Issue1543

>>> from jarray import array
>>> from java.lang import Object
>>> x = array([0,1,2], Object)
>>> x.append(3)
>>> x
array(java.lang.Object, [0, 1, 2, 3])
>>> y = array([x], Object)
>>> y
array(java.lang.Object, [array(java.lang.Object, [0, 1, 2, 3, None, None, None])])
>>> x == y[0]
False

The append on x causes some extra array space to be allocated, but apparently the end of the array isn't remembered when assigned into y. However, if we use PyObject instead of Object...

>>> x = array([0,1,2], object)
>>> x.append(3)
>>> x
array(org.python.core.PyObject, [0, 1, 2, 3])
>>> y = array([x], object)
>>> y
array(org.python.core.PyObject, [array(org.python.core.PyObject, [0, 1, 2, 3])])
>>> x == y[0]
True

In case it matters:
$ java -version
java version "1.6.0_14"
Java(TM) SE Runtime Environment (build 1.6.0_14-b08)
Java HotSpot(TM) 64-Bit Server VM (build 14.0-b16, mixed mode)

The problem here is the underlying array x does indeed get resized to a size of 7 when appended to, because the capacity increases of arrays always double the size of the array first (I'm not sure why, or if that's really desirable for the array type in the first place)

This is exposed in y because extending an array with an iterable apparently calls Py.tojava() on all the items of the iterable. That turns the x array into a real java array, then back to a Python array again, but without the precise size information it originally had

Fixed in r7147