[JacobStephens2]

Summary

 Release targets
   x86_64-unknown-linux-musl  → ubuntu-24.04
   aarch64-unknown-linux-musl → ubuntu-24.04-arm
-  aarch64-apple-darwin      → macos-26
 CI
   test, test-linux-arm, console-test, site-test
-  macos-plan, test-macos-gate, test-macos-shard, test-macos
-  test:macos label handling and macos-gate nextest profile

Preserves Linux job definitions, ARM opt-in policy, concurrency and release smoke checks. Updates active CI/testing documentation; historical research stays untouched. Broader platform documentation and ADR 0003 are assigned to #1668.

Closes #1666

Evidence

• Before: Three release targets and four macOS CI jobs, plus a manually maintained macOS nextest filter.
  After: dist plan --output-format=json emits exactly the two static Linux build targets on their Ubuntu runners; installer target triples are Linux-only. A structural comparison confirms existing Linux jobs and concurrency are unchanged.
• python3 scripts/generate-release.py and python3 scripts/generate-release.py --check passed. Generated release.yml is byte-identical: its matrix comes from the runtime release plan.
• python3 -m unittest discover -s scripts -p test_generate_release.py: 3 tests passed.
• scripts/affected-tests main: passed, including site, Console and the full Rust suite. Local cargo test took 4,378.39 seconds, including 3m 27s compilation, with RUST_TEST_THREADS=3. Local setup uses Python 3.11 and the installed Playwright browser cache.
• One local Console copy-path test timed out; that exact case passed in isolation, then the full Console rerun passed all 126 tests. No product changes were needed.
• CI and Release passed on 828cdfa7, including both Linux release builds and ARM, Rocky and WSL smoke checks. ARM CI remains intentionally skipped for this PR into a Spec branch.

Existing validation seams: release generator CLI and cargo-dist release plan; CI job/concurrency definitions; repository affected-test selector. No new tests were needed for these configuration removals.

Review coverage

Independent thirdshift-code-review reviews used issue-1665 as the fixed point. Standards: 0 findings. Spec: 0 findings. Both reviewers read all five changed files: .config/nextest.toml, .github/workflows/ci.yml, dist-workspace.toml, docs/ci-performance.md, and docs/testing.md.

Unread changed files: none on either axis. Generated-output exclusions: none.

Reports and each axis's full files-read list are retained in .thirdshift-review-AHJC1Q/standards.md and .thirdshift-review-AHJC1Q/spec.md, with counts in summary.json.

Unaddressed findings

Standards

None.

Spec

None.

Security

None. Guidance-mode security review against pinned merge base 86c8887f557b225573a32552bb3fb0927bc04edf covered all five changed files and supporting CI, release generation, artifact handoffs, publication hooks and update code. No introduced security findings were identified.

Source comparison confirmed retained Linux CI job content and concurrency are unchanged; release permissions, smoke checks and publication gates are unchanged. Reused passing CI and Release checks on unchanged head 828cdfa7, as docs/testing.md permits. No source changes were required.

Security review outcome

No unaddressed introduced findings.

Merge Danger

Door: two-way

Configuration can be reverted. Future releases stop publishing native macOS binaries; existing release artifacts are unaffected.

Blast Radius: CI

Removes native macOS validation and release builds. Linux build/test policies and publication hooks remain in place.

Built with codex codex-cli 0.166.0 · gpt-6.1-sol · xhigh