Milter-greylist

On Thu, Aug 15, 2013 at 04:14:49PM +0200, manu@... wrote:
> Jim Klimov <jimklimov@...> wrote:
> 
> > # This is in wrong order by a couple of lines - you should first
> > set the priv->priv_hostname value with strncpy(), then fix it :)
> 
> Right.
> 
> > # Also, this is geared towards exactly resolving a bracketed name.
> > My code could also do resolution in the (theoretical) case that a
> > numeric IP address was passed without brackets. Then again, this
> > routine is called "unbracketing" for a reason ;)
> 
> I prefer to address real problems rather than potential ones.
>  
> > # In the routine itself you have an ifdef... I think it warrants
> > an #else to report that unbracketing was needed and requested, but
> > the program was configured and built without proper DNS support.
> 
> Updated version: http://ftp.espci.fr/shadow/manu/unbracket2.patch

Just want to note, that the bracketed version sendmail provides may look like

[IPv6:2a01:1b0:7999:446:0:2:8ed:2c78]

This is the form sendmail falls back to a hostname if the ip addr does not resolve.
In a MG log it looks like 

Aug 16 06:11:23 neelix milter-greylist: r7G4BN1J030057: skipping greylist because this is the default action, (from=<XXXXXX@...>, rcpt=<xxxxxxx@...>, addr=[IPv6:2a01:1b0:7999:446:0:2:8ed:2c78][2a01:1b0:7999:446:0:2:8ed:2c78]) ACL 614

as opposed to 

Aug 16 08:06:07 neelix milter-greylist: r7G667ba013248: skipping greylist because this is the default action, (from==netbsd.org@...>, rcpt=<xxx@...>, addr=mail.NetBSD.org[2001:4f8:3:7::25]) ACL 614


It would be interessting, what Oracle CommunicationServer defines ...

Maybe the function resolve_bracketed() should handle this case too.


Johann