* Added timeout values to configs
* Fixed possible crash when using delay larger than 2 years
* Connection timeout with hardcoded value
* Using provided timeout value
* tcp client requiring timeout value at construction
* Client using updated tcp client
* Mixnode using updated tcp client
* Healthchecker having separate timeout + new client config field
* Updated validator with connection timeout field
* Added connection_timeout to validator config template
* Preventing multiplication overflow for reconnection backoff
* Check for addition overflow + helpful-ish error message
* Not resetting delay twice
* Lowered healtcheck reconnection backoff to something way more reasonable but still impossible to reach
* Moved 'addressing' to nymsphinx + slightly better API + tests
* Replaced all usages of above in the project
* Removed 'split_and_encapsulate_message' from chunking API to get rid of circular dependency + to prefer usage of TopologyAccessor
* changes in Cargo.lock
* formatting
* Filtering out early eof errors
* Wrapping ConnectionManager in Abortable
* Decreased pathchecker logging level
* Client aborting all connection futures on drop
* Moved AbortHandle to connection_managers HashMap to couple them closer together
* Integrated new tcp client into healthchecker
* Printing detailed healtcheck score if in debug
* Decreased logging level for when establishing new connection
* Slightly better formatting for printing healthcheck scores
Basic support for message chunking and reconstruction. It's now possible to take arbitrarily-sized input and send it through the mixnet. The sending Nym client will automatically chunk the input into Sphinx packets, and the receiving Nym client will then reconstruct the message at the other end.
Dave Hrycyszyn