# LaunchBenchmark report — POST-PHASE-5.2 (gate removal) snapshot
# Captured AFTER the Main.kt bootstrap-gate removal. NOTE: the slim
# benchmark does not actually exercise the gated code path (the gate is
# inside the App() composable, which this harness does not drive). The
# row remains a harness-floor sanity check; the gate-removal payoff
# manifests during real App() boot when no relay has connected yet and
# the previous 30s `withTimeoutOrNull` would otherwise idle the
# subscription.
# date            2026-06-18T07:57:51.519608Z
# jvm             21.0.9 Homebrew
# os              Mac OS X 26.5 aarch64
# cpus            10
# max-heap-mb     512
# git-sha         48726ee4df23bf97c6e302766af42e048c3060f0
# iterations      5 (after 2 warmup, discarded)
# fork-mode       single-JVM (cold-fork driver deferred)

t_account_logged_in            n=5  min=  0.21ms  q1=  0.48ms  median=  0.72ms  q3=  0.87ms  max=  1.15ms
t_first_event                  n=5  min=  1.17ms  q1=  1.34ms  median=  1.46ms  q3=  2.06ms  max=  2.10ms
t_n_events                     n=5  min=  1.51ms  q1=  1.76ms  median=  2.63ms  q3=  2.75ms  max=  3.55ms
# events-consumed per iteration: [23, 50, 15, 28, 23]