fix(ci): tolerate a missing target/debug/deps in the cargo-test prune step

[jdalton]

Summary

Fixes a CI script bug that has been failing main's scheduled and push runs repeatedly, unrelated to any specific PR's code changes.

What broke and why

The full cargo-test job (nightly cron, release tags, workflow_dispatch) prunes linked test binaries between packages to avoid exhausting runner disk, using a bare find target/debug/deps -maxdepth 1 -type f -perm -111 ! -name '*.so' -delete.

That directory only exists once something has actually built into it. When the job's scope builds straight to --release (the perry/perry-stdlib branch just above the first prune call) and no cargo test -p <package> has run yet, target/debug/deps is absent, find exits 1, and the whole job fails, even when every test in it passed. I confirmed this against a real run: cargo test reported 2595 passed; 0 failed, and the job still failed on this exact line right after.

Changes

Both prune calls in .github/workflows/test.yml (find target/debug/deps ...) now redirect stderr and fall back to true, so a missing directory is treated as a no-op instead of a job failure. The behavior when the directory does exist is unchanged.

Related issue

n/a

Test plan

I confirmed the failure mode against a real CI run at https://github.com/PerryTS/perry/actions/runs/32553699612, where the cargo-test job passed every test and then failed on this exact line with find: 'target/debug/deps': No such file or directory. I ran actionlint against the edited workflow file and it reported no new findings, only two pre-existing unrelated shellcheck notes elsewhere in the file that this change does not touch. cargo build --release and the workspace test suite are not applicable here, since no Rust source changed.

Checklist

I have not bumped the workspace version or edited CLAUDE.md or CHANGELOG.md. My commit follows the fix: prefix convention used in the log. I have read CONTRIBUTING.md and agree to the Code of Conduct.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: ce4901b8-53b1-4d24-96ac-ba0fc695fa8a

📥 Commits

Reviewing files that changed from the base of the PR and between 9963215 and 6ca23fb.

📒 Files selected for processing (1)

• .github/workflows/test.yml

Included review availability: Your plan provides up to 8 included reviews per hour; 7 remain after this review.

---

📝 Walkthrough

Walkthrough

The full cargo-test workflow now suppresses executable cleanup errors and continues when deletion fails in both cleanup locations.

Changes

Test workflow cleanup

Layer / File(s)	Summary
Error-tolerant cleanup commands .github/workflows/test.yml	Both executable cleanup commands suppress stderr and continue successfully when deletion fails.

Estimated code review effort: 1 (Trivial) | ~2 minutes

Merge Risk: ⚪ Minimal · up to 6ca23

This localized CI change makes pruning a missing directory a no-op, preventing otherwise successful test jobs from failing; no actionable merge-blocking risk remains after normal checks and review.

Suggested reviewers: proggeramlug, thehypnoo

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Description check	⚠️ Warning	The description explains the bug and test results, but it omits the required Changes, Related issue, and Checklist sections.	Add the missing template sections, use "n/a" for Related issue if applicable, and complete the required checklist items.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly identifies the CI fix and the missing target/debug/deps directory that caused the prune step to fail.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check. Docstring coverage is scoped to functions touched by this diff. Analyzed 0 functions across 0 files. (1 skipped: 1 unsupported.)
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[proggeramlug]

Landed on main in 38dac3b via #8631, together with the changelog fragment this PR was missing.

Verified your actionlint claim rather than taking it on trust: findings are identical on origin/main and on the merged branch (same 3, at lines 1011 and 3489 — this PR touches 1078/1088). Also confirmed the mechanism locally: find on a missing directory is rc 1, rc 0 with the guard.

One thing noted in the changelog fragment, not a blocker: 2>/dev/null || true suppresses genuine find errors too, so if the prune ever silently stops working the symptom becomes a later disk exhaustion rather than a pointed error. Right tradeoff for a disk-hygiene step; just worth having written down.

Thanks — this was failing main repeatedly.

[jdalton]

Appreciate the independent verification, and good catch on the tradeoff.

On the failure-mode tradeoff

Agreed the failure mode changes shape rather than disappearing: a genuinely broken prune now surfaces as disk exhaustion later instead of a pointed find error at the point of failure. For a best-effort disk-hygiene step, that's the right side to fail on - a red job now would cost every consumer of main a false failure on their own unrelated change. Thanks for landing it in 38dac3b4c.