Re: [Mldonkey-users] Re: URL with login & pass

[Lionel Bouton]

Robin Hill wrote the following on 04/30/2004 06:06 PM :

> RFC 2396 only says that _some_

Not quite, this isn't really clear, the user:password syntax is clearly 
scheme specific, but the userinfo is not clearly optional. Remember that 
we are only speaking of a description that must be interpreted on the 
client side and this is not part of a *protocol*. So as soon as the 
protocol underneath can carry user info, the URI can and should define 
an appropriate userinfo syntax (which isn't standardised in the RFC, 
only the character set is defined).

> URI schemes support this syntax, it does
> not specify which schemes do/do not, and the URI format provided is only
> the generic format.  RFC 2616 details HTTP 1.1 and this does not support
> a username/password, nor does HTTP 1.0 (RFC 1945).
>  

After re-reading 2616, I understand why there is confusion on the subject :
3.2 Uniform Resource Identifiers
[...]
As far as HTTP is concerned, Uniform Resource Identifiers are simply 
formatted strings which identify--via name, location, or any other 
characteristic--a resource.
3.2.1 General Syntax
[...]
For definitive information on URL syntax and semantics, see "Uniform 
Resource Identifiers (URI): Generic Syntax and Semantics," RFC 2396
[...]
3.2.2 http URL
[...]
http_URL = "http:" "//" host [ ":" port ] [ abs_path [ "?" query ]]

So the 2616 more or less contradicts itself : http URLs are on one hand 
based on the general syntax described in 2396, but on the other hand 
miss the "userinfo" part ...

In the end, your interpretation seems the right one : there is no RFC 
standard for user:pass in http, but from 2396, there is an incentive for 
the browser market to make one and that is what happened.

Wow, thanks for triggering my refreshing RFC-related memories, things 
are more obsc ^H^H^H^Hclear to me now !

Best regards,

Lionel.