Cannot load a GLTF model in a unit test

Questions
, ,
#1

Hi,

I’m currently writing unit tests for my app and am running into some issues getting the models to load. Here is what I’m trying to achieve:

import React from 'react';
import ReactThreeTestRenderer from '@react-three/test-renderer';
import { useGLTF } from '@react-three/drei';
import MyComponent from './MyComponent';

let model;

beforeAll(() => {
  model = useGLTF('my_model.gltf').nodes;
});

test('mesh to have one child', async () => {
  const renderer = await ReactThreeTestRenderer.create(<MyComponent model={model} />);

  const mesh = renderer.scene.children[0].allChildren;
  expect(mesh.length).toBe(1);
});

Running this test yields the output of:

 FAIL  src/components/MyComponent.test.js
  ✕ mesh to have one child (1 ms)

  ● mesh to have one child

    thrown: Promise {}

       6 | let model;
       7 |
    >  8 | beforeAll(() => {
         | ^
       9 |   model = useGLTF('my_model.gltf').nodes;
      10 | });
      11 |

      at Object.<anonymous> (src/components/MyComponent.test.js:8:1

I tried using useLoader, but got similar results.

For context, MyComponent in this example is the component that takes the loaded model as a prop and applies various mesh properties and methods to it, rendering it in the process. For various reasons, useGLTF is not called inside MyComponent. Rather, it is called in a separate function that loads other models as well.

I tried many iterations, but couldn’t get anything to work. I worst comes to worst, I can always just fallback on a default Sphere or something as the model to be passed, but mainly curious as to why an approach like this is not working.

Thanks!

#2

May be wrong here - but isn’t useGLTF required to be wrapped around in <Suspense>, which in this case it wouldn’t be ?

#3

useGLTF is used multiple times elsewhere without issue. I have a function that uses this hook, which when called by the app returns the expected loaded models. Even calling that exact same function when running tests returns the same error as above :confused: