[prisis]

Summary

Two fixes from the review of the refactor, applied on top of current main. The refactor itself is already in main via #25, so this PR does not touch the module split.

• Run teardown: services booted by a run are now torn down when setup fails or when the run is a dry run. Cleanup is installed right after services are wired, and finalize, dry-run, runtime setup and the run:before hook run inside the try/finally. Regression test: run-teardown-on-throw.test.ts, which fails on the previous handler.
• Publish preflight: remote-provider and repo lookups in the release-PR check and the push-rule check record a _preflight failure instead of rejecting publishContext. Regression test: orchestrator-preflight-guard.test.ts, which fails on the previous preflight.

Behaviour change: a dry run now also runs teardown, and a failed remote lookup returns a failure result instead of throwing.

Testing

• lint:types and check:schemas-drift pass.
• Full suite: 73 failing, unchanged from the base run. One perf test is timing-sensitive under load: audit-offline fails in the full run and passes alone. The release-plan-perf test that failed in the base run passes here.

Needs manual check before merge

• Ctrl+C during run setup: signal cleanup now starts earlier. Service cleanup runs, but Node may not exit. No test covers this.

🤖 Generated with Claude Code

[github-actions]

Thank you for following the naming conventions! 🙏

[pkg-pr-new]

Open in StackBlitz

npm i https://pkg.pr.new/@visulima/vis@30

npm i https://pkg.pr.new/@visulima/vis-mcp@30

commit: 36f9628

[codspeed]

Merging this PR will not alter performance

⚠️ 8 benchmarks spent significant time in system calls

System calls cannot be consistently instrumented, so they are not included in the measure, which understates the real cost. Please switch to the Walltime instrument to accurately measure system calls.

Measurement and system calls

• detectLockfileDrift · aube over foreign lockfile
• resolveInstaller · auto (detect → pnpm)
• detectPm · bun workspace
• detectPm · pnpm workspace
• resolveInstaller · explicit backend (no fs probe)
• resolveLocalBin · miss (full ancestor walk → PM fallback)
• detectPm · npm workspace
• × 2800 packages

✅ 12 untouched benchmarks
⏩ 101 skipped benchmarks¹

---

_{Comparing fix/vis-review-fixes (6d7abdb) with main (ccb03ee)²}

Footnotes

1. 101 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

2. No successful run was found on main (27fe29c) during the generation of this report, so ccb03ee was used instead as the comparison base. There might be some changes unrelated to this pull request in this report. ↩

[github-actions]

🚀 Great job, @prisis! You've merged your 11th PR! 🎊

Your contributions to vis are making a real difference. Keep up the fantastic work! 💪
Remember, every PR counts and helps improve the project. What will you tackle next? 🤔

@prisis is here if you need any guidance. Onward and upward! 🚀