Details
- Create a session between devices A and B
- On B comment out sending confirmation.
- Send some messages from A to B
- On B simulate Olm error.
- Test that all messages are sent again and decrypted successfully.
Diff Detail
- Repository
- rCOMM Comm
- Branch
- land
- Lint
No Lint Coverage - Unit
No Test Coverage
This logic is a bit confusing.
I guess the reason we decided to store a separate deviceToken per keyserver is because we want to know if we've shared our deviceToken to that keyserver yet, not because we expect to have a different deviceToken per keyserver.
It would probably make more sense if we have a deviceTokenHasBeenUploaded: boolean in KeyserverInfo, and a deviceToken field at the top level of Redux. But we'll likely end up removing deviceTokens from keyservers soon anyways...