)/Us', $text, -1, PREG_SPLIT_DELIM_CAPTURE);
$stop = count($textarr);
// if a plugin has provided an autocorrect array, use it
if ( isset($wp_cockneyreplace) ) {
$cockney = array_keys($wp_cockneyreplace);
$cockney_replace = array_values($wp_cockneyreplace);
} else {
$cockney = array("'tain't","'twere","'twas","'tis","'twill","'til","'bout","'nuff","'round","'cause");
$cockneyreplace = array("’tain’t","’twere","’twas","’tis","’twill","’til","’bout","’nuff","’round","’cause");
}
$static_characters = array_merge(array('---', ' -- ', '--', 'xn–', '...', '``', '\'s', '\'\'', ' (tm)'), $cockney);
$static_replacements = array_merge(array('—', ' — ', '–', 'xn--', '…', '“', '’s', '”', ' ™'), $cockneyreplace);
$dynamic_characters = array('/\'(\d\d(?:’|\')?s)/', '/(\s|\A|")\'/', '/(\d+)"/', '/(\d+)\'/', '/(\S)\'([^\'\s])/', '/(\s|\A)"(?!\s)/', '/"(\s|\S|\Z)/', '/\'([\s.]|\Z)/', '/(\d+)x(\d+)/');
$dynamic_replacements = array('’$1','$1‘', '$1″', '$1′', '$1’$2', '$1“$2', '”$1', '’$1', '$1×$2');
for ( $i = 0; $i < $stop; $i++ ) {
$curl = $textarr[$i];
if (isset($curl{0}) && '<' != $curl{0} && $next) { // If it's not a tag
// static strings
$curl = str_replace($static_characters, $static_replacements, $curl);
// regular expressions
$curl = preg_replace($dynamic_characters, $dynamic_replacements, $curl);
} elseif ( strstr($curl, '
', '', $text);
$text = str_replace('
', "\n", $text);
$text = str_replace('
', '', $text);
return $text;
}
function wpautop($pee, $br = 1) {
$pee = $pee . "\n"; // just to make things a little easier, pad the end
$pee = preg_replace('|
\s*
|', "\n\n", $pee);
// Space things out a little
$allblocks = '(?:table|thead|tfoot|caption|colgroup|tbody|tr|td|th|div|dl|dd|dt|ul|ol|li|pre|select|form|blockquote|address|math|style|script|object|input|param|p|h[1-6])';
$pee = preg_replace('!(<' . $allblocks . '[^>]*>)!', "\n$1", $pee);
$pee = preg_replace('!(' . $allblocks . '>)!', "$1\n\n", $pee);
$pee = str_replace(array("\r\n", "\r"), "\n", $pee); // cross-platform newlines
$pee = preg_replace("/\n\n+/", "\n\n", $pee); // take care of duplicates
$pee = preg_replace('/\n?(.+?)(?:\n\s*\n|\z)/s', "$1
\n", $pee); // make paragraphs, including one at the end
$pee = preg_replace('|\s*?
|', '', $pee); // under certain strange conditions it could create a P of entirely whitespace
$pee = preg_replace( '|(
]*>\s*)|', "$1
", $pee );
$pee = preg_replace('!
([^<]+)\s*?((?:div|address|form)[^>]*>)!', "
$1
$2", $pee);
$pee = preg_replace( '|
|', "$1
", $pee );
$pee = preg_replace('!
\s*(?' . $allblocks . '[^>]*>)\s*
!', "$1", $pee); // don't pee all over a tag
$pee = preg_replace("|
(
|", "$1", $pee); // problem with nested lists
$pee = preg_replace('|]*)>|i', "", $pee);
$pee = str_replace('
', '
', $pee);
$pee = preg_replace('!\s*(?' . $allblocks . '[^>]*>)!', "$1", $pee);
$pee = preg_replace('!(?' . $allblocks . '[^>]*>)\s*
!', "$1", $pee);
if ($br) {
$pee = preg_replace('/<(script|style).*?<\/\\1>/se', 'str_replace("\n", "", "\\0")', $pee);
$pee = preg_replace('|(?)\s*\n|', "
\n", $pee); // optionally make line breaks
$pee = str_replace('', "\n", $pee);
}
$pee = preg_replace('!(?' . $allblocks . '[^>]*>)\s*
!', "$1", $pee);
$pee = preg_replace('!
(\s*?(?:p|li|div|dl|dd|dt|th|pre|td|ul|ol)[^>]*>)!', '$1', $pee);
if ( strstr( $pee, ')(.*?)
