When trying to connect a node (client or vault) from one of our Linux test machines to an existing running network, and an invalid peer endpoint is passed in, it tries to bootstrap off itself because its IP is in the list of hard-coded contacts. This will make the join succeed because the node itself started listening on port 5483 before attempting to join. Subsequent RPCs will of course fail.
Expected behaviour: Fail right away when trying to connect to a non-existing network, no matter if hard-coded contact or not. Perhaps port 5483 should only ever be opened once we've got at least one connection to another node.
Migrated from https://github.com/maidsafe/MaidSafe/issues/87
When trying to connect a node (client or vault) from one of our Linux test machines to an existing running network, and an invalid peer endpoint is passed in, it tries to bootstrap off itself because its IP is in the list of hard-coded contacts. This will make the join succeed because the node itself started listening on port 5483 before attempting to join. Subsequent RPCs will of course fail.
Expected behaviour: Fail right away when trying to connect to a non-existing network, no matter if hard-coded contact or not. Perhaps port 5483 should only ever be opened once we've got at least one connection to another node.