Re: [Exim] Address rewriting vs. CNAME records?

Página Principal
Apagar esta mensagem
Responder a esta mensagem
Autor: Exim Users Mailing List
Data:  
Para: Matthew Byng-Maddick
CC: Exim Users Mailing List
Assunto: Re: [Exim] Address rewriting vs. CNAME records?
[ On Friday, February 7, 2003 at 20:39:06 (+0000), Matthew Byng-Maddick wrote: ]
> Subject: Re: [Exim] Address rewriting vs. CNAME records?
>
> On Fri, Feb 07, 2003 at 03:04:39PM -0500, Greg A. Woods wrote:
> >
> > I'm sorry you feel that way -- obviously you've never read the code, at
> > least not with deep enough comprehension to understand what it does.
>
> Perhaps you're talking rubbish?


I highly doubt it. I've not only done work directly on the related code
in TCP Wrappers, I have, as I've said, also written several chunks of
code that do the same thing for several different kinds of service
daemons, and have done so in a few different programming languages too.

Furthermore I've been intently studying the correct use of the DNS, and
the common mistakes people make in attempting to use the DNS, for almost
a decade now. I know in great detail what causes problems and what does
not, and I know when people think there's no problem but are only
fooling themselves and anyone else foolish enough to believe them.

I'm not talking about just RFCs and such -- I'm talking about deployed
code and I'm talking about real-world interoperability problems.

Would you rather risk some unknown percentage of problems with your
e-mail, or would you rather avoid all of that risk by properly deploying
your DNS in the most complete and consistent manner possible? It's your
risk, and your decision -- just don't even try to point the finger
elsewhere when you've caused your own problems.


> The box in question (colon.colondot.net) does not have PTRs for every
> single A record that points to it. Far from it.
>
> Now, your point?


Where did I say anything about my mailer requiring PTRs? I said that it
would reject clients with mis-configured or incomplete PTRs, not that it
required a PTR of every client.

Indeed I have now at least twice said quite clearly that _NOT_ having
PTRs is _BETTER_ than having broken ones. Have you not been paying
close enough attention? All you've done so far is proven that point.

(Isn't the English language a wonderful tool? :-)

--
                                Greg A. Woods


+1 416 218-0098;            <g.a.woods@???>;           <woods@???>
Planix, Inc. <woods@???>; VE3TCP; Secrets of the Weird <woods@???>