Repository navigation
test: fix the last nightly lifecycle failure and a layout test that measured mid-animation - #2108
Merged
Merged
Conversation
…cleanup "cleans up the remote connection" closed the dashboard window the wizard ran in and then waited for the remote session to end. Before launch-on-card (#2074) that dashboard window became the remote connection's window, so closing it ended the session. Now a build started from the dashboard opens in its own window, the dashboard window stays a dashboard, and the remote session kept running: the wait timed out on both passes of the nightly. The failure screenshots show the "Remote Loopback" window still open next to the local ComfyUI window. The test now closes the window that shows the remote connection, found by the webContents the previous test drove, as well as the dashboard window, and waits for both to be gone before it opens a dashboard from the local ComfyUI window. The dashboard window's id is renamed to say what it is.
|
No actionable comments were generated in the recent review. Walkthrough
Merge Risk:
|
…s stopped animating The Test check failed on this PR in releaseChannelSwitchLayout.test.ts, "keeps the dialog and every row in place from the click until the other channel has loaded". Every box in the baseline was 0.98 of its later size and 4px lower: the baseline was taken at the first keyframe of the setup dialog's pop-in animation (scale .98, translateY 4px, 160ms), and the layouts sampled afterwards were at rest. finishEntranceAnimations took one snapshot of document.getAnimations() and awaited those. An animation that starts after the snapshot, or one that is cancelled and restarted so its finished promise rejects, let the helper return while the dialog was still moving. It now waits frame by frame, awaits whatever finite animation is running, and returns only after three consecutive frames with none.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Follows #2092 and #2101, both merged. Two test-only commits: the last failure in the nightly lifecycle chain on
main, and a unit test that failed this PR's own Test check.1. Lifecycle: the remote connection's window
Lifecycle Tests dispatched on
mainat9f5a178f(run 43651779492, Windows): 66 passed, 1 failed, 2 did not run.@sec-bootwindowtests pass on both passes, including "restart clicked during a FRESH chooser boot", which fix(picker): offer Restart from the dashboard for a build starting with no window #2101's picker change targets.e2e/lifecycle.test.ts:2649, "cleans up the remote connection: window closed, record untracked, local install untouched", fails on both passes. The two delete tests after it did not run.Root cause: the test is wrong. It closed the dashboard window the wizard ran in and then waited for the remote session to end. Before launch-on-card (#2074) that dashboard window became the remote connection's window when its tile was clicked, so closing it ended the session. Now a build started from the dashboard opens in its own window and the dashboard window stays a dashboard. The remote session kept running and the wait timed out. The failure screenshots show the "Remote Loopback" window still open next to the local ComfyUI window.
Fix. The cleanup closes the window that shows the remote connection, found by the webContents the previous test drove, as well as the dashboard window. It waits for both to be gone before it opens a dashboard from the local ComfyUI window, so one ComfyUI window is left, as the rest of the test expects.
_remoteWindowIdis renamed_remoteDashboardWindowIdto say what it holds.2. Unit: the release-channel layout test
The Test check failed on the first push here in
releaseChannelSwitchLayout.test.ts(from #2097), which this PR had not touched. Every box in its baseline was 0.98 of its later size and 4px lower: the baseline was taken at the first keyframe of the setup dialog'spop-inanimation (scale .98, translateY 4px, 160ms), and the layouts sampled afterwards were at rest.finishEntranceAnimationstook one snapshot ofdocument.getAnimations()and awaited those. An animation that starts after the snapshot, or one that is cancelled and restarted, let it return while the dialog was still moving. It now waits frame by frame, awaits whatever finite animation is running, and returns only after three consecutive frames with none.I did not reproduce the failure locally; it depends on timing. The diagnosis comes from the numbers in the CI failure.
Validation
pnpm run typecheck,pnpm run lint,pnpm run format:checkpass. The layout test file passes locally three times in a row.mainsince 9 Nov. The second commit only touches a unit test.