http://bugzilla.spamassassin.org/show_bug.cgi?id=4191
------- Additional Comments From [EMAIL PROTECTED] 2005-03-14 11:52 ------- Subject: Re: New: uri_to_domain() is broken for urls with empty port. On Mon, Mar 14, 2005 at 08:43:12AM -0800, [EMAIL PROTECTED] wrote: > uri_to_domain() does not properly handle urls in the form: > > http://blah.blah.com:/ I was going to say this was fixed in 3.1, but apparently it's not, uri_to_domain() indeed doesn't DTRT thing. :( I committed a fix to trunk, r157454. > In particular, the ":" at the end is not handled properly. Spammers are > apparentlyi aware of this and are taking advantage of this problem to > circumvent > URIBL checks. I get about 30 messages a day with URL's that look like this: > > http://crazyrxl0wprices.net:/ Interestingly enough, the URIBL plugin does deal with this in 3.1 already... debug: uridnsbl: domains to query: crazyrxl0wprices.com [...] debug: uri: uri found, http://crazyrxl0wprices.com/ debug: uri: uri found, http://crazyrxl0wprices.com:/ Aha. A different bug fixes this one, in a round-about way: # bug 3186: If in a sentence, we might pick up odd characters ... # ie: "visit http://example.biz."; or "visit http://example.biz!!!"; # the host portion should end in some form of alpha-numeric, strip off # the rest. if ($host =~ s/[^0-9A-Za-z]+$//) { push(@nuris, join ('', $proto, $host, $rest)); } ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee.
