Re: [OAUTH-WG] Possible alternative resolution to issue 26

[Julian Reschke <[e-mail]>]

On 2011-09-30 14:03, Buhake Sindi wrote:
> Hi everyone,
>
> As for encoding, my understanding is that the scope parameter were scope
> fields provided by the service provider and that scope should match the
> service provider scope. Fair enough, we could argue that non-UTF-8
> characters can't be sent over HTTP response headers, so a better
> solution, I would suggest is illustrated in section 4.2, Error Handling,
> of RFC 5987.
>
> My solution: add an encoding as a value of the scope and error-desc,
> followed by a colon (:) and then the percentage-encoded message.

Why invent yet another format that's close to RFC 5987 but not the same?

> So, the message, is first encoded to the encoding specified, and then
> percentage-encoded:
>
> Example:
>
> error-descr="UTF-8: <percentage-encoded of an UTF-8 message>"
>
> The client would then:
>
>  1. Percentage decode the string and get the string based on the encoding.
>
> Problem is though, the length of the string value.

How is that a problem?

 > ...

Best regards, Julian