[java-gnome-hackers] Problem with bindings generator

[Stefan P. <st...@pr...>]

Hi all,

I just stumbled over something in the generated source code for
GtkTreePath.c

1. GtkTreePath.getIndices(TreePath) always returns null
========================================================
The method 

JNIEXPORT jintArray JNICALL
Java_org_gnome_gtk_GtkTreePath_gtk_1tree_1path_1get_1indices

is supposed to return an integer array. Calling it from GtkTreePath.java
only returned a NullPointer. Inspecting the code I found
	// translate return value to JNI type
	_result = (jintArray) NULL;
which I think explains my problems :)
So, question number 1: Why was the line generated that way?

2. Modified GtkTreePath.getIndices(TreePath) succeeds,
   but later int array usage emits NullPointerException
========================================================
When I change the line to
	_result = (jintArray) result;
the application crashes. JUnit tests for TreePath.getIndices() fail
in the third line of the following code:
        int[] ind = path.getIndices();
        if (ind==null) System.out.println("Indices are null");
        System.out.println(ind);
with the following message:
  Encoutered an error at ValidateTreeStore.java:430
  The following exception was thrown:

     java.lang.NullPointerException

  Normally exceptions in unit tests are trapped and reported as
  failures;
  this was unexpected. It could be a bug or something deeper with your
  environment or setup.

Running the test in Eclipse gives a stacktrace that reveals that the
NullPointer originates in java.lang.String.valueOf(Object) in the one
and only line
    public static String valueOf(Object obj) {
	return (obj == null) ? "null" : obj.toString();
    }

The only chance that this can produce a NullPointerException is
Object.toString() failing on the int array.

My only idea is that the int[] returned from get_indices has been
cleaned up in the meantime.

So question 2. is simply for ideas on this