Repository navigation
quest: plan the TS SI revision burst investigation - #7810
Conversation
While #7729 was tested, the two-exporter TS fixture dropped a queued TDT/TOT revision with an unlimited SI max-age. Unconfirmed and maybe fixture-only; the quest proves a burst reaches its stamped slots or fixes the fixture. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
Warning Review limit reachedYou've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository. Next included review available in 26 minutes. View limit details
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. |
kixelated
left a comment
kixelated
left a comment
There was a problem hiding this comment.
Automated review by review (OpenAI)
Reviewed head 208cb21.
[P2] Bound the goal by the SI track's freshness policy. quest/m1/ts-reliability/clock-revision-burst.md:5-8, 42-45 promises every queued revision "whatever the SI track's max-age", but the proposed test uses an unlimited track. On a zero-age SI track, a newer group makes the older one stale before the ordered subscriber reads it; the max-age policy deliberately permits that skip, as the linked #7729 regression demonstrates. The unlimited test could pass while the stated goal remains impossible under that policy, or a future fix could try to override the publisher's bound. Scope the guarantee and test to revisions still within both the publisher's max-age and the export delay, and distinguish intentional expiry from a lost fresh revision.
Verification: GitHub-only static review of the quest, ordered SI subscription and the linked max-age change. I did not run the proposed burst test.
(Written by review (OpenAI))
![pullfrog[bot]](https://avatars.githubusercontent.com/in/1768019?s=60&v=4){kind=small}
There was a problem hiding this comment.
✅ No new issues found.
Reviewed changes
- Quest plan: adds
quest/m1/ts-reliability/clock-revision-burst.md, scoping a burst test for queued TDT/TOT revisions and either fixing the exporter or proving the loss is theexport_twicefixture's. - Epic link: adds the quest to the TS fidelity and recovery
## Requiredlist.
The plan's fixture references check out: export_twice is used by exactly the three named tests, its clock tracks come from timed() (no max age, so unlimited on main), and revisions arrive between the two exporters' drains ahead of their stamped slots. The only recorded repro in #7729 is the zero-max-age regression, so the plan's "reproduce first" step is the right starting point for the unconfirmed burst loss.
deepseek-v4.1-flash-1017 (free via Pullfrog for OSS) ｜ 𝕏
Plans one follow-up from testing moq-dev/moq#7729. The maintainer approved it for m1 on 2026-11-04, so
/quest-planran without an interview.Quest
[S]quest/m1/ts-reliability/clock-revision-burst.md, under TS fidelity and recovery: "A TS export never loses a queued clock revision". While feat(net)!: authored tracks default to a zero max age #7729 was tested, the two-exporter fixture in moq-mux's TS export tests dropped a queued TDT/TOT revision even though the SI track's max-age was unlimited. That fixture isexport_twiceinrs/moq-mux/src/container/ts/export_test.rs, which three late-join tests use. The loss is unconfirmed and may come from the fixture alone. The quest is done when either a burst test shows every queued revision reaching its stamped slot, or the loss is traced to the fixture and the fixture is fixed.a_superseded_si_revision_stays_readable_within_the_export_budget. So the quest starts by writing a burst test.Overlap search
The epic's open children (skip-flag and republish-announce) don't touch clock tables.
tdt-placement.mdwas deleted when moq-dev/moq#7779 placed revisions on the media grid. No open PR or quest covers losing a queued revision.Merge order
Any order works. #7785 and #7780 delete their own quests and the README links to them. Their edits to the epic README sit next to the line this PR adds, so a local merge with
queston PATH resolves the conflict.Checks
quest check --base origin/main(597 documents ok) andsh/markdown.sh checkpass. Nothing was built.Review
Round 1:
quest reviewwith Codex (gpt-6.1-sol, medium effort) on head208cb21ca78e4d9fe986886914ae49a7cd974b38, baseeacff68011. It reported no findings.(Written by Claude Opus 5.5)
🤖 Generated with Claude Code