diff --git a/.github/workflows/cache-retention.yml b/.github/workflows/cache-retention.yml index 580f2d1fa3..c704777fd5 100644 --- a/.github/workflows/cache-retention.yml +++ b/.github/workflows/cache-retention.yml @@ -4,6 +4,7 @@ on: workflow_run: workflows: ["CI"] types: [completed] + branches: [main] # Qualification calls ci.yml with actions: read. A nested workflow cannot raise # that token, so cache deletion stays in this workflow. @@ -21,13 +22,13 @@ jobs: # Restrict the triggering-source checkout to same-repository default-branch # CI events. The helper runs from this trusted workflow revision; triggering # source is used only for cache identity and is never executed. Static and - # build save snapshots before later lanes finish, so failed, timed-out, or - # cancelled runs still retain those snapshots. + # build save snapshots before later lanes finish, so failed or timed-out + # runs still retain those snapshots, while queue-superseded cancelled runs + # never published caches. if: >- (github.event.workflow_run.conclusion == 'success' || github.event.workflow_run.conclusion == 'failure' || - github.event.workflow_run.conclusion == 'timed_out' || - github.event.workflow_run.conclusion == 'cancelled') && + github.event.workflow_run.conclusion == 'timed_out') && github.event.workflow_run.head_repository.full_name == github.repository && github.event.workflow_run.head_branch == github.event.repository.default_branch && (github.event.workflow_run.event == 'push' || @@ -60,5 +61,6 @@ jobs: CACHE_KNIP_PREFIX: knip-meta-v1-${{ runner.os }}- CACHE_KNIP_KEEP_KEY: knip-meta-v1-${{ runner.os }}-${{ hashFiles('triggering-source/knip.json', 'triggering-source/tsconfig.json', 'triggering-source/package.json', 'triggering-source/package-lock.json') }}-${{ github.event.workflow_run.head_sha }} CACHE_PORTABLE_PREFIX: rea-portable-build-v2-${{ runner.os }}-${{ runner.arch }}- + CACHE_WINDOWS_BUILD_PREFIX: rea-windows-build-v1-Windows-X64- DEDUPE_NPM_DOWNLOADS: "true" run: node trusted-retention/scripts/ci/cache-retention.mjs diff --git a/.github/workflows/ci-timings.yml b/.github/workflows/ci-timings.yml index 74f27af32c..682dd09b70 100644 --- a/.github/workflows/ci-timings.yml +++ b/.github/workflows/ci-timings.yml @@ -2,7 +2,7 @@ name: CI timings on: schedule: - - cron: "17 * * * *" + - cron: "17 */6 * * *" workflow_dispatch: concurrency: diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml index ecad1fded4..bc00c5e565 100644 --- a/.github/workflows/ci.yml +++ b/.github/workflows/ci.yml @@ -84,6 +84,8 @@ jobs: agentenv: ${{ steps.plan.outputs.agentenv }} nativeaot: ${{ steps.plan.outputs.nativeaot }} ghidra_bridge: ${{ steps.plan.outputs.ghidra_bridge }} + go_binary: ${{ steps.plan.outputs.go_binary }} + android_metadata: ${{ steps.plan.outputs.android_metadata }} website: ${{ steps.plan.outputs.website }} matrix: ${{ steps.plan.outputs.matrix }} test_selection: ${{ steps.plan.outputs.test_selection }} @@ -94,6 +96,7 @@ jobs: with: ref: ${{ inputs.source_sha || github.sha }} fetch-depth: 0 + filter: blob:none - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0 with: node-version-file: .nvmrc @@ -334,7 +337,7 @@ jobs: fail-fast: ${{ github.event_name == 'pull_request' && needs.changes.outputs.tier == 'fast' }} matrix: # scripts/ci/test-selection.mjs owns the shard counts: twelve fast - # built shards, six full-qualification shards with coverage, or one + # built shards, twelve full-qualification shards with coverage, or one # job per group of a selected PR subset. shard: ${{ fromJSON(needs.changes.outputs.test_shards) }} steps: @@ -429,7 +432,8 @@ jobs: needs.changes.outputs.web_source_map == 'true' || needs.changes.outputs.agentenv == 'true' || needs.changes.outputs.packet_capture == 'true' || - needs.changes.outputs.signal_recording == 'true' + needs.changes.outputs.signal_recording == 'true' || + needs.changes.outputs.go_binary == 'true' runs-on: ubuntu-latest timeout-minutes: 10 permissions: @@ -688,7 +692,28 @@ jobs: - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0 with: node-version-file: .nvmrc + # Include runner ImageOS/ImageVersion and a UTC date bucket so runner image + # rollouts invalidate immediately and cached MinGW outputs refresh daily. + - name: Compute Windows native cache key + id: windows-native-key + env: + SOURCE_HASH: ${{ hashFiles('native/windows/**', 'src/domain/windowsNativeArtifact*.ts', 'src/domain/dosMz*.ts', 'src/domain/peCoffHeader*.ts', 'scripts/build-windows-native.mjs', 'scripts/verify-windows-native-artifact.mjs', 'scripts/lib/verifier-source-loader.mjs', 'tests/fixtures/windows-native-inspection-failure.cc', 'tests/fixtures/windows/processBoundary.cc', 'package.json') }} + run: | + utc_date="$(date -u +%Y-%m-%d)" + printf 'key=rea-windows-native-v2-%s-%s-%s-%s-%s-%s\n' \ + "${RUNNER_OS}" "${RUNNER_ARCH}" "${ImageOS:-ubuntu}" "${ImageVersion:-rolling}" "${utc_date}" "${SOURCE_HASH}" \ + >> "${GITHUB_OUTPUT}" + - name: Restore Windows native build outputs + id: windows-native-cache + uses: actions/cache/restore@0057852bfaa89a56745cba8c7296529d2fc39830 # v4.3.0 + with: + path: | + native/windows/build/ + .cache/windows-native-inspection-fixture/ + .cache/windows-process-boundary/ + key: ${{ steps.windows-native-key.outputs.key }} - name: Prepare cross compiler + if: steps.windows-native-cache.outputs.cache-hit != 'true' run: | # Slow Azure archive transfers can exhaust the job before compilation. # Keep the same signed packages and avoid slow HTTP mirror fallback. @@ -699,9 +724,11 @@ jobs: done sudo apt-get update sudo apt-get install --no-install-recommends -y g++-mingw-w64-x86-64 - - run: npm run build:windows-native + - if: steps.windows-native-cache.outputs.cache-hit != 'true' + run: npm run build:windows-native - run: npm run verify:windows-native:artifact - - run: node scripts/build-windows-native.mjs --inspection-fixture + - if: steps.windows-native-cache.outputs.cache-hit != 'true' + run: node scripts/build-windows-native.mjs --inspection-fixture - uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2 with: name: windows-native-inspection-fixture @@ -712,11 +739,30 @@ jobs: # results from a separate executable. Cross-build it in the lane that # already has MinGW so the Windows job can observe them too. - name: Build Windows process boundary fixture + if: steps.windows-native-cache.outputs.cache-hit != 'true' run: | mkdir -p .cache/windows-process-boundary x86_64-w64-mingw32-g++ -municode -std=c++17 -O2 -static -Wall -Wextra \ -o .cache/windows-process-boundary/process-boundary.exe \ tests/fixtures/windows/processBoundary.cc -ladvapi32 + - name: Verify cross-built Windows test fixtures + run: >- + node -e "const fs = require('node:fs'); + for (const p of ['.cache/windows-native-inspection-fixture/manifest.json', '.cache/windows-process-boundary/process-boundary.exe']) { + if (fs.statSync(p).size <= 0) throw new Error('empty Windows fixture: ' + p); }" + - name: Save validated Windows native build outputs + if: >- + success() && + steps.windows-native-cache.outputs.cache-hit != 'true' && + github.ref == format('refs/heads/{0}', github.event.repository.default_branch) && + (github.event_name == 'push' || github.event_name == 'schedule' || github.event_name == 'workflow_dispatch') + uses: actions/cache/save@0057852bfaa89a56745cba8c7296529d2fc39830 # v4.3.0 + with: + path: | + native/windows/build/ + .cache/windows-native-inspection-fixture/ + .cache/windows-process-boundary/ + key: ${{ steps.windows-native-cache.outputs.cache-primary-key }} - uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2 with: name: windows-process-boundary-fixture @@ -766,9 +812,17 @@ jobs: console.log(JSON.stringify({ platform: process.platform, architecture: process.arch, node: process.version }));" - run: npm ci - uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093 # v4.3.0 + if: matrix.os != 'windows-latest' with: name: rea-build-cache path: .turbo/cache/ + - name: Restore Windows Turbo build cache + if: matrix.os == 'windows-latest' + uses: actions/cache/restore@0057852bfaa89a56745cba8c7296529d2fc39830 # v4.3.0 + with: + path: .turbo/cache/ + key: rea-windows-build-v1-${{ runner.os }}-${{ runner.arch }}-${{ hashFiles('.nvmrc', '.npmrc', 'package-lock.json') }}-${{ github.sha }} + restore-keys: rea-windows-build-v1-${{ runner.os }}-${{ runner.arch }}-${{ hashFiles('.nvmrc', '.npmrc', 'package-lock.json') }}- - uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093 # v4.3.0 with: name: windows-native-controls @@ -819,21 +873,19 @@ jobs: env: REA_ARTIFACT_NODE_COMPILE_CACHE: ${{ runner.temp }}/rea-artifact-node-cache run: | - npm run verify:keyed-archive - npm run verify:asset-catalog - npm run verify:macos-bundle - npm run verify:legacy-simulator - npm run verify:apple-dispatch - npm run verify:interface-builder - - name: Verify portable iNES CLI, MCP and source boundaries + npm run build:cached + node scripts/verify/apple/keyed-archive.mjs + node scripts/verify/apple/asset-catalog.mjs + node scripts/verify/apple/macos-bundle.mjs + node scripts/verify/apple/legacy-simulator.mjs + node scripts/verify/apple/dispatch.mjs + node scripts/verify/apple/interface-builder.mjs + - name: Run native Apple, iNES, and process boundary regressions run: >- npm run test:focused -- tests/acceptance/inesContainer.test.ts tests/boundary/filesystem/inesArtifact.test.ts tests/boundary/filesystem/inesCancellation.test.ts - - name: Run native Apple and process boundary regressions - run: >- - npm run test:focused -- tests/boundary/providers/hopper/hopperMacLaunchArguments.test.ts tests/acceptance/providers/hopperStartupRecovery.test.ts tests/boundary/providers/hopper/hopperClient.helperFailure.test.ts @@ -922,16 +974,27 @@ jobs: GOMAXPROCS: "2" GOMEMLIMIT: "3GiB" run: npm run typecheck -- --checkers 2 - - uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093 # v4.3.0 + - name: Restore Windows Turbo build cache + id: windows-build-cache + uses: actions/cache/restore@0057852bfaa89a56745cba8c7296529d2fc39830 # v4.3.0 with: - name: rea-build-cache path: .turbo/cache/ + key: rea-windows-build-v1-${{ runner.os }}-${{ runner.arch }}-${{ hashFiles('.nvmrc', '.npmrc', 'package-lock.json') }}-${{ github.sha }} + restore-keys: rea-windows-build-v1-${{ runner.os }}-${{ runner.arch }}-${{ hashFiles('.nvmrc', '.npmrc', 'package-lock.json') }}- - uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093 # v4.3.0 with: name: windows-native-controls path: native/windows/build/ - run: npm run verify:windows-native:artifact - run: npm run build:cached + - name: Save validated Windows Turbo build cache + if: >- + success() && steps.npm-cache.outputs.can-save == 'true' && + steps.windows-build-cache.outputs.cache-hit != 'true' + uses: actions/cache/save@0057852bfaa89a56745cba8c7296529d2fc39830 # v4.3.0 + with: + path: .turbo/cache/ + key: ${{ steps.windows-build-cache.outputs.cache-primary-key }} - name: Record named Windows host capabilities run: node scripts/windows-capabilities.mjs | Out-File -Encoding utf8NoBOM windows-capabilities.json - name: Validate Windows host capability report @@ -1022,6 +1085,9 @@ jobs: tests/process-global/runnerCancellation.test.ts tests/process-global/windowsOwnedCommand.test.ts tests/conformance/providers/ghidra/realGhidraWindowsWorkflow.test.ts + tests/boundary/filesystem/directoryArtifactReaderOpen.test.ts + tests/boundary/filesystem/javascriptPackagePrecedence.test.ts + tests/boundary/filesystem/javascriptArtifactContentIdentity.test.ts - name: Verify real Node Inspector discovery and observation timeout-minutes: 10 env: @@ -1029,12 +1095,6 @@ jobs: UV_THREADPOOL_SIZE: "2" run: npm run verify:inspector - run: npm run verify:package:windows - - name: Verify native directory identity on Node 24 - run: >- - npx vitest run --coverage.enabled=false - tests/boundary/filesystem/directoryArtifactReaderOpen.test.ts - tests/boundary/filesystem/javascriptPackagePrecedence.test.ts - tests/boundary/filesystem/javascriptArtifactContentIdentity.test.ts - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0 with: node-version: 22.23.4 @@ -1067,6 +1127,8 @@ jobs: id: npm-cache uses: ./.github/actions/npm-cache - run: npm ci + - name: Reuse portable build outputs across runs + uses: ./.github/actions/portable-build-cache - run: npm run docs:check - name: Verify published routes, source links and local anchors run: npm run verify:docs:routes @@ -1202,6 +1264,16 @@ jobs: build_cache_artifact: rea-build-cache remote_cache_write: ${{ inputs.remote_cache_write == true || (github.ref == format('refs/heads/{0}', github.event.repository.default_branch) && (github.event_name == 'push' || github.event_name == 'schedule' || github.event_name == 'workflow_dispatch')) }} + go_binary: + needs: [changes, static, build] + if: needs.changes.outputs.go_binary == 'true' + secrets: + TURBO_TOKEN: ${{ secrets.TURBO_TOKEN }} + uses: ./.github/workflows/real-go-binary.yml + with: + build_cache_artifact: rea-build-cache + remote_cache_write: ${{ inputs.remote_cache_write == true || (github.ref == format('refs/heads/{0}', github.event.repository.default_branch) && (github.event_name == 'push' || github.event_name == 'schedule' || github.event_name == 'workflow_dispatch')) }} + nativeaot: needs: [changes, static] if: needs.changes.outputs.nativeaot == 'true' @@ -1212,6 +1284,11 @@ jobs: if: needs.changes.outputs.ghidra_bridge == 'true' uses: ./.github/workflows/ghidra-bridge-sources.yml + android_metadata: + needs: [changes, static] + if: needs.changes.outputs.android_metadata == 'true' + uses: ./.github/workflows/real-android-metadata.yml + website: needs: [changes, static] if: needs.changes.outputs.website == 'true' @@ -1252,8 +1329,10 @@ jobs: web_runtime, web_source_map, agentenv, + go_binary, nativeaot, ghidra_bridge, + android_metadata, website, ] runs-on: ubuntu-24.04 diff --git a/.github/workflows/codeql.yml b/.github/workflows/codeql.yml index 2a6342aabd..74dc8952fb 100644 --- a/.github/workflows/codeql.yml +++ b/.github/workflows/codeql.yml @@ -3,8 +3,32 @@ name: CodeQL production on: push: branches: [main] + paths-ignore: + - "tests/**" + - "scripts/fixtures/**" + - "**/*.test.