Variance Summary

Variance Concepts

• Describes subtyping relationships between complex types
• Three types of variance:
  1. Covariant: If S <: T then C(S) <: C(T)
  2. Contravariant: If S <: T then C(T) <: C(S)
  3. Invariant: No subtyping relationship

Java Array Covariance

• Arrays are covariant in Java
• Can lead to runtime errors
• Example:

  Integer[] intArray = new Integer[2];
  Object[] objArray = intArray;        // OK due to covariance
  objArray[0] = "Hello";              // Runtime error!
  

Problems with Array Covariance

• Type safety compromised
• Runtime errors possible
• Example of unsafe code:

  Shape[] shapes = new Circle[2];    // allowed
  shapes[0] = new Square(1.0);       // runtime error
  

Variance in Generics

• Generic types are invariant by default
• Example:

  List<Circle> circles = new ArrayList<>();
  List<Shape> shapes = circles;  // compile error
  

Safe Use of Variance

• Use generics instead of arrays when possible
• Be cautious with array covariance
• Consider using wildcards for flexibility
• Example:

  // Safe generic usage
  List<Circle> circles = new ArrayList<>();
  List<? extends Shape> shapes = circles;  // OK with wildcard
  

Best Practices

• Prefer generics over arrays
• Be aware of array covariance pitfalls
• Use wildcards appropriately
• Design for type safety
• Test edge cases with different types
• Document variance relationships