Bgbot — a MediaWiki robot Code

--- a/trunk/utils.php
+++ b/trunk/utils.php
@@ -20,8 +20,8 @@
 #  59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
 #  http://www.gnu.org/copyleft/gpl.html
 #
-#  Author: Borislav Manolov <b.manolov at web.de>
-#          http://purl.oclc.org/NET/manolov/
+#  Author: Borislav Manolov <b.manolov at gmail dot com>
+#          http://purl.org/NET/borislav/
 #############################################################################
 
 
@@ -36,150 +36,160 @@
 
 
 $lats = array(
-  'a', 'b', 'v', 'g', 'd', 'e', 'zh', 'z', 'i', 'j',
-  'k', 'l', 'm', 'n', 'o', 'p', 'r', 's', 't', 'u',
-  'f', 'h', 'c', 'ch', 'sh', 'sht', 'j', 'y', 'ju', 'ja',
-  'A', 'B', 'V', 'G', 'D', 'E', 'Zh', 'Z', 'I', 'J',
-  'K', 'L', 'M', 'N', 'O', 'P', 'R', 'S', 'T', 'U',
-  'F', 'H', 'C', 'Ch', 'Sh', 'Sht', 'J', 'Y', 'Ju', 'Ja',
-  ' ');
+'a', 'b', 'v', 'g', 'd', 'e', 'zh', 'z', 'i', 'j',
+'k', 'l', 'm', 'n', 'o', 'p', 'r', 's', 't', 'u',
+'f', 'h', 'c', 'ch', 'sh', 'sht', 'j', 'y', 'ju', 'ja',
+'A', 'B', 'V', 'G', 'D', 'E', 'Zh', 'Z', 'I', 'J',
+'K', 'L', 'M', 'N', 'O', 'P', 'R', 'S', 'T', 'U',
+'F', 'H', 'C', 'Ch', 'Sh', 'Sht', 'J', 'Y', 'Ju', 'Ja',
+' ');
 
 
 # change the string encoding from windows-1251 to UTF-8
 function utf8($str) {
-  return iconv('windows-1251', 'UTF-8', $str);
+	return iconv('windows-1251', 'UTF-8', $str);
 }
 
 
 function cyr2hex($str) {
-  return urlencode( iconv('windows-1251', 'UTF-8', $str) );
+	return urlencode( iconv('windows-1251', 'UTF-8', $str) );
 }
 
 
 function cyr2lat($str) {
-  global $cyrs, $lats;
-  return str_replace($cyrs, $lats, $str);
+	global $cyrs, $lats;
+	return str_replace($cyrs, $lats, $str);
 }
 
 
 # read a line from the console
 function my_readline($str = 'Enter required data: ') {
-  echo "\n$str";
+	echo "\n$str";
 
-  $fp = fopen('php://stdin', 'r');
-  $resp = trim( fgets($fp) );
-  fclose($fp);
-  return $resp;
+	$fp = fopen('php://stdin', 'r');
+	$resp = trim( fgets($fp) );
+	fclose($fp);
+	return $resp;
 }
 
 
 function my_fwrite($file, $text, $mode='a+') {
-  $myFile = @fopen($file, $mode);
-  if (! $myFile) return 0;
-  if (! @fputs($myFile, $text)) return 0;
-  if (! @fclose($myFile)) return 0;
-  return 1;
+	$myFile = @fopen($file, $mode);
+	if (! $myFile) return false;
+	flock($myFile, LOCK_EX);
+	if (! @fputs($myFile, $text)) return false;
+	flock($myFile, LOCK_UN);
+	if (! @fclose($myFile)) return false;
+	return true;
 }
 
 
 function my_urlencode($str) {
-  $len = strlen($str);
+	$len = strlen($str);
 	for ($i = 0; $i < $len; $i++) {
 		$nstr .= $str{$i} != '%' ? urlencode($str{$i}) : '%';
 	}
-  return $nstr;
+	return $nstr;
 }
 
 
 # returns the text inside the $element element
 function getCDATA($element, $data) {
-  preg_match("/<$element>(.+)<\/$element>/s", $data, $matches);
-  return $matches[1];
+	preg_match("/<$element.*>(.+)<\/$element>/Us", $data, $matches);
+	return $matches[1];
 }
 
 # returns the text inside the $element elements
 function getCDATAs($element, $data) {
-  preg_match_all("/<$element>(.+)<\/$element>/s", $data, $matches);
-  return $matches[1];
+	preg_match_all("/<$element.*>(.+)<\/$element>/Us", $data, $matches);
+	return $matches[1];
 }
 
 # original source: http://www.randomchaos.com/document.php?source=php_and_unicode
 function utf82unicode($str) {
 
-  $unicode = '';
-  $values = array();
-  $lookingFor = 1;
+	$unicode = '';
+	$values = array();
+	$lookingFor = 1;
 
-  for ($i = 0; $i < strlen($str); $i++ ) {
-    $thisValue = ord( $str{$i} );
-    if ( $thisValue < 128 ) {
-      $unicode .= $str{$i}; # don't convert ASCII characters
-    } else {
-      if ( count($values) == 0 )
-        $lookingFor = ( $thisValue < 224 ) ? 2 : 3;
+	for ($i = 0; $i < strlen($str); $i++ ) {
+		$thisValue = ord( $str{$i} );
+		if ( $thisValue < 128 ) {
+			$unicode .= $str{$i}; # don't convert ASCII characters
+		} else {
+			if ( count($values) == 0 ) {
+				$lookingFor = ( $thisValue < 224 ) ? 2 : 3;
+			}
 
