On Fri, Jan 11, 2019 at 04:19:26PM +0200, Henrik K wrote:
>
> On Fri, Jan 11, 2019 at 08:53:43AM -0500, Bill Cole wrote:
> > On 11 Jan 2019, at 3:53, Henrik K wrote:
> >
> > >Thanks, explaining things directly will help.
> > >
> > >I committed a simpler fix:
> > >http://svn.apache.org/viewvc?view=revision&sortby=date&revision=1851021
> >
> > Are you sure that it is doing the same thing?
> >
> > With a *cursory* look I think your fix is harmless, as it catches a style of
> > expressing a failure to find a valid PTR. However, I am somewhat familiar
> > with the inelegant complexity of the thousand-line parse_received_line() and
> > why it has so many baroque special cases, and I'm concerned that you may be
> > doing the exact opposite of Henry's patch. If amavisd-milter can't get a
> > name for its client (always localhost, yes?) that's not the same as any
> > actual MTA not getting a name, is it?
>
> Amavisd-milter gets what MTA provides as milter attributes, it makes no
> resolving itself. It's localhost only if the client connection originates
> from localhost.
And Postfix/Sendmail milter specification provides unknown as [1.2.3.4]
/* Transform unknown hostname from Postfix to Sendmail form. */
.... (ptr = concatenate("[", client_addr, "]", (char *) 0)));
There is absolutely no reason to pass rdns=!1.2.3.4! string to X-Relays-*
rules. It seems to just confuse the existing rules.
