ws: Concurrent Error Handling Model

[niaow]

So I have realized that some of my concurrent error handling model was flawed. My main goal is to find the root cause error and return it everywhere relevant. However, there are many edge cases which can complicate this, like both ends attempting to close the connection at once - where it is entirely possible that a read error will cause the write end to act as if it completed a successful closure. I may have to rethink some of this.