# Faster joins testing
Using #megolm:matrix.org as a big room test ground.
The resync process is stalled for 120s by hacking a `sleep` in the code.
## Encrypted room
### Element Android
Seems to behave mostly perfectly:
- Receiving is fine
- Sending is properly waiting for the room to un-partial-state (just wait in `/members` I guess, does it need a retry logic ? TBC)
- Then OTK requests are sent (not parrallelized so still quite slow for big rooms)
- Re-joining a room (with full leaving of the server, and a new fast join) seems to keep devices (or members ?) list around, and sending works directly. Is is a trouble ? how does it know that the list is up to date ?
### Hydrogen
- Receiving is blocked sometimes, seems stuck on a backfill request
- Sending is properly waiting for the room to un-partial-state (just wait in `/members` I guess, does it need a retry logic ? TBC)
- Then OTK requests are sent (not parrallelized so still quite slow for big rooms)
### Element Web
Seems to behave mostly perfectly:
- Receiving is fine
- Sending is properly waiting for the room to un-partial-state (just wait in `/members` I guess, does it need a retry logic ? TBC)
- Then OTK requests are sent (not parrallelized so still quite slow for big rooms)