diff -rbc FuzzyOcr-2.3e/CHANGES FuzzyOcr-2.3f/CHANGES
*** FuzzyOcr-2.3e/CHANGES	Tue Sep 12 08:38:42 2006
--- FuzzyOcr-2.3f/CHANGES	Tue Sep 12 11:56:52 2006
***************
*** 1,3 ****
--- 1,17 ----
+ version 2.3f:
+     Fixed:
+         Properly initialized $h and $w to zero so that when getting the height and width
+         from an image, if the size parameters cannot be parsed, they can get properly tested.
+ 
+     Fixed:
+         Hashing now works. $digest was getting reset because it went out of scope. grrr.
+ 
+     Fixed:
+         $efile was only being replaced for first occurrence in complex scansets.
+ 
+     Fixed:
+         Various bugs where: Use of uninitialized values were reported.
+ 
  version 2.3e:
      Fixed:
          Option: 'focr_db_safe'
diff -rbc FuzzyOcr-2.3e/FuzzyOcr.cf FuzzyOcr-2.3f/FuzzyOcr.cf
*** FuzzyOcr-2.3e/FuzzyOcr.cf	Tue Sep 12 08:28:36 2006
--- FuzzyOcr-2.3f/FuzzyOcr.cf	Tue Sep 12 11:43:48 2006
***************
*** 128,131 ****
--- 128,134 ----
  # Automatically add hashes of spam images recognized by OCR to the Image Hash database, to disable, set to 0.0 (Default value: 1)
  #focr_hashing_learn_scanned 1
  #
+ # If enabled, this will save image files in tempdir when errors occur to further debug the plugin. (Default: 0.0).
+ #focr_keep_bad_images 0.0
+ #
  ######################################################################
diff -rbc FuzzyOcr-2.3e/FuzzyOcr.pm FuzzyOcr-2.3f/FuzzyOcr.pm
*** FuzzyOcr-2.3e/FuzzyOcr.pm	Tue Sep 12 08:28:48 2006
--- FuzzyOcr-2.3f/FuzzyOcr.pm	Tue Sep 12 12:10:31 2006
***************
*** 90,96 ****
  
  our @pgm_opts = qw/personal_wordlist global_wordlist logfile
      threshold counts_required verbose timeout gif_max_frames
!     db_safe db_hash db_max_days path_bin scansets keep_bad_images
      enable_image_hashing digest_db hashing_learn_scanned/;
  
  our @paths = qw(/usr/local/netpbm/bin /usr/local/bin /usr/bin);
--- 90,96 ----
  
  our @pgm_opts = qw/personal_wordlist global_wordlist logfile
      threshold counts_required verbose timeout gif_max_frames
!     db_hash db_safe db_max_days path_bin scansets keep_bad_images
      enable_image_hashing digest_db hashing_learn_scanned/;
  
  our @paths = qw(/usr/local/netpbm/bin /usr/local/bin /usr/bin);
***************
*** 501,514 ****
  }
              
  sub add_image_hash_db {
!     my $digest = shift;
!     my $score  = shift;
      my $ret = 0;
  
      if ($Option{enable_image_hashing} == 2) {
!         my $dbfile = shift || $Option{db_hash};
!         my $fname  = shift;
!         my $ctype  = shift;
          my %DB = ();
          tie %DB, 'MLDBM', $dbfile or $ret++;
          if ($ret>0) {
--- 501,514 ----
  }
              
  sub add_image_hash_db {
!     my $digest = $_[0];
!     my $score  = $_[1];
      my $ret = 0;
  
      if ($Option{enable_image_hashing} == 2) {
!         my $dbfile = $_[2] || $Option{db_hash};
!         my $fname  = $_[3];
!         my $ctype  = $_[4];
          my %DB = ();
          tie %DB, 'MLDBM', $dbfile or $ret++;
          if ($ret>0) {
***************
*** 549,555 ****
  
  sub calc_image_hash {
      my $pfile = $_[0];
!     my ($rcode, $hash, $h, $w);
      my $s = -s $pfile;
  
      foreach my $a (qw/identify ppmhist/) {
--- 549,555 ----
  
  sub calc_image_hash {
      my $pfile = $_[0];
!     my ($rcode, $hash);
      my $s = -s $pfile;
  
