Re: #537: Remove segments (consensus call)

Explicit acknowledgements are required for extensions which change semantics, which would include (re-)defining a flag on the DATA frame.  On the other hand, without acknowledgement, an END_SEGMENT frame could be defined with no payload and would be discarded by implementations that don't understand it.

Sent from Windows Mail

From: Yutaka Hirano<mailto:yhirano@google.com>
Sent: ?Tuesday?, ?July? ?1?, ?2014 ?10?:?32? ?PM
To: Mark Nottingham<mailto:mnot@mnot.net>
Cc: HTTP Working Group<mailto:ietf-http-wg@w3.org>, Martin Thomson<mailto:martin.thomson@gmail.com>

It is OK for me if an "explicit acknowledgement" (including proxies) mechanism for extensions is specified in the HTTP/2 spec. Otherwise, not.
IIUC such mechanism is not specified, right? Is there any discussion?

Thanks,



On Wed, Jul 2, 2014 at 2:02 PM, Mark Nottingham <mnot@mnot.net<mailto:mnot@mnot.net>> wrote:
<https://github.com/http2/http2-spec/issues/537>

On 1 Jul 2014, at 4:40 am, Martin Thomson <martin.thomson@gmail.com<mailto:martin.thomson@gmail.com>> wrote:

> On 30 June 2014 11:23, David Krauss <potswa@gmail.com<mailto:potswa@gmail.com>> wrote:
>>> That's an argument for the new application negotiation token.
>>
>> Such a proxy should only be put in a place where no negotiation is necessary. Bailing out at any END_SEGMENT would be acceptable then.
>
> There's an obvious counterargument to that one...
>
> That's fine, but if you want to operate sans-standard, then you can
> add your own END_SEGMENT.
>
> I really don't care either way here.  I'm just enumerating the
> options, and noting that what is currently specified isn't
> particularly well-supported.  Our responsibility is to either more
> clearly define it, or remove it.

Agreed.

It sounds like we're leaning towards removing it - can people live with that?


--
Mark Nottingham   https://www.mnot.net/

Received on Wednesday, 2 July 2014 15:43:48 UTC