Re: [hybi] Straw poll: Do you miss interjectable WebSocket level control frame? (was: Re: Discontinuation of mux ...)

[Yutaka Hirano <[e-mail]>]

>
> * Close: Can be used to perform a "clean" close in the middle of sending a
> message. However, I cannot think of any applications that would consider a
> situation in which a in-transit message was aborted to close the connection
> to be a "clean close" scenario.

An endpoint should be able to cancel the ongoing transmission, i.e. The
close call should abort transferring very_very_long_message in the
following code.


ws = new WebSocket(...);
ws.onopen = function() {
  ws.send(very_very_long_message);
  ws.close();
};

It is possible to _Close the WebSocket connection_ without the closing
handshake in such cases, but since the WebSocket API doesn't expose the
transmission status, doing so makes the closing handshake completely
unreliable.



On Wed, Feb 19, 2014 at 4:09 PM, Adam Rice <[e-mail]> wrote:

> With the current set of control frames, I would not miss the ability to
> interject them in the middle of a data message.
>
> Breaking them down one by one:
>
> * Pong: useless in the middle of a data message (the peer already knows
> that it is receiving data, and useless for RTT measurement because it has
> an unknown amount of OS queueing and buffer bloat in front of it).
> * Ping: almost always useless in the middle of a data message. It could be
> useful if, for example, you had a middle box which closed "idle"
> connections, and considered any connection "idle" if there was no TCP/IP
> payload traffic originating from the client, and the server was in the
> middle of sending a huge message. However, such a middle box would also
> break HTTP downloads so seems unlikely to exist in practice. Even if it
> existed, how would the server detect it to know that it needed to send a
> Ping?
> * Close: Can be used to perform a "clean" close in the middle of sending a
> message. However, I cannot think of any applications that would consider a
> situation in which a in-transit message was aborted to close the connection
> to be a "clean close" scenario.
>
>
> On 19 February 2014 13:24, Takeshi Yoshino <[e-mail]> wrote:
>
>> I'm inclined to say "Yes" for this question so said yes for #2 of
>> Yutaka's original question [1], but I could also cope with if we decide to
>> drop this.
>>
>> [1] http://www.ietf.org/mail-archive/web/hybi/current/msg10403.html
>>
>> _______________________________________________
>> hybi mailing list
>> [e-mail]
>> https://www.ietf.org/mailman/listinfo/hybi
>>
>>
>