Repository navigation
fix(relay,video): unbreak the Windows platform build - #6821
Conversation
The relay's module-wide clippy::disallowed_methods expectation is only fulfilled by the Unix-only signal test, so Windows reports it unfulfilled and the denied-warnings test build fails. Move it onto that test. moq-video compiled code on Windows that nothing there calls: I420::from_bgra (only PipeWire uses it), pump::Read::Idle (only X11 returns it), and Applied::new under default features (NVENC and VA-API backends are Linux-only). Gate each to where it is used. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
No actionable comments were generated in the recent review.
|
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. |
kixelated
left a comment
There was a problem hiding this comment.
Automated review by review (OpenAI)
Reviewed commit: 910417e
No actionable bugs found. The narrower lint expectation is attached to the Unix-only test that actually reads Instant::now(), and the video cfg changes match the existing PipeWire/X11 callers and platform-gated encoder backends. This is a suitably small build fix rather than a broad warning suppression; no runtime or wire change is needed.
Verification: static inspection of the diff, surrounding tests, caller sites, backend gates and the apple alias. I did not run Clippy, cross-compilation, tests or the platform workflow.
The module-wide expectation also covered the stun-gated test's tokio::time::timeout, so the all-features shape now needs its own allowance. Also take rustfmt's spacing on the gated match arm. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
kixelated
left a comment
There was a problem hiding this comment.
Automated review by review (OpenAI)
Reviewed commit: ec14978
No actionable bugs found in the update since 910417e. The STUN test also calls the explicitly disallowed tokio::time::timeout; its new function-scoped expectation correctly restores that feature-enabled lint shape without restoring the overly broad module expectation. My earlier review missed this second caller. Keeping both exemptions with their gated tests is the right direction; the pump change is formatting-only.
Verification: static comparison and inspection of the test and crate Clippy configuration. I did not run Clippy, builds or tests.
Nightly's Platform Windows job fails on main (run):
#[expect(clippy::disallowed_methods)]onrelay::testsis only fulfilled by the#[cfg(unix)]signal test'sInstant::now()and thestun-gated test'stokio::time::timeout, so a Windows build withoutstunreports it unfulfilled. Moved onto those two tests.cargo checkdoes not deny warnings). Gated each to its users:I420::from_bgra: PipeWire only.pump::Read::Idle: only X11 returns it.Applied::new's dead-code allowance now matches the backend gates (NVENC and VA-API are Linux-only, VideoToolbox isapple); a default-feature Windows build failed-D warningson it.The relay and
from_bgrafixes mirror the ones in moq-dev/moq#6805, written against current main (#6805's relay change misses thestuntest, which the all-features clippy shape catches).Impact
No public API or wire impact.
Checks
cargo clippy --target x86_64-pc-windows-msvc -- -D warningsfor moq-video (default,--no-default-features,capture,capture,render). moq-relay does not cross-build here (ring and aws-lc-sys need an MSVC C toolchain); covered by the Platform dispatch below.just check; clippy-D warningsfor moq-relay (default, all features) and moq-video (default, no default,capture,capture,pipewire);relay::testswith all features.(Written by Claude Opus 5.5)
🤖 Generated with Claude Code