httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Roy T. Fielding" <field...@kiwi.ICS.UCI.EDU>
Subject Re: [STATUS] Sun Apr 13 19:08:09 PDT 1997
Date Tue, 15 Apr 1997 03:01:52 GMT
>On Mon, 14 Apr 1997, Rodent of Unusual Size wrote:
>>      o Doesn't issue "Vary: Host" in responses because it would appear
>>        in *all* responses and would therefore chew up bandwidth
>>        unnecessarily.  [I'm still not entirely comfortable with this;
>>        just because Apache's (current) implementation means we'd send
>>        the header every time doesn't == it's okey to violate the RFC in
>>        the interests of performance.  Just IMHO.]
>It doesn't issue "Vary: Host" because any compliant proxy would have to
>use the host header anyhow to direct the hit to the correct origin server,
>and so all requests implicitly vary on Host.  I think Roy considers
>this an oversight in RFC2068, but don't let me put words in his mouth...
>Hey Roy are you keeping a list of deficiencies of RFC2068?  I know you've
>cursed a few times when we've tried to implement things here :)

I decided that we are already compliant with HTTP/1.1, since we are
doing resource selection using Host and not negotiation.  However, I will
need to suggest a fix to section 14.43 of RFC 2068, since it incorrectly
refers to the Request-URI instead of the "requested resource."

I haven't been keeping an apache-specific list, since all of our issues
are going into the general HTTP-WG issues list (kept by Jim Gettys).


View raw message