Repository navigation
quest: Nightly's qmux benchmark handshake is already fixed - #8427
Conversation
Quest-Claim: 18d5f623dfb89b7f
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
No actionable comments were generated in the recent review. WalkthroughThe Nightly README no longer lists the WebSocket fallback benchmark handshake requirement, including its timeout and Audio quality job outcome description. The qmux-bench handshake-timeout investigation document was deleted. Estimated code review effort: Merge Risk:
|
There was a problem hiding this comment.
✅ No new issues found.
Reviewed changes
- Quest deletion —
quest/m0/nightly/qmux-bench-handshake.mdis removed in full; the slug no longer appears anywhere in the tree, so nothing dangles. - Epic README link — the matching
## Requiredbullet inquest/m0/nightly/README.mdis removed, leavingmacos.mdandwindows.md.
Verified the premise against the tree rather than the body: js/net/bench/qmux.ts now builds its fake peer with Session.accept, whose comment records that it sends transport parameters first, which is exactly the cause #8054 introduced and #8351 fixed. Grepping the checkout for qmux-bench-handshake returns no matches, and the Quest job passed, so quest check is clean on the reduced tree. #8351 also deleted its own parallel quest for the same failure, so this removes the duplicate that #8408 planned after the fix had already landed.
deepseek-v4.1-flash-1017 (free via Pullfrog for OSS) | 𝕏
kixelated
left a comment
There was a problem hiding this comment.
Automated review by review (OpenAI)
Reviewed commit 19071ad.
No actionable defects found in the two-file documentation cleanup. The deleted quest's corresponding Required link is removed, and the benchmark at this commit uses Session.accept for its server peer, consistent with the stated handshake fix. Overall direction: retiring this completed task looks sound; no runtime code changes are included.
Verification limits: GitHub source/diff review only. I did not run quest check, the benchmark, or Nightly, and have not independently confirmed the reported timing results.

Finishes
quest/m0/nightly/qmux-bench-handshake.mdby deleting it: the regression is already fixed onmain.Summary
The scheduled Nightly's
just js qmux-benchstep timed out on the handshake from 2026-11-07 on. The cause was moq-dev/moq#8054:Session.readystarted waiting for the peer's transport parameters, and the benchmark's fake peer never sent any. moq-dev/moq#8351 (merged 2026-11-09 18:00 UTC) made the fake peer a realSession.accept, so it sends its parameters. The last failing Nightly (run 42659882487, headb3fbce18) predates that merge, and no Nightly has run since.Local repro on current
main:just js qmux-benchpasses all 12 rows (eager ~5.6 ms / lazy ~22 ms at 0 ms latency, ~56 / ~123 ms at 50 ms).This PR only removes the quest and its link in
quest/m0/nightly/README.md. Public API: none. Wire: none.Caused by
readywaits for peer parameters, which the benchmark's fake peer did not send (fixed by moq-dev/moq#8351)Choices made without asking
Review
Skipped: the diff only deletes a finished quest and its link;
quest checkpasses (503 documents ok).just checknot run, since no code changed.Models
Size: S
Implementer: Claude Code, Claude Opus 5.5, low effort
🤖 Generated with Claude Code
(Written by Claude Opus 5.5 with low)