*" + - "**/*.fixture.*" + - "**/test_*.py" + - "**/*_test.py" + - "**/*.md" + - "**/*.mermaid" + - "**/*.csv" + - "licenses/**" + - ".git-blame-ignore-revs" pull_request: branches: [main] + paths-ignore: + - "tests/**" + - "scripts/fixtures/**" + - "**/*.test.*" + - "**/*.fixture.*" + - "**/test_*.py" + - "**/*_test.py" + - "**/*.md" + - "**/*.mermaid" + - "**/*.csv" + - "licenses/**" + - ".git-blame-ignore-revs" schedule: - cron: "37 5 * * 1" workflow_dispatch: @@ -18,10 +42,35 @@ concurrency: cancel-in-progress: ${{ github.event_name == 'pull_request' }} jobs: - analyze: + scope: # A repository administrator must disable default setup before enabling # this workflow; default setup rejects advanced CodeQL result uploads. if: vars.REA_CODEQL_ADVANCED == 'true' + name: Select production languages + runs-on: ubuntu-24.04 + timeout-minutes: 5 + outputs: + run: ${{ steps.scope.outputs.run }} + languages: ${{ steps.scope.outputs.languages }} + steps: + - uses: actions/checkout@d23441a48e516b6c34aea4fa41551a30e30af803 # v6.1.0 + with: + fetch-depth: ${{ github.event_name == 'pull_request' && 0 || 1 }} + filter: blob:none + persist-credentials: false + - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0 + with: + node-version-file: .nvmrc + - name: Select CodeQL languages + id: scope + env: + BASE_SHA: ${{ github.event.pull_request.base.sha }} + HEAD_SHA: ${{ github.event.pull_request.head.sha || github.sha }} + run: node scripts/ci/codeql-scope.mjs >> "$GITHUB_OUTPUT" + + analyze: + needs: scope + if: vars.REA_CODEQL_ADVANCED == 'true' && needs.scope.outputs.run == 'true' name: Analyze production (${{ matrix.language }}) runs-on: ubuntu-latest timeout-minutes: 45 @@ -32,7 +81,7 @@ jobs: fail-fast: false max-parallel: 1 matrix: - language: [javascript-typescript, python, actions] + language: ${{ fromJSON(needs.scope.outputs.languages) }} steps: - uses: actions/checkout@d23441a48e516b6c34aea4fa41551a30e30af803 # v6.1.0 with: diff --git a/.github/workflows/real-android-metadata.yml b/.github/workflows/real-android-metadata.yml index a799da1479..8566068504 100644 --- a/.github/workflows/real-android-metadata.yml +++ b/.github/workflows/real-android-metadata.yml @@ -1,28 +1,12 @@ name: Real Android metadata bridge on: - push: - branches: ["**"] - paths: - - .github/workflows/real-android-metadata.yml - - bridge/android/** - - src/android/JadxRelease.ts - - scripts/fixtures/android/** - - tests/conformance/android/** - - tests/fixtures/android/static-initializers/** - - tests/fixtures/android/direct-invokes/** - - tests/fixtures/android/external-invokes/** - - .nvmrc workflow_dispatch: workflow_call: permissions: contents: read -concurrency: - group: android-metadata-${{ github.ref }} - cancel-in-progress: true - jobs: android-metadata: runs-on: ubuntu-24.04 diff --git a/.github/workflows/real-go-binary.yml b/.github/workflows/real-go-binary.yml index a50c3db396..02dbe312ef 100644 --- a/.github/workflows/real-go-binary.yml +++ b/.github/workflows/real-go-binary.yml @@ -1,38 +1,37 @@ name: Real Go binary metadata on: + workflow_call: + inputs: + build_cache_artifact: + description: Portable build cache artifact from the calling CI run. + required: false + type: string + default: "" + remote_cache_write: + description: Trusted release qualification may publish the remote cache. + required: false + type: boolean + default: false + secrets: + TURBO_TOKEN: + description: Remote cache token. Only trusted callers pass it. + required: false workflow_dispatch: - pull_request: - paths: - - ".github/workflows/real-go-binary.yml" - - "src/go/**" - - "src/domain/go/**" - - "src/domain/dosMz.ts" - - "src/application/go/**" - - "src/contracts/go/**" - - "src/composition/go.ts" - - "src/cli/goCommands.ts" - - "src/server/contractToolBindings.ts" - - "src/server/serviceToolBindings.ts" - - "src/server/createServer.ts" - - "scripts/verify-go-binary.mjs" - - "scripts/lib/mcp-verifier-results.mjs" permissions: contents: read -concurrency: - group: real-go-binary-${{ github.event.pull_request.number || github.ref }} - cancel-in-progress: true - env: NODE_OPTIONS: --max-old-space-size=768 HUSKY: "0" - # Writes only on the default branch for push, schedule, manual dispatch, - # and workflow_call. Pull requests get an empty token and remote read-only mode. - TURBO_TOKEN: "${{ (github.ref == format('refs/heads/{0}', github.event.repository.default_branch) && (github.event_name == 'push' || github.event_name == 'schedule' || github.event_name == 'workflow_dispatch' || github.event_name == 'workflow_call')) && secrets.TURBO_TOKEN || '' }}" + # A called workflow sees the caller event, so workflow_call is not an event + # name here. Writes when remote_cache_write is set, or on the default branch + # for push, schedule, and manual dispatch. Open pull requests stay empty-token + # and use the local cache only. + TURBO_TOKEN: "${{ (inputs.remote_cache_write == true || (github.ref == format('refs/heads/{0}', github.event.repository.default_branch) && (github.event_name == 'push' || github.event_name == 'schedule' || github.event_name == 'workflow_dispatch'))) && secrets.TURBO_TOKEN || '' }}" TURBO_TEAM: ${{ vars.TURBO_TEAM }} - TURBO_CACHE: "${{ (github.ref == format('refs/heads/{0}', github.event.repository.default_branch) && (github.event_name == 'push' || github.event_name == 'schedule' || github.event_name == 'workflow_dispatch' || github.event_name == 'workflow_call')) && 'local:rw,remote:rw' || 'local:rw,remote:r' }}" + TURBO_CACHE: "${{ (inputs.remote_cache_write == true || (github.ref == format('refs/heads/{0}', github.event.repository.default_branch) && (github.event_name == 'push' || github.event_name == 'schedule' || github.event_name == 'workflow_dispatch'))) && 'local:rw,remote:rw' || 'local:rw' }}" jobs: real-go-binary: @@ -51,5 +50,8 @@ jobs: go-version: "1.26.1" cache: false - run: npm ci + - uses: ./.github/actions/portable-build-cache + with: + build_cache_artifact: ${{ inputs.build_cache_artifact }} - name: Verify compiler-produced ELF, PE and Mach-O through public CLI and MCP run: npm run verify:go:binary diff --git a/docs/development/testing.md b/docs/development/testing.md index 5876c82855..17ca443826 100644 --- a/docs/development/testing.md +++ b/docs/development/testing.md @@ -544,9 +544,14 @@ relying on the new scope. While the variable is unset, the existing default setup continues scanning. Adding the configuration file alone does not change default setup in this user-owned repository. -Pull-request scans are cancelled when superseded. Active scans on `main` finish -so frequent pushes cannot prevent a complete baseline from being uploaded. -Cancelled runs skip the SARIF artifact upload when no report was produced. +Pull-request scans are cancelled when superseded, skip pure test/prose diffs via +`paths-ignore`, and select only the CodeQL languages (`javascript-typescript`, +`python`, `actions`) whose production inputs changed (failing closed to all +three languages on shared CodeQL configuration or unclassified paths). Active +scans on `main`, weekly schedules, and manual dispatches always analyze all +three languages so frequent pushes cannot prevent a complete baseline from being +uploaded. Cancelled runs skip the SARIF artifact upload when no report was +produced. Each analysis uses two threads and a 4096 MiB CodeQL memory budget; language jobs run one at a time. Complete SARIF reports are retained for seven days. @@ -1168,7 +1173,7 @@ or qualify the release revision for mutation coverage. Fast CI starts the source-only projects after classification, independently of build and static checks, and runs twelve built-dependent shards after the build. -Those thirteen test jobs still select every portable test. Full qualification keeps six +Those thirteen test jobs still select every portable test. Full qualification keeps twelve complete Vitest shards with coverage. Neither tier retries tests. Canonical project ownership lives in `scripts/lib/test-projects.mjs`, shared by Vitest, CI group execution, and development-test preparation. A test importing compiled @@ -1249,11 +1254,13 @@ separate and grows when more shard jobs compete for runners. The npm download cache stays the setup cache rather than a shared `node_modules` artifact. `process-boundary` stays serial because process-tree sampling shares the host. Hosted CI jobs use four workers; other hosts stay at two -because they are shared with other checks. Full qualification keeps six covered +because they are shared with other checks. Full qualification keeps twelve covered shards; its longer instrumentation is outside this fast budget. The separate `CI timings` workflow batches recent completed CI and Release runs -hourly or on manual dispatch. It uploads `ci-job-timings-batch` with raw API pages and +every six hours or on manual dispatch, and snapshots pull-request CI health across +the preceding twelve hours using full ISO 8601 timestamps within a ten-page +(1,000-run) API budget. It uploads `ci-job-timings-batch` with raw API pages and a normalized point-in-time report for each collected workflow attempt, plus a batch manifest describing examined coverage, failures, and limits. Overlapping lookback windows may collect an attempt again; identify records by run ID and @@ -1283,9 +1290,18 @@ rebuilds misses. A restore prefix with the same platform and toolchain permits partial task reuse when one generator's inputs change; Turbo still invalidates that generator's entry. Native outputs are not stored in this cache. -Native package jobs, the Windows curated lane, and the Apple artifact lane reuse -the same workflow's Turbo cache for portable compiled JavaScript, skills, and -catalogs. The key parser retains the validated `test:prepare` task identities in +POSIX native package jobs, documentation checks, Go binary verification, and the +Apple artifact lane reuse the same workflow's portable Turbo cache +(`rea-build-cache`) for compiled JavaScript, skills, and catalogs. Windows jobs +(`windows-curated` and Windows `package-e2e`) compile with `typescript` rather +than Linux's accelerated `tsrs` checker, so they restore and publish a separate +Windows `.turbo/cache/` family (`rea-windows-build-v1-Windows-X64-*`, pruned to +two generations by `cache-retention.mjs`). `windows-native-build` caches +cross-compiled MinGW outputs keyed by source/fixture hashes, runner +`ImageOS`/`ImageVersion`, and UTC date (bounding lifetime to 24 hours so +upstream toolchain updates refresh daily) while running +`npm run verify:windows-native:artifact` unconditionally on every cache hit. The +key parser retains the validated `test:prepare` task identities in `.cache/ci-build-plan.json`; staging reads this plan without starting another Turbo run that could evict cache entries. The