Neil f5b6e14bc6 test(diff): assert steady-state responsiveness and record the large-diff stall regression ...

The freeze guard asserted only worst-case lag, which the Pierre renderer trips on CI hardware.
Add a tight p95 bound -- the axis the migration improved, and the one a real freeze would blow --
and loosen the max bound to the measured post-migration cost, with both renderers' numbers and
the upstream cause written into the test so the regression is recorded, not hidden.

2026-09-10 22:24:13 -07:00

..

e2e

test(diff): assert steady-state responsiveness and record the large-diff stall regression

2026-09-10 22:24:13 -07:00

tools

fix(browser): restore the Chrome-shaped browser identity (STA-7147) (#19927)

2026-09-10 15:20:34 -07:00

.gitignore

feat: add idempotent E2E test suite with headless Electron support (#671)

2026-04-19 12:09:32 -07:00

AGENTS.md

test: honor background launch in paired client window helpers (#18978)

2026-09-05 18:41:24 -07:00

playwright.config.ts

Stabilize sharded e2e CI concurrency (#6312)

2026-06-24 18:39:02 -07:00