[kixelated]

Problem

Two js/net tests fail intermittently under CPU load. moq-dev/moq#8326 saw them in two of five whole-package runs under 48 CPU burners, without naming them. Rerunning the whole js/net package 30 times under 48 burners on 24 cores (on 7d91a5b634) failed 3 runs, all on these two:

• js/net/src/group.test.ts: "a compressed frame is refused unless its worst-case slice fits a peer's frame cap", 2/30 runs.
• js/net/src/lite/publisher.test.ts: "lite draft-07: a re-offer dropped again before it opens stays stale at the end", 1/30 runs.

Approach

Quarantine both with test.skip and a comment pointing at a follow-up quest, per the flaky-test rule. No code or assertion changes.

Impact

None; tests only. No public API, no wire change.

Follow-ups

Plan one quest per test under quest/m0/test-flakes-2/:

• compressed-frame-cap.md: the test hit bun's 5000 ms per-test timeout (5840 ms and 10442 ms under load; about 0.85 s alone with the host near load 50). It deflates and inflates a random payload of nearly 32 MiB, so it looks CPU-bound rather than racy, like the flate tail-shed test fixed in moq-dev/moq#8137. Suspected fix: prove the worst-case-slice cap without a full-size payload, for example by deriving the bound from a small input or checking encodable directly. Recommendation: plan it.
• redropped-reoffer.md: failed in 22 ms with holesToEnd returning [] where [["drop", 1n, 1n, 3n]] was expected (publisher.test.ts:1572), so no stale drop stood at the end. Budget.poll is already stubbed, so the write budget is not the cause. The test sends update 2, yields once (flush() is one yieldNow()), then sends update 3, relying on update 3 arriving inside the task the publisher yields after update 2, before the re-offer of group 1 opens. Suspected cause, not confirmed: under load that one yield is not enough ordering, so the re-offer opens (and is served) before update 3 re-drops it. A fix should wait on an event, such as the re-offer being queued, rather than a host-task count. Recommendation: plan it.

A targeted --rerun-each 300 of the re-offer test under load did not run: it waited in quest gate behind a host-wide hold until its time limit.

Review

Skipped at ba0889d: the change only turns two test(...) calls into test.skip(...) with a comment each.

Models

Size: S
Implementer: Claude Code, Opus 5.5, high effort

🤖 Generated with Claude Code

(Written by Opus 5.5 with high)

[coderabbitai]

Warning

Review limit reached

You've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository.

Next included review available in 19 minutes.

Check out review usage here.

View limit details

Limit details: You’ve used all 4 included reviews currently available.

Learn how review limits work.

Review configuration:

Run configuration

• Source mode: Auto
• Repository: Unavailable
• Branch: Unavailable
• File: No configuration file from this repository applied
• Commit: Unavailable
• Layers: Organization UI
• Review profile: CHILL
• Plan: Advanced
• Run ID: 78c45f93-209a-44fd-9a82-e26ecbadcc15

Commits

Reviewing files that changed from the base of the PR and between 3e458ec and ba0889d.

Files selected for processing: 2

• js/net/src/group.test.ts
• js/net/src/lite/publisher.test.ts

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

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

[pullfrog]

ℹ️ No critical issues — one tracking gap inline.

Reviewed changes

• Quarantined compressed-frame cap test — js/net/src/group.test.ts turns the 32 MiB worst-case-slice test into test.skip, citing bun's 5 s per-test timeout under load.
• Quarantined re-offer test — js/net/src/lite/publisher.test.ts skips the draft-07 redropped-reoffer test, citing its dependence on a single flush() ordering update 3 ahead of the re-offer opening.

Quarantining rather than loosening a timeout or budget matches the flaky-test rule, and the PR body's investigation is specific enough to act on.

  ^{｜ Fix all ➔ ｜ Fix 👍s ➔ ｜ View workflow run ｜ Using deepseek-v4.1-flash-1017 (free via Pullfrog for OSS) ｜ 𝕏}

[kixelated]

Automated review by review (OpenAI)

Reviewed ba0889d.

No actionable findings. The diff only quarantines the two identified load-sensitive tests with test.skip and follow-up references; assertions and production behavior are unchanged. This removes their CI coverage rather than fixing the suspected timeout/order issues. Verification: static diff review only; the reported load reproductions were not independently rerun.