chore: add jest config for ui package, convert Button tests to ts and fix existing assertions

[brandongregoryscott]

What kind of change does this PR introduce?

Chore - reinstates tests for the ui package.

What is the current behavior?

See #13540 (comment) - it sounds like tests were started at some point (maybe in the https://github.com/supabase/ui repo) but weren't maintained once ported back into the monorepo.

What is the new behavior?

This PR adds a minimal jest config + refactors the Button component tests to pass based on its current code. It also adds a Github Action for running the tests in the packages/ui directory when any file in it changes.

 PASS  src/components/Button/Button.test.tsx
  #Button
    ✓ should render button correctly (37 ms)
    ✓ should render different text (24 ms)
    ✓ should ignore events when disabled (16 ms)
    ✓ should ignore events when loading (20 ms)
    ✓ should have "w-full" class when block is true (68 ms)
    ✓ should have "primary" class from theme (29 ms)
    ✓ should have "default" class from theme (20 ms)
    ✓ should have "secondary" class from theme (9 ms)
    ✓ should have "alternative" class from theme (11 ms)
    ✓ should have "outline" class from theme (13 ms)
    ✓ should have "dashed" class from theme (14 ms)
    ✓ should have "link" class from theme (18 ms)
    ✓ should have "text" class from theme (13 ms)
    ✓ should have "danger" class from theme (5 ms)
    ✓ should have "warning" class from theme (6 ms)
    ✓ should have "tiny" class from theme (15 ms)
    ✓ should have "small" class from theme (8 ms)
    ✓ should have "medium" class from theme (7 ms)
    ✓ should have "large" class from theme (6 ms)
    ✓ should have "xlarge" class from theme (9 ms)

Additional context

In an effort to keep this PR smaller and incrementally reinstate test suites / add new ones, I restricted the testRegex to just *.test.tsx files. I left a comment in the config file on why, but the existing tests are likely going to need some refactoring to pass anyway.

This PR also introduces a change to the way the Button component exposes its ref, since it was causing issues when wrapped with the Next Link component - I added a test for this as well. It's kinda hard to tell where/if the current ref was being used, but I think this should work.

[vercel]

@brandongregoryscott is attempting to deploy a commit to the Supabase Team on Vercel.

A member of the Team first needs to authorize it.

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Comments	Updated (UTC)
docs	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	May 16, 2023 5:04pm
studio	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	May 16, 2023 5:04pm
studio-self-hosted	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	May 16, 2023 5:04pm
studio-staging	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	May 16, 2023 5:04pm
supabase-studio-prod	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	May 16, 2023 5:04pm
supabase-studio-staging	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	May 16, 2023 5:04pm
zone-www-dot-com	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	May 16, 2023 5:04pm

[brandongregoryscott]

This test was failing before the change to the ref - doesn't seem related to passing a custom element with the as prop as originally theorized in #13540 (comment)

[saltcod]

Hey @brandongregoryscott ! had a look at this — your ui tests look good! But studio tests are failing on this branch. Any ideas?

[brandongregoryscott]

Hey @saltcod - the studio tests are passing for me locally. Looks like they failed in Github Actions but not for the same reason (JavaScript heap error)

Edit: Alright, seeing the same heap errors on my fork. Might be a resource limitation I'm not seeing on my local machine. Let me take another look at these. I'm seeing it happen on the latest master from my fork too: https://github.com/brandongregoryscott/supabase/actions/runs/4986094913/attempts/1 and https://github.com/brandongregoryscott/supabase/actions/runs/4986094913

Those tests look time based - do we think they're just flaky? Not sure I have permission to retry the tests

[brandongregoryscott]

Following-up: The heap issue was happening on my fork off of the fresh master, too. Not sure why it's happening all of a sudden, could just be a tipping point. Following the instructions in this comment, this seems to resolve the CI issues: https://github.com/brandongregoryscott/supabase/actions/runs/4986366186/jobs/8927026785