|
Neohapsis is currently accepting applications for employment. For more information, please visit our website www.neohapsis.com or email hr@neohapsis.com |
Re: header_if_reject
From: Wietse Venema (wietse
porcupine.org)
Date: Mon Feb 07 2005 - 10:00:10 CST
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Piotr KUCHARSKI:
> On Mon, Feb 07, 2005 at 09:25:01AM -0500, Wietse Venema wrote:
> > In order to be technically correct:
> > - Non-ASCII and control characters need to be sanitized.
>
> printable(), done.
>
> > - Long header lines need to be folded (as in: "broken up into
> > multiple prepend records").
>
> Hm, I don't think it is needed, important information is within
> first 80 chars. I'll simply cut error_text at var_line_limit-20.
> Unless there already exist some function to do such folding.
var_line_limit specifies INPUT line length. It is 2 kbytes,
twice the length of what SMTP requires.
It IS NOT the length of a header line.
> > In order to be architecturally correct:
> > - The label/reject/warn action needs to be separated from the match
> > operation (DNSBL, unknown helo name, etc.). so that we can avoid
> > foo_if_reject madness.
>
> I don't quite understand, could you explain?
I don't want to pollute Postfix with turds: xx_if_reject, yy_if_reject,
zz_if_reject, and so on.
Wietse
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]