Re: [apps-discuss] Fwd: I-D Action: draft-yevstifeyev-ftp-uri-scheme-04.txt

[Mykyta Yevstifeyev <[e-mail]>]

15.07.2011 2:44, Frank Ellermann wrote:
> On 14 July 2011 05:29, Mykyta Yevstifeyev wrote:
>
>> "@" is in<gen-delims>, so we can safely assume that it won't
>> apper anywhere else except the authority non-percent-encoded.
> Quoting RFC 3986 section 3.3, page 23:
>
>   segment-nz-nc = 1*( unreserved / pct-encoded / sub-delims / "@" )
>                     ; non-zero-length segment without any colon ":"
>
>   pchar         = unreserved / pct-encoded / sub-delims / ":" / "@"
>
>
> That allows an unencoded '@' in the<path>, not the '@' you are
> interested in while trying to skip any<userinfo>@.
>
>   query         = *( pchar / "/" / "?" )
>
> Ditto, unencoded '@' in the<query>  (section 3.4, page 24).
>
>    fragment     = *( pchar / "/" / "?" )
>
> Ditto, unencoded '@' in the<fragment>  (section 3.5, page 24).
I see what you mean.  I'll correct the expression as you proposed.  
However, I actually don't understand why "@" is either in <gen-delims>, 
which is in <reserved> and doesn't allow it to be present in queries and 
fragments, and in <pchar>, which is OK with this?
> [ . . . ]
> See above wrt<path>,<query>, and<fragment>.  I recall that
> it took me more than a year to grok the fine print in STD 66.
Mykyta Yevstifeyev
>
> -Frank
>