Fix conversion from float to quint16 in SunSpecDataPoint::convertFromFloatWithSSF

Calling this function with a negative value and DataType::Int16 always gives a 0 result on ARM. Need to convert to qint16 first to avoid undefined behavior (cf. https://embeddeduse.com/2013/08/25/casting-a-negative-float-to-an-unsigned-int/)

libnymea-sunspec/sunspecdatapoint.cpp

@@ -714,7 +714,7 @@ QVector<quint16> SunSpecDataPoint::convertFromFloatWithSSF(float value, qint16 s
         break;
     }
     case Int16: {
-        quint16 rawValue = static_cast<quint16>(value * pow(10, -1 * scaleFactor));
+        quint16 rawValue = static_cast<qint16>(value * pow(10, -1 * scaleFactor));
         rawData << rawValue;
         break;
     }