Skip to content

quest: record the maintainer's 2026-11-15 answers to waiting items - #8870

Merged
kixelated merged 2 commits into
mainfrom
quest-complete/wait-answers-2026-11-15
Nov 15, 2026
Merged

kixelated merged 2 commits into
mainfrom
quest-complete/wait-answers-2026-11-15

Conversation

@kixelated

@kixelated kixelated commented Nov 15, 2026 •

Copy link
Copy Markdown
Collaborator

Problem

On 2026-11-15 the maintainer answered the quest ## Waiting items walked by /quest-complete. Each answer has to be recorded in its quest's Plan and the item removed, and quests whose only purpose was a wait have to go.

Approach

  • quest/m3/gpu-ci.md: records the sandboxed-container runner and drops the VFIO passthrough text.
  • quest/m3/cli-packaging.md: records wait-then-fork; quest/m3/cpal-alsa-runtime.md's fork fallback now cites it instead of "the maintainer decides".
  • quest/m3/capture-linux.md, quest/m3/capture-windows.md: record window capture only. Source::App already returns Unsupported("application capture") off macOS (rs/moq-video/src/capture/mod.rs), so both Goals drop app capture as work.
  • quest/m3/lite-tail-drop-count.md: records the max-delay wait, and an error on lite-05/06 and lite-07 when no max delay is set.
  • quest/m3/webrtc-dial-policy.md: records the moq-net value type reached through the two #[non_exhaustive] configs.
  • quest/m3/quic-ack-hook.md, moved with quest mv to quest/m4/quic-ack-hook.md (links in quic-deadline.md follow): Goal and Plan rewritten to the trap tests on the crate-private methods; the public poll_acked/acked/offset() API is dropped. Kept at [S] (the item suggested XS), since it is about nine trap cases through two drivers.
  • quest/m4/resume-redesign.md: deleted with quest rm; quest/m4/resume-trace-validation.md's check-the-redesign bullet now says the per-reader path stays.
  • quest/m4/announce-interest-memory.md: rewritten from the shared-snapshot redesign to a session::Limits cap on lite-05/06 and moq-transport interest streams ([M] to [S]; Limits is #[non_exhaustive], so the field is additive).
  • quest/m4/binding-export-coverage.md: builds the list only from the exporter and gains quest/m3/ffi-exporter.md in ## Required.
  • quest/m4/browser-reset-reports.md: records approval to file at crbug.com and bugzilla.mozilla.org after each tracker's LLM policy check.
  • quest/m4/upstream-reports.md: drops the GStreamer faad, retina, and mr-boxington reports. The approved Chromium and Firefox reports and the already-filed Biome follow-through remain, so the quest stays (retitled "Report approved upstream bugs") and its approved filings get the same LLM-policy check.
  • quest/m4/quic-interop-listing.md: records approval to register once every supported case passes against two peers.
  • quest/m4/intra-refresh/recovery-boundary.md: deleted with quest rm; its decision lives in the epic README's shared decisions, and its leftover test and codec-review guidance moved into h264-import.md and h265-import.md. consumer-warmup.md notes why a stale warmup still leaves the transcoder's one group of history clean.
  • quest/m5/video-hardware.md: the Windows GPU probes and dx12 export tests split into the new quest/m4/windows-gpu-tests.md ([M], ## Hosts windows), linked Related both ways. mediafoundation_camera_texture stays in the parent, since the windows host has no camera. quest/m5/video-hardware-access.md points at the split.
  • quest/m9/stream-transports/relay.md: records reshaping moq_tokio::websocket::accept, split into the new quest/m2/websocket-accept-driver.md ([S], top-level m2 since no m2 epic fits), which relay.md now lists in Required. The new quest notes that SETUP is read with the driver running, so the driver comes back once QMux is established rather than at the end, and that its PR is ! with a migration note in Impact.
  • quest/m1/vr-demo-try.md: the Maintainer item becomes a nuts-rice: item naming the ask in moq-dev/moq#8781.
  • quest/m3/ffi-exporter.md: keeps its Waiting item and gains a Deferred 2026-11-15 line.

