3 namespace Drupal\Component\Utility;
6 * Provides Unicode-related conversions and operations.
13 * Matches Unicode characters that are word boundaries.
15 * Characters with the following General_category (gc) property values are used
16 * as word boundaries. While this does not fully conform to the Word Boundaries
17 * algorithm described in http://unicode.org/reports/tr29, as PCRE does not
18 * contain the Word_Break property table, this simpler algorithm has to do.
19 * - Cc, Cf, Cn, Co, Cs: Other.
20 * - Pc, Pd, Pe, Pf, Pi, Po, Ps: Punctuation.
21 * - Sc, Sk, Sm, So: Symbols.
22 * - Zl, Zp, Zs: Separators.
24 * Non-boundary characters include the following General_category (gc) property
26 * - Ll, Lm, Lo, Lt, Lu: Letters.
27 * - Mc, Me, Mn: Combining Marks.
28 * - Nd, Nl, No: Numbers.
30 * Note that the PCRE property matcher is not used because we wanted to be
31 * compatible with Unicode 5.2.0 regardless of the PCRE version used (and any
32 * bugs in PCRE property tables).
34 * @see http://unicode.org/glossary
36 const PREG_CLASS_WORD_BOUNDARY = <<<'EOD'
37 \x{0}-\x{2F}\x{3A}-\x{40}\x{5B}-\x{60}\x{7B}-\x{A9}\x{AB}-\x{B1}\x{B4}
38 \x{B6}-\x{B8}\x{BB}\x{BF}\x{D7}\x{F7}\x{2C2}-\x{2C5}\x{2D2}-\x{2DF}
39 \x{2E5}-\x{2EB}\x{2ED}\x{2EF}-\x{2FF}\x{375}\x{37E}-\x{385}\x{387}\x{3F6}
40 \x{482}\x{55A}-\x{55F}\x{589}-\x{58A}\x{5BE}\x{5C0}\x{5C3}\x{5C6}
41 \x{5F3}-\x{60F}\x{61B}-\x{61F}\x{66A}-\x{66D}\x{6D4}\x{6DD}\x{6E9}
42 \x{6FD}-\x{6FE}\x{700}-\x{70F}\x{7F6}-\x{7F9}\x{830}-\x{83E}
43 \x{964}-\x{965}\x{970}\x{9F2}-\x{9F3}\x{9FA}-\x{9FB}\x{AF1}\x{B70}
44 \x{BF3}-\x{BFA}\x{C7F}\x{CF1}-\x{CF2}\x{D79}\x{DF4}\x{E3F}\x{E4F}
45 \x{E5A}-\x{E5B}\x{F01}-\x{F17}\x{F1A}-\x{F1F}\x{F34}\x{F36}\x{F38}
46 \x{F3A}-\x{F3D}\x{F85}\x{FBE}-\x{FC5}\x{FC7}-\x{FD8}\x{104A}-\x{104F}
47 \x{109E}-\x{109F}\x{10FB}\x{1360}-\x{1368}\x{1390}-\x{1399}\x{1400}
48 \x{166D}-\x{166E}\x{1680}\x{169B}-\x{169C}\x{16EB}-\x{16ED}
49 \x{1735}-\x{1736}\x{17B4}-\x{17B5}\x{17D4}-\x{17D6}\x{17D8}-\x{17DB}
50 \x{1800}-\x{180A}\x{180E}\x{1940}-\x{1945}\x{19DE}-\x{19FF}
51 \x{1A1E}-\x{1A1F}\x{1AA0}-\x{1AA6}\x{1AA8}-\x{1AAD}\x{1B5A}-\x{1B6A}
52 \x{1B74}-\x{1B7C}\x{1C3B}-\x{1C3F}\x{1C7E}-\x{1C7F}\x{1CD3}\x{1FBD}
53 \x{1FBF}-\x{1FC1}\x{1FCD}-\x{1FCF}\x{1FDD}-\x{1FDF}\x{1FED}-\x{1FEF}
54 \x{1FFD}-\x{206F}\x{207A}-\x{207E}\x{208A}-\x{208E}\x{20A0}-\x{20B8}
55 \x{2100}-\x{2101}\x{2103}-\x{2106}\x{2108}-\x{2109}\x{2114}
56 \x{2116}-\x{2118}\x{211E}-\x{2123}\x{2125}\x{2127}\x{2129}\x{212E}
57 \x{213A}-\x{213B}\x{2140}-\x{2144}\x{214A}-\x{214D}\x{214F}
58 \x{2190}-\x{244A}\x{249C}-\x{24E9}\x{2500}-\x{2775}\x{2794}-\x{2B59}
59 \x{2CE5}-\x{2CEA}\x{2CF9}-\x{2CFC}\x{2CFE}-\x{2CFF}\x{2E00}-\x{2E2E}
60 \x{2E30}-\x{3004}\x{3008}-\x{3020}\x{3030}\x{3036}-\x{3037}
61 \x{303D}-\x{303F}\x{309B}-\x{309C}\x{30A0}\x{30FB}\x{3190}-\x{3191}