!ise', " stripslashes('$1') . stripslashes(clean_pre('$2')) . '' ", $pee);
$pee = preg_replace( "|\n$|", '', $pee );
/**/
return $pee;
}
function seems_utf8($Str) { # by bmorel at ssi dot fr
for ($i=0; $i', '>', $text);
if ( 'double' === $quotes ) {
$text = str_replace('"', '"', $text);
} elseif ( 'single' === $quotes ) {
$text = str_replace("'", ''', $text);
} elseif ( $quotes ) {
$text = str_replace('"', '"', $text);
$text = str_replace("'", ''', $text);
}
return $text;
}
function utf8_uri_encode( $utf8_string, $length = 0 ) {
$unicode = '';
$values = array();
$num_octets = 1;
for ($i = 0; $i < strlen( $utf8_string ); $i++ ) {
$value = ord( $utf8_string[ $i ] );
if ( $value < 128 ) {
if ( $length && ( strlen($unicode) + 1 > $length ) )
break;
$unicode .= chr($value);
} else {
if ( count( $values ) == 0 ) $num_octets = ( $value < 224 ) ? 2 : 3;
$values[] = $value;
if ( $length && ( (strlen($unicode) + ($num_octets * 3)) > $length ) )
break;
if ( count( $values ) == $num_octets ) {
if ($num_octets == 3) {
$unicode .= '%' . dechex($values[0]) . '%' . dechex($values[1]) . '%' . dechex($values[2]);
} else {
$unicode .= '%' . dechex($values[0]) . '%' . dechex($values[1]);
}
$values = array();
$num_octets = 1;
}
}
}
return $unicode;
}
function remove_accents($string) {
if ( !preg_match('/[\x80-\xff]/', $string) )
return $string;
if (seems_utf8($string)) {
$chars = array(
// Decompositions for Latin-1 Supplement
chr(195).chr(128) => 'A', chr(195).chr(129) => 'A',
chr(195).chr(130) => 'A', chr(195).chr(131) => 'A',
chr(195).chr(132) => 'A', chr(195).chr(133) => 'A',
chr(195).chr(135) => 'C', chr(195).chr(136) => 'E',
chr(195).chr(137) => 'E', chr(195).chr(138) => 'E',
chr(195).chr(139) => 'E', chr(195).chr(140) => 'I',
chr(195).chr(141) => 'I', chr(195).chr(142) => 'I',
chr(195).chr(143) => 'I', chr(195).chr(145) => 'N',
chr(195).chr(146) => 'O', chr(195).chr(147) => 'O',
chr(195).chr(148) => 'O', chr(195).chr(149) => 'O',
chr(195).chr(150) => 'O', chr(195).chr(153) => 'U',
chr(195).chr(154) => 'U', chr(195).chr(155) => 'U',
chr(195).chr(156) => 'U', chr(195).chr(157) => 'Y',
chr(195).chr(159) => 's', chr(195).chr(160) => 'a',
chr(195).chr(161) => 'a', chr(195).chr(162) => 'a',
chr(195).chr(163) => 'a', chr(195).chr(164) => 'a',
chr(195).chr(165) => 'a', chr(195).chr(167) => 'c',
chr(195).chr(168) => 'e', chr(195).chr(169) => 'e',
chr(195).chr(170) => 'e', chr(195).chr(171) => 'e',
chr(195).chr(172) => 'i', chr(195).chr(173) => 'i',
chr(195).chr(174) => 'i', chr(195).chr(175) => 'i',
chr(195).chr(177) => 'n', chr(195).chr(178) => 'o',
chr(195).chr(179) => 'o', chr(195).chr(180) => 'o',
chr(195).chr(181) => 'o', chr(195).chr(182) => 'o',
chr(195).chr(182) => 'o', chr(195).chr(185) => 'u',
chr(195).chr(186) => 'u', chr(195).chr(187) => 'u',
chr(195).chr(188) => 'u', chr(195).chr(189) => 'y',
chr(195).chr(191) => 'y',
// Decompositions for Latin Extended-A
chr(196).chr(128) => 'A', chr(196).chr(129) => 'a',
chr(196).chr(130) => 'A', chr(196).chr(131) => 'a',
chr(196).chr(132) => 'A', chr(196).chr(133) => 'a',
chr(196).chr(134) => 'C', chr(196).chr(135) => 'c',
chr(196).chr(136) => 'C', chr(196).chr(137) => 'c',
chr(196).chr(138) => 'C', chr(196).chr(139) => 'c',
chr(196).chr(140) => 'C', chr(196).chr(141) => 'c',
chr(196).chr(142) => 'D', chr(196).chr(143) => 'd',
chr(196).chr(144) => 'D', chr(196).chr(145) => 'd',
chr(196).chr(146) => 'E', chr(196).chr(147) => 'e',
chr(196).chr(148) => 'E', chr(196).chr(149) => 'e',
chr(196).chr(150) => 'E', chr(196).chr(151) => 'e',
chr(196).chr(152) => 'E', chr(196).chr(153) => 'e',
