Re: draft-montenegro-httpbis-uri-encoding

On 21/03/2014 7:18 p.m., Mark Nottingham wrote:
> 
> On 21 Mar 2014, at 4:43 pm, Amos Jeffries wrote:
> 
>> What should servers (and intermediaries) receiving non-UTF-8 values in
>> these headers do when they rely on receiving UTF-8 URLs?
> 
> I don't think it makes any difference to them; again, the URL on the wire is still ASCII, and that's still the cache key.

I have this nasty idea about client implementations starting to send
these headers with values indicating things like Unicode or Windows
charsets. That could lead to strange happenings in places like security
ACLs, background auth protocols, log processing software or TTY output
devices if the recipient intermediary was translating from url-encoding
to non-ASCII outputs.

Amos

Received on Friday, 21 March 2014 08:15:37 UTC