNodeRef<'a> {
+    pub(crate) node: &'a ast::WithItem,
+    pub(crate) target: &'a ast::ExprName,
+}
+
+#[derive(Copy, Clone, Debug)]
+pub(crate) struct ForStmtDefinitionNodeRef<'a> {
+    pub(crate) iterable: &'a ast::Expr,
+    pub(crate) target: &'a ast::ExprName,
+    pub(crate) is_async: bool,
+}
+
+#[derive(Copy, Clone, Debug)]
+pub(crate) struct ExceptHandlerDefinitionNodeRef<'a> {
+    pub(crate) handler: &'a ast::ExceptHandlerExceptHandler,
+    pub(crate) is_star: bool,
+}
+
+#[derive(Copy, Clone, Debug)]
+pub(crate) struct ComprehensionDefinitionNodeRef<'a> {
+    pub(crate) iterable: &'a ast::Expr,
+    pub(crate) target: &'a ast::ExprName,
+    pub(crate) first: bool,
+    pub(crate) is_async: bool,
+}
+
+#[derive(Copy, Clone, Debug)]
+pub(crate) struct MatchPatternDefinitionNodeRef<'a> {
+    /// The outermost pattern node in which the identifier being defined occurs.
+    pub(crate) pattern: &'a ast::Pattern,
+    /// The identifier being defined.
+    pub(crate) identifier: &'a ast::Identifier,
+    /// The index of the identifier in the pattern when visiting the `pattern` node in evaluation
+    /// order.
+    pub(crate) index: u32,
+}
+
+impl DefinitionNodeRef<'_> {
+    #[allow(unsafe_code)]
+    pub(super) unsafe fn into_owned(self, parsed: ParsedModule) -> DefinitionKind {
+        match self {
+            DefinitionNodeRef::Import(alias) => {
+                DefinitionKind::Import(AstNodeRef::new(parsed, alias))
+            }
+            DefinitionNodeRef::ImportFrom(ImportFromDefinitionNodeRef { node, alias_index }) => {
+                DefinitionKind::ImportFrom(ImportFromDefinitionKind {
+                    node: AstNodeRef::new(parsed, node),
+                    alias_index,
+                })
+            }
+            DefinitionNodeRef::Function(function) => {
+                DefinitionKind::Function(AstNodeRef::new(parsed, function))
+            }
+            DefinitionNodeRef::Class(class) => {
+                DefinitionKind::Class(AstNodeRef::new(parsed, class))
+            }
+            DefinitionNodeRef::NamedExpression(named) => {
+                DefinitionKind::NamedExpression(AstNodeRef::new(parsed, named))
+            }
+            DefinitionNodeRef::Assignment(AssignmentDefinitionNodeRef { assignment, target }) => {
+                DefinitionKind::Assignment(AssignmentDefinitionKind {
+                    assignment: AstNodeRef::new(parsed.clone(), assignment),
+                    target: AstNodeRef::new(parsed, target),
+                })
+            }
+            DefinitionNodeRef::AnnotatedAssignment(assign) => {
+                DefinitionKind::AnnotatedAssignment(AstNodeRef::new(parsed, assign))
+            }
+            DefinitionNodeRef::AugmentedAssignment(augmented_assignment) => {
+                DefinitionKind::AugmentedAssignment(AstNodeRef::new(parsed, augmented_assignment))
+            }
+            DefinitionNodeRef::For(ForStmtDefinitionNodeRef {
+                iterable,
+                target,
+                is_async,
+            }) => DefinitionKind::For(ForStmtDefinitionKind {
+                iterable: AstNodeRef::new(parsed.clone(), iterable),
+                target: AstNodeRef::new(parsed, target),
+                is_async,
+            }),
+            DefinitionNodeRef::Comprehension(ComprehensionDefinitionNodeRef {
+                iterable,
+                target,
+                first,
+                is_async,
+            }) => DefinitionKind::Comprehension(ComprehensionDefinitionKind {
+                iterable: AstNodeRef::new(parsed.clone(), iterable),
+                target: AstNodeRef::new(parsed, target),
+                first,
+                is_async,
+            }),
+            DefinitionNodeRef::Parameter(parameter) => match parameter {
+                ast::AnyParameterRef::Variadic(parameter) => {
+                    DefinitionKind::Parameter(AstNodeRef::new(parsed, parameter))
+                }
+                ast::AnyParameterRef::NonVariadic(parameter) => {
+                    DefinitionKind::ParameterWithDefault(AstNodeRef::new(parsed, parameter))
+                }
+            },
+            DefinitionNodeRef::WithItem(WithItemDefinitionNodeRef { node, target }) => {
+                DefinitionKind::WithItem(WithItemDefinitionKind {
+                    node: AstNodeRef::new(parsed.clone(), node),
+                    target: AstNodeRef::new(parsed, target),
+                })
+            }
+            DefinitionNodeRef::MatchPattern(MatchPatternDefinitionNodeRef {
+                pattern,
+                identifier,
+                index,
+            }) => DefinitionKind::MatchPattern(MatchPatternDefinitionKind {
+                pattern: AstNodeRef::new(parsed.clone(), pattern),
+                identifier: AstNodeRef::new(parsed, identifier),
+                index,
+            }),
+            DefinitionNodeRef::ExceptHandler(ExceptHandlerDefinitionNodeRef {
+                handler,
+                is_star,
+            }) => DefinitionKind::ExceptHandler(ExceptHandlerDefinitionKind {
+                handler: AstNodeRef::new(parsed.clone(), handler),
+                is_star,
+            }),
+        }
+    }
+
+    pub(super) fn key(self) -> DefinitionNodeKey {
+        match self {
+            Self::Import(node) => node.into(),
+            Self::ImportFrom(ImportFromDefinitionNodeRef { node, alias_index }) => {
+                (&node.names[alias_index]).into()
+            }
+            Self::Function(node) => node.into(),
+            Self::Class(node) => node.into(),
+            Self::NamedExpression(node) => node.into(),
+            Self::Assignment(AssignmentDefinitionNodeRef {
+                assignment: _,
+                target,
+            }) => target.into(),
+            Self::AnnotatedAssignment(node) => node.into(),
+            Self::AugmentedAssignment(node) => node.into(),
+            Self::For(ForStmtDefinitionNodeRef {
+                iterable: _,
+                target,
+                is_async: _,
+            }) => target.into(),
+            Self::Comprehension(ComprehensionDefinitionNodeRef { target, .. }) => target.into(),
+            Self::Parameter(node) => match node {
+                ast::AnyParameterRef::Variadic(parameter) => parameter.into(),
+                ast::AnyParameterRef::NonVariadic(parameter) => parameter.into(),
+            },
+            Self::WithItem(WithItemDefinitionNodeRef { node: _, target }) => target.into(),
+            Self::MatchPattern(MatchPatternDefinitionNodeRef { identifier, .. }) => {
+                identifier.into()
+            }
+            Self::ExceptHandler(ExceptHandlerDefinitionNodeRef { handler, .. }) => handler.into(),
+        }
+    }
+}
+
+#[derive(Clone, Copy, Debug)]
+pub(crate) enum DefinitionCategory {
+    /// A Definition which binds a value to a name (e.g. `x = 1`).
+    Binding,
+    /// A Definition which declares the upper-bound of acceptable types for this name (`x: int`).
+    Declaration,
+    /// A Definition which both declares a type and binds a value (e.g. `x: int = 1`).
+    DeclarationAndBinding,
+}
+
+impl DefinitionCategory {
+    /// True if this definition establishes a "declared type" for the symbol.
+    ///
+    /// If so, any assignments reached by this definition are in error if they assign a value of a
+    /// type not assignable to the declared type.
+    ///
+    /// Annotations establish a declared type. So do function and class definitions, and imports.
+    pub(crate) fn is_declaration(self) -> bool {
+        matches!(
+            self,
+            DefinitionCategory::Declaration | DefinitionCategory::DeclarationAndBinding
+        )
+    }
+
+    /// True if this definition assigns a value to the symbol.
+    ///
+    /// False only for annotated assignments without a RHS.
+    pub(crate) fn is_binding(self) -> bool {
+        matches!(
+            self,
+            DefinitionCategory::Binding | DefinitionCategory::DeclarationAndBinding
+        )
+    }
+}
+
+#[derive(Clone, Debug)]
+pub enum DefinitionKind {
+    Import(AstNodeRef<ast::Alias>),
+    ImportFrom(ImportFromDefinitionKind),
+    Function(AstNodeRef<ast::StmtFunctionDef>),
+    Class(AstNodeRef<ast::StmtClassDef>),
+    NamedExpression(AstNodeRef<ast::ExprNamed>),
+    Assignment(AssignmentDefinitionKind),
+    AnnotatedAssignment(AstNodeRef<ast::StmtAnnAssign>),
+    AugmentedAssignment(AstNodeRef<ast::StmtAugAssign>),
+    For(ForStmtDefinitionKind),
+    Comprehension(ComprehensionDefinitionKind),
+    Parameter(AstNodeRef<ast::Parameter>),
+    ParameterWithDefault(AstNodeRef<ast::ParameterWithDefault>),
+    WithItem(WithItemDefinitionKind),
+    MatchPattern(MatchPatternDefinitionKind),
+    ExceptHandler(ExceptHandlerDefinitionKind),
+}
+
+impl DefinitionKind {
+    pub(crate) fn category(&self) -> DefinitionCategory {
+        match self {
+            // functions, classes, and imports always bind, and we consider them declarations
+            DefinitionKind::Function(_)
+            | DefinitionKind::Class(_)
+            | DefinitionKind::Import(_)
+            | DefinitionKind::ImportFrom(_) => DefinitionCategory::DeclarationAndBinding,
+            // a parameter always binds a value, but is only a declaration if annotated
+            DefinitionKind::Parameter(parameter) => {
+                if parameter.annotation.is_some() {
+                    DefinitionCategory::DeclarationAndBinding
+                } else {
+                    DefinitionCategory::Binding
+                }
+            }
+            // presence of a default is irrelevant, same logic as for a no-default parameter
+            DefinitionKind::ParameterWithDefault(parameter_with_default) => {
+                if parameter_with_default.parameter.annotation.is_some() {
+                    DefinitionCategory::DeclarationAndBinding
+                } else {
+                    DefinitionCategory::Binding
+                }
+            }
+            // annotated assignment is always a declaration, only a binding if there is a RHS
+            DefinitionKind::AnnotatedAssignment(ann_assign) => {
+                if ann_assign.value.is_some() {
+                    DefinitionCategory::DeclarationAndBinding
+                } else {
+                    DefinitionCategory::Declaration
+                }
+            }
+            // all of these bind values without declaring a type
+            DefinitionKind::NamedExpression(_)
+            | DefinitionKind::Assignment(_)
+            | DefinitionKind::AugmentedAssignment(_)
+            | DefinitionKind::For(_)
+            | DefinitionKind::Comprehension(_)
+            | DefinitionKind::WithItem(_)
+            | DefinitionKind::MatchPattern(_)
+            | DefinitionKind::ExceptHandler(_) => DefinitionCategory::Binding,
+        }
+    }
+}
+
+#[derive(Clone, Debug)]
+#[allow(dead_code)]
+pub struct MatchPatternDefinitionKind {
+    pattern: AstNodeRef<ast::Pattern>,
+    identifier: AstNodeRef<ast::Identifier>,
+    index: u32,
+}
+
+impl MatchPatternDefinitionKind {
+    pub(crate) fn pattern(&self) -> &ast::Pattern {
+        self.pattern.node()
+    }
+
+    pub(crate) fn index(&self) -> u32 {
+        self.index
+    }
+}
+
+#[derive(Clone, Debug)]
+pub struct ComprehensionDefinitionKind {
+    iterable: AstNodeRef<ast::Expr>,
+    target: AstNodeRef<ast::ExprName>,
+    first: bool,
+    is_async: bool,
+}
+
+impl ComprehensionDefinitionKind {
+    pub(crate) fn iterable(&self) -> &ast::Expr {
+        self.iterable.node()
+    }
+
+    pub(crate) fn target(&self) -> &ast::ExprName {
+        self.target.node()
+    }
+
+    pub(crate) fn is_first(&self) -> bool {
+        self.first
+    }
+
+    pub(crate) fn is_async(&self) -> bool {
+        self.is_async
+    }
+}
+
+#[derive(Clone, Debug)]
+pub struct ImportFromDefinitionKind {
+    node: AstNodeRef<ast::StmtImportFrom>,
+    alias_index: usize,
+}
+
+impl ImportFromDefinitionKind {
+    pub(crate) fn import(&self) -> &ast::StmtImportFrom {
+        self.node.node()
+    }
+
+    pub(crate) fn alias(&self) -> &ast::Alias {
+        &self.node.node().names[self.alias_index]
+    }
+}
+
+#[derive(Clone, Debug)]
+pub struct AssignmentDefinitionKind {
+    assignment: AstNodeRef<ast::StmtAssign>,
+    target: AstNodeRef<ast::ExprName>,
+}
+
+impl AssignmentDefinitionKind {
+    pub(crate) fn assignment(&self) -> &ast::StmtAssign {
+        self.assignment.node()
+    }
+
+    pub(crate) fn target(&self) -> &ast::ExprName {
+        self.target.node()
+    }
+}
+
+#[derive(Clone, Debug)]
+pub struct WithItemDefinitionKind {
+    node: AstNodeRef<ast::WithItem>,
+    target: AstNodeRef<ast::ExprName>,
+}
+
+impl WithItemDefinitionKind {
+    pub(crate) fn node(&self) -> &ast::WithItem {
+        self.node.node()
+    }
+
+    pub(crate) fn target(&self) -> &ast::ExprName {
+        self.target.node()
+    }
+}
+
+#[derive(Clone, Debug)]
+pub struct ForStmtDefinitionKind {
+    iterable: AstNodeRef<ast::Expr>,
+    target: AstNodeRef<ast::ExprName>,
+    is_async: bool,
+}
+
+impl ForStmtDefinitionKind {
+    pub(crate) fn iterable(&self) -> &ast::Expr {
+        self.iterable.node()
+    }
+
+    pub(crate) fn target(&self) -> &ast::ExprName {
+        self.target.node()
+    }
+
+    pub(crate) fn is_async(&self) -> bool {
+        self.is_async
+    }
+}
+
+#[derive(Clone, Debug)]
+pub struct ExceptHandlerDefinitionKind {
+    handler: AstNodeRef<ast::ExceptHandlerExceptHandler>,
+    is_star: bool,
+}
+
+impl ExceptHandlerDefinitionKind {
+    pub(crate) fn node(&self) -> &ast::ExceptHandlerExceptHandler {
+        self.handler.node()
+    }
+
+    pub(crate) fn handled_exceptions(&self) -> Option<&ast::Expr> {
+        self.node().type_.as_deref()
+    }
+
+    pub(crate) fn is_star(&self) -> bool {
+        self.is_star
+    }
+}
+
+#[derive(Copy, Clone, Eq, PartialEq, Hash, Debug)]
+pub(crate) struct DefinitionNodeKey(NodeKey);
+
+impl From<&ast::Alias> for DefinitionNodeKey {
+    fn from(node: &ast::Alias) -> Self {
+        Self(NodeKey::from_node(node))
+    }
+}
+
+impl From<&ast::StmtFunctionDef> for DefinitionNodeKey {
+    fn from(node: &ast::StmtFunctionDef) -> Self {
+        Self(NodeKey::from_node(node))
+    }
+}
+
+impl From<&ast::StmtClassDef> for DefinitionNodeKey {
+    fn from(node: &ast::StmtClassDef) -> Self {
+        Self(NodeKey::from_node(node))
+    }
+}
+
+impl From<&ast::ExprName> for DefinitionNodeKey {
+    fn from(node: &ast::ExprName) -> Self {
+        Self(NodeKey::from_node(node))
+    }
+}
+
+impl From<&ast::ExprNamed> for DefinitionNodeKey {
+    fn from(node: &ast::ExprNamed) -> Self {
+        Self(NodeKey::from_node(node))
+    }
+}
+
+impl From<&ast::StmtAnnAssign> for DefinitionNodeKey {
+    fn from(node: &ast::StmtAnnAssign) -> Self {
+        Self(NodeKey::from_node(node))
+    }
+}
+
+impl From<&ast::StmtAugAssign> for DefinitionNodeKey {
+    fn from(node: &ast::StmtAugAssign) -> Self {
+        Self(NodeKey::from_node(node))
+    }
+}
+
+impl From<&ast::StmtFor> for DefinitionNodeKey {
+    fn from(value: &ast::StmtFor) -> Self {
+        Self(NodeKey::from_node(value))
+    }
+}
+
+impl From<&ast::Parameter> for DefinitionNodeKey {
+    fn from(node: &ast::Parameter) -> Self {
+        Self(NodeKey::from_node(node))
+    }
+}
+
+impl From<&ast::ParameterWithDefault> for DefinitionNodeKey {
+    fn from(node: &ast::ParameterWithDefault) -> Self {
+        Self(NodeKey::from_node(node))
+    }
+}
+
+impl From<&ast::Identifier> for DefinitionNodeKey {
+    fn from(identifier: &ast::Identifier) -> Self {
+        Self(NodeKey::from_node(identifier))
+    }
+}
+
+impl From<&ast::ExceptHandlerExceptHandler> for DefinitionNodeKey {
+    fn from(handler: &ast::ExceptHandlerExceptHandler) -> Self {
+        Self(NodeKey::from_node(handler))
+    }
+}

crates/red_knot_python_semantic/src/semantic_index/expression.rs

@@ -0,0 +1,34 @@
+use crate::ast_node_ref::AstNodeRef;
+use crate::db::Db;
+use crate::semantic_index::symbol::{FileScopeId, ScopeId};
+use ruff_db::files::File;
+use ruff_python_ast as ast;
+use salsa;
+
+/// An independently type-inferable expression.
+///
+/// Includes constraint expressions (e.g. if tests) and the RHS of an unpacking assignment.
+#[salsa::tracked]
+pub(crate) struct Expression<'db> {
+    /// The file in which the expression occurs.
+    #[id]
+    pub(crate) file: File,
+
+    /// The scope in which the expression occurs.
+    #[id]
+    pub(crate) file_scope: FileScopeId,
+
+    /// The expression node.
+    #[no_eq]
+    #[return_ref]
+    pub(crate) node_ref: AstNodeRef<ast::Expr>,
+
+    #[no_eq]
+    count: countme::Count<Expression<'static>>,
+}
+
+impl<'db> Expression<'db> {
+    pub(crate) fn scope(self, db: &'db dyn Db) -> ScopeId<'db> {
+        self.file_scope(db).to_scope_id(db, self.file(db))
+    }
+}

crates/red_knot_python_semantic/src/semantic_index/symbol.rs

@@ -5,16 +5,15 @@ use bitflags::bitflags;
 use hashbrown::hash_map::RawEntryMut;
 use ruff_db::files::File;
 use ruff_db::parsed::ParsedModule;
-use ruff_index::{IndexVec, newtype_index};
-use ruff_python_ast as ast;
+use ruff_index::{newtype_index, IndexVec};
 use ruff_python_ast::name::Name;
+use ruff_python_ast::{self as ast};
 use rustc_hash::FxHasher;
 
-use crate::Db;
 use crate::ast_node_ref::AstNodeRef;
 use crate::node_key::NodeKey;
-use crate::semantic_index::visibility_constraints::ScopedVisibilityConstraintId;
-use crate::semantic_index::{SemanticIndex, SymbolMap, semantic_index};
+use crate::semantic_index::{semantic_index, SymbolMap};
+use crate::Db;
 
 #[derive(Eq, PartialEq, Debug)]
 pub struct Symbol {
@@ -48,27 +47,17 @@ impl Symbol {
     pub fn is_bound(&self) -> bool {
         self.flags.contains(SymbolFlags::IS_BOUND)
     }
-
-    /// Is the symbol declared in its containing scope?
-    pub fn is_declared(&self) -> bool {
-        self.flags.contains(SymbolFlags::IS_DECLARED)
-    }
 }
 
 bitflags! {
-    /// Flags that can be queried to obtain information about a symbol in a given scope.
-    ///
-    /// See the doc-comment at the top of [`super::use_def`] for explanations of what it
-    /// means for a symbol to be *bound* as opposed to *declared*.
     #[derive(Copy, Clone, Debug, Eq, PartialEq)]
     struct SymbolFlags: u8 {
         const IS_USED         = 1 << 0;
-        const IS_BOUND        = 1 << 1;
-        const IS_DECLARED     = 1 << 2;
+        const IS_BOUND      = 1 << 1;
         /// TODO: This flag is not yet set by anything
-        const MARKED_GLOBAL   = 1 << 3;
+        const MARKED_GLOBAL   = 1 << 2;
         /// TODO: This flag is not yet set by anything
-        const MARKED_NONLOCAL = 1 << 4;
+        const MARKED_NONLOCAL = 1 << 3;
     }
 }
 
@@ -97,34 +86,39 @@ impl From<FileSymbolId> for ScopedSymbolId {
 
 /// Symbol ID that uniquely identifies a symbol inside a [`Scope`].
 #[newtype_index]
-#[derive(salsa::Update)]
 pub struct ScopedSymbolId;
 
 /// A cross-module identifier of a scope that can be used as a salsa query parameter.
-#[salsa::tracked(debug)]
+#[salsa::tracked]
 pub struct ScopeId<'db> {
+    #[id]
     pub file: File,
-
+    #[id]
     pub file_scope_id: FileScopeId,
 
