[codex] Stabilize ccusage timeout cleanup test

[zergzorg]

Summary

• make ccusage_timeout_kills_descendant_and_closes_pipes wait for the descendant pid file with a short retry loop
• give the fake runner a 1s timeout budget instead of 100ms so the shell has time to create descendant.pid before the timeout kill path runs
• keep the production ccusage timeout/kill behavior unchanged

Why

The test was racing its own setup. On a slower or busy machine, the 100ms timeout can kill the fake shell runner before it reaches:

echo $! > descendant.pid

When that happens, the assertion fails with:

read descendant pid: Os { code: 2, kind: NotFound, message: "No such file or directory" }

The test is meant to verify timeout cleanup of descendant processes and pipe draining, not whether a shell script can always initialize in under 100ms. The new version still exercises the timeout path, keeps the cleanup hang guard, and then verifies that the descendant is killed.

Validation

• cargo test plugin_engine::host_api::tests::ccusage_timeout_kills_descendant_and_closes_pipes -- --nocapture
• cargo test

---

Summary by cubic

Stabilizes the ccusage_timeout_kills_descendant_and_closes_pipes test to remove timing flakiness on slower machines while keeping production timeout/kill behavior unchanged.

• Bug Fixes
  • Poll for descendant.pid with a 20ms interval and 1s deadline before asserting.
  • Increase the fake runner timeout from 100ms to 1s to avoid setup races; still exercises the timeout path and verifies descendant kill and pipe cleanup.

^{Written for commit bbc999a. Summary will update on new commits. Review in cubic}

[cubic-dev-ai]

No issues found across 1 file

_{Re-trigger cubic}

[validatedev]

@codex review

[Copilot]

Pull request overview

This PR stabilizes the Unix-only ccusage_timeout_kills_descendant_and_closes_pipes test in src-tauri by reducing timing flakiness while keeping production timeout/kill behavior unchanged.

Changes:

• Add a small retry loop helper to wait for descendant.pid to appear before parsing it.
• Increase the fake runner timeout budget from 100ms to 1s, while still exercising the timeout path and verifying cleanup.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[chatgpt-codex-connector]

Codex Review: Didn't find any major issues. Another round soon, please!

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[robinebers]

LGTM — clean fix for the test race. Bumping the runner budget to 1s and adding the pid-file poll loop keeps the timeout path exercised without depending on sub-100ms shell startup. No production code touched.