The voice context records for a selected flight or military contact
rounded the track without wrapping, so a contact on 359.5° or more was
described to the assistant as heading "360°". Both now wrap after
rounding: 359.6° reads "0°", 359.4° still reads "359°".
The mission orbit lookup reads each loaded satellite's launch year from its
international designator, but satellite.js never parses that field, so the
year check never ran and a new payload could take an older namesake's orbit
and position. The loaded catalog now keeps the designator from the TLE line.
The live flow status derived its wording from the rounded coverage
percent, so 1 matched dot among 300 rounded to 0% and read
"SIMULATED ... (no matches)", and 1 simulated dot among 300 rounded to
100% and dropped "Unmatched: simulated". Coverage now stays within 1-99%
whenever both matched and simulated dots are shown.
Placed route, arrow and drawn-line labels picked metres before rounding
to ten metres, so anything from 995 m to 1 km read "1000 m", even where
the draw hint for the same line read "1.0 km". The label now rounds
first, then picks the unit, and does the same at the 10 km cut, where
9,960 m read "10.0 km".
The draw hint picked m², ha or km² from the raw area and then rounded,
so 9,999.5 m² up to a hectare read "10000 m²" and about 999,500 m² up
to a square kilometre read "100.0 ha". It now rounds first, as the line
measure already does.
dump1090-fa and readsb file an aircraft's vertical rate under geom_rate
instead of baro_rate when its source is GNSS. The decoder-feed adapter
read only baro_rate, so those aircraft showed "V/S —" on the Local ADS-B
card. It now falls back to geom_rate; a sent baro_rate still wins.
Moving the replay-speed slider during the pre-launch hold or countdown
rescaled the time left to liftoff, so 0.25x turned T-8 into T-32 and 4x
into T-2. The speed change now skips a replay that has not lifted off;
mission time already flown is still preserved as before.
routeDistanceText picked one decimal while the raw kilometres were under
10, then rounded them for display, so 9.95 km or more read "10.0 km".
It now picks the unit from the rounded tenth, as it already does at the
metre/kilometre boundary.
The selected ALPR card rounded a mapped camera:direction without
wrapping, so a camera facing 359.5° or more read "DIRECTION 360°". The
rounded bearing now wraps to 0°.
TeleGeography stores 17 cables (2Africa, Echo, Hawaiki...) as several
features under one id, and the tool made a row per feature, so "Echo"
read "3 submarine cables" and area counts ran high. Rows are now unique
by cable id, and a cable still matches when any of its parts does.
Replay Ascent schedules liftoff 15 s out: a 5 s tile-preparation hold,
then a 10 s countdown. The hold ended 5 s before liftoff instead of 10,
so the replay sat on the pad for 10 s and the countdown began at T-5.
The hold now ends at T-10, and the countdown runs T-10 to T-1.
A station whose feed publishes no capacity froze its capacity at the
first status poll's bikes + docks, or at the placeholder 15 when that
poll had no row for it, so later cards read "10 avail · 20 docks · 15
cap" and coloured by the wrong ratio. Each poll now resolves it afresh.
The selected-vehicle card rounded the operator's reported bearing without
wrapping, so a GTFS-Realtime vehicle heading 359.5° or more read
"reported hdg 360°". The card now wraps after rounding: 359.6° reads
"reported hdg 0°", 359.4° still reads "reported hdg 359°".
The CCTV HUD heading token and the camera panel's meta line rounded the
bearing without wrapping, so a camera facing 359.5° or more (a feed
bearing, or a heading calibrated just west of north) read "HDG 360°".
Both now wrap after rounding: 359.6° reads "HDG 0°", 359.4° "HDG 359°".
readsb reports an aircraft's vertical rate as geom_rate instead of baro_rate
when its source is GNSS. The military normaliser read only baro_rate, so about
one in six live /v2/mil contacts had a null vertical rate in analyst and
tracked-aircraft data. Fall back to geom_rate, as the adsb.lol flights
fallback already does; a sent baro_rate still wins.
The on-globe Contacts markers picked metres before rounding, so a contact
999.5 m or more away read "1000 m" while the Contacts panel read
"1.0 km". The marker label now rounds first, as the panel does.
select_nearest_aircraft passes altitudeM: null for an airborne contact
without a barometric altitude (military rows from adsb.lol can lack
alt_baro). The speech builder coerced it with Number(), and Number(null)
is 0, so the reply said "..., 0 feet." and the voice card showed
"Altitude 0 feet". Read the altitude only when it is a finite number,
so an unknown one is left out as the builder already intends.
Vessel cards and the AIS HUD rounded the direction without wrapping, so
a vessel reporting no heading and a course over ground of 359.5° or more
read "360°" on its cards and "HDG: 360DEG" in the HUD. The rounded
direction now wraps to 0°.
The Street Level viewer caption rounded the photo's bearing without
wrapping, so a view facing 359.5° or more read "360°". The rounded
bearing now wraps to 0°.
readsb omits lat/lon for a contact heard without a current position; about a
third of every /v2/mil response is such rows. They made every military
snapshot partial, so a contact that left the feed was retained at full
opacity for five minutes instead of fading and leaving on the missed-poll
limits. A row with a hex and no lat/lon now leaves the snapshot complete and
is listed in retainIds, as OpenSky's null-position rows already are.
