[codex] fix moq-srt negative pacing offsets

[kixelated]

Summary

• Fix moq-srt egress pacing so frames with presentation timestamps before the current media anchor pace before the anchor instead of collapsing onto it.
• Keep the unavoidable startup/Instant underflow case panic-free by saturating only that boundary condition.
• Add a regression assertion for a reordered B-frame whose PTS trails the re-anchored live edge.

Root Cause

pace() computed anchor + Duration::from(ts).saturating_sub(Duration::from(base)). When ts < base, that flattened the negative media offset to zero, so reordered frames could be stamped at the live edge instead of their earlier media instant.

Validation

• cargo test -p moq-srt pace_re_anchors_to_live_edge
• cargo check -p moq-srt
• cargo fmt --package moq-srt
• git diff --check

Refs #1917.

(Written by Codex)

[coderabbitai]

Caution

Review failed

The pull request is closed.

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

Run ID: 6ca3d158-2380-43db-9a60-9b4d6d21defc

📥 Commits

Reviewing files that changed from the base of the PR and between 86ff316 and 806e0db.

📒 Files selected for processing (1)

• rs/moq-srt/src/server.rs

---

Walkthrough

pace now computes send_at with logic that can place it earlier than anchor when ts is before base. The test pace_re_anchors_to_live_edge adds a reordered-frame case that checks an earlier send_at while keeping the re-anchored base behavior unchanged.

✨ Finishing Touches

✨ Simplify code

• Create PR with simplified code
• Commit simplified code in branch codex/fix-srt-negative-pacing

---

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.}