feat(window): no minimum size

The window would not go below 1100x700. That was a fair guess at where
the layout stops being comfortable and no business of ours to enforce:
an operator wanting OpsLog as a narrow strip beside a decoder, parked on
a second screen, or on a small laptop, hit a wall with nothing to show
for it. The panels already scroll and collapse.

Zero is how Wails says "do not constrain" — winc fills PtMinTrackSize
only when the value is above zero, checked in the library rather than
assumed — so Windows applies its own floor, about the width of the
caption buttons, and nothing here adds to it.

Removing the constant meant three other users of it had to be sorted
out, since each wanted something different from "the minimum":

- The saved-geometry sanity check now has its own floor, 200x150. That
  is not a minimum the operator feels; it decides when a stored size is
  corrupt rather than chosen. A window restored at 0x0 cannot be grabbed
  to fix it, and that is the one state there is no way back from.
- The maximised-corner monitor check was passing the minimum AS a size,
  which would now be a zero-sized rectangle. It uses the saved size.
- Leaving compact mode validated its captured geometry against the
  minimum; it uses the sane floor.

SetCompactHeight keeps its 900 minimum: it returns early unless compact
is on, and leaving compact resets the minimum to none.
This commit is contained in:
2026-09-10 20:20:05 +02:00
parent fd93036f86
commit 9e67ddcea4
4 changed files with 62 additions and 13 deletions
+4 -2
View File
@@ -13,7 +13,8 @@
"The satellite ground track no longer draws a straight line across the map when it crosses the antimeridian.",
"A correction you make on the transmit VFO is now kept for the whole pass, and remembered for that transponder. Doppler tracking used to undo it a second later.",
"The Green Heron RT-21 is named in the rotator list. It speaks the DCU-1 command set OpsLog already drives, over its COM port or straight over TCP with the Ethernet option — set the controller to DCU-1 / Rotor-EZ.",
"Retract and Calibrate on a SteppIR now show the elements moving, and inhibit the transmitter while they do. Neither said anything before — the same was missing on an Ultrabeam retract, where the element lengths counting down hid it."
"Retract and Calibrate on a SteppIR now show the elements moving, and inhibit the transmitter while they do. Neither said anything before — the same was missing on an Ultrabeam retract, where the element lengths counting down hid it.",
"The window can be made as small as you like. The 1100x700 floor is gone — only Windows own limit remains."
],
"fr": [
"Le tracé du ciel et la position passent dans leur propre colonne, à gauche de la carte. Largeur réglable, et la colonne se replie comme celle de droite.",
@@ -27,7 +28,8 @@
"Le tracé au sol du satellite ne trace plus une ligne droite en travers de la carte lorsquil franchit lantiméridien.",
"Une correction faite sur le VFO d’émission est désormais conservée pour toute la passe, et mémorisée pour ce transpondeur. Le suivi Doppler leffaçait une seconde plus tard.",
"Le Green Heron RT-21 est nommé dans la liste des rotators. Il parle le jeu de commandes DCU-1 que OpsLog pilote déjà, via son port COM ou directement en TCP avec loption Ethernet — régler le contrôleur sur DCU-1 / Rotor-EZ.",
"Rétracter et Calibrer sur une SteppIR montrent désormais les éléments en mouvement, et inhibent l’émission pendant ce temps. Ni lun ni lautre ne le signalait — même manque sur la rétraction dune Ultrabeam, où le défilement des longueurs d’éléments le masquait."
"Rétracter et Calibrer sur une SteppIR montrent désormais les éléments en mouvement, et inhibent l’émission pendant ce temps. Ni lun ni lautre ne le signalait — même manque sur la rétraction dune Ultrabeam, où le défilement des longueurs d’éléments le masquait.",
"La fenêtre peut être réduite autant que vous voulez. Le plancher de 1100x700 disparaît — il ne reste que la limite propre à Windows."
]
},
{