Its row gains a date field for a manual entry, and a flex row makes room
by shrinking its children — so the widest box, which is the one the eye
is on while typing, was the one that visibly narrowed.
A notch narrower and shrink-0: the size it will always be, with or
without the date, and the slack now comes from the boxes beside it.
"I changed an award and the column stayed empty" cannot be answered
without knowing whether the recompute ran at all, how long it took, and
whether it changed anything — and zero rows changed is a real answer
(the definition matches nothing in the log) that looks exactly like a
recompute which never happened.
Preferences is one component holding two hundred pieces of state, and its
biggest panels are eight hundred lines of form — external services, CAT.
A plain controlled input sends every keystroke into that state, so every
character re-rendered the whole dialog, and the letter arrived after the
finger had left the key.
BufferedInput keeps the text where it is typed and hands it up 120 ms
later. A drop-in for Input, so the change is one import rather than a
hundred call sites — which means it has to be right in every shape those
call sites take:
• Blur flushes at once, so clicking Save cannot lose the last word, and
so does unmounting: a panel changed mid-word still hands up what was
there.
• A value that comes back DIFFERENT from what was sent up is adopted,
even with the focus in the field. That is what keeps the fields that
normalise as you type working — the callsign box that upper-cases,
the port box that drops everything but digits. They echo a corrected
value, and the correction wins.
• A value changed from outside while the field is idle wins too: that
is how loading the settings, or switching profile, refills the form.
• Checkboxes, colour pickers and file fields pass straight through.
0.27.17 went out, so everything committed after its release commit needs
a block of its own — the padlock, the duplicated section name, the
digital row of the matrix and the MQTT chip.
The backdrop-blur fix goes with them rather than staying in 0.27.17. It
had been folded into that version's typing-lag entry, but it is not in
the binary 0.27.17 shipped: what that release actually carries is the
cluster-macro half, which is what its entry now says again.
Three things reported together.
Preferences said the section name twice: a small line above every panel
repeated the heading right underneath it — "GENERAL" over "General" —
while the sidebar next to them already shows which section is open,
highlighted. Two lines and a highlight for one fact; the small line goes.
The band matrix opens on the digital row the operator chooses. Its
digital row has always rotated — DIGI, then each digital mode in their
own list — but it always STARTED on DIGI, so somebody who only works FT8
clicked through to their own mode on every callsign. Settings ▸ General
now says where the rotation starts, and the dropdown offers exactly the
modes the matrix rotates through (the phone-mode rule is now shared
rather than copied, so the two cannot drift). DIGI stays the default: it
is the right answer for anyone working several digital modes.
And the MQTT chip is gone from the status bar. That is the name of a
message protocol, not of anything an operator has — a chip in the status
bar has to say what it is about, and this one told nobody anything. What
it carried is shown in the Chase New panel, which is the place that uses
it; its poller went with it.
Logging a contact from a piece of paper — a contest sheet, a friend's
report, a QSO worked on another radio — means the frequency, the band,
the mode, the date and both times all have to stop following the rig and
the clock at once. That is a single decision, and it was five clicks in
five different places, each of which had to be found first.
The padlock beside Start UTC now holds all five and releases all five,
with the same behaviour as before: locking pre-fills the times so the
fields are not empty, unlocking hands them back to the clock and snaps a
QSO already in progress to now.
The five per-field locks stay underneath. Everything downstream reads
them, and individually they say the right thing — "this value is
decoupled from the rig". Only the control is one.
A backdrop-filter covers the whole window and is recomputed every time
anything above it repaints. Behind these dialogs is an application that
never stops moving — CAT polls four times a second, spots arrive, meters
sweep, maps redraw — so the filter was being recomputed continuously, and
each keystroke's repaint dragged a full-window blur with it. That is why
the lag was felt in Preferences and nowhere else, and why fixing the
cluster macros did not end it.
Worse in one place: the cluster server editor opens FROM Preferences, so
its overlay was the SECOND full-window filter stacked over the first.
That is exactly where the delay was first reported.
