X-Git-Url: http://gb7djk.dxcluster.net/gitweb/gitweb.cgi?a=blobdiff_plain;f=perl%2FBadWords.pm;h=312a04082c705fac550f9b0fcb065388cca2e564;hb=6ec22e78a4a344ce675645fabf18b2a1971f364a;hp=2336bb0468e22b1b95e3b517f5750c6397dd9fc2;hpb=2b58ccdf81685a1167a43c38705a0d84b9d8d661;p=spider.git

diff --git a/perl/BadWords.pm b/perl/BadWords.pm
index 2336bb04..312a0408 100644
--- a/perl/BadWords.pm
+++ b/perl/BadWords.pm
@@ -3,7 +3,7 @@
 #
 # Copyright (c) 2000 Dirk Koopman
 #
-# $Id$
+#
 #
 
 package BadWords;
@@ -12,49 +12,85 @@ use strict;
 
 use DXUtil;
 use DXVars;
-use IO::File;
+use DXHash;
+use DXDebug;
 
-use vars qw(%badwords $fn);
+use IO::File;
 
-$fn = "$main::data/badwords";
-%badwords = ();
+use vars qw($badword $regexcode);
 
-use vars qw($VERSION $BRANCH);
-$VERSION = sprintf( "%d.%03d", q$Revision$ =~ /(\d+)\.(\d+)/ );
-$BRANCH = sprintf( "%d.%03d", q$Revision$ =~ /\d+\.\d+\.(\d+)\.(\d+)/ ) || 0;
-$main::build += $VERSION;
-$main::branch += $BRANCH;
+our $regex;
 
 # load the badwords file
 sub load
 {
+	my $bwfn = localdata("badword");
+	filecopy("$main::data.issue", $bwfn) unless -e $bwfn;
+	
 	my @out;
-	return unless -e $fn;
-	my $fh = new IO::File $fn;
+
+	$badword = new DXHash "badword";
+	
+	push @out, create_regex(); 
+	return @out;
+}
+
+sub create_regex
+{
+	$regex = localdata("badw_regex");
+	filecopy("$regex.gb.issue", $regex) unless -e $regex;
+	
+	my @out;
+	my $fh = new IO::File $regex;
 	
 	if ($fh) {
-		%badwords = ();
+		my $s = "sub { my \$str = shift; my \@out; \n";
 		while (<$fh>) {
 			chomp;
 			next if /^\s*\#/;
 			my @list = split " ";
 			for (@list) {
-				$badwords{lc $_}++;
+				# create a closure for each word so that it matches stuff with spaces/punctuation
+				# and repeated characters in it
+				my $w = uc $_;
+				my @l = split //, $w;
+				my $e = join '+[\s\W]*', @l;
+				$s .= qq{push \@out, \$1 if \$str =~ m|\\b($e+)|;\n};
 			}
 		}
+		$s .= "return \@out;\n}";
+		$regexcode = eval $s;
+		dbg($s) if isdbg('badword');
+		if ($@) {
+			@out = ($@);
+			dbg($@);
+			return @out;
+		}
 		$fh->close;
 	} else {
-		my $l = "can't open $fn $!";
-		dbg('err', $l);
+		my $l = "can't open $regex $!";
+		dbg($l);
 		push @out, $l;
 	}
+	
 	return @out;
 }
 
 # check the text against the badwords list
 sub check
 {
-	return grep { $badwords{$_} } split(/\b/, lc shift);
+	my $s = uc shift;
+	my @out;
+
+	push @out, &$regexcode($s) if $regexcode;
+	
+	return @out if @out;
+	
+	for (split(/\b/, $s)) {
+		push @out, $_ if $badword->in($_);
+	}
+
+	return @out;
 }
 
 1;