QState::addTransition(sender, signal, target) should accept null as target

States with no target are completely legitmate and the docs clearly state that.
It is however an error to use

QSignalTransition *QState::addTransition(QObject *sender, const char *signal, AbstractState *target)

without setting a target, although if one creates new QSignalTransition (which is what the function does anyways) and does not set it's target (or set it to null) it is just fine.

This is unexpected asymmetry, one should also be able to pass a null to addTransition.