62 \x{3196}-\x{319F}\x{31C0}-\x{31E3}\x{3200}-\x{321E}\x{322A}-\x{3250}
63 \x{3260}-\x{327F}\x{328A}-\x{32B0}\x{32C0}-\x{33FF}\x{4DC0}-\x{4DFF}
64 \x{A490}-\x{A4C6}\x{A4FE}-\x{A4FF}\x{A60D}-\x{A60F}\x{A673}\x{A67E}
65 \x{A6F2}-\x{A716}\x{A720}-\x{A721}\x{A789}-\x{A78A}\x{A828}-\x{A82B}
66 \x{A836}-\x{A839}\x{A874}-\x{A877}\x{A8CE}-\x{A8CF}\x{A8F8}-\x{A8FA}
67 \x{A92E}-\x{A92F}\x{A95F}\x{A9C1}-\x{A9CD}\x{A9DE}-\x{A9DF}
68 \x{AA5C}-\x{AA5F}\x{AA77}-\x{AA79}\x{AADE}-\x{AADF}\x{ABEB}
69 \x{E000}-\x{F8FF}\x{FB29}\x{FD3E}-\x{FD3F}\x{FDFC}-\x{FDFD}
70 \x{FE10}-\x{FE19}\x{FE30}-\x{FE6B}\x{FEFF}-\x{FF0F}\x{FF1A}-\x{FF20}
71 \x{FF3B}-\x{FF40}\x{FF5B}-\x{FF65}\x{FFE0}-\x{FFFD}
75 * Indicates that standard PHP (emulated) unicode support is being used.
77 const STATUS_SINGLEBYTE = 0;
80 * Indicates that full unicode support with the PHP mbstring extension is
83 const STATUS_MULTIBYTE = 1;
86 * Indicates an error during check for PHP unicode support.
88 const STATUS_ERROR = -1;
91 * Holds the multibyte capabilities of the current environment.
95 protected static $status = 0;
98 * Gets the current status of unicode/multibyte support on this environment.
101 * The status of multibyte support. It can be one of:
102 * - \Drupal\Component\Utility\Unicode::STATUS_MULTIBYTE
103 * Full unicode support using an extension.
104 * - \Drupal\Component\Utility\Unicode::STATUS_SINGLEBYTE
105 * Standard PHP (emulated) unicode support.
106 * - \Drupal\Component\Utility\Unicode::STATUS_ERROR
107 * An error occurred. No unicode support.
109 public static function getStatus() {
110 return static::$status;
114 * Sets the value for multibyte support status for the current environment.
116 * The following status keys are supported:
117 * - \Drupal\Component\Utility\Unicode::STATUS_MULTIBYTE
118 * Full unicode support using an extension.
119 * - \Drupal\Component\Utility\Unicode::STATUS_SINGLEBYTE
120 * Standard PHP (emulated) unicode support.
121 * - \Drupal\Component\Utility\Unicode::STATUS_ERROR
122 * An error occurred. No unicode support.
125 * The new status of multibyte support.
127 public static function setStatus($status) {
128 if (!in_array($status, [static::STATUS_SINGLEBYTE, static::STATUS_MULTIBYTE, static::STATUS_ERROR])) {
129 throw new \InvalidArgumentException('Invalid status value for unicode support.');
131 static::$status = $status;
135 * Checks for Unicode support in PHP and sets the proper settings if possible.
137 * Because of the need to be able to handle text in various encodings, we do
138 * not support mbstring function overloading. HTTP input/output conversion
139 * must be disabled for similar reasons.
142 * A string identifier of a failed multibyte extension check, if any.
143 * Otherwise, an empty string.
