NameVirtualHost

[Brian Behlendorf <br...@hyperreal.org>]

Okay, I intercepted Dean at a bar last night and we talked about this.  While
I still don't like this, I've accepted it with the given explanation: it's
a double-check that your virtual host configurations don't produce any
surprises, due to a slight misconfiguration.  It's basically an admission
that 1) our current configuration syntax isn't explicit enough, and 2) the
whole situation regarding the Host: header makes it difficult to abstract
to the layer we were aiming to abstract to.  Using NameVirtualHost allows
us to support the legacy lame <VirtualHost> definition with the least
amount of work.

We should also aim to have a new syntax in the future which makes things
completely explicit, so the apparent redundancy is not necessary.  Dean's
<IpAddr> proposal from awhile back was a step in the right direction; I've
got another one but I'll hold off as it's definitely a 2.0 thing. 

I would at least like to propose, though, that NameVirtualHost could be
better named.  Perhaps AllowNameVhostOn?  Something to distinguish it from
<VirtualHost>, or I'm sure we'll start seeing people try <NameVirtualHost>
and wonder why it doesn't work :)

	Brian


--=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--
"it's a big world, with lots of records to play."-sig   brian@hyperreal.org