Re: HTTP/2.0 issue: magic method registration

On 22/08/2013 05:23, Amos Jeffries wrote:
> When processed by an HTTP/1 server the HTTP/2 magic string 
> intentionally appears to be a valid request-line with method "PRI".
>
> I am thinking this method needs a registration in the registry created 
> by HTTPbis 1.1 documents. That will allow preventing its use in any 
> situations other than HTTP/2.0 setup and help point out to 
> implementers that they may need to handle it differently to other 
> extension methods (silent pass-thru or reject but no fancy "fixing 
> missing headers").
+1.

Received on Thursday, 22 August 2013 09:27:10 UTC