Can't think of a scenario where we don't want to close one at the same
time as the other, but let's be conservative and keep it a very low
number for now.
* socks5 inbound: stop reading when closing connection
* Wait for lane at select top-level
* Allow closing connection while waiting for lanes to clear
* Some tidy in inbound.rs
* Put chained future back inline in the select
* Remove commented out line
* Disable the read data branch on is_finished
* Rename to nym-task
* Update imports to use nym_task
* Update to nym_task in explicit module prefix
* Touchups and rustfmt
* rustfmt
* crate metadata
* fix nym-connect-android
* cleaned up MixProcessingError
* Added Error impl to (hopefully) all error enums in the codebase
* Replaced all occurences of error("{0}") with error(transparent)
* Changelog entry
* Implementation of 'Debug' on 'RealMessage'
* expect with failed channel name instead of throwing empty panics
* Introduced Debug trait constraint in ProxyRunner
* Derive Debug for socks5_requests::Message
* Return and handle ClientRequest::LaneQueueLenghts
* Pass lane queue lengths to inbound future
* Remove unused self reference
* Request lane queue lengths periodically for all open connections
* Add timeouts
* Rename to ConnectionCommandSender and Receiver
* Rename to client_connection_tx/rx
* Fix wasm build
* Replace bool with enum
* WIP: QA network details
* Initial implementation to multiplex socks5-client sends
* Introduce TransmissionLane enum
* WIP
* WIP: client requests connection id
* WIP
* mulitplex somewhat done
* Remove closed lanes
* WIP: connection handling over ws
* Remove unused published active connections shared data
* Start on status timer
* Max number of connections, and prune
* Some tidy
* Remove commented out code and tweak log
* Tidy
* Tweak log output
* Rename to TransmissionBuffer
* Use number of msg sent instead of time to rank age of lanes
* Create client-connections crate
* Remove waker call that probably are not needed
* Extract out some types from real traffic stream module
* Revert to develop qa.env
* Tweak comments, tidy for getting ready to merge
* Update changelog
* wasm client compile fixes
* rustfmt
* Basic storage stub
* New models for new node status api
* Route handling
* Mounting new routes
* Missing selective commit
* Moved network monitor related files to separate package
* Starting to see some sqlx action
* Schema updates
* Log statement upon finished migration
* Removed old diesel related imports
* Converted mixnode cache initialisation into a fairing
* Moved cache related functionalities to separate package
Also defined staging there
* Created run method for validator cache + removed unwrap
* Removed old node-status-api types and left bunch of todo placeholders in their place
* Fixed managing validatorcache
* Status reports are starting to get constructed
* Submitting some dummy results to the database
* Removing duplicate code for generating reports
* Removed statuses older than 48h
* Initial attempt at trying to obtain reports for all active nodes
* Removed duplicates from the full report
* Grabbing uptime history
* Updating historical uptimes of active nodes
* Updated sqlx-data.json
* Removed all placeholder foomp owner values
* Changed Layer serde behaviour for easier usage
* Extended validator api config
* Initial (seems working !) integration with network monitor
* Added database path configuration to config
* Using ValidatorCache in NetworkMonitor
* Flag indicating whether validator cache has been initialised
* Introduced a locla-only route for reward script to perform daily chores
* Flag to save config to a file
* Moved spawning of receiving future to run method rather than new
* Removed arguments that dont make sense to be configured via CLI
* Removed dead code from config file
* More dead code removal
* Added validator API to CI
* Corrected manifest-path arguments
* Constructing network monitor by passing config
* Combined validator API CI with the main CI file
* Using query_as for NodeStatus
* Checking if historical uptimes were already calculated on particular day
* Making id field NOT NULL
* More query_as! action
* Updated sqlx-data.json
* Removed unused chrono feature
* Renamed the migration file
* Changed default validator endpoint to point to local validator
* Removing unnecessary clone
* More appropriate naming
* Removed dead code
* Lock file updates
* Updated network monitor address in contract code
* Don't stage node status api if network monitor is disabled
* cargo fmt
* Updated all license notices to SPDX
* Updated all non-breaking dependencies
* Updated common/crypto dependencies
* Updated all tokio [and associated] dependencies to most recent version
* Bumped version of rand_distr
* Fixed api changes in tests
* Made clippy happier about the acronym
* Fixed the type while trying to make clippy even happier
* nightly cargo fmt
* Increased default connection buffer size to help with packet bursts
* Informing remote about failure to establish connection
* Properly closing outbound proxy on failure to read from socket
* Split proxy runner into separate inbound and outbound files
* Flag indicating whether socks client actually started proxying
* Clearing stale proxies if other side is closed
* Decreased log level on not a warning
* Closing proxy if no messages were received in a while
* Dont print harmless false-positive errors
* Formatting + future notes
* Fixed packet delay in higher volume traffic
* Don't attempt writing empty messages to connection
Jędrzej Stuczyński