I wrote:
>
> Ian Holsman wrote:
> >
> > On 14/11/2003, at 7:38 AM, Cahya Wirawan wrote:
> >
> > > On Tue, Nov 11, 2003 at 03:38:31PM +0200, Eli Marmor wrote:
> > >> I wrote:
> > >>
> > >>> ...
> > >
> >
> > two points.
> > where are you specifying *which* ip you want to bind to.
> > are you relying on the request coming in on the same ip# your request
> > should be going out on?
>
> After his patch is working, the next step is to define a directive to
> control this behavior.
>
> This directive will define the IP (or hostname), and will have 2
> special keywords, one for the default IP of the machine (i.e. "don't
> bind to anything"), and the second for the IP of the current vhost
> (i.e. what his patch is doing currently).
>
> The only decision that will remain after that, is what to define as the
> default value of this directive; If you ask me, I think that the IP of
> the vhost should be the default. It makes more sense than the default
> IP of the machine.
I take my words back:
If bind() involves a sys-call, then the current behavior ("don't bind
to anything") should remain the default. Because of performance reasons
of course.
--
Eli Marmor
[EMAIL PROTECTED]
CTO, Founder
Netmask (El-Mar) Internet Technologies Ltd.
__________________________________________________________
Tel.: +972-9-766-1020 8 Yad-Harutzim St.
Fax.: +972-9-766-1314 P.O.B. 7004
Mobile: +972-50-23-7338 Kfar-Saba 44641, Israel
