Re: UTF-8 in URIs

On 2014-01-17 10:02, Poul-Henning Kamp wrote:
> In message <aba0c024677d0d0bf4dc944ef64da6b7.squirrel@arekh.dyndns.org>, "Nicol
> as Mailhot" writes:
>
>> I'd rather have clear interop rules and simple to understand breakage at
>> http2 adoption time rather than diffuse heisenbugs and sites that work by
>> sheer chance as long as you chain components with the same unwritten
>> assumptions. See the python 2 unicode debacle and how much fuzzy encoding
>> rules cost them.
>
> Seconded.
>
> This is one clear way HTTP/2 can improve over HTTP/1.1:  Eliminate
> unnecessary and pointless fuzziness.

There is no fuzziness here.

If you really want to improve over HTTP/1.1, there are more interesting 
issues, such as character encoding in header field values.

Best regards, Julian

Received on Friday, 17 January 2014 09:10:55 UTC