Re: Getting to Consensus on 1xx Status Codes (#535)

On 19 July 2014 03:19, Roy T. Fielding <fielding@gbiv.com> wrote:

> All of these codes have been implemented in practice.  I have used
> them, interoperably, in both non-browser user agents and origin servers.
> I would have used them in javascript as well if they were passed through
> to the script consistently.
>
> AFAICT, this working group is short on experience and too focused on
> browser usage of HTTP.  Get over it.
>

Well there is experience of 1xx codes usage in the working group.  As I
have pointed out, Jetty has a good implementation of 100 support, 101 and
also 102.    We see real usage of all the codes (ok not so much for 102,
but we'll send it if you want it).

We wish to continue supporting 100 semantic at least and our preferred way
to achieve that is by following rfc723x semantics as defined - but could
live with a new mechanism if it was well designed and motivated (but have
not seen such a proposal yet).      We also wish to support 101 and am very
keen to try an upgrade to waka (but can't find the detailed specification).

The option b) proposed is to define a document to deprecate all 1xx.   If
such a document could be produced and agreed to, then I would not weep on
the passing of 1xx - so long as it was widely supported and done for the
entire web.   However I am dubious that this group - or at least this
subset of the WG - has the authority or backing to propose such a wide
spread deprecation (ie deprecating the preferred access mode of little
sites like Amazon web services).

I admit to befuddlement at how we are tying ourselves in knots to support
streaming huge headers yet are prepared to discard a much more widely used
semantic.

I really don't understand the resistance to Julian's 1xx proposal is.  As
Roy says,  h1 implementations today have to handle 1xx at least to the
ignore it or fake it level.  Changing the transport for this semantic does
not add complexity, but making the semantic conditional on transport does.

cheers





-- 
Greg Wilkins <gregw@intalio.com>
http://eclipse.org/jetty HTTP, SPDY, Websocket server and client that scales
http://www.webtide.com  advice and support for jetty and cometd.