The dialogs an operator types in for minutes — Preferences, the cluster
editor, the QSO editor, bulk edit, alert rules, award definitions — now
dim the background harder instead of blurring it. Everything else keeps
the blur: a confirmation you click through in a second costs nothing.
This tab began as the relay and rotator dashboard and stopped there: the
three things an operator touches most — the radio, the CW keyer, the
voice keyer — were the ones missing from the page that claims to show the
station.
The radio card carries the frequency and the mode large, because that is
what gets glanced at, and the split pair only when there IS a split: a
second frequency shown at all times is one more number to read past. When
CAT is down it says which kind of down — switched off, or on and not
answering.
The CW keyer card carries the speed, which is the control an operator
reaches for mid-QSO when a station comes back faster than expected, and
Stop beside it because a message going to the wrong callsign has to end
now. The voice keyer card carries the recorded messages themselves: a
card that only said "idle" would be a light, not a control.
Each polls its own binding and holds its own state, like the supply card
above them, so they drop into the grid and reorder with everything else.
The two keyers appear only when there is something behind them — a port
configured, a message actually recorded — because an operator who works
neither should not be handed two dead cards. That question is asked once
on opening the tab: a keyer is bought and wired, not something that
appears mid-session.
Typing a cluster macro cost two things per character. The state lived on
SettingsModal, so every keystroke in one of those twenty-four boxes
re-rendered the WHOLE preferences dialog — every list, every form, every
panel. And the save wrote through to Go and into the database on each
one: a round trip per character.
The editor is now its own module-scoped component with its own state, so
a keystroke re-renders twelve rows. And writeUiPrefDebounced holds the
database write until the typing stops, while the local cache — which is
what everything reads back — is still written at once. Pending values are
flushed when the page goes away, so typing and immediately closing does
not lose the last word.
The behaviour is unchanged: still saved as you type, still no Save
button, because a text box whose contents only take effect on some other
button is how work gets lost.
Also folds the satellite changelog into one [NEW] entry. Satellites are
new in this version — nobody reading the notes has seen any of it — so a
running account of how it was built, tab then tracking then rotator then
where the settings moved to, is the wrong shape. One entry saying what it
does.
Setting up a telnet cluster is the step operators get stuck on. The
address and the port are two pieces of information nobody has to hand,
and a typo in either looks exactly like a node that is down — there is
nothing to read, only a connection that never comes.
So the editor offers a list, and choosing one fills the fields. Six to
begin with: F4BPO, DXFun, SOTA, POTA, and the two Reverse Beacon feeds —
listed separately and named for what they carry, because they are one
network on two ports where 7000 is CW and RTTY and 7001 is FT8 and FT4,
and no amount of trying will tell you which is which.
The preset then gets out of the way. Everything stays editable, the name
is only filled when the operator has not chosen one of their own, and a
node typed in by hand behaves exactly the same. Reopening a node created
from a preset shows it selected, so the list also answers "which one is
this".
Click the heading. Within each PERIOD and never across them: the slots
are what this panel is — what was on the air in one fifteen-second
window — and a list sorted end to end would mix three minutes of decodes
into one column of numbers with no way to tell which window any of them
came from.
One click sorts the way that column is worth reading — strongest signal,
lowest frequency, furthest DX, A to Z, most wanted — the second reverses
it, and the third gives arrival order back. Arrival order stays the
default and stays one click away, because it mirrors the decoder's own
window line for line, which is what makes the two screens comparable at a
glance.
Status ranks by the cluster's own order, so the two views rank the same
things the same way, with the markers that are orthogonal to the entity —
a new county on a worked country — sorted above the plain duplicates.
A station that never sent a grid cannot be placed, and an unresolved
country is not a name: both sort to the end whichever way the column
goes, rather than pretending to a distance of zero and heading the list
under "nearest first".
The map answers "where is the satellite over the earth". This answers
"where do I look", which during a pass is the question that matters.
The projection is the one every tracker uses and every operator already
reads: the centre is the zenith, the rim is the horizon, north is up. So
the radius is (90 − elevation), not the elevation — a bird overhead is a
dot in the middle, and a pass that hugs the rim is one that never rises.
