Fix the flaky mine-in-hold player test

[nulltask]

Summary

devel CI failed in packages/player/src/index.test.ts ("mine notes inside an active LN hold deal silent gauge damage"): expected 16 to be less than 16.

The implementation is right; the test was timing-sensitive. It schedules real-time setTimeout taps and relied on a single tap at exactly 750 ms to resolve the mine. On a loaded runner that tap can fall outside the mine's window, so the mine expires unpressed via markExpiredLandmines and, as in beatoraja, an unpressed mine deals no damage. The gauge then matches the mine-free baseline.

The test now taps every 50 ms across 600–950 ms, so one tap lands inside the window even when timers drift. Re-taps while the LN is held are neutral (no verdict, no empty POOR), and the baseline replays the same schedule, so the verdict comparison is unchanged.

Test plan

• The test passes five times in a row.
• The whole packages/player/src/index.test.ts passes (95 tests) with four copies running in parallel to load the machine.

🤖 Generated with Claude Code

[cloudflare-workers-and-pages]

Deploying with    Cloudflare Workers

The latest updates on your project. Learn more about integrating Git with Workers.

Status	Name	Latest Commit	Preview URL	Updated (UTC)
✅ Deployment successful! View logs	be-music-player-demo	b441e11	Commit Preview URL Branch Preview URL	Oct 11 2026, 11:38 AM

[github-actions]

Exports Benchmark

• Base SHA: bdb740514028
• Head SHA: b441e11cf5eb
• Comparable cases: 92
• Regression threshold: 8.00%
• Overall verdict uses the median change across cases.
• Per-case lists compare median ops/s, not mean.
• Base runs: median of 3
• Head runs: median of 3

Summary

Metric	Value
Overall	unchanged
Median change	-0.51%
Mean change	-0.49%
Cases improved (>= threshold)	0
Cases regressed (<= -threshold)	1
Cases unchanged	91
Cases excluded (sub-timer-resolution)	211
Head benchmarked cases	303
Head skipped cases	21

Top Regressions

API	Base median ops/s	Head median ops/s	Change
audio-renderer.writeAudioFile	3808.20	3109.80	-18.34%

Top Improvements

No improvement over threshold.

Excluded (sub-timer-resolution)

Per-call latency at or below 0.001ms on at least one side — the reported time is measurement noise, not the case's real cost, so no percent change is shown.

API	Base median ops/s	Head median ops/s
audio-renderer.createTimingBuildContext	1923076.94	1923076.92
chart.compareEvents	10416666.69	10416666.69
chart.createBeatResolver	20833332.60	20833332.60
chart.eventToBeat	13888889.10	13888889.10
chart.exWavVolumeCentibelsToLinearGain	15625000.04	15625000.04
chart.getMeasureBeats	20833332.60	20833332.60
chart.isBmsBgmVolumeChangeChannel	17857142.61	17857142.61
chart.isBmsDynamicVolumeChangeChannel	17857142.61	17857142.61
chart.isBmsKeyVolumeChangeChannel	17857142.61	17857142.61
chart.isBmsLongNoteChannel	20833332.60	20833332.60
chart.isLandmineChannel	20833332.60	20833332.60
chart.isPlayableChannel	20833332.60	20833332.60