[Blends-commit] r3105 - /blends/trunk/team_analysis_tools/get-archive-pages

[tille at users.alioth.debian.org]

Author: tille
Date: Thu Jan  5 22:13:31 2012
New Revision: 3105

URL: http://svn.debian.org/wsvn/blends/?sc=1&rev=3105
Log:
Handle spammy subjects

Modified:
    blends/trunk/team_analysis_tools/get-archive-pages

Modified: blends/trunk/team_analysis_tools/get-archive-pages
URL: http://svn.debian.org/wsvn/blends/blends/trunk/team_analysis_tools/get-archive-pages?rev=3105&op=diff
==============================================================================
--- blends/trunk/team_analysis_tools/get-archive-pages (original)
+++ blends/trunk/team_analysis_tools/get-archive-pages Thu Jan  5 22:13:31 2012
@@ -333,11 +333,20 @@
 					last;
 				    }
 				}
+			        my $spamsubject;
+			        my $numspamsubjects = 0;
+				foreach $spamsubject (@SPAMSUBJECTS) {
+				    if ( $subject =~ /$spamauthor/ ) {
+					$numspamsubjects = 1;
+					last;
+				    }
+				}
+				
 				# authors with strange names also provide spam
 			        my $countstrangechars = 0;
 			        while ($author =~ /;\s*&#x[\dA-F][\dA-F][\dA-F]/g) { $countstrangechars++ }
 			        
-				if ( $author =~ /^[-&#x\d;\sA-F\?:,]+$/ || $countstrangechars > 7 || $numspamauthors > 0 ) {
+				if ( $author =~ /^[-&#x\d;\sA-F\?:,]+$/ || $countstrangechars > 7 || $numspamauthors > 0 || $numspamsubjects > 0 ) {
                                     # $author =~ /info/i ) { # never had a non-spam message from an author whos name contains info
                                     # This is not reliable: We have for instance
                                     #   http://lists.debian.org/debian-security/2003/07/msg00054.html