Fix IETF subscriber race cancelling groups before consumers attach

[kixelated]

Summary

• The Producer/Consumer refactor (a0f55d06) made unused() fire immediately when consumers == 0, which is the initial state for GroupProducer and FrameProducer
• The IETF subscriber's tokio::select! with unused() was a race condition: groups were randomly cancelled before any downstream publisher could call next_group() to create a consumer
• This broke cross-version relay forwarding (e.g. IETF Draft14 publisher → relay → moq-lite-03 subscriber), with catalog.json groups being silently dropped
• Removed the group.unused() and frame.unused() checks from the IETF subscriber, matching the lite subscriber which already skips them ("written to a cache")

Test plan

• Run just dev with publisher using moq-transport-14 and subscriber using moq-lite-03 — catalog.json should reliably reach the subscriber
• Run just check to verify no regressions

🤖 Generated with Claude Code

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between 28fc538 and bb98c20.

📒 Files selected for processing (1)

• rs/moq-lite/src/ietf/subscriber.rs

---

Walkthrough

This change refactors the control flow in the subscriber module by replacing tokio::select! blocks with direct await calls. Previously, the code used select! to concurrently monitor producer.unused() and frame.unused() futures as cancellation triggers. The refactor removes this intermediate branching, instead directly awaiting run_group() and run_frame() calls. The linear await-based approach maintains identical error handling semantics where errors close frames or groups and successful completions finish them. No public API changes were made.

🚥 Pre-merge checks | ✅ 2 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (2 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title directly addresses the main bug fix: removing race condition logic that was cancelling groups before consumers could attach.
Description check	✅ Passed	The description clearly explains the root cause (Producer/Consumer refactor), the bug (race condition with tokio::select!), impact (cross-version relay failures), and the solution (removing unused() checks).

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

• 📝 Generate docstrings (stacked PR)
• 📝 Generate docstrings (commit on current branch)

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment
• Commit unit tests in branch fix/ietf-subscriber-group-unused

---

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 and usage tips.}