Whether it comes over the roof or along the treeline is something no
amount of azimuth and elevation digits conveys, and one glance settles.
The whole pass is drawn: a dashed track with arrowheads for the direction
of travel, a hollow circle where it rises, a filled one where it sets,
and a cross where the satellite is now — green above the horizon, grey
below, because the numbers are still right down there and nothing can be
worked through the earth.
The track is fetched once a minute, not once a second: the SHAPE of a
pass does not change while it happens. Only the marker moves, and that
rides on the tuning poll that was already running.
Headings, because a column of numbers with nothing above it makes an
operator work out what they are looking at every time. A real table, so
the satellite column takes the width the longest name needs — "ZHUHAI-1
OVS-1A" was cut to eight characters in a fixed one.
Colour where it carries meaning, and nowhere else. The maximum elevation
is the quality of the pass, so it is coloured like one: a bird 70°
overhead and one scraping 12° along the horizon are not the same evening,
and the table should say so without the operator reading every number. A
pass in progress is green, one starting within five minutes is amber. And
a dot for the mode: FM and SSB call for a completely different set-up,
and which the next pass is decides whether you reach for a handheld or
for the whole station.
The rotator's baud rate is a dropdown, like every other one in OpsLog.
Three things reported together from the tab.
The shipped frequency plan went from eleven satellites to twenty-five:
the eight Tevel FM cubesats, EO-88, AO-109, CAS-4A and 4B, TO-108,
GreenCube's single-frequency digipeater, and QO-100's wideband
transponder beside its narrowband one. It remains a starting point in a
file the operator can correct — a transponder gets switched and no
release should be needed to follow it — and the picker still lists every
bird in the element set when the "with a plan" filter is unticked.
The readout column drags to any width between 240 and 720 pixels,
double-clicks back to its default, and folds away entirely. How much map
against how much detail is the operator's call: watching a footprint
cross an ocean and working a pass want opposite things.
And the locator is no longer asked for twice. Passes are predicted from
the station locator, which is set once in Station information; the field
here was only ever for an antenna at another site, so it says so and sits
folded away. Nobody should have to wonder which of two locators is in
use.
Also: "Driven by" is two columns wide. "OpsLog (EasyComm II)" did not fit
in a third of the row, and a truncated choice is a choice that cannot be
read.
Reported from a station running MSHV, WSJT-X and JTDX together: click a
call in MSHV and the entry field filled, emptied, refilled — once a
second — with the map zooming in and out to match.
Two causes, both about reading one program's statement as another's.
The clear was tracked per LISTENER. Several decoders commonly share one
multicast group, so an idle WSJT-X reporting no DX Call — which is simply
true, and which it repeats every second — was read as MSHV abandoning the
station it was calling. "The operator cleared the DX Call" is a statement
about one program, never about a socket, so it is now tracked per
program, and a clear carries the id of whoever made it.
And nothing arbitrated between them. The program that announces a station
now holds the entry field, and the others cannot touch it until it lets
go: it clears its own call, it stops sending (closed), or the QSO is
logged. That is the operator's own suggestion, and it is the right one —
between overs there is no way to tell "I have nothing" from "I am not the
one you are working" except by remembering who was.
Refusing another program's callsign is logged once per focus, not once a
second: an operator whose second decoder "stopped filling the call" needs
something to read.
The world map and the grid-square map shared a single key, so picking
satellite imagery to look at grids repainted the main map too, and there
was no way to have terrain on one and streets on the other. They are
different maps answering different questions, and the imagery that suits
one is not the imagery that suits the next.
Four keys now, one per map, in lib/mapBase beside the remembered views —
named in one place so a rename cannot silently orphan somebody's choice —
and portable, so a copied data folder brings them along. The grid map
inherits whatever was set under the old shared key rather than being
reset to the default: an operator who chose imagery there keeps it.
It handles azimuth and elevation, and a great many stations already run
it in front of a controller OpsLog has never heard of. For those,
OpsLog talking to the controller itself would be a second program
fighting PstRotator over the same cable — so it hands over the bearing
instead, and lets PstRotator turn the mast.
