|
| 1 | +import java.lang.reflect.Field; |
| 2 | + |
| 3 | +public class FinalFieldDemo { |
| 4 | + |
| 5 | + public static void main(String[] args) { |
| 6 | + |
| 7 | + try { |
| 8 | + User user = new User("Somesh"); |
| 9 | + |
| 10 | + System.out.println("Before reflection: " + user.name); |
| 11 | + |
| 12 | + Field field = User.class.getDeclaredField("name"); |
| 13 | + field.setAccessible(true); // deep reflection |
| 14 | + field.set(user, "Changed Name"); // JDK 26 warns by default |
| 15 | + |
| 16 | + System.out.println("After reflection: " + user.name); |
| 17 | + |
| 18 | + } catch (Exception e) { |
| 19 | + e.printStackTrace(); |
| 20 | + } |
| 21 | + } |
| 22 | +} |
| 23 | + |
| 24 | +class User { |
| 25 | + final String name; |
| 26 | + |
| 27 | + User(String name) { |
| 28 | + this.name = name; |
| 29 | + } |
| 30 | +} |
| 31 | + |
| 32 | +/* |
| 33 | +What changed: Previous vs New |
| 34 | +
|
| 35 | +Previous Java style: |
| 36 | +- final fields could still be changed using deep reflection |
| 37 | +- setAccessible(true) + Field.set(...) could bypass immutability |
| 38 | +- final was not fully enforced at runtime |
| 39 | +
|
| 40 | +New Java 26 style: |
| 41 | +- JDK 26 warns when code mutates final fields through deep reflection |
| 42 | +- This prepares for a future release where such mutation will be denied by default |
| 43 | +- Goal: make final actually mean final |
| 44 | +
|
| 45 | +Why the new approach is better: |
| 46 | +- Better integrity |
| 47 | +- Better security |
| 48 | +- More reliable immutability |
| 49 | +- Allows stronger JVM optimizations |
| 50 | +
|
| 51 | +Pros: |
| 52 | +1. Safer code |
| 53 | + - Prevents unexpected mutation of final fields |
| 54 | +
|
| 55 | +2. Better reliability |
| 56 | + - final fields behave more like developers expect |
| 57 | +
|
| 58 | +3. Better JVM optimization potential |
| 59 | + - Immutable fields are easier to optimize |
| 60 | +
|
| 61 | +4. Stronger platform integrity |
| 62 | + - Reflection loophole is being closed |
| 63 | +
|
| 64 | +Cons: |
| 65 | +1. Legacy libraries may warn |
| 66 | + - Old frameworks using reflection may need updates |
| 67 | +
|
| 68 | +2. Migration effort |
| 69 | + - Some tools depend on final-field mutation |
| 70 | +
|
| 71 | +3. Behavior change over time |
| 72 | + - Warning now, stricter denial later |
| 73 | +
|
| 74 | +Best use case: |
| 75 | +- Understanding JEP 500 |
| 76 | +- Learning why reflection-based mutation is discouraged |
| 77 | +- Testing migration impact in older codebases |
| 78 | +
|
| 79 | +Compile and run: |
| 80 | +javac FinalFieldDemo.java |
| 81 | +java FinalFieldDemo |
| 82 | +
|
| 83 | +Helpful testing options in JDK 26: |
| 84 | +java --illegal-final-field-mutation=warn FinalFieldDemo |
| 85 | +java --illegal-final-field-mutation=debug FinalFieldDemo |
| 86 | +java --illegal-final-field-mutation=deny FinalFieldDemo |
| 87 | +java --enable-final-field-mutation=ALL-UNNAMED FinalFieldDemo |
| 88 | +*/ |
0 commit comments