Using Linq for statically-typed reflection

I discovered the TypedReflector
on Daniel Cazzulino’s Blog. In essence, a simplified version of the code looks like:

public static class Reflector
{
 public static PropertyInfo GetProperty<T, TRet>(Expression<Func<T, TRet>> expression)
 {
  return GetMemberInfo(expression) as PropertyInfo;
 }

 public static FieldInfo GetField<T, TRet>(Expression<Func<T, TRet>> expression)
 {
  return GetMemberInfo(expression) as FieldInfo;
 }

 public static MethodInfo GetMethod<T, TRet>(Expression<Func<T, TRet>> expression)
 {
  return GetMemberInfo(expression) as MethodInfo;
 }

 public static MemberInfo GetMemberInfo(Expression memberInfoExpression)
 {
  Expression lambdaBodyExpression = ((LambdaExpression)memberInfoExpression).Body;
  switch (lambdaBodyExpression.NodeType)
  {
   case ExpressionType.MemberAccess:
    return ((MemberExpression)lambdaBodyExpression).Member;
   case ExpressionType.Call:
    return ((MethodCallExpression)lambdaBodyExpression).Method;
   default:
    throw new ArgumentException("Unsupported NodeType");
  }
 }
}

This class allows me to write the following:

PropertyInfo namePropertyInfo = Reflector.GetProperty<Person, string>(p => p.Surname);

Not only can i be sure that the lambda (and thus the namePropertyInfo.Invoke) will always return a string, i can also use it to avoid ‘hardcoded’ propertynames:

//const string SurnameProperty = "Surname";
//comboBox1.DisplayMember = SurnameProperty;
comboBox1.DisplayMember = Reflector.GetProperty<Person, string>(p => p.Surname).Name;