Change makes sense, but from our chat it seems like the test plan failed

(I'll handle this, already working on it)

Rebase

Rebase

Got it, thanks for explaining!

Not looking too closely, seems rote

Lint fixes

Great!!

Updated to cover all diffs landed to master. Didn't end up being that hard actually

Rebase

There's actually a lot more I have to do here after the most recently landed diffs

Lint fixes

Lint fixes

1. Prettier fixes for last revision
2. Update tryflow link in socket.react.js

Rebase

Rebase

Rebase

Update promiseAll after seeing type errors from other invocations

Is reverting this hack tracked in any task? We should have a task that explicitly references the part of the code where the hack needs to be reverted.

Your tests aren't testing what I wanted to test, and I'm concered about their quality. You can keep them if you want, but you'll have to address the issues I highlighted inline. Might be easier to just get rid of them.

Confused by the word boundaries... passing back but let me know if I'm missing something

Make sure you use the SQL template

I don't understand why you're using word boundaries

Makes sense. What did you think of my test plan suggestion?

I'm worried that the test plan isn't thorough enough to make sure all of the permissions match up exactly. Would it be possible to write some unit tests that test creating some test chats (covering all thread types), then adding / removing each user-surfaced permission to each community-level role and making sure no changes occur?

Will this break logging out on older clients?

That makes sense, thanks for explaining. I think an important thing to note here is that we would need to introduce a new Provider here to handle caching globally (rather than per-component) in order to have parity with the old createSelector approach

Let me know if you want me to run this in production!