producer stages matching archives, metadata, and available manifests into `.cache/ci-build-cache/`. @@ -1355,24 +1371,32 @@ revision and selects: fake provider process; - tests whose text names a changed path. -Verifier scripts (`scripts/verify-*.mjs`, `scripts/verify/**/*.mjs`) are traced -like source modules. Changed Markdown outside `src/` and `tests/` selects only -the tests that name it. Declarations, guard baselines -(`scripts/verify/baselines/*.json`) and advertised-schema summaries select no -tests, because the static and build jobs check them. +Verifier scripts (`scripts/verify-*.mjs`, `scripts/verify/**/*.mjs`) and shared +script libraries (`scripts/lib/*`, when no untraced consumer imports or names +them) are traced like source modules. Changed Markdown, non-code `docs/` assets, +issue templates, `LICENSE`, and `licenses/` inventories outside `src/` and +`tests/` select only the tests that name them. Declarations, guard and layer +baselines (`scripts/verify/baselines/*.json`, +`scripts/lib/module-layers.baseline.json`), `.git-blame-ignore-revs`, Python +analysis-efficiency test scripts, and advertised-schema summaries select no +portable Vitest tests unless named by path, because static, Python, and build +jobs check them. Non-module CLI parity fixtures with a single known test suite +are bound in `EXPLICIT_FIXTURE_OWNERS` and fail closed if their owner is missing. The PR falls back to the complete fast suite when it cannot decide: -- a shared boundary (`src/process/`, `src/filesystem/`, `src/contracts/tool*`, - `src/server/createServer*`, `src/cli.ts`, `scripts/ci/`) changed; +- a shared production boundary (`src/process/`, `src/filesystem/`, + `src/contracts/tool*`, `src/server/createServer*`, `src/cli.ts`, + `scripts/ci/`) changed; - any other changed input is not traced, such as workflows, package metadata, - build and test-runner scripts, `scripts/lib/`, fixtures or data; + build and test-runner scripts, untraced `scripts/lib/` consumers, or unmapped + fixtures and data; - no test is selected; -- more than 150 test files are selected; +- more than 400 test files are selected; - a selected file has no single owning test project. -A selection runs as one `Test shard (k/n)` job per group of up to 40 files, at -most four jobs, using exact file identity (`scripts/ci/test-group.mjs selected`). +A selection runs as one `Test shard (k/n)` job per group of up to 50 files, at +most eight jobs, using exact file identity (`scripts/ci/test-group.mjs selected`). The separate `Source tests` job does not run, because the groups include the selected source tests. The `Classify changes` summary gives the mode and its reason. diff --git a/scripts/ci/cache-retention.mjs b/scripts/ci/cache-retention.mjs index 24abf4fdae..019433615a 100644 --- a/scripts/ci/cache-retention.mjs +++ b/scripts/ci/cache-retention.mjs @@ -18,6 +18,7 @@ const PREFIX_PATTERNS = [ /^rea-static-v1-[A-Za-z0-9]+-[A-Za-z0-9]+-$/u, /^rea-static-deep-v1-[A-Za-z0-9]+-[A-Za-z0-9]+-$/u, /^rea-portable-build-v2-[A-Za-z0-9]+-[A-Za-z0-9]+-$/u, + /^rea-windows-build-v1-[A-Za-z0-9]+-[A-Za-z0-9]+-$/u, /^knip-meta-v1-[A-Za-z0-9]+-$/u, ]; @@ -272,6 +273,10 @@ function familiesFromEnv(env) { env.CACHE_PORTABLE_PREFIX.length > 0 ? { prefix: env.CACHE_PORTABLE_PREFIX } : undefined, + typeof env.CACHE_WINDOWS_BUILD_PREFIX === "string" && + env.CACHE_WINDOWS_BUILD_PREFIX.length > 0 + ? { prefix: env.CACHE_WINDOWS_BUILD_PREFIX } + : undefined, ].filter((family) => family !== undefined); if (families.length === 0 && env.DEDUPE_NPM_DOWNLOADS !== "true") throw new Error( diff --git a/scripts/ci/change-plan.mjs b/scripts/ci/change-plan.mjs index 112db3889f..1cc2bf4f80 100644 --- a/scripts/ci/change-plan.mjs +++ b/scripts/ci/change-plan.mjs @@ -22,6 +22,7 @@ const GUARDED_WORKFLOWS = new Set([ ".github/workflows/real-ghidra-windows.yml", ".github/workflows/sqlite-runtime.yml", ".github/workflows/ci-timings.yml", + ".github/workflows/cache-retention.yml", ".github/workflows/mutation.yml", ]); // Sources compiled by build-conformance-fixtures.mjs. The readiness journey @@ -189,6 +190,17 @@ export function planChanges(request, repository) { selected[owner] = true; // The documentation check scans living prose outside docs/ too. if (isLivingProsePath(file)) selected.docs = true; + if (file === ".git-blame-ignore-revs") return; + if ( + file.startsWith("licenses/") || + file === "THIRD_PARTY_NOTICES.md" || + file === "DISCLAIMER.md" || + file === "LICENSE" + ) { + selected.docs = true; + selected.code = true; + return; + } if ( /^(README(?:_[^/]+)?\.md|AGENTS\.md|CONTRIBUTING\.md|CHANGELOG\.md|SECURITY\.md|native\/windows\/README\.md|\.github\/pull_request_template\.md)$/.test( file, diff --git a/scripts/ci/codeql-scope.d.mts b/scripts/ci/codeql-scope.d.mts new file mode 100644 index 0000000000..67f82b0de4 --- /dev/null +++ b/scripts/ci/codeql-scope.d.mts @@ -0,0 +1,25 @@ +export type CodeqlLanguage = "javascript-typescript" | "python" | "actions"; + +export interface CodeqlScopeRequest { + readonly event: string | undefined; + readonly base: string | undefined; + readonly head: string | undefined; +} + +export interface CodeqlScopeRepository { + readonly mergeBase: (base: string, head: string) => string; + readonly changedFiles: (base: string, head: string) => readonly string[]; +} + +export interface CodeqlScopePlan { + readonly run: boolean; + readonly languages: readonly CodeqlLanguage[]; + readonly reason: string; +} + +export const CODEQL_LANGUAGES: readonly CodeqlLanguage[]; + +export function planCodeqlLanguages( + request: CodeqlScopeRequest, + repository: CodeqlScopeRepository, +): CodeqlScopePlan; diff --git a/scripts/ci/codeql-scope.mjs b/scripts/ci/codeql-scope.mjs new file mode 100644 index 0000000000..4874e3673b --- /dev/null +++ b/scripts/ci/codeql-scope.mjs @@ -0,0 +1,169 @@ +import { execFileSync } from "node:child_process"; +import { matchesGlob } from "node:path"; +import { pathToFileURL } from "node:url"; + +/** @type {readonly import("./codeql-scope.d.mts").CodeqlLanguage[]} */ +export const CODEQL_LANGUAGES = ["javascript-typescript", "python", "actions"]; + +// Inputs excluded by .github/codeql/production.yml plus pure prose, assets, and +// source languages outside CodeQL's javascript-typescript/python/actions extractors. +const EXCLUDED_PATTERNS = [ + "tests/**", + "scripts/fixtures/**", + "**/*.test.*", + "**/*.fixture.*", + "**/test_*.py", + "**/*_test.py", + "**/*.md", + "**/*.mermaid", + "**/*.csv", + "licenses/**", + "LICENSE", + ".git-blame-ignore-revs", + "website/locales/**/*.po", + "website/locales/**/*.pot", + "**/*.{png,jpg,jpeg,gif,svg,ico,webp}", + "bridge/**/*.{java,swift}", + "native/windows/**/*.{cc,cpp,c,h,hpp}", + "**/*.{go,sol,rs}", +]; + +const isExcludedFromCodeql = (file) => + !file.startsWith("docs/.vitepress/") && + EXCLUDED_PATTERNS.some((pattern) => matchesGlob(file, pattern)); + +const isActionsInput = (file) => + /^\.github\/workflows\/[^/]+\.ya?ml$/u.test(file) || + /^\.github\/actions\/.+\/action\.ya?ml$/u.test(file); + +const isPythonInput = (file) => file.endsWith(".py"); + +const isJavaScriptTypeScriptInput = (file) => + file.startsWith("docs/.vitepress/") || + file === ".npmrc" || + file === ".nvmrc" || + /\.(?:[cm]?[jt]s|[jt]sx|html?|xhtml|vue|jsonc?)$/u.test(file); + +/** + * Select the CodeQL languages required for one event. Non-PR baselines and any + * unrecognized or scan-configuration inputs select all configured languages. + * + * @param {import("./codeql-scope.d.mts").CodeqlScopeRequest} request + * @param {import("./codeql-scope.d.mts").CodeqlScopeRepository} repository + * @returns {import("./codeql-scope.d.mts").CodeqlScopePlan} + */ +export function planCodeqlLanguages(request, repository) { + if (request.event !== "pull_request") { + return { + run: true, + languages: [...CODEQL_LANGUAGES], + reason: `Complete CodeQL baseline for ${request.event ?? "unknown"} event`, + }; + } + const { base, head } = request; + if ( + !base || + !head || + !/^[0-9a-f]{40}$/iu.test(base) || + !/^[0-9a-f]{40}$/iu.test(head) + ) { + return { + run: true, + languages: [...CODEQL_LANGUAGES], + reason: "Missing or invalid pull-request revision; running all languages", + }; + } + let files; + try { + const mergeBase = repository.mergeBase(base, head); + files = repository.changedFiles(mergeBase, head); + } catch (error) { + const message = error instanceof Error ? error.message : String(error); + return { + run: true, + languages: [...CODEQL_LANGUAGES], + reason: `Git diff failed (${message}); running all languages`, + }; + } + + /** @type {Set} */ + const selected = new Set(); + for (const file of files) { + if ( + file.startsWith(".github/codeql/") || + file === ".github/workflows/codeql.yml" + ) { + return { + run: true, + languages: [...CODEQL_LANGUAGES], + reason: `CodeQL configuration changed: ${file}`, + }; + } + if (isExcludedFromCodeql(file)) continue; + let matched = false; + if (isActionsInput(file)) { + selected.add("actions"); + matched = true; + } + if (isPythonInput(file)) { + selected.add("python"); + matched = true; + } + if (isJavaScriptTypeScriptInput(file)) { + selected.add("javascript-typescript"); + matched = true; + } + if (!matched) { + return { + run: true, + languages: [...CODEQL_LANGUAGES], + reason: `Unclassified input selects all CodeQL languages: ${file}`, + }; + } + } + const languages = CODEQL_LANGUAGES.filter((language) => + selected.has(language), + ); + return { + run: languages.length > 0, + languages, + reason: + languages.length > 0 + ? `Selected CodeQL languages: ${languages.join(", ")}` + : "No CodeQL-extracted production files changed", + }; +} + +function gitCodeqlRepository(cwd = process.cwd()) { + const git = (args) => + execFileSync("git", args, { + cwd, + encoding: "utf8", + maxBuffer: 16 * 1024 * 1024, + }); + return { + mergeBase: (base, head) => git(["merge-base", base, head]).trim(), + changedFiles: (base, head) => + git(["diff", "--name-only", "-z", "--no-renames", `${base}..${head}`]) + .split("\0") + .filter(Boolean), + }; +} + +if ( + process.argv[1] && + import.meta.url === pathToFileURL(process.argv[1]).href +) { + const plan = planCodeqlLanguages( + { + event: process.env.GITHUB_EVENT_NAME, + base: process.env.BASE_SHA, + head: process.env.HEAD_SHA, + }, + gitCodeqlRepository(), + ); + process.stderr.write(`CodeQL language selection: ${plan.reason}\n`); + process.stdout.write( + `run=${String(plan.run)}\nlanguages=${JSON.stringify(plan.languages)}\n`, + ); +} diff --git a/scripts/ci/job-timings-batch.mjs b/scripts/ci/job-timings-batch.mjs index f2dca7143c..0aded74ceb 100644 --- a/scripts/ci/job-timings-batch.mjs +++ b/scripts/ci/job-timings-batch.mjs @@ -39,7 +39,7 @@ export async function collectJobTimingsBatch({ status: "completed", createdAfter: cutoff, lookbackHours: 24, - overlap: "Each hourly run re-reads the preceding 24 hours.", + overlap: "Each scheduled run re-reads the preceding 24 hours.", pagination: "Run listings use the first 100; jobs are paginated within the remaining time and 16 MiB response limits.", maxRunsPerWorkflow: MAX_RUNS_PER_WORKFLOW, diff --git a/scripts/ci/pr-ci-health.d.mts b/scripts/ci/pr-ci-health.d.mts index b00a74aeba..de7ef00028 100644 --- a/scripts/ci/pr-ci-health.d.mts +++ b/scripts/ci/pr-ci-health.d.mts @@ -45,6 +45,7 @@ export interface PrCiHealthReport { repositoryId: number | null; defaultBranch: string; lookbackDays: number; + lookbackHours?: number; scan: { status: "complete" | "partial"; cutoff: string; @@ -67,6 +68,7 @@ export function collectPrCiHealth(options: { repositoryId?: number | string | null; defaultBranch: string; lookbackDays?: number; + lookbackHours?: number; now?: number; apiRequest?: GitHubApiRequest; }): Promise; diff --git a/scripts/ci/pr-ci-health.mjs b/scripts/ci/pr-ci-health.mjs index a982e792f5..9330963e3e 100644 --- a/scripts/ci/pr-ci-health.mjs +++ b/scripts/ci/pr-ci-health.mjs @@ -7,12 +7,13 @@ import { createGhApiRequest } from "./github-api.mjs"; const OUTPUT_DIRECTORY = join(".cache", "ci-pr-health"); const PAGE_SIZE = 100; const MAX_PULL_REQUEST_PAGES = 3; -const MAX_WORKFLOW_RUN_PAGES = 5; +const MAX_WORKFLOW_RUN_PAGES = 10; const MAX_PULL_REQUEST_DETAILS = 40; const MAX_API_REQUESTS = MAX_PULL_REQUEST_PAGES + MAX_WORKFLOW_RUN_PAGES + MAX_PULL_REQUEST_DETAILS; -const LOOKBACK_DAYS = 14; +const LOOKBACK_HOURS = 12; const MAX_LOOKBACK_DAYS = 30; +const MAX_LOOKBACK_HOURS = MAX_LOOKBACK_DAYS * 24; const MAX_TOTAL_TIME_MS = 