Both kinds sit behind one small interface, chosen in Settings. Neither is
more correct than the other: the right one is whichever the station
already has working.
The 450° overlap is deliberately NOT applied on the PstRotator path.
PstRotator knows which machine is on the other end and does its own; two
programs each deciding to go the long way round is exactly how an antenna
unwinds in the middle of a pass.
Position queries are asked at most every three seconds rather than on
every tick. A PstRotator query binds a socket and waits up to a second
and a half, and many setups answer nothing at all — so one silence is
enough and it stops asking, reporting the commanded position instead and
saying that is what it is.
Settings ▸ Lists ▸ Satellites is gone. It was a text box an operator
typed their birds into by hand, and it had nothing to do with the
satellites the tracker knew: the same station kept two lists of the same
satellites and they drifted apart. The SAT_NAME box on the entry form now
offers the followed set — or every satellite with a frequency plan when
none is followed — merged with whatever that old list still holds, read
and never written, because what it holds is somebody's past work.
And the satellite section moved out of Hardware, which it never was.
Which birds you chase, where your antenna stands and how old your
elements are is operating; the rotator inside it is the only hardware
there, and one block does not make a section a device.
Two things belong in different places, and they were in one.
Settings → Satellites now holds the setup: which satellites to follow —
the same two-column shape as the awards, for the same reason, since a
feed carries two hundred birds and an operator works six — and the
orbital elements, their age, the fetch, and pasting your own. Following
none still means every satellite with both elements and a frequency plan,
so somebody who has not chosen yet is not handed an empty tab.
The panel keeps only what a pass needs. A countdown to AOS, or to LOS
once it is up, because that is the number that decides whether you sit
down; a bar for where in the pass you are, since mid-pass the useful
question is not the clock but whether you are past the peak; rise, peak
and set with compass directions, because "rises SW" is a direction to
look in and 213° is arithmetic. Distance, altitude and footprint. And
approaching or receding, which is the sign of the whole Doppler
correction and the only thing that explains why the frequencies are
moving the way they are.
The countdowns run in the browser from two timestamps. Predicting a pass
steps the orbit across a day thirty seconds at a time, which is not
something to do once a second for a clock the page can keep itself.
Two things reported together on the new tab.
The locator: the station's is on the ACTIVE PROFILE, not in a settings
key. keyStationMyGrid is a legacy key that EnsureDefault migrated into
the profile long ago and nothing writes any more, so reading it told an
operator with a perfectly good locator on screen that he had not set one
— and refused every pass prediction on the strength of it.
The map: Leaflet stacks its panes and controls up to z-index 1000, which
without a stacking context of their own float over the whole application
— Preferences opened behind the map, its Save and Close buttons under it.
The other three maps in OpsLog each carry `isolate` for exactly this
reason; this one was missing it.
EasyComm is what satellite rotator controllers agreed on, so a box that
works with SatPC32, Gpredict or Hamlib works here. Serial or TCP, and its
own settings rather than the HF rotator's: an az/el pair is a different
machine on a different port, and an operator who has both must not have
to choose.
A great many EasyComm controllers — the Arduino trackers above all —
accept commands and never say a word back. That is legal and common, so a
silent controller is not treated as a broken one: it is still driven, and
the last commanded position is reported in its place, marked as commanded
rather than read. A stuck rotator must not be able to hide behind an
order it never carried out, which is why the panel shows the antenna's
position beside the satellite's.
The 450° overlap is the reason a satellite rotator is worth having, so it
is used: a pass crossing north continues past 360 instead of unwinding
three quarters of a turn with the antenna sweeping the ground. Below the
configured elevation the mast is left alone — the numbers are right all
the way round the orbit, but a rotator that chases a satellite through
the far side of the earth spends the night turning, and a mast has a
finite number of turns in it.
The hard part of satellite tuning is not the arithmetic, it is deciding
who owns the dial. A tracker that forces both frequencies fights the
operator every time they turn the knob to follow a station across a
linear transponder; one that never touches the receiver leaves them
chasing a signal that slides nine kilohertz across a 70 cm pass.