      foreach my $a (qw/identify ppmhist/) {
***************
*** 567,574 ****
      if ($rcode) {
          chomp $rcode;
          debuglog("$App{identify}: Timed out [$rcode], skipping...");
!         return (1, $hash);
      }
      foreach (@stdout_data) {
          if ($_ =~ /(\d+)x(\d+)/) {
              $h = $1;
--- 567,576 ----
      if ($rcode) {
          chomp $rcode;
          debuglog("$App{identify}: Timed out [$rcode], skipping...");
!         return (1, '');
      }
+ 
+     my ($h,$w) = (0,0);
      foreach (@stdout_data) {
          if ($_ =~ /(\d+)x(\d+)/) {
              $h = $1;
***************
*** 578,584 ****
      }
      if ($h == 0 or $w == 0) {
          debuglog("Unable to determine size of image, skipping...");
!         return(1,$hash);
      }
      $rcode = $t->run_and_catch(sub {
          @stdout_data = qx($App{ppmhist} -noheader $pfile 2>/dev/null);
--- 580,586 ----
      }
      if ($h == 0 or $w == 0) {
          debuglog("Unable to determine size of image, skipping...");
!         return(1,'');
      }
      $rcode = $t->run_and_catch(sub {
          @stdout_data = qx($App{ppmhist} -noheader $pfile 2>/dev/null);
***************
*** 586,592 ****
      if ($rcode) {
          chomp $rcode;
          debuglog("$App{ppmhist}: Timed out [$rcode], skipping...");
!         return (1, $hash);
      }
      my $cnt = 0;
      my $c = scalar(@stdout_data);
--- 588,594 ----
      if ($rcode) {
          chomp $rcode;
          debuglog("$App{ppmhist}: Timed out [$rcode], skipping...");
!         return (1, '');
      }
      my $cnt = 0;
      my $c = scalar(@stdout_data);
***************
*** 617,623 ****
  
  sub wrong_ctype {
      my ( $format, $ctype ) = @_;
!     if ($Score{wrongctype}) {
          my $debuginfo = "";
          if ( $Option{verbose} > 0 ) {
              $debuginfo =
--- 619,625 ----
  
  sub wrong_ctype {
      my ( $format, $ctype ) = @_;
!     if ($Score{wctypescore}) {
          my $debuginfo = "";
          if ( $Option{verbose} > 0 ) {
              $debuginfo =
***************
*** 626,634 ****
          }
          for my $set ( 0 .. 3 ) {
              $pms->{conf}->{scoreset}->[$set]->{"FUZZY_OCR_WRONG_CTYPE"} =
!               sprintf( "%0.3f", $Score{wrongctype} );
          }
!         $pms->_handle_hit( "FUZZY_OCR_WRONG_CTYPE", $Score{wrongctype}, "BODY: ",
              $pms->{conf}->{descriptions}->{FUZZY_OCR_WRONG_CTYPE} . "\n$debuginfo" );
      }
  }
--- 628,636 ----
          }
          for my $set ( 0 .. 3 ) {
              $pms->{conf}->{scoreset}->[$set]->{"FUZZY_OCR_WRONG_CTYPE"} =
!               sprintf( "%0.3f", $Score{wctypescore} );
          }
!         $pms->_handle_hit( "FUZZY_OCR_WRONG_CTYPE", $Score{wctypescore}, "BODY: ",
              $pms->{conf}->{descriptions}->{FUZZY_OCR_WRONG_CTYPE} . "\n$debuginfo" );
      }
  }
***************
*** 693,705 ****
          debuglog("Scan canceled, message has already more than $Score{autodisable} points.");
          return 0;
      }
-     my $imgdir = Mail::SpamAssassin::Util::secure_tmpdir();
-     unless ($imgdir) {
-         debuglog("Scan canceled, cannot create Image TMPDIR.");
-         return 0;
-     }
-     debuglog("Using: $imgdir");
  
      my %imgfiles = ();
      my @found    = ();
      my @hashes   = ();
--- 695,702 ----
          debuglog("Scan canceled, message has already more than $Score{autodisable} points.");
          return 0;
      }
  
+     my $imgdir;
      my %imgfiles = ();
      my @found    = ();
      my @hashes   = ();
***************
*** 732,737 ****
--- 729,739 ----
              next;
          }
  
+         $imgdir = Mail::SpamAssassin::Util::secure_tmpdir() unless ($imgdir);
+         unless ($imgdir) {
+             debuglog("Scan canceled, cannot create Image TMPDIR.");
+             return 0;
+         }
          my $imgfilename = $imgdir . "/" . $fname;
          my $unique = 0;
          while (-e $imgfilename) {
***************
*** 742,763 ****
              debuglog("Cannot write \"$imgfilename\", skipping...");
              next;
          }
-         binmode PICT;
          my $pdata = $p->decode();
          print PICT $pdata;
          close PICT;
          $imgfiles{$imgfilename}{header} = substr($pdata,0,6);
          $imgfiles{$imgfilename}{ctype} = $ctype;
          $imgfiles{$imgfilename}{fname} = $fname;
      }
  
