Why does javac not catch this ternary error?

Question

Clearly returning null for a boolean type is an error. Why doesn't javac catch this?

public class Test {
    public static void main(String[] args) {
        Object a = null;
        System.out.println(getThing(a));
    }

    private static boolean getThing(Object o) {
        return o == null ? null : Boolean.TRUE;
    }
}

Possible duplicate of [Tricky ternary operator in Java - autoboxing](https://stackoverflow.com/questions/8098953/tricky-ternary-operator-in-java-autoboxing) — 11thdimension, Jul 14 '17 at 19:14
This is also discussed here, but I didn't include autoboxing in my search so I didn't find it. [Java autoboxing and ternary operator madness](https://stackoverflow.com/questions/25417438/java-autoboxing-and-ternary-operator-madness) — macawm, Jul 14 '17 at 19:28

score 1 · Accepted Answer · answered Aug 07 '22 at 22:17

The question

Clearly returning null for a boolean type is an error. Why doesn't javac catch this?

Given that question, here's the relevant part of the code:

private static boolean getThing(Object o) {
    return o == null ? null : Boolean.TRUE;
}

That method includes a conditional expression (JLS 15.25) where:

the first expression is o == null
the second expression is null
the third expression is Boolean.TRUE

Why it compiles

The reason javac allows it is:

The conditional expression is classified as a reference conditional expression (JLS 15.25.3) due to the second and third expressions "null" and "Boolean.TRUE", excerpted from JLS 15.25:

If both the second and the third operand expressions are boolean expressions, the conditional expression is a boolean conditional expression.

If both the second and the third operand expressions are numeric expressions, the conditional expression is a numeric conditional expression.

Otherwise, the conditional expression is a reference conditional expression.
The entire conditional expression is classified as type java.lang.Boolean
The conditional expression is of reference type Boolean, and the method return type – primitive "boolean" – can be unboxed from Boolean

The compiler kind of "allowing" a code path where "null" might be unboxed to "boolean" isn't all that different from various other obvious programming errors.

For example, here's code that creates a new object, gives it a value of null, then tries to invoke a method on that object. This code compiles fine but, unsurprisingly, throws an exception at runtime:

Object o = null;
System.out.println(o.hashCode());

Exception in thread "main" java.lang.NullPointerException:
Cannot invoke "Object.hashCode()" because "o" is null

It compiles, but doesn't run

Though your code compiles, the following exception is thrown at runtime when null is unboxed to boolean:

Exception in thread "main" java.lang.NullPointerException:
Cannot invoke "java.lang.Boolean.booleanValue()"

While the return value of the statement is null – a valid value for Boolean – unboxing to a boolean primitive (to line up with the method return type) throws an exception. Unboxing tries to call null.booleanValue() which won't work.

You weren't asking, but – an easy fix would be to change the return type of the method from primitive boolean to Boolean. After doing that, your program compiles and runs, printing "null" (from System.out.println(getThing(a)).

Why does javac not catch this ternary error?

1 Answers1

The question

Why it compiles

It compiles, but doesn't run