Repository navigation
fix(ci): switch the release e2e back to the web window after the native first launch - #816
Merged
Merged
Conversation
…ve first launch The first launch of a release opens the native window since #809, and the macOS and Windows end-to-end checks expected the web one on it.
… back On the slower Intel runner the switch came before the native window had attached, so lich had no window of it to close and both stayed open. lich now logs the attach, and the macOS and Windows checks wait for it.
…d a freed port, serially #813 ran them in parallel. A test that writes a script and execs it fails with ETXTBSY when another test forks while the script is still open for writing (golang/go#22315), and the port-retry test can lose its freed port to a parallel test's 127.0.0.1:0. Measured on the package: 1 failure in 15 runs before, 0 in 25 after.
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The
v0.70.0release run failed on both macOS jobs and on Windows: since #809 the first launch of a release opens the native window, and the end-to-end checks expected the web one on it.mac-e2e.sh: asserts the first launch opens the native window, waits for it to attach, then switches back the way Settings does it (store.SetSetting window.kind=web,system.OpenWindow web), checks the native window closed and goes on with the web checks as before.Windows step in
release.yml: the same switch. On the runner the native window exits before it attaches and lich falls back to the web one, so it waits for the attach or the exit instead of asserting the native window opened.internal/system:AttachNativelogsnative window attached, the counterpart of the existingnative window exited. The first validation run showed why the checks need it: on the slower Intel runner the switch came before the native window attached, lich had no window of it to close, and both stayed open.internal/terminaltests: the validation run's linux job failed ontext file busyinTestSessionAccountReadsWhatTheWrapperExecsInto, a flake from test(terminal): run the tests that touch no process-wide state in parallel #813. A test that writes a script and execs it hits ETXTBSY when a parallel test forks while the script is still open for writing (os: StartProcess ETXTBSY race on Unix systems golang/go#22315);TestListenRetriesUntilPortFreesWithoutWaitMarkercan lose its freed port to a parallel127.0.0.1:0. Those 32 tests run serially again; nothing they assert changed.go test -count=N ./internal/terminal/: 1 failure in 15 runs before, 0 in 25 after.Validated with
workflow_dispatchon this branch (thereleasejob is skipped off a tag): run 39069991006 has the three end-to-end jobs green; its linux job failed on the flake above. After merge,v0.70.0is re-tagged on the merge commit; nothing was published from the first tag.