!     unless (keys %imgfiles) {
!         rmdir $imgdir;
          debuglog("Skipping OCR, no image files found...");
          return 0;
      }
! 
! 
      my $t = Mail::SpamAssassin::Timeout->new({ secs => $Option{timeout} });
      my $retcode;
  
--- 744,764 ----
              debuglog("Cannot write \"$imgfilename\", skipping...");
              next;
          }
          my $pdata = $p->decode();
+         binmode PICT;
          print PICT $pdata;
          close PICT;
+         debuglog("Wrote: $imgfilename"); $cnt++;
          $imgfiles{$imgfilename}{header} = substr($pdata,0,6);
          $imgfiles{$imgfilename}{ctype} = $ctype;
          $imgfiles{$imgfilename}{fname} = $fname;
      }
  
!     if ($cnt == 0) {
          debuglog("Skipping OCR, no image files found...");
          return 0;
      }
!     debuglog("Found: $cnt images"); $cnt = 0;
      my $t = Mail::SpamAssassin::Timeout->new({ secs => $Option{timeout} });
      my $retcode;
  
***************
*** 771,777 ****
          my $ptype = 0;
          my $tfile = $file;
          my $pfile = $file . ".pnm";
!         my $efile = $file . ".err";
  
          if ( substr($$pic{header},0,3) eq "\x47\x49\x46" ) {
              debuglog("Found GIF header name=\"$$pic{fname}\"");
--- 772,778 ----
          my $ptype = 0;
          my $tfile = $file;
          my $pfile = $file . ".pnm";
!         my $efile = $file . ".stderr";
  
          if ( substr($$pic{header},0,3) eq "\x47\x49\x46" ) {
              debuglog("Found GIF header name=\"$$pic{fname}\"");
***************
*** 823,829 ****
                      debuglog("$App{giffix}: Timed out [$retcode], skipping...");
                      ++$imgerr if $Option{keep_bad_images}; next;
                  }
!                 open ERR, "< $efile";
                  @stderr_data = <ERR>;
                  close ERR;
                  foreach (@stderr_data) {
--- 824,830 ----
                      debuglog("$App{giffix}: Timed out [$retcode], skipping...");
                      ++$imgerr if $Option{keep_bad_images}; next;
                  }
!                 if (open ERR,$efile) {
                      @stderr_data = <ERR>;
                      close ERR;
                      foreach (@stderr_data) {
***************
*** 833,838 ****
--- 834,840 ----
                          }
                      }
                  }
+             }
  
              if ($corrupt) {
                  if ($interlaced_gif or ($image_count gt 1)) {
***************
*** 879,886 ****
                      };
                      my $fs = 0;
                      foreach my $n (0 .. $image_count) {
!                         my $f = sprintf "%s/out%02d",$imgdir,$n;
!                         $tfile = $f if ($fs < -s $f);
                      }
                  }
              }
--- 881,891 ----
                      };
                      my $fs = 0;
                      foreach my $n (0 .. $image_count) {
!                         my $f = Mail::SpamAssassin::Util::untaint_file_path(
!                             sprintf("%s/out%02d",$imgdir,$n)
!                         );
!                         my $s = -s $f || 0;
!                         $tfile = $f if ($fs < $s);
                      }
                  }
              }
***************
*** 1001,1008 ****
  
          if($Option{enable_image_hashing}) {
              debuglog("Calculating the image hash...");
!             my ($rcode, $digest) = calc_image_hash($pfile);
!             if ($rcode) {
                  debuglog("Error calculating the image hash, skipping hash check...");
              } else {
                  if (my $score = check_image_hash_db($digest, $Option{db_hash}, $$pic{fname}, $$pic{ctype})) {
--- 1006,1013 ----
  
          if($Option{enable_image_hashing}) {
              debuglog("Calculating the image hash...");
!             ($corrupt, $digest) = calc_image_hash($pfile);
!             if ($corrupt) {
                  debuglog("Error calculating the image hash, skipping hash check...");
              } else {
                  if (my $score = check_image_hash_db($digest, $Option{db_hash}, $$pic{fname}, $$pic{ctype})) {
***************
*** 1028,1034 ****
              my $scan = $scanset;
              $scan =~ s/\$gocr/$App{gocr}/;
              $scan =~ s/\$pfile/$pfile/;
!             $scan =~ s/\$efile/$efile/;
              unlink $efile if -e $efile;
              debuglog("Trying: $scanset");
              my @ocrdata;
--- 1033,1039 ----
              my $scan = $scanset;
              $scan =~ s/\$gocr/$App{gocr}/;
              $scan =~ s/\$pfile/$pfile/;
!             $scan =~ s/\$efile/$efile/g;
              unlink $efile if -e $efile;
              debuglog("Trying: $scanset");
              my @ocrdata;