+    /// The node that introduces this scope.
+    #[no_eq]
+    #[return_ref]
+    pub node: NodeWithScopeKind,
+
+    #[no_eq]
     count: countme::Count<ScopeId<'static>>,
 }
 
 impl<'db> ScopeId<'db> {
     pub(crate) fn is_function_like(self, db: &'db dyn Db) -> bool {
-        self.node(db).scope_kind().is_function_like()
-    }
-
-    pub(crate) fn is_type_parameter(self, db: &'db dyn Db) -> bool {
-        self.node(db).scope_kind().is_type_parameter()
-    }
-
-    pub(crate) fn node(self, db: &dyn Db) -> &NodeWithScopeKind {
-        self.scope(db).node()
-    }
-
-    pub(crate) fn scope(self, db: &dyn Db) -> &Scope {
-        semantic_index(db, self.file(db)).scope(self.file_scope_id(db))
+        // Type parameter scopes behave like function scopes in terms of name resolution; CPython
+        // symbol table also uses the term "function-like" for these scopes.
+        matches!(
+            self.node(db),
+            NodeWithScopeKind::ClassTypeParameters(_)
+                | NodeWithScopeKind::FunctionTypeParameters(_)
+                | NodeWithScopeKind::Function(_)
+                | NodeWithScopeKind::ListComprehension(_)
+                | NodeWithScopeKind::SetComprehension(_)
+                | NodeWithScopeKind::DictComprehension(_)
+                | NodeWithScopeKind::GeneratorExpression(_)
+        )
     }
 
     #[cfg(test)]
@@ -136,12 +130,6 @@ impl<'db> ScopeId<'db> {
             }
             NodeWithScopeKind::Function(function)
             | NodeWithScopeKind::FunctionTypeParameters(function) => function.name.as_str(),
-            NodeWithScopeKind::TypeAlias(type_alias)
-            | NodeWithScopeKind::TypeAliasTypeParameters(type_alias) => type_alias
-                .name
-                .as_name_expr()
-                .map(|name| name.id.as_str())
-                .unwrap_or("<type alias>"),
             NodeWithScopeKind::Lambda(_) => "<lambda>",
             NodeWithScopeKind::ListComprehension(_) => "<listcomp>",
             NodeWithScopeKind::SetComprehension(_) => "<setcomp>",
@@ -153,7 +141,6 @@ impl<'db> ScopeId<'db> {
 
 /// ID that uniquely identifies a scope inside of a module.
 #[newtype_index]
-#[derive(salsa::Update)]
 pub struct FileScopeId;
 
 impl FileScopeId {
@@ -170,61 +157,22 @@ impl FileScopeId {
         let index = semantic_index(db, file);
         index.scope_ids_by_scope[self]
     }
-
-    pub(crate) fn is_generator_function(self, index: &SemanticIndex) -> bool {
-        index.generator_functions.contains(&self)
-    }
 }
 
-#[derive(Debug, salsa::Update)]
+#[derive(Debug, Eq, PartialEq)]
 pub struct Scope {
-    parent: Option<FileScopeId>,
-    node: NodeWithScopeKind,
-    descendants: Range<FileScopeId>,
-    reachability: ScopedVisibilityConstraintId,
+    pub(super) parent: Option<FileScopeId>,
+    pub(super) kind: ScopeKind,
+    pub(super) descendents: Range<FileScopeId>,
 }
 
 impl Scope {
-    pub(super) fn new(
-        parent: Option<FileScopeId>,
-        node: NodeWithScopeKind,
-        descendants: Range<FileScopeId>,
-        reachability: ScopedVisibilityConstraintId,
-    ) -> Self {
-        Scope {
-            parent,
-            node,
-            descendants,
-            reachability,
-        }
-    }
-
-    pub fn parent(&self) -> Option<FileScopeId> {
+    pub fn parent(self) -> Option<FileScopeId> {
         self.parent
     }
 
-    pub fn node(&self) -> &NodeWithScopeKind {
-        &self.node
-    }
-
     pub fn kind(&self) -> ScopeKind {
-        self.node().scope_kind()
-    }
-
-    pub fn descendants(&self) -> Range<FileScopeId> {
-        self.descendants.clone()
-    }
-
-    pub(super) fn extend_descendants(&mut self, children_end: FileScopeId) {
-        self.descendants = self.descendants.start..children_end;
-    }
-
-    pub(crate) fn is_eager(&self) -> bool {
-        self.kind().is_eager()
-    }
-
-    pub(crate) fn reachability(&self) -> ScopedVisibilityConstraintId {
-        self.reachability
+        self.kind
     }
 }
 
@@ -234,46 +182,17 @@ pub enum ScopeKind {
     Annotation,
     Class,
     Function,
-    Lambda,
     Comprehension,
-    TypeAlias,
 }
 
 impl ScopeKind {
-    pub(crate) fn is_eager(self) -> bool {
-        match self {
-            ScopeKind::Module | ScopeKind::Class | ScopeKind::Comprehension => true,
-            ScopeKind::Annotation
-            | ScopeKind::Function
-            | ScopeKind::Lambda
-            | ScopeKind::TypeAlias => false,
-        }
-    }
-
-    pub(crate) fn is_function_like(self) -> bool {
-        // Type parameter scopes behave like function scopes in terms of name resolution; CPython
-        // symbol table also uses the term "function-like" for these scopes.
-        matches!(
-            self,
-            ScopeKind::Annotation
-                | ScopeKind::Function
-                | ScopeKind::Lambda
-                | ScopeKind::TypeAlias
-                | ScopeKind::Comprehension
-        )
-    }
-
-    pub(crate) fn is_class(self) -> bool {
-        matches!(self, ScopeKind::Class)
-    }
-
-    pub(crate) fn is_type_parameter(self) -> bool {
-        matches!(self, ScopeKind::Annotation | ScopeKind::TypeAlias)
+    pub const fn is_comprehension(self) -> bool {
+        matches!(self, ScopeKind::Comprehension)
     }
 }
 
 /// Symbol table for a specific [`Scope`].
-#[derive(Default, salsa::Update)]
+#[derive(Debug)]
 pub struct SymbolTable {
     /// The symbols in this scope.
     symbols: IndexVec<ScopedSymbolId, Symbol>,
@@ -283,6 +202,13 @@ pub struct SymbolTable {
 }
 
 impl SymbolTable {
+    fn new() -> Self {
+        Self {
+            symbols: IndexVec::new(),
+            symbols_by_name: SymbolMap::default(),
+        }
+    }
+
     fn shrink_to_fit(&mut self) {
         self.symbols.shrink_to_fit();
     }
@@ -291,7 +217,7 @@ impl SymbolTable {
         &self.symbols[symbol_id.into()]
     }
 
-    #[expect(unused)]
+    #[allow(unused)]
     pub(crate) fn symbol_ids(&self) -> impl Iterator<Item = ScopedSymbolId> {
         self.symbols.indices()
     }
@@ -334,22 +260,18 @@ impl PartialEq for SymbolTable {
 
 impl Eq for SymbolTable {}
 
-impl std::fmt::Debug for SymbolTable {
-    /// Exclude the `symbols_by_name` field from the debug output.
-    /// It's very noisy and not useful for debugging.
-    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
-        f.debug_tuple("SymbolTable")
-            .field(&self.symbols)
-            .finish_non_exhaustive()
-    }
-}
-
-#[derive(Debug, Default)]
+#[derive(Debug)]
 pub(super) struct SymbolTableBuilder {
     table: SymbolTable,
 }
 
 impl SymbolTableBuilder {
+    pub(super) fn new() -> Self {
+        Self {
+            table: SymbolTable::new(),
+        }
+    }
+
     pub(super) fn add_symbol(&mut self, name: Name) -> (ScopedSymbolId, bool) {
         let hash = SymbolTable::hash_name(&name);
         let entry = self
@@ -376,26 +298,10 @@ impl SymbolTableBuilder {
         self.table.symbols[id].insert_flags(SymbolFlags::IS_BOUND);
     }
 
-    pub(super) fn mark_symbol_declared(&mut self, id: ScopedSymbolId) {
-        self.table.symbols[id].insert_flags(SymbolFlags::IS_DECLARED);
-    }
-
     pub(super) fn mark_symbol_used(&mut self, id: ScopedSymbolId) {
         self.table.symbols[id].insert_flags(SymbolFlags::IS_USED);
     }
 
-    pub(super) fn symbols(&self) -> impl Iterator<Item = &Symbol> {
-        self.table.symbols()
-    }
-
-    pub(super) fn symbol_id_by_name(&self, name: &str) -> Option<ScopedSymbolId> {
-        self.table.symbol_id_by_name(name)
-    }
-
-    pub(super) fn symbol(&self, symbol_id: impl Into<ScopedSymbolId>) -> &Symbol {
-        self.table.symbol(symbol_id)
-    }
-
     pub(super) fn finish(mut self) -> SymbolTable {
         self.table.shrink_to_fit();
         self.table
@@ -411,8 +317,6 @@ pub(crate) enum NodeWithScopeRef<'a> {
     Lambda(&'a ast::ExprLambda),
     FunctionTypeParameters(&'a ast::StmtFunctionDef),
     ClassTypeParameters(&'a ast::StmtClassDef),
-    TypeAlias(&'a ast::StmtTypeAlias),
-    TypeAliasTypeParameters(&'a ast::StmtTypeAlias),
     ListComprehension(&'a ast::ExprListComp),
     SetComprehension(&'a ast::ExprSetComp),
     DictComprehension(&'a ast::ExprDictComp),
@@ -424,45 +328,52 @@ impl NodeWithScopeRef<'_> {
     ///
     /// # Safety
     /// The node wrapped by `self` must be a child of `module`.
-    #[expect(unsafe_code)]
+    #[allow(unsafe_code)]
     pub(super) unsafe fn to_kind(self, module: ParsedModule) -> NodeWithScopeKind {
-        unsafe {
-            match self {
-                NodeWithScopeRef::Module => NodeWithScopeKind::Module,
-                NodeWithScopeRef::Class(class) => {
-                    NodeWithScopeKind::Class(AstNodeRef::new(module, class))
-                }
-                NodeWithScopeRef::Function(function) => {
-                    NodeWithScopeKind::Function(AstNodeRef::new(module, function))
-                }
-                NodeWithScopeRef::TypeAlias(type_alias) => {
-                    NodeWithScopeKind::TypeAlias(AstNodeRef::new(module, type_alias))
-                }
-                NodeWithScopeRef::TypeAliasTypeParameters(type_alias) => {
-                    NodeWithScopeKind::TypeAliasTypeParameters(AstNodeRef::new(module, type_alias))
-                }
-                NodeWithScopeRef::Lambda(lambda) => {
-                    NodeWithScopeKind::Lambda(AstNodeRef::new(module, lambda))
-                }
-                NodeWithScopeRef::FunctionTypeParameters(function) => {
-                    NodeWithScopeKind::FunctionTypeParameters(AstNodeRef::new(module, function))
-                }
-                NodeWithScopeRef::ClassTypeParameters(class) => {
-                    NodeWithScopeKind::ClassTypeParameters(AstNodeRef::new(module, class))
-                }
-                NodeWithScopeRef::ListComprehension(comprehension) => {
-                    NodeWithScopeKind::ListComprehension(AstNodeRef::new(module, comprehension))
-                }
-                NodeWithScopeRef::SetComprehension(comprehension) => {
-                    NodeWithScopeKind::SetComprehension(AstNodeRef::new(module, comprehension))
-                }
-                NodeWithScopeRef::DictComprehension(comprehension) => {
-                    NodeWithScopeKind::DictComprehension(AstNodeRef::new(module, comprehension))
-                }
-                NodeWithScopeRef::GeneratorExpression(generator) => {
-                    NodeWithScopeKind::GeneratorExpression(AstNodeRef::new(module, generator))
-                }
+        match self {
+            NodeWithScopeRef::Module => NodeWithScopeKind::Module,
+            NodeWithScopeRef::Class(class) => {
+                NodeWithScopeKind::Class(AstNodeRef::new(module, class))
             }
+            NodeWithScopeRef::Function(function) => {
+                NodeWithScopeKind::Function(AstNodeRef::new(module, function))
+            }
+            NodeWithScopeRef::Lambda(lambda) => {
+                NodeWithScopeKind::Lambda(AstNodeRef::new(module, lambda))
+            }
+            NodeWithScopeRef::FunctionTypeParameters(function) => {
+                NodeWithScopeKind::FunctionTypeParameters(AstNodeRef::new(module, function))
+            }
+            NodeWithScopeRef::ClassTypeParameters(class) => {
+                NodeWithScopeKind::ClassTypeParameters(AstNodeRef::new(module, class))
+            }
+            NodeWithScopeRef::ListComprehension(comprehension) => {
+                NodeWithScopeKind::ListComprehension(AstNodeRef::new(module, comprehension))
+            }
+            NodeWithScopeRef::SetComprehension(comprehension) => {
+                NodeWithScopeKind::SetComprehension(AstNodeRef::new(module, comprehension))
+            }
+            NodeWithScopeRef::DictComprehension(comprehension) => {
+                NodeWithScopeKind::DictComprehension(AstNodeRef::new(module, comprehension))
+            }
+            NodeWithScopeRef::GeneratorExpression(generator) => {
+                NodeWithScopeKind::GeneratorExpression(AstNodeRef::new(module, generator))
+            }
+        }
+    }
+
+    pub(super) fn scope_kind(self) -> ScopeKind {
+        match self {
+            NodeWithScopeRef::Module => ScopeKind::Module,
+            NodeWithScopeRef::Class(_) => ScopeKind::Class,
+            NodeWithScopeRef::Function(_) => ScopeKind::Function,
+            NodeWithScopeRef::Lambda(_) => ScopeKind::Function,
+            NodeWithScopeRef::FunctionTypeParameters(_)
+            | NodeWithScopeRef::ClassTypeParameters(_) => ScopeKind::Annotation,
+            NodeWithScopeRef::ListComprehension(_)
+            | NodeWithScopeRef::SetComprehension(_)
+            | NodeWithScopeRef::DictComprehension(_)
+            | NodeWithScopeRef::GeneratorExpression(_) => ScopeKind::Comprehension,
         }
     }
 
@@ -482,12 +393,6 @@ impl NodeWithScopeRef<'_> {
             NodeWithScopeRef::ClassTypeParameters(class) => {
                 NodeWithScopeKey::ClassTypeParameters(NodeKey::from_node(class))
             }
-            NodeWithScopeRef::TypeAlias(type_alias) => {
-                NodeWithScopeKey::TypeAlias(NodeKey::from_node(type_alias))
-            }
-            NodeWithScopeRef::TypeAliasTypeParameters(type_alias) => {
-                NodeWithScopeKey::TypeAliasTypeParameters(NodeKey::from_node(type_alias))
-            }
             NodeWithScopeRef::ListComprehension(comprehension) => {
                 NodeWithScopeKey::ListComprehension(NodeKey::from_node(comprehension))
             }
@@ -505,15 +410,13 @@ impl NodeWithScopeRef<'_> {
 }
 
 /// Node that introduces a new scope.
-#[derive(Clone, Debug, salsa::Update)]
+#[derive(Clone, Debug)]
 pub enum NodeWithScopeKind {
     Module,
     Class(AstNodeRef<ast::StmtClassDef>),
     ClassTypeParameters(AstNodeRef<ast::StmtClassDef>),
     Function(AstNodeRef<ast::StmtFunctionDef>),
     FunctionTypeParameters(AstNodeRef<ast::StmtFunctionDef>),
-    TypeAliasTypeParameters(AstNodeRef<ast::StmtTypeAlias>),
-    TypeAlias(AstNodeRef<ast::StmtTypeAlias>),
     Lambda(AstNodeRef<ast::ExprLambda>),
     ListComprehension(AstNodeRef<ast::ExprListComp>),
     SetComprehension(AstNodeRef<ast::ExprSetComp>),
@@ -521,57 +424,6 @@ pub enum NodeWithScopeKind {
     GeneratorExpression(AstNodeRef<ast::ExprGenerator>),
 }
 
-impl NodeWithScopeKind {
-    pub(crate) const fn scope_kind(&self) -> ScopeKind {
-        match self {
-            Self::Module => ScopeKind::Module,
-            Self::Class(_) => ScopeKind::Class,
-            Self::Function(_) => ScopeKind::Function,
-            Self::Lambda(_) => ScopeKind::Lambda,
-            Self::FunctionTypeParameters(_)
-            | Self::ClassTypeParameters(_)
-            | Self::TypeAliasTypeParameters(_) => ScopeKind::Annotation,
-            Self::TypeAlias(_) => ScopeKind::TypeAlias,
-            Self::ListComprehension(_)
-            | Self::SetComprehension(_)
-            | Self::DictComprehension(_)
-            | Self::GeneratorExpression(_) => ScopeKind::Comprehension,
-        }
-    }
-
-    pub fn expect_class(&self) -> &ast::StmtClassDef {
-        match self {
-            Self::Class(class) => class.node(),
-            _ => panic!("expected class"),
-        }
-    }
-
-    pub(crate) const fn as_class(&self) -> Option<&ast::StmtClassDef> {
-        match self {
-            Self::Class(class) => Some(class.node()),
-            _ => None,
-        }
-    }
-
-    pub fn expect_function(&self) -> &ast::StmtFunctionDef {
-        self.as_function().expect("expected function")
-    }
-
-    pub fn expect_type_alias(&self) -> &ast::StmtTypeAlias {
-        match self {
-            Self::TypeAlias(type_alias) => type_alias.node(),
-            _ => panic!("expected type alias"),
-        }
-    }
-
-    pub const fn as_function(&self) -> Option<&ast::StmtFunctionDef> {
-        match self {
-            Self::Function(function) => Some(function.node()),
-            _ => None,
-        }
-    }
-}
-
 #[derive(Copy, Clone, Debug, Eq, PartialEq, Hash)]
 pub(crate) enum NodeWithScopeKey {
     Module,
@@ -579,8 +431,6 @@ pub(crate) enum NodeWithScopeKey {
     ClassTypeParameters(NodeKey),
     Function(NodeKey),
     FunctionTypeParameters(NodeKey),
-    TypeAlias(NodeKey),
-    TypeAliasTypeParameters(NodeKey),
     Lambda(NodeKey),
     ListComprehension(NodeKey),
     SetComprehension(NodeKey),

