You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Flaky test: markdown.test.ts 'looks once for the end of a formula that never closes' exceeds its timing ratio on a loaded runner #2084
Seen on #2081, which changes only scripts/chat-virtual-browser-qa.ts, in Fast checks (run 40565565799, job 122405886230, self-hosted runner):
error: expect(received).toBeLessThan(expected)
Expected: < 30
Received: 32.17235984354594
at linear (src/lib/markdown.test.ts:273:19)
at <anonymous> (src/lib/markdown.test.ts:328:20)
(fail) text no one wrote by hand > looks once for the end of a formula that never closes [301.29ms]
linear() passes when the best of three time ratios (large input / small input) is under 30. On a runner shared by many jobs, all three rounds came out above 30, though the parser hadn't changed. The same file passed locally in bun run check fast on the same head.
Expected: the growth check doesn't depend on how busy the runner is (for example, it counts the parser's work instead of timing it, as #1974 did for another case).
Activity
fix(tests): name the omo fixture with a shell herdr can read, and tim…
fix(tests): measure what the code controls in three tests that failed…
fix(codex): count a loaded thread one read of the kernel's lock table…
fix(files): keep a large scratch file's length when many wait to be c…