145 public static function check() {
146 // Check for mbstring extension.
147 if (!function_exists('mb_strlen')) {
148 static::$status = static::STATUS_SINGLEBYTE;
152 // Check mbstring configuration.
153 if (ini_get('mbstring.func_overload') != 0) {
154 static::$status = static::STATUS_ERROR;
155 return 'mbstring.func_overload';
157 if (ini_get('mbstring.encoding_translation') != 0) {
158 static::$status = static::STATUS_ERROR;
159 return 'mbstring.encoding_translation';
161 // mbstring.http_input and mbstring.http_output are deprecated and empty by
162 // default in PHP 5.6.
163 if (version_compare(PHP_VERSION, '5.6.0') == -1) {
164 if (ini_get('mbstring.http_input') != 'pass') {
165 static::$status = static::STATUS_ERROR;
166 return 'mbstring.http_input';
168 if (ini_get('mbstring.http_output') != 'pass') {
169 static::$status = static::STATUS_ERROR;
170 return 'mbstring.http_output';
174 // Set appropriate configuration.
175 mb_internal_encoding('utf-8');
177 static::$status = static::STATUS_MULTIBYTE;
182 * Decodes UTF byte-order mark (BOM) into the encoding's name.
184 * @param string $data
185 * The data possibly containing a BOM. This can be the entire contents of
186 * a file, or just a fragment containing at least the first five bytes.
188 * @return string|bool
189 * The name of the encoding, or FALSE if no byte order mark was present.
191 public static function encodingFromBOM($data) {
193 "\xEF\xBB\xBF" => 'UTF-8',
194 "\xFE\xFF" => 'UTF-16BE',
195 "\xFF\xFE" => 'UTF-16LE',
196 "\x00\x00\xFE\xFF" => 'UTF-32BE',
197 "\xFF\xFE\x00\x00" => 'UTF-32LE',
198 "\x2B\x2F\x76\x38" => 'UTF-7',
199 "\x2B\x2F\x76\x39" => 'UTF-7',
200 "\x2B\x2F\x76\x2B" => 'UTF-7',
201 "\x2B\x2F\x76\x2F" => 'UTF-7',
202 "\x2B\x2F\x76\x38\x2D" => 'UTF-7',
205 foreach ($bomMap as $bom => $encoding) {
206 if (strpos($data, $bom) === 0) {
214 * Converts data to UTF-8.
216 * Requires the iconv, GNU recode or mbstring PHP extension.
218 * @param string $data
219 * The data to be converted.
220 * @param string $encoding
221 * The encoding that the data is in.
223 * @return string|bool
224 * Converted data or FALSE.
226 public static function convertToUtf8($data, $encoding) {
227 if (function_exists('iconv')) {
228 return @iconv($encoding, 'utf-8', $data);
230 elseif (function_exists('mb_convert_encoding')) {
231 return @mb_convert_encoding($data, 'utf-8', $encoding);
233 elseif (function_exists('recode_string')) {
234 return @recode_string($encoding . '..utf-8', $data);
241 * Truncates a UTF-8-encoded string safely to a number of bytes.
243 * If the end position is in the middle of a UTF-8 sequence, it scans backwards
244 * until the beginning of the byte sequence.
246 * Use this function whenever you want to chop off a string at an unsure
247 * location. On the other hand, if you're sure that you're splitting on a
248 * character boundary (e.g. after using strpos() or similar), you can safely
249 * use substr() instead.
251 * @param string $string
252 * The string to truncate.
254 * An upper limit on the returned string length.
257 * The truncated string.
259 public static function truncateBytes($string, $len) {
260 if (strlen($string) <= $len) {
263 if ((ord($string[$len]) < 0x80) || (ord($string[$len]) >= 0xC0)) {
264 return substr($string, 0, $len);
266 // Scan backwards to beginning of the byte sequence.
267 while (--$len >= 0 && ord($string[$len]) >= 0x80 && ord($string[$len]) < 0xC0);
269 return substr($string, 0, $len);
273 * Counts the number of characters in a UTF-8 string.
275 * This is less than or equal to the byte count.
277 * @param string $text
278 * The string to run the operation on.
281 * The length of the string.
283 public static function strlen($text) {
284 if (static::getStatus() == static::STATUS_MULTIBYTE) {
285 return mb_strlen($text);
288 // Do not count UTF-8 continuation bytes.