60_000; const SHA_PATTERN = /^[0-9a-f]{40}$/iu; const COMPLETED_CATEGORIES = new Map([ @@ -27,7 +28,8 @@ export async function collectPrCiHealth({ repository, repositoryId, defaultBranch, - lookbackDays = LOOKBACK_DAYS, + lookbackDays, + lookbackHours, now = Date.now(), apiRequest = createGhApiRequest(), }) { @@ -36,10 +38,11 @@ export async function collectPrCiHealth({ repositoryId, defaultBranch, lookbackDays, + lookbackHours, now, }); const observedAt = new Date(now).toISOString(); - const cutoff = now - identity.lookbackDays * 24 * 60 * 60 * 1_000; + const cutoff = now - identity.lookbackHours * 60 * 60 * 1_000; const snapshot = await collectApiSnapshot({ apiRequest, identity, cutoff }); const classified = classifyPullRequests(snapshot); const { pulls, runs, budget } = snapshot; @@ -53,6 +56,7 @@ export async function collectPrCiHealth({ repositoryId: identity.repositoryId, defaultBranch: identity.defaultBranch, lookbackDays: identity.lookbackDays, + lookbackHours: identity.lookbackHours, scan: { status: errors.length === 0 && pulls.complete && runs.complete @@ -518,8 +522,8 @@ async function listRecentWorkflowRuns({ requestGet, identity, cutoff }) { let error = null; for (let page = 1; page <= MAX_WORKFLOW_RUN_PAGES; page += 1) { try { - const cutoffDate = new Date(cutoff).toISOString().slice(0, 10); - const endpoint = `repos/${identity.repository}/actions/workflows/ci.yml/runs?event=pull_request&created=%3E%3D${cutoffDate}&per_page=${PAGE_SIZE}&page=${page}`; + const cutoffIso = new Date(cutoff).toISOString(); + const endpoint = `repos/${identity.repository}/actions/workflows/ci.yml/runs?event=pull_request&created=%3E%3D${encodeURIComponent(cutoffIso)}&per_page=${PAGE_SIZE}&page=${page}`; const response = await requestGet(endpoint); if (!isRecord(response) || !Array.isArray(response.workflow_runs)) throw new Error("workflow_run_page_malformed"); @@ -576,6 +580,7 @@ function validateInputs({ repositoryId, defaultBranch, lookbackDays, + lookbackHours, now, }) { if ( @@ -603,19 +608,33 @@ function validateInputs({ ) throw new Error("CI_REPOSITORY_ID must be a positive safe integer"); if ( - !Number.isInteger(lookbackDays) || - lookbackDays < 1 || - lookbackDays > MAX_LOOKBACK_DAYS + lookbackDays !== undefined && + (!Number.isInteger(lookbackDays) || + lookbackDays < 1 || + lookbackDays > MAX_LOOKBACK_DAYS) ) throw new Error( `lookbackDays must be an integer from 1 to ${String(MAX_LOOKBACK_DAYS)}`, ); + if ( + lookbackHours !== undefined && + (!Number.isInteger(lookbackHours) || + lookbackHours < 1 || + lookbackHours > MAX_LOOKBACK_HOURS) + ) + throw new Error( + `lookbackHours must be an integer from 1 to ${String(MAX_LOOKBACK_HOURS)}`, + ); + const resolvedHours = + lookbackHours ?? + (lookbackDays !== undefined ? lookbackDays * 24 : LOOKBACK_HOURS); if (!Number.isFinite(now)) throw new Error("now must be a finite timestamp"); return { repository, repositoryId: parsedRepositoryId, defaultBranch, - lookbackDays, + lookbackDays: resolvedHours / 24, + lookbackHours: resolvedHours, }; } @@ -664,7 +683,7 @@ function renderMarkdownSummary(report) { return [ "# Pull request CI health snapshot", "", - `Observed ${report.observedAt} for \`${report.repository}\`, target branch \`${escapeCell(report.defaultBranch)}\`; recent window ${String(report.lookbackDays)} days. Scan: **${report.scan.status}**.`, + `Observed ${report.observedAt} for \`${report.repository}\`, target branch \`${escapeCell(report.defaultBranch)}\`; recent window ${String(report.lookbackHours ?? report.lookbackDays * 24)} hours. Scan: **${report.scan.status}**.`, "", "This is read-only point-in-time telemetry. Missing is reported only when PR and workflow-run inventories are complete and the current immutable PR head identity is unambiguous. Incomplete lookups remain deferred or unknown.", "", diff --git a/scripts/ci/scopes.mjs b/scripts/ci/scopes.mjs index e57b91fe18..10f39b2a56 100644 --- a/scripts/ci/scopes.mjs +++ b/scripts/ci/scopes.mjs @@ -682,6 +682,31 @@ const PATH_SCOPES = { "tests/conformance/ghidra/loader-relocation*.c", ".github/workflows/ghidra-bridge-sources.yml", ], + go_binary: [ + ".github/workflows/real-go-binary.yml", + "src/go/**", + "src/domain/go/**", + "src/domain/dosMz.ts", + "src/application/go/**", + "src/contracts/go/**", + "src/composition/go.ts", + "src/cli/goCommands.ts", + "src/server/contractToolBindings.ts", + "src/server/serviceToolBindings.ts", + "src/server/createServer.ts", + "scripts/verify-go-binary.mjs", + "scripts/lib/mcp-verifier-results.mjs", + ], + android_metadata: [ + ".github/workflows/real-android-metadata.yml", + "bridge/android/**", + "src/android/JadxRelease.ts", + "scripts/fixtures/android/**", + "tests/conformance/android/**", + "tests/fixtures/android/static-initializers/**", + "tests/fixtures/android/direct-invokes/**", + "tests/fixtures/android/external-invokes/**", + ], website: [ "website/**", "scripts/*website*.py", @@ -754,6 +779,12 @@ export const CI_LANES = { agentenv: sameJobs(PATH_SCOPES.agentenv, ["agentenv"]), nativeaot: sameJobs(PATH_SCOPES.nativeaot, ["nativeaot"]), ghidra_bridge: sameJobs(PATH_SCOPES.ghidra_bridge, ["ghidra_bridge"]), + go_binary: sameJobs(PATH_SCOPES.go_binary, ["go_binary"]), + android_metadata: sameJobs( + PATH_SCOPES.android_metadata, + ["android_metadata"], + { testInputs: true }, + ), website: sameJobs(PATH_SCOPES.website, ["website"], { earlyInputs: [".agents/skills/reverse-engineer-anything/**"], }), diff --git a/scripts/ci/test-selection.mjs b/scripts/ci/test-selection.mjs index 9359ab1900..504d315f32 100644 --- a/scripts/ci/test-selection.mjs +++ b/scripts/ci/test-selection.mjs @@ -9,12 +9,12 @@ import { SHARED_BOUNDARY } from "./scopes.mjs"; // Selection falls back to the complete suite whenever it cannot decide. /** More selected files than this run the complete suite instead. */ -const SELECTION_CAP = 150; +const SELECTION_CAP = 400; // Selected files per test job, and the most jobs one selection uses. Each job // runs one `--shard` of the selection; the CI sequencer balances the shards by // main's recorded per-file durations when that history is available. -const FILES_PER_SHARD = 40; -const MAX_SHARDS = 4; +const FILES_PER_SHARD = 50; +const MAX_SHARDS = 8; // Leaf adapters whose tests also reach their sources through the built CLI, // MCP server or a fake provider process rather than an import. An adapter @@ -51,13 +51,33 @@ const isPresentationOwner = (path) => // paths (package subpaths such as `ajv/dist/2020.js`, fixture apps) are not. const LAUNCHES_BUILT_RUNTIME = /scripts\/rea\.mjs|\bdist\/(?:main|cli)\.js|["'`]dist["'`]\s*,\s*["'`](?:main|cli)\.js/u; -// Inputs that only a non-test check reads: declarations (typecheck), guard -// baselines (architecture guards, static job) and the advertised-schema -// summaries (build job). They select no tests. +// Inputs that only a non-test check reads: declarations (typecheck), guard and +// layer baselines (architecture/boundary guards, static job), git-blame ignore +// revisions, and the advertised-schema summaries (build job). Any test that +// directly names one by path still runs via `names`. const isCheckerInput = (path) => /\.d\.m?ts$/u.test(path) || + path === "scripts/lib/module-layers.baseline.json" || + path === ".git-blame-ignore-revs" || + path === "scripts/run-python-tests.py" || + /^tests\/support\/analysis-efficiency\/.+\.py$/u.test(path) || /^scripts\/verify\/baselines\/[^/]+\.json$/u.test(path) || /^scripts\/verify\/advertised-schemas\/tools\/[^/]+\.json$/u.test(path); +// Non-module fixtures with a single known owning test suite. +const EXPLICIT_FIXTURE_OWNERS = [ + [ + "tests/fixtures/cli-parity/*.txt", + "tests/boundary/cli/cliParityGoldens.test.ts", + ], + [ + "tests/boundary/cli/cliContractSchemaParity.uncomparable.json", + "tests/boundary/cli/cliContractSchemaParity.test.ts", + ], +]; +const explicitFixtureOwners = (path) => + EXPLICIT_FIXTURE_OWNERS.filter(([pattern]) => matchesGlob(path, pattern)).map( + ([, owner]) => owner, + ); // Verifier scripts are imported by their tests like source modules. const isVerifier = (path) => /^scripts\/(verify-[^/]+|verify\/.+)\.mjs$/u.test(path); @@ -80,24 +100,40 @@ const isHelper = (path) => /^src\/.+\.fixture[^/]*\.ts$/u.test(path) || isVerifier(path) || isScriptLibrary(path)); +const STATIC_CHECKER_CONFIG = + /^(?:tsconfig(?:\..+)?\.json|\.oxlintrc\.json|\.oxfmtrc\.json|\.jscpd\.json|knip\.json|stryker(?:\..+)?\.config\.mjs)$/u; // Files outside src/ and tests/ that can run or load a shared script module: -// other scripts, root configuration, package metadata, hooks and workflows. +// other scripts, runtime/test configuration, package metadata, hooks and workflows. const isScriptConsumer = (path) => !isHelper(path) && !isCheckerInput(path) && + !STATIC_CHECKER_CONFIG.test(path) && + !path.startsWith(".github/ISSUE_TEMPLATE/") && (/^scripts\/.+\.(?:m?[jt]s|cjs)$/u.test(path) || (/^[^/]+\.(?:m?[jt]s|cjs|json)$/u.test(path) && path !== "package-lock.json") || /^\.github\/.+\.ya?ml$/u.test(path) || /^\.husky\/[^/]+$/u.test(path)); -// Prose that tests may read by path. Other non-module inputs (fixtures, data, -// scripts, workflows, package metadata) are not traced, so they run everything. -const isProse = (path) => path.endsWith(".md") && !/^(src|tests)\//u.test(path); +const DOCS_CODE = /\.(?:[cm]?[jt]s|[jt]sx)$/u; +// Prose and documentation/license metadata that tests may read by path. Other +// non-module inputs (unmapped fixtures, data, scripts, workflows, package +// metadata) are not traced, so they run everything. +const isProse = (path) => + path === "LICENSE" || + path.startsWith(".github/ISSUE_TEMPLATE/") || + (path.endsWith(".md") && !/^(src|tests)\//u.test(path)) || + (path.startsWith("docs/") && + !DOCS_CODE.test(path) && + !path.endsWith("/package.json")) || + path.startsWith("licenses/"); /** Why a changed path cannot be traced to tests, if it cannot. */ const untracedInput = (path) => - SHARED_BOUNDARY.test(path) + !isTest(path) && SHARED_BOUNDARY.test(path) ? `Shared boundary: ${path}` - : !isModule(path) && !isProse(path) && !isCheckerInput(path) + : !isModule(path) && + !isProse(path) && + !isCheckerInput(path) && + explicitFixtureOwners(path).length === 0 ? `Input is not traced to tests: ${path}` : undefined; @@ -145,7 +181,14 @@ export function traceAffectedTests({ files, snapshot }) { const changedModules = files.filter( (path) => isModule(path) && !isTest(path), ); - for (const path of files) if (isTest(path)) select(path, "changed"); + for (const path of files) { + if (isTest(path)) select(path, "changed"); + for (const owner of explicitFixtureOwners(path)) { + if (!trackedSet.has(owner)) + untraced.push(`Missing test owner ${owner} for ${path}`); + else select(owner, "owner"); + } + } // A seed and every helper that imports it, directly or through another. const throughHelpers = (seed) => { @@ -196,7 +239,15 @@ export function traceAffectedTests({ files, snapshot }) { } // A shared script module that another script, configuration file, hook or - // workflow imports or names by file name can shape every run. + // workflow imports or names as a standalone file name can shape every run. + // Require a token boundary before the basename so `verify-module-boundaries.mjs` + // in package.json does not falsely match `scripts/lib/module-boundaries.mjs`. + const namesScriptBasename = (source, libraryPath) => { + const escaped = posix + .basename(libraryPath) + .replace(/[.*+?^${}()|[\]\\]/gu, "\\$&"); + return new RegExp(`(? targets.has(path) || source.includes(posix.basename(path)), + (path) => + targets.has(path) || + (consumer !== "scripts/ci/scopes.mjs" && + namesScriptBasename(source, path)), ); if (library !== undefined) untraced.push( @@ -336,15 +390,14 @@ export function parseTestSelection(value) { return plan; } -// Fast feedback uses one source group and twelve built groups. Shards are -// balanced by recorded duration, but hosted runner speed varies, so the -// slowest shard sets the critical path: at eight shards it ran 24% above the -// mean shard. Smaller shards bound that excess at a few percent more runner -// time. Full qualification retains six complete shards with coverage. A -// selection runs its own shard count. +// Fast feedback uses one source group and twelve built groups; full +// qualification uses twelve complete shards with coverage. Shards are balanced +// by recorded duration, and twelve shards halve the full-qualification critical +// path (~14.5m -> ~7.5m) while bounding runner-speed variance at negligible +// extra setup (~22s/shard). A selection runs its own shard count. const shardsOf = (count) => Array.from({ length: count }, (_, index) => `${index + 1}/${count}`); /** The `test-shard` matrix labels for a tier and its test selection. */ -export const testShards = (tier, plan) => - shardsOf(plan.mode === "selected" ? plan.shards : tier === "full" ? 