Repository navigation
test(media): keep the mute-glide check off playback and transitionrun timing - #917
Conversation
… timing On CI's Linux WebKit the 2s review clip autoplays while the hover waits on hidden controls, ends, and restarts when the test mutes it. The knob's transition then never stays paused and the poll times out. Pause the clip for this test, and pause the stretched transition from the poll instead of a one-shot transitionrun listener. Co-authored-by: claude00 <fb8b76a01257932b8ca35060862c29bd09b2137638573731c2d5074fc02c5385@buzz.block.builderlab.xyz> Co-Authored-By: Claude Opus 5.5 <[email protected]> Signed-off-by: Bradley Axen <[email protected]>
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. |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: d5c16cbe0c
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".
| }); | ||
| // Playback hides the controls and, on CI's Linux WebKit, restarts the ended | ||
| // clip when it is muted. Keep the clip still so only the volume UI moves. | ||
| await dialog.locator("[data-review-media]").evaluate((video) => { |
There was a problem hiding this comment.
This commit is authored and committed as Codex, but its message contains neither a Signed-off-by nor a Co-authored-by trailer. Any PR containing it will fail the hosted DCO check, and an agent-authored commit must also credit its human collaborator, so add both trailers before publishing the commit.
AGENTS.md reference: AGENTS.md:L194-L200
Useful? React with 👍 / 👎.
wesbillman
left a comment
wesbillman
left a comment
There was a problem hiding this comment.
No changes requested: the test still exercises the production transform transition and retains keyboard/reduced-motion assertions, with cleanup restoring the duration token. Hosted Linux logs show all 16 cases in this file passing across Chromium/WebKit on the PR test merge; overall CI was incomplete at the snapshot, and one run does not establish flake elimination. Star Lord’s automated source review via Wes’s account; no tests or app executed locally.
Head: d5c16cbe0c259d7503ebc49d28945fd8c2b9705a · Base: 55b71bf1db0ff6f5700dafc0d9f347a7d52ec8ed.
wesbillman
left a comment
wesbillman
left a comment
There was a problem hiding this comment.
🤖 No changes requested. Carl, an automated reviewer, commenting via Wes’s GitHub account.
The test still proves a real transform transition on the production knob: the stretch resolves on .volumeKnob itself (VideoPlayer.module.css:240-242), an absent animation still fails the poll, and the keyboard and reduced-motion 0s assertions are unchanged. Pausing playback is the right scope, since this test checks volume CSS, not playback. Cleanup finishes the animation before removing the override, so the later assertions don't see a stray transition.
- The Codex DCO/trailer comment is incorrect. The commit carries both
Co-authored-byandSigned-off-by, and DCO Check passes. - Non-blocking follow-up: if muting an ended clip really restarts it from 0 on Linux WebKit (likely WebKit re-running
autoplayonce the clip is muted), users would see that too. This PR now steps around it, so it may deserve its own issue. - Validation: all CI on this head is green, including 6/6 WebKit shards. One green run doesn't prove the flake is gone. Watching main afterwards, as the PR plans, is the real exit check. No local runs.
Head: d5c16cbe0c259d7503ebc49d28945fd8c2b9705a
Summary
media-controls.spec.mjs› "mute glides the volume knob and keyboard changes stay immediate" failed on webkit in 14 CI runs over the last day (5 onmain). It is now the most frequent browser flake on main.The CI traces from the 23:17 and 21:22
mainruns show the same sequence each time:hoverretries on "<video>intercepts pointer events" for about 3.4s.This PR changes only the test:
transitionrunlistener, and stretches--duration-faston the knob while it does, so a slow runner can't finish the 120ms glide first. The override is removed infinally.No assertions were removed or loosened.
Testing
pnpm test:browser --project webkit --project chromium --no-deps tests/browser/media-controls.spec.mjspasses: 16/16 locally.--repeat-each 10across both engines passes: 20/20.🤖 Generated with Claude Code