Re: International Domain Characters and Header Checks

2022-05-30 Thread Viktor Dukhovni
On Thu, May 26, 2022 at 11:35:05PM +0200, Benny Pedersen wrote: > > SMTPUTF8 is enabled. > > so there is dns servers with domains in utf8 existing ? > > all my known dns servers uses idn, not eai, postfix imho need to convert > eai to idn dns to know if domain exists on dns > > to block domain

Re: International Domain Characters and Header Checks

2022-05-30 Thread Viktor Dukhovni
On Thu, May 26, 2022 at 03:39:08PM -0500, Bryan K. Walton wrote: > For example, one of our domain names is courseleaf.com. We want to > block any mail that has similar domain names in the From header. An > example might be: coǔrṣeleaf.com 1. Note that corresponding IDN name is: "xn--coreleaf-r

Re: RSA and ECDSA - warning: No certs for key at index 1

2022-05-30 Thread Viktor Dukhovni
On Mon, May 30, 2022 at 08:52:21AM +0200, Maurizio Caloro wrote: > try to install RSA and ECDSA, but it's don't run like normal mode. Simplest in most cases (and quite sufficient) to stick to just one algorithm. Multiple algorithms require a deeper understanding of what you're doing. > Generate

Re: IPv6 DNSRBLs

2022-05-30 Thread Peter
On 31/05/22 7:58 am, Wietse Venema wrote: Something in the postscreen_dnsbl_sites list that says "this DNSXL does/doesnot answer queries about a certain class of IP address".For example instead of site[filter]*weight, whether another suffix, or something like {site[filter]*weight, q=a, client=ipv

Re: IPv6 DNSRBLs

2022-05-30 Thread Wietse Venema
Peter: > On 31/05/22 7:14 am, Wietse Venema wrote: > > How urgent is this? > > Not very, targetting 3.8.0 should be sufficient. Implementing this for a stable release would never be an option. > > I think that we can do better than maintaining multiple parallel > > lists. For example, use one li

Re: IPv6 DNSRBLs

2022-05-30 Thread Peter
On 31/05/22 7:14 am, Wietse Venema wrote: How urgent is this? Not very, targetting 3.8.0 should be sufficient. At this time, I see only an optimization to avoid sending a query about an IPv4 client address to a DNSXL service that provides reputation only about IPv6 addresses (and vice versa).

Re: IPv6 DNSRBLs

2022-05-30 Thread Wietse Venema
How urgent is this? At this time, I see only an optimization to avoid sending a query about an IPv4 client address to a DNSXL service that provides reputation only about IPv6 addresses (and vice versa). With postscreen parallel DNSXL lookups that is not a disaster. I am concerned about the potenti

Re: devnull

2022-05-30 Thread Viktor Dukhovni
> On 30 May 2022, at 2:03 pm, Juerg Reimann wrote: > >> In virtual_alias_maps: >> >> donotreply@some.domain devnull@localhost >> >> Assuming that localhost is listed in $mydestination. >> >> Wietse > > Thanks a lot, the @localhost did the trick! Using local aliases(5) to /dev/null for t

Re: IPv6 DNSRBLs

2022-05-30 Thread Peter
On 31/05/22 1:44 am, Wietse Venema wrote: With 'IPv4 requests' and 'IPv6 requests' did you mean: - The type of the (reversed) client IP address? Yes. - The type (A, ) of the DNS query and response? Also yes, but only if that ends up being required by the IPv6 DNSRBLs. At present all th

Re: devnull

2022-05-30 Thread Viktor Dukhovni
On Mon, May 30, 2022 at 11:38:19AM -0400, Bill Cole wrote: > Just use a check_sender_access map (in one of the smtpd_*_restrictions > lists) that maps the address to DISCARD. That discards the message for *all* recipients, which is not the stated goal. -- Viktor.

Re: devnull

2022-05-30 Thread Bill Cole
On 2022-05-30 at 11:15:56 UTC-0400 (Mon, 30 May 2022 17:15:56 +0200) Juerg Reimann is rumored to have said: Hello everybody, I'm trying to setup a mail address that basically gets devnulled. I'll skip my standard rant about how terrible that almost always is and assume that you have good re

Re: devnull

2022-05-30 Thread Wietse Venema
Juerg Reimann: > Hello everybody, > > I'm trying to setup a mail address that basically gets devnulled. I have > virtual domains, so I did the usual: > > Excerpt from my main.cf: > > 30 alias_database = hash:/opt/local/etc/postfix/aliases > 31 alias_maps = $alias_database As documented - The

Re: devnull

2022-05-30 Thread Viktor Dukhovni
On Mon, May 30, 2022 at 05:15:56PM +0200, Juerg Reimann wrote: > I'm trying to setup a mail address that basically gets devnulled. I > have virtual domains, so I did the usual: > > Excerpt from my main.cf: > > 30 alias_database = hash:/opt/local/etc/postfix/aliases > 31 alias_maps = $alias_datab

Re: IPv6 DNSRBLs

2022-05-30 Thread Wietse Venema
Peter: > We're now starting to see some IPv6 DNSRBLs (eg: > bl.ipv6.spameatingmonkey.net). It occurs to me that postscreen and > postfix should only be sending IPv4 requests to IPv4-specific DNSRBLs > and IPv6 requests to IPv6-specific lists. I would therefore hope for > consideration of the

Re: postscreen_dnsbl_sites questions about multiple matches.

2022-05-30 Thread Matus UHLAR - fantomas
On 30.05.22 14:02, Peter wrote: Next question: What happens if zen returns multiple responses: 127.0.0.10 127.0.0.3 postscreen_dnsbl_sites = zen.spamhaus.org=127.0.0.[1..2]*3 zen.spamhaus.org=127.0.0.3*2 zen.spamhaus.org=127.0.0.[4..255]*3 On 30.05.22 10:06, Matus UHLAR - fantomas wrote: th

Re: postscreen_dnsbl_sites questions about multiple matches.

2022-05-30 Thread Matus UHLAR - fantomas
On 30.05.22 14:02, Peter wrote: First off my goal is that I want all zen.spamhaus.org entries to have a score of 3 except for CSS entries which should have a score of 2. zen returns 127.0.0.n for all entries and CSS specifically returns 127.0.0.3. What I think I can do is this: postscreen_d