289 return strlen(preg_replace("/[\x80-\xBF]/", '', $text));
294 * Converts a UTF-8 string to uppercase.
296 * @param string $text
297 * The string to run the operation on.
300 * The string in uppercase.
302 public static function strtoupper($text) {
303 if (static::getStatus() == static::STATUS_MULTIBYTE) {
304 return mb_strtoupper($text);
307 // Use C-locale for ASCII-only uppercase.
308 $text = strtoupper($text);
309 // Case flip Latin-1 accented letters.
310 $text = preg_replace_callback('/\xC3[\xA0-\xB6\xB8-\xBE]/', '\Drupal\Component\Utility\Unicode::caseFlip', $text);
316 * Converts a UTF-8 string to lowercase.
318 * @param string $text
319 * The string to run the operation on.
322 * The string in lowercase.
324 public static function strtolower($text) {
325 if (static::getStatus() == static::STATUS_MULTIBYTE) {
326 return mb_strtolower($text);
329 // Use C-locale for ASCII-only lowercase.
330 $text = strtolower($text);
331 // Case flip Latin-1 accented letters.
332 $text = preg_replace_callback('/\xC3[\x80-\x96\x98-\x9E]/', '\Drupal\Component\Utility\Unicode::caseFlip', $text);
338 * Capitalizes the first character of a UTF-8 string.
340 * @param string $text
341 * The string to convert.
344 * The string with the first character as uppercase.
346 public static function ucfirst($text) {
347 return static::strtoupper(static::substr($text, 0, 1)) . static::substr($text, 1);
351 * Converts the first character of a UTF-8 string to lowercase.
353 * @param string $text
354 * The string that will be converted.
357 * The string with the first character as lowercase.
359 * @ingroup php_wrappers
361 public static function lcfirst($text) {
362 // Note: no mbstring equivalent!
363 return static::strtolower(static::substr($text, 0, 1)) . static::substr($text, 1);
367 * Capitalizes the first character of each word in a UTF-8 string.
369 * @param string $text
370 * The text that will be converted.
373 * The input $text with each word capitalized.
375 * @ingroup php_wrappers
377 public static function ucwords($text) {
378 $regex = '/(^|[' . static::PREG_CLASS_WORD_BOUNDARY . '])([^' . static::PREG_CLASS_WORD_BOUNDARY . '])/u';
379 return preg_replace_callback($regex, function(array $matches) {
380 return $matches[1] . Unicode::strtoupper($matches[2]);
385 * Cuts off a piece of a string based on character indices and counts.
387 * Follows the same behavior as PHP's own substr() function. Note that for
388 * cutting off a string at a known character/substring location, the usage of
389 * PHP's normal strpos/substr is safe and much faster.
391 * @param string $text
394 * The position at which to start reading.
396 * The number of characters to read.
399 * The shortened string.
401 public static function substr($text, $start, $length = NULL) {
402 if (static::getStatus() == static::STATUS_MULTIBYTE) {
403 return $length === NULL ? mb_substr($text, $start) : mb_substr($text, $start, $length);
406 $strlen = strlen($text);
407 // Find the starting byte offset.
410 // Count all the characters except continuation bytes from the start
411 // until we have found $start characters or the end of the string.
412 $bytes = -1; $chars = -1;
413 while ($bytes < $strlen - 1 && $chars < $start) {
415 $c = ord($text[$bytes]);
416 if ($c < 0x80 || $c >= 0xC0) {
421 elseif ($start < 0) {
422 // Count all the characters except continuation bytes from the end
423 // until we have found abs($start) characters.
424 $start = abs($start);
425 $bytes = $strlen; $chars = 0;
426 while ($bytes > 0 && $chars < $start) {
428 $c = ord($text[$bytes]);
429 if ($c < 0x80 || $c >= 0xC0) {
436 // Find the ending byte offset.
437 if ($length === NULL) {
440 elseif ($length > 0) {
441 // Count all the characters except continuation bytes from the starting
442 // index until we have found $length characters or reached the end of
443 // the string, then backtrace one byte.
447 while ($iend < $strlen - 1 && $chars < $length) {
449 $c = ord($text[$iend]);
451 if ($c < 0x80 || $c >= 0xC0) {
456 // Backtrace one byte if the last character we found was a real
457 // character and we don't need it.
