Fix float-to-int truncation in QmlSensorOutputRange setters

QmlSensorOutputRange stores and exposes minimum, maximum and accuracy
as qreal, but its setters took int parameters. Passing the qreal values
from QSensor::outputRanges() therefore truncated them to integers,
losing all fractional precision.

Change setMinimum(), setMaximum() and setAccuracy() to take qreal.

Pick-to: 6.12
Change-Id: Ieae14469c1f65a6ee85cc7509d08d1cf04a8f4c9
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>