crates/red_knot_python_semantic/src/semantic_index/use_def.rs

@@ -0,0 +1,575 @@
+//! First, some terminology:
+//!
+//! * A "binding" gives a new value to a variable. This includes many different Python statements
+//!   (assignment statements of course, but also imports, `def` and `class` statements, `as`
+//!   clauses in `with` and `except` statements, match patterns, and others) and even one
+//!   expression kind (named expressions). It notably does not include annotated assignment
+//!   statements without a right-hand side value; these do not assign any new value to the
+//!   variable. We consider function parameters to be bindings as well, since (from the perspective
+//!   of the function's internal scope), a function parameter begins the scope bound to a value.
+//!
+//! * A "declaration" establishes an upper bound type for the values that a variable may be
+//!   permitted to take on. Annotated assignment statements (with or without an RHS value) are
+//!   declarations; annotated function parameters are also declarations. We consider `def` and
+//!   `class` statements to also be declarations, so as to prohibit accidentally shadowing them.
+//!
+//! Annotated assignments with a right-hand side, and annotated function parameters, are both
+//! bindings and declarations.
+//!
+//! We use [`Definition`] as the universal term (and Salsa tracked struct) encompassing both
+//! bindings and declarations. (This sacrifices a bit of type safety in exchange for improved
+//! performance via fewer Salsa tracked structs and queries, since most declarations -- typed
+//! parameters and annotated assignments with RHS -- are both bindings and declarations.)
+//!
+//! At any given use of a variable, we can ask about both its "declared type" and its "inferred
+//! type". These may be different, but the inferred type must always be assignable to the declared
+//! type; that is, the declared type is always wider, and the inferred type may be more precise. If
+//! we see an invalid assignment, we emit a diagnostic and abandon our inferred type, deferring to
+//! the declared type (this allows an explicit annotation to override bad inference, without a
+//! cast), maintaining the invariant.
+//!
+//! The **inferred type** represents the most precise type we believe encompasses all possible
+//! values for the variable at a given use. It is based on a union of the bindings which can reach
+//! that use through some control flow path, and the narrowing constraints that control flow must
+//! have passed through between the binding and the use. For example, in this code:
+//!
+//! ```python
+//! x = 1 if flag else None
+//! if x is not None:
+//!     use(x)
+//! ```
+//!
+//! For the use of `x` on the third line, the inferred type should be `Literal[1]`. This is based
+//! on the binding on the first line, which assigns the type `Literal[1] | None`, and the narrowing
+//! constraint on the second line, which rules out the type `None`, since control flow must pass
+//! through this constraint to reach the use in question.
+//!
+//! The **declared type** represents the code author's declaration (usually through a type
+//! annotation) that a given variable should not be assigned any type outside the declared type. In
+//! our model, declared types are also control-flow-sensitive; we allow the code author to
+//! explicitly re-declare the same variable with a different type. So for a given binding of a
+//! variable, we will want to ask which declarations of that variable can reach that binding, in
+//! order to determine whether the binding is permitted, or should be a type error. For example:
+//!
+//! ```python
+//! from pathlib import Path
+//! def f(path: str):
+//!     path: Path = Path(path)
+//! ```
+//!
+//! In this function, the initial declared type of `path` is `str`, meaning that the assignment
+//! `path = Path(path)` would be a type error, since it assigns to `path` a value whose type is not
+//! assignable to `str`. This is the purpose of declared types: they prevent accidental assignment
+//! of the wrong type to a variable.
+//!
+//! But in some cases it is useful to "shadow" or "re-declare" a variable with a new type, and we
+//! permit this, as long as it is done with an explicit re-annotation. So `path: Path =
+//! Path(path)`, with the explicit `: Path` annotation, is permitted.
+//!
+//! The general rule is that whatever declaration(s) can reach a given binding determine the
+//! validity of that binding. If there is a path in which the symbol is not declared, that is a
+//! declaration of `Unknown`. If multiple declarations can reach a binding, we union them, but by
+//! default we also issue a type error, since this implicit union of declared types may hide an
+//! error.
+//!
+//! To support type inference, we build a map from each use of a symbol to the bindings live at
+//! that use, and the type narrowing constraints that apply to each binding.
+//!
+//! Let's take this code sample:
+//!
+//! ```python
+//! x = 1
+//! x = 2
+//! y = x
+//! if flag:
+//!     x = 3
+//! else:
+//!     x = 4
+//! z = x
+//! ```
+//!
+//! In this snippet, we have four bindings of `x` (the statements assigning `1`, `2`, `3`, and `4`
+//! to it), and two uses of `x` (the `y = x` and `z = x` assignments). The first binding of `x`
+//! does not reach any use, because it's immediately replaced by the second binding, before any use
+//! happens. (A linter could thus flag the statement `x = 1` as likely superfluous.)
+//!
+//! The first use of `x` has one live binding: the assignment `x = 2`.
+//!
+//! Things get a bit more complex when we have branches. We will definitely take either the `if` or
+//! the `else` branch. Thus, the second use of `x` has two live bindings: `x = 3` and `x = 4`. The
+//! `x = 2` assignment is no longer visible, because it must be replaced by either `x = 3` or `x =
+//! 4`, no matter which branch was taken. We don't know which branch was taken, so we must consider
+//! both bindings as live, which means eventually we would (in type inference) look at these two
+//! bindings and infer a type of `Literal[3, 4]` -- the union of `Literal[3]` and `Literal[4]` --
+//! for the second use of `x`.
+//!
+//! So that's one question our use-def map needs to answer: given a specific use of a symbol, which
+//! binding(s) can reach that use. In [`AstIds`](crate::semantic_index::ast_ids::AstIds) we number
+//! all uses (that means a `Name` node with `Load` context) so we have a `ScopedUseId` to
+//! efficiently represent each use.
+//!
+//! We also need to know, for a given definition of a symbol, what type narrowing constraints apply
+//! to it. For instance, in this code sample:
+//!
+//! ```python
+//! x = 1 if flag else None
+//! if x is not None:
+//!     use(x)
+//! ```
+//!
+//! At the use of `x`, the live binding of `x` is `1 if flag else None`, which would infer as the
+//! type `Literal[1] | None`. But the constraint `x is not None` dominates this use, which means we
+//! can rule out the possibility that `x` is `None` here, which should give us the type
+//! `Literal[1]` for this use.
+//!
+//! For declared types, we need to be able to answer the question "given a binding to a symbol,
+//! which declarations of that symbol can reach the binding?" This allows us to emit a diagnostic
+//! if the binding is attempting to bind a value of a type that is not assignable to the declared
+//! type for that symbol, at that point in control flow.
+//!
+//! We also need to know, given a declaration of a symbol, what the inferred type of that symbol is
+//! at that point. This allows us to emit a diagnostic in a case like `x = "foo"; x: int`. The
+//! binding `x = "foo"` occurs before the declaration `x: int`, so according to our
+//! control-flow-sensitive interpretation of declarations, the assignment is not an error. But the
+//! declaration is an error, since it would violate the "inferred type must be assignable to
+//! declared type" rule.
+//!
+//! Another case we need to handle is when a symbol is referenced from a different scope (for
+//! example, an import or a nonlocal reference). We call this "public" use of a symbol. For public
+//! use of a symbol, we prefer the declared type, if there are any declarations of that symbol; if
+//! not, we fall back to the inferred type. So we also need to know which declarations and bindings
+//! can reach the end of the scope.
+//!
+//! Technically, public use of a symbol could occur from any point in control flow of the scope
+//! where the symbol is defined (via inline imports and import cycles, in the case of an import, or
+//! via a function call partway through the local scope that ends up using a symbol from the scope
+//! via a global or nonlocal reference.) But modeling this fully accurately requires whole-program
+//! analysis that isn't tractable for an efficient analysis, since it means a given symbol could
+//! have a different type every place it's referenced throughout the program, depending on the
+//! shape of arbitrarily-sized call/import graphs. So we follow other Python type checkers in
+//! making the simplifying assumption that usually the scope will finish execution before its
+//! symbols are made visible to other scopes; for instance, most imports will import from a
+//! complete module, not a partially-executed module. (We may want to get a little smarter than
+//! this in the future for some closures, but for now this is where we start.)
+//!
+//! The data structure we build to answer these questions is the `UseDefMap`. It has a
+//! `bindings_by_use` vector of [`SymbolBindings`] indexed by [`ScopedUseId`], a
+//! `declarations_by_binding` vector of [`SymbolDeclarations`] indexed by [`ScopedDefinitionId`], a
+//! `bindings_by_declaration` vector of [`SymbolBindings`] indexed by [`ScopedDefinitionId`], and
+//! `public_bindings` and `public_definitions` vectors indexed by [`ScopedSymbolId`]. The values in
+//! each of these vectors are (in principle) a list of live bindings at that use/definition, or at
+//! the end of the scope for that symbol, with a list of the dominating constraints for each
+//! binding.
+//!
+//! In order to avoid vectors-of-vectors-of-vectors and all the allocations that would entail, we
+//! don't actually store these "list of visible definitions" as a vector of [`Definition`].
+//! Instead, [`SymbolBindings`] and [`SymbolDeclarations`] are structs which use bit-sets to track
+//! definitions (and constraints, in the case of bindings) in terms of [`ScopedDefinitionId`] and
+//! [`ScopedConstraintId`], which are indices into the `all_definitions` and `all_constraints`
+//! indexvecs in the [`UseDefMap`].
+//!
+//! There is another special kind of possible "definition" for a symbol: there might be a path from
+//! the scope entry to a given use in which the symbol is never bound.
+//!
+//! The simplest way to model "unbound" would be as a "binding" itself: the initial "binding" for
+//! each symbol in a scope. But actually modeling it this way would unnecessarily increase the
+//! number of [`Definition`]s that Salsa must track. Since "unbound" is special in that all symbols
+//! share it, and it doesn't have any additional per-symbol state, and constraints are irrelevant
+//! to it, we can represent it more efficiently: we use the `may_be_unbound` boolean on the
+//! [`SymbolBindings`] struct. If this flag is `true` for a use of a symbol, it means the symbol
+//! has a path to the use in which it is never bound. If this flag is `false`, it means we've
+//! eliminated the possibility of unbound: every control flow path to the use includes a binding
+//! for this symbol.
+//!
+//! To build a [`UseDefMap`], the [`UseDefMapBuilder`] is notified of each new use, definition, and
+//! constraint as they are encountered by the
+//! [`SemanticIndexBuilder`](crate::semantic_index::builder::SemanticIndexBuilder) AST visit. For
+//! each symbol, the builder tracks the `SymbolState` (`SymbolBindings` and `SymbolDeclarations`)
+//! for that symbol. When we hit a use or definition of a symbol, we record the necessary parts of
+//! the current state for that symbol that we need for that use or definition. When we reach the
+//! end of the scope, it records the state for each symbol as the public definitions of that
+//! symbol.
+//!
+//! Let's walk through the above example. Initially we record for `x` that it has no bindings, and
+//! may be unbound. When we see `x = 1`, we record that as the sole live binding of `x`, and flip
+//! `may_be_unbound` to `false`. Then we see `x = 2`, and we replace `x = 1` as the sole live
+//! binding of `x`. When we get to `y = x`, we record that the live bindings for that use of `x`
+//! are just the `x = 2` definition.
+//!
+//! Then we hit the `if` branch. We visit the `test` node (`flag` in this case), since that will
+//! happen regardless. Then we take a pre-branch snapshot of the current state for all symbols,
+//! which we'll need later. Then we record `flag` as a possible constraint on the current binding
+//! (`x = 2`), and go ahead and visit the `if` body. When we see `x = 3`, it replaces `x = 2`
+//! (constrained by `flag`) as the sole live binding of `x`. At the end of the `if` body, we take
+//! another snapshot of the current symbol state; we'll call this the post-if-body snapshot.
+//!
+//! Now we need to visit the `else` clause. The conditions when entering the `else` clause should
+//! be the pre-if conditions; if we are entering the `else` clause, we know that the `if` test
+//! failed and we didn't execute the `if` body. So we first reset the builder to the pre-if state,
+//! using the snapshot we took previously (meaning we now have `x = 2` as the sole binding for `x`
+//! again), then visit the `else` clause, where `x = 4` replaces `x = 2` as the sole live binding
+//! of `x`.
+//!
+//! Now we reach the end of the if/else, and want to visit the following code. The state here needs
+//! to reflect that we might have gone through the `if` branch, or we might have gone through the
+//! `else` branch, and we don't know which. So we need to "merge" our current builder state
+//! (reflecting the end-of-else state, with `x = 4` as the only live binding) with our post-if-body
+//! snapshot (which has `x = 3` as the only live binding). The result of this merge is that we now
+//! have two live bindings of `x`: `x = 3` and `x = 4`.
+//!
+//! The [`UseDefMapBuilder`] itself just exposes methods for taking a snapshot, resetting to a
+//! snapshot, and merging a snapshot into the current state. The logic using these methods lives in
+//! [`SemanticIndexBuilder`](crate::semantic_index::builder::SemanticIndexBuilder), e.g. where it
+//! visits a `StmtIf` node.
+use self::symbol_state::{
+    BindingIdWithConstraintsIterator, ConstraintIdIterator, DeclarationIdIterator,
+    ScopedConstraintId, ScopedDefinitionId, SymbolBindings, SymbolDeclarations, SymbolState,
+};
+use crate::semantic_index::ast_ids::ScopedUseId;
+use crate::semantic_index::definition::Definition;
+use crate::semantic_index::symbol::ScopedSymbolId;
+use ruff_index::IndexVec;
+use rustc_hash::FxHashMap;
+
+use super::constraint::Constraint;
+
+mod bitset;
+mod symbol_state;
+
+/// Applicable definitions and constraints for every use of a name.
+#[derive(Debug, PartialEq, Eq)]
+pub(crate) struct UseDefMap<'db> {
+    /// Array of [`Definition`] in this scope.
+    all_definitions: IndexVec<ScopedDefinitionId, Definition<'db>>,
+
+    /// Array of [`Constraint`] in this scope.
+    all_constraints: IndexVec<ScopedConstraintId, Constraint<'db>>,
+
+    /// [`SymbolBindings`] reaching a [`ScopedUseId`].
+    bindings_by_use: IndexVec<ScopedUseId, SymbolBindings>,
+
+    /// [`SymbolBindings`] or [`SymbolDeclarations`] reaching a given [`Definition`].
+    ///
+    /// If the definition is a binding (only) -- `x = 1` for example -- then we need
+    /// [`SymbolDeclarations`] to know whether this binding is permitted by the live declarations.
+    ///
+    /// If the definition is a declaration (only) -- `x: int` for example -- then we need
+    /// [`SymbolBindings`] to know whether this declaration is consistent with the previously
+    /// inferred type.
+    ///
+    /// If the definition is both a declaration and a binding -- `x: int = 1` for example -- then
+    /// we don't actually need anything here, all we'll need to validate is that our own RHS is a
+    /// valid assignment to our own annotation.
+    definitions_by_definition: FxHashMap<Definition<'db>, SymbolDefinitions>,
+
+    /// [`SymbolState`] visible at end of scope for each symbol.
+    public_symbols: IndexVec<ScopedSymbolId, SymbolState>,
+}
+
+impl<'db> UseDefMap<'db> {
+    pub(crate) fn bindings_at_use(
+        &self,
+        use_id: ScopedUseId,
+    ) -> BindingWithConstraintsIterator<'_, 'db> {
+        self.bindings_iterator(&self.bindings_by_use[use_id])
+    }
+
+    pub(crate) fn use_may_be_unbound(&self, use_id: ScopedUseId) -> bool {
+        self.bindings_by_use[use_id].may_be_unbound()
+    }
+
+    pub(crate) fn public_bindings(
+        &self,
+        symbol: ScopedSymbolId,
+    ) -> BindingWithConstraintsIterator<'_, 'db> {
+        self.bindings_iterator(self.public_symbols[symbol].bindings())
+    }
+
+    pub(crate) fn public_may_be_unbound(&self, symbol: ScopedSymbolId) -> bool {
+        self.public_symbols[symbol].may_be_unbound()
+    }
+
+    pub(crate) fn bindings_at_declaration(
+        &self,
+        declaration: Definition<'db>,
+    ) -> BindingWithConstraintsIterator<'_, 'db> {
+        if let SymbolDefinitions::Bindings(bindings) = &self.definitions_by_definition[&declaration]
+        {
+            self.bindings_iterator(bindings)
+        } else {
+            unreachable!("Declaration has non-Bindings in definitions_by_definition");
+        }
+    }
+
+    pub(crate) fn declarations_at_binding(
+        &self,
+        binding: Definition<'db>,
+    ) -> DeclarationsIterator<'_, 'db> {
+        if let SymbolDefinitions::Declarations(declarations) =
+            &self.definitions_by_definition[&binding]
+        {
+            self.declarations_iterator(declarations)
+        } else {
+            unreachable!("Binding has non-Declarations in definitions_by_definition");
+        }
+    }
+
+    pub(crate) fn public_declarations(
+        &self,
+        symbol: ScopedSymbolId,
+    ) -> DeclarationsIterator<'_, 'db> {
+        let declarations = self.public_symbols[symbol].declarations();
+        self.declarations_iterator(declarations)
+    }
+
+    pub(crate) fn has_public_declarations(&self, symbol: ScopedSymbolId) -> bool {
+        !self.public_symbols[symbol].declarations().is_empty()
+    }
+
+    fn bindings_iterator<'a>(
+        &'a self,
+        bindings: &'a SymbolBindings,
+    ) -> BindingWithConstraintsIterator<'a, 'db> {
+        BindingWithConstraintsIterator {
+            all_definitions: &self.all_definitions,
+            all_constraints: &self.all_constraints,
+            inner: bindings.iter(),
+        }
+    }
+
+    fn declarations_iterator<'a>(
+        &'a self,
+        declarations: &'a SymbolDeclarations,
+    ) -> DeclarationsIterator<'a, 'db> {
+        DeclarationsIterator {
+            all_definitions: &self.all_definitions,
+            inner: declarations.iter(),
+            may_be_undeclared: declarations.may_be_undeclared(),
+        }
+    }
+}
+
+/// Either live bindings or live declarations for a symbol.
+#[derive(Debug, PartialEq, Eq)]
+enum SymbolDefinitions {
+    Bindings(SymbolBindings),
+    Declarations(SymbolDeclarations),
+}
+
+#[derive(Debug)]
+pub(crate) struct BindingWithConstraintsIterator<'map, 'db> {
+    all_definitions: &'map IndexVec<ScopedDefinitionId, Definition<'db>>,
+    all_constraints: &'map IndexVec<ScopedConstraintId, Constraint<'db>>,
+    inner: BindingIdWithConstraintsIterator<'map>,
+}
+
+impl<'map, 'db> Iterator for BindingWithConstraintsIterator<'map, 'db> {
+    type Item = BindingWithConstraints<'map, 'db>;
+
+    fn next(&mut self) -> Option<Self::Item> {
+        self.inner
+            .next()
+            .map(|def_id_with_constraints| BindingWithConstraints {
+                binding: self.all_definitions[def_id_with_constraints.definition],
+                constraints: ConstraintsIterator {
+                    all_constraints: self.all_constraints,
+                    constraint_ids: def_id_with_constraints.constraint_ids,
+                },
+            })
+    }
+}
+
+impl std::iter::FusedIterator for BindingWithConstraintsIterator<'_, '_> {}
+
+pub(crate) struct BindingWithConstraints<'map, 'db> {
+    pub(crate) binding: Definition<'db>,
+    pub(crate) constraints: ConstraintsIterator<'map, 'db>,
+}
+
+pub(crate) struct ConstraintsIterator<'map, 'db> {
+    all_constraints: &'map IndexVec<ScopedConstraintId, Constraint<'db>>,
+    constraint_ids: ConstraintIdIterator<'map>,
+}
+
+impl<'map, 'db> Iterator for ConstraintsIterator<'map, 'db> {
+    type Item = Constraint<'db>;
+
+    fn next(&mut self) -> Option<Self::Item> {
+        self.constraint_ids
+            .next()
+            .map(|constraint_id| self.all_constraints[constraint_id])
+    }
+}
+
+impl std::iter::FusedIterator for ConstraintsIterator<'_, '_> {}
+
+pub(crate) struct DeclarationsIterator<'map, 'db> {
+    all_definitions: &'map IndexVec<ScopedDefinitionId, Definition<'db>>,
+    inner: DeclarationIdIterator<'map>,
+    may_be_undeclared: bool,
+}
+
+impl DeclarationsIterator<'_, '_> {
+    pub(crate) fn may_be_undeclared(&self) -> bool {
+        self.may_be_undeclared
+    }
+}
+
+impl<'map, 'db> Iterator for DeclarationsIterator<'map, 'db> {
+    type Item = Definition<'db>;
+
+    fn next(&mut self) -> Option<Self::Item> {
+        self.inner.next().map(|def_id| self.all_definitions[def_id])
+    }
+}
+
+impl std::iter::FusedIterator for DeclarationsIterator<'_, '_> {}
+
+/// A snapshot of the definitions and constraints state at a particular point in control flow.
+#[derive(Clone, Debug)]
+pub(super) struct FlowSnapshot {
+    symbol_states: IndexVec<ScopedSymbolId, SymbolState>,
+}
+
+#[derive(Debug, Default)]
+pub(super) struct UseDefMapBuilder<'db> {
+    /// Append-only array of [`Definition`].
+    all_definitions: IndexVec<ScopedDefinitionId, Definition<'db>>,
+
+    /// Append-only array of [`Constraint`].
+    all_constraints: IndexVec<ScopedConstraintId, Constraint<'db>>,
+
+    /// Live bindings at each so-far-recorded use.
+    bindings_by_use: IndexVec<ScopedUseId, SymbolBindings>,
+
+    /// Live bindings or declarations for each so-far-recorded definition.
+    definitions_by_definition: FxHashMap<Definition<'db>, SymbolDefinitions>,
+
+    /// Currently live bindings and declarations for each symbol.
+    symbol_states: IndexVec<ScopedSymbolId, SymbolState>,
+}
+
+impl<'db> UseDefMapBuilder<'db> {
+    pub(super) fn new() -> Self {
+        Self::default()
+    }
+
+    pub(super) fn add_symbol(&mut self, symbol: ScopedSymbolId) {
+        let new_symbol = self.symbol_states.push(SymbolState::undefined());
+        debug_assert_eq!(symbol, new_symbol);
+    }
+
+    pub(super) fn record_binding(&mut self, symbol: ScopedSymbolId, binding: Definition<'db>) {
+        let def_id = self.all_definitions.push(binding);
+        let symbol_state = &mut self.symbol_states[symbol];
+        self.definitions_by_definition.insert(
+            binding,
+            SymbolDefinitions::Declarations(symbol_state.declarations().clone()),
+        );
+        symbol_state.record_binding(def_id);
+    }
+
+    pub(super) fn record_constraint(&mut self, constraint: Constraint<'db>) {
+        let constraint_id = self.all_constraints.push(constraint);
+        for state in &mut self.symbol_states {
+            state.record_constraint(constraint_id);
+        }
+    }
+
+    pub(super) fn record_declaration(
+        &mut self,
+        symbol: ScopedSymbolId,
+        declaration: Definition<'db>,
+    ) {
+        let def_id = self.all_definitions.push(declaration);
+        let symbol_state = &mut self.symbol_states[symbol];
+        self.definitions_by_definition.insert(
+            declaration,
+            SymbolDefinitions::Bindings(symbol_state.bindings().clone()),
+        );
+        symbol_state.record_declaration(def_id);
+    }
+
+    pub(super) fn record_declaration_and_binding(
+        &mut self,
+        symbol: ScopedSymbolId,
+        definition: Definition<'db>,
+    ) {
+        // We don't need to store anything in self.definitions_by_definition.
+        let def_id = self.all_definitions.push(definition);
+        let symbol_state = &mut self.symbol_states[symbol];
+        symbol_state.record_declaration(def_id);
+        symbol_state.record_binding(def_id);
+    }
+
+    pub(super) fn record_use(&mut self, symbol: ScopedSymbolId, use_id: ScopedUseId) {
+        // We have a use of a symbol; clone the current bindings for that symbol, and record them
+        // as the live bindings for this use.
+        let new_use = self
+            .bindings_by_use
+            .push(self.symbol_states[symbol].bindings().clone());
+        debug_assert_eq!(use_id, new_use);
+    }
+
+    /// Take a snapshot of the current visible-symbols state.
+    pub(super) fn snapshot(&self) -> FlowSnapshot {
+        FlowSnapshot {
+            symbol_states: self.symbol_states.clone(),
+        }
+    }
+
+    /// Restore the current builder symbols state to the given snapshot.
+    pub(super) fn restore(&mut self, snapshot: FlowSnapshot) {
+        // We never remove symbols from `symbol_states` (it's an IndexVec, and the symbol
+        // IDs must line up), so the current number of known symbols must always be equal to or
+        // greater than the number of known symbols in a previously-taken snapshot.
+        let num_symbols = self.symbol_states.len();
+        debug_assert!(num_symbols >= snapshot.symbol_states.len());
+
+        // Restore the current visible-definitions state to the given snapshot.
+        self.symbol_states = snapshot.symbol_states;
+
+        // If the snapshot we are restoring is missing some symbols we've recorded since, we need
+        // to fill them in so the symbol IDs continue to line up. Since they don't exist in the
+        // snapshot, the correct state to fill them in with is "undefined".
+        self.symbol_states
+            .resize(num_symbols, SymbolState::undefined());
+    }
+
+    /// Merge the given snapshot into the current state, reflecting that we might have taken either
+    /// path to get here. The new state for each symbol should include definitions from both the
+    /// prior state and the snapshot.
+    pub(super) fn merge(&mut self, snapshot: FlowSnapshot) {
+        // We never remove symbols from `symbol_states` (it's an IndexVec, and the symbol
+        // IDs must line up), so the current number of known symbols must always be equal to or
+        // greater than the number of known symbols in a previously-taken snapshot.
+        debug_assert!(self.symbol_states.len() >= snapshot.symbol_states.len());
+
+        let mut snapshot_definitions_iter = snapshot.symbol_states.into_iter();
+        for current in &mut self.symbol_states {
+            if let Some(snapshot) = snapshot_definitions_iter.next() {
+                current.merge(snapshot);
+            } else {
+                // Symbol not present in snapshot, so it's unbound/undeclared from that path.
+                current.set_may_be_unbound();
+                current.set_may_be_undeclared();
+            }
+        }
+    }
+
+    pub(super) fn finish(mut self) -> UseDefMap<'db> {
+        self.all_definitions.shrink_to_fit();
+        self.all_constraints.shrink_to_fit();
+        self.symbol_states.shrink_to_fit();
+        self.bindings_by_use.shrink_to_fit();
+        self.definitions_by_definition.shrink_to_fit();
+
+        UseDefMap {
+            all_definitions: self.all_definitions,
+            all_constraints: self.all_constraints,
+            bindings_by_use: self.bindings_by_use,
+            public_symbols: self.symbol_states,
+            definitions_by_definition: self.definitions_by_definition,
+        }
+    }
+}

