[test] disable animations in unit testing

[lauri865]

• I have followed (at least) the PR section of the contributing guide.

An experiment to remove transitions from unit tests in order to make tests more stable. Lots of tests breaking due to act errors, which more often than not are related to the Transition component of react-transition-group. Would be good to disable it globally, and get rid of unnecessary waitFor calls as a result to make testing faster as well as more stable.

I don't have more time to dig into this now, so will leave it at this, if anyone wants to explore it further. Some tests broke as a result.

[mui-bot]

Deploy preview: https://deploy-preview-16288--material-ui-x.netlify.app/

Generated by 🚫 dangerJS against 1e8be24

[LukasTy]

Suggested change

import * as React from 'react';

[LukasTy]

@lauri865 @mui/xgrid WDYT about the updated approach of https://reactcommunity.org/react-transition-group/testing/? 🤔
Some problematic tests would have to be adjusted. 🙈

[lauri865]

@lauri865 @mui/xgrid WDYT about the updated approach of https://reactcommunity.org/react-transition-group/testing/? 🤔 Some problematic tests would have to be adjusted. 🙈

Ah, nice find! Seems to break less tests it seems as well.

Wonder why they suggested an alternative approach here? 🤔
https://testing-library.com/docs/example-react-transition-group/

[LukasTy]

Ah, nice find! Seems to break less tests it seems as well.

Wonder why they suggested an alternative approach here? 🤔 https://testing-library.com/docs/example-react-transition-group/

I am not sure; maybe they are suggesting going for maximum efficiency and completely mocking the module to avoid any potential overhead.
I've used sinon.stub in the latest implementation to assert this idea. 😉
There are some Data Grid browser tests failing after this. 🤔

[lauri865]

Ah, nice find! Seems to break less tests it seems as well.
Wonder why they suggested an alternative approach here? 🤔 https://testing-library.com/docs/example-react-transition-group/

I am not sure; maybe they are suggesting going for maximum efficiency and completely mocking the module to avoid any potential overhead. I've used sinon.stub in the latest implementation to assert this idea. 😉 There are some Data Grid browser tests failing after this. 🤔

I'm not insinuating it's better. I found it difficult to fake the components due to some state being dependent on transition life-cycle events, hence I overrode props instead.

Are you use the latest one works? Transition component children are functions in many instances, and this should error out ("functions are not valid children" yaddayadda).

mui-x/test/utils/mochaHooks.js

Line 24 in 583b078

function FakeTransition({ children }) { return <React.Fragment>{children}</React.Fragment>; };

If the config prop works, I think it's probably the best bet.