land #8880: restore the end/close arm on the stdin provider path

[proggeramlug]

Lands #8880 — restores the end/close arm on the stdin listener provider path.

The fix is correct and needed. stdin_on_op / stdin_off_op are what the stdin object's native on / once / addListener delegate to — every registration that is not codegen's literal process.stdin.x(…) shape, i.e. an alias (const s = process.stdin; s.once("end", …)) or stdin passed as a parameter (helper(process.stdin)). Claude Code's print-mode reader is exactly the parameter form, so those registrations hit _ => return and were silently discarded, and the end half of race(once("end"), timeout(3000)) could never win.

Verified on main before merging: stdin_off_op has only data / readable arms — no end/close. So the gap is real.

One correction to the PR description

It attributes the gap to work being "lost when #8861 was batch-landed". That is not what happened, and the record is worth keeping straight:

• The ref I merged (0bf52fbbf) is byte-identical to fix(stdin): deliver piped stdin to -p — lower once, keep every end listener, buffer pull-mode bytes #8861's head today — it did not move after I fetched it.
• stdin_on_op in fix(stdin): deliver piped stdin to -p — lower once, keep every end listener, buffer pull-mode bytes #8861 as submitted has exactly "data" / "readable" / "keypress" / _ => return. The "end" | "close" arm was never present.
• Neither provider_path_registers_end_listeners nor provider_path_removes_end_listeners appears anywhere in fix(stdin): deliver piped stdin to -p — lower once, keep every end listener, buffer pull-mode bytes #8861 (0 occurrences).
• The mod_tests.rs split I did while landing preserved every test: 15 #[test] in fix(stdin): deliver piped stdin to -p — lower once, keep every end listener, buffer pull-mode bytes #8861's mod.rs, 15 across mod.rs + mod_tests.rs on main.

So this is completing #8861's third layer, not restoring something dropped. #8861 landed the two layers it actually contained — the STDIN_END_CALLBACKS GC rooting (which I added while landing, since the new list was missing from the root scanner) and STDIN_PULL_MODE.

Where I was incomplete: while landing #8864 I checked that the "end" | "close" arm survived at readline/mod.rs:1651 and reported the interaction as safe. That line is the syntactic extern's arm, not the provider's. Checking one of two paths and reporting on both is the gap worth recording.

Validation

• all 30 lint-job gates pass
• both new tests pass: provider_path_registers_end_listeners, provider_path_removes_end_listeners — and they assert the provider entry points directly, so this cannot silently regress again
• perry-stdlib 124, perry-runtime 2716 — 0 failed

Summary by CodeRabbit

• Bug Fixes

  • Fixed piped input commands that could hang when listening for stdin completion events.
  • Restored support for end and close events when stdin is accessed through aliases or passed as a parameter.
  • Ensured these listeners are properly removed, preventing stale callbacks from accumulating.

• Tests

  • Added coverage for registering, triggering, and removing stdin completion listeners.

[coderabbitai]

Caution

Review failed

The pull request is closed.

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: c6b4dac2-a32e-4696-8401-0e17f0f5f947

📥 Commits

Reviewing files that changed from the base of the PR and between ea39b9c and 2ed020c.

📒 Files selected for processing (3)

• changelog.d/8879-stdin-end-provider-arm.md
• crates/perry-stdlib/src/readline/mod.rs
• crates/perry-stdlib/src/readline/mod_tests.rs

---

📝 Walkthrough

Walkthrough

The stdin provider now handles "end" and "close" listeners. Removal clears registered callbacks. Tests cover registration, EOF firing, and removal. The changelog records the fix and benchmark results.

Changes

Stdin provider listener lifecycle

Layer / File(s)	Summary
Provider listener registration and removal crates/perry-stdlib/src/readline/mod.rs, crates/perry-stdlib/src/readline/mod_tests.rs, changelog.d/8879-stdin-end-provider-arm.md	stdin_on_op stores "end" and "close" callbacks. stdin_off_op removes them and clears the matching close callback. Tests verify callback registration, EOF execution, and removal. The changelog documents the fix and benchmark results.

Estimated code review effort: 2 (Simple) | ~10 minutes

✨ 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 land-8880-stdin-provider

---

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.}