458 if ($last_real && $chars >= $length) {
462 elseif ($length < 0) {
463 // Count all the characters except continuation bytes from the end
464 // until we have found abs($start) characters, then backtrace one byte.
465 $length = abs($length);
466 $iend = $strlen; $chars = 0;
467 while ($iend > 0 && $chars < $length) {
469 $c = ord($text[$iend]);
470 if ($c < 0x80 || $c >= 0xC0) {
474 // Backtrace one byte if we are not at the beginning of the string.
480 // $length == 0, return an empty string.
484 return substr($text, $istart, max(0, $iend - $istart + 1));
489 * Truncates a UTF-8-encoded string safely to a number of characters.
491 * @param string $string
492 * The string to truncate.
493 * @param int $max_length
494 * An upper limit on the returned string length, including trailing ellipsis
495 * if $add_ellipsis is TRUE.
496 * @param bool $wordsafe
497 * If TRUE, attempt to truncate on a word boundary. Word boundaries are
498 * spaces, punctuation, and Unicode characters used as word boundaries in
499 * non-Latin languages; see Unicode::PREG_CLASS_WORD_BOUNDARY for more
500 * information. If a word boundary cannot be found that would make the length
501 * of the returned string fall within length guidelines (see parameters
502 * $max_length and $min_wordsafe_length), word boundaries are ignored.
503 * @param bool $add_ellipsis
504 * If TRUE, add '...' to the end of the truncated string (defaults to
505 * FALSE). The string length will still fall within $max_length.
506 * @param int $min_wordsafe_length
507 * If $wordsafe is TRUE, the minimum acceptable length for truncation (before
508 * adding an ellipsis, if $add_ellipsis is TRUE). Has no effect if $wordsafe
509 * is FALSE. This can be used to prevent having a very short resulting string
510 * that will not be understandable. For instance, if you are truncating the
511 * string "See myverylongurlexample.com for more information" to a word-safe
512 * return length of 20, the only available word boundary within 20 characters
513 * is after the word "See", which wouldn't leave a very informative string. If
514 * you had set $min_wordsafe_length to 10, though, the function would realise
515 * that "See" alone is too short, and would then just truncate ignoring word
516 * boundaries, giving you "See myverylongurl..." (assuming you had set
517 * $add_ellipses to TRUE).
520 * The truncated string.
522 public static function truncate($string, $max_length, $wordsafe = FALSE, $add_ellipsis = FALSE, $min_wordsafe_length = 1) {
524 $max_length = max($max_length, 0);
525 $min_wordsafe_length = max($min_wordsafe_length, 0);
527 if (static::strlen($string) <= $max_length) {
528 // No truncation needed, so don't add ellipsis, just return.
533 // Truncate ellipsis in case $max_length is small.
534 $ellipsis = static::substr('…', 0, $max_length);
535 $max_length -= static::strlen($ellipsis);
536 $max_length = max($max_length, 0);
539 if ($max_length <= $min_wordsafe_length) {
540 // Do not attempt word-safe if lengths are bad.
546 // Find the last word boundary, if there is one within $min_wordsafe_length
547 // to $max_length characters. preg_match() is always greedy, so it will
548 // find the longest string possible.
549 $found = preg_match('/^(.{' . $min_wordsafe_length . ',' . $max_length . '})[' . Unicode::PREG_CLASS_WORD_BOUNDARY . ']/u', $string, $matches);
551 $string = $matches[1];
554 $string = static::substr($string, 0, $max_length);
558 $string = static::substr($string, 0, $max_length);
562 // If we're adding an ellipsis, remove any trailing periods.
563 $string = rtrim($string, '.');
565 $string .= $ellipsis;
572 * Compares UTF-8-encoded strings in a binary safe case-insensitive manner.
574 * @param string $str1
576 * @param string $str2
580 * Returns < 0 if $str1 is less than $str2; > 0 if $str1 is greater than
581 * $str2, and 0 if they are equal.
583 public static function strcasecmp($str1 , $str2) {
584 return strcmp(static::strtoupper($str1), static::strtoupper($str2));
588 * Encodes MIME/HTTP headers that contain incorrectly encoded characters.
590 * For example, Unicode::mimeHeaderEncode('tést.txt') returns
591 * "=?UTF-8?B?dMOpc3QudHh0?=".
593 * See http://www.rfc-editor.org/rfc/rfc2047.txt for more information.
