Fix incorrect rebootstrap #7238
Conversation
…during actualization
Pull Request Test Coverage Report for Build daniel.tabacaru_696
💛 - Coveralls |
| return SyncClientHookAction::NoAction; | ||
| }; | ||
| auto realm = Realm::get_shared_realm(triggered_config); | ||
| interrupted.get(); |
There was a problem hiding this comment.
Don't we need to wait for the bootstrap to finish here for the test to really validate whether this fix has worked?
There was a problem hiding this comment.
As a matter of fact the test does finish, but it's not immediately obvious why. I think it helps the readability if we fulfil the promise after the message is integrated (I'll add some comments too).
There was a problem hiding this comment.
Should we check here that the subscription set version is zero and the subscription set state is Complete?
There was a problem hiding this comment.
thinking about it more, do we even need the promise in that case. could the test outside the event hook just wait for the subscription set to be complete?
There was a problem hiding this comment.
no. the subscription set is complete as part of the first session, but the second session and its session wrapper used to read the state of the subscription store at the wrong time.
What, How & Why?
Pausing and resuming the sync session at a very specific time may cause the next download message to be mistaken for a bootstrap message and crash the application.
Fixes #6611.
☑️ ToDos
[ ] C-API, if public C++ API changed[ ]bindgen/spec.yml, if public C++ API changed