So the operator owns the receiver and the tracker follows them. Every
second it asks the radio where the receiver actually is. Where it put it,
nothing has changed. Further than a dial-turn's tolerance, and the
operator has chosen a station: what they landed on is converted back into
a nominal frequency, and the transmitter is derived from that. Which is
the division of labour on a linear bird — the operator listens, the radio
does the sums.
Three ways to reach the radio, because a satellite pair is a shape of
operating rather than a manufacturer's feature. An IC-9700 or IC-9100 is
asked for its OWN satellite mode: it pairs main and sub, gives full
duplex, and keeps the dials linked the way its designers meant, which is
always better than an imitation built out of split. A Flex gets two
slices, A the downlink and B the uplink, created when missing, because
"slice B does not exist" is not something to make an operator fix at the
start of a ten-minute pass. Everything else gets the downlink, and is
told so — half the job announced beats half the job hidden.
What goes in the log is the NOMINAL pair. Two stations working each other
through a transponder read different numbers off their dials at the same
instant; the only figure they can both agree on is the transponder's own.
FREQ is the uplink and FREQ_RX the downlink — the one place a satellite
QSO differs from every other kind, and the reason FREQ alone cannot
describe one.
Three questions answered at once, because on a pass there is no time to
go looking for any of them: where the bird is, when the next one comes,
and what to tune. The map draws each satellite's footprint — the honest
answer to "can I hear it", since everything inside the circle has the
satellite above its horizon — and the selected one's path over the
ground. The pass list is every favourite in time order, the one in
progress in green.
The readout shows the corrected frequency large and the nominal one
beneath it. Only one of them, and an operator cannot tell a Doppler
correction from a mistuned transponder.
The map opens on the station rather than the Atlantic, and remembers
where it was left like the others. The panel is mounted only while its
tab is visible: it asks for the tuning once a second, and there is no
reason to compute an orbit nobody is looking at.
What internal/sat could not know: where the antenna is, which birds the
operator cares about, and where the files live.
Startup reads the cached elements and the frequency plan from disk and
nothing else — one file and a few hundred parses, so the tab is full the
moment it is opened, on a shack PC with no internet as much as on one
with. Fetching is the slow, optional half and never blocks a launch; it
happens on its own only when the set is stale and the operator asked for
it.
Elements pasted in by hand go in their own file. The feed cache is
replaced wholesale on every refresh, so a freshly launched satellite —
whose elements circulate on a mailing list days before any feed carries
it, which is exactly the week everybody wants to hear it — would
otherwise be wiped by the first automatic update.
The list joins both halves and shows what is missing on either side. A
bird with elements and no plan is one the operator can still track; a
bird with a plan and no elements is the visible symptom of an element set
that is too old. Dropping either turns a fixable configuration problem
into a satellite that "does not exist".
GetSatelliteTuning is the working answer, and everything that will later
drive a radio is built on top of it rather than beside it, so the display
and the rig can never disagree. It keeps the operator's frequency
nominal and applies Doppler only on the way out: on a linear pass the
station being answered stays put on the dial while both radios chase the
shift. A geostationary bird is corrected by nothing at all.
Two things the tracker cannot work without, both kept apart from the
orbital maths on purpose.
The elements come from Celestrak's amateur group, with PE0SAT as the
fallback for the hour when Celestrak is rate-limiting a hundred trackers
at once. A malformed satellite is skipped rather than fatal — a feed of
two hundred birds with one bad checksum must still give the operator the
other hundred and ninety-nine — and the count is returned so the app can
say so. The cache is plain TLE text in the data directory, written
beside and renamed, and only replaced once a feed has produced usable
elements: a captive portal must not take away the set the station
already had. Loading it first is what makes the satellite tab full on a
shack PC with no internet.
The frequency plan is separate because it changes for different reasons:
elements every few days from a feed, a transponder when the satellite is
commanded into another mode. The shipped list is a starting point, copied
to the data directory on first use and read from there afterwards, so an
operator can correct a frequency without waiting for a release and keep
the correction across updates — and a file they have broken is reported,
not overwritten.
