Re: Milter header order

2018-11-12 Thread Henrik K
On Mon, Nov 12, 2018 at 11:03:09AM -0500, Viktor Dukhovni wrote: > > With that, the position of the header should be less critical. Thanks for the info, I'll tune forthcoming AuthRes parser for SpamAssassin with this in mind. Also found out that amavisd-milter naively inserts it's synthetic Rece

Re: Milter header order

2018-11-12 Thread Viktor Dukhovni
> On Nov 11, 2018, at 5:23 AM, Henrik K wrote: > > As I understand, the correct positioning of milter inserted internal headers > would be above postfix's own. But it seems all Authentication-Results are > added below it, so SpamAssassin won't consider them as internal. While RFC7001 defines "A

Re: Milter header order

2018-11-11 Thread Henrik K
On Sun, Nov 11, 2018 at 01:16:08PM +0200, Henrik K wrote: > > Compiled new opendmarc that inserts to index 0 and it ends up correctly > above postfix's header. I guess many people have this problem without > realizing it. Sigh, yet more custom patched packages to run.. FYI they are fixed in dev

Re: Milter header order

2018-11-11 Thread Henrik K
On Sun, Nov 11, 2018 at 12:23:53PM +0200, Henrik K wrote: > > I have milter chain opendkim->opendmarc->amavisd-milter for incoming > external mail. Postfix 3.1.0 from Ubuntu 16.04.5. > > As I understand, the correct positioning of milter inserted internal headers > would be above postfix's own.