Re: [hybi] how do we move forward on agreeing on framing?

[Ian Fette (イアンフェッティ) <[e-mail]>]

On Thu, Aug 19, 2010 at 12:43 AM, Dave Cridland <[e-mail]> wrote:

> On Thu Aug 19 06:28:28 2010, John Tamplin wrote:
>
>> Is this something we can all get behind, even if it doesn't do everything
>> we
>> want, in the interest of getting something useful done soon?  If not, what
>> specific changes would be necessary for you to agree?
>>
>
> I'd feel more comfortable with knowing that the extension framing (not
> actual support) was done. Otherwise, any extension mechanism is going to
> smell like a real hack, even if it's possible. I also worry about actors
> seeing extensions and choking on them, whether that's intermediaries or
> whatever. Finally, it's not clear to me if extension data needs to be
> per-message or per-frame - I can't think of a per-frame use-case, but still,
> this effects possible extension mechanisms heavily.
>
>
There are different definitions of "need", but I think the people interested
in multiplexing have made a good case that that's extension data that should
be present per-frame assuming that a message gets fragmented. (The "control
frame" argument that switches the state of the stream did not get people
excited.)

I also would be a bit happier if we had a better sense for how the
extensibility worked. For instance, let's say we have an intermediary that
is unaware of extensions, and so it re-fragments a frame that contained e.g.
a stream ID, and now all of a sudden we have one frame that contains an
stream ID and another frame that doesn't. Oops. The receiver is probably
going to choke. If you have to deal with this, then you have to keep state
around and now have all the problems that were brought up with trying to do
multiplexing via a control frame.

That said, I don't think anyone is going to rush to produce intermediaries
for WS, so maybe we do have some time (6m timeframe) to figure some of that
out and circle back.

-Ian


> It could be my fears are misplaced, of course, and I'm willing to be
> persuaded, but this is my current stance.
>
> Aside from this one issue, this design - or the previous one, in fact - are
> fine with me. I marginally prefer this one, modulo the absence of an
> extension data block.
>
> It'd be useful at this point, though, to produce a new draft containing
> this design, which seems to have broad support if not yet actual consensus -
> ie, this should become the group's current working design.
>
> Dave.
> --
> Dave Cridland - mailto:[e-mail] - xmpp:[e-mail]<[e-mail]>
>  - acap://acap.dave.cridland.net/byowner/user/dwd/bookmarks/
>  - http://dave.cridland.net/
> Infotrope Polymer - ACAP, IMAP, ESMTP, and Lemonade
> _______________________________________________
> hybi mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/hybi
>