Behavior on unknown chainslug

[vaibhavchellani]

We cannot assume a socket is aware of all other socket's so it makes sense to not maintain any kind of data store that maintains all other socket's as it would make it really hard to be permissionless, it should be possible that anyone can deploy a socket on a new chain, wire to it and get going

However, what is the desired behaviour if the outbound is called with a toChainSlug which hasn't been configured? I am not sure if we fail silently or revert rn

[ameeshaagrawal]

Currently socket stores the plug config for each toChainSlug which is used at outbound to get the capacitor address where the messages are queued in the packets which are transmitted to toChainSlug.

We have 2 cases here:

1. If there is no switchboard for toChainSlug, plugs cannot register for that slug and hence call will revert.
2. If someone registers the switchboard for this unsupported toChainSlug, plugs will be able to register and hence outbound call will go through. But the transmission will not happen.

[arthcp]

I think plugs cannot connect in case there is no capacitor. So unless someone connects a switchboard to some remote chain, plugs cannot send messages.

[vaibhavchellani]

ok makes sense and its basically upto the transmtiter managers to decide if they serve the toChainSlug or not, people can keep sending messages all day long

[ameeshaagrawal]

Only if transmitter has the role. Transmitters are chain specific, so they can only seal/propose if we grant roles for this new toChainSlug.

[vaibhavchellani]

makes sense, I can totally imagine a transmitManager has an open function that allows anyone to seal for unregistered destChainSlug