Incorrect code analysis (unbalanced-hint)

[blakemcbride]

Apache NetBeans version

Apache NetBeans 15 release candidate

What happened

Java variable is being used but NB says it isn't. See two images .

How to reproduce

No response

Did this work correctly in an earlier version?

No / Don't know

Operating System

Fedora Linux

JDK

8

Apache NetBeans packaging

Apache NetBeans provided installer

Anything else

No response

Are you willing to submit a pull request?

No

Code of Conduct

Yes

[neilcsmith-net]

Confirmed. This appears to be a regression caused by #4421 as far as I can tell. cc @mbien can you take a look?

[mbien]

seems to be specific to final array fields with write ops.

    private final static char[] bar1 = {'y', 'n'};  // marked unused
    private final static char[] bar2 = new char[2]; // ok

    public static void main(String[] args) {
        
        final char[] foo1 = {'y', 'n'};             // marked unused
        final char[] foo3 = new char[] {'y', 'n'};  // marked unused
        
              char[] foo2 = {'y', 'n'};             // ok
        final char[] foo4 = new char[2];            // ok
        
        char[] foo5;                                // ok
        foo5 = new char[] {'Y', 'n'};
        
        System.out.println(foo1);
        System.out.println(foo2);
        System.out.println(foo3);
        System.out.println(foo4);
        System.out.println(foo5);
        System.out.println(bar1);
        System.out.println(bar2);
        
    }

reverting the array handler of #4421 fixes this issue but it would break #4402 again for arrays. Can't come up with a quick way of fixing this atm. Shame we haven't noticed this earlier so we could have simply reverted the PR.

diff --git a/java/java.hints/src/org/netbeans/modules/java/hints/bugs/Unbalanced.java b/java/java.hints/src/org/netbeans/modules/java/hints/bugs/Unbalanced.java
index 51829b5..3c3d3ae 100644
--- a/java/java.hints/src/org/netbeans/modules/java/hints/bugs/Unbalanced.java
+++ b/java/java.hints/src/org/netbeans/modules/java/hints/bugs/Unbalanced.java
@@ -152,7 +152,7 @@
                 if (secondAccess != null) {
                     record(ctx.getInfo(), var, secondAccess);
                 }
-            } else if (!var.getModifiers().contains(Modifier.FINAL)) {
+            } else {
                 record(ctx.getInfo(), var, State.WRITE, State.READ);
             }

[neilcsmith-net]

Thanks @mbien I think the whole PR might need reverting - there's a problem with collections as well as arrays. This shows a warning in 15 but not in 14.

public class Test1 {

    private final List<String> collection = new ArrayList<>();

    public Test1() {
        use(collection);
    }

    public boolean isEmpty() {
        return collection.isEmpty();
    }
    
    private void use(List<String> list) {
        list.add("FOO");
    }
}

We might need some tests for not containing this warning too? :smile:

Shame we haven't noticed this earlier so we could have simply reverted the PR.

Shame none of the 4 requested reviewers looked at it either!

[errael]

Here's something which is probably the same issue.

public void foo(List<?> list) {
    final Object[] listeners = list.toArray(); //<<<<<<<<<< hint on listeners
    Objects.requireNonNull(listeners);
}

The first listeners get the hint

The array is only written to, never read from

Remove the "final" and the hint goes away

[swpalmer-cl]

Might be related to this issue:

  static record Thing(boolean isOkay) {}
  void neverRead(Thing thing) {
    List<Thing> things = new ArrayList<>(); // NB claims only added to, never read
    things.add(thing);
    things.forEach(System.out::println); // but here it is read
  }