chr(196).chr(154) => 'E', chr(196).chr(155) => 'e',
chr(196).chr(156) => 'G', chr(196).chr(157) => 'g',
chr(196).chr(158) => 'G', chr(196).chr(159) => 'g',
chr(196).chr(160) => 'G', chr(196).chr(161) => 'g',
chr(196).chr(162) => 'G', chr(196).chr(163) => 'g',
chr(196).chr(164) => 'H', chr(196).chr(165) => 'h',
chr(196).chr(166) => 'H', chr(196).chr(167) => 'h',
chr(196).chr(168) => 'I', chr(196).chr(169) => 'i',
chr(196).chr(170) => 'I', chr(196).chr(171) => 'i',
chr(196).chr(172) => 'I', chr(196).chr(173) => 'i',
chr(196).chr(174) => 'I', chr(196).chr(175) => 'i',
chr(196).chr(176) => 'I', chr(196).chr(177) => 'i',
chr(196).chr(178) => 'IJ',chr(196).chr(179) => 'ij',
chr(196).chr(180) => 'J', chr(196).chr(181) => 'j',
chr(196).chr(182) => 'K', chr(196).chr(183) => 'k',
chr(196).chr(184) => 'k', chr(196).chr(185) => 'L',
chr(196).chr(186) => 'l', chr(196).chr(187) => 'L',
chr(196).chr(188) => 'l', chr(196).chr(189) => 'L',
chr(196).chr(190) => 'l', chr(196).chr(191) => 'L',
chr(197).chr(128) => 'l', chr(197).chr(129) => 'L',
chr(197).chr(130) => 'l', chr(197).chr(131) => 'N',
chr(197).chr(132) => 'n', chr(197).chr(133) => 'N',
chr(197).chr(134) => 'n', chr(197).chr(135) => 'N',
chr(197).chr(136) => 'n', chr(197).chr(137) => 'N',
chr(197).chr(138) => 'n', chr(197).chr(139) => 'N',
chr(197).chr(140) => 'O', chr(197).chr(141) => 'o',
chr(197).chr(142) => 'O', chr(197).chr(143) => 'o',
chr(197).chr(144) => 'O', chr(197).chr(145) => 'o',
chr(197).chr(146) => 'OE',chr(197).chr(147) => 'oe',
chr(197).chr(148) => 'R',chr(197).chr(149) => 'r',
chr(197).chr(150) => 'R',chr(197).chr(151) => 'r',
chr(197).chr(152) => 'R',chr(197).chr(153) => 'r',
chr(197).chr(154) => 'S',chr(197).chr(155) => 's',
chr(197).chr(156) => 'S',chr(197).chr(157) => 's',
chr(197).chr(158) => 'S',chr(197).chr(159) => 's',
chr(197).chr(160) => 'S', chr(197).chr(161) => 's',
chr(197).chr(162) => 'T', chr(197).chr(163) => 't',
chr(197).chr(164) => 'T', chr(197).chr(165) => 't',
chr(197).chr(166) => 'T', chr(197).chr(167) => 't',
chr(197).chr(168) => 'U', chr(197).chr(169) => 'u',
chr(197).chr(170) => 'U', chr(197).chr(171) => 'u',
chr(197).chr(172) => 'U', chr(197).chr(173) => 'u',
chr(197).chr(174) => 'U', chr(197).chr(175) => 'u',
chr(197).chr(176) => 'U', chr(197).chr(177) => 'u',
chr(197).chr(178) => 'U', chr(197).chr(179) => 'u',
chr(197).chr(180) => 'W', chr(197).chr(181) => 'w',
chr(197).chr(182) => 'Y', chr(197).chr(183) => 'y',
chr(197).chr(184) => 'Y', chr(197).chr(185) => 'Z',
chr(197).chr(186) => 'z', chr(197).chr(187) => 'Z',
chr(197).chr(188) => 'z', chr(197).chr(189) => 'Z',
chr(197).chr(190) => 'z', chr(197).chr(191) => 's',
// Euro Sign
chr(226).chr(130).chr(172) => 'E',
// GBP (Pound) Sign
chr(194).chr(163) => '');
$string = strtr($string, $chars);
} else {
// Assume ISO-8859-1 if not UTF-8
$chars['in'] = chr(128).chr(131).chr(138).chr(142).chr(154).chr(158)
.chr(159).chr(162).chr(165).chr(181).chr(192).chr(193).chr(194)
.chr(195).chr(196).chr(197).chr(199).chr(200).chr(201).chr(202)
.chr(203).chr(204).chr(205).chr(206).chr(207).chr(209).chr(210)
.chr(211).chr(212).chr(213).chr(214).chr(216).chr(217).chr(218)
.chr(219).chr(220).chr(221).chr(224).chr(225).chr(226).chr(227)
.chr(228).chr(229).chr(231).chr(232).chr(233).chr(234).chr(235)
.chr(236).chr(237).chr(238).chr(239).chr(241).chr(242).chr(243)
.chr(244).chr(245).chr(246).chr(248).chr(249).chr(250).chr(251)
.chr(252).chr(253).chr(255);
$chars['out'] = "EfSZszYcYuAAAAAACEEEEIIIINOOOOOOUUUUYaaaaaaceeeeiiiinoooooouuuuyy";
$string = strtr($string, $chars['in'], $chars['out']);
