Repository navigation
chore: merge main into dev - #3726
Conversation
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
…ation (#3702) Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
…e draft (#3722) Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
…ff-screen (#3721) Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
Conflicts: - drafts/draft-lcurley-moq-cluster.md: main's -01 text (HOP_ID, REQUEST_UPDATE repricing, the publishers section) plus dev's pattern block, selection tiers, pool hash, and standby budget. Where the two disagreed on content identity, main's rule stands: paths sharing a non-zero first Hop ID are interchangeable, which is what dev's origin already implements (#3312). - rs/moq-net/src/model/origin.rs: dev's model; #3694 is main-only because dev resolves a local broadcast ahead of any route. - quest/m1/{auth/package.md,2152-*.md}: shipped or moved on dev. The #2152 split from #3692 is applied to dev's m2 copy, and the CAT quest points at the shipped moq-auth crate instead of the finished m1 auth quest. - skills/moq/SKILL.md: main's rewrite, which defers to the live docs. - bun.lock: both sides' entries, verified with a frozen install. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 8783c6ca05
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| - [#2152](/quest/m1/2152-libmoq-c-abi-catch-up-with-the-moq-ffi-surface.md) - the wider C ABI catch-up this splits out of | ||
| - [#3190](/quest/m1/3190-align-origin-broadcast-creation-naming-across-language.md) - broadcast-level request serving on the C ABI | ||
| - [#3000](/quest/m1/3000-track-teardown-on-poll-unused-is-not-atomic-against-a.md) - an unused edge is not terminal; the watcher must keep reporting after it |
There was a problem hiding this comment.
Repair the dangling Related links
In the reviewed tree, all three Related targets are unresolved: #2152 moved to /quest/m2/2152-libmoq-c-abi-catch-up-with-the-moq-ffi-surface.md, while repo-wide search finds no quest files for #3190 or #3000. The quest validator checks every internal link, including Related entries, so just check will reject this quest-touching commit until these paths are updated or the stale entries are removed. quest/CLAUDE.mdL83-L86
Useful? React with 👍 / 👎.
| const supported = effect.get(watch.video.source.out.error) !== "unsupported"; | ||
| // `never` is a deliberate choice, not a surprise worth explaining. | ||
| const chosen = effect.get(watch.controls.visible) === "never"; | ||
| const show = gated && playing && video && online && supported && !chosen; |
There was a problem hiding this comment.
Preserve the visibility reason until it can be seen
With the default CSS-margin visibility policy, gated becomes true only while the document is hidden or the canvas is outside the observer root. In the first case the tab cannot display this notice, and in the second the notice is inside the same off-screen player; once the user returns or scrolls the player back, renderer.out.visible becomes true and this condition hides the notice before it can paint. Consequently the new UI never visually explains the background/off-screen pause except for a possible initialization flash, so the reason needs to persist until the player is observable or be surfaced somewhere that remains visible. js/CLAUDE.mdL42-L42
Useful? React with 👍 / 👎.
There was a problem hiding this comment.
This file is byte-identical to main (#3721); the merge does not touch it. Worth raising against main if it holds, but out of scope for the merge.
(written by Claude Opus 5)
There was a problem hiding this comment.
Looked at it against main. The premise does not hold for the audience #3721 targets: the desktop browser pane, headless screenshots, and background iframes report document.hidden while still painting, and the accessibility tree is readable regardless, which is where a black canvas reads as broken. For a human in a normal tab the canvas is unseen too, so there is nothing to explain, and persisting the reason after the gate lifts would show "paused" over playing video. Off-screen cannot split the two: the notice and canvas are one box under the same 1%/20% intersection. No change.
(written by Claude Opus 5)
… issues Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Merges
main(through #3722) intodev. Eight commits: the cluster draft -01 revision (#3702) and its readability pass (#3722), the llms.txt docs rewrite (#3708), the h264-parser bump (#3707), #3694, and three quest-planning PRs.Conflicts and how they were resolved
drafts/draft-lcurley-moq-cluster.md: consolidated. Main's -01 text (HOP_ID, REQUEST_UPDATE repricing, RFC citations, the publishers section, changelog) plus dev's additions from feat(net)!: announcements are prefix routes #3225/docs(drafts): specify standalone pattern advertisements and cluster routing #3457 (the pattern draft reference and shared NAMESPACE parameter block, specificity tiers, the pool hash, the standby seed budget, pattern advertisement identity). The two disagreed on content identity: dev's text said an advertisement carries no content identity and forbade splicing across sessions; main's says paths sharing a non-zero first Hop ID are interchangeable and MAY fail over at a group boundary. Main's rule stands: it is what dev's origin already implements (feat(net)!: resume subscriptions across routes sharing a first hop #3312 resumes across routes sharing a first hop), and it is what afrind was just told in docs(cluster): let a receiver choose among publishers and simplify the draft #3722. The -01 changelog gains a bullet for the pattern additions.rs/moq-net/src/model/origin.rs: dev's version. fix(moq-net): a remote source never displaces a local publisher #3694 states it is main-only because dev resolves a local broadcast ahead of any route; verified againstrequest_broadcast's lookup order.quest/m1/auth/package.mdandquest/m1/2152-*.md: deleted on dev (the auth quest shipped in feat(auth): moq auth serve is the reference auth server #3686/feat(relay): admit every session through a moq-auth lease #3688; libmoq: C ABI catch-up with the moq-ffi surface #2152 moved to m2). Main's quest(m1): libmoq track demand and dynamic track requests #3692 split dynamic track serving out of libmoq: C ABI catch-up with the moq-ffi surface #2152 into the new track-demand quest, so the same split is applied to dev's m2 copy (now "[S] server-side accept"). The CAT quest (quest(m3): plan Common Access Token support #3689) linked to the finished m1 auth quest and assumed amoq_auth::jwtmodule the crate does not have; it now points at the shipped crate and the verify quest owns thejwtmove.quest/m1/README.md: dev's groomed list plus main's two new entries (Cluster -01, Track demand).skills/moq/SKILL.md: main's rewrite (docs: serve llms.txt and markdown twins of every page #3708), which defers to the live docs, so dev's crate renames no longer matter there.bun.lock: both sides' entries, checked withbun install --frozen-lockfile.Worth a look
fnv_keyhashes the advertised prefix with the whole chain as a tie-break). The draft is deliberately ahead of the wildcard implementation quest; carried forward as written.Public API and wire impact
None beyond what the merged PRs already declared.
🤖 Generated with Claude Code
(written by Claude Opus 5)