fix issue with stale hardware timepoints after a failed short timepoint interval run

src/main/java/org/micromanager/lightsheetmanager/model/UserSettings.java

@@ -118,6 +118,11 @@ public void load() {
     public void loadFromJson(final String json, final boolean notify) {
         // TODO: switch this based on microscope geometry type
         var settings = ScapeAcquisitionSettings.fromJson(json, ScapeAcquisitionSettings.class);
+        // useHardwareTimePoints is a DERIVED value. It is serialized as part of the acquisition settings,
+        // but must never be trusted when reloaded: a stale "true" persisted from an earlier short-interval run
+        // would otherwise seed acqSettings_ across a plugin restart and be read before the next run recomputes it.
+        // Force it false on load; the engine recomputes it each run.
+        settings = new ScapeAcquisitionSettings.Builder(settings).useHardwareTimePoints(false).build();
         model_.acquisitions().updateSettings(settings);
         if (notify) {
             notifyListeners(settings); // update the ui

src/main/java/org/micromanager/lightsheetmanager/model/acquisitions/AcquisitionEngineDispim.java

@@ -701,6 +701,12 @@ private boolean doHardwareCalculations(PLogicDispim plc) {
 //            }
 //        }
 
+        // FREEZE POINT: (mirror of AcquisitionEngineScape): rebuild the run-time snapshot from asb_
+        // before arming the controller, so the hardware is armed from the same settings the AcqEngJ event
+        // loop uses. See AcquisitionEngineScape.doHardwareCalculations() for the full rationale.
+        // TODO(IMMUTABLE-RUN): remove once acqSettings_ is the single run-time source of truth.
+        updateSettings();
+
         double extraChannelOffset = 0.0;
         plc.prepareControllerForAcquisition(acqSettings_, extraChannelOffset);
         return true;

src/main/java/org/micromanager/lightsheetmanager/model/acquisitions/AcquisitionEngineScape.java

@@ -1021,6 +1021,17 @@ private boolean doHardwareCalculations(PLogicScape plc) {
            camera.setExposure(exposureTime);
         }
 
+        // FREEZE POINT: rebuild the run-time snapshot from asb_ before arming the controller.
+        // run() reaches this method (via doHardwareCalculations) BEFORE it rebuilds acqSettings_ with
+        // updateSettings(). Without this rebuild the controller is armed from a stale acqSettings_, e.g.
+        // useHardwareTimePoints left true by a prior short-interval reject, while the AcqEngJ event loop
+        // later runs from the fresh (false) value, so the hardware is armed for hardware timepoints while
+        // the software issues software timepoints ("stopped at first timepoint"). Rebuilding here is
+        // strictly fresher: doHardwareCalculations only mutates asb_ (the flag reset above + recalculated
+        // timing), never acqSettings_ directly.
+        // TODO(IMMUTABLE-RUN): remove once acqSettings_ is the single run-time source of truth.
+        updateSettings();
+
         double extraChannelOffset = 0.0;
         return plc.prepareControllerForAcquisition(acqSettings_, extraChannelOffset);
     }