Mailing List Archive

On May 1, 8:54am, Rob Hartill wrote:
} Subject: multi home
}
} Are there any plans for HTTP/1.1 to send the hostname that
} the client thinks it is talking to, via a HTTP header ?
}
} e.g
}
} Server-name: www.apache.org
}
} This'll obsolete the virtual host stuff we have, and make it
} easy for servers to offer virtual host support without the
} need for multiple IP addresses. We'd only need to add an
} aliasing feature based on hostname and URL, e.g

And this will be a very good thing...it's an ugly hack now.
The correct place to address this is in the protocols. As
I understand it (info from Brian) 1.1 will fit this bill. Now
there are other implimentation issues we will need to address
such as are server aliases host specific? heh...only the begining ;)

}
} HostAlias www.apache.org/ /export/apache/
} HostAlias www.apache.org/foo /wherever/foo
} HostAlias www.hyperreal.com/ /wherever/
}
}
} -=-=
}
} Where can I find more info on HTTP/1.1 ?
}
}
}
} robh
}-- End of excerpt from Rob Hartill

On Mon, 1 May 1995, Rob Hartill wrote:
> I wrote something along these lines last week, but don't remember
> it ever coming back to me. It may have been lost trying
> to find apache.org

Mail should not have been lost....

> Are there any plans for HTTP/1.1 to send the hostname that
> the client thinks it is talking to, via a HTTP header ?

Yes. The mechanism will be something like an

Original-URI:

header which is essentially the complete URL. The reason for this and
not just "Server-name" is that when we have URC's and URN's it is
potentially useful to the server to know what the URL was mapped from.

> This'll obsolete the virtual host stuff we have, and make it
> easy for servers to offer virtual host support without the
> need for multiple IP addresses.

Yes yes yes and oooh boy yes. :)

> Where can I find more info on HTTP/1.1 ?

Hmmm - I thought there was a proposal somewhere. Roy's HTTP spec page
<URL:http://www.ics.uci.edu/pub/ietf/http/> has some pointers to
particular proposals but no codified hard proposal, like HTML 3.0 has.
Though we all know Roy has been just a little busy recently :)

Brian


--=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--
brian@organic.com brian@hyperreal.com http://www.[hyperreal,organic].com/

> Yes. The mechanism will be something like an
>
> Original-URI:

Orig-URI: (the prefix orig- is common in mail)

> header which is essentially the complete URL. The reason for this and
> not just "Server-name" is that when we have URC's and URN's it is
> potentially useful to the server to know what the URL was mapped from.

and because the digest AA will need it anyway. Errr, wait a minute,
that'll break the privacy of the request...bugger. Pardon me while
I go strangle someone....

> Hmmm - I thought there was a proposal somewhere. Roy's HTTP spec page
> <URL:http://www.ics.uci.edu/pub/ietf/http/> has some pointers to
> particular proposals but no codified hard proposal, like HTML 3.0 has.
> Though we all know Roy has been just a little busy recently :)

Yep -- see the http-wg mailing list for snippets; the full proposal
won't be ready until late this month.

......Roy
604 down, 1146 to go