Question

I’ve seen numerous implementations of the INotifyPropertyChanged (Automatically INotifyPropertyChanged) that have support for the UnaryExpression in addition to the MemberExpression. Examples of the UnaryExpression include uses of unary operators within the expression such as,

RiasePropertyChanged(() => !MyPropertyName);

Is there reason enough to warrant its inclusion in the property name extraction? I had supposed the support for the UnaryExpression is because you could extract the MemberExpression from this.

Does the community have any pro/cons related this this implementation as it differs from the Prism NotificationObject?

Answer 1

I’ve only seen this in situations where someone wanted to support a wide variety of ways to refer to the member. Limiting it to UnaryExpression limits you slightly.

It doesn’t really matter… adding support for both will be primarily a performance concern, but it will likely not make much difference in the end. I’d say support both just so you don’t risk having the usage be more of a leaky abstraction than it already is.

Here’s what we use:

public void OnPropertyChanged<TProperty>(Expression<Func<TProperty>> property)
{
    var lambda = (LambdaExpression)property;
    MemberExpression memberExpression;

    if ( lambda.Body is UnaryExpression )
    {
      var unaryExpression = (UnaryExpression)lambda.Body;
      memberExpression = (MemberExpression)unaryExpression.Operand;
    }
    else
    {
      memberExpression = (MemberExpression)lambda.Body;
    }
    OnPropertyChanged(memberExpression.Member.Name);
}