pm/testing baseline: 98 runs, 7 regressions (v5.15-rc1)