On Wed, 2014-10-22 at 13:40 -0500, Jesse Stroik wrote: > I noticed URLs from the TLD .link aren't properly classified on my mail > server. I wrote a simple URI rule to recognize that TLD which never > matched. I wrote a similar body rule, which did properly match. > Interestingly, I do see DNS queries going out for the URLs in question. > > This is sa 3.3.2-4 -- is it a known issue? The URL in question is on a > single line and is easily pulled out with egrep and properly parsed with > the body rule. > As others have already said, URI body rules use a list of valid TLDs to help with recognising URIs embedded in body text and this list is currently hardcoded into SA.
However, this doesn't affect any rules you might write to match domain names in headers, so rules that use a regex to look for .link domains in, for instance, Received or Reply-to headers will work as you'd expect them to. So, If you don't want to mess around with replacing the RegistrarBoundaries.pm file in your installation, you may care to write a few rules that work with the headers and use them until a version of SA with a configurable TLD list is released. I'm currently using this meta-rule: describe MG_LINK_TLD Messages from or containing a URL with the .link TLD uri __MG_LTD1 /\.link/i header __MG_LTD2 From =~ /\.link>/ header __MG_LTD3 Received =~ /from.*\.link/ header __MG_LTD4 Return-Path =~ /\.link/ meta MG_LINK_TLD (__MG_LTD1 || __MG_LTD2 || __MG_LTD3 || __MG_LTD4) score MG_LINK_TLD 7.5 which I've tested fairly carefully. All the subrules except __MG_TLD1 work exactly as I wanted them to. I can live with __MG_TLD1 not working until either a current SA maintenance version is released with an extended list of hardcoded TLDs or a version using a configurable list appears. HTH Martin