Decisions

All by the maintainer on 2026-11-15 in /quest-complete:

  • GPU CI runner: ✅ (a) a sandboxed runner container on the CI host with the NVIDIA device nodes and host driver libraries bound in (rejected: (b) a VFIO-passthrough VM).
  • cli-packaging while cpal has no runtime-ALSA release: ✅ (A) wait, and (D) fork cpal, alsa, and alsa-sys only once it becomes a blocker (rejected: (B) split by OS; (C) load-time libasound).
  • App capture off macOS: ✅ (A) window capture only; Source::App stays macOS-only (rejected: (B) compose app capture).
  • lite-05+ tail when a group is never accounted for: ✅ (c) wait out the max delay where set, and (a) error on lite-07 when none is set (rejected: (a) alone; (b) wait until close). ✅ A lite-05/06 tail with no max delay errors too (rejected: keep the 1 s constant).
  • Dial policy home: ✅ (a) a plain value type with no I/O in moq-net, through moq-webrtc's and moq-rtmp's #[non_exhaustive] configs (rejected: (b) moq-tokio; (c) its own crate).
  • QUIC driver ACK wait: ✅ keep only the trap tests and drop the public API (rejected: make it public; delete the quest). ✅ Move it to m4, since it is tests only (rejected: keep it in m3).
  • Resume redesign: ✅ delete it. Removed with quest rm: it only waited on that decision.
  • Legacy interest streams: ✅ a session::Limits cap for lite-05/06 and moq-transport now, deleted once lite-07 is the oldest version served (rejected: the shared snapshot redesign; deleting later with no cap).
  • FFI item list: ✅ only from the exporter, which joins Required (rejected: from UniFFI metadata now).
  • Browser reset reports: ✅ approved at crbug.com and bugzilla.mozilla.org, each after checking that tracker's LLM-report policy; a tracker that bans them gets a draft for the maintainer instead.
  • Unapproved upstream reports: ✅ (b) drop GStreamer faad, retina idle Pending, and mr-boxington eviction (rejected: (a) approve).
  • QUIC interop listing: ✅ register once every supported case passes against at least two peers.
  • Intra-refresh warmup: ✅ a hint nothing enforces (rejected: refuse the stream; a published-ahead bound; a fixed config bound). Since moq-dev/moq#8857 a longer sweep only shows stripes a little longer. That left recovery-boundary with no work of its own, so it was removed with quest rm and its Required links dropped from h264-import.md and consumer-warmup.md.
  • Video hardware: ✅ split the windows-host items into a new m4 quest (rejected: the lab; keep together).
  • WebSocket driver handoff: ✅ reshape moq_tokio::websocket::accept (rejected: an additive sibling), ✅ as its own m2 quest that relay.md requires (rejected: land it with the rust2 migration).
  • VR demo: ✅ ask nuts-rice, who started it in moq-dev/moq#6373, to try it in a headset.
  • ffi-exporter binding renames: deferred, not decided; the binding API changes anyway in the rust2 migration.

Impact

None: quest text only. No public API or wire change.

Validation

nix develop --command bash sh/quest.sh check (pinned quest 3122d76): quest: 488 documents ok on the second commit. Checked that Source::App already refuses off macOS, that moq-webrtc and moq-rtmp both depend on moq-net, that session::Limits is #[non_exhaustive], that every test the new Windows quest names exists, and that the comment asking nuts-rice exists.

Models

Size: M
Implementer: Claude Code, claude-opus-5-5, medium effort

(Written by Claude Opus 5.5)

Review

Rewritten by quest review after each round.

  • Head: b447433

  • Base: origin/main at 813f7a0

  • Risk: medium, the base has no RISK.md to name what is always high or customer-facing, so every change is medium

  • Classified: medium by default

  • Reviewer: Codex, gpt-6.1-sol, high effort

  • Session: 01a1e105-0994-7802-a6b9-677f80ee9120, round 2

  • Score: 10/10, The incremental quest changes consistently separate the WebSocket API work, clarify tail handling, and update references for the moved ACK validation plan; no actionable defects were found.

  • Rounds: 2; first round 0; findings 0: fixed 0, refuted 0, open 0

  • Findings: none

Settle sixteen Waiting items from /quest-complete, delete resume-redesign
and recovery-boundary, split the Windows GPU tests into their own m4 quest,
and defer ffi-exporter's decision.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@kixelated
kixelated marked this pull request as ready for review November 15, 2026 21:03
@coderabbitai

coderabbitai Bot commented Nov 15, 2026 •

Copy link
Copy Markdown
Contributor

Warning

Review limit reached

You'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 30 minutes.

