Re: [hybi] Clarify the role of closing handshake

[Takeshi Yoshino <[e-mail]>]

On Mon, Feb 14, 2011 at 12:34, Gabriel Montenegro <
[e-mail]> wrote:

> In A-2 and B, it seems you reversed the Pros and Cons.
>

Oops. Yes, it's reversed...


> I’m not clear on the algorithm for B. Could you formulate in terms of peers
> A and B and clarify which side is doing what?
>

Here's pseudo code for [B].

private closeWriteChannel() {
  socket.send(CLOSE_FRAME);
  writeChannelClosed = true;
}

public initiateClose() {
  closeWriteChannel();
}

public onReceivedCloseFrame() {
  if (!writeChannelClosed) {
    closeWriteChannel();
  }
  socket.close();
}

Here's what happens on each peer.

Normal case:

1. Peer A initiates close by sending a Close frame to peer B
2. Peer B receives a Close frame
3. Peer B sends a Close frame to peer A
4. Peer B closes the underlying transport channel
5. Peer A receives a Close frame
6. Peer A closes the underlying transport channel

In case both side initiate close at approximately same time:

1.
Peer A initiates close by sending a Close frame to peer B
Peer B initiates close by sending a Close frame to peer A
2.
Peer A receives a Close frame
Peer B receives a Close frame
3.
Peer A closes the underlying transport channel
Peer B closes the underlying transport channel


> At any rate, the cons of adding a wait state and additional complexity in
> the protocol is why I prefer the simplicity of A.
>
>
>
> Most of the time, there won’t be a gracious Close, as users will kill the
> browser window, exit out of the tab, etc. So I fear we’re complicating the
> protocol for dubious benefits. We have to deal with sudden termination
> anyway.  At any rate, even if the RST prevents the other peer from receiving
> the WS_Close, the intermediaries will see it.
>

Yes, as far as intermediaries are built to deal with such RST.


>
>
> This is a case of added complexity for very little benefit.
>

I'm not so eager to adopt B. Hmm, ok, how about putting only A-1 as required
spec, and some note for implementor explaining the RST issue and A-2
(Yutaka's idea) as optional solution?

Takeshi