Known connectivity issues
This is more a note than a real issue, but used to group what's known, in progress and missing
Connecting status
Maybe still there. But for now, the scenarios I see are not related (last week was due to TURN changes). So for now, wontfix til no new precise scenario. Too much time lost for non reproducible scenarios
UPnP
UPnP candidate selection
Generally not chosen. The current question is to determine why it's not the case Observation: Lot of timeout when adding candidates
Improve the current code
Add libpcp?
Unanswered for now
IPv6
TURN not added in IPv6
This patch solves current state for daemon + pjsip: https://review.jami.net/c/ring-daemon/+/15606
However still a TODO: update our TURN configuration to allow both IPv6 + IPv4 selection as specified in the issue
IPv6 only to IPv4
All is there: ring-client-ios#46
iOS
IPv4 addresses are not always resolved
Solved by https://review.jami.net/c/ring-daemon/+/15655 but still in discussion with pjsip
Android
Wake up after a night
Some devices are not correctly awoken. First lead was a lock solved by https://github.com/savoirfairelinux/opendht/commit/23a75a3f504938771649a5791451207b366fa351 However still some tests to know if there is more than one issue, cause the OS can be pretty aggressive with Doze Mode and constructors (cf https://dontkillmyapp.com/).
Ice possible evolutions
- Implement latest RFCs
- Get candidates as soon as available to avoid any delay in TURN resolution for example
- Follow a connectivity change to update sockets when going from wifi to 3g for example
- #288
Misc
- Merge current stack into pjsip master to avoid to manage patches
- Investigate: #274 (closed) but it's been a while I don't see this, so for now, I think I will close