-      $values[] = $thisValue;
+			$values[] = $thisValue;
 
-      if ( count( $values ) == $lookingFor ) {
-        $number = ( $lookingFor == 3 )
-                ?
-          ( ( $values[0] % 16 ) * 4096 ) + ( ( $values[1] % 64 ) * 64 ) + ( $values[2] % 64 )
-                :
-          ( ( $values[0] % 32 ) * 64 ) + ( $values[1] % 64 );
+			if ( count( $values ) == $lookingFor ) {
+				$number = ( $lookingFor == 3 )
+					?
+				( ( $values[0] % 16 ) * 4096 ) + ( ( $values[1] % 64 ) * 64 ) + ( $values[2] % 64 )
+					:
+				( ( $values[0] % 32 ) * 64 ) + ( $values[1] % 64 );
 
-        $unicode .= "&#$number;";
-        $values = array();
-        $lookingFor = 1;
-        } // if
-      } // if
-  } // for
-  return $unicode;
-} // utf82unicode
+				$unicode .= "&#$number;";
+				$values = array();
+				$lookingFor = 1;
+			}
+		}
+	}
+	return $unicode;
+}
 
 
 # encodes a unicode sequence as UTF-8
 function unicode2utf8($unicode_str) {
-  $base = 10;
-  if ($unicode_str[0] == 'x') {
-    $unicode_str = substr($unicode_str, 1);
-    $base = 16;
-  }
-  $bin = base_convert($unicode_str, $base, 2);
-  $len = strlen($bin);
-  if ( $len < 8 ) {
-    // 0xxxxxxx
-    $bin = '0' . str_pad($bin, 7, '0', STR_PAD_LEFT);
-  } elseif ( $len < 12 ) {
-    // 110xxxxx 10xxxxxx
-    $bin = str_pad($bin, 11, '0', STR_PAD_LEFT);
-    $bin = substr_replace($bin, '10', 5, 0);
-    $bin = '110' . $bin;
-  } elseif ( $len < 17 ) {
-    // 1110xxxx 10xxxxxx 10xxxxxx
-    $bin = str_pad($bin, 16, '0', STR_PAD_LEFT);
-    $bin = substr_replace($bin, '10', 4, 0);
-    $bin = substr_replace($bin, '10', 12, 0);
-    $bin = '1110' . $bin;
-  } else {
-    // 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx
-    $bin = str_pad($bin, 21, '0', STR_PAD_LEFT);
-    $bin = substr_replace($bin, '10', 3, 0);
-    $bin = substr_replace($bin, '10', 11, 0);
-    $bin = substr_replace($bin, '10', 19, 0);
-    $bin = '11110' . $bin;
-  }
+	$base = 10;
+	if ($unicode_str[0] == 'x') {
+		$unicode_str = substr($unicode_str, 1);
+		$base = 16;
+	}
+	$bin = base_convert($unicode_str, $base, 2);
+	$len = strlen($bin);
+	if ( $len < 8 ) {
+		// 0xxxxxxx
+		$bin = '0' . str_pad($bin, 7, '0', STR_PAD_LEFT);
+	} elseif ( $len < 12 ) {
+		// 110xxxxx 10xxxxxx
+		$bin = str_pad($bin, 11, '0', STR_PAD_LEFT);
+		$bin = substr_replace($bin, '10', 5, 0);
+		$bin = '110' . $bin;
+	} elseif ( $len < 17 ) {
+		// 1110xxxx 10xxxxxx 10xxxxxx
+		$bin = str_pad($bin, 16, '0', STR_PAD_LEFT);
+		$bin = substr_replace($bin, '10', 4, 0);
+		$bin = substr_replace($bin, '10', 12, 0);
+		$bin = '1110' . $bin;
+	} else {
+		// 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx
+		$bin = str_pad($bin, 21, '0', STR_PAD_LEFT);
+		$bin = substr_replace($bin, '10', 3, 0);
+		$bin = substr_replace($bin, '10', 11, 0);
+		$bin = substr_replace($bin, '10', 19, 0);
+		$bin = '11110' . $bin;
+	}
 
-  $hex = base_convert($bin, 2, 16);
-  $hex = strtoupper($hex);
-  $len = strlen($hex);
-  if ($len == 1) {
-    $utf8 = '%0' . $hex;
-  } else {
-    for ($i = 0; $i < strlen($hex); $i+=2) {
-      $utf8 .= '%' . $hex[$i] . $hex[$i+1];
-    }
-  }
+	$hex = base_convert($bin, 2, 16);
+	$hex = strtoupper($hex);
+	$len = strlen($hex);
+	if ($len == 1) {
+		$utf8 = '%0' . $hex;
+	} else {
+		for ($i = 0; $i < strlen($hex); $i+=2) {
+			$utf8 .= '%' . $hex[$i] . $hex[$i+1];
+		}
+	}
 
-  return $utf8;
+	return $utf8;
 }
-?>
+
+# escape all regexp meta-characters in a string
+function escape_regexp($str) {
+	return str_replace(
+		array('\\', '^', '$', '.', '[', ']', '|', '(', ')', '?', '*', '+', '{', ']', '/'),
+		array('\\\\', '\^', '\$', '\.', '\[', '\]', '\|', '\(', '\)', '\?', '\*', '\+', '\{', '\]', '\/'),
+		$str);
+}