Can you retitle all of the diffs in this stack to explicitly reference message deletion? Just realized that they're just generically titled.

Great work figuring this out! I have a lot of comments, but none of them are about core logic... mostly asking for more code comments, some renames, and some refactors

In D14462#402560, @tomek wrote:

There is one place I'm not sure of: threadPermissionsRemovedForGenesisMembers - we include threadPermissions.EDIT_MESSAGE there, but I'm not sure why. In the most recent revision I didn't add the new permissions there.

Your test plan is super incomplete. Can you please update the test plan to go one-by-one through ever scenario we'd have to contemplate while working on this? Eg. fresh registration, log in of existing user with Farcaster connection, log in of existing user with no Farcaster connection, existing log-in that hasn't logged out since before the new registration flow (never got prompted to connect Farcaster), etc.

One more comment update

Update comment

• don't we need to document it somewhere in some README?

I don't remember all the places where we need to add a new permission. Wondering if you modelled this on another diff that introduced a different permission – if so, would be helpful to share a link to that diff.

In D14444#401325, @varun wrote:

Feels like there are probably ways to dismiss the modal without setting the fid to null here. Ideally we can catch all methods of closing the modal, but if that's really hard, one alternative would just be to set the fid to null as soon as we show the modal.

the onClosed callback gets called by BottomSheet's onChange callback:

const onChange = React.useCallback(
  (index: number) => {
    if (index === -1) {
      onClosed();
    }
  },
  [onClosed],
);

i think this should handle all methods of closing the modal

Does it not occur on every single login because the NUX doesn't display on subsequent logins?

Feels like there are probably ways to dismiss the modal without setting the fid to null here. Ideally we can catch all methods of closing the modal, but if that's really hard, one alternative would just be to set the fid to null as soon as we show the modal.

Emscripten failure is weird, not sure why that's happening, but definitely unrelated to this diff (perhaps the issue is live on master). Created ENG-10380 to investigate

CI - not sure if I need to test something else; not familiar with Android configuration

One thing that might be worth checking is whether the outsideButtons logic is still correct

Preemptively publishing in case CI has issues, as it recently has been timing out

Preemptively publishing in case CI has issues, as it recently has been timing out