[popfile-commit] engine/Classifier MailParse.pm,1.117,1.118

[<jgr...@us...>]

Update of /cvsroot/popfile/engine/Classifier
In directory sc8-pr-cvs1:/tmp/cvs-serv27871/Classifier

Modified Files:
	MailParse.pm 
Log Message:
Added pseudowords html:imgremotesrc to catch The Big Picture spams

Index: MailParse.pm
===================================================================
RCS file: /cvsroot/popfile/engine/Classifier/MailParse.pm,v
retrieving revision 1.117
retrieving revision 1.118
diff -C2 -d -r1.117 -r1.118
*** MailParse.pm	27 Apr 2003 00:11:57 -0000	1.117
--- MailParse.pm	27 Apr 2003 00:20:51 -0000	1.118
***************
*** 408,412 ****
               ( ( $tag =~ /^img|frame|iframe$/i )
                 || ( $tag =~ /^script$/i && $parse_script_uri ) ) ) {
!             add_url( $self, $value, $encoded, $quote, $end_quote, '' );
              next;
          }
--- 408,423 ----
               ( ( $tag =~ /^img|frame|iframe$/i )
                 || ( $tag =~ /^script$/i && $parse_script_uri ) ) ) {
!             my $host = add_url( $self, $value, $encoded, $quote, $end_quote, '' );
! 
! 	    # If the host name is not blank (i.e. there was a hostname in the url
! 	    # and it was an image, then if the host was not this host then report
! 	    # an off machine image
! 
!             if ( ( $host ne '' ) && ( $tag =~ /^img$/i ) ) {
! 	        if ( $host ne 'localhost' ) {
!                     $self->update_pseudoword( 'html', 'imgremotesrc' );
! 	        }
! 	    }
! 
              next;
          }
***************
*** 539,542 ****
--- 550,555 ----
  #               identification of values found in for example the subject line
  #
+ # Returns the hostname
+ #
  # ---------------------------------------------------------------------------------------------
  sub add_url
***************
*** 642,648 ****
      }
  
!     if ( !defined $host || $host eq '' ) {
          print "no hostname found: [$temp_url]\n" if ($self->{debug});
!         return 0;
      }
  
--- 655,661 ----
      }
  
!     if ( !defined( $host ) || ( $host eq '' ) ) {
          print "no hostname found: [$temp_url]\n" if ($self->{debug});
!         return '';
      }
  
***************
*** 652,656 ****
      $hash = $1 if ( $url =~ s/^[\#](.*)$// );
  
!     if ( !defined $protocol || $protocol =~ /^(http|https)$/ ) {
          $temp_before = $before;
          $temp_before = "\:\/\/" if (defined $protocol);
--- 665,669 ----
      $hash = $1 if ( $url =~ s/^[\#](.*)$// );
  
!     if ( !defined( $protocol ) || ( $protocol =~ /^(http|https)$/ ) ) {
          $temp_before = $before;
          $temp_before = "\:\/\/" if (defined $protocol);
***************
*** 678,682 ****
  
      # $protocol $authinfo $host $port $query $hash may be processed below if desired
!     return 1;
  }
  
--- 691,695 ----
  
      # $protocol $authinfo $host $port $query $hash may be processed below if desired
!     return $host;
  }