Use qreal and not double when appropriate

A qreal is a float on ARM architecture which may cause problems when qreals are used with types such as QPointF and then converted to double by QtLocation code. Such problems will only be noticed by ARM users since a qreal is a double on other architectures.