I am learning Java generics from a book. The book says that "Class Type Variables Are Not Valid in Static Contexts" and explains it with the following example.
Consider a generic class with type variables, such as Entry. You cannot use the type variables K and V with static variables or methods. For example, the following does not work:
public class Entry<K, V> {
// Compiler error - V in static context ("Entry.this' cannot be referenced from a static context")
private static V defaultValue;
// Compiler error - V in static context ("Entry.this' cannot be referenced from a static context")
public static void setDefault(V value) {
defaultValue = value;
}
}
After all, type erasure means there is only one such variable or method in the erased Entry class, and not one for each K and V.
I don't understand the above explanation. I tried to create the same code for K also and I got the same compile errors. Why is the above code illegal ?
Designers of Java generics chose to implement it using a mechanism called "type erasure". It means that generic specializations like Entry<String,Integer> and Entry<Integer,String> do not exist as separate classes. The type parameters are erased.
After you erase the type parameters from Entry<String,Integer> and Entry<Integer,String> you're left with just the Entry class.
If it were possible to have a static variable like defaultValue you would expect Entry<String,Integer>.defaultValue to be a Integer. And you would expect Entry<Integer,String>.defaultValue to be a String. But after type erasure only one Entry class with only one defaultValue variable, which now has to be both Integer and String. That's impossible. That's why you can't have a static variable of the generic type.