An unsampled base sits at ellipsoid height 0. Over high terrain seen at an
angle that point projects below the screen while the ground above it is in
view, so the gate skipped ground the camera could see and never sampled it.
The gate now tests loaded globe terrain when there is some, and otherwise
the whole span of heights a sample can accept (0 to 9,000 m).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The HUD builds place context for its AI summary whenever the camera
settles, and each build paid for up to four reverse geocodes plus a Nearby
Places search, even when no OpenAI key was configured to use them.
- Ask GET /api/openai/hud-summary/status once; with no summary configured,
or after an unconfigured reply, make no place lookups for the page.
- Reverse-geocode at most two points: the view center and one viewport
point.
- Reuse a lookup until the view center moves a quarter of the visible
width, the camera's distance to it doubles or halves, or the camera
crosses a lookup height. The reused points are asked again, so only a
timeout, failure or empty answer spends.
- Repaint the last AI reply when a small pan leaves the context unchanged.
Cached-only hosts and voice context lookups are unchanged.
- The trailing visibility frame uses the layer's injectable runtime timer and
is cancelled by resetState as well as disable and destroy.
- An empty-feed regular pass records the camera pose, so later in-cadence
frames do not keep scheduling trailing frames.
- New renderGovernor.test.mjs: a parked camera schedules nothing, a stopped
camera gets exactly one trailing frame and no self-wake, a forced pass
requests its frame, destroy cancels the timer.
- The CHANGELOG entry moves to changelog.d/871-vessels-idle-render.md; docs
say the hold also covers vessels dimmed under a live target.
The route proxy also refuses with 'routing busy — too many routes at once'
(429) and 'route proxy error'; both showed in English in every locale.
Mark them in ROUTING_REFUSALS and add es, ru, zh-CN and zh-Hant entries.
errorText is now cleared on a new connection, a mode switch and stop(), so
it never outlives the error it describes. Tests cover the two refusals,
'Routing failed: rate limited' (pattern order) and the stop() reset.
A Retry-After longer than the 10 s retry budget used to sleep the whole
budget out and then give up, so the proxy's 429 pause started 10 s late
and every new-point poll in that window still went upstream and logged.
The retry loop now stops at once in that case.
Tests: the retry loop no longer sleeps for an unreachable Retry-After;
fully cached points, including stale ones, are served during the pause
with no upstream call; a 503 does not pause; a shorter concurrent
Retry-After never cuts the pause short. CURRENT-STATE now says which
requests are still served and which log.
Said Turksever