docs(quest): fold #3946's tune-in design and repro into its quest

[kixelated]

Problem

#3946 carries only a red regression test for the moq play tune-in stall, and the quest now waits on the play harness (which waits on #3966), so nothing there can land yet.

Approach

Record #3946's chosen design in quest/m1/play-tunein-backpressure.md: buffer encoded frames for the whole max_age window, decode a few ahead, evict the oldest decoded frame when that queue fills. Note the repro's shape and that it stays on #3946's branch for reuse. Resize the quest to [M], since the decoder split is now in scope.

Impact

• Public API / wire: none

Alternatives

Keep #3946 open as a draft until its prerequisites land.

Follow-ups

None.

(written by Claude Opus 5.5)

🤖 Generated with Claude Code

[coderabbitai]

Warning

Review limit reached

Next included review available in 10 minutes.

Check out review usage here.

View limit details

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

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

Learn how review limits work.

Review configuration:

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Advanced

Run ID: fe5dfbb1-6896-4030-94f1-afd2af10cb33

📥 Commits

Reviewing files that changed from the base of the PR and between aee97ca and 3d318d8.

📒 Files selected for processing (1)

• quest/m1/play-tunein-backpressure.md

---

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.}

[kixelated]

Recommendation: MERGE

Positive improvement? Yes. #3946 is stuck behind the play harness / #3966, so folding its chosen tune-in design and repro shape into the quest unblocks planning without waiting on that stack. The old open A/V alternatives are replaced with a concrete policy (encode-buffer for max_age, decode a few ahead, evict oldest decoded when full), which is what a quest should carry once a choice is made.

Worth the complexity? Yes — and there is almost none. One markdown file, +15/−14, size bump [S] → [M] because the decoder split is now in scope. No public API or wire impact.

Different approach? Keeping #3946 open as a draft until prerequisites land (the PR's own alternative) would leave the design knowledge stranded on a blocked branch. Capturing it on the quest is the better split: design + test intent land now; implementation stays on #3946 / follow-ups.

Nits (non-blocking):

• The repro branch name is quoted as quest/main/play-tunein-backpressure; worth a quick check that matches the actual test(cli): reproduce wide-delay play tune-in stall #3946 branch so a later porter does not hunt the wrong ref.
• "Port it onto the harness, and also cover …" expands Required scope a fair bit for an [M]; that sizing looks intentional, just confirm the extra cases (speaker-owned anchors, discontinuity, tail flush) are meant as part of this quest and not a separate follow-up.

Otherwise this is a clean docs fold. Merge it.

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

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 3d318d8d36

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".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Preserve future-due frames when the decoded queue fills

With a playout delay longer than the small decoded queue's media span, the decoder can fill that queue while every frame is still scheduled in the future. Unconditionally evicting its oldest entry then discards the next frame that should be presented and can produce repeated jumps during normal catch-up or decoder bursts. Keep this queue backpressured while the independent encoded reader continues draining, or restrict eviction to frames that the current anchor says are already past due.

Useful? React with 👍 / 👎.

[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 2026-09-23T21:29:06.879333Z	3d318d8	PR opened

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.