MilitaryFlightRecords coerced a missing course to 0 (`courseDeg || 0`)
and a missing speed to 0 (`speedMps ?? 0`) before stickyNumber, so the
hold-last-known-good merge never ran. When an adsb.lol row arrived
without `track` or `gs`, the aircraft's icon swung to north and the
position history got a 0 kt / 0 degree fix, so dead reckoning stopped
or turned it north. The civil FlightRecords holds the last value for
the same input.
Pass the raw values to stickyNumber, with 0 as the fallback only until a
value is known, as before. A grounded row without a speed still reads
0, so a parked aircraft doesn't coast at its landing speed.
In the keyless setup a single transient failure while loading the default
map layers locked in the degraded fallback until reload.
_getImageryProvider cached the construction-fallback resolution (OSM) under
the Esri stack id, so every later "Esri Satellite" selection showed OSM.
createKeylessTerrain turns any Re:Earth failure into a flat
EllipsoidTerrainProvider; the controller cached it and set _terrainMode to
'keyless', so a boot-time network blip left the globe flat.
Fallbacks are still cached and shown at once, and no switch waits on a
retry. One controller timer now retries the real provider behind whichever
fallback is on screen (Esri's OSM stand-in, or flat terrain on a shown
globe), starting at 30 seconds and doubling to a 10-minute cap while it
keeps failing. A failed retry leaves the cached fallback and the installed
terrain untouched. A success replaces the cache entry and applies it live:
terrain is installed in place with a render request (or, behind a 3D
tileset, picked up by the next globe activation), and Esri is reselected
automatically, after any in-flight switch settles. A switch that read the
flat fallback re-reads the terrain cache before installing, and destroy()
cancels the timer and releases a provider built after it.
overpassLooksRuntimeError and overpassLooksRateLimited searched the whole
200 body, element tags included. A result containing a place named
"Timed Out Escape Rooms" (or any tag with "runtime error", "out of
memory", "too many requests" or "rate_limited") was refused as an
upstream failure, and the endpoint was put on cooldown for every other
query.
Parse the body once. A data reply (one with an elements array) reports
failure in its remark, so the phrase checks run on the remark. Any other
body, an error page or JSON without elements, is matched whole as
before.
The flights ground-floor pipeline subtracted longitudes without wrapping.
approxDistanceKm, and its copy in the mesh-floor sampler, measured two
points 0.2° apart either side of 180° (21 km at Fiji) as about 38,000 km,
so low and grounded aircraft just across the line from the viewer were
never floored or mesh-sampled. corridorFloorCells walked a segment across
180° the long way round, filling its cell budget with ground behind the
contact and never requesting the cells it was about to cross, and
stickyFloorCell dropped its hysteresis for a contact on the seam. Each now
takes the short way round, and corridor points are normalised back into
[-180, 180].
SEARCH NEARBY SITES centred its Google Places search on the mean of west
and east, which for a view across 180° (west > east) is longitude 0, and
sized the radius from the view's height alone. It now measures the view's
width through 180° and searches its real centre.
A shot action seeks to the target shot's startProgress, which
sceneTimingForShot computes as start / total. sceneSeekState multiplied
that back into seconds and compared it with each shot's end, but
(start / total) * total can round one ulp below start. The comparison
then put the seek at the very end of the previous shot. With five equal
3.9 s shots, "go to" shots 2, 3 and 5 landed on shots 1, 2 and 4.
Compare the requested progress against each shot's end / total instead,
the same division the caller used, so a shot's own start always selects
that shot. Exact boundaries still select the next shot, and anything
below a start still selects the previous one.
OpenSky's state-vector spec allows null longitude/latitude, and every live
/states/all response carries such rows (about 50 of 7,800). The normaliser
rejects them, so admitRecords reported complete: false on effectively every
poll. While a snapshot is incomplete, FlightRecords.absence() retains any
aircraft received in the last five minutes, which meant the stale fade,
MISSING_POLL_LIMIT and the landed fast-cull never ran: vanished and landed
aircraft stayed on the globe at full opacity for up to five minutes.
admitRecords now takes an optional unpositionedId(row). A rejected row it
names is a contact without a position: its id is returned in retainIds and
the row does not count against completeness. OpenSky opts in for rows with an
ICAO24 address and strictly null longitude and latitude; the id comes from
the same helper the normaliser uses. Rows with no usable id, out-of-range or
half-null coordinates, truncated rows and positioned duplicates still make
the snapshot incomplete, and a nonempty feed with no positioned rows is
still malformed.
absence() retains exactly the retainIds aircraft under the existing
five-minute bound, except likely-landed ones, which keep the landed limit.
Everything else ages out normally. Military readsb admission does not opt
in and is unchanged.
fetchPinnedRadioResponse built `new Response(stream, { status })` inside
the raw https response callback, which runs outside the surrounding
promise. A Radio Browser mirror answering 204, 205 or 304 (null-body
statuses), or any status outside 200-599, made the Response constructor
throw. Nothing caught it, so the whole dev/preview server process exited
and every layer went down, not only radio. The catalog's mirror failover
never saw the error.
Mirror the guard local-receivers.js already uses: null-body statuses get
a null body, statuses outside 200-599 are rejected, and anything thrown
in the callback rejects the promise so the existing error handling and
mirror failover run.
Aditya