[AlexSchmidt1999]

Today still publishes partial day results between database reads. Rest, HR, steps, Effort, and sleep can each trigger another dashboard update during one load. The existing assignments now run together after the final read and the existing cancellation/day check. A superseded load leaves the displayed state intact. Queries, formulas, cache contents, and the chart renderer stay the same.

The local build-451 archive looks better than build 450. In Today windows after the first ten seconds, with sync inactive, late callbacks fell from 1.87% to 0.74%. The worst interval fell from 517.9 to 98.5 ms, over roughly three sampled minutes per build. These are ordinary-use captures with different interactions. They do not prove that the previous training fix caused the improvement. The new summaries do not identify active workouts, and the four-second Trends sample is too short for a useful comparison.

A reduced native SwiftUI chart probe ran five times in optimized mode. Six staged publications produced seven body evaluations each time, including initial layout. A final batched publication produced two. This tests coalescing behavior, not production Today evaluation counts or iPhone scrolling speed. The local usage report has the measurements and limits. Raw archives and preference copies remain local.

Verification:

• The macOS app compiled and 75 focused Release tests passed for Rest freshness, HR zoom, hero layout, explainability, and workout sample processing. The English text fixtures were run with -testLanguage en. The first run on this German host had 15 language-only expectation failures.
• All 338 Python tool tests passed with canonical temporary paths on macOS.
• The iOS Release build passed, including widgets and Watch components.
• Source hygiene, translation coverage, source policy, and the four-bundle binary policy passed.

Apple UI only. No new dependencies, settings, capture mode, BLE commands, or storage changes. This is a fork test follow-up stacked on the workout sample optimization. Build 452 includes both changes. An iPhone comparison of this new batching fix is still pending.

[AlexSchmidt1999]

Replaced by #31, which retains this draft's commits and includes the tested build-454 UI update fix. The current device comparison will be recorded there.