The client opts into turbotunnel mode by sending a magic token at the
beginning of each WebSocket connection (before sending even the
ClientID). The token is just a random byte string I generated. The
server peeks at the token and, if it matches, uses turbotunnel mode.
Otherwise, it unreads the token and continues in the old
one-session-per-WebSocket mode.
Unless something externally called Write after Close, the
writeLoop(ws, pr2) goroutine would run forever, because nothing would
ever close pw2/pr2.
https://bugs.torproject.org/33367#comment:4
We had already implemented Read, Write, and Close. Pass RemoteAddr,
LocalAddr, SetReadDeadline, and SetWriteDeadline through to the
underlying *websocket.Conn. Implement SetDeadline by calling both
SetReadDeadline and SetWriteDeadline.
https://bugs.torproject.org/33144
Rename websocketconn.WebSocketConn to websocketconn.Conn, and
websocketconn.NewWebSocketConn to websocketconn.New
Following the guidelines at
https://blog.golang.org/package-names#TOC_3%2e
Proxies now include information about what type they are when they poll
for client offers. The broker saves this information along with
snowflake ids and outputs it on the /debug page.
Switch to containing all communication between the proxy and the broker
in the HTTP response body. This will make things easier if we ever use
something other than HTTP communicate between different actors in the
snowflake system.
Other changes to the protocol are as follows:
- requests are accompanied by a version number so the broker can be
backwards compatable if desired in the future
- all responses are 200 OK unless the request was badly formatted
The safelog Write function can be called from multiple go routines, and
it was not thread safe. These locks in particular allow us to pass the
logscrubber's output io.Writer to other libraries, such as pion.
Our log scrubber was a bit over-zealous and was scrubbing timestamps as
well. While we're still over-scrubbing, we're now more precise and
ensure that compressed IPv6 addresses have "::"
