Repository navigation
quest: plan Nightly's Windows platform job passing - #8405
Conversation
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 7 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: f5d83d1
Base: 66a0b1f
[P3] Remove already-fixed causes from the implementation plan — quest/m0/nightly/windows.md:16–19
Both listed failures were fixed by #8258, already present in this PR's base: rs/moq-vaapi/src/lib.rs has #![cfg(target_os = "linux")], and memory.rs gates both Duration and run on test-support. Presenting these as current causes sends the future implementer back into solved work. Mark them as resolved history and update the parent README's requirement to focus on the remaining Windows failures.
The cited newer Windows job does fail on the unused encode_all helper in moq-video, so investigating that diagnostic on main remains useful, as the plan requests.
Verification limits: GitHub-only review of the two planning files, relevant workflow/source/history, and cited job logs. No builds, tests, quest check, or new workflow runs were executed; current main's runtime outcome was not reproduced. COMMENT only, not approval.
![pullfrog[bot]](https://avatars.githubusercontent.com/in/1768019?s=60&v=4){kind=small}
There was a problem hiding this comment.
ℹ️ No blocking issues — one stale section in the new quest's Plan.
Reviewed changes
quest/m0/nightly/windows.md(new) — plans makingplatform.yml's Windows job pass onmain, listing three known causes and the reproduction route.quest/m0/nightly/README.md— links the new child from## Required, beside the macOS sibling.
The goal, m0 placement, and the m0/nightly/ location all match the questline rules and the macOS quest. The only concern is that two of the three bullets in the Plan no longer reproduce on main.
deepseek-v4.1-flash-1017 (free via Pullfrog for OSS) ｜ 𝕏
| - `moq-vaapi` is not gated to Linux under `--all`, so the Windows build tries | ||
| to compile it. | ||
| - `moq-relay`'s memory integration test leaves items unused on Windows, | ||
| which `build.warnings = deny` turns into errors. |
There was a problem hiding this comment.
Both of these were fixed on main by moq-dev/moq#8258, merged 2026-11-08: rs/moq-vaapi/src/lib.rs:20 now has #![cfg(target_os = "linux")], and rs/moq-relay/tests/integration/memory.rs:8-9 gates the Duration import (with run gated too). The 2026-11-09 dispatch cited just below confirms it: the compile and moq-relay test steps pass there, and only the moq-video warning fails the job, so these bullets send the implementer to reproduce work that is already done.
Technical details
# Scope the Plan to the remaining Windows failure
## Affected sites
- `quest/m0/nightly/windows.md:16-19` — the two stale causes.
- `quest/m0/nightly/README.md:54` — the summary repeats the same two fixed items and omits the `moq-video` cause.
## Required outcome
- The Plan names only causes that still reproduce on a dispatch of this branch, and records #8258 as the reason the first two no longer do.
## Evidence
- `rs/moq-vaapi/src/lib.rs:20` — `#![cfg(target_os = "linux")]`, added by #8258.
- `rs/moq-relay/tests/integration/memory.rs:8-9` — `Duration` import behind `test-support`, also from #8258; `run` at line 367 is gated the same way.
- Run 42635646987 on `6c0d45e9`: `just rs windows --all` and the `moq-relay` nextest step pass; only `just rs test -p moq-video --no-default-features --features render` fails on `encode_all` at `rs/moq-video/src/encode/backend/mod.rs:84`.| - `moq-vaapi` is not gated to Linux under `--all`, so the Windows build tries | |
| to compile it. | |
| - `moq-relay`'s memory integration test leaves items unused on Windows, | |
| which `build.warnings = deny` turns into errors. | |
| - `moq-vaapi`'s Linux gating and `moq-relay`'s unused items already landed | |
| in [moq-dev/moq#8258](https://github.com/moq-dev/moq/pull/8258); confirm | |
| they stay fixed on the dispatch, but they need no work. The remaining | |
| cause: |
Plans the follow-up moq-dev/moq#8251 suggested:
platform.yml's Windows job is red onmain, which fails every Platform dispatch, release PR, and Nightly'splatform / Windows, and hides real Windows breaks.Changes
quest/m0/nightly/windows.md[S] under Nightly passes on main, beside its macOS sibling, since Nightly callsplatform.yml. Its Plan lists the three known causes:moq-vaapinot gated to Linux under--all, Windows-only unused items inmoq-relay's memory test, and adead_codewarning inmoq-video's encode backend seen on the latest dispatch.Decisions
m0/nightly/rather than them0/ci/the prompt named, because Nightly runsplatform.ymland the macOS job's quest already lives there.Checks
quest check: pass.Impact
Public API: none. Wire: none.
Models
Size: XS
Implementer: Claude Code, Opus 5.5, high effort
(Written by Opus 5.5 with high effort)
🤖 Generated with Claude Code
Review
Rewritten by
quest reviewafter each round.01a1c33f-f03c-76a3-b584-93555b07678e, round 1The change only adds a scoped Windows CI follow-up plan and links it from the parent quest, with no executable behavior changes or actionable defects.