[popfile-commit] engine/Classifier Bayes.pm,1.138,1.139 MailParse.pm,1.116,1.117

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

Update of /cvsroot/popfile/engine/Classifier
In directory sc8-pr-cvs1:/tmp/cvs-serv22470/Classifier

Modified Files:
	Bayes.pm MailParse.pm 
Log Message:
Added new pseudowords html:imgwidthX and html:imgheightY used to catch web bugs

Index: Bayes.pm
===================================================================
RCS file: /cvsroot/popfile/engine/Classifier/Bayes.pm,v
retrieving revision 1.138
retrieving revision 1.139
diff -C2 -d -r1.138 -r1.139
*** Bayes.pm	26 Apr 2003 09:43:20 -0000	1.138
--- Bayes.pm	27 Apr 2003 00:11:57 -0000	1.139
***************
*** 1287,1291 ****
  
              if ( /([^\s]+) (\d+)/ ) {
!                 my $word = $self->{mangler__}->mangle($1,1);
                  my $value = $2;
                  $value =~ s/[\r\n]//g;
--- 1287,1291 ----
  
              if ( /([^\s]+) (\d+)/ ) {
!                 my $word  = $1;
                  my $value = $2;
                  $value =~ s/[\r\n]//g;
***************
*** 1345,1349 ****
  
              if ( /([^\s]+) (\d+)/ ) {
!                 my $word = $self->{mangler__}->mangle($1,1);
                  my $value = $2;
                  $value =~ s/[\r\n]//g;
--- 1345,1349 ----
  
              if ( /([^\s]+) (\d+)/ ) {
!                 my $word  = $1;
                  my $value = $2;
                  $value =~ s/[\r\n]//g;

Index: MailParse.pm
===================================================================
RCS file: /cvsroot/popfile/engine/Classifier/MailParse.pm,v
retrieving revision 1.116
retrieving revision 1.117
diff -C2 -d -r1.116 -r1.117
*** MailParse.pm	25 Apr 2003 07:02:06 -0000	1.116
--- MailParse.pm	27 Apr 2003 00:11:57 -0000	1.117
***************
*** 424,433 ****
                  }
              } else {
                  # Anything that isn't a mailto is probably an URL
!                 
                  $self->add_url($value, $encoded, $quote, $end_quote, '');
              }
!             
!             next;            
          }
  
--- 424,434 ----
                  }
              } else {
+ 
                  # Anything that isn't a mailto is probably an URL
! 
                  $self->add_url($value, $encoded, $quote, $end_quote, '');
              }
! 
!             next;
          }
  
***************
*** 455,458 ****
--- 456,460 ----
  
          # Tags with colors in them
+ 
          if ( ( $attribute =~ /^color$/i ) && ( $tag =~ /^font$/i ) ) {
              update_word( $self, $value, $encoded, $quote, $end_quote, '' );
***************
*** 467,471 ****
--- 469,481 ----
          }
  
+         # The width and height of images
+ 
+         if ( ( $attribute =~ /^(width|height)$/i ) && ( $tag =~ /^img$/i ) ) {
+             $attribute = lc( $attribute );
+             $self->update_pseudoword( 'html', "img$attribute$value" );
+ 	}
+ 
          # Font sizes
+ 
          if ( ( $attribute =~ /^size$/i ) && ( $tag =~ /^font$/i ) ) {
              $self->update_pseudoword( 'html', "fontsize$value" );