Skip to content

Add deterministic timing stress harness - #11

Merged
datashaman merged 1 commit into
mainfrom
agent/issue-5-ui-stress
Aug 10, 2026
Merged

datashaman merged 1 commit into
mainfrom
agent/issue-5-ui-stress

Conversation

@datashaman

Copy link
Copy Markdown
Owner

Closes #5.

What changed

  • add a deterministic fake-clock harness that compares a normal run with a 10-minute run containing repeated UI stalls and bridge reconnects
  • record scheduling watermark, late MIDI events, maximum engine lateness, and reconnect count in native engine diagnostics
  • expose those metrics through the bridge and application status panel
  • fail the harness on phase or timestamp drift, unpaired notes, duplicate or missing loop boundaries, backward watermarks, late events, or missing reconnects
  • document the 1 microsecond timing tolerance and print a diagnosable report in CI

Why

Transport and MIDI scheduling must remain native-engine concerns even when the web UI stalls, reloads, or reconnects. This provides a deterministic regression gate for that boundary without relying on wall-clock timing or a physical MIDI device.

Impact

Developers get a fast, reproducible timing report in the normal verification path. Users get visible runtime scheduling diagnostics without any change to transport or phrase behavior.

Validation

  • ./scripts/verify.sh
  • Debug CTest: 2/2 passed
  • Release CTest: 2/2 passed
  • 600-second stress simulation: zero phase drift, zero timestamp drift, zero late MIDI events
  • packaged Release app smoke test
  • rendered UI inspection with no browser warnings or errors
  • git diff --check

@datashaman
datashaman marked this pull request as ready for review August 10, 2026 07:29
@datashaman
datashaman merged commit ed47167 into main Aug 10, 2026
2 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Prove transport stability under simulated UI stress

1 participant