Test passing locally, some are flaky in CI

[joepio]

When running on my macbook, the e2e tests are running fine. However, on the (far slower) CI machine some of them fail pretty often.

Runs

• https://686d0f5d62e047882bb195b8--atomic-tests.netlify.app

Problems

• [ ] ChatRoom second window doesn't show the new message. Maybe because the subscribe is processed after the commit has been fired. I think a robust solution is to send all the relevant, potentially missed commits (from the past x time) to the client. Or maybe just send the current RESOURCE to the client on subscribe? Could also work. A dirty hack is to add a timeout or something, or maybe add some retries if its flaky.
• [ ] Bookmark opens the wrong drive, atomic instead of localtest-something. Doesn't seem to happen when running locally. Maybe some SERVER_URL / FRONTEND_URL related issue?
• [ ] Sveltekit pnpm link failing
• [ ] New Drive can't be found, because there are too many in the dropdown. We should clean these up.
• [ ] Filepicker can't find txt file. I think this is because our testFilePath function is unreliable / not properly configured. I'd like a more robust solution.
• [ ] editProfileAndCommit fails when list of drives is too long, the click doesnt work. Try scrollIntoViewIfNeeded. Maybe a cleanup would be better though.