UplinkFor is the part that matters on the air. On an inverting linear
transponder, tuning up the downlink means going down the uplink; get it
backwards and you transmit at the far end of the passband from the
station you can hear, which is the classic first evening on a linear
bird.
Names are matched on letters and digits alone. Celestrak says
"RADFXSAT (FOX-1B)" where every operator says AO-91, and nobody spells
Es'hail the same way twice.
The foundation of the satellite branch, and nothing above it yet: where a
satellite is (SGP4 from akhenakh/sgp4, Apache-2.0 and pure Go, so the
no-cgo rule holds), where it will be (passes with an elevation floor,
because a three-degree scrape is a line in a table that will never be a
QSO), and what its motion does to a frequency.
The Doppler pair is the part worth being careful about: the two
corrections go in OPPOSITE directions. The downlink arrives shifted and
we tune to meet it; the uplink must LEAVE shifted the other way to land
on the transponder's nominal input. A test pins the signs and the size —
7 km/s on 2 m is about 3.4 kHz.
Elements keep their raw lines beside the parsed form: that is what the
cache stores and what an operator pastes by hand for a bird no feed
carries yet, which is exactly when everyone wants to hear it.
Ticking "RTTY on USB" changed nothing: the flag is deliberately absent
from catLinkSig — none of these preferences is worth dropping the CAT
link, and with it WSJT-X's rigctl session, to apply — so saving the
settings left the link alone and the running client kept its old answer
until the next launch.
Preferences that the link does not depend on are now pushed to the
connected rig when the settings are saved, through the Yaesu escape on
the manager. SetRTTYUpper joins the controller interface for that: it is
a preference rather than a command, but it has to be reachable on a rig
that is already talking.
The selector row appears above the dial when there is more than one
rotor, and the widget's height is not its own to take — it sits in a
strip sized by the entry form beside it. The row was simply added, so the
SP/LP pair and half the Stop button went off the end.
The dial (24 px), the three button rows (8 px each) and the padding
(4 px) now give that height back between them, which is a selector row
almost exactly. The dial and the controls column stay the same height as
each other, so the two columns still line up.
Windows Defender removed 0.27.14 from a station as
Trojan:Script/Wacatac.H!ml. That detection is machine-learning, not a
signature, and the behaviour it scored is ours: an unsigned binary
replaces itself on disk, clears the mark-of-the-web, and spawns a
windowless PowerShell that waits for its own process to die before
starting another executable. Byte for byte, that is a dropper; the model
reads the shape, not the intention, and "Script/" names the PowerShell.
The wait it was written for is not needed. --post-update already makes
the new instance patient with the single-instance mutex — twenty seconds
of it — so the new exe can be started directly while this one is still
shutting down and simply wait its turn.
The deferred-swap fallback keeps its helper: nothing else on a stock
Windows can wait for a pid and then move a file over an image that is
still running. It is reached only when the rename failed, never on the
ordinary path.
The eight-digit rigs, the typed watering hole and the RTTY sideband all
landed after the 0.27.15 release commit, so they were sitting in a block
that had already shipped. 0.27.15 says what was released again.
His log settles the 28.074 case: "SetCATFrequency 28.074 MHz" and the
state still reads mode=USB — nothing sent a mode, so the rig simply
stayed where it was. A spot click has always carried one; a frequency
typed by hand carried none. It now uses the same table and the same
tolerance as a spot (±3 kHz of a known FT8/FT4/JS8 frequency), and only
towards the digital modes: tuning AWAY from one leaves the mode alone,
because there the frequency says nothing about what the operator means.
RTTY on Yaesu is a choice the log cannot make: ADIF records "RTTY" and
the rig has MD06 (RTTY-L) and MD09 (RTTY-U). The older lower sideband
stays the default and a station whose FSK controller wants the upper one
says so once in Settings → CAT.
An FTDX3000 rejects every FA command: the FTDX10 family writes a
frequency in nine digits and everything before it — FTDX3000, FTDX5000,
FTDX1200, FT-2000, FT-950, FT-450 — writes eight, answering the longer
form with "?;". The rig would not follow and nothing said why.
