Repository navigation
RAL-205: browser tests hold up under load - #294
Merged
Merged
Conversation
RAL-203's local/check failed in a_segfault_at_start_is_tried_once_more_ and_a_second_fails at load ~7: after the stand-in's crash at start, the retry's real Chrome hadn't started before the run's 45 s timer fired. - tests/web.rs: one real Chrome at a time. The 12 browser runs ran in parallel test threads, so several flatpak Chromes started at once. A process-wide lock is held only around the browser's run, not the viewer's setup, and a test that panicked holding it doesn't poison it. - tests/browser/sign-in.mjs: the retry after a crash at start gets a fresh 45 s budget instead of what the crashed start left, and a 90 s cap from the script's start ends any run, so a Chrome that never starts holds CI for two budgets at most. Nothing the tests prove changes: one stated retry, a second crash fails, nothing of the browser survives. Agreed with upbeat-anole (web). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Sent back by upbeat-anole (5522). The retry's timer is set only while the first 45 s timer is pending, so the retry's deadline is always under 90 s from the start, and the cap's finish would only ever no-op. The comment now says why a run ends within two budgets without it. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
RAL-205 (claim 5506), from the watchman, after RAL-203's local/check failed in
a_segfault_at_start_is_tried_once_more_and_a_second_failsat load ~7. After the stand-in's crash at start, the retry's real Chrome hadn't started before the run's 45 s timer fired. Agreed with upbeat-anole, since web is their side.tests/web.rs: one real Chrome at a time. The 12 browser runs ran in parallel test threads, so several flatpak Chromes started at once. The lock is held only around the browser's run, not the viewer's setup. It's taken withunwrap_or_else(|p| p.into_inner()), so a panicking test doesn't poison the rest.tests/browser/sign-in.mjs: the retry after a crash at start gets a fresh 45 s budget, and a 90 s cap from the script's start ends any run.docs/roadmap.md: the RAL-205 entry.The assertions are unchanged: one stated retry, a second crash fails, and nothing of the browser survives.
Evidence at aa78c78:
cargo fmt --checkandcargo clippy --all-targets -- -D warningsare clean.cargo test --no-fail-fastpasses 763 tests, 0 failures, with real Chrome (RALLY_TEST_BROWSER_REQUIRED=1).cargo test --test webpasses 54/54 twice at load average 10.9–12.5, in 43 s and 39 s (the failing run on main took 52 s).🤖 Generated with Claude Code