596 * - Only encode strings that contain non-ASCII characters.
597 * - We progressively cut-off a chunk with self::truncateBytes(). This ensures
598 * each chunk starts and ends on a character boundary.
599 * - Using \n as the chunk separator may cause problems on some systems and
600 * may have to be changed to \r\n or \r.
602 * @param string $string
603 * The header to encode.
606 * The mime-encoded header.
608 public static function mimeHeaderEncode($string) {
609 if (preg_match('/[^\x20-\x7E]/', $string)) {
610 $chunk_size = 47; // floor((75 - strlen("=?UTF-8?B??=")) * 0.75);
611 $len = strlen($string);
614 $chunk = static::truncateBytes($string, $chunk_size);
615 $output .= ' =?UTF-8?B?' . base64_encode($chunk) . "?=\n";
617 $string = substr($string, $c);
620 return trim($output);
626 * Decodes MIME/HTTP encoded header values.
628 * @param string $header
629 * The header to decode.
632 * The mime-decoded header.
634 public static function mimeHeaderDecode($header) {
635 $callback = function ($matches) {
636 $data = ($matches[2] == 'B') ? base64_decode($matches[3]) : str_replace('_', ' ', quoted_printable_decode($matches[3]));
637 if (strtolower($matches[1]) != 'utf-8') {
638 $data = static::convertToUtf8($data, $matches[1]);
642 // First step: encoded chunks followed by other encoded chunks (need to collapse whitespace)
643 $header = preg_replace_callback('/=\?([^?]+)\?(Q|B)\?([^?]+|\?(?!=))\?=\s+(?==\?)/', $callback, $header);
644 // Second step: remaining chunks (do not collapse whitespace)
645 return preg_replace_callback('/=\?([^?]+)\?(Q|B)\?([^?]+|\?(?!=))\?=/', $callback, $header);
649 * Flip U+C0-U+DE to U+E0-U+FD and back. Can be used as preg_replace callback.
651 * @param array $matches
652 * An array of matches by preg_replace_callback().
657 public static function caseFlip($matches) {
658 return $matches[0][0] . chr(ord($matches[0][1]) ^ 32);
662 * Checks whether a string is valid UTF-8.
664 * All functions designed to filter input should use drupal_validate_utf8
665 * to ensure they operate on valid UTF-8 strings to prevent bypass of the
668 * When text containing an invalid UTF-8 lead byte (0xC0 - 0xFF) is presented
669 * as UTF-8 to Internet Explorer 6, the program may misinterpret subsequent
670 * bytes. When these subsequent bytes are HTML control characters such as
671 * quotes or angle brackets, parts of the text that were deemed safe by filters
672 * end up in locations that are potentially unsafe; An onerror attribute that
673 * is outside of a tag, and thus deemed safe by a filter, can be interpreted
674 * by the browser as if it were inside the tag.
676 * The function does not return FALSE for strings containing character codes
677 * above U+10FFFF, even though these are prohibited by RFC 3629.
679 * @param string $text
683 * TRUE if the text is valid UTF-8, FALSE if not.
685 public static function validateUtf8($text) {
686 if (strlen($text) == 0) {
689 // With the PCRE_UTF8 modifier 'u', preg_match() fails silently on strings
690 // containing invalid UTF-8 byte sequences. It does not reject character
691 // codes above U+10FFFF (represented by 4 or more octets), though.
692 return (preg_match('/^./us', $text) == 1);
696 * Finds the position of the first occurrence of a string in another string.
698 * @param string $haystack
699 * The string to search in.
700 * @param string $needle
701 * The string to find in $haystack.
703 * If specified, start the search at this number of characters from the
704 * beginning (default 0).
707 * The position where $needle occurs in $haystack, always relative to the
708 * beginning (independent of $offset), or FALSE if not found. Note that
709 * a return value of 0 is not the same as FALSE.
711 public static function strpos($haystack, $needle, $offset = 0) {
712 if (static::getStatus() == static::STATUS_MULTIBYTE) {
713 return mb_strpos($haystack, $needle, $offset);
716 // Remove Unicode continuation characters, to be compatible with
717 // Unicode::strlen() and Unicode::substr().
718 $haystack = preg_replace("/[\x80-\xBF]/", '', $haystack);
719 $needle = preg_replace("/[\x80-\xBF]/", '', $needle);
720 return strpos($haystack, $needle, $offset);