(WIP) Exploration/prototyping of test utils package

[LFDanLu]

Closes

✅ Pull Request Checklist:

• Included link to corresponding React Spectrum GitHub Issue.
• Added/updated unit tests and storybook for this change (for new code or code which already has tests).
• Filled out test instructions.
• Updated documentation (if it already exists for this component).
• Looked at the Accessibility Practices for this feature - Aria Practices

📝 Test Instructions:

🧢 Your Project:

RSP

[jluyau]

i think these utils would definitely make writing tests for rsp components much cleaner. another utility class that would help would be interacting with the current dialog (methods to get elements within the content body as well as pressing the dialog buttons)

[jluyau]

this is probably on your radar, but having these mocks and such callable with 1 line of code would be nice too!

[LFDanLu]

Yep, I've been mulling over whether or not these mocks are too broad. Ideally we would limit what it actually gets applied on, similar to

react-spectrum/packages/@react-aria/overlays/test/useOverlayPosition.test.tsx

Lines 244 to 252 in 1c32cd5

	jest.spyOn(window, 'getComputedStyle').mockImplementation(element => {
	if (element.attributes.getNamedItem('data-testid')?.value === 'container') {
	const sty = realGetComputedStyle(element);
	sty.position = 'relative';
	return sty;
	} else {
	return realGetComputedStyle(element);
	}
	})

, but I'm worried that users may need to have finer control over the mocks.

Perhaps we can provide a basic util that does the simple mock and offer a way to override/customize

[LFDanLu]

Another interesting idea that @devongovett brought up was making our virtualized components render as non-virtualized components when rendered in a test environment specifically. That means we won't have to mock any of this stuff, thus avoid possible mock collisions and assumptions. RSP would still write tests with the virtualized components still being virtualized just to make sure things work as expected

[jluyau]

one helper function that i think would help (since it's come up a bunch) is a method to select a given cell, though i guess it'd require the column and row indexes or text

[LFDanLu]

sounds good, we have a util for that that we use internally as well actually haha. I think we can start off with searching for cell text first since that is what a user would look for typically and see about supporting a (x row, y column) alternative maybe.

[LFDanLu]

Closing this for now, will incorporate the feedback when work resumes on this portion of the testing util work