Simply, when the example starts up, the behavior is as expected and wanted. But there are two (or three) problems. Once you zoom in, the filled rectangle actually disappears (possibly in negative z-space) which makes the grid, axis and box become uncleared and thus "redrawn". Zooming back out redraws the rectangle, but zooming too far out makes the 3D scene (box, grid, axis, ...) being split by the rectangle (clearly showing the rectangle here). Zooming any further out, again, makes the rectangle disappear and everything pile up.

My guess is that the rectangle is being drawn in 3D, while it should be 2D. Using background() instead clears the issue, but defies the use of an offscreen buffer (and using it as an alpha fill).

An ideal situation would be, in the draw function:

1. canvas.beginDraw(); // GLGraphicsOffscreen
2. 
   
3. // Begin 2D drawing // Even using beginScreenDrawing() does not work
4. canvas.fill(0xFFFF0000, 204); // 80% alpha
5. canvas.rect(0, 0, canvas.width, canvas.height); // Not applet.rect() or just rect(), but using the canvas, although this seems to work just fine too
6. 
   
7. // Begin 3D drawing
8. scene.beginDraw();
9. 
   
10. canvas.fill(204, 102, 0);
11. canvas.box(20, 30, 50);
12. scene.endDraw();
13. 
    
14. canvas.endDraw();
15. image(canvas.getTexture(), 0, 0, canvas.width / 2, canvas.height / 2);

I'm trying to add new vertices to an already existing GLModel, but it's giving me the following error:

Invalid memory access of location 0x0 eip=0x949c43f8.

I can add vertices in the same function as where I create the model, but not for example 5 seconds later in a different function. Whenever I call a function on the model that alters it, the sketch crashes (even when using the same function that works right after instantiating the model).

For instance:

_faceModel.beginUpdateVertices(); => crash

_faceModel.updateVertices(verts); => crash

...

But only when called after x seconds and from a different function.

Any tips on where to go from here? Some example code below.

I don't know if creating hundres of GLModels is the way too go? I originally thought of putting all of them in one model, but if this doesn't work... Just want to render a hundred spheres or so with OpenGL instead of the basic renderer.

—

So again: calling createSphere right after instantiating the model works fine, even when called 50 times. Calling createSphere after some time crashes the application. I've tried putting it between renderer.beginGL() or putting the same code somewhere else, but to no avail.

1. // Enough particles to hold all future vertices
2. _faceModel = new GLModel(_applet, 100000, GLModel.TRIANGLES, GLModel.STREAM);
3. createSphere(_applet.random(0, 200), _applet.random(0, 200), _applet.random(-200, 200), DEFAULT_NODE_START_DIAMETER, DEFAULT_SPHERE_RESOLUTION);

1. private void createSphere(float $x, float $y, float $z, int $radius, int $resolution) {
2. Sphere sph = new Sphere(new Vec3D($x, $y, $z), $radius);
3. TriangleMesh tm = (TriangleMesh) sph.toMesh($resolution);
4. 
   
5. float[] verts = tm.getMeshAsVertexArray();
6. int numV = verts.length / 4;
7. 
   
8. _faceModel.beginUpdateVertices(); // <<<< Crash here, but not on first runs
9. for(int i = 0; i < numV; i++)
10.       _faceModel.updateVertex(i + numVertices, verts[4 * i], verts[4 * i + 1], verts[4 * i + 2]);
11. _faceModel.endUpdateVertices();
12. 
    
13. numVertices += numV;

Using the Carnivore library, I'm listening to all traffic in my network. Now, I want to take the bytes I receive and re-compose them into an image, either complete or corrupt. The important thing is that I can visualise the data, if it's an image.

Now, I thought this was going to be easier (I always do). Apparently, and logically, the image doesn't come through in one pass. It loads in several, creating multiple requests etc. Basically, I can't get the image out in one piece. So it seems I have two options:

1 – Take whatever bytes I receive and try to create an image out of them. But this doesn't seem te be working. Even though I've got the headers, the ImageIO can't make something out of it. I've searched far and wide, but neither one works. For example:

1. try { 
2. BufferedImage abdf = ImageIO.read(new ByteArrayInputStream(packet.data));
3. System.out.println(abdf);
4. } catch(IOException e) {
5. e.printStackTrace();
6. }

2 – Store all data unique to a single transmission and, when finished, re-compose it. But it doesn't seem the library has any properties that define a single transmission, just the packets. And even then, what if a user loads a website? I can't reconstruct those images if it's one big blob of various items.

So, where to go from here? My deadline is approaching and I've only begun :D Cheers!

EDIT See http://www.creativeapplications.net/linux/fohnseher-linux/ for an example.