6 : 12); +export const testShards = (_tier, plan) => + shardsOf(plan.mode === "selected" ? plan.shards : 12); diff --git a/tests/boundary/filesystem/ciCachePolicy.test.ts b/tests/boundary/filesystem/ciCachePolicy.test.ts index 8d95eb49de..2d04746639 100644 --- a/tests/boundary/filesystem/ciCachePolicy.test.ts +++ b/tests/boundary/filesystem/ciCachePolicy.test.ts @@ -64,9 +64,7 @@ const assertRetentionPolicy = (expression: string): void => { repositoryName === currentRepository && branch === defaultBranch && ["push", "schedule", "workflow_dispatch"].includes(event) && - ["success", "failure", "timed_out", "cancelled"].includes( - conclusion, - ); + ["success", "failure", "timed_out"].includes(conclusion); expect( evaluatePolicy(expression, { github: { @@ -556,6 +554,7 @@ it("deletes superseded caches only from the trusted follow-up workflow", async ( workflow_run: z.object({ workflows: z.array(z.literal("CI")), types: z.array(z.literal("completed")), + branches: z.array(z.literal("main")), }), workflow_call: z.unknown().optional(), pull_request: z.unknown().optional(), @@ -605,7 +604,6 @@ it("deletes superseded caches only from the trusted follow-up workflow", async ( "persist-credentials": false, }); expect(checkouts).toHaveLength(2); - expect(steps.some((step) => step.run === "npm ci")).toBe(false); const setupNode = steps.find( (step) => step.uses === @@ -635,6 +633,9 @@ it("deletes superseded caches only from the trusted follow-up workflow", async ( expect(retentionEnv.CACHE_PORTABLE_PREFIX).toBe( "rea-portable-build-v2-${{ runner.os }}-${{ runner.arch }}-", ); + expect(retentionEnv.CACHE_WINDOWS_BUILD_PREFIX).toBe( + "rea-windows-build-v1-Windows-X64-", + ); expect(retentionEnv.DEDUPE_NPM_DOWNLOADS).toBe("true"); expect(retentionEnv.GH_TOKEN).toBe("${{ github.token }}"); @@ -667,6 +668,14 @@ it("deletes superseded caches only from the trusted follow-up workflow", async ( expect(retentionEnv.CACHE_KNIP_PREFIX).toBe(knipPrefix); expect(retentionEnv.CACHE_KNIP_KEEP_KEY).toBe(targetSourceKey("knip-cache")); expect(publishedKey("knip-cache").startsWith(knipPrefix)).toBe(true); + const windowsNativeSteps = + ci?.workflow.jobs?.["windows-native-build"]?.steps ?? []; + const windowsNativeKey = windowsNativeSteps.find( + (step) => step.id === "windows-native-key", + ); + expect(String(windowsNativeKey?.run)).toContain("ImageOS"); + expect(String(windowsNativeKey?.run)).toContain("ImageVersion"); + expect(String(windowsNativeKey?.run)).toContain("date -u +%Y-%m-%d"); }); it("publishes each hosted npm download cache from one writer", async () => { diff --git a/tests/boundary/filesystem/ciCacheRetention.test.ts b/tests/boundary/filesystem/ciCacheRetention.test.ts index 571d05bd22..53a4eaac2d 100644 --- a/tests/boundary/filesystem/ciCacheRetention.test.ts +++ b/tests/boundary/filesystem/ciCacheRetention.test.ts @@ -89,8 +89,9 @@ describe("Turbo snapshot retention", () => { ).toEqual([4]); }); - it("retains the newest portable snapshots without a caller keep key", () => { + it("retains the newest portable and Windows build snapshots without a caller keep key", () => { const portable = "rea-portable-build-v2-Linux-X64-"; + const windowsBuild = "rea-windows-build-v1-Windows-X64-"; const deletions = selectPrefixDeletions( [ cache(7, `${portable}one`, "2026-10-20T01:00:00Z"), @@ -100,6 +101,16 @@ describe("Turbo snapshot retention", () => { { prefix: portable }, ); expect(deletions.map((entry) => entry.id)).toEqual([7]); + expect( + selectPrefixDeletions( + [ + cache(10, `${windowsBuild}one`, "2026-10-20T01:00:00Z"), + cache(11, `${windowsBuild}two`, "2026-10-20T02:00:00Z"), + cache(12, `${windowsBuild}three`, "2026-10-20T03:00:00Z"), + ], + { prefix: windowsBuild }, + ).map((entry) => entry.id), + ).toEqual([10]); }); it("refuses prefixes and lists that can reach download caches", () => { diff --git a/tests/boundary/filesystem/ciChangeScope.test.ts b/tests/boundary/filesystem/ciChangeScope.test.ts index e0ffbdd643..bc30f0753b 100644 --- a/tests/boundary/filesystem/ciChangeScope.test.ts +++ b/tests/boundary/filesystem/ciChangeScope.test.ts @@ -258,6 +258,14 @@ it.each([ ["src/ports/apkSignature/ApkSignaturePort.ts", "apk_signature"], ["src/ports/javascript/JavaScriptRecoveryPort.ts", "javascript_recovery"], ["src/ports/WebRuntimePort.ts", "web_runtime"], + ["src/go/GoBinaryProvider.ts", "go_binary"], + ["scripts/verify-go-binary.mjs", "go_binary"], + ["bridge/android/ReaJadxBridge.java", "android_metadata"], + ["src/android/JadxRelease.ts", "android_metadata"], + [ + "tests/conformance/android/ReaJadxMetadataRegression.java", + "android_metadata", + ], ["src/ports/WebSourceLocationPorts.ts", "web_source_map"], ["src/ports/captures/PacketCapturePort.ts", "packet_capture"], ["src/ports/signals/SignalRecordingPort.ts", "signal_recording"], @@ -470,7 +478,7 @@ it("selects the tests a pull request affects and runs everything when it cannot await runPlanScript(directory, base, head, "pull_request", true), ).toMatchObject({ testSelection: { version: 3, mode: "full" }, - testShards: ["1/6", "2/6", "3/6", "4/6", "5/6", "6/6"], + testShards: expect.arrayContaining(["12/12"]), }); expect(await classify(directory, base, head, "push")).toMatchObject({ testSelection: { mode: "full" }, @@ -506,14 +514,14 @@ it("selects the tests a pull request affects and runs everything when it cannot }); // Above the cap, selection gives way to the complete suite. const many = Array.from( - { length: 151 }, + { length: 401 }, (_, index) => `tests/composition/many${String(index)}.test.ts`, ); const wide = await change(directory, many); expect(await classify(directory, isolated, wide)).toMatchObject({ testSelection: { mode: "full", - reason: "151 affected test files exceed the 150-file selection cap", + reason: "401 affected test files exceed the 400-file selection cap", }, }); const capped = await change(directory, many.slice(0, 81), "// edited\n"); @@ -523,6 +531,33 @@ it("selects the tests a pull request affects and runs everything when it cannot }); }); +it("routes license inventories to docs checks and keeps blame-ignore edits formatting-only", async () => { + expect(await classifyPaths([".git-blame-ignore-revs"])).toMatchObject({ + code: false, + docs: false, + runtime: false, + package: false, + }); + expect(await classifyPaths(["licenses/npm-dependencies.txt"])).toMatchObject({ + code: true, + docs: true, + runtime: false, + package: false, + }); + expect(await classifyPaths(["THIRD_PARTY_NOTICES.md"])).toMatchObject({ + code: true, + docs: true, + runtime: false, + package: false, + }); + expect(await classifyPaths(["LICENSE"])).toMatchObject({ + code: true, + docs: true, + runtime: false, + package: false, + }); +}); + it("passes the workflow environment through the entrypoint unchanged", async () => { const { directory, base } = await fixture(); const head = await change(directory, ["src/evm/Example.ts"]); diff --git a/tests/boundary/filesystem/ciCodeqlScope.test.ts b/tests/boundary/filesystem/ciCodeqlScope.test.ts index ddc15c85c3..a4e84be670 100644 --- a/tests/boundary/filesystem/ciCodeqlScope.test.ts +++ b/tests/boundary/filesystem/ciCodeqlScope.test.ts @@ -3,6 +3,10 @@ import { matchesGlob } from "node:path"; import { expect, it } from "vitest"; import { parse } from "yaml"; import { z } from "zod"; +import { + CODEQL_LANGUAGES, + planCodeqlLanguages, +} from "../../../scripts/ci/codeql-scope.mjs"; const github = new URL("../../../.github/", import.meta.url); const configuration = z @@ -16,14 +20,77 @@ const configuration = z parse(await readFile(new URL("codeql/production.yml", github), "utf8")), ); -const excluded = (path: string): boolean => - configuration["paths-ignore"].some( +const triggerPathsSchema = z.object({ + branches: z.array(z.string()), + "paths-ignore": z.array(z.string()), +}); +const workflowDocument = z + .object({ + on: z.object({ + push: triggerPathsSchema, + pull_request: triggerPathsSchema, + schedule: z.array(z.object({ cron: z.string() })), + workflow_dispatch: z.unknown(), + }), + permissions: z.object({ contents: z.literal("read") }), + concurrency: z.object({ + "cancel-in-progress": z.string(), + }), + jobs: z.object({ + scope: z.object({ + if: z.string(), + outputs: z.object({ + run: z.string(), + languages: z.string(), + }), + steps: z.array( + z.object({ + id: z.string().optional(), + uses: z.string().optional(), + run: z.string().optional(), + with: z.record(z.string(), z.unknown()).optional(), + }), + ), + }), + analyze: z.object({ + needs: z.literal("scope"), + if: z.string(), + permissions: z.object({ + contents: z.literal("read"), + "security-events": z.literal("write"), + }), + strategy: z.object({ + "max-parallel": z.literal(1), + matrix: z.object({ language: z.string() }), + }), + steps: z.array( + z.object({ + uses: z.string(), + with: z.record(z.string(), z.unknown()).optional(), + }), + ), + }), + }), + }) + .parse( + parse(await readFile(new URL("workflows/codeql.yml", github), "utf8")), + ); + +const matchesPatterns = (path: string, patterns: readonly string[]): boolean => + patterns.some( (pattern) => path === pattern || path.startsWith(`${pattern}/`) || matchesGlob(path, pattern), ); +const excluded = (path: string): boolean => + matchesPatterns(path, configuration["paths-ignore"]); + +const triggerIgnored = (path: string): boolean => + matchesPatterns(path, workflowDocument.on.push["paths-ignore"]) || + matchesPatterns(path, workflowDocument.on.pull_request["paths-ignore"]); + it.each([ "tests/fixtures/fakeHopper.mjs", "tests/conformance/ghidra/riscv-checks.mjs", @@ -37,6 +104,7 @@ it.each([ "bridge/example_test.py", ])("excludes test-only input %s", (path) => { expect(excluded(path)).toBe(true); + expect(triggerIgnored(path)).toBe(true); }); it.each([ @@ -60,6 +128,7 @@ it.each([ ".github/actions/publish-mcp/action.yml", ])("retains production or release input %s", (path) => { expect(excluded(path)).toBe(false); + expect(triggerIgnored(path)).toBe(false); }); it("retains all packaged entrypoints and the existing local-input threat model", async () => { @@ -74,6 +143,12 @@ it("retains all packaged entrypoints and the existing local-input threat model", ), ); expect(manifest.files.filter(excluded)).toEqual([]); + // Non-prose packaged files remain active triggers on main and pull requests. + expect( + manifest.files + .filter((path) => !path.endsWith(".md")) + .filter(triggerIgnored), + ).toEqual([]); expect(configuration["threat-models"]).toEqual(["local"]); // A narrow allowlist would silently drop new production roots. Explicitly // traverse the hidden configuration directory without excluding the root. @@ -81,46 +156,23 @@ it("retains all packaged entrypoints and the existing local-input threat model", expect(configuration.paths).toContain("docs/.vitepress"); }); -it("uses the scope in a bounded workflow gated on administrator activation", async () => { - const workflow = z - .object({ - permissions: z.object({ contents: z.literal("read") }), - concurrency: z.object({ - "cancel-in-progress": z.string(), - }), - jobs: z.object({ - analyze: z.object({ - if: z.string(), - permissions: z.object({ - contents: z.literal("read"), - "security-events": z.literal("write"), - }), - strategy: z.object({ - "max-parallel": z.literal(1), - matrix: z.object({ language: z.array(z.string()) }), - }), - steps: z.array( - z.object({ - uses: z.string(), - with: z.record(z.string(), z.unknown()).optional(), - }), - ), - }), - }), - }) - .parse( - parse(await readFile(new URL("workflows/codeql.yml", github), "utf8")), - ); - expect(workflow.concurrency["cancel-in-progress"]).toBe( +it("uses the scope in a bounded workflow gated on administrator activation", () => { + expect(workflowDocument.concurrency["cancel-in-progress"]).toBe( "${{ github.event_name == 'pull_request' }}", ); - const job = workflow.jobs.analyze; - expect(job.if).toBe("vars.REA_CODEQL_ADVANCED == 'true'"); - expect(job.strategy.matrix.language).toEqual([ - "javascript-typescript", - "python", - "actions", - ]); + expect(workflowDocument.on.push["paths-ignore"]).toEqual( + workflowDocument.on.pull_request["paths-ignore"], + ); + expect(workflowDocument.jobs.scope.if).toBe( + "vars.REA_CODEQL_ADVANCED == 'true'", + ); + const job = workflowDocument.jobs.analyze; + expect(job.if).toBe( + "vars.REA_CODEQL_ADVANCED == 'true' && needs.scope.outputs.run == 'true'", + ); + expect(job.strategy.matrix.language).toBe( + "${{ fromJSON(needs.scope.outputs.languages) }}", + ); const init = job.steps.find((step) => step.uses.startsWith("github/codeql-action/init@"), ); @@ -140,3 +192,71 @@ it("uses the scope in a bounded workflow gated on administrator activation", asy }); expect(analyze?.with).not.toHaveProperty("upload"); }); + +const SHA_BASE = "a".repeat(40); +const SHA_HEAD = "b".repeat(40); +const selectPr = (files: readonly string[]) => + planCodeqlLanguages( + { event: "pull_request", base: SHA_BASE, head: SHA_HEAD }, + { mergeBase: (base) => base, changedFiles: () => files }, + ); + +it("selects affected CodeQL languages on pull requests and fails closed on shared or unknown inputs", () => { + for (const event of ["push", "schedule", "workflow_dispatch"]) { + expect( + planCodeqlLanguages( + { event, base: undefined, head: SHA_HEAD }, + { + mergeBase: () => { + throw new Error("Non-PR events must not diff"); + }, + changedFiles: () => [], + }, + ), + ).toMatchObject({ run: true, languages: CODEQL_LANGUAGES }); + } + expect( + selectPr([ + "tests/boundary/filesystem/ciCodeqlScope.test.ts", + "src/process/ExecFileOutput.test.ts", + "docs/development/testing.md", + "bridge/ghidra/ReaGhidraBridge.java", + ]), + ).toMatchObject({ run: false, languages: [] }); + expect( + selectPr(["src/process/ExecFileOutput.ts", "docs/.vitepress/config.ts"]), + ).toMatchObject({ + run: true, + languages: ["javascript-typescript"], + }); + expect(selectPr(["bridge/pwndbg/launch.py"])).toMatchObject({ + run: true, + languages: ["python"], + }); + expect( + selectPr([ + ".github/workflows/release.yml", + ".github/actions/publish-mcp/action.yml", + ]), + ).toMatchObject({ + run: true, + languages: ["actions"], + }); + for (const broad of [ + ".github/codeql/production.yml", + ".github/workflows/codeql.yml", + "install.sh", + "scripts/unknown-entrypoint", + ]) { + expect(selectPr([broad]), broad).toMatchObject({ + run: true, + languages: CODEQL_LANGUAGES, + }); + } + expect( + planCodeqlLanguages( + { event: "pull_request", base: "not-a-sha", head: SHA_HEAD }, + { mergeBase: (base) => base, changedFiles: () => [] }, + ), + ).toMatchObject({ run: true, languages: CODEQL_LANGUAGES }); +}); diff --git a/tests/boundary/filesystem/ciPrHealth.test.ts b/tests/boundary/filesystem/ciPrHealth.test.ts index ca1ace52aa..6571325e9e 100644 --- a/tests/boundary/filesystem/ciPrHealth.test.ts +++ b/tests/boundary/filesystem/ciPrHealth.test.ts @@ -354,8 +354,9 @@ it.each([false, true])( now: observedAt, apiRequest, }); - expect(report.scan.workflowRunPages).toBe(5); + expect(report.scan.workflowRunPages).toBe(10); expect(report.scan.workflowRunInventoryComplete).toBe(false); + expect(report.lookbackHours).toBe(12); const firstPull = report.pullRequests[0]; if (firstPull === undefined) throw new Error("fixture PR was not scanned"); expect(firstPull.ci).toMatchObject({ @@ -598,6 +599,32 @@ it("keeps a proven newest run even when two older matched runs have equal timest expect(result.runs).toHaveLength(3); }); +it("queries workflow runs with full ISO timestamp precision across the configured hour window", async () => { + const endpoints: string[] = []; + const midDay = Date.parse("2026-10-14T18:17:00Z"); + const report = await collectPrCiHealth({ + repository: "owner/rea", + repositoryId: 11, + defaultBranch: "main", + now: midDay, + apiRequest: async (_method, endpoint) => { + endpoints.push(endpoint); + if (endpoint.includes("/pulls?")) + return [ + rawPull(91, { mergeable: true, created_at: "2026-10-14T12:00:00Z" }), + rawPull(92, { mergeable: true, created_at: "2026-10-14T05:00:00Z" }), + ]; + return { workflow_runs: [] }; + }, + }); + expect(report.lookbackHours).toBe(12); + expect(report.scan.cutoff).toBe("2026-10-14T06:17:00.000Z"); + expect(report.pullRequests.map((pull) => pull.number)).toEqual([91]); + expect(endpoints).toContain( + "repos/owner/rea/actions/workflows/ci.yml/runs?event=pull_request&created=%3E%3D2026-10-14T06%3A17%3A00.000Z&per_page=100&page=1", + ); +}); + function run(overrides: Partial = {}): PrCiWorkflowRun { return { id: 77, @@ -617,7 +644,7 @@ function rawPull(number: number, overrides: Record = {}) { return { number, draft: false, - created_at: "2026-10-13T00:00:00Z", + created_at: "2026-10-13T18:00:00Z", html_url: `https://github.com/owner/rea/pull/${String(number)}`, base: { ref: "main", repo: { id: 11 } }, head: { diff --git a/tests/boundary/filesystem/ciRequiredLanes.test.ts b/tests/boundary/filesystem/ciRequiredLanes.test.ts index 619d62bc0f..ddcafd1193 100644 --- a/tests/boundary/filesystem/ciRequiredLanes.test.ts +++ b/tests/boundary/filesystem/ciRequiredLanes.test.ts @@ -375,6 +375,7 @@ async function assertProviderCacheCallers( "sqlite-runtime", "nativeaot", "ghidra_bridge", + "android_metadata", "website", ]) expect(jobs[name]?.secrets, name).toBeUndefined(); @@ -430,6 +431,7 @@ it("binds provider build reuse to the current run while keeping standalone dispa "agentenv", "packet_capture", "signal_recording", + "go_binary", ]; const buildTerms = z .string() @@ -928,8 +930,10 @@ it("binds each canonical lane to its explicit fast and full workflow jobs", asyn "web_runtime", "web_source_map", "agentenv", + "go_binary", "nativeaot", "ghidra_bridge", + "android_metadata", "website", ].map((name) => [name, { fast: [name], full: [name] }]), ), diff --git a/tests/boundary/mcp/toolSchemaValidity.test.ts b/tests/boundary/mcp/toolSchemaValidity.test.ts index 976824652d..d78a9a43cd 100644 --- a/tests/boundary/mcp/toolSchemaValidity.test.ts +++ b/tests/boundary/mcp/toolSchemaValidity.test.ts @@ -41,7 +41,7 @@ interface ToolSchemas { function expectStrictInputSchemaParity( tools: readonly ToolSchemas[], - ajv: Ajv2020, + validators: ReadonlyMap>, ): void { const advertised = new Map(tools.map((tool) => [tool.name, tool])); const names = [ @@ -62,15 +62,51 @@ function expectStrictInputSchemaParity( ({ name: toolName }) => toolName === name, ); const tool = advertised.get(name); + const validate = validators.get(name); const example = contract?.examples[0]; - if (contract === undefined || tool === undefined || example === undefined) + if ( + contract === undefined || + tool === undefined || + validate === undefined || + example === undefined + ) throw new Error(`${name} did not have an advertised example`); const malformed = { ...example.input, __unexpected_root_key__: true }; expect(contract.inputSchema.safeParse(malformed).success, name).toBe(false); - expect(ajv.compile(tool.inputSchema)(malformed), name).toBe(false); + expect(validate(malformed), name).toBe(false); } } +let cachedCanonicalTools: + | Promise>["tools"]> + | undefined; + +function listCanonicalAdvertisedTools() { + cachedCanonicalTools ??= (async () => { + const server = new McpServer({ name: "schema-validation", version: "0" }); + const client = new Client( + { name: "schema-validation", version: "0" }, + { jsonSchemaValidator: createMcpSchemaValidation() }, + ); + const [clientTransport, serverTransport] = + InMemoryTransport.createLinkedPair(); + for (const contract of TOOL_CONTRACTS) + server.registerTool( + contract.name, + toolRegistrationOptions(contract), + async () => ({ content: [] }), + ); + try { + await server.connect(serverTransport); + await client.connect(clientTransport); + return (await client.listTools()).tools; + } finally { + await Promise.allSettled([client.close(), server.close()]); + } + })(); + return cachedCanonicalTools; +} + function expectKnownAuthorityHints(tools: readonly ToolSchemas[]): void { const advertised = new Map(tools.map((tool) => [tool.name, tool])); const expected = { @@ -467,188 +503,159 @@ describe("MCP JSON Schema validity", () => { }); it("advertises valid input and output schemas for every canonical tool", async () => { - const server = new McpServer({ name: "schema-validation", version: "0" }); - const client = new Client( - { name: "schema-validation", version: "0" }, - { jsonSchemaValidator: createMcpSchemaValidation() }, + const tools = await listCanonicalAdvertisedTools(); + expect(tools.map(({ name }) => name).sort()).toEqual( + TOOL_CONTRACTS.map(({ name }) => name).sort(), ); - const [clientTransport, serverTransport] = - InMemoryTransport.createLinkedPair(); - for (const contract of TOOL_CONTRACTS) - server.registerTool( - contract.name, - toolRegistrationOptions(contract), - async () => ({ - content: [], - }), + expect(advertisedSchemaErrors(tools)).toEqual([]); + expect(localInputReferences(tools)).toEqual([]); + expect(recursiveReferences(tools)).toEqual([]); + expectKnownAuthorityHints(tools); + const byName = new Map(tools.map((tool) => [tool.name, tool])); + expectManagedLocalSignatureSchema(byName.get("inspect_managed_members")); + const comparison = byName.get("compare_web_captures")?.inputSchema; + for (const side of ["before", "after"]) + expect(comparison).toHaveProperty( + ["properties", side, "properties", "inspection"], + { + type: "object", + description: expect.stringContaining( + "complete canonical observation", + ), + }, ); - try { - await server.connect(serverTransport); - await client.connect(clientTransport); - const { tools } = await client.listTools(); - expect(tools.map(({ name }) => name).sort()).toEqual( - TOOL_CONTRACTS.map(({ name }) => name).sort(), + for (const contract of TOOL_CONTRACTS) { + const tool = byName.get(contract.name); + expect(tool?.title?.trim(), contract.name).toBeTruthy(); + expect(tool?.description?.trim(), contract.name).toBeTruthy(); + const examples = advertisedInputExamples(contract); + expect(tool?.inputSchema.examples, contract.name).toEqual( + examples.length === 0 ? undefined : examples, ); - expect(advertisedSchemaErrors(tools)).toEqual([]); - expect(localInputReferences(tools)).toEqual([]); - expect(recursiveReferences(tools)).toEqual([]); - expectKnownAuthorityHints(tools); - const byName = new Map(tools.map((tool) => [tool.name, tool])); - expectManagedLocalSignatureSchema(byName.get("inspect_managed_members")); - const comparison = byName.get("compare_web_captures")?.inputSchema; - for (const side of ["before", "after"]) - expect(comparison).toHaveProperty( - ["properties", side, "properties", "inspection"], - { - type: "object", - description: expect.stringContaining( - "complete canonical observation", - ), - }, - ); - for (const contract of TOOL_CONTRACTS) { - const tool = byName.get(contract.name); - expect(tool?.title?.trim(), contract.name).toBeTruthy(); - expect(tool?.description?.trim(), contract.name).toBeTruthy(); - const examples = advertisedInputExamples(contract); - expect(tool?.inputSchema.examples, contract.name).toEqual( - examples.length === 0 ? undefined : examples, - ); - for (const example of contract.examples) - expect( - contract.inputSchema.safeParse(example.input).success, - `${contract.name}: ${example.title}`, - ).toBe(true); - expectMeaningfulPropertyDescriptions(tool?.inputSchema, contract.name); - } - } finally { - await Promise.allSettled([client.close(), server.close()]); + for (const example of contract.examples) + expect( + contract.inputSchema.safeParse(example.input).success, + `${contract.name}: ${example.title}`, + ).toBe(true); + expectMeaningfulPropertyDescriptions(tool?.inputSchema, contract.name); } }); }); describe("MCP root input schemas", () => { it("advertises object roots while enforcing union requirements at runtime", async () => { - const server = new McpServer({ name: "schema-constraints", version: "0" }); - const client = new Client( - { name: "schema-constraints", version: "0" }, - { jsonSchemaValidator: createMcpSchemaValidation() }, + const tools = await listCanonicalAdvertisedTools(); + const advertised = new Map(tools.map((tool) => [tool.name, tool])); + const ajv = new Ajv2020({ + strict: false, + validateFormats: false, + inlineRefs: false, + code: { optimize: false }, + }); + const compiledValidators = new Map< + string, + ReturnType + >(); + const graphContract = TOOL_CONTRACTS.find( + ({ name }) => name === "project_managed_application_graph", ); - const [clientTransport, serverTransport] = - InMemoryTransport.createLinkedPair(); - for (const