The width is LEARNED from the rig's own replies rather than tabulated: it
announces the format in every answer to FA;, so it comes from the radio
in front of the operator instead of from a model list that is always one
release behind — and a Yaesu this backend has never heard of is right on
the first read. Nine until the first reply lands, which is what the
modern rigs use and what this backend was written against.
The five older ID codes are named too, so the console says FTDX3000
rather than a bare number.
The main hint named three fixed colours, which stopped being true the
moment the operator could choose them. And the paragraph under "Mark
stations already worked" explained a trade-off nobody reads standing at
the radio; the switch says what it does.
Same order as the last block: what is new leads, each fix follows the
thing it belongs to — the antenna pair, then the Icom network run, then
the rest.
The shortened transmit gag was written for an Ultrabeam that is now polled
four times a second while it moves. The SteppIR was still on two seconds
and reported nothing at all until its own poll came round, so the gap
between the 900 ms grace and the first poll that would have seen the
movement was a hole in "block TX while the elements travel" — the
transmitter released in the middle of a move.
It now reports a commanded move at once (bounded, so an antenna that
never answers cannot latch the inhibit on), on a COPY of the cached
status so the flag cannot leak into what the poll goroutine owns, and
follows the motors with its poll rate exactly as the Ultrabeam does.
The transmit gag starts when the move is commanded; the indicator waited
for a status poll to say the elements were travelling, so the two
disagreed by a second or more — and on an automatic follow (band change
with tracking on) there was no sign at all until a poll landed. The
backend now announces the order as it goes out, the interface shows it at
once and re-reads the antenna immediately, and the real status takes over
the moment it arrives.
The WSJT-X "grey out stations already worked" switch is renamed "Mark
stations already worked": with the colour now chosen by the operator the
old name described the wrong thing, but the switch still answers a
question the palette cannot — WHETHER dupes should be marked at all. On a
well-filled log they are most of a period, and a window where nearly
every line is coloured has stopped saying anything.
The motorised antenna gagged the transmitter for a second or two after it
had finished moving. Three delays were stacked: the antenna polled every
two seconds, the gag held for three after the command whatever the
antenna said, and the widget refreshed every three. The antenna is now
asked four times a second WHILE IT MOVES — an idle one has nothing to say
and stays at two seconds — the gag only bridges the command itself
(900 ms), and the widget follows at half a second while moving.
WSJT-X highlighting:
- A watch-list station already worked on this band and mode is no longer
painted as one to call. The list is a statement of intent, not of what
is left to do, and its pink outranked every other verdict including the
log's, so a worked station stayed pink for the session with nothing to
tell it from one still needed.
- The four colours are the operator's to choose (Settings → UDP). Only
the background: the text colour is derived by luma, so a dark blue
cannot come back as black-on-black in somebody else's window. Changing
one clears the installed highlights, or the de-duplication would keep
showing yesterday's colour until a callsign changed verdict.
The square ring put the ticks in the corners where the room is, and made
every distance from the centre depend on direction — a marker at 45 sat
further out than one at north. A dial is read by angle, so the ring it is
read against is the same distance away all the way round now, and the
markers need one radius instead of a per-direction one.
The antenna is a sector that fades outwards rather than an arrow: an
antenna does not look along a line, it looks through a lobe, and the
arrow claimed a precision the beamwidth does not have. Where the mouse
would send it is drawn the same way, and its azimuth cross-fades into the
big readout while aiming — the figure is read at the moment the beam is,
so it belongs in the same place rather than somewhere to look away to.
Three colours, three meanings, kept apart: green where the antenna IS
(as everywhere else in OpsLog), orange where a click would send it,
yellow what was ORDERED. His revision drew all of them orange and yellow,
which loses the distinction the dial exists for — and drew the second
lobe of a bidirectional Ultrabeam at full strength, indistinguishable
from the main one. It is dimmed again; the dashed boom, the REV/BI badge
and the compact form are untouched.
