fix(lint): underscore the Linux arm's stack_lo too — #8298 fixed only the macOS arm

[proggeramlug]

One-liner unbreaking the warnings job (-D warnings, host-compatible scope) on main: #8298 underscore-fixed the unused stack_lo on the macOS arm of gc/native_stack_scan.rs but left the bare name on the #[cfg(not(target_os = "macos"))] arm — the one Linux CI actually compiles. Caught on #8291's rerun (job 95306309639's sibling run 32015928121); verified still live on main tip 85a60d4a3.

Summary by CodeRabbit

• Refactor
  • Updated internal stack-bound handling for non-macOS builds without changing stack-scanning behavior.

[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: ab666ed6-5070-4b8e-b58e-45e1b9c282ab

📥 Commits

Reviewing files that changed from the base of the PR and between 85a60d4 and 658482b.

📒 Files selected for processing (1)

• crates/perry-runtime/src/gc/native_stack_scan.rs

Included review availability: Your plan includes up to 8 reviews per rolling hour; 1 remains after this review.

---

📝 Walkthrough

Walkthrough

The non-macOS native stack scanner renames its unused lower-bound binding to _stack_lo. Stack range calculation and scanning behavior remain unchanged.

Changes

Native stack scanning

Layer / File(s)	Summary
Rename unused stack bound crates/perry-runtime/src/gc/native_stack_scan.rs	The non-macOS binding changes from stack_lo to _stack_lo. The computed stack range and scan behavior remain unchanged.

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

Merge Risk: ⚪ Minimal · up to 65848

This is a localized lint-only fix for the Linux arm code path, with no actionable merge-blocking risk remaining beyond normal checks and review.

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Description check	⚠️ Warning	The description explains the change but omits the required Summary, Changes, Related issue, Test plan, Screenshots/output, and Checklist sections.	Use the repository template and include the required sections, issue reference, test verification, and checklist status.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly identifies the Linux stack_lo lint fix and distinguishes it from the earlier macOS-only fix.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
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 docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch fix/warnings-stack-lo-linux-arm

---

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]

Merging. This is a correction to my #8298, and the failure mode is worth naming because it is not obvious.

#8298 fixed the unused stack_lo on the #[cfg(target_os = "macos")] arm and I verified it with RUSTFLAGS="-D warnings" cargo check — on macOS. That check structurally cannot see the #[cfg(not(target_os = "macos"))] arm, because rustc never compiles it here. So a green local -D warnings run said nothing at all about the arm Linux CI actually builds, and warnings stayed red on main.

Generalising: a warnings check on one host is not evidence about cfg-gated code for another. Same class as a #[cfg(windows)] thread-local slipping past a census I ran on macOS earlier today (#8270). Worth remembering that "I ran the gate" and "the gate covered this code" are different claims whenever cfg is involved.

I checked the rest of the file for the same shape before merging, so this should not need a third round: native_stack_scan.rs has two cfg(not(target_os = "macos")) arms and the only destructuring binding in either is the one this PR fixes.

Thanks for catching it — and for the precise provenance (job 95306309639's sibling run, verified live on tip). That made it a thirty-second confirmation instead of a hunt.