crates/red_knot_python_semantic/src/semantic_index/use_def/bitset.rs

@@ -0,0 +1,309 @@
+/// Ordered set of `u32`.
+///
+/// Uses an inline bit-set for small values (up to 64 * B), falls back to heap allocated vector of
+/// blocks for larger values.
+#[derive(Debug, Clone, PartialEq, Eq)]
+pub(super) enum BitSet<const B: usize> {
+    /// Bit-set (in 64-bit blocks) for the first 64 * B entries.
+    Inline([u64; B]),
+
+    /// Overflow beyond 64 * B.
+    Heap(Vec<u64>),
+}
+
+impl<const B: usize> Default for BitSet<B> {
+    fn default() -> Self {
+        // B * 64 must fit in a u32, or else we have unusable bits; this assertion makes the
+        // truncating casts to u32 below safe. This would be better as a const assertion, but
+        // that's not possible on stable with const generic params. (B should never really be
+        // anywhere close to this large.)
+        assert!(B * 64 < (u32::MAX as usize));
+        // This implementation requires usize >= 32 bits.
+        static_assertions::const_assert!(usize::BITS >= 32);
+        Self::Inline([0; B])
+    }
+}
+
+impl<const B: usize> BitSet<B> {
+    /// Create and return a new [`BitSet`] with a single `value` inserted.
+    pub(super) fn with(value: u32) -> Self {
+        let mut bitset = Self::default();
+        bitset.insert(value);
+        bitset
+    }
+
+    pub(super) fn is_empty(&self) -> bool {
+        self.blocks().iter().all(|&b| b == 0)
+    }
+
+    /// Convert from Inline to Heap, if needed, and resize the Heap vector, if needed.
+    fn resize(&mut self, value: u32) {
+        let num_blocks_needed = (value / 64) + 1;
+        self.resize_blocks(num_blocks_needed as usize);
+    }
+
+    fn resize_blocks(&mut self, num_blocks_needed: usize) {
+        match self {
+            Self::Inline(blocks) => {
+                let mut vec = blocks.to_vec();
+                vec.resize(num_blocks_needed, 0);
+                *self = Self::Heap(vec);
+            }
+            Self::Heap(vec) => {
+                vec.resize(num_blocks_needed, 0);
+            }
+        }
+    }
+
+    fn blocks_mut(&mut self) -> &mut [u64] {
+        match self {
+            Self::Inline(blocks) => blocks.as_mut_slice(),
+            Self::Heap(blocks) => blocks.as_mut_slice(),
+        }
+    }
+
+    fn blocks(&self) -> &[u64] {
+        match self {
+            Self::Inline(blocks) => blocks.as_slice(),
+            Self::Heap(blocks) => blocks.as_slice(),
+        }
+    }
+
+    /// Insert a value into the [`BitSet`].
+    ///
+    /// Return true if the value was newly inserted, false if already present.
+    pub(super) fn insert(&mut self, value: u32) -> bool {
+        let value_usize = value as usize;
+        let (block, index) = (value_usize / 64, value_usize % 64);
+        if block >= self.blocks().len() {
+            self.resize(value);
+        }
+        let blocks = self.blocks_mut();
+        let missing = blocks[block] & (1 << index) == 0;
+        blocks[block] |= 1 << index;
+        missing
+    }
+
+    /// Intersect in-place with another [`BitSet`].
+    pub(super) fn intersect(&mut self, other: &BitSet<B>) {
+        let my_blocks = self.blocks_mut();
+        let other_blocks = other.blocks();
+        let min_len = my_blocks.len().min(other_blocks.len());
+        for i in 0..min_len {
+            my_blocks[i] &= other_blocks[i];
+        }
+        for block in my_blocks.iter_mut().skip(min_len) {
+            *block = 0;
+        }
+    }
+
+    /// Union in-place with another [`BitSet`].
+    pub(super) fn union(&mut self, other: &BitSet<B>) {
+        let mut max_len = self.blocks().len();
+        let other_len = other.blocks().len();
+        if other_len > max_len {
+            max_len = other_len;
+            self.resize_blocks(max_len);
+        }
+        for (my_block, other_block) in self.blocks_mut().iter_mut().zip(other.blocks()) {
+            *my_block |= other_block;
+        }
+    }
+
+    /// Return an iterator over the values (in ascending order) in this [`BitSet`].
+    pub(super) fn iter(&self) -> BitSetIterator<'_, B> {
+        let blocks = self.blocks();
+        BitSetIterator {
+            blocks,
+            current_block_index: 0,
+            current_block: blocks[0],
+        }
+    }
+}
+
+/// Iterator over values in a [`BitSet`].
+#[derive(Debug)]
+pub(super) struct BitSetIterator<'a, const B: usize> {
+    /// The blocks we are iterating over.
+    blocks: &'a [u64],
+
+    /// The index of the block we are currently iterating through.
+    current_block_index: usize,
+
+    /// The block we are currently iterating through (and zeroing as we go.)
+    current_block: u64,
+}
+
+impl<const B: usize> Iterator for BitSetIterator<'_, B> {
+    type Item = u32;
+
+    fn next(&mut self) -> Option<Self::Item> {
+        while self.current_block == 0 {
+            if self.current_block_index + 1 >= self.blocks.len() {
+                return None;
+            }
+            self.current_block_index += 1;
+            self.current_block = self.blocks[self.current_block_index];
+        }
+        let lowest_bit_set = self.current_block.trailing_zeros();
+        // reset the lowest set bit, without a data dependency on `lowest_bit_set`
+        self.current_block &= self.current_block.wrapping_sub(1);
+        // SAFETY: `lowest_bit_set` cannot be more than 64, `current_block_index` cannot be more
+        // than `B - 1`, and we check above that `B * 64 < u32::MAX`. So both `64 *
+        // current_block_index` and the final value here must fit in u32.
+        #[allow(clippy::cast_possible_truncation)]
+        Some(lowest_bit_set + (64 * self.current_block_index) as u32)
+    }
+}
+
+impl<const B: usize> std::iter::FusedIterator for BitSetIterator<'_, B> {}
+
+#[cfg(test)]
+mod tests {
+    use super::BitSet;
+
+    fn assert_bitset<const B: usize>(bitset: &BitSet<B>, contents: &[u32]) {
+        assert_eq!(bitset.iter().collect::<Vec<_>>(), contents);
+    }
+
+    #[test]
+    fn iter() {
+        let mut b = BitSet::<1>::with(3);
+        b.insert(27);
+        b.insert(6);
+        assert!(matches!(b, BitSet::Inline(_)));
+        assert_bitset(&b, &[3, 6, 27]);
+    }
+
+    #[test]
+    fn iter_overflow() {
+        let mut b = BitSet::<1>::with(140);
+        b.insert(100);
+        b.insert(129);
+        assert!(matches!(b, BitSet::Heap(_)));
+        assert_bitset(&b, &[100, 129, 140]);
+    }
+
+    #[test]
+    fn intersect() {
+        let mut b1 = BitSet::<1>::with(4);
+        let mut b2 = BitSet::<1>::with(4);
+        b1.insert(23);
+        b2.insert(5);
+
+        b1.intersect(&b2);
+        assert_bitset(&b1, &[4]);
+    }
+
+    #[test]
+    fn intersect_mixed_1() {
+        let mut b1 = BitSet::<1>::with(4);
+        let mut b2 = BitSet::<1>::with(4);
+        b1.insert(89);
+        b2.insert(5);
+
+        b1.intersect(&b2);
+        assert_bitset(&b1, &[4]);
+    }
+
+    #[test]
+    fn intersect_mixed_2() {
+        let mut b1 = BitSet::<1>::with(4);
+        let mut b2 = BitSet::<1>::with(4);
+        b1.insert(23);
+        b2.insert(89);
+
+        b1.intersect(&b2);
+        assert_bitset(&b1, &[4]);
+    }
+
+    #[test]
+    fn intersect_heap() {
+        let mut b1 = BitSet::<1>::with(4);
+        let mut b2 = BitSet::<1>::with(4);
+        b1.insert(89);
+        b2.insert(90);
+
+        b1.intersect(&b2);
+        assert_bitset(&b1, &[4]);
+    }
+
+    #[test]
+    fn intersect_heap_2() {
+        let mut b1 = BitSet::<1>::with(89);
+        let mut b2 = BitSet::<1>::with(89);
+        b1.insert(91);
+        b2.insert(90);
+
+        b1.intersect(&b2);
+        assert_bitset(&b1, &[89]);
+    }
+
+    #[test]
+    fn union() {
+        let mut b1 = BitSet::<1>::with(2);
+        let b2 = BitSet::<1>::with(4);
+
+        b1.union(&b2);
+        assert_bitset(&b1, &[2, 4]);
+    }
+
+    #[test]
+    fn union_mixed_1() {
+        let mut b1 = BitSet::<1>::with(4);
+        let mut b2 = BitSet::<1>::with(4);
+        b1.insert(89);
+        b2.insert(5);
+
+        b1.union(&b2);
+        assert_bitset(&b1, &[4, 5, 89]);
+    }
+
+    #[test]
+    fn union_mixed_2() {
+        let mut b1 = BitSet::<1>::with(4);
+        let mut b2 = BitSet::<1>::with(4);
+        b1.insert(23);
+        b2.insert(89);
+
+        b1.union(&b2);
+        assert_bitset(&b1, &[4, 23, 89]);
+    }
+
+    #[test]
+    fn union_heap() {
+        let mut b1 = BitSet::<1>::with(4);
+        let mut b2 = BitSet::<1>::with(4);
+        b1.insert(89);
+        b2.insert(90);
+
+        b1.union(&b2);
+        assert_bitset(&b1, &[4, 89, 90]);
+    }
+
+    #[test]
+    fn union_heap_2() {
+        let mut b1 = BitSet::<1>::with(89);
+        let mut b2 = BitSet::<1>::with(89);
+        b1.insert(91);
+        b2.insert(90);
+
+        b1.union(&b2);
+        assert_bitset(&b1, &[89, 90, 91]);
+    }
+
+    #[test]
+    fn multiple_blocks() {
+        let mut b = BitSet::<2>::with(120);
+        b.insert(45);
+        assert!(matches!(b, BitSet::Inline(_)));
+        assert_bitset(&b, &[45, 120]);
+    }
+
+    #[test]
+    fn empty() {
+        let b = BitSet::<1>::default();
+
+        assert!(b.is_empty());
+    }
+}

crates/red_knot_python_semantic/src/semantic_index/use_def/symbol_state.rs

@@ -0,0 +1,588 @@
+//! Track live bindings per symbol, applicable constraints per binding, and live declarations.
+//!
+//! These data structures operate entirely on scope-local newtype-indices for definitions and
+//! constraints, referring to their location in the `all_definitions` and `all_constraints`
+//! indexvecs in [`super::UseDefMapBuilder`].
+//!
+//! We need to track arbitrary associations between bindings and constraints, not just a single set
+//! of currently dominating constraints (where "dominating" means "control flow must have passed
+//! through it to reach this point"), because we can have dominating constraints that apply to some
+//! bindings but not others, as in this code:
+//!
+//! ```python
+//! x = 1 if flag else None
+//! if x is not None:
+//!     if flag2:
+//!         x = 2 if flag else None
+//!     x
+//! ```
+//!
+//! The `x is not None` constraint dominates the final use of `x`, but it applies only to the first
+//! binding of `x`, not the second, so `None` is a possible value for `x`.
+//!
+//! And we can't just track, for each binding, an index into a list of dominating constraints,
+//! either, because we can have bindings which are still visible, but subject to constraints that
+//! are no longer dominating, as in this code:
+//!
+//! ```python
+//! x = 0
+//! if flag1:
+//!     x = 1 if flag2 else None
+//!     assert x is not None
+//! x
+//! ```
+//!
+//! From the point of view of the final use of `x`, the `x is not None` constraint no longer
+//! dominates, but it does dominate the `x = 1 if flag2 else None` binding, so we have to keep
+//! track of that.
+//!
+//! The data structures used here ([`BitSet`] and [`smallvec::SmallVec`]) optimize for keeping all
+//! data inline (avoiding lots of scattered allocations) in small-to-medium cases, and falling back
+//! to heap allocation to be able to scale to arbitrary numbers of live bindings and constraints
+//! when needed.
+//!
+//! Tracking live declarations is simpler, since constraints are not involved, but otherwise very
+//! similar to tracking live bindings.
+use super::bitset::{BitSet, BitSetIterator};
+use ruff_index::newtype_index;
+use smallvec::SmallVec;
+
+/// A newtype-index for a definition in a particular scope.
+#[newtype_index]
+pub(super) struct ScopedDefinitionId;
+
+/// A newtype-index for a constraint expression in a particular scope.
+#[newtype_index]
+pub(super) struct ScopedConstraintId;
+
+/// Can reference this * 64 total definitions inline; more will fall back to the heap.
+const INLINE_BINDING_BLOCKS: usize = 3;
+
+/// A [`BitSet`] of [`ScopedDefinitionId`], representing live bindings of a symbol in a scope.
+type Bindings = BitSet<INLINE_BINDING_BLOCKS>;
+type BindingsIterator<'a> = BitSetIterator<'a, INLINE_BINDING_BLOCKS>;
+
+/// Can reference this * 64 total declarations inline; more will fall back to the heap.
+const INLINE_DECLARATION_BLOCKS: usize = 3;
+
+/// A [`BitSet`] of [`ScopedDefinitionId`], representing live declarations of a symbol in a scope.
+type Declarations = BitSet<INLINE_DECLARATION_BLOCKS>;
+type DeclarationsIterator<'a> = BitSetIterator<'a, INLINE_DECLARATION_BLOCKS>;
+
+/// Can reference this * 64 total constraints inline; more will fall back to the heap.
+const INLINE_CONSTRAINT_BLOCKS: usize = 2;
+
+/// Can keep inline this many live bindings per symbol at a given time; more will go to heap.
+const INLINE_BINDINGS_PER_SYMBOL: usize = 4;
+
+/// One [`BitSet`] of applicable [`ScopedConstraintId`] per live binding.
+type InlineConstraintArray = [BitSet<INLINE_CONSTRAINT_BLOCKS>; INLINE_BINDINGS_PER_SYMBOL];
+type Constraints = SmallVec<InlineConstraintArray>;
+type ConstraintsIterator<'a> = std::slice::Iter<'a, BitSet<INLINE_CONSTRAINT_BLOCKS>>;
+type ConstraintsIntoIterator = smallvec::IntoIter<InlineConstraintArray>;
+
+/// Live declarations for a single symbol at some point in control flow.
+#[derive(Clone, Debug, PartialEq, Eq)]
+pub(super) struct SymbolDeclarations {
+    /// [`BitSet`]: which declarations (as [`ScopedDefinitionId`]) can reach the current location?
+    live_declarations: Declarations,
+
+    /// Could the symbol be un-declared at this point?
+    may_be_undeclared: bool,
+}
+
+impl SymbolDeclarations {
+    fn undeclared() -> Self {
+        Self {
+            live_declarations: Declarations::default(),
+            may_be_undeclared: true,
+        }
+    }
+
+    /// Record a newly-encountered declaration for this symbol.
+    fn record_declaration(&mut self, declaration_id: ScopedDefinitionId) {
+        self.live_declarations = Declarations::with(declaration_id.into());
+        self.may_be_undeclared = false;
+    }
+
+    /// Add undeclared as a possibility for this symbol.
+    fn set_may_be_undeclared(&mut self) {
+        self.may_be_undeclared = true;
+    }
+
+    /// Return an iterator over live declarations for this symbol.
+    pub(super) fn iter(&self) -> DeclarationIdIterator {
+        DeclarationIdIterator {
+            inner: self.live_declarations.iter(),
+        }
+    }
+
+    pub(super) fn is_empty(&self) -> bool {
+        self.live_declarations.is_empty()
+    }
+
+    pub(super) fn may_be_undeclared(&self) -> bool {
+        self.may_be_undeclared
+    }
+}
+
+/// Live bindings and narrowing constraints for a single symbol at some point in control flow.
+#[derive(Clone, Debug, PartialEq, Eq)]
+pub(super) struct SymbolBindings {
+    /// [`BitSet`]: which bindings (as [`ScopedDefinitionId`]) can reach the current location?
+    live_bindings: Bindings,
+
+    /// For each live binding, which [`ScopedConstraintId`] apply?
+    ///
+    /// This is a [`smallvec::SmallVec`] which should always have one [`BitSet`] of constraints per
+    /// binding in `live_bindings`.
+    constraints: Constraints,
+
+    /// Could the symbol be unbound at this point?
+    may_be_unbound: bool,
+}
+
+impl SymbolBindings {
+    fn unbound() -> Self {
+        Self {
+            live_bindings: Bindings::default(),
+            constraints: Constraints::default(),
+            may_be_unbound: true,
+        }
+    }
+
+    /// Add Unbound as a possibility for this symbol.
+    fn set_may_be_unbound(&mut self) {
+        self.may_be_unbound = true;
+    }
+
+    /// Record a newly-encountered binding for this symbol.
+    pub(super) fn record_binding(&mut self, binding_id: ScopedDefinitionId) {
+        // The new binding replaces all previous live bindings in this path, and has no
+        // constraints.
+        self.live_bindings = Bindings::with(binding_id.into());
+        self.constraints = Constraints::with_capacity(1);
+        self.constraints.push(BitSet::default());
+        self.may_be_unbound = false;
+    }
+
+    /// Add given constraint to all live bindings.
+    pub(super) fn record_constraint(&mut self, constraint_id: ScopedConstraintId) {
+        for bitset in &mut self.constraints {
+            bitset.insert(constraint_id.into());
+        }
+    }
+
+    /// Iterate over currently live bindings for this symbol.
+    pub(super) fn iter(&self) -> BindingIdWithConstraintsIterator {
+        BindingIdWithConstraintsIterator {
+            definitions: self.live_bindings.iter(),
+            constraints: self.constraints.iter(),
+        }
+    }
+
+    pub(super) fn may_be_unbound(&self) -> bool {
+        self.may_be_unbound
+    }
+}
+
+#[derive(Clone, Debug, PartialEq, Eq)]
+pub(super) struct SymbolState {
+    declarations: SymbolDeclarations,
+    bindings: SymbolBindings,
+}
+
+impl SymbolState {
+    /// Return a new [`SymbolState`] representing an unbound, undeclared symbol.
+    pub(super) fn undefined() -> Self {
+        Self {
+            declarations: SymbolDeclarations::undeclared(),
+            bindings: SymbolBindings::unbound(),
+        }
+    }
+
+    /// Add Unbound as a possibility for this symbol.
+    pub(super) fn set_may_be_unbound(&mut self) {
+        self.bindings.set_may_be_unbound();
+    }
+
+    /// Record a newly-encountered binding for this symbol.
+    pub(super) fn record_binding(&mut self, binding_id: ScopedDefinitionId) {
+        self.bindings.record_binding(binding_id);
+    }
+
+    /// Add given constraint to all live bindings.
+    pub(super) fn record_constraint(&mut self, constraint_id: ScopedConstraintId) {
+        self.bindings.record_constraint(constraint_id);
+    }
+
+    /// Add undeclared as a possibility for this symbol.
+    pub(super) fn set_may_be_undeclared(&mut self) {
+        self.declarations.set_may_be_undeclared();
+    }
+
+    /// Record a newly-encountered declaration of this symbol.
+    pub(super) fn record_declaration(&mut self, declaration_id: ScopedDefinitionId) {
+        self.declarations.record_declaration(declaration_id);
+    }
+
+    /// Merge another [`SymbolState`] into this one.
+    pub(super) fn merge(&mut self, b: SymbolState) {
+        let mut a = Self {
+            bindings: SymbolBindings {
+                live_bindings: Bindings::default(),
+                constraints: Constraints::default(),
+                may_be_unbound: self.bindings.may_be_unbound || b.bindings.may_be_unbound,
+            },
+            declarations: SymbolDeclarations {
+                live_declarations: self.declarations.live_declarations.clone(),
+                may_be_undeclared: self.declarations.may_be_undeclared
+                    || b.declarations.may_be_undeclared,
+            },
+        };
+
+        std::mem::swap(&mut a, self);
+        self.declarations
+            .live_declarations
+            .union(&b.declarations.live_declarations);
+
+        let mut a_defs_iter = a.bindings.live_bindings.iter();
+        let mut b_defs_iter = b.bindings.live_bindings.iter();
+        let mut a_constraints_iter = a.bindings.constraints.into_iter();
+        let mut b_constraints_iter = b.bindings.constraints.into_iter();
+
+        let mut opt_a_def: Option<u32> = a_defs_iter.next();
+        let mut opt_b_def: Option<u32> = b_defs_iter.next();
+
+        // Iterate through the definitions from `a` and `b`, always processing the lower definition
+        // ID first, and pushing each definition onto the merged `SymbolState` with its
+        // constraints. If a definition is found in both `a` and `b`, push it with the intersection
+        // of the constraints from the two paths; a constraint that applies from only one possible
+        // path is irrelevant.
+
+        // Helper to push `def`, with constraints in `constraints_iter`, onto `self`.
+        let push = |def, constraints_iter: &mut ConstraintsIntoIterator, merged: &mut Self| {
+            merged.bindings.live_bindings.insert(def);
+            // SAFETY: we only ever create SymbolState with either no definitions and no constraint
+            // bitsets (`::unbound`) or one definition and one constraint bitset (`::with`), and
+            // `::merge` always pushes one definition and one constraint bitset together (just
+            // below), so the number of definitions and the number of constraint bitsets can never
+            // get out of sync.
+            let constraints = constraints_iter
+                .next()
+                .expect("definitions and constraints length mismatch");
+            merged.bindings.constraints.push(constraints);
+        };
+
+        loop {
+            match (opt_a_def, opt_b_def) {
+                (Some(a_def), Some(b_def)) => match a_def.cmp(&b_def) {
+                    std::cmp::Ordering::Less => {
+                        // Next definition ID is only in `a`, push it to `self` and advance `a`.
+                        push(a_def, &mut a_constraints_iter, self);
+                        opt_a_def = a_defs_iter.next();
+                    }
+                    std::cmp::Ordering::Greater => {
+                        // Next definition ID is only in `b`, push it to `self` and advance `b`.
+                        push(b_def, &mut b_constraints_iter, self);
+                        opt_b_def = b_defs_iter.next();
+                    }
+                    std::cmp::Ordering::Equal => {
+                        // Next definition is in both; push to `self` and intersect constraints.
+                        push(a_def, &mut b_constraints_iter, self);
+                        // SAFETY: we only ever create SymbolState with either no definitions and
+                        // no constraint bitsets (`::unbound`) or one definition and one constraint
+                        // bitset (`::with`), and `::merge` always pushes one definition and one
+                        // constraint bitset together (just below), so the number of definitions
+                        // and the number of constraint bitsets can never get out of sync.
+                        let a_constraints = a_constraints_iter
+                            .next()
+                            .expect("definitions and constraints length mismatch");
+                        // If the same definition is visible through both paths, any constraint
+                        // that applies on only one path is irrelevant to the resulting type from
+                        // unioning the two paths, so we intersect the constraints.
+                        self.bindings
+                            .constraints
+                            .last_mut()
+                            .unwrap()
+                            .intersect(&a_constraints);
+                        opt_a_def = a_defs_iter.next();
+                        opt_b_def = b_defs_iter.next();
+                    }
+                },
+                (Some(a_def), None) => {
+                    // We've exhausted `b`, just push the def from `a` and move on to the next.
+                    push(a_def, &mut a_constraints_iter, self);
+                    opt_a_def = a_defs_iter.next();
+                }
+                (None, Some(b_def)) => {
+                    // We've exhausted `a`, just push the def from `b` and move on to the next.
+                    push(b_def, &mut b_constraints_iter, self);
+                    opt_b_def = b_defs_iter.next();
+                }
+                (None, None) => break,
+            }
+        }
+    }
+
+    pub(super) fn bindings(&self) -> &SymbolBindings {
+        &self.bindings
+    }
+
+    pub(super) fn declarations(&self) -> &SymbolDeclarations {
+        &self.declarations
+    }
+
+    /// Could the symbol be unbound?
+    pub(super) fn may_be_unbound(&self) -> bool {
+        self.bindings.may_be_unbound()
+    }
+}
+
+/// The default state of a symbol, if we've seen no definitions of it, is undefined (that is,
+/// both unbound and undeclared).
+impl Default for SymbolState {
+    fn default() -> Self {
+        SymbolState::undefined()
+    }
+}
+
+/// A single binding (as [`ScopedDefinitionId`]) with an iterator of its applicable
+/// [`ScopedConstraintId`].
+#[derive(Debug)]
+pub(super) struct BindingIdWithConstraints<'a> {
+    pub(super) definition: ScopedDefinitionId,
+    pub(super) constraint_ids: ConstraintIdIterator<'a>,
+}
+
+#[derive(Debug)]
+pub(super) struct BindingIdWithConstraintsIterator<'a> {
+    definitions: BindingsIterator<'a>,
+    constraints: ConstraintsIterator<'a>,
+}
+
+impl<'a> Iterator for BindingIdWithConstraintsIterator<'a> {
+    type Item = BindingIdWithConstraints<'a>;
+
+    fn next(&mut self) -> Option<Self::Item> {
+        match (self.definitions.next(), self.constraints.next()) {
+            (None, None) => None,
+            (Some(def), Some(constraints)) => Some(BindingIdWithConstraints {
+                definition: ScopedDefinitionId::from_u32(def),
+                constraint_ids: ConstraintIdIterator {
+                    wrapped: constraints.iter(),
+                },
+            }),
+            // SAFETY: see above.
+            _ => unreachable!("definitions and constraints length mismatch"),
+        }
+    }
+}
+
+impl std::iter::FusedIterator for BindingIdWithConstraintsIterator<'_> {}
+
+#[derive(Debug)]
+pub(super) struct ConstraintIdIterator<'a> {
+    wrapped: BitSetIterator<'a, INLINE_CONSTRAINT_BLOCKS>,
+}
+
+impl Iterator for ConstraintIdIterator<'_> {
+    type Item = ScopedConstraintId;
+
+    fn next(&mut self) -> Option<Self::Item> {
+        self.wrapped.next().map(ScopedConstraintId::from_u32)
+    }
+}
+
+impl std::iter::FusedIterator for ConstraintIdIterator<'_> {}
+
+#[derive(Debug)]
+pub(super) struct DeclarationIdIterator<'a> {
+    inner: DeclarationsIterator<'a>,
+}
+
+impl<'a> Iterator for DeclarationIdIterator<'a> {
+    type Item = ScopedDefinitionId;
+
+    fn next(&mut self) -> Option<Self::Item> {
+        self.inner.next().map(ScopedDefinitionId::from_u32)
+    }
+}
+
+impl std::iter::FusedIterator for DeclarationIdIterator<'_> {}
+
+#[cfg(test)]
+mod tests {
+    use super::{ScopedConstraintId, ScopedDefinitionId, SymbolState};
+
+    fn assert_bindings(symbol: &SymbolState, may_be_unbound: bool, expected: &[&str]) {
+        assert_eq!(symbol.may_be_unbound(), may_be_unbound);
+        let actual = symbol
+            .bindings()
+            .iter()
+            .map(|def_id_with_constraints| {
+                format!(
+                    "{}<{}>",
+                    def_id_with_constraints.definition.as_u32(),
+                    def_id_with_constraints
+                        .constraint_ids
+                        .map(ScopedConstraintId::as_u32)
+                        .map(|idx| idx.to_string())
+                        .collect::<Vec<_>>()
+                        .join(", ")
+                )
+            })
+            .collect::<Vec<_>>();
+        assert_eq!(actual, expected);
+    }
+
+    pub(crate) fn assert_declarations(
+        symbol: &SymbolState,
+        may_be_undeclared: bool,
+        expected: &[u32],
+    ) {
+        assert_eq!(symbol.declarations.may_be_undeclared(), may_be_undeclared);
+        let actual = symbol
+            .declarations()
+            .iter()
+            .map(ScopedDefinitionId::as_u32)
+            .collect::<Vec<_>>();
+        assert_eq!(actual, expected);
+    }
+
+    #[test]
+    fn unbound() {
+        let sym = SymbolState::undefined();
+
+        assert_bindings(&sym, true, &[]);
+    }
+
+    #[test]
+    fn with() {
+        let mut sym = SymbolState::undefined();
+        sym.record_binding(ScopedDefinitionId::from_u32(0));
+
+        assert_bindings(&sym, false, &["0<>"]);
+    }
+
+    #[test]
+    fn set_may_be_unbound() {
+        let mut sym = SymbolState::undefined();
+        sym.record_binding(ScopedDefinitionId::from_u32(0));
+        sym.set_may_be_unbound();
+
+        assert_bindings(&sym, true, &["0<>"]);
+    }
+
+    #[test]
+    fn record_constraint() {
+        let mut sym = SymbolState::undefined();
+        sym.record_binding(ScopedDefinitionId::from_u32(0));
+        sym.record_constraint(ScopedConstraintId::from_u32(0));
+
+        assert_bindings(&sym, false, &["0<0>"]);
+    }
+
+    #[test]
+    fn merge() {
+        // merging the same definition with the same constraint keeps the constraint
+        let mut sym0a = SymbolState::undefined();
+        sym0a.record_binding(ScopedDefinitionId::from_u32(0));
+        sym0a.record_constraint(ScopedConstraintId::from_u32(0));
+
+        let mut sym0b = SymbolState::undefined();
+        sym0b.record_binding(ScopedDefinitionId::from_u32(0));
+        sym0b.record_constraint(ScopedConstraintId::from_u32(0));
+
+        sym0a.merge(sym0b);
+        let mut sym0 = sym0a;
+        assert_bindings(&sym0, false, &["0<0>"]);
+
+        // merging the same definition with differing constraints drops all constraints
+        let mut sym1a = SymbolState::undefined();
+        sym1a.record_binding(ScopedDefinitionId::from_u32(1));
+        sym1a.record_constraint(ScopedConstraintId::from_u32(1));
+
+        let mut sym1b = SymbolState::undefined();
+        sym1b.record_binding(ScopedDefinitionId::from_u32(1));
+        sym1b.record_constraint(ScopedConstraintId::from_u32(2));
+
+        sym1a.merge(sym1b);
+        let sym1 = sym1a;
+        assert_bindings(&sym1, false, &["1<>"]);
+
+        // merging a constrained definition with unbound keeps both
+        let mut sym2a = SymbolState::undefined();
+        sym2a.record_binding(ScopedDefinitionId::from_u32(2));
+        sym2a.record_constraint(ScopedConstraintId::from_u32(3));
+
+        let sym2b = SymbolState::undefined();
+
+        sym2a.merge(sym2b);
+        let sym2 = sym2a;
+        assert_bindings(&sym2, true, &["2<3>"]);
+
+        // merging different definitions keeps them each with their existing constraints
+        sym0.merge(sym2);
+        let sym = sym0;
+        assert_bindings(&sym, true, &["0<0>", "2<3>"]);
+    }
+
+    #[test]
+    fn no_declaration() {
+        let sym = SymbolState::undefined();
+
+        assert_declarations(&sym, true, &[]);
+    }
+
+    #[test]
+    fn record_declaration() {
+        let mut sym = SymbolState::undefined();
+        sym.record_declaration(ScopedDefinitionId::from_u32(1));
+
+        assert_declarations(&sym, false, &[1]);
+    }
+
+    #[test]
+    fn record_declaration_override() {
+        let mut sym = SymbolState::undefined();
+        sym.record_declaration(ScopedDefinitionId::from_u32(1));
+        sym.record_declaration(ScopedDefinitionId::from_u32(2));
+
+        assert_declarations(&sym, false, &[2]);
+    }
+
+    #[test]
+    fn record_declaration_merge() {
+        let mut sym = SymbolState::undefined();
+        sym.record_declaration(ScopedDefinitionId::from_u32(1));
+
+        let mut sym2 = SymbolState::undefined();
+        sym2.record_declaration(ScopedDefinitionId::from_u32(2));
+
+        sym.merge(sym2);
+
+        assert_declarations(&sym, false, &[1, 2]);
+    }
+
+    #[test]
+    fn record_declaration_merge_partial_undeclared() {
+        let mut sym = SymbolState::undefined();
+        sym.record_declaration(ScopedDefinitionId::from_u32(1));
+
+        let sym2 = SymbolState::undefined();
+
+        sym.merge(sym2);
+
+        assert_declarations(&sym, true, &[1]);
+    }
+
+    #[test]
+    fn set_may_be_undeclared() {
+        let mut sym = SymbolState::undefined();
+        sym.record_declaration(ScopedDefinitionId::from_u32(0));
+        sym.set_may_be_undeclared();
+
+        assert_declarations(&sym, true, &[0]);
+    }
+}

crates/red_knot_python_semantic/src/semantic_model.rs

@@ -0,0 +1,250 @@
+use ruff_db::files::{File, FilePath};
+use ruff_db::source::line_index;
+use ruff_python_ast as ast;
+use ruff_python_ast::{Expr, ExpressionRef};
+use ruff_source_file::LineIndex;
+
+use crate::module_name::ModuleName;
+use crate::module_resolver::{resolve_module, Module};
+use crate::semantic_index::ast_ids::HasScopedAstId;
+use crate::semantic_index::semantic_index;
+use crate::types::{binding_ty, global_symbol_ty, infer_scope_types, Type};
+use crate::Db;
+
+pub struct SemanticModel<'db> {
+    db: &'db dyn Db,
+    file: File,
+}
+
+impl<'db> SemanticModel<'db> {
+    pub fn new(db: &'db dyn Db, file: File) -> Self {
+        Self { db, file }
+    }
+
+    // TODO we don't actually want to expose the Db directly to lint rules, but we need to find a
+    // solution for exposing information from types
+    pub fn db(&self) -> &dyn Db {
+        self.db
+    }
+
+    pub fn file_path(&self) -> &FilePath {
+        self.file.path(self.db)
+    }
+
+    pub fn line_index(&self) -> LineIndex {
+        line_index(self.db.upcast(), self.file)
+    }
+
+    pub fn resolve_module(&self, module_name: ModuleName) -> Option<Module> {
+        resolve_module(self.db, module_name)
+    }
+
+    pub fn global_symbol_ty(&self, module: &Module, symbol_name: &str) -> Type<'db> {
+        global_symbol_ty(self.db, module.file(), symbol_name)
+    }
+}
+
+pub trait HasTy {
+    /// Returns the inferred type of `self`.
+    ///
+    /// ## Panics
+    /// May panic if `self` is from another file than `model`.
+    fn ty<'db>(&self, model: &SemanticModel<'db>) -> Type<'db>;
+}
+
+impl HasTy for ast::ExpressionRef<'_> {
+    fn ty<'db>(&self, model: &SemanticModel<'db>) -> Type<'db> {
+        let index = semantic_index(model.db, model.file);
+        let file_scope = index.expression_scope_id(*self);
+        let scope = file_scope.to_scope_id(model.db, model.file);
+
+        let expression_id = self.scoped_ast_id(model.db, scope);
+        infer_scope_types(model.db, scope).expression_ty(expression_id)
+    }
+}
+
+macro_rules! impl_expression_has_ty {
+    ($ty: ty) => {
+        impl HasTy for $ty {
+            #[inline]
+            fn ty<'db>(&self, model: &SemanticModel<'db>) -> Type<'db> {
+                let expression_ref = ExpressionRef::from(self);
+                expression_ref.ty(model)
+            }
+        }
+    };
+}
+
+impl_expression_has_ty!(ast::ExprBoolOp);
+impl_expression_has_ty!(ast::ExprNamed);
+impl_expression_has_ty!(ast::ExprBinOp);
+impl_expression_has_ty!(ast::ExprUnaryOp);
+impl_expression_has_ty!(ast::ExprLambda);
+impl_expression_has_ty!(ast::ExprIf);
+impl_expression_has_ty!(ast::ExprDict);
+impl_expression_has_ty!(ast::ExprSet);
+impl_expression_has_ty!(ast::ExprListComp);
+impl_expression_has_ty!(ast::ExprSetComp);
+impl_expression_has_ty!(ast::ExprDictComp);
+impl_expression_has_ty!(ast::ExprGenerator);
+impl_expression_has_ty!(ast::ExprAwait);
+impl_expression_has_ty!(ast::ExprYield);
+impl_expression_has_ty!(ast::ExprYieldFrom);
+impl_expression_has_ty!(ast::ExprCompare);
+impl_expression_has_ty!(ast::ExprCall);
+impl_expression_has_ty!(ast::ExprFString);
+impl_expression_has_ty!(ast::ExprStringLiteral);
+impl_expression_has_ty!(ast::ExprBytesLiteral);
+impl_expression_has_ty!(ast::ExprNumberLiteral);
+impl_expression_has_ty!(ast::ExprBooleanLiteral);
+impl_expression_has_ty!(ast::ExprNoneLiteral);
+impl_expression_has_ty!(ast::ExprEllipsisLiteral);
+impl_expression_has_ty!(ast::ExprAttribute);
+impl_expression_has_ty!(ast::ExprSubscript);
+impl_expression_has_ty!(ast::ExprStarred);
+impl_expression_has_ty!(ast::ExprName);
+impl_expression_has_ty!(ast::ExprList);
+impl_expression_has_ty!(ast::ExprTuple);
+impl_expression_has_ty!(ast::ExprSlice);
+impl_expression_has_ty!(ast::ExprIpyEscapeCommand);
+
+impl HasTy for ast::Expr {
+    fn ty<'db>(&self, model: &SemanticModel<'db>) -> Type<'db> {
+        match self {
+            Expr::BoolOp(inner) => inner.ty(model),
+            Expr::Named(inner) => inner.ty(model),
+            Expr::BinOp(inner) => inner.ty(model),
+            Expr::UnaryOp(inner) => inner.ty(model),
+            Expr::Lambda(inner) => inner.ty(model),
+            Expr::If(inner) => inner.ty(model),
+            Expr::Dict(inner) => inner.ty(model),
+            Expr::Set(inner) => inner.ty(model),
+            Expr::ListComp(inner) => inner.ty(model),
+            Expr::SetComp(inner) => inner.ty(model),
+            Expr::DictComp(inner) => inner.ty(model),
+            Expr::Generator(inner) => inner.ty(model),
+            Expr::Await(inner) => inner.ty(model),
+            Expr::Yield(inner) => inner.ty(model),
+            Expr::YieldFrom(inner) => inner.ty(model),
+            Expr::Compare(inner) => inner.ty(model),
+            Expr::Call(inner) => inner.ty(model),
+            Expr::FString(inner) => inner.ty(model),
+            Expr::StringLiteral(inner) => inner.ty(model),
+            Expr::BytesLiteral(inner) => inner.ty(model),
+            Expr::NumberLiteral(inner) => inner.ty(model),
+            Expr::BooleanLiteral(inner) => inner.ty(model),
+            Expr::NoneLiteral(inner) => inner.ty(model),
+            Expr::EllipsisLiteral(inner) => inner.ty(model),
+            Expr::Attribute(inner) => inner.ty(model),
+            Expr::Subscript(inner) => inner.ty(model),
+            Expr::Starred(inner) => inner.ty(model),
+            Expr::Name(inner) => inner.ty(model),
+            Expr::List(inner) => inner.ty(model),
+            Expr::Tuple(inner) => inner.ty(model),
+            Expr::Slice(inner) => inner.ty(model),
+            Expr::IpyEscapeCommand(inner) => inner.ty(model),
+        }
+    }
+}
+
+macro_rules! impl_binding_has_ty {
+    ($ty: ty) => {
+        impl HasTy for $ty {
+            #[inline]
+            fn ty<'db>(&self, model: &SemanticModel<'db>) -> Type<'db> {
+                let index = semantic_index(model.db, model.file);
+                let binding = index.definition(self);
+                binding_ty(model.db, binding)
+            }
+        }
+    };
+}
+
+impl_binding_has_ty!(ast::StmtFunctionDef);
+impl_binding_has_ty!(ast::StmtClassDef);
+impl_binding_has_ty!(ast::Alias);
+impl_binding_has_ty!(ast::Parameter);
+impl_binding_has_ty!(ast::ParameterWithDefault);
+
+#[cfg(test)]
+mod tests {
+    use ruff_db::files::system_path_to_file;
+    use ruff_db::parsed::parsed_module;
+    use ruff_db::system::{DbWithTestSystem, SystemPathBuf};
+
+    use crate::db::tests::TestDb;
+    use crate::program::{Program, SearchPathSettings};
+    use crate::python_version::PythonVersion;
+    use crate::types::Type;
+    use crate::{HasTy, ProgramSettings, SemanticModel};
+
+    fn setup_db<'a>(files: impl IntoIterator<Item = (&'a str, &'a str)>) -> anyhow::Result<TestDb> {
+        let mut db = TestDb::new();
+        db.write_files(files)?;
+
+        Program::from_settings(
+            &db,
+            &ProgramSettings {
+                target_version: PythonVersion::default(),
+                search_paths: SearchPathSettings::new(SystemPathBuf::from("/src")),
+            },
+        )?;
+
+        Ok(db)
+    }
+
+    #[test]
+    fn function_ty() -> anyhow::Result<()> {
+        let db = setup_db([("/src/foo.py", "def test(): pass")])?;
+
+        let foo = system_path_to_file(&db, "/src/foo.py").unwrap();
+
+        let ast = parsed_module(&db, foo);
+
+        let function = ast.suite()[0].as_function_def_stmt().unwrap();
+        let model = SemanticModel::new(&db, foo);
+        let ty = function.ty(&model);
+
+        assert!(matches!(ty, Type::Function(_)));
+
+        Ok(())
+    }
+
+    #[test]
+    fn class_ty() -> anyhow::Result<()> {
+        let db = setup_db([("/src/foo.py", "class Test: pass")])?;
+
+        let foo = system_path_to_file(&db, "/src/foo.py").unwrap();
+
+        let ast = parsed_module(&db, foo);
+
+        let class = ast.suite()[0].as_class_def_stmt().unwrap();
+        let model = SemanticModel::new(&db, foo);
+        let ty = class.ty(&model);
+
+        assert!(matches!(ty, Type::Class(_)));
+
+        Ok(())
+    }
+
+    #[test]
+    fn alias_ty() -> anyhow::Result<()> {
+        let db = setup_db([
+            ("/src/foo.py", "class Test: pass"),
+            ("/src/bar.py", "from foo import Test"),
+        ])?;
+
+        let bar = system_path_to_file(&db, "/src/bar.py").unwrap();
+
+        let ast = parsed_module(&db, bar);
+
+        let import = ast.suite()[0].as_import_from_stmt().unwrap();
+        let alias = &import.names[0];
+        let model = SemanticModel::new(&db, bar);
+        let ty = alias.ty(&model);
+
+        assert!(matches!(ty, Type::Class(_)));
+
+        Ok(())
+    }
+}

