Re: Questions in draft-09

Hi,


On Wed, Jan 8, 2014 at 1:15 AM, Shigeki Ohtsu <ohtsu@iij.ad.jp> wrote:

> Hi, I have a few questions on draft-09.
>
> They are trivial and not serious but need to be clarified
> in the draft for implementation.
>
> 1. Synchronization of HTTP2-Settings
>  HTTP Upgrade is still in the spec.
>  Are the values of HTTP2-Settings is acked and synchronized at
>  the time of 101 response?  There is no need for endpoints to
>  send and wait for a SETTINGS-ACK of HTTP2-Settings Header.
>
>
My understanding is HTTP2-Settings are implicitly ACKed.
The whole point of HTTP2-Settings is inform settings to the remote endpoint
without
delay.
I think it is not so hard for the implementation to apply the settings at
the initialization.


> 2. How to refuse a large SETTINGS_HEADER_TABLE_SIZE
>  When an endpoint received a large SETTINGS_HEADER_TABLE_SIZE
>   and do not want it, what to do it?
>
>  a) The endpoint does not send SEETINGS-ACK, but it may cause
>     a SETTINGS_TIMEOUT error.
>  b) The endpoint sends SETTINGS-ACK but does not allocate a large
>     header table and use only LiteralHeaderWithoutIndexing in
>     encoding instead.
>  c) others?
>
>
We discussed this topic some time ago:

http://lists.w3.org/Archives/Public/ietf-http-wg/2013OctDec/0317.html



> 3. Header Ordering(8.1.3.3) and Cookie Compressing(8.1.3.4)
>  An encoded cookie header can contain 0x0 in its value though it is
>  not comma-separated lists and its ordering is not significant.
>  In this case, is the cookie value splited and joined with '; '
>  into a single cookie header?
>
>
My understanding is that the spec only requires concatenation of cookie
values with "; "
if they are sent separately in independent header field.
If they are ever sent in concatenated with 0x0 delimiter in one header
field, then
rule 8.1.3.3<http://tools.ietf.org/html/draft-ietf-httpbis-http2-09#section-8.1.3.3>
is
applied and we are not going to concatenate them.

Best regards,

Tatsuhiro Tsujikawa


> Regards,
>
>