[vtk-developers] bug in java-wrapping for vtkHyperOctree

[Amy Squillacote]

As you said, the problem is in VTK-generated code, so I can't just 
re-write it.

The possible fixes I had thought of for this are as follows.

    * Move vtkCompactHyperOctreeCursor into its own .h/.cxx files. Then
      it would be java-wrapped.
    * Somehow change the java-wrapping for the NewCellCursor method of
      vtkHyperOctree so that it does tempObj.NewInstance() instead of
      trying to create a new class from java given the class name. (How
      hard would it be to change the wrapping just for that method?)
    * Keep the vtkCompactHyperOctreeCursor generated by the
      NewCellCursor method internal to vtkHyperOctree (i.e., not return
      a pointer to it). Then the SubdivideLeaf and CollapseTerminalNode
      could use this internal vtkCompactHyperOctreeCursor* instead of
      being passed an instance of it.

- Amy

David Cole wrote:
> Scratch the stupid response. Never mind. I thought that was java code 
> you had written, not the generated stuff....
>
> <sheepish_grin />
>
>
> On Mon, Nov 3, 2008 at 3:59 PM, David Cole <david.cole at kitware.com 
> <mailto:david.cole at kitware.com>> wrote:
>
>     Can you call "tempObj.NewInstance()" from java? (Instead of
>     explicitly creating an object of a given class, use the existing
>     instance to give you an "empty" new one of the same type... does
>     that work?)
>
>
>
>     On Mon, Nov 3, 2008 at 3:55 PM, Amy Squillacote <ahs at cfdrc.com
>     <mailto:ahs at cfdrc.com>> wrote:
>
>         Hi all,
>
>         We are trying to use vtkHyperOctree from java. We have run
>         into problems with the NewCellCursor method. In c++, this
>         method returns a vtkCompactHyperOctreeCursor, a subclass of
>         vtkHyperOctreeCursor. The code below is copied from the
>         NewCellCursor method in the java-wrapped version of
>         vtkHyperOctree.
>
>          public vtkHyperOctreeCursor NewCellCursor() {
>           long temp = NewCellCursor_22();
>
>           if (temp == 0) return null;
>           vtkHyperOctreeCursor obj = null;
>           java.lang.ref.WeakReference ref =
>         (java.lang.ref.WeakReference)vtkGlobalJavaHash.PointerToReference.get(new
>         Long(temp));
>           if (ref != null) {
>             obj = (vtkHyperOctreeCursor)ref.get();
>           }
>           if (obj == null) {
>             vtkHyperOctreeCursor tempObj = new vtkHyperOctreeCursor(temp);
>             String className = tempObj.GetClassName();
>             try {
>               Class c = Class.forName("vtk." + className);
>               java.lang.reflect.Constructor cons =
>         c.getConstructor(new Class[] {long.class} );
>               obj = (vtkHyperOctreeCursor)cons.newInstance(new
>         Object[] {new Long(temp)});
>             } catch (Exception e) {
>               e.printStackTrace();
>             }
>             tempObj.Delete();
>           }
>           return obj;
>          }
>
>         The problem occurs on the following line: Class c =
>         Class.forName("vtk." + className);. The className variable is
>         set to vtkCompactHyperOctreeCursor, but this class is not
>         java-wrapped; it is internal to vtkHyperOctree.cxx. Thus this
>         line causes a java  ClassNotFoundException.
>
>         We need to be able to call NewCellCursor() from java in order
>         to use the SubdivideLeaf and CollapseTerminalNode methods,
>         both of which take a vtkHyperOctreeCursor as an argument. It
>         is not possible to directly create an instance of
>         vtkHyperOctreeCursor because it is an abstract class.
>
>         Ideas? Suggestions?
>
>         - Amy
>
>         -- 
>         Amy Squillacote                    Phone: (256) 726-4839
>         Computer Scientist                 Fax: (256) 726-4806
>         CFD Research Corporation           Web: http://www.cfdrc.com
>         215 Wynn Drive, Suite 501
>         Huntsville, AL  35805
>
>
>         _______________________________________________
>         vtk-developers mailing list
>         vtk-developers at vtk.org <mailto:vtk-developers at vtk.org>
>         http://www.vtk.org/mailman/listinfo/vtk-developers
>
>
>

-- 
Amy Squillacote                    Phone: (256) 726-4839
Computer Scientist                 Fax: (256) 726-4806
CFD Research Corporation           Web: http://www.cfdrc.com
215 Wynn Drive, Suite 501
Huntsville, AL  35805