Page MenuHomePhabricator

[native] Don't bother with keyserver auth in BackgroundIdentityLoginHandler
ClosedPublic

Authored by ashoat on Jul 6 2024, 2:39 PM.
Tags
None
Referenced Files
F3396833: D12678.diff
Sun, Dec 1, 2:41 PM
Unknown Object (File)
Tue, Nov 26, 5:01 AM
Unknown Object (File)
Tue, Nov 12, 7:03 AM
Unknown Object (File)
Tue, Nov 12, 6:39 AM
Unknown Object (File)
Tue, Nov 12, 3:05 AM
Unknown Object (File)
Mon, Nov 11, 9:52 PM
Unknown Object (File)
Oct 11 2024, 4:52 AM
Unknown Object (File)
Oct 11 2024, 4:52 AM
Subscribers
None

Details

Summary

We don't need to do it here, since we're already auth'd with the authoritative keyserver.

We just need to perform identity auth.

This mitigates ENG-8725 and ENG-8726, and addresses ENG-8779.

Test Plan

Confirm that I'm able to go from a release build running with usingCSAT=false to one running with usingCSAT=true. Exact instructions here. Make sure that the two "mitigated" issues are no longer present, and that the upgrade goes without issue (except for ENG-8778, which is a separate issue)

Diff Detail

Repository
rCOMM Comm
Lint
Lint Not Applicable
Unit
Tests Not Applicable

Event Timeline

ashoat requested review of this revision.Jul 6 2024, 2:56 PM

we're already auth'd with the authoritative keyserver

Why is it guaranteed?

This revision is now accepted and ready to land.Jul 8 2024, 1:40 AM

we're already auth'd with the authoritative keyserver

Why is it guaranteed?

That's a good question. The dataLoaded condition generally guarantees this, but I'm currently investigating a situation in ENG-8772 where dataLoaded is true despite the user being logged out (currentUserInfo is anonymous).

As part of that investigation, I'll probably conclude that we need to block BackgroundIdentityLoginHandler on both conditions in LogInHandler: loggedIn and navLoggedIn. I've created ENG-8785, and we can track this work there.