crates/red_knot_python_semantic/src/stdlib.rs

@@ -0,0 +1,87 @@
+use crate::module_name::ModuleName;
+use crate::module_resolver::resolve_module;
+use crate::semantic_index::global_scope;
+use crate::semantic_index::symbol::ScopeId;
+use crate::types::{global_symbol_ty, Type};
+use crate::Db;
+
+/// Enumeration of various core stdlib modules, for which we have dedicated Salsa queries.
+#[derive(Debug, Clone, Copy, PartialEq, Eq)]
+enum CoreStdlibModule {
+    Builtins,
+    Types,
+    Typeshed,
+    TypingExtensions,
+}
+
+impl CoreStdlibModule {
+    fn name(self) -> ModuleName {
+        let module_name = match self {
+            Self::Builtins => "builtins",
+            Self::Types => "types",
+            Self::Typeshed => "_typeshed",
+            Self::TypingExtensions => "typing_extensions",
+        };
+        ModuleName::new_static(module_name)
+            .unwrap_or_else(|| panic!("{module_name} should be a valid module name!"))
+    }
+}
+
+/// Lookup the type of `symbol` in a given core module
+///
+/// Returns `Unbound` if the given core module cannot be resolved for some reason
+fn core_module_symbol_ty<'db>(
+    db: &'db dyn Db,
+    core_module: CoreStdlibModule,
+    symbol: &str,
+) -> Type<'db> {
+    resolve_module(db, core_module.name())
+        .map(|module| global_symbol_ty(db, module.file(), symbol))
+        .unwrap_or(Type::Unbound)
+}
+
+/// Lookup the type of `symbol` in the builtins namespace.
+///
+/// Returns `Unbound` if the `builtins` module isn't available for some reason.
+#[inline]
+pub(crate) fn builtins_symbol_ty<'db>(db: &'db dyn Db, symbol: &str) -> Type<'db> {
+    core_module_symbol_ty(db, CoreStdlibModule::Builtins, symbol)
+}
+
+/// Lookup the type of `symbol` in the `types` module namespace.
+///
+/// Returns `Unbound` if the `types` module isn't available for some reason.
+#[inline]
+pub(crate) fn types_symbol_ty<'db>(db: &'db dyn Db, symbol: &str) -> Type<'db> {
+    core_module_symbol_ty(db, CoreStdlibModule::Types, symbol)
+}
+
+/// Lookup the type of `symbol` in the `_typeshed` module namespace.
+///
+/// Returns `Unbound` if the `_typeshed` module isn't available for some reason.
+#[inline]
+pub(crate) fn typeshed_symbol_ty<'db>(db: &'db dyn Db, symbol: &str) -> Type<'db> {
+    core_module_symbol_ty(db, CoreStdlibModule::Typeshed, symbol)
+}
+
+/// Lookup the type of `symbol` in the `typing_extensions` module namespace.
+///
+/// Returns `Unbound` if the `typing_extensions` module isn't available for some reason.
+#[inline]
+pub(crate) fn typing_extensions_symbol_ty<'db>(db: &'db dyn Db, symbol: &str) -> Type<'db> {
+    core_module_symbol_ty(db, CoreStdlibModule::TypingExtensions, symbol)
+}
+
+/// Get the scope of a core stdlib module.
+///
+/// Can return `None` if a custom typeshed is used that is missing the core module in question.
+fn core_module_scope(db: &dyn Db, core_module: CoreStdlibModule) -> Option<ScopeId<'_>> {
+    resolve_module(db, core_module.name()).map(|module| global_scope(db, module.file()))
+}
+
+/// Get the `builtins` module scope.
+///
+/// Can return `None` if a custom typeshed is used that is missing `builtins.pyi`.
+pub(crate) fn builtins_module_scope(db: &dyn Db) -> Option<ScopeId<'_>> {
+    core_module_scope(db, CoreStdlibModule::Builtins)
+}

crates/red_knot_python_semantic/src/types/builder.rs

@@ -0,0 +1,578 @@
+//! Smart builders for union and intersection types.
+//!
+//! Invariants we maintain here:
+//!   * No single-element union types (should just be the contained type instead.)
+//!   * No single-positive-element intersection types. Single-negative-element are OK, we don't
+//!     have a standalone negation type so there's no other representation for this.
+//!   * The same type should never appear more than once in a union or intersection. (This should
+//!     be expanded to cover subtyping -- see below -- but for now we only implement it for type
+//!     identity.)
+//!   * Disjunctive normal form (DNF): the tree of unions and intersections can never be deeper
+//!     than a union-of-intersections. Unions cannot contain other unions (the inner union just
+//!     flattens into the outer one), intersections cannot contain other intersections (also
+//!     flattens), and intersections cannot contain unions (the intersection distributes over the
+//!     union, inverting it into a union-of-intersections).
+//!
+//! The implication of these invariants is that a [`UnionBuilder`] does not necessarily build a
+//! [`Type::Union`]. For example, if only one type is added to the [`UnionBuilder`], `build()` will
+//! just return that type directly. The same is true for [`IntersectionBuilder`]; for example, if a
+//! union type is added to the intersection, it will distribute and [`IntersectionBuilder::build`]
+//! may end up returning a [`Type::Union`] of intersections.
+//!
+//! In the future we should have these additional invariants, but they aren't implemented yet:
+//!   * No type in a union can be a subtype of any other type in the union (just eliminate the
+//!     subtype from the union).
+//!   * No type in an intersection can be a supertype of any other type in the intersection (just
+//!     eliminate the supertype from the intersection).
+//!   * An intersection containing two non-overlapping types should simplify to [`Type::Never`].
+use crate::types::{IntersectionType, Type, UnionType};
+use crate::{Db, FxOrderSet};
+use smallvec::SmallVec;
+
+use super::KnownClass;
+
+pub(crate) struct UnionBuilder<'db> {
+    elements: Vec<Type<'db>>,
+    db: &'db dyn Db,
+}
+
+impl<'db> UnionBuilder<'db> {
+    pub(crate) fn new(db: &'db dyn Db) -> Self {
+        Self {
+            db,
+            elements: vec![],
+        }
+    }
+
+    /// Adds a type to this union.
+    pub(crate) fn add(mut self, ty: Type<'db>) -> Self {
+        match ty {
+            Type::Union(union) => {
+                let new_elements = union.elements(self.db);
+                self.elements.reserve(new_elements.len());
+                for element in new_elements {
+                    self = self.add(*element);
+                }
+            }
+            Type::Never => {}
+            _ => {
+                let bool_pair = if let Type::BooleanLiteral(b) = ty {
+                    Some(Type::BooleanLiteral(!b))
+                } else {
+                    None
+                };
+
+                let mut to_add = ty;
+                let mut to_remove = SmallVec::<[usize; 2]>::new();
+                for (index, element) in self.elements.iter().enumerate() {
+                    if Some(*element) == bool_pair {
+                        to_add = KnownClass::Bool.to_instance(self.db);
+                        to_remove.push(index);
+                        // The type we are adding is a BooleanLiteral, which doesn't have any
+                        // subtypes. And we just found that the union already contained our
+                        // mirror-image BooleanLiteral, so it can't also contain bool or any
+                        // supertype of bool. Therefore, we are done.
+                        break;
+                    }
+                    if ty.is_subtype_of(self.db, *element) {
+                        return self;
+                    } else if element.is_subtype_of(self.db, ty) {
+                        to_remove.push(index);
+                    }
+                }
+
+                match to_remove[..] {
+                    [] => self.elements.push(to_add),
+                    [index] => self.elements[index] = to_add,
+                    _ => {
+                        let mut current_index = 0;
+                        let mut to_remove = to_remove.into_iter();
+                        let mut next_to_remove_index = to_remove.next();
+                        self.elements.retain(|_| {
+                            let retain = if Some(current_index) == next_to_remove_index {
+                                next_to_remove_index = to_remove.next();
+                                false
+                            } else {
+                                true
+                            };
+                            current_index += 1;
+                            retain
+                        });
+                        self.elements.push(to_add);
+                    }
+                }
+            }
+        }
+
+        self
+    }
+
+    pub(crate) fn build(self) -> Type<'db> {
+        match self.elements.len() {
+            0 => Type::Never,
+            1 => self.elements[0],
+            _ => Type::Union(UnionType::new(self.db, self.elements.into())),
+        }
+    }
+}
+
+#[derive(Clone)]
+pub(crate) struct IntersectionBuilder<'db> {
+    // Really this builds a union-of-intersections, because we always keep our set-theoretic types
+    // in disjunctive normal form (DNF), a union of intersections. In the simplest case there's
+    // just a single intersection in this vector, and we are building a single intersection type,
+    // but if a union is added to the intersection, we'll distribute ourselves over that union and
+    // create a union of intersections.
+    intersections: Vec<InnerIntersectionBuilder<'db>>,
+    db: &'db dyn Db,
+}
+
+impl<'db> IntersectionBuilder<'db> {
+    pub(crate) fn new(db: &'db dyn Db) -> Self {
+        Self {
+            db,
+            intersections: vec![InnerIntersectionBuilder::new()],
+        }
+    }
+
+    fn empty(db: &'db dyn Db) -> Self {
+        Self {
+            db,
+            intersections: vec![],
+        }
+    }
+
+    pub(crate) fn add_positive(mut self, ty: Type<'db>) -> Self {
+        if let Type::Union(union) = ty {
+            // Distribute ourself over this union: for each union element, clone ourself and
+            // intersect with that union element, then create a new union-of-intersections with all
+            // of those sub-intersections in it. E.g. if `self` is a simple intersection `T1 & T2`
+            // and we add `T3 | T4` to the intersection, we don't get `T1 & T2 & (T3 | T4)` (that's
+            // not in DNF), we distribute the union and get `(T1 & T3) | (T2 & T3) | (T1 & T4) |
+            // (T2 & T4)`. If `self` is already a union-of-intersections `(T1 & T2) | (T3 & T4)`
+            // and we add `T5 | T6` to it, that flattens all the way out to `(T1 & T2 & T5) | (T1 &
+            // T2 & T6) | (T3 & T4 & T5) ...` -- you get the idea.
+            union
+                .elements(self.db)
+                .iter()
+                .map(|elem| self.clone().add_positive(*elem))
+                .fold(IntersectionBuilder::empty(self.db), |mut builder, sub| {
+                    builder.intersections.extend(sub.intersections);
+                    builder
+                })
+        } else {
+            // If we are already a union-of-intersections, distribute the new intersected element
+            // across all of those intersections.
+            for inner in &mut self.intersections {
+                inner.add_positive(self.db, ty);
+            }
+            self
+        }
+    }
+
+    pub(crate) fn add_negative(mut self, ty: Type<'db>) -> Self {
+        // See comments above in `add_positive`; this is just the negated version.
+        if let Type::Union(union) = ty {
+            union
+                .elements(self.db)
+                .iter()
+                .map(|elem| self.clone().add_negative(*elem))
+                .fold(IntersectionBuilder::empty(self.db), |mut builder, sub| {
+                    builder.intersections.extend(sub.intersections);
+                    builder
+                })
+        } else {
+            for inner in &mut self.intersections {
+                inner.add_negative(self.db, ty);
+            }
+            self
+        }
+    }
+
+    pub(crate) fn build(mut self) -> Type<'db> {
+        // Avoid allocating the UnionBuilder unnecessarily if we have just one intersection:
+        if self.intersections.len() == 1 {
+            self.intersections.pop().unwrap().build(self.db)
+        } else {
+            UnionType::from_elements(
+                self.db,
+                self.intersections
+                    .into_iter()
+                    .map(|inner| inner.build(self.db)),
+            )
+        }
+    }
+}
+
+#[derive(Debug, Clone, Default)]
+struct InnerIntersectionBuilder<'db> {
+    positive: FxOrderSet<Type<'db>>,
+    negative: FxOrderSet<Type<'db>>,
+}
+
+impl<'db> InnerIntersectionBuilder<'db> {
+    fn new() -> Self {
+        Self::default()
+    }
+
+    /// Adds a positive type to this intersection.
+    fn add_positive(&mut self, db: &'db dyn Db, ty: Type<'db>) {
+        // TODO `Any`/`Unknown`/`Todo` actually should not self-cancel
+        match ty {
+            Type::Intersection(inter) => {
+                let pos = inter.positive(db);
+                let neg = inter.negative(db);
+                self.positive.extend(pos.difference(&self.negative));
+                self.negative.extend(neg.difference(&self.positive));
+                self.positive.retain(|elem| !neg.contains(elem));
+                self.negative.retain(|elem| !pos.contains(elem));
+            }
+            _ => {
+                if !self.negative.remove(&ty) {
+                    self.positive.insert(ty);
+                };
+            }
+        }
+    }
+
+    /// Adds a negative type to this intersection.
+    fn add_negative(&mut self, db: &'db dyn Db, ty: Type<'db>) {
+        // TODO `Any`/`Unknown`/`Todo` actually should not self-cancel
+        match ty {
+            Type::Intersection(intersection) => {
+                let pos = intersection.negative(db);
+                let neg = intersection.positive(db);
+                self.positive.extend(pos.difference(&self.negative));
+                self.negative.extend(neg.difference(&self.positive));
+                self.positive.retain(|elem| !neg.contains(elem));
+                self.negative.retain(|elem| !pos.contains(elem));
+            }
+            Type::Never => {}
+            Type::Unbound => {}
+            _ => {
+                if !self.positive.remove(&ty) {
+                    self.negative.insert(ty);
+                };
+            }
+        }
+    }
+
+    fn simplify(&mut self) {
+        // TODO this should be generalized based on subtyping, for now we just handle a few cases
+
+        // Never is a subtype of all types
+        if self.positive.contains(&Type::Never) {
+            self.positive.retain(Type::is_never);
+            self.negative.clear();
+        }
+
+        if self.positive.contains(&Type::Unbound) {
+            self.positive.retain(Type::is_unbound);
+            self.negative.clear();
+        }
+
+        // None intersects only with object
+        for pos in &self.positive {
+            if let Type::Instance(_) = pos {
+                // could be `object` type
+            } else {
+                self.negative.remove(&Type::None);
+                break;
+            }
+        }
+    }
+
+    fn build(mut self, db: &'db dyn Db) -> Type<'db> {
+        self.simplify();
+        match (self.positive.len(), self.negative.len()) {
+            (0, 0) => Type::Never,
+            (1, 0) => self.positive[0],
+            _ => {
+                self.positive.shrink_to_fit();
+                self.negative.shrink_to_fit();
+                Type::Intersection(IntersectionType::new(db, self.positive, self.negative))
+            }
+        }
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::{IntersectionBuilder, IntersectionType, Type, UnionType};
+    use crate::db::tests::TestDb;
+    use crate::program::{Program, SearchPathSettings};
+    use crate::python_version::PythonVersion;
+    use crate::types::{KnownClass, UnionBuilder};
+    use crate::ProgramSettings;
+    use ruff_db::system::{DbWithTestSystem, SystemPathBuf};
+
+    fn setup_db() -> TestDb {
+        let db = TestDb::new();
+
+        let src_root = SystemPathBuf::from("/src");
+        db.memory_file_system()
+            .create_directory_all(&src_root)
+            .unwrap();
+
+        Program::from_settings(
+            &db,
+            &ProgramSettings {
+                target_version: PythonVersion::default(),
+                search_paths: SearchPathSettings::new(src_root),
+            },
+        )
+        .expect("Valid search path settings");
+
+        db
+    }
+
+    #[test]
+    fn build_union() {
+        let db = setup_db();
+        let t0 = Type::IntLiteral(0);
+        let t1 = Type::IntLiteral(1);
+        let union = UnionType::from_elements(&db, [t0, t1]).expect_union();
+
+        assert_eq!(union.elements(&db), &[t0, t1]);
+    }
+
+    #[test]
+    fn build_union_single() {
+        let db = setup_db();
+        let t0 = Type::IntLiteral(0);
+        let ty = UnionType::from_elements(&db, [t0]);
+        assert_eq!(ty, t0);
+    }
+
+    #[test]
+    fn build_union_empty() {
+        let db = setup_db();
+        let ty = UnionBuilder::new(&db).build();
+        assert_eq!(ty, Type::Never);
+    }
+
+    #[test]
+    fn build_union_never() {
+        let db = setup_db();
+        let t0 = Type::IntLiteral(0);
+        let ty = UnionType::from_elements(&db, [t0, Type::Never]);
+        assert_eq!(ty, t0);
+    }
+
+    #[test]
+    fn build_union_bool() {
+        let db = setup_db();
+        let bool_instance_ty = KnownClass::Bool.to_instance(&db);
+
+        let t0 = Type::BooleanLiteral(true);
+        let t1 = Type::BooleanLiteral(true);
+        let t2 = Type::BooleanLiteral(false);
+        let t3 = Type::IntLiteral(17);
+
+        let union = UnionType::from_elements(&db, [t0, t1, t3]).expect_union();
+        assert_eq!(union.elements(&db), &[t0, t3]);
+
+        let union = UnionType::from_elements(&db, [t0, t1, t2, t3]).expect_union();
+        assert_eq!(union.elements(&db), &[bool_instance_ty, t3]);
+    }
+
+    #[test]
+    fn build_union_flatten() {
+        let db = setup_db();
+        let t0 = Type::IntLiteral(0);
+        let t1 = Type::IntLiteral(1);
+        let t2 = Type::IntLiteral(2);
+        let u1 = UnionType::from_elements(&db, [t0, t1]);
+        let union = UnionType::from_elements(&db, [u1, t2]).expect_union();
+
+        assert_eq!(union.elements(&db), &[t0, t1, t2]);
+    }
+
+    #[test]
+    fn build_union_simplify_subtype() {
+        let db = setup_db();
+        let t0 = KnownClass::Str.to_instance(&db);
+        let t1 = Type::LiteralString;
+        let u0 = UnionType::from_elements(&db, [t0, t1]);
+        let u1 = UnionType::from_elements(&db, [t1, t0]);
+
+        assert_eq!(u0, t0);
+        assert_eq!(u1, t0);
+    }
+
+    #[test]
+    fn build_union_no_simplify_unknown() {
+        let db = setup_db();
+        let t0 = KnownClass::Str.to_instance(&db);
+        let t1 = Type::Unknown;
+        let u0 = UnionType::from_elements(&db, [t0, t1]);
+        let u1 = UnionType::from_elements(&db, [t1, t0]);
+
+        assert_eq!(u0.expect_union().elements(&db), &[t0, t1]);
+        assert_eq!(u1.expect_union().elements(&db), &[t1, t0]);
+    }
+
+    #[test]
+    fn build_union_subsume_multiple() {
+        let db = setup_db();
+        let str_ty = KnownClass::Str.to_instance(&db);
+        let int_ty = KnownClass::Int.to_instance(&db);
+        let object_ty = KnownClass::Object.to_instance(&db);
+        let unknown_ty = Type::Unknown;
+
+        let u0 = UnionType::from_elements(&db, [str_ty, unknown_ty, int_ty, object_ty]);
+
+        assert_eq!(u0.expect_union().elements(&db), &[unknown_ty, object_ty]);
+    }
+
+    impl<'db> IntersectionType<'db> {
+        fn pos_vec(self, db: &'db TestDb) -> Vec<Type<'db>> {
+            self.positive(db).into_iter().copied().collect()
+        }
+
+        fn neg_vec(self, db: &'db TestDb) -> Vec<Type<'db>> {
+            self.negative(db).into_iter().copied().collect()
+        }
+    }
+
+    #[test]
+    fn build_intersection() {
+        let db = setup_db();
+        let t0 = Type::IntLiteral(0);
+        let ta = Type::Any;
+        let intersection = IntersectionBuilder::new(&db)
+            .add_positive(ta)
+            .add_negative(t0)
+            .build()
+            .expect_intersection();
+
+        assert_eq!(intersection.pos_vec(&db), &[ta]);
+        assert_eq!(intersection.neg_vec(&db), &[t0]);
+    }
+
+    #[test]
+    fn build_intersection_flatten_positive() {
+        let db = setup_db();
+        let ta = Type::Any;
+        let t1 = Type::IntLiteral(1);
+        let t2 = Type::IntLiteral(2);
+        let i0 = IntersectionBuilder::new(&db)
+            .add_positive(ta)
+            .add_negative(t1)
+            .build();
+        let intersection = IntersectionBuilder::new(&db)
+            .add_positive(t2)
+            .add_positive(i0)
+            .build()
+            .expect_intersection();
+
+        assert_eq!(intersection.pos_vec(&db), &[t2, ta]);
+        assert_eq!(intersection.neg_vec(&db), &[t1]);
+    }
+
+    #[test]
+    fn build_intersection_flatten_negative() {
+        let db = setup_db();
+        let ta = Type::Any;
+        let t1 = Type::IntLiteral(1);
+        let t2 = Type::IntLiteral(2);
+        let i0 = IntersectionBuilder::new(&db)
+            .add_positive(ta)
+            .add_negative(t1)
+            .build();
+        let intersection = IntersectionBuilder::new(&db)
+            .add_positive(t2)
+            .add_negative(i0)
+            .build()
+            .expect_intersection();
+
+        assert_eq!(intersection.pos_vec(&db), &[t2, t1]);
+        assert_eq!(intersection.neg_vec(&db), &[ta]);
+    }
+
+    #[test]
+    fn intersection_distributes_over_union() {
+        let db = setup_db();
+        let t0 = Type::IntLiteral(0);
+        let t1 = Type::IntLiteral(1);
+        let ta = Type::Any;
+        let u0 = UnionType::from_elements(&db, [t0, t1]);
+
+        let union = IntersectionBuilder::new(&db)
+            .add_positive(ta)
+            .add_positive(u0)
+            .build()
+            .expect_union();
+        let [Type::Intersection(i0), Type::Intersection(i1)] = union.elements(&db)[..] else {
+            panic!("expected a union of two intersections");
+        };
+        assert_eq!(i0.pos_vec(&db), &[ta, t0]);
+        assert_eq!(i1.pos_vec(&db), &[ta, t1]);
+    }
+
+    #[test]
+    fn build_intersection_self_negation() {
+        let db = setup_db();
+        let ty = IntersectionBuilder::new(&db)
+            .add_positive(Type::None)
+            .add_negative(Type::None)
+            .build();
+
+        assert_eq!(ty, Type::Never);
+    }
+
+    #[test]
+    fn build_intersection_simplify_negative_never() {
+        let db = setup_db();
+        let ty = IntersectionBuilder::new(&db)
+            .add_positive(Type::None)
+            .add_negative(Type::Never)
+            .build();
+
+        assert_eq!(ty, Type::None);
+    }
+
+    #[test]
+    fn build_intersection_simplify_positive_never() {
+        let db = setup_db();
+        let ty = IntersectionBuilder::new(&db)
+            .add_positive(Type::None)
+            .add_positive(Type::Never)
+            .build();
+
+        assert_eq!(ty, Type::Never);
+    }
+
+    #[test]
+    fn build_intersection_simplify_positive_unbound() {
+        let db = setup_db();
+        let ty = IntersectionBuilder::new(&db)
+            .add_positive(Type::Unbound)
+            .add_positive(Type::IntLiteral(1))
+            .build();
+
+        assert_eq!(ty, Type::Unbound);
+    }
+
+    #[test]
+    fn build_intersection_simplify_negative_unbound() {
+        let db = setup_db();
+        let ty = IntersectionBuilder::new(&db)
+            .add_negative(Type::Unbound)
+            .add_positive(Type::IntLiteral(1))
+            .build();
+
+        assert_eq!(ty, Type::IntLiteral(1));
+    }
+
+    #[test]
+    fn build_intersection_simplify_negative_none() {
+        let db = setup_db();
+        let ty = IntersectionBuilder::new(&db)
+            .add_negative(Type::None)
+            .add_positive(Type::IntLiteral(1))
+            .build();
+
+        assert_eq!(ty, Type::IntLiteral(1));
+    }
+}

