Posted by Oli Filth on 11/18/05 00:22

M. Trausch said the following on 17/11/2005 19:17:
> Oli Filth wrote:
>
>>>That's because the browser treats %23 as #... e.g., you cannot use that
>>>character, as it's local to the browser as a named-anchor.
>>
>>What? No it doesn't.
>>
>
>
> According to RFC 1738, the characters that should be escaped in a URL
> are {\$&+,:/=?@}. It defines the following characters to be unsafe, and
> hence undefined as to how the application can/will handle it and remain
> compliant with the RFC: { "`'<>#%\{\}|\\\^~[]}
>
> So... you're partially right. It doesn't *have* to. The point is that
> the intended use of the hash mark is as a named-anchor reference. It
> doesn't have to be passed. It depends on the application's
> implementation of RFC 1738.
>

My point was that the browse *never* treats %23 in a URL as #, it treats
it as %23.


--
Oli

[Back to original message]