chore: release

[moq-bot]

🤖 New release

• moq-tokio: 0.19.22 -> 0.19.23 (✓ API compatible changes)
• moq-stats: 0.2.11 -> 0.2.12 (✓ API compatible changes)
• moq-relay: 0.17.1 -> 0.17.2 (✓ API compatible changes)
• libmoq: 0.6.11 -> 0.6.12
• moq-boy: 0.5.11 -> 0.5.12
• moq-rtc: 0.3.11 -> 0.3.12
• moq-cli: 0.14.1 -> 0.14.2
• moq-ffi: 0.4.11 -> 0.4.12
• moq-gst: 0.4.11 -> 0.4.12

Changelog

moq-tokio

0.19.23 - 2026-10-06

Fixed

• (moq-tokio) finite default connection receive window (backport fix(moq-tokio): finite default connection receive window (noq reassembly cap, part 1) #4605) (#4858)

moq-stats

0.2.12 - 2026-10-06

Added

• (stats) linger an empty group broadcast before unannouncing it (backport feat(stats): linger an empty group broadcast before unannouncing it #4871)

moq-relay

0.17.2 - 2026-10-06

Added

• (stats) linger an empty group broadcast before unannouncing it (backport feat(stats): linger an empty group broadcast before unannouncing it #4871)

libmoq

0.6.12 - 2026-10-06

Other

• updated the following local packages: moq-tokio

moq-boy

0.5.12 - 2026-10-06

Other

• updated the following local packages: moq-tokio

moq-rtc

0.3.12 - 2026-10-06

Other

• updated the following local packages: moq-tokio

moq-cli

0.14.2 - 2026-10-06

Other

• updated the following local packages: moq-tokio, moq-relay, moq-rtc

moq-ffi

0.4.12 - 2026-10-06

Other

• updated the following local packages: moq-tokio

moq-gst

0.4.12 - 2026-10-06

Other

• updated the following local packages: moq-tokio

---

This PR was generated with release-plz.

[kixelated]

Automated review of #4867 (release-plz), head 5c80abff53a3e5c3ca4646df39ea77812f52fd59

This PR doesn't actually bump anything. All 25 version bumps listed in the body already landed on release in #4763 (merged 15:04 PT, commit 4810d476). The real diff against release is only two files: a spki 0.8.0 → 0.8.1 lockfile bump and a second moq-tokio changelog entry.

Blocking

1. Duplicate 0.19.22 heading in rs/moq-tokio/CHANGELOG.md. After this PR, the file has two ## [0.19.22] ... - 2026-10-05 sections back to back (lines 10 and 16 on head). The new one holds just one bullet, release ([#4763]), which is the release commit listing itself as a change. Merging this leaves a broken changelog for a version that the in-flight Release RS run is about to publish to crates.io.
   • Likely cause (a hunch, not verified): just rs release runs release-plz release-pr before release-plz release. So on the push from chore: release #4763, the PR step ran while moq-tokio 0.19.22 was still unpublished; crates.io still shows 0.19.21 while moq-net 0.3.10 is already up. It then treated the chore: release #4763 commit as a new unreleased change. Running release before release-pr in rs/justfile would probably stop this follow-up PR from appearing after every release merge.
   • Fix: close this PR, or delete lines 10-15 of rs/moq-tokio/CHANGELOG.md. Then let release-plz regenerate it after Release RS run 37380041632 finishes.

Non-blocking

2. The PR body is stale and misleading. It repeats chore: release #4763's 25-crate bump list and changelogs, none of which this diff contains. Anyone skimming it will think this is a new release.
3. The spki 0.8.0 → 0.8.1 lockfile bump is fine. It's a transitive dependency of elliptic-curve/ecdsa, published to crates.io at 15:12 PT, not yanked, and the checksum matches. It can ride along with the next real release.
4. CI is still pending (Check, Test, WASM, Android, Swift, OBS, Windows, macOS).

Verdict: ITERATE. Drop the duplicate changelog section, or close this PR and let release-plz regenerate it once the current publish finishes.

This is an automated review, not the maintainer's decision
(Written by Grok)

[kixelated]

Automated review by review (OpenAI)

Reviewed commit: 5c80abf.

Direction: iterate on the generated release metadata. The actual two-file diff contains a spki 0.8.0 → 0.8.1 lockfile update and a changelog insertion, not the 25 version bumps described in the PR body.

Confirmed the existing finding in #4867 (comment): rs/moq-tokio/CHANGELOG.md:10–20 has duplicate 0.19.22 sections, with the new section merely listing the prior release PR. Remove the redundant section or regenerate the release PR, and refresh its body to match the actual diff. No duplicate inline finding added.

Verification limits: reviewed the complete diff, head changelog, and existing discussion; no build, release-plz run, registry checksum verification, or publishing was performed.

[kixelated]

Automated review by review (OpenAI)

Reviewed commit: 050b918.

The previous duplicate-changelog/body findings are addressed. This is now a real nine-crate release: manifests, workspace requirements, lockfile versions and new changelog headings agree with the body. moq-tokio has one 0.19.22 heading followed by the new 0.19.23 entry for the receive-window backport; the stats/relay entries describe the linger backport. No new actionable correctness findings in the full 20-file release diff.

Direction: the regenerated release metadata is coherent. Keep this as the release-branch publication step; it does not include the unrelated breaking work on main.

Verification: static review of every manifest/changelog/lockfile change, current release baseline and prior discussion. No release-plz execution, publish dry run, registry checksum validation or local build was run. Android passed; Check, Swift, WASM and OBS were queued and Platform was running when checked, so this is not a release-readiness confirmation.

(Written by OpenAI)