Check out review usage here.

View limit details

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

Learn how review limits work.

Review configuration:

Run configuration

  • Source mode: Auto
  • Repository: Unavailable
  • Branch: Unavailable
  • File: No configuration file from this repository applied
  • Commit: Unavailable
  • Layers: Organization UI
  • Review profile: CHILL
  • Plan: Advanced
  • Run ID: 49091a3e-d33e-43ab-a242-2b38670692a6
Commits

Reviewing files that changed from the base of the PR and between c6991c4 and b447433.

Files selected for processing: 28

  • quest/m1/vr-demo-try.md
  • quest/m2/websocket-accept-driver.md
  • quest/m3/capture-linux.md
  • quest/m3/capture-windows.md
  • quest/m3/cli-packaging.md
  • quest/m3/cpal-alsa-runtime.md
  • quest/m3/ffi-exporter.md
  • quest/m3/gpu-ci.md
  • quest/m3/lite-tail-drop-count.md
  • quest/m3/webrtc-dial-policy.md
  • quest/m4/announce-interest-memory.md
  • quest/m4/binding-export-coverage.md
  • quest/m4/browser-reset-reports.md
  • quest/m4/intra-refresh/README.md
  • quest/m4/intra-refresh/consumer-warmup.md
  • quest/m4/intra-refresh/h264-import.md
  • quest/m4/intra-refresh/h265-import.md
  • quest/m4/intra-refresh/recovery-boundary.md
  • quest/m4/quic-ack-hook.md
  • quest/m4/quic-deadline.md
  • quest/m4/quic-interop-listing.md
  • quest/m4/resume-redesign.md
  • quest/m4/resume-trace-validation.md
  • quest/m4/upstream-reports.md
  • quest/m4/windows-gpu-tests.md
  • quest/m5/video-hardware-access.md
  • quest/m5/video-hardware.md
  • quest/m9/stream-transports/relay.md
  • Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

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

pullfrog[bot]
pullfrog Bot previously approved these changes Nov 15, 2026

@pullfrog pullfrog Bot left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

✅ No new issues found.

Reviewed changes — quest-only text edits recording the maintainer's 2026-11-15 answers. I spot-checked for dangling references only, not the full diff.

  • Deleted quests — resume-redesign.md and intra-refresh/recovery-boundary.md are gone, and nothing in quest/ or doc/ still references them.
  • New quest — quest/m4/windows-gpu-tests.md exists, split out of video-hardware.md.
  • Rewrites — the other edits record decisions in their quests' Plans (gpu-ci, cli-packaging, announce-interest-memory, quic-ack-hook and others).

Pullfrog  | View workflow run | Using claude-sonnet-5-5 | 𝕏

Split the websocket::accept reshape into its own m2 quest, end lite-05/06
tails with an error when no max delay is set, and move quic-ack-hook to m4.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

@pullfrog pullfrog Bot left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

✅ No new issues found.

Reviewed changes

Incremental review of the follow-up commit since the prior pullfrog review.

  • Split the WebSocket driver handoff into the new quest/m2/websocket-accept-driver.md, listed in relay.md Required.
  • Moved quic-ack-hook to quest/m4/ and updated its links in quic-deadline.md. No references to the old m3 path remain.
  • Tightened lite-tail-drop-count.md so a lite-05/06 tail with no max delay errors too.
  • Recorded the nuts-rice ask in vr-demo-try.md.

Pullfrog  | View workflow run | Using claude-sonnet-5-5 | 𝕏

@kixelated
kixelated added this pull request to the merge queue Nov 15, 2026
Merged via the queue into main with commit 2f0fa57 Nov 15, 2026
9 checks passed
@kixelated
kixelated deleted the quest-complete/wait-answers-2026-11-15 branch November 15, 2026 21:22
kixelated added a commit that referenced this pull request Nov 15, 2026
#8870 deleted quest/m4/intra-refresh/recovery-boundary.md on the maintainer's
answers; the deletion stands over this branch's reword.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant