Currently the server names are only resolved once, to optimize performance. However, this doesn't support dynamic environments well in which the name should be re-resolved.
I think that I can work around this by wrapping the library calls, and creating a new connection upon error (or maybe creating something which detects name resolution changes), but this doesn't seem ideal.
I considered periodically running SetServers, but I believe the ServerList isn't public in the client.
Currently the server names are only resolved once, to optimize performance. However, this doesn't support dynamic environments well in which the name should be re-resolved.
I think that I can work around this by wrapping the library calls, and creating a new connection upon error (or maybe creating something which detects name resolution changes), but this doesn't seem ideal.
I considered periodically running SetServers, but I believe the ServerList isn't public in the client.