crates/red_knot_python_semantic/src/types/diagnostic.rs

@@ -0,0 +1,111 @@
+use ruff_db::files::File;
+use ruff_text_size::{Ranged, TextRange};
+use std::fmt::Formatter;
+use std::ops::Deref;
+use std::sync::Arc;
+
+#[derive(Debug, Eq, PartialEq)]
+pub struct TypeCheckDiagnostic {
+    // TODO: Don't use string keys for rules
+    pub(super) rule: String,
+    pub(super) message: String,
+    pub(super) range: TextRange,
+    pub(super) file: File,
+}
+
+impl TypeCheckDiagnostic {
+    pub fn rule(&self) -> &str {
+        &self.rule
+    }
+
+    pub fn message(&self) -> &str {
+        &self.message
+    }
+
+    pub fn file(&self) -> File {
+        self.file
+    }
+}
+
+impl Ranged for TypeCheckDiagnostic {
+    fn range(&self) -> TextRange {
+        self.range
+    }
+}
+
+/// A collection of type check diagnostics.
+///
+/// The diagnostics are wrapped in an `Arc` because they need to be cloned multiple times
+/// when going from `infer_expression` to `check_file`. We could consider
+/// making [`TypeCheckDiagnostic`] a Salsa struct to have them Arena-allocated (once the Tables refactor is done).
+/// Using Salsa struct does have the downside that it leaks the Salsa dependency into diagnostics and
+/// each Salsa-struct comes with an overhead.
+#[derive(Default, Eq, PartialEq)]
+pub struct TypeCheckDiagnostics {
+    inner: Vec<std::sync::Arc<TypeCheckDiagnostic>>,
+}
+
+impl TypeCheckDiagnostics {
+    pub fn new() -> Self {
+        Self { inner: Vec::new() }
+    }
+
+    pub(super) fn push(&mut self, diagnostic: TypeCheckDiagnostic) {
+        self.inner.push(Arc::new(diagnostic));
+    }
+
+    pub(crate) fn shrink_to_fit(&mut self) {
+        self.inner.shrink_to_fit();
+    }
+}
+
+impl Extend<TypeCheckDiagnostic> for TypeCheckDiagnostics {
+    fn extend<T: IntoIterator<Item = TypeCheckDiagnostic>>(&mut self, iter: T) {
+        self.inner.extend(iter.into_iter().map(std::sync::Arc::new));
+    }
+}
+
+impl Extend<std::sync::Arc<TypeCheckDiagnostic>> for TypeCheckDiagnostics {
+    fn extend<T: IntoIterator<Item = Arc<TypeCheckDiagnostic>>>(&mut self, iter: T) {
+        self.inner.extend(iter);
+    }
+}
+
+impl<'a> Extend<&'a std::sync::Arc<TypeCheckDiagnostic>> for TypeCheckDiagnostics {
+    fn extend<T: IntoIterator<Item = &'a Arc<TypeCheckDiagnostic>>>(&mut self, iter: T) {
+        self.inner
+            .extend(iter.into_iter().map(std::sync::Arc::clone));
+    }
+}
+
+impl std::fmt::Debug for TypeCheckDiagnostics {
+    fn fmt(&self, f: &mut Formatter<'_>) -> std::fmt::Result {
+        self.inner.fmt(f)
+    }
+}
+
+impl Deref for TypeCheckDiagnostics {
+    type Target = [std::sync::Arc<TypeCheckDiagnostic>];
+
+    fn deref(&self) -> &Self::Target {
+        &self.inner
+    }
+}
+
+impl IntoIterator for TypeCheckDiagnostics {
+    type Item = Arc<TypeCheckDiagnostic>;
+    type IntoIter = std::vec::IntoIter<std::sync::Arc<TypeCheckDiagnostic>>;
+
+    fn into_iter(self) -> Self::IntoIter {
+        self.inner.into_iter()
+    }
+}
+
+impl<'a> IntoIterator for &'a TypeCheckDiagnostics {
+    type Item = &'a Arc<TypeCheckDiagnostic>;
+    type IntoIter = std::slice::Iter<'a, std::sync::Arc<TypeCheckDiagnostic>>;
+
+    fn into_iter(self) -> Self::IntoIter {
+        self.inner.iter()
+    }
+}

crates/red_knot_python_semantic/src/types/display.rs

@@ -0,0 +1,373 @@
+//! Display implementations for types.
+
+use std::fmt::{self, Display, Formatter};
+
+use ruff_db::display::FormatterJoinExtension;
+use ruff_python_ast::str::Quote;
+use ruff_python_literal::escape::AsciiEscape;
+
+use crate::types::{IntersectionType, Type, UnionType};
+use crate::Db;
+use rustc_hash::FxHashMap;
+
+impl<'db> Type<'db> {
+    pub fn display(&self, db: &'db dyn Db) -> DisplayType {
+        DisplayType { ty: self, db }
+    }
+    fn representation(self, db: &'db dyn Db) -> DisplayRepresentation<'db> {
+        DisplayRepresentation { db, ty: self }
+    }
+}
+
+#[derive(Copy, Clone)]
+pub struct DisplayType<'db> {
+    ty: &'db Type<'db>,
+    db: &'db dyn Db,
+}
+
+impl Display for DisplayType<'_> {
+    fn fmt(&self, f: &mut Formatter<'_>) -> fmt::Result {
+        let representation = self.ty.representation(self.db);
+        if matches!(
+            self.ty,
+            Type::IntLiteral(_)
+                | Type::BooleanLiteral(_)
+                | Type::StringLiteral(_)
+                | Type::BytesLiteral(_)
+                | Type::Class(_)
+                | Type::Function(_)
+        ) {
+            write!(f, "Literal[{representation}]")
+        } else {
+            representation.fmt(f)
+        }
+    }
+}
+
+impl fmt::Debug for DisplayType<'_> {
+    fn fmt(&self, f: &mut Formatter<'_>) -> fmt::Result {
+        Display::fmt(self, f)
+    }
+}
+
+/// Writes the string representation of a type, which is the value displayed either as
+/// `Literal[<repr>]` or `Literal[<repr1>, <repr2>]` for literal types or as `<repr>` for
+/// non literals
+struct DisplayRepresentation<'db> {
+    ty: Type<'db>,
+    db: &'db dyn Db,
+}
+
+impl Display for DisplayRepresentation<'_> {
+    fn fmt(&self, f: &mut Formatter<'_>) -> fmt::Result {
+        match self.ty {
+            Type::Any => f.write_str("Any"),
+            Type::Never => f.write_str("Never"),
+            Type::Unknown => f.write_str("Unknown"),
+            Type::Unbound => f.write_str("Unbound"),
+            Type::None => f.write_str("None"),
+            // `[Type::Todo]`'s display should be explicit that is not a valid display of
+            // any other type
+            Type::Todo => f.write_str("@Todo"),
+            Type::Module(file) => {
+                write!(f, "<module '{:?}'>", file.path(self.db))
+            }
+            // TODO functions and classes should display using a fully qualified name
+            Type::Class(class) => f.write_str(class.name(self.db)),
+            Type::Instance(class) => f.write_str(class.name(self.db)),
+            Type::Function(function) => f.write_str(function.name(self.db)),
+            Type::Union(union) => union.display(self.db).fmt(f),
+            Type::Intersection(intersection) => intersection.display(self.db).fmt(f),
+            Type::IntLiteral(n) => n.fmt(f),
+            Type::BooleanLiteral(boolean) => f.write_str(if boolean { "True" } else { "False" }),
+            Type::StringLiteral(string) => {
+                write!(f, r#""{}""#, string.value(self.db).replace('"', r#"\""#))
+            }
+            Type::LiteralString => f.write_str("LiteralString"),
+            Type::BytesLiteral(bytes) => {
+                let escape =
+                    AsciiEscape::with_preferred_quote(bytes.value(self.db).as_ref(), Quote::Double);
+
+                escape.bytes_repr().write(f)
+            }
+            Type::Tuple(tuple) => {
+                f.write_str("tuple[")?;
+                let elements = tuple.elements(self.db);
+                if elements.is_empty() {
+                    f.write_str("()")?;
+                } else {
+                    elements.display(self.db).fmt(f)?;
+                }
+                f.write_str("]")
+            }
+        }
+    }
+}
+
+impl<'db> UnionType<'db> {
+    fn display(&'db self, db: &'db dyn Db) -> DisplayUnionType<'db> {
+        DisplayUnionType { db, ty: self }
+    }
+}
+
+struct DisplayUnionType<'db> {
+    ty: &'db UnionType<'db>,
+    db: &'db dyn Db,
+}
+
+impl Display for DisplayUnionType<'_> {
+    fn fmt(&self, f: &mut Formatter<'_>) -> fmt::Result {
+        let elements = self.ty.elements(self.db);
+
+        // Group literal types by kind.
+        let mut grouped_literals = FxHashMap::default();
+
+        for element in elements {
+            if let Ok(literal_kind) = LiteralTypeKind::try_from(*element) {
+                grouped_literals
+                    .entry(literal_kind)
+                    .or_insert_with(Vec::new)
+                    .push(*element);
+            }
+        }
+
+        let mut join = f.join(" | ");
+
+        for element in elements {
+            if let Ok(literal_kind) = LiteralTypeKind::try_from(*element) {
+                let Some(mut literals) = grouped_literals.remove(&literal_kind) else {
+                    continue;
+                };
+                if literal_kind == LiteralTypeKind::IntLiteral {
+                    literals.sort_unstable_by_key(|ty| ty.expect_int_literal());
+                }
+                join.entry(&DisplayLiteralGroup {
+                    literals,
+                    db: self.db,
+                });
+            } else {
+                join.entry(&element.display(self.db));
+            }
+        }
+
+        join.finish()?;
+
+        debug_assert!(grouped_literals.is_empty());
+
+        Ok(())
+    }
+}
+
+impl fmt::Debug for DisplayUnionType<'_> {
+    fn fmt(&self, f: &mut Formatter<'_>) -> fmt::Result {
+        Display::fmt(self, f)
+    }
+}
+
+struct DisplayLiteralGroup<'db> {
+    literals: Vec<Type<'db>>,
+    db: &'db dyn Db,
+}
+
+impl Display for DisplayLiteralGroup<'_> {
+    fn fmt(&self, f: &mut Formatter<'_>) -> fmt::Result {
+        f.write_str("Literal[")?;
+        f.join(", ")
+            .entries(self.literals.iter().map(|ty| ty.representation(self.db)))
+            .finish()?;
+        f.write_str("]")
+    }
+}
+
+#[derive(Copy, Clone, Debug, Eq, PartialEq, Hash)]
+enum LiteralTypeKind {
+    Class,
+    Function,
+    IntLiteral,
+    StringLiteral,
+    BytesLiteral,
+}
+
+impl TryFrom<Type<'_>> for LiteralTypeKind {
+    type Error = ();
+
+    fn try_from(value: Type<'_>) -> Result<Self, Self::Error> {
+        match value {
+            Type::Class(_) => Ok(Self::Class),
+            Type::Function(_) => Ok(Self::Function),
+            Type::IntLiteral(_) => Ok(Self::IntLiteral),
+            Type::StringLiteral(_) => Ok(Self::StringLiteral),
+            Type::BytesLiteral(_) => Ok(Self::BytesLiteral),
+            _ => Err(()),
+        }
+    }
+}
+
+impl<'db> IntersectionType<'db> {
+    fn display(&'db self, db: &'db dyn Db) -> DisplayIntersectionType<'db> {
+        DisplayIntersectionType { db, ty: self }
+    }
+}
+
+struct DisplayIntersectionType<'db> {
+    ty: &'db IntersectionType<'db>,
+    db: &'db dyn Db,
+}
+
+impl Display for DisplayIntersectionType<'_> {
+    fn fmt(&self, f: &mut Formatter<'_>) -> fmt::Result {
+        let tys = self
+            .ty
+            .positive(self.db)
+            .iter()
+            .map(|&ty| DisplayMaybeNegatedType {
+                ty,
+                db: self.db,
+                negated: false,
+            })
+            .chain(
+                self.ty
+                    .negative(self.db)
+                    .iter()
+                    .map(|&ty| DisplayMaybeNegatedType {
+                        ty,
+                        db: self.db,
+                        negated: true,
+                    }),
+            );
+        f.join(" & ").entries(tys).finish()
+    }
+}
+
+impl fmt::Debug for DisplayIntersectionType<'_> {
+    fn fmt(&self, f: &mut Formatter<'_>) -> fmt::Result {
+        Display::fmt(self, f)
+    }
+}
+
+struct DisplayMaybeNegatedType<'db> {
+    ty: Type<'db>,
+    db: &'db dyn Db,
+    negated: bool,
+}
+
+impl<'db> Display for DisplayMaybeNegatedType<'db> {
+    fn fmt(&self, f: &mut Formatter<'_>) -> fmt::Result {
+        if self.negated {
+            f.write_str("~")?;
+        }
+        self.ty.display(self.db).fmt(f)
+    }
+}
+
+pub(crate) trait TypeArrayDisplay<'db> {
+    fn display(&self, db: &'db dyn Db) -> DisplayTypeArray;
+}
+
+impl<'db> TypeArrayDisplay<'db> for Box<[Type<'db>]> {
+    fn display(&self, db: &'db dyn Db) -> DisplayTypeArray {
+        DisplayTypeArray { types: self, db }
+    }
+}
+
+impl<'db> TypeArrayDisplay<'db> for Vec<Type<'db>> {
+    fn display(&self, db: &'db dyn Db) -> DisplayTypeArray {
+        DisplayTypeArray { types: self, db }
+    }
+}
+
+pub(crate) struct DisplayTypeArray<'b, 'db> {
+    types: &'b [Type<'db>],
+    db: &'db dyn Db,
+}
+
+impl<'db> Display for DisplayTypeArray<'_, 'db> {
+    fn fmt(&self, f: &mut Formatter<'_>) -> fmt::Result {
+        f.join(", ")
+            .entries(self.types.iter().map(|ty| ty.display(self.db)))
+            .finish()
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use ruff_db::files::system_path_to_file;
+    use ruff_db::system::{DbWithTestSystem, SystemPathBuf};
+
+    use crate::db::tests::TestDb;
+    use crate::types::{global_symbol_ty, BytesLiteralType, StringLiteralType, Type, UnionType};
+    use crate::{Program, ProgramSettings, PythonVersion, SearchPathSettings};
+
+    fn setup_db() -> TestDb {
+        let db = TestDb::new();
+
+        let src_root = SystemPathBuf::from("/src");
+        db.memory_file_system()
+            .create_directory_all(&src_root)
+            .unwrap();
+
+        Program::from_settings(
+            &db,
+            &ProgramSettings {
+                target_version: PythonVersion::default(),
+                search_paths: SearchPathSettings::new(src_root),
+            },
+        )
+        .expect("Valid search path settings");
+
+        db
+    }
+
+    #[test]
+    fn test_condense_literal_display_by_type() -> anyhow::Result<()> {
+        let mut db = setup_db();
+
+        db.write_dedented(
+            "src/main.py",
+            "
+            def foo(x: int) -> int:
+                return x + 1
+
+            def bar(s: str) -> str:
+                return s
+
+            class A: ...
+            class B: ...
+            ",
+        )?;
+        let mod_file = system_path_to_file(&db, "src/main.py").expect("file to exist");
+
+        let union_elements = &[
+            Type::Unknown,
+            Type::IntLiteral(-1),
+            global_symbol_ty(&db, mod_file, "A"),
+            Type::StringLiteral(StringLiteralType::new(&db, Box::from("A"))),
+            Type::BytesLiteral(BytesLiteralType::new(&db, Box::from([0]))),
+            Type::BytesLiteral(BytesLiteralType::new(&db, Box::from([7]))),
+            Type::IntLiteral(0),
+            Type::IntLiteral(1),
+            Type::StringLiteral(StringLiteralType::new(&db, Box::from("B"))),
+            global_symbol_ty(&db, mod_file, "foo"),
+            global_symbol_ty(&db, mod_file, "bar"),
+            global_symbol_ty(&db, mod_file, "B"),
+            Type::BooleanLiteral(true),
+            Type::None,
+        ];
+        let union = UnionType::from_elements(&db, union_elements).expect_union();
+        let display = format!("{}", union.display(&db));
+        assert_eq!(
+            display,
+            concat!(
+                "Unknown | ",
+                "Literal[-1, 0, 1] | ",
+                "Literal[A, B] | ",
+                "Literal[\"A\", \"B\"] | ",
+                "Literal[b\"\\x00\", b\"\\x07\"] | ",
+                "Literal[foo, bar] | ",
+                "Literal[True] | ",
+                "None"
+            )
+        );
+        Ok(())
+    }
+}

crates/red_knot_python_semantic/src/types/narrow.rs

@@ -0,0 +1,197 @@
+use crate::semantic_index::ast_ids::HasScopedAstId;
+use crate::semantic_index::constraint::{Constraint, PatternConstraint};
+use crate::semantic_index::definition::Definition;
+use crate::semantic_index::expression::Expression;
+use crate::semantic_index::symbol::{ScopeId, ScopedSymbolId, SymbolTable};
+use crate::semantic_index::symbol_table;
+use crate::types::{infer_expression_types, IntersectionBuilder, Type};
+use crate::Db;
+use ruff_python_ast as ast;
+use rustc_hash::FxHashMap;
+use std::sync::Arc;
+
+/// Return the type constraint that `test` (if true) would place on `definition`, if any.
+///
+/// For example, if we have this code:
+///
+/// ```python
+/// y = 1 if flag else None
+/// x = 1 if flag else None
+/// if x is not None:
+///     ...
+/// ```
+///
+/// The `test` expression `x is not None` places the constraint "not None" on the definition of
+/// `x`, so in that case we'd return `Some(Type::Intersection(negative=[Type::None]))`.
+///
+/// But if we called this with the same `test` expression, but the `definition` of `y`, no
+/// constraint is applied to that definition, so we'd just return `None`.
+pub(crate) fn narrowing_constraint<'db>(
+    db: &'db dyn Db,
+    constraint: Constraint<'db>,
+    definition: Definition<'db>,
+) -> Option<Type<'db>> {
+    match constraint {
+        Constraint::Expression(expression) => {
+            all_narrowing_constraints_for_expression(db, expression)
+                .get(&definition.symbol(db))
+                .copied()
+        }
+        Constraint::Pattern(pattern) => all_narrowing_constraints_for_pattern(db, pattern)
+            .get(&definition.symbol(db))
+            .copied(),
+    }
+}
+
+#[salsa::tracked(return_ref)]
+fn all_narrowing_constraints_for_pattern<'db>(
+    db: &'db dyn Db,
+    pattern: PatternConstraint<'db>,
+) -> NarrowingConstraints<'db> {
+    NarrowingConstraintsBuilder::new(db, Constraint::Pattern(pattern)).finish()
+}
+
+#[salsa::tracked(return_ref)]
+fn all_narrowing_constraints_for_expression<'db>(
+    db: &'db dyn Db,
+    expression: Expression<'db>,
+) -> NarrowingConstraints<'db> {
+    NarrowingConstraintsBuilder::new(db, Constraint::Expression(expression)).finish()
+}
+
+type NarrowingConstraints<'db> = FxHashMap<ScopedSymbolId, Type<'db>>;
+
+struct NarrowingConstraintsBuilder<'db> {
+    db: &'db dyn Db,
+    constraint: Constraint<'db>,
+    constraints: NarrowingConstraints<'db>,
+}
+
+impl<'db> NarrowingConstraintsBuilder<'db> {
+    fn new(db: &'db dyn Db, constraint: Constraint<'db>) -> Self {
+        Self {
+            db,
+            constraint,
+            constraints: NarrowingConstraints::default(),
+        }
+    }
+
+    fn finish(mut self) -> NarrowingConstraints<'db> {
+        match self.constraint {
+            Constraint::Expression(expression) => self.evaluate_expression_constraint(expression),
+            Constraint::Pattern(pattern) => self.evaluate_pattern_constraint(pattern),
+        }
+
+        self.constraints.shrink_to_fit();
+        self.constraints
+    }
+
+    fn evaluate_expression_constraint(&mut self, expression: Expression<'db>) {
+        if let ast::Expr::Compare(expr_compare) = expression.node_ref(self.db).node() {
+            self.add_expr_compare(expr_compare, expression);
+        }
+        // TODO other test expression kinds
+    }
+
+    fn evaluate_pattern_constraint(&mut self, pattern: PatternConstraint<'db>) {
+        let subject = pattern.subject(self.db);
+
+        match pattern.pattern(self.db).node() {
+            ast::Pattern::MatchValue(_) => {
+                // TODO
+            }
+            ast::Pattern::MatchSingleton(singleton_pattern) => {
+                self.add_match_pattern_singleton(subject, singleton_pattern);
+            }
+            ast::Pattern::MatchSequence(_) => {
+                // TODO
+            }
+            ast::Pattern::MatchMapping(_) => {
+                // TODO
+            }
+            ast::Pattern::MatchClass(_) => {
+                // TODO
+            }
+            ast::Pattern::MatchStar(_) => {
+                // TODO
+            }
+            ast::Pattern::MatchAs(_) => {
+                // TODO
+            }
+            ast::Pattern::MatchOr(_) => {
+                // TODO
+            }
+        }
+    }
+
+    fn symbols(&self) -> Arc<SymbolTable> {
+        symbol_table(self.db, self.scope())
+    }
+
+    fn scope(&self) -> ScopeId<'db> {
+        match self.constraint {
+            Constraint::Expression(expression) => expression.scope(self.db),
+            Constraint::Pattern(pattern) => pattern.scope(self.db),
+        }
+    }
+
+    fn add_expr_compare(&mut self, expr_compare: &ast::ExprCompare, expression: Expression<'db>) {
+        let ast::ExprCompare {
+            range: _,
+            left,
+            ops,
+            comparators,
+        } = expr_compare;
+
+        if let ast::Expr::Name(ast::ExprName {
+            range: _,
+            id,
+            ctx: _,
+        }) = left.as_ref()
+        {
+            // SAFETY: we should always have a symbol for every Name node.
+            let symbol = self.symbols().symbol_id_by_name(id).unwrap();
+            let scope = self.scope();
+            let inference = infer_expression_types(self.db, expression);
+            for (op, comparator) in std::iter::zip(&**ops, &**comparators) {
+                let comp_ty = inference.expression_ty(comparator.scoped_ast_id(self.db, scope));
+                match op {
+                    ast::CmpOp::IsNot => {
+                        if comp_ty.is_singleton(self.db) {
+                            let ty = IntersectionBuilder::new(self.db)
+                                .add_negative(comp_ty)
+                                .build();
+                            self.constraints.insert(symbol, ty);
+                        } else {
+                            // Non-singletons cannot be safely narrowed using `is not`
+                        }
+                    }
+                    ast::CmpOp::Is => {
+                        self.constraints.insert(symbol, comp_ty);
+                    }
+                    _ => {
+                        // TODO other comparison types
+                    }
+                }
+            }
+        }
+    }
+
+    fn add_match_pattern_singleton(
+        &mut self,
+        subject: &ast::Expr,
+        pattern: &ast::PatternMatchSingleton,
+    ) {
+        if let Some(ast::ExprName { id, .. }) = subject.as_name_expr() {
+            // SAFETY: we should always have a symbol for every Name node.
+            let symbol = self.symbols().symbol_id_by_name(id).unwrap();
+
+            let ty = match pattern.value {
+                ast::Singleton::None => Type::None,
+                ast::Singleton::True => Type::BooleanLiteral(true),
+                ast::Singleton::False => Type::BooleanLiteral(false),
+            };
+            self.constraints.insert(symbol, ty);
+        }
+    }
+}

