Intent: /map's Server-Timing names its phases in test/map-cache.test.ts goes red in pulls that never touch it (pull #5178), so good work is parked and the board shows red noise.
Done:
[ ] the pull names the cause (wall clock, a real sleep, a shared port or directory, a fake clock mixed with Date.now(), a budget measured end to end instead of in the code it guards)
[ ] bun test --rerun-each 30 test/map-cache.test.ts passes, and also passes with other test files running beside it (bun test test/map-cache.test.ts test/apps.test.ts). The pull says how it was checked
[ ] the test still fails when the behaviour it guards is broken; what it promises about the product is unchanged. A time or resource budget may only move to what the code itself spends, never just get wider
Files: test/map-cache.test.ts; the source only if the flake is a real race there (then the pull says so).
Hint: it asserts the phases sum to within 0.4 ms of the total (line 257) and a p95 under 5 s; check the sum against the rounding Server-Timing actually uses, not a fixed 0.4 ms.
Intent:
/map's Server-Timing names its phasesin test/map-cache.test.ts goes red in pulls that never touch it (pull #5178), so good work is parked and the board shows red noise.Done:
bun test --rerun-each 30 test/map-cache.test.tspasses, and also passes with other test files running beside it (bun test test/map-cache.test.ts test/apps.test.ts). The pull says how it was checkedFiles: test/map-cache.test.ts; the source only if the flake is a real race there (then the pull says so). Hint: it asserts the phases sum to within 0.4 ms of the total (line 257) and a p95 under 5 s; check the sum against the rounding Server-Timing actually uses, not a fixed 0.4 ms.