[kixelated]

Problem

After #5376, alternating named and anonymous retractions at one clock reading still queues a prefix repeatedly. A regression reproduced 100 queued entries for one prefix.

Approach

Backport the anonymous-retraction fix from #5391 (61c7ac2). Keep an anonymous quarantine marker with no publisher so it deduplicates queue entries without delaying announcements. Extend the regression to alternate named and anonymous routes and retain expiry coverage.

Impact

• No public API or wire changes.

Alternatives

A separate queued-prefix set would duplicate bookkeeping already held by the quarantine map.

Follow-ups

Merge after #5391 lands on main.

Checks

• just check passes, including 5,080 Rust tests.
• origin/withdraw completes all four publisher/subscriber shapes. Timings are noisy alongside concurrent builds; no speedup claim.

(Written by GPT-6)

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-13T04:54:16.533761Z	e3e7ab5	Draft marked ready

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" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[kixelated]

Automated review by review (OpenAI)

Reviewed commit: e3e7ab5

No actionable findings in the full one-file diff. This backports the anonymous-retraction fix from #5391 and addresses the remaining queue-growth issue reported in #5376. Publisher-less markers preserve same-reading deduplication without holding announcements; the alternating named/anonymous regression retains eventual-expiry coverage.

Direction: focused reuse of existing per-prefix state, correctly adapted to release’s announcement API. No public API or wire-format changes observed.

Verification: static review of the full patch and surrounding announcement/coalescing, quarantine, and timer paths. Tests and benchmarks were not run; reported results were not independently reproduced.