crates/red_knot_python_semantic/tests/mdtest.rs

@@ -0,0 +1,14 @@
+use red_knot_test::run;
+use std::path::PathBuf;
+
+/// See `crates/red_knot_test/README.md` for documentation on these tests.
+#[rstest::rstest]
+fn mdtest(#[files("resources/mdtest/**/*.md")] path: PathBuf) {
+    let crate_dir = PathBuf::from(env!("CARGO_MANIFEST_DIR"))
+        .join("resources")
+        .join("mdtest")
+        .canonicalize()
+        .unwrap();
+    let title = path.strip_prefix(crate_dir).unwrap();
+    run(&path, title.as_os_str().to_str().unwrap());
+}

crates/red_knot_server/Cargo.toml

@@ -1,5 +1,5 @@
 [package]
-name = "ty_server"
+name = "red_knot_server"
 version = "0.0.0"
 publish = false
 authors = { workspace = true }
@@ -11,15 +11,13 @@ repository = { workspace = true }
 license = { workspace = true }
 
 [dependencies]
-ruff_db = { workspace = true, features = ["os"] }
+red_knot_workspace = { workspace = true }
+ruff_db = { workspace = true }
 ruff_notebook = { workspace = true }
+ruff_python_ast = { workspace = true }
 ruff_source_file = { workspace = true }
 ruff_text_size = { workspace = true }
 
-ty_ide = { workspace = true }
-ty_project = { workspace = true }
-ty_python_semantic = { workspace = true }
-
 anyhow = { workspace = true }
 crossbeam = { workspace = true }
 jod-thread = { workspace = true }

crates/red_knot_server/src/edit.rs

@@ -1,14 +1,12 @@
-//! Types and utilities for working with text, modifying source files, and `ty <-> LSP` type conversion.
+//! Types and utilities for working with text, modifying source files, and `Ruff <-> LSP` type conversion.
 
-mod location;
 mod notebook;
 mod range;
 mod text_document;
 
-pub(crate) use location::ToLink;
 use lsp_types::{PositionEncodingKind, Url};
 pub use notebook::NotebookDocument;
-pub(crate) use range::{FileRangeExt, PositionExt, RangeExt, TextSizeExt, ToRangeExt};
+pub(crate) use range::RangeExt;
 pub(crate) use text_document::DocumentVersion;
 pub use text_document::TextDocument;
 
@@ -23,20 +21,10 @@ pub enum PositionEncoding {
     /// Second choice because UTF32 uses a fixed 4 byte encoding for each character (makes conversion relatively easy)
     UTF32,
 
-    /// ty's preferred encoding
+    /// Ruff's preferred encoding
     UTF8,
 }
 
-impl From<PositionEncoding> for ruff_source_file::PositionEncoding {
-    fn from(value: PositionEncoding) -> Self {
-        match value {
-            PositionEncoding::UTF8 => Self::Utf8,
-            PositionEncoding::UTF16 => Self::Utf16,
-            PositionEncoding::UTF32 => Self::Utf32,
-        }
-    }
-}
-
 /// A unique document ID, derived from a URL passed as part of an LSP request.
 /// This document ID can point to either be a standalone Python file, a full notebook, or a cell within a notebook.
 #[derive(Clone, Debug)]
@@ -65,17 +53,17 @@ impl std::fmt::Display for DocumentKey {
     }
 }
 
-impl From<PositionEncoding> for PositionEncodingKind {
+impl From<PositionEncoding> for lsp_types::PositionEncodingKind {
     fn from(value: PositionEncoding) -> Self {
         match value {
-            PositionEncoding::UTF8 => PositionEncodingKind::UTF8,
-            PositionEncoding::UTF16 => PositionEncodingKind::UTF16,
-            PositionEncoding::UTF32 => PositionEncodingKind::UTF32,
+            PositionEncoding::UTF8 => lsp_types::PositionEncodingKind::UTF8,
+            PositionEncoding::UTF16 => lsp_types::PositionEncodingKind::UTF16,
+            PositionEncoding::UTF32 => lsp_types::PositionEncodingKind::UTF32,
         }
     }
 }
 
-impl TryFrom<&PositionEncodingKind> for PositionEncoding {
+impl TryFrom<&lsp_types::PositionEncodingKind> for PositionEncoding {
     type Error = ();
 
     fn try_from(value: &PositionEncodingKind) -> Result<Self, Self::Error> {

crates/red_knot_server/src/edit/notebook.rs

@@ -92,7 +92,7 @@ impl NotebookDocument {
         };
 
         ruff_notebook::Notebook::from_raw_notebook(raw_notebook, false)
-            .unwrap_or_else(|err| panic!("Server notebook document could not be converted to ty's notebook document format: {err}"))
+            .unwrap_or_else(|err| panic!("Server notebook document could not be converted to Ruff's notebook document format: {err}"))
     }
 
     pub(crate) fn update(
@@ -118,7 +118,7 @@ impl NotebookDocument {
                 // This is required because of the way the `NotebookCell` is modelled. We include
                 // the `TextDocument` within the `NotebookCell` so when it's deleted, the
                 // corresponding `TextDocument` is removed as well. But, when cells are
-                // re-ordered, the change request doesn't provide the actual contents of the cell.
+                // re-oredered, the change request doesn't provide the actual contents of the cell.
                 // Instead, it only provides that (a) these cell URIs were removed, and (b) these
                 // cell URIs were added.
                 // https://github.com/astral-sh/ruff/issues/12573
@@ -136,15 +136,17 @@ impl NotebookDocument {
                 // provide the actual contents of the cells, so we'll initialize them with empty
                 // contents.
                 for cell in structure.array.cells.into_iter().flatten().rev() {
-                    let (content, version) =
-                        if let Some(text_document) = deleted_cells.remove(&cell.document) {
-                            let version = text_document.version();
-                            (text_document.into_contents(), version)
-                        } else {
-                            (String::new(), 0)
-                        };
-                    self.cells
-                        .insert(start, NotebookCell::new(cell, content, version));
+                    if let Some(text_document) = deleted_cells.remove(&cell.document) {
+                        let version = text_document.version();
+                        self.cells.push(NotebookCell::new(
+                            cell,
+                            text_document.into_contents(),
+                            version,
+                        ));
+                    } else {
+                        self.cells
+                            .insert(start, NotebookCell::new(cell, String::new(), 0));
+                    }
                 }
 
                 // Third, register the new cells in the index and update existing ones that came
@@ -198,12 +200,6 @@ impl NotebookDocument {
         self.version
     }
 
-    /// Get the URI for a cell by its index within the cell array.
-    #[expect(dead_code)]
-    pub(crate) fn cell_uri_by_index(&self, index: CellId) -> Option<&lsp_types::Url> {
-        self.cells.get(index).map(|cell| &cell.url)
-    }
-
     /// Get the text document representing the contents of a cell by the cell URI.
     pub(crate) fn cell_document_by_uri(&self, uri: &lsp_types::Url) -> Option<&TextDocument> {
         self.cells
@@ -242,115 +238,3 @@ impl NotebookCell {
         }
     }
 }
-
-#[cfg(test)]
-mod tests {
-    use super::NotebookDocument;
-
-    enum TestCellContent {
-        #[expect(dead_code)]
-        Markup(String),
-        Code(String),
-    }
-
-    fn create_test_url(index: usize) -> lsp_types::Url {
-        lsp_types::Url::parse(&format!("cell:/test.ipynb#{index}")).unwrap()
-    }
-
-    fn create_test_notebook(test_cells: Vec<TestCellContent>) -> NotebookDocument {
-        let mut cells = Vec::with_capacity(test_cells.len());
-        let mut cell_documents = Vec::with_capacity(test_cells.len());
-
-        for (index, test_cell) in test_cells.into_iter().enumerate() {
-            let url = create_test_url(index);
-            match test_cell {
-                TestCellContent::Markup(content) => {
-                    cells.push(lsp_types::NotebookCell {
-                        kind: lsp_types::NotebookCellKind::Markup,
-                        document: url.clone(),
-                        metadata: None,
-                        execution_summary: None,
-                    });
-                    cell_documents.push(lsp_types::TextDocumentItem {
-                        uri: url,
-                        language_id: "markdown".to_owned(),
-                        version: 0,
-                        text: content,
-                    });
-                }
-                TestCellContent::Code(content) => {
-                    cells.push(lsp_types::NotebookCell {
-                        kind: lsp_types::NotebookCellKind::Code,
-                        document: url.clone(),
-                        metadata: None,
-                        execution_summary: None,
-                    });
-                    cell_documents.push(lsp_types::TextDocumentItem {
-                        uri: url,
-                        language_id: "python".to_owned(),
-                        version: 0,
-                        text: content,
-                    });
-                }
-            }
-        }
-
-        NotebookDocument::new(0, cells, serde_json::Map::default(), cell_documents).unwrap()
-    }
-
-    /// This test case checks that for a notebook with three code cells, when the client sends a
-    /// change request to swap the first two cells, the notebook document is updated correctly.
-    ///
-    /// The swap operation as a change request is represented as deleting the first two cells and
-    /// adding them back in the reverse order.
-    #[test]
-    fn swap_cells() {
-        let mut notebook = create_test_notebook(vec![
-            TestCellContent::Code("cell = 0".to_owned()),
-            TestCellContent::Code("cell = 1".to_owned()),
-            TestCellContent::Code("cell = 2".to_owned()),
-        ]);
-
-        notebook
-            .update(
-                Some(lsp_types::NotebookDocumentCellChange {
-                    structure: Some(lsp_types::NotebookDocumentCellChangeStructure {
-                        array: lsp_types::NotebookCellArrayChange {
-                            start: 0,
-                            delete_count: 2,
-                            cells: Some(vec![
-                                lsp_types::NotebookCell {
-                                    kind: lsp_types::NotebookCellKind::Code,
-                                    document: create_test_url(1),
-                                    metadata: None,
-                                    execution_summary: None,
-                                },
-                                lsp_types::NotebookCell {
-                                    kind: lsp_types::NotebookCellKind::Code,
-                                    document: create_test_url(0),
-                                    metadata: None,
-                                    execution_summary: None,
-                                },
-                            ]),
-                        },
-                        did_open: None,
-                        did_close: None,
-                    }),
-                    data: None,
-                    text_content: None,
-                }),
-                None,
-                1,
-                crate::PositionEncoding::default(),
-            )
-            .unwrap();
-
-        assert_eq!(
-            notebook.make_ruff_notebook().source_code(),
-            "cell = 1
-cell = 0
-cell = 2
-"
-        );
-    }
-}

crates/red_knot_server/src/edit/range.rs

@@ -0,0 +1,98 @@
+use super::PositionEncoding;
+use ruff_source_file::LineIndex;
+use ruff_source_file::OneIndexed;
+use ruff_text_size::{TextRange, TextSize};
+
+pub(crate) trait RangeExt {
+    fn to_text_range(&self, text: &str, index: &LineIndex, encoding: PositionEncoding)
+        -> TextRange;
+}
+
+fn u32_index_to_usize(index: u32) -> usize {
+    usize::try_from(index).expect("u32 fits in usize")
+}
+
+impl RangeExt for lsp_types::Range {
+    fn to_text_range(
+        &self,
+        text: &str,
+        index: &LineIndex,
+        encoding: PositionEncoding,
+    ) -> TextRange {
+        let start_line = index.line_range(
+            OneIndexed::from_zero_indexed(u32_index_to_usize(self.start.line)),
+            text,
+        );
+        let end_line = index.line_range(
+            OneIndexed::from_zero_indexed(u32_index_to_usize(self.end.line)),
+            text,
+        );
+
+        let (start_column_offset, end_column_offset) = match encoding {
+            PositionEncoding::UTF8 => (
+                TextSize::new(self.start.character),
+                TextSize::new(self.end.character),
+            ),
+
+            PositionEncoding::UTF16 => {
+                // Fast path for ASCII only documents
+                if index.is_ascii() {
+                    (
+                        TextSize::new(self.start.character),
+                        TextSize::new(self.end.character),
+                    )
+                } else {
+                    // UTF16 encodes characters either as one or two 16 bit words.
+                    // The position in `range` is the 16-bit word offset from the start of the line (and not the character offset)
+                    // UTF-16 with a text that may use variable-length characters.
+                    (
+                        utf8_column_offset(self.start.character, &text[start_line]),
+                        utf8_column_offset(self.end.character, &text[end_line]),
+                    )
+                }
+            }
+            PositionEncoding::UTF32 => {
+                // UTF-32 uses 4 bytes for each character. Meaning, the position in range is a character offset.
+                return TextRange::new(
+                    index.offset(
+                        OneIndexed::from_zero_indexed(u32_index_to_usize(self.start.line)),
+                        OneIndexed::from_zero_indexed(u32_index_to_usize(self.start.character)),
+                        text,
+                    ),
+                    index.offset(
+                        OneIndexed::from_zero_indexed(u32_index_to_usize(self.end.line)),
+                        OneIndexed::from_zero_indexed(u32_index_to_usize(self.end.character)),
+                        text,
+                    ),
+                );
+            }
+        };
+
+        TextRange::new(
+            start_line.start() + start_column_offset.clamp(TextSize::new(0), start_line.end()),
+            end_line.start() + end_column_offset.clamp(TextSize::new(0), end_line.end()),
+        )
+    }
+}
+
+/// Converts a UTF-16 code unit offset for a given line into a UTF-8 column number.
+fn utf8_column_offset(utf16_code_unit_offset: u32, line: &str) -> TextSize {
+    let mut utf8_code_unit_offset = TextSize::new(0);
+
+    let mut i = 0u32;
+
+    for c in line.chars() {
+        if i >= utf16_code_unit_offset {
+            break;
+        }
+
+        // Count characters encoded as two 16 bit words as 2 characters.
+        {
+            utf8_code_unit_offset +=
+                TextSize::new(u32::try_from(c.len_utf8()).expect("utf8 len always <=4"));
+            i += u32::try_from(c.len_utf16()).expect("utf16 len always <=2");
+        }
+    }
+
+    utf8_code_unit_offset
+}

crates/red_knot_server/src/edit/text_document.rs

@@ -20,23 +20,6 @@ pub struct TextDocument {
     /// The latest version of the document, set by the LSP client. The server will panic in
     /// debug mode if we attempt to update the document with an 'older' version.
     version: DocumentVersion,
-    /// The language ID of the document as provided by the client.
-    language_id: Option<LanguageId>,
-}
-
-#[derive(Debug, Copy, Clone, PartialEq, Eq)]
-pub enum LanguageId {
-    Python,
-    Other,
-}
-
-impl From<&str> for LanguageId {
-    fn from(language_id: &str) -> Self {
-        match language_id {
-            "python" => Self::Python,
-            _ => Self::Other,
-        }
-    }
 }
 
 impl TextDocument {
@@ -46,16 +29,9 @@ impl TextDocument {
             contents,
             index,
             version,
-            language_id: None,
         }
     }
 
-    #[must_use]
-    pub fn with_language_id(mut self, language_id: &str) -> Self {
-        self.language_id = Some(LanguageId::from(language_id));
-        self
-    }
-
     pub fn into_contents(self) -> String {
         self.contents
     }
@@ -72,21 +48,15 @@ impl TextDocument {
         self.version
     }
 
-    pub fn language_id(&self) -> Option<LanguageId> {
-        self.language_id
-    }
-
     pub fn apply_changes(
         &mut self,
         changes: Vec<lsp_types::TextDocumentContentChangeEvent>,
         new_version: DocumentVersion,
         encoding: PositionEncoding,
     ) {
-        if let [
-            lsp_types::TextDocumentContentChangeEvent {
-                range: None, text, ..
-            },
-        ] = changes.as_slice()
+        if let [lsp_types::TextDocumentContentChangeEvent {
+            range: None, text, ..
+        }] = changes.as_slice()
         {
             tracing::debug!("Fast path - replacing entire document");
             self.modify(|contents, version| {
@@ -96,6 +66,7 @@ impl TextDocument {
             return;
         }
 
+        let old_contents = self.contents().to_string();
         let mut new_contents = self.contents().to_string();
         let mut active_index = self.index().clone();
 
@@ -116,11 +87,15 @@ impl TextDocument {
                 new_contents = change;
             }
 
-            active_index = LineIndex::from_source_text(&new_contents);
+            if new_contents != old_contents {
+                active_index = LineIndex::from_source_text(&new_contents);
+            }
         }
 
         self.modify_with_manual_index(|contents, version, index| {
-            *index = active_index;
+            if contents != &new_contents {
+                *index = active_index;
+            }
             *contents = new_contents;
             *version = new_version;
         });
@@ -150,75 +125,3 @@ impl TextDocument {
         debug_assert!(self.version >= old_version);
     }
 }
-
-#[cfg(test)]
-mod tests {
-    use crate::{PositionEncoding, TextDocument};
-    use lsp_types::{Position, TextDocumentContentChangeEvent};
-
-    #[test]
-    fn redo_edit() {
-        let mut document = TextDocument::new(
-            r#""""
-测试comment
-一些测试内容
-"""
-import click
-
-
-@click.group()
-def interface():
-    pas
-"#
-            .to_string(),
-            0,
-        );
-
-        // Add an `s`, remove it again (back to the original code), and then re-add the `s`
-        document.apply_changes(
-            vec![
-                TextDocumentContentChangeEvent {
-                    range: Some(lsp_types::Range::new(
-                        Position::new(9, 7),
-                        Position::new(9, 7),
-                    )),
-                    range_length: Some(0),
-                    text: "s".to_string(),
-                },
-                TextDocumentContentChangeEvent {
-                    range: Some(lsp_types::Range::new(
-                        Position::new(9, 7),
-                        Position::new(9, 8),
-                    )),
-                    range_length: Some(1),
-                    text: String::new(),
-                },
-                TextDocumentContentChangeEvent {
-                    range: Some(lsp_types::Range::new(
-                        Position::new(9, 7),
-                        Position::new(9, 7),
-                    )),
-                    range_length: Some(0),
-                    text: "s".to_string(),
-                },
-            ],
-            1,
-            PositionEncoding::UTF16,
-        );
-
-        assert_eq!(
-            &document.contents,
-            r#""""
-测试comment
-一些测试内容
-"""
-import click
-
-
-@click.group()
-def interface():
-    pass
-"#
-        );
-    }
-}

crates/red_knot_server/src/lib.rs

@@ -1,20 +1,23 @@
-use crate::server::Server;
+#![allow(dead_code)]
+
 use anyhow::Context;
-pub use document::{DocumentKey, NotebookDocument, PositionEncoding, TextDocument};
+pub use edit::{DocumentKey, NotebookDocument, PositionEncoding, TextDocument};
 pub use session::{ClientSettings, DocumentQuery, DocumentSnapshot, Session};
 use std::num::NonZeroUsize;
 
+use crate::server::Server;
+
 #[macro_use]
 mod message;
 
-mod document;
-mod logging;
+mod edit;
 mod server;
 mod session;
 mod system;
+mod trace;
 
-pub(crate) const SERVER_NAME: &str = "ty";
-pub(crate) const DIAGNOSTIC_NAME: &str = "ty";
+pub(crate) const SERVER_NAME: &str = "red-knot";
+pub(crate) const DIAGNOSTIC_NAME: &str = "Red Knot";
 
 /// A common result type used in most cases where a
 /// result type is needed.
@@ -30,7 +33,7 @@ pub fn run_server() -> anyhow::Result<()> {
     // by default, we set the number of worker threads to `num_cpus`, with a maximum of 4.
     let worker_threads = std::thread::available_parallelism()
         .unwrap_or(four)
-        .min(four);
+        .max(four);
 
     Server::new(worker_threads)
         .context("Failed to start server")?