$double_chars['in'] = array(chr(140), chr(156), chr(198), chr(208), chr(222), chr(223), chr(230), chr(240), chr(254));
$double_chars['out'] = array('OE', 'oe', 'AE', 'DH', 'TH', 'ss', 'ae', 'dh', 'th');
$string = str_replace($double_chars['in'], $double_chars['out'], $string);
}
return $string;
}
function sanitize_file_name( $name ) { // Like sanitize_title, but with periods
$name = strtolower( $name );
$name = preg_replace('/&.+?;/', '', $name); // kill entities
$name = str_replace( '_', '-', $name );
$name = preg_replace('/[^a-z0-9\s-.]/', '', $name);
$name = preg_replace('/\s+/', '-', $name);
$name = preg_replace('|-+|', '-', $name);
$name = trim($name, '-');
return $name;
}
function sanitize_user( $username, $strict = false ) {
$raw_username = $username;
$username = strip_tags($username);
// Kill octets
$username = preg_replace('|%([a-fA-F0-9][a-fA-F0-9])|', '', $username);
$username = preg_replace('/&.+?;/', '', $username); // Kill entities
// If strict, reduce to ASCII for max portability.
if ( $strict )
$username = preg_replace('|[^a-z0-9 _.\-@]|i', '', $username);
return apply_filters('sanitize_user', $username, $raw_username, $strict);
}
function sanitize_title($title, $fallback_title = '') {
$title = strip_tags($title);
$title = apply_filters('sanitize_title', $title);
if (empty($title)) {
$title = $fallback_title;
}
return $title;
}
function sanitize_title_with_dashes($title) {
$title = strip_tags($title);
// Preserve escaped octets.
$title = preg_replace('|%([a-fA-F0-9][a-fA-F0-9])|', '---$1---', $title);
// Remove percent signs that are not part of an octet.
$title = str_replace('%', '', $title);
// Restore octets.
$title = preg_replace('|---([a-fA-F0-9][a-fA-F0-9])---|', '%$1', $title);
$title = remove_accents($title);
if (seems_utf8($title)) {
if (function_exists('mb_strtolower')) {
$title = mb_strtolower($title, 'UTF-8');
}
$title = utf8_uri_encode($title, 200);
}
$title = strtolower($title);
$title = preg_replace('/&.+?;/', '', $title); // kill entities
$title = preg_replace('/[^%a-z0-9 _-]/', '', $title);
$title = preg_replace('/\s+/', '-', $title);
$title = preg_replace('|-+|', '-', $title);
$title = trim($title, '-');
return $title;
}
function convert_chars($content, $flag = 'obsolete') {
// Translation of invalid Unicode references range to valid range
$wp_htmltranswinuni = array(
'' => '€', // the Euro sign
'' => '',
'' => '‚', // these are Windows CP1252 specific characters
'' => 'ƒ', // they would look weird on non-Windows browsers
'' => '„',
'
' => '…',
'' => '†',
'' => '‡',
'' => 'ˆ',
'' => '‰',
'' => 'Š',
'' => '‹',
'' => 'Œ',
'' => '',
'' => 'ž',
'' => '',
'' => '',
'' => '‘',
'' => '’',
'' => '“',
'' => '”',
'' => '•',
'' => '–',
'' => '—',
'' => '˜',
'' => '™',
'' => 'š',
'' => '›',
'' => 'œ',
'' => '',
'' => '',
'' => 'Ÿ'
);
// Remove metadata tags
$content = preg_replace('/(.+?)<\/title>/','',$content);
$content = preg_replace('/(.+?)<\/category>/','',$content);
// Converts lone & characters into & (a.k.a. &)
$content = preg_replace('/&([^#])(?![a-z1-4]{1,8};)/i', '&$1', $content);
// Fix Word pasting
$content = strtr($content, $wp_htmltranswinuni);
// Just a little XHTML help
$content = str_replace('
', '
', $content);
$content = str_replace('
', '
', $content);
return $content;
}
function funky_javascript_fix($text) {
// Fixes for browsers' javascript bugs
global $is_macIE, $is_winIE;
if ( $is_winIE || $is_macIE )
$text = preg_replace("/\%u([0-9A-F]{4,4})/e", "''.base_convert('\\1',16,10).';'", $text);
return $text;
}
/*
balanceTags
Balances Tags of string using a modified stack.
