Main now listens to camera.moveStart in IntelHUD's constructor (#1384), so
the fake camera in src/hudPageVisibility.test.mjs needs that event too, as
hudAltitudeDatum.test.mjs already has. No change to the fix itself.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Main now has #1282, which pauses periodic layer polling in a hidden page
with its own catch-up and listener cleanup, so this PR drops its duplicate
lifecycle.js change and test in favor of main's.
Review fixes:
- data.js passes panelAwareVisibility(document) (new in app/embed.js) as
the lifecycle's visibility target, so an inline panel whose host reports
it hidden keeps polling. The HUD's 250 ms readout follows the same
document.hidden && !isEmbeddedInline() rule as tools.js.
- Tests: embed.test.mjs (layers poll in an inline panel reported hidden,
pause without the flag) and hudPageVisibility.test.mjs (HUD readout).
Both use mock timers and clean up in finally.
- changelog.d/1344-inline-panel-visibility.md and a note in
docs/CURRENT-STATE.md.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A hidden tab shows nobody the result, so polled layers (flights, quakes,
weather...) skip their refresh while the page is hidden and catch up with
one refresh when it is shown again. The lifecycle takes an optional
pageVisibility service ({ isHidden, subscribe }) so src/data/lifecycle.js
stays free of browser globals; src/app/data.js supplies the document-backed
one. Without the service, refreshes run as before. Every place that stopped
a layer's interval now goes through one helper that also drops the
catch-up listener.
The HUD's 250 ms camera readout also skips hidden tabs.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
