[regression] camera position changes when re-rendering after camera modification

[w1nklr]

The regression is between SubsurfaceViewer 1.14.15 and 1.15.0

The initial behavior is that after any user interaction with the camera, re-rendering (without new props) should NOT modify the camera position.
As of 1.15.0, there is a slight offset every time now :/

This is unfortunately not tested here, but in client code...

[hkfb]

@w1nklr Is this caused by the new View controller inertia default value? Could you try disabling it?

[w1nklr]

@w1nklr Is this caused by the new View controller inertia default value? Could you try disabling it?

I tried setting inertia to false, but it didn't change anything :(

For clarification, the tests is:

• display a scene
• modify camera position (mouse wheel)
• wait a bit
• compare with reference image
• re-render the scene
• compare with same reference image

Starting with version 1.15.0, the second image comparison fails :/
Re-rendering should NOT modify the point of view, so even some inertia should not have an impact.

[danielsantiago1002]

Hi, I tried to do a test for this. I see that the code has a function called "forceUpdate", so that's the function I'm using to test the rendering.

https://github.com/equinor/webviz-subsurface-components/blob/master/typescript/packages/subsurface-viewer/src/components/Map.tsx#L496

I created two branches in my repository for testing, one with version 1.15 of the subsurface viewer, and another with version 1.14.13.

• regression/camera-position-re-rendering https://github.com/danielsantiago1002/webviz-subsurface-components/tree/regression/camera-position-re-rendering
• test-subsurface-viewer-1.14.13 https://github.com/danielsantiago1002/webviz-subsurface-components/tree/test-subsurface-viewer-1.14.13

I created a function to take a screenshot of the DeckGL layer, and also two buttons to force the render and take the screenshot.

    const takeScreenshot = () => {
        if (deckRef.current && deckRef.current.deck) {
            const deck = deckRef.current.deck;
            // eslint-disable-next-line @typescript-eslint/no-explicit-any
            const canvas = (deck as any).canvas as HTMLCanvasElement | null;
            if (canvas) {
                const dataUrl = canvas.toDataURL("image/png");
                const link = document.createElement("a");
                link.download = "subsurface_viewer_snapshot.png";
                link.href = dataUrl;
                document.body.appendChild(link);
                link.click();
                document.body.removeChild(link);
            }
        }
    };

Is this a valid approach to test the problem? I would like you to verify it before drawing any conclusions.