netmod-tcp: fixing: issue where dropped connections wouldn't resume
The solution to this problem has been to send explicit hello-responses from the receiving stream end. When sending a HELLO a task tries to read from the sending stream to wait for an ACK. If this response times out (TODO: make these values configurable), it will officially drop the sending connection, meaning that the next send re-initialises the stream. With some initial testing this has shown _most_ connection drops to recover gracefully, but not all. Further testing is required.
Showing with 111 additions and 36 deletions