Page MenuHomePhabricator

[identity] add user ID to account ownership message
ClosedPublic

Authored by varun on Oct 12 2023, 10:38 PM.
Tags
None
Referenced Files
Unknown Object (File)
Sat, Nov 23, 10:29 AM
Unknown Object (File)
Fri, Nov 22, 7:33 AM
Unknown Object (File)
Thu, Nov 21, 8:41 PM
Unknown Object (File)
Tue, Nov 12, 3:03 PM
Unknown Object (File)
Mon, Nov 11, 3:05 PM
Unknown Object (File)
Fri, Nov 1, 10:51 AM
Unknown Object (File)
Mon, Oct 28, 11:56 AM
Unknown Object (File)
Sat, Oct 26, 10:35 PM
Subscribers

Details

Summary

add user ID to the account ownership message. we use this user ID instead of minting a new one for existing users if they're already registered on keyserver but haven't yet registered with the identity service.

also added a condition expression to make sure we don't overwrite any user data accidentally.

Depends on D9480

Test Plan

successfully registered a user with my local identity service who was previously registered with my local keyserver.

called the add_user_to_users_table function twice with the same user ID to test the condition expression. second call failed as expected.

Diff Detail

Repository
rCOMM Comm
Lint
No Lint Coverage
Unit
No Test Coverage

Event Timeline

bartek added inline comments.
services/identity/src/database.rs
245–247

👍
By the way, we should look for other places where we are missing such checks.

services/identity/src/reserved_users.rs
24

To match our naming conventions

This revision is now accepted and ready to land.Oct 16 2023, 1:01 AM
services/identity/src/reserved_users.rs
87

nit: this could be const

Will this require some sort of synced rollout between keyserver and identity service?

Will this require some sort of synced rollout between keyserver and identity service?

doesn't need to be perfectly in sync since we have the cron job that runs daily. i'll still let you know before i deploy a new version of the identity service

services/identity/src/database.rs
245–247

the other two places we call put_item are adding a nonce to the nonces table and adding an access token to the access token table. i think it's fine if we don't have condition expressions there. what do you think?

services/identity/src/database.rs
245–247

Yeah that's okay. I mean we should check places where the checks are missing but it'd be recommended to have them based on the functionality.
Adding a nonce doesn't look like something that requires them