Mark Boyns (email@example.com)
16 Apr 1998 10:24:15 -0700
On 16 Apr 1998 firstname.lastname@example.org writes:
> Mark Boyns wrote:
> > Andrew, does htdig support virtual hosts via the HTTP Host header?
> > I'm having a strange problem with one of James Pei's servers and I
> > suspect he must have mucked something up....
> > If you're interested, trying searching for "housing". It returns
> > http://career.sdsu.edu/hrlo/ which should really be
> > http://starter.sdsu.edu/hrlo or maybe even
> > http://www.sa.sdsu.edu/hrlo/. Note that career and www.sa cnames of
> > starter.
> > Argh, I might have to call him :)
> Better you than me, buddy! :-)
You're telling me...
> I think the problem might be the conflict between the need to map aliases to a
> single host and the ability to support virtual hosts.
> If the allow_virtual_hosts attribute is set to true, htdig doesn't perform the
> hostname canonicalization. When I started working on ht://Dig that presented
> a big problem with machines that had aliases but not virtual hosts. (I don't
> know if you remember, but most of the life sciences web pages were indexed 2
> or three times because of that...)
> So the problem is the mixing of the two modes.
> htdig sends out the "Host:" header in all cases, but only when
> "allow_virtual_hosts" is set to true will it actually attempt to send the
> correct value for the "Host:" header.
> If "allow_virtual_hosts" is false, then hostnames are compared by IP address
> and the "Host:" header will only ever get the first hostname used for any
> particular IP address.
Ok, I'll check it out. From the htdig output I don't see any
reference to starter or www.sa so I'm thinking that
allow_virtual_hosts might fix it. (I'm using 3.0.8b1)
> I don't know what version of ht://Dig you're running. I added the
> "allow_virtual_hosts" attribute in 3.0.8b1, I believe (Goddamn disk crash
> whiped out my CVS tree!!!)
You don't have 80GB completely mirrored like rohan? :)
This archive was generated by hypermail 2.0b3 on Sat Jan 02 1999 - 16:26:01 PST