The conninfo that authorises the RX stream goes out during the login,
before the audio socket exists — it has to, since it is what authorises
it. So the rig is told to send audio to :50003 while nothing is bound
there, gets an ICMP port-unreachable back, backs off, and the audio turns
up only when its own retry timer comes round: twenty to thirty seconds by
the operator's watch, ninety in one log.
It is sent once more the moment the port is listening, which is the same
message the session already carries — RS-BA1 repeats it too. The audio
dial moved above the pumps, because after ctrlPump is running the
control-stream auth state belongs to it.
"I turned the sound back on and nothing comes out of the speakers", with
a log that reports success at every step: the network audio stream up,
664-byte packets arriving, the monitor started. The render goroutine's
error was thrown away — a device unplugged, renamed by Windows or unable
to open at 16 kHz fails exactly there, silently.
It is logged now, with the device id. Two once-only lines either side of
it say whether the decoded audio is reaching a running monitor or
arriving with nobody listening, which separates a device fault from the
speakers simply being switched off.
The guard I added for services with no credentials demanded one, and
nobody has ever set it: OpsLog carries its own Club Log APPLICATION key
(clublogAppAPIKey), so the account is an email, a password and the
logbook callsign. An operator whose live upload had been working for
months was told the service was not configured the moment he sent QSOs by
hand after an import.
Written in app.go, the rules drifted from the uploaders on the first try.
They now live in internal/extsvc beside the Upload* functions that
enforce them, each case mirroring that function's own guard — which also
caught Cloudlog, where the station profile is required and the check did
not ask for it. The message names the fields actually missing rather than
listing everything the service takes.
Audio off, and the IC-7760 still answers at connect and then never again:
twenty-eight commands sent, not one reply, packets still arriving on the
stream. A count says how much went unanswered and never which command
went out last — which is the one thing that can identify a frame this rig
does not tolerate, on a radio nobody here has.
The transport now keeps the command headers of the last eight frames and
prints them with the silence report.
The RX audio stream is experimental and shares the rig's session with
CI-V. The shape in an operator's log is unmistakable: audio packets
arriving by the hundred while not one CI-V reply comes back, the watchdog
tearing the session down, twenty seconds' pause, and the whole thing
again — read from outside as "the Icom keeps disconnecting", with
nothing pointing at the switch that would end it.
The log now says it, and names the setting. icomAudio counts what it has
delivered so the line distinguishes "audio is enabled" from "audio is
arriving", which is the half that makes it a suspect.
From an operator's log: an IC-7760 in standby, and OpsLog dialling and
dropping every forty seconds for as long as it was left there — control
link up, login OK, token renewed, and not one CI-V answer.
lastGoodAt bounds "the link answers but no CI-V comes back". It belongs
to a session and was never cleared when a new one opened, so a rig that
went to standby half an hour ago handed every fresh session a
half-hour-old last good read: past the grace before the first command was
even sent. Torn down at once, redialled twenty seconds later, torn down
again. Cleared on connect, the rule reads as it was written — silent
since connect is a rig in standby, and the session is kept so it can be
woken.
The console it is woken from was missing too. It appeared only once the
live CAT state said "icom", which a sleeping rig never says, so the ON
button was absent at the one moment it exists for. It now follows the
CONFIGURED radio — which also had to start following a radio switched
from the status bar, instead of waiting for a trip through Settings and a
Save that changed nothing.
An exported file was full of blank gaps: a record ran down a dozen lines
and the next appeared to start in the middle of the page. ADDRESS is a
multi-line field by the standard and callbooks and other loggers fill it
that way — "Kabul", four blank lines, "Afghanistan" — and the writer
copied the value out as it stood. The files were always valid, since ADIF
counts bytes; they were unreadable, and so was anything that quoted them.
Line breaks inside a value are now joined with a comma and tabs become
spaces, which is how an address reads on one line anyway. The length
prefix is computed after the flattening, so the record stays exact, and
every path through the writer gets it: the file exports, the uploads and
the record forwarded over UDP.
The changelog's 0.27.15 block also takes back the FT-map hover fix, which
landed after the 0.27.14 release commit and was sitting in that block.
rouggy