contract of TOOL_CONTRACTS) - server.registerTool( - contract.name, - toolRegistrationOptions(contract), - async () => ({ content: [] }), - ); - try { - await server.connect(serverTransport); - await client.connect(clientTransport); - const { tools } = await client.listTools(); - const advertised = new Map(tools.map((tool) => [tool.name, tool])); - const ajv = new Ajv2020({ strict: false, validateFormats: false }); - const graphContract = TOOL_CONTRACTS.find( - ({ name }) => name === "project_managed_application_graph", - ); - const graphTool = advertised.get("project_managed_application_graph"); - if (graphContract === undefined || graphTool === undefined) - throw new Error("Managed application graph tool was not advertised"); - expect(graphContract.inputSchema.safeParse({}).success).toBe(false); - expect(graphTool.inputSchema).toMatchObject({ type: "object" }); - expect(ajv.compile(graphTool.inputSchema)({})).toBe(false); - expect(ajv.compile(graphTool.inputSchema)({ unrelated: true })).toBe( - false, - ); - expect(graphTool.inputSchema.description).toContain( - "Provide a complete input group", - ); - expect(graphTool.inputSchema.description).not.toContain("exactly one"); + const graphTool = advertised.get("project_managed_application_graph"); + if (graphContract === undefined || graphTool === undefined) + throw new Error("Managed application graph tool was not advertised"); + expect(graphContract.inputSchema.safeParse({}).success).toBe(false); + expect(graphTool.inputSchema).toMatchObject({ type: "object" }); + const validateGraph = ajv.compile(graphTool.inputSchema); + compiledValidators.set("project_managed_application_graph", validateGraph); + expect(validateGraph({})).toBe(false); + expect(validateGraph({ unrelated: true })).toBe(false); + expect(graphTool.inputSchema.description).toContain( + "Provide a complete input group", + ); + expect(graphTool.inputSchema.description).not.toContain("exactly one"); - for (const contract of TOOL_CONTRACTS) { - const inputSchema = advertised.get(contract.name)!.inputSchema; - expect(inputSchema.type, contract.name).toBe("object"); - expect(inputSchema.properties, contract.name).toBeDefined(); - // Anthropic tool input schemas reject root combinators. - for (const combinator of ["anyOf", "oneOf", "allOf"]) - expect(inputSchema, contract.name).not.toHaveProperty(combinator); - const validate = ajv.compile(inputSchema); - for (const example of contract.examples) - expect( - validate(example.input), - `${contract.name}: ${example.title}`, - ).toBe(true); - } - - expectStrictInputSchemaParity(tools, ajv); - - for (const name of ["inspect_web_page", "analyze_web_bundle"]) { - const contract = TOOL_CONTRACTS.find((tool) => tool.name === name); - const tool = advertised.get(name); - if (contract === undefined || tool === undefined) - throw new Error(`${name} was not advertised`); - const validate = ajv.compile( - advertiseTool( - { inputSchema: tool.inputSchema }, - advertisedToolPresentation("off"), - ).inputSchema ?? {}, - ); - const base = { - cdp_endpoint: "http://127.0.0.1:9222", - target_id: "target", - }; - for (const flag of [undefined, false, true, "false", null]) { - const input = - flag === undefined - ? base - : { ...base, include_script_sources: flag }; - const accepted = - flag === undefined || - flag === true || - (name === "inspect_web_page" && flag === false); - const parsed = contract.inputSchema.safeParse(input); - expect(parsed.success, `${name}: ${String(flag)}`).toBe(accepted); - // The advertised dialect relaxes Boolean literals; the canonical parser - // still rejects false when this operation requires script sources. - expect(validate(input), `${name}: ${String(flag)}`).toBe( - flag === undefined || typeof flag === "boolean", - ); - if (parsed.success && flag === undefined) - expect(parsed.data).toMatchObject({ - include_script_sources: name === "analyze_web_bundle", - }); - } - } - - const nativeObservation = TOOL_CONTRACTS.find( - ({ name }) => name === "observe_native_ui", - ); - const nativeScenario = TOOL_CONTRACTS.find( - ({ name }) => name === "capture_native_ui_scenario", - ); - const nativeObservationTool = advertised.get("observe_native_ui"); - const nativeScenarioTool = advertised.get("capture_native_ui_scenario"); - if ( - nativeObservation === undefined || - nativeScenario === undefined || - nativeObservationTool === undefined || - nativeScenarioTool === undefined - ) - throw new Error("Native UI tools were not registered"); - const target = { pid: 123, window_id: 456 }; - const scenario = { - ...target, - steps: [{ kind: "wait", milliseconds: 100 }], - }; - const observationResult = nativeObservation.inputSchema.safeParse(target); - const scenarioResult = nativeScenario.inputSchema.safeParse(scenario); - expect(observationResult.success).toBe(true); - expect(scenarioResult.success).toBe(true); - expect(ajv.compile(nativeObservationTool.inputSchema)(target)).toBe(true); - expect(ajv.compile(nativeScenarioTool.inputSchema)(scenario)).toBe(true); - } finally { - await Promise.allSettled([client.close(), server.close()]); + for (const contract of TOOL_CONTRACTS) { + const inputSchema = advertised.get(contract.name)!.inputSchema; + expect(inputSchema.type, contract.name).toBe("object"); + expect(inputSchema.properties, contract.name).toBeDefined(); + // Anthropic tool input schemas reject root combinators. + for (const combinator of ["anyOf", "oneOf", "allOf"]) + expect(inputSchema, contract.name).not.toHaveProperty(combinator); + const validate = + compiledValidators.get(contract.name) ?? ajv.compile(inputSchema); + compiledValidators.set(contract.name, validate); + for (const example of contract.examples) + expect( + validate(example.input), + `${contract.name}: ${example.title}`, + ).toBe(true); } + + expectStrictInputSchemaParity(tools, compiledValidators); + + for (const name of ["inspect_web_page", "analyze_web_bundle"]) { + const contract = TOOL_CONTRACTS.find((tool) => tool.name === name); + const tool = advertised.get(name); + if (contract === undefined || tool === undefined) + throw new Error(`${name} was not advertised`); + const validate = ajv.compile( + advertiseTool( + { inputSchema: tool.inputSchema }, + advertisedToolPresentation("off"), + ).inputSchema ?? {}, + ); + const base = { + cdp_endpoint: "http://127.0.0.1:9222", + target_id: "target", + }; + for (const flag of [undefined, false, true, "false", null]) { + const input = + flag === undefined ? base : { ...base, include_script_sources: flag }; + const accepted = + flag === undefined || + flag === true || + (name === "inspect_web_page" && flag === false); + const parsed = contract.inputSchema.safeParse(input); + expect(parsed.success, `${name}: ${String(flag)}`).toBe(accepted); + // The advertised dialect relaxes Boolean literals; the canonical parser + // still rejects false when this operation requires script sources. + expect(validate(input), `${name}: ${String(flag)}`).toBe( + flag === undefined || typeof flag === "boolean", + ); + if (parsed.success && flag === undefined) + expect(parsed.data).toMatchObject({ + include_script_sources: name === "analyze_web_bundle", + }); + } + } + + const nativeObservation = TOOL_CONTRACTS.find( + ({ name }) => name === "observe_native_ui", + ); + const nativeScenario = TOOL_CONTRACTS.find( + ({ name }) => name === "capture_native_ui_scenario", + ); + const nativeObservationTool = advertised.get("observe_native_ui"); + const nativeScenarioTool = advertised.get("capture_native_ui_scenario"); + if ( + nativeObservation === undefined || + nativeScenario === undefined || + nativeObservationTool === undefined || + nativeScenarioTool === undefined + ) + throw new Error("Native UI tools were not registered"); + const target = { pid: 123, window_id: 456 }; + const scenario = { + ...target, + steps: [{ kind: "wait", milliseconds: 100 }], + }; + const observationResult = nativeObservation.inputSchema.safeParse(target); + const scenarioResult = nativeScenario.inputSchema.safeParse(scenario); + expect(observationResult.success).toBe(true); + expect(scenarioResult.success).toBe(true); + expect(compiledValidators.get("observe_native_ui")!(target)).toBe(true); + expect( + compiledValidators.get("capture_native_ui_scenario")!(scenario), + ).toBe(true); }); }); @@ -665,82 +672,55 @@ const advertiseAndEnforceItemSelectorCollectionConstraints = if (contract === undefined) throw new Error("inspect_analysis_view contract was not registered"); - const server = new McpServer({ - name: "item-selector-schema", - version: "0", - }); - server.registerTool( - contract.name, - toolRegistrationOptions(contract), - async () => ({ - content: [], - }), + const advertised = (await listCanonicalAdvertisedTools()).find( + ({ name }) => name === contract.name, ); - const client = new Client( - { name: "item-selector-schema", version: "0" }, - { jsonSchemaValidator: createMcpSchemaValidation() }, - ); - const [clientTransport, serverTransport] = - InMemoryTransport.createLinkedPair(); - try { - await server.connect(serverTransport); - await client.connect(clientTransport); - const advertised = (await client.listTools()).tools.find( - ({ name }) => name === contract.name, - ); - if (advertised === undefined) - throw new Error("inspect_analysis_view tool was not advertised"); + if (advertised === undefined) + throw new Error("inspect_analysis_view tool was not advertised"); - const validate = new Ajv2020({ - strict: false, - validateFormats: false, - }).compile(advertised.inputSchema); - const source = { - kind: "retained-evidence", - evidence_id: `ev_${"a".repeat(64)}`, - }; - const itemInput = (collection: string, selector: unknown) => ({ - source, - view: { kind: "item", collection, selector }, - }); - const nodeId = `jag_node_${"b".repeat(64)}`; - // Collection-specific selector constraints must survive JSON Schema - // conversion: the advertised schema rejects exactly what runtime - // input validation rejects. - const invalid: ReadonlyArray = [ - ["modules", { index: 0 }], - ["modules", { name: "entry" }], - ["sections", { path: "/src/entry.js" }], - ["symbols", { node_id: nodeId }], - ]; - const valid: ReadonlyArray = [ - ["modules", { path: "/src/entry.js" }], - ["modules", { node_id: nodeId }], - ["sections", { index: 0 }], - ["sections", { name: ".text" }], - ["symbols", { index: 0 }], - ["symbols", { name: "main" }], - ]; - for (const [collection, selector] of invalid) { - const input = itemInput(collection, selector); - const label = `${collection}/${Object.keys(selector as object)[0]}`; - expect(contract.inputSchema.safeParse(input).success, label).toBe( - false, - ); - expect(validate(input), `advertised schema rejects ${label}`).toBe( - false, - ); - } - for (const [collection, selector] of valid) { - const input = itemInput(collection, selector); - const label = `${collection}/${Object.keys(selector as object)[0]}`; - expect(contract.inputSchema.safeParse(input).success, label).toBe(true); - expect(validate(input), `advertised schema accepts ${label}`).toBe( - true, - ); - } - } finally { - await Promise.allSettled([client.close(), server.close()]); + const validate = new Ajv2020({ + strict: false, + validateFormats: false, + inlineRefs: false, + code: { optimize: false }, + }).compile(advertised.inputSchema); + const source = { + kind: "retained-evidence", + evidence_id: `ev_${"a".repeat(64)}`, + }; + const itemInput = (collection: string, selector: unknown) => ({ + source, + view: { kind: "item", collection, selector }, + }); + const nodeId = `jag_node_${"b".repeat(64)}`; + // Collection-specific selector constraints must survive JSON Schema + // conversion: the advertised schema rejects exactly what runtime + // input validation rejects. + const invalid: ReadonlyArray = [ + ["modules", { index: 0 }], + ["modules", { name: "entry" }], + ["sections", { path: "/src/entry.js" }], + ["symbols", { node_id: nodeId }], + ]; + const valid: ReadonlyArray = [ + ["modules", { path: "/src/entry.js" }], + ["modules", { node_id: nodeId }], + ["sections", { index: 0 }], + ["sections", { name: ".text" }], + ["symbols", { index: 0 }], + ["symbols", { name: "main" }], + ]; + for (const [collection, selector] of invalid) { + const input = itemInput(collection, selector); + const label = `${collection}/${Object.keys(selector as object)[0]}`; + expect(contract.inputSchema.safeParse(input).success, label).toBe(false); + expect(validate(input), `advertised schema rejects ${label}`).toBe(false); + } + for (const [collection, selector] of valid) { + const input = itemInput(collection, selector); + const label = `${collection}/${Object.keys(selector as object)[0]}`; + expect(contract.inputSchema.safeParse(input).success, label).toBe(true); + expect(validate(input), `advertised schema accepts ${label}`).toBe(true); } }; @@ -751,64 +731,38 @@ it( describe("MCP process input JSON Schema", () => { it("advertises process strings without weakening terminal input", async () => { - const server = new McpServer({ - name: "process-schema-validation", - version: "0", - }); - const client = new Client( - { - name: "process-schema-validation", - version: "0", - }, - { jsonSchemaValidator: createMcpSchemaValidation() }, + const tools = await listCanonicalAdvertisedTools(); + const processContract = TOOL_CONTRACTS.find( + ({ name }) => name === "capture_process_scenario", ); - const [clientTransport, serverTransport] = - InMemoryTransport.createLinkedPair(); - for (const contract of TOOL_CONTRACTS) - server.registerTool( - contract.name, - toolRegistrationOptions(contract), - async () => ({ content: [] }), - ); - try { - await server.connect(serverTransport); - await client.connect(clientTransport); - const processContract = TOOL_CONTRACTS.find( - ({ name }) => name === "capture_process_scenario", - ); - const processTool = (await client.listTools()).tools.find( - ({ name }) => name === "capture_process_scenario", - ); - if (processContract === undefined || processTool === undefined) - throw new Error("Process scenario tool was not advertised"); - const validate = new Ajv2020({ - strict: false, - validateFormats: false, - }).compile(processTool.inputSchema); - const base = { executable: "/usr/bin/true" }; - for (const input of [ - { ...base, executable: "/usr/bin/true\0" }, - { ...base, arguments: ["\0"] }, - { ...base, working_directory: "/tmp\0" }, - { ...base, environment: { KEY: "value\0" } }, - { ...base, filesystem_observation_paths: ["/tmp\0"] }, - ]) { - expect(processContract.inputSchema.safeParse(input).success).toBe( - false, - ); - expect(validate(input)).toBe(false); - } - const terminalInput = { - ...base, - events: [{ type: "input", at_ms: 0, data: "\0" }], - }; - expect(processContract.inputSchema.safeParse(terminalInput).success).toBe( - true, - ); - expect(validate(terminalInput)).toBe(true); - } finally { - await Promise.allSettled([client.close(), server.close()]); + const processTool = tools.find( + ({ name }) => name === "capture_process_scenario", + ); + if (processContract === undefined || processTool === undefined) + throw new Error("Process scenario tool was not advertised"); + const validate = new Ajv2020({ + strict: false, + validateFormats: false, + }).compile(processTool.inputSchema); + const base = { executable: "/usr/bin/true" }; + for (const input of [ + { ...base, executable: "/usr/bin/true\0" }, + { ...base, arguments: ["\0"] }, + { ...base, working_directory: "/tmp\0" }, + { ...base, environment: { KEY: "value\0" } }, + { ...base, filesystem_observation_paths: ["/tmp\0"] }, + ]) { + expect(processContract.inputSchema.safeParse(input).success).toBe(false); + expect(validate(input)).toBe(false); } + const terminalInput = { + ...base, + events: [{ type: "input", at_ms: 0, data: "\0" }], + }; + expect(processContract.inputSchema.safeParse(terminalInput).success).toBe( + true, + ); + expect(validate(terminalInput)).toBe(true); }); }); diff --git a/tests/evaluation/ciTestSelection.test.ts b/tests/evaluation/ciTestSelection.test.ts index eec26a8922..131fca2bf5 100644 --- a/tests/evaluation/ciTestSelection.test.ts +++ b/tests/evaluation/ciTestSelection.test.ts @@ -102,14 +102,22 @@ it("lets checker-only inputs and prose select nothing by themselves", () => { const sources = { "src/domain/Value.test.ts": "", "tests/boundary/filesystem/docs.test.ts": 'read("docs/guide.md");', + "tests/boundary/filesystem/mermaid.test.ts": + 'read("docs/architecture/architecture.mermaid");', + "tests/boundary/filesystem/licenses.test.ts": + 'read("licenses/npm-dependencies.txt"); read("LICENSE");', + "tests/boundary/filesystem/issueTemplate.test.ts": + 'read(".github/ISSUE_TEMPLATE/bug_report.yml");', }; - // A declaration, guard baseline or schema summary adds no tests… + // A declaration, guard/layer baseline, blame-ignore list or schema summary adds no tests… expect( selectedFiles( plan( [ "src/domain/Value.test.ts", "scripts/lib/helper.d.mts", + "scripts/lib/module-layers.baseline.json", + ".git-blame-ignore-revs", "scripts/verify/baselines/some-guard.json", "scripts/verify/advertised-schemas/tools/tool.json", ], @@ -117,15 +125,70 @@ it("lets checker-only inputs and prose select nothing by themselves", () => { ), ), ).toEqual(["src/domain/Value.test.ts"]); - // …and prose selects the tests that read it by path. + // …and prose, non-code docs assets, and license inventories select the tests that read them by path. expect(selectedFiles(plan(["docs/guide.md"], sources))).toEqual([ "tests/boundary/filesystem/docs.test.ts", ]); + expect( + selectedFiles(plan(["docs/architecture/architecture.mermaid"], sources)), + ).toEqual(["tests/boundary/filesystem/mermaid.test.ts"]); + expect( + selectedFiles(plan(["licenses/npm-dependencies.txt"], sources)), + ).toEqual(["tests/boundary/filesystem/licenses.test.ts"]); + expect(selectedFiles(plan(["LICENSE"], sources))).toEqual([ + "tests/boundary/filesystem/licenses.test.ts", + ]); + expect( + selectedFiles(plan([".github/ISSUE_TEMPLATE/bug_report.yml"], sources)), + ).toEqual(["tests/boundary/filesystem/issueTemplate.test.ts"]); + // Authored code under docs/ stays untraced so it falls back to the full suite. + expect(plan(["docs/.vitepress/config.ts"], sources)).toMatchObject({ + mode: "full", + reason: "Input is not traced to tests: docs/.vitepress/config.ts", + }); expect(plan(["scripts/verify/baselines/x.json"], sources)).toMatchObject({ mode: "full", }); }); +it("selects tests under shared-boundary directories and declared fixture owners, failing closed on missing owners", () => { + const sources = { + "src/process/OwnedCommand.test.ts": "", + "tests/boundary/cli/cliParityGoldens.test.ts": "", + "tests/boundary/cli/cliContractSchemaParity.test.ts": "", + }; + expect( + selectedFiles(plan(["src/process/OwnedCommand.test.ts"], sources)), + ).toEqual(["src/process/OwnedCommand.test.ts"]); + expect(plan(["src/process/OwnedCommand.ts"], sources)).toMatchObject({ + mode: "full", + reason: "Shared boundary: src/process/OwnedCommand.ts", + }); + expect( + selectedFiles( + plan(["tests/fixtures/cli-parity/error-unknown-command.txt"], sources), + ), + ).toEqual(["tests/boundary/cli/cliParityGoldens.test.ts"]); + expect( + selectedFiles( + plan( + ["tests/boundary/cli/cliContractSchemaParity.uncomparable.json"], + sources, + ), + ), + ).toEqual(["tests/boundary/cli/cliContractSchemaParity.test.ts"]); + // If an explicit fixture owner is missing or renamed, selection fails closed. + expect( + plan(["tests/fixtures/cli-parity/error-unknown-command.txt"], { + "src/process/OwnedCommand.test.ts": "", + }), + ).toMatchObject({ + mode: "full", + reason: + "Missing test owner tests/boundary/cli/cliParityGoldens.test.ts for tests/fixtures/cli-parity/error-unknown-command.txt", + }); +}); + it("traces shared script modules to their tests and falls back for other consumers", () => { const sources = { "scripts/lib/shared.mjs": "", @@ -194,6 +257,41 @@ it("traces shared script modules to their tests and falls back for other consume "tests/boundary/filesystem/runtimeVerifier.test.ts", "tests/boundary/filesystem/startup.test.ts", ]); + // Static checker JSON/Stryker configs, hyphen-prefixed script names in + // package.json (such as verify-shared.mjs vs scripts/lib/shared.mjs), and + // path-string tables in scopes.mjs do not force a full-suite fallback. + expect( + selectedFiles( + plan( + [ + "scripts/lib/shared.mjs", + "tests/support/analysis-efficiency/test_eval.py", + ], + { + ...sources, + "package.json": + '{"scripts":{"verify:shared":"node scripts/verify-shared.mjs"}}', + "tsconfig.scripts.json": '{"include":["scripts/lib/shared.mjs"]}', + "stryker.config.mjs": 'import "./scripts/lib/shared.mjs";', + "scripts/ci/scopes.mjs": + 'export const S = ["scripts/lib/shared.mjs"];', + }, + ), + ), + ).toEqual([ + "tests/boundary/filesystem/shared.test.ts", + "tests/boundary/filesystem/verifier.test.ts", + ]); + expect( + plan(["scripts/lib/shared.mjs"], { + ...sources, + "scripts/ci/scopes.mjs": 'import "../lib/shared.mjs";', + }), + ).toMatchObject({ + mode: "full", + reason: + "Script consumer is not traced: scripts/ci/scopes.mjs uses scripts/lib/shared.mjs", + }); // Other scripts remain untraced inputs. expect(plan(["scripts/build.mjs"], sources)).toMatchObject({ mode: "full", @@ -212,9 +310,27 @@ it("parses only well-formed plans and derives the selected shard count", () => { expect( testShards("fast", parseTestSelection(JSON.stringify(selected))), ).toEqual(["1/2", "2/2"]); + const maxSelectedFiles = Object.fromEntries( + Array.from({ length: 400 }, (_, i) => [ + `src/feature-${String(i)}.test.ts`, + "", + ]), + ); + expect(plan(Object.keys(maxSelectedFiles), maxSelectedFiles)).toMatchObject({ + mode: "selected", + shards: 8, + }); + const overCapFiles = { + ...maxSelectedFiles, + "src/feature-400.test.ts": "", + }; + expect(plan(Object.keys(overCapFiles), overCapFiles)).toMatchObject({ + mode: "full", + reason: "401 affected test files exceed the 400-file selection cap", + }); const full = { version: 3, mode: "full", reason: "Shared" }; expect(testShards("fast", parseTestSelection(full))).toHaveLength(12); - expect(testShards("full", parseTestSelection(full))).toHaveLength(6); + expect(testShards("full", parseTestSelection(full))).toHaveLength(12); for (const value of [ "", "null", @@ -229,8 +345,8 @@ it("parses only well-formed plans and derives the selected shard count", () => { { ...selected, shards: 0 }, { ...selected, - files: Array.from({ length: 5 }, (_, i) => `src/${i}.test.ts`), - shards: 5, + files: Array.from({ length: 9 }, (_, i) => `src/${i}.test.ts`), + shards: 9, }, { ...full, shards: 1 }, ]) diff --git a/tests/fixtures/mcpAdvertisedSchemaValidity.ts b/tests/fixtures/mcpAdvertisedSchemaValidity.ts index 08e083529b..2c86bb14bd 100644 --- a/tests/fixtures/mcpAdvertisedSchemaValidity.ts +++ b/tests/fixtures/mcpAdvertisedSchemaValidity.ts @@ -31,6 +31,9 @@ export function advertisedSchemaErrors( strict: false, validateFormats: false, addUsedSchema: false, + // Keep shared definitions as compiled functions instead of expanding them + // at every reference, while still compiling every $ref target. + inlineRefs: false, // This check compiles schemas to detect invalid references but never runs // their validators. Generated-code optimization adds substantial work here // without checking another contract; actual SDK validators keep defaults. diff --git a/tests/fixtures/minimalContractResult.ts b/tests/fixtures/minimalContractResult.ts index d84568475e..8977c561f4 100644 --- a/tests/fixtures/minimalContractResult.ts +++ b/tests/fixtures/minimalContractResult.ts @@ -27,6 +27,9 @@ const ajv = new Ajv2020({ strict: false, allErrors: false, validateFormats: false, + addUsedSchema: false, + inlineRefs: false, + code: { optimize: false }, }); type Root = Readonly>; diff --git a/tests/process-global/curatedWindowsLane.test.ts b/tests/process-global/curatedWindowsLane.test.ts index cdea873b03..230720c67c 100644 --- a/tests/process-global/curatedWindowsLane.test.ts +++ b/tests/process-global/curatedWindowsLane.test.ts @@ -213,4 +213,21 @@ it("checks the complete source with the fallback compiler before Windows build r const built = steps.findIndex((step) => step.run === "npm run build:cached"); expect(checked).toBeGreaterThan(-1); expect(built).toBeGreaterThan(checked); + const vitestRuns = steps + .map((step) => step.run) + .filter((value) => value?.includes("npx vitest run")); + // One Node 24 curated run and one Node 22 path-boundary run; both cover the + // three native directory-identity suites without booting a second Node 24 Vitest process. + expect(vitestRuns).toHaveLength(2); + for (const run of vitestRuns) { + expect(run).toContain( + "tests/boundary/filesystem/directoryArtifactReaderOpen.test.ts", + ); + expect(run).toContain( + "tests/boundary/filesystem/javascriptPackagePrecedence.test.ts", + ); + expect(run).toContain( + "tests/boundary/filesystem/javascriptArtifactContentIdentity.test.ts", + ); + } });