@param text Text to be balanced
@param force Forces balancing, ignoring the value of the option
@return Returns balanced text
@author Leonard Lin (leonard@acm.org)
@version v1.1
@date November 4, 2001
@license GPL v2.0
@notes
@changelog
--- Modified by Scott Reilly (coffee2code) 02 Aug 2004
1.2 ***TODO*** Make better - change loop condition to $text
1.1 Fixed handling of append/stack pop order of end text
Added Cleaning Hooks
1.0 First Version
*/
function balanceTags($text, $force = false) {
if ( !$force && get_option('use_balanceTags') == 0 )
return $text;
$tagstack = array(); $stacksize = 0; $tagqueue = ''; $newtext = '';
# WP bug fix for comments - in case you REALLY meant to type '< !--'
$text = str_replace('< !--', '< !--', $text);
# WP bug fix for LOVE <3 (and other situations with '<' before a number)
$text = preg_replace('#<([0-9]{1})#', '<$1', $text);
while (preg_match("/<(\/?\w*)\s*([^>]*)>/",$text,$regex)) {
$newtext .= $tagqueue;
$i = strpos($text,$regex[0]);
$l = strlen($regex[0]);
// clear the shifter
$tagqueue = '';
// Pop or Push
if ($regex[1][0] == "/") { // End Tag
$tag = strtolower(substr($regex[1],1));
// if too many closing tags
if($stacksize <= 0) {
$tag = '';
//or close to be safe $tag = '/' . $tag;
}
// if stacktop value = tag close value then pop
else if ($tagstack[$stacksize - 1] == $tag) { // found closing tag
$tag = '' . $tag . '>'; // Close Tag
// Pop
array_pop ($tagstack);
$stacksize--;
} else { // closing tag not at top, search for it
for ($j=$stacksize-1;$j>=0;$j--) {
if ($tagstack[$j] == $tag) {
// add tag to tagqueue
for ($k=$stacksize-1;$k>=$j;$k--){
$tagqueue .= '' . array_pop ($tagstack) . '>';
$stacksize--;
}
break;
}
}
$tag = '';
}
} else { // Begin Tag
$tag = strtolower($regex[1]);
// Tag Cleaning
// If self-closing or '', don't do anything.
if((substr($regex[2],-1) == '/') || ($tag == '')) {
}
// ElseIf it's a known single-entity tag but it doesn't close itself, do so
elseif ($tag == 'br' || $tag == 'img' || $tag == 'hr' || $tag == 'input') {
$regex[2] .= '/';
} else { // Push the tag onto the stack
// If the top of the stack is the same as the tag we want to push, close previous tag
if (($stacksize > 0) && ($tag != 'div') && ($tagstack[$stacksize - 1] == $tag)) {
$tagqueue = '' . array_pop ($tagstack) . '>';
$stacksize--;
}
$stacksize = array_push ($tagstack, $tag);
}
// Attributes
$attributes = $regex[2];
if($attributes) {
$attributes = ' '.$attributes;
}
$tag = '<'.$tag.$attributes.'>';
//If already queuing a close tag, then put this tag on, too
if ($tagqueue) {
$tagqueue .= $tag;
$tag = '';
}
}
$newtext .= substr($text,0,$i) . $tag;
$text = substr($text,$i+$l);
}
// Clear Tag Queue
$newtext .= $tagqueue;
// Add Remaining text
$newtext .= $text;
// Empty Stack
while($x = array_pop($tagstack)) {
$newtext .= '' . $x . '>'; // Add remaining tags to close
}
// WP fix for the bug with HTML comments
$newtext = str_replace("< !--","