Yeah, that's a bit strange indeed.

I'll produce a simple test for you. But it's interesting that it doesn't happen on Intel, right?

Looks more like some AI hallucination to me. The only part in which jPCT is using GL_SCISSOR_TEST is in the context of render targets, which doesn't apply here.

I'll try to find some time to look into it during the holiday season. Do you actually have a basic test case that shows the issue?

I have a surprising piece of information: I've been trying hardware for running offline AI and I got an excellent computer with an Intel GPU. On AMD and NVidia that problem happens. On the Intel GPU it does not. So I asked that very computer and it suggested the following:

// After you have moved the camera or otherwise changed the viewport size:
GL11.glDisable(GL11.GL_SCISSOR_TEST);          // make sure the old scissor is gone
GL11.glViewport(0, 0, canvas.getWidth(), canvas.getHeight());
GL11.glEnable(GL11.GL_SCISSOR_TEST);           // re‑enable with the *new* dimensions
GL11.glScissor(0, 0, canvas.getWidth(), canvas.getHeight());


Maybe...I'll have to look into it to see what the overlay code does exactly.

The code is only the following. Is it possible that the Overlay class needs attention after this change to AWTGlCanvas?

       fullScreenOverlay = new Overlay(theWorld, 0, 50, (int)(buffer.getWidth()), (int)(buffer.getHeight()-50), "FullInterface");
       fullScreenOverlay.setTransparencyMode(Object3D.TRANSPARENCY_MODE_DEFAULT);
       fullScreenOverlay.setTransparency(250);


It's bizarre that you compiled. I'll see if I did something weird to the jar.

Yeah, it's like our repair just extended the drawing area. I'll show you the code tomorrow, but it should draw at the ends of the frame, but instead it's not centralized.

The last change regarding blit was 8 years ago when I moved from int-parameters to floats. No signature change other than that happened. I tried your line with the latest version, it compiles just fine. I'm not sure why you are having issues with it. What part of it is causing which problem exactly?

About the canvas: I don't know, if I even got the issue? Is it that it's somehow shifted to the left when it's supposed to be centered?

That line compiles with old jpct versions and not with the latest one.

What about the canvas?

I'm not sure what you mean by the missing blit call!? Missing compared to what? Actually, nothing has changed for ages in that regard.