1 <?php

     2 

     3 /**

     4 * @file

     5 * Provides Unicode-related conversions and operations.

     6 */

     7 

     8 /**

     9  * Indicates an error during check for PHP unicode support.

    10  */

    11 define('UNICODE_ERROR', -1);

    12 

    13 /**

    14  * Indicates that standard PHP (emulated) unicode support is being used.

    15  */

    16 define('UNICODE_SINGLEBYTE', 0);

    17 

    18 /**

    19  * Indicates that full unicode support with the PHP mbstring extension is being

    20  * used.

    21  */

    22 define('UNICODE_MULTIBYTE', 1);

    23 

    24 /**

    25  * Matches Unicode characters that are word boundaries.

    26  *

    27  * Characters with the following General_category (gc) property values are used

    28  * as word boundaries. While this does not fully conform to the Word Boundaries

    29  * algorithm described in http://unicode.org/reports/tr29, as PCRE does not

    30  * contain the Word_Break property table, this simpler algorithm has to do.

    31  * - Cc, Cf, Cn, Co, Cs: Other.

    32  * - Pc, Pd, Pe, Pf, Pi, Po, Ps: Punctuation.

    33  * - Sc, Sk, Sm, So: Symbols.

    34  * - Zl, Zp, Zs: Separators.

    35  *

    36  * Non-boundary characters include the following General_category (gc) property

    37  * values:

    38  * - Ll, Lm, Lo, Lt, Lu: Letters.

    39  * - Mc, Me, Mn: Combining Marks.

    40  * - Nd, Nl, No: Numbers.

    41  *

    42  * Note that the PCRE property matcher is not used because we wanted to be

    43  * compatible with Unicode 5.2.0 regardless of the PCRE version used (and any

    44  * bugs in PCRE property tables).

    45  *

    46  * @see http://unicode.org/glossary

    47  */

    48 define('PREG_CLASS_UNICODE_WORD_BOUNDARY',

    49   '\x{0}-\x{2F}\x{3A}-\x{40}\x{5B}-\x{60}\x{7B}-\x{A9}\x{AB}-\x{B1}\x{B4}' .

    50   '\x{B6}-\x{B8}\x{BB}\x{BF}\x{D7}\x{F7}\x{2C2}-\x{2C5}\x{2D2}-\x{2DF}' .

    51   '\x{2E5}-\x{2EB}\x{2ED}\x{2EF}-\x{2FF}\x{375}\x{37E}-\x{385}\x{387}\x{3F6}' .

    52   '\x{482}\x{55A}-\x{55F}\x{589}-\x{58A}\x{5BE}\x{5C0}\x{5C3}\x{5C6}' .

    53   '\x{5F3}-\x{60F}\x{61B}-\x{61F}\x{66A}-\x{66D}\x{6D4}\x{6DD}\x{6E9}' .

    54   '\x{6FD}-\x{6FE}\x{700}-\x{70F}\x{7F6}-\x{7F9}\x{830}-\x{83E}' .

    55   '\x{964}-\x{965}\x{970}\x{9F2}-\x{9F3}\x{9FA}-\x{9FB}\x{AF1}\x{B70}' .

    56   '\x{BF3}-\x{BFA}\x{C7F}\x{CF1}-\x{CF2}\x{D79}\x{DF4}\x{E3F}\x{E4F}' .

    57   '\x{E5A}-\x{E5B}\x{F01}-\x{F17}\x{F1A}-\x{F1F}\x{F34}\x{F36}\x{F38}' .

    58   '\x{F3A}-\x{F3D}\x{F85}\x{FBE}-\x{FC5}\x{FC7}-\x{FD8}\x{104A}-\x{104F}' .

    59   '\x{109E}-\x{109F}\x{10FB}\x{1360}-\x{1368}\x{1390}-\x{1399}\x{1400}' .

    60   '\x{166D}-\x{166E}\x{1680}\x{169B}-\x{169C}\x{16EB}-\x{16ED}' .

    61   '\x{1735}-\x{1736}\x{17B4}-\x{17B5}\x{17D4}-\x{17D6}\x{17D8}-\x{17DB}' .

    62   '\x{1800}-\x{180A}\x{180E}\x{1940}-\x{1945}\x{19DE}-\x{19FF}' .

    63   '\x{1A1E}-\x{1A1F}\x{1AA0}-\x{1AA6}\x{1AA8}-\x{1AAD}\x{1B5A}-\x{1B6A}' .

    64   '\x{1B74}-\x{1B7C}\x{1C3B}-\x{1C3F}\x{1C7E}-\x{1C7F}\x{1CD3}\x{1FBD}' .

    65   '\x{1FBF}-\x{1FC1}\x{1FCD}-\x{1FCF}\x{1FDD}-\x{1FDF}\x{1FED}-\x{1FEF}' .

    66   '\x{1FFD}-\x{206F}\x{207A}-\x{207E}\x{208A}-\x{208E}\x{20A0}-\x{20B8}' .

    67   '\x{2100}-\x{2101}\x{2103}-\x{2106}\x{2108}-\x{2109}\x{2114}' .

    68   '\x{2116}-\x{2118}\x{211E}-\x{2123}\x{2125}\x{2127}\x{2129}\x{212E}' .

    69   '\x{213A}-\x{213B}\x{2140}-\x{2144}\x{214A}-\x{214D}\x{214F}' .

    70   '\x{2190}-\x{244A}\x{249C}-\x{24E9}\x{2500}-\x{2775}\x{2794}-\x{2B59}' .

    71   '\x{2CE5}-\x{2CEA}\x{2CF9}-\x{2CFC}\x{2CFE}-\x{2CFF}\x{2E00}-\x{2E2E}' .

    72   '\x{2E30}-\x{3004}\x{3008}-\x{3020}\x{3030}\x{3036}-\x{3037}' .

    73   '\x{303D}-\x{303F}\x{309B}-\x{309C}\x{30A0}\x{30FB}\x{3190}-\x{3191}' .

    74   '\x{3196}-\x{319F}\x{31C0}-\x{31E3}\x{3200}-\x{321E}\x{322A}-\x{3250}' .

    75   '\x{3260}-\x{327F}\x{328A}-\x{32B0}\x{32C0}-\x{33FF}\x{4DC0}-\x{4DFF}' .

    76   '\x{A490}-\x{A4C6}\x{A4FE}-\x{A4FF}\x{A60D}-\x{A60F}\x{A673}\x{A67E}' .

    77   '\x{A6F2}-\x{A716}\x{A720}-\x{A721}\x{A789}-\x{A78A}\x{A828}-\x{A82B}' .

    78   '\x{A836}-\x{A839}\x{A874}-\x{A877}\x{A8CE}-\x{A8CF}\x{A8F8}-\x{A8FA}' .

    79   '\x{A92E}-\x{A92F}\x{A95F}\x{A9C1}-\x{A9CD}\x{A9DE}-\x{A9DF}' .

    80   '\x{AA5C}-\x{AA5F}\x{AA77}-\x{AA79}\x{AADE}-\x{AADF}\x{ABEB}' .

    81   '\x{E000}-\x{F8FF}\x{FB29}\x{FD3E}-\x{FD3F}\x{FDFC}-\x{FDFD}' .

    82   '\x{FE10}-\x{FE19}\x{FE30}-\x{FE6B}\x{FEFF}-\x{FF0F}\x{FF1A}-\x{FF20}' .

    83   '\x{FF3B}-\x{FF40}\x{FF5B}-\x{FF65}\x{FFE0}-\x{FFFD}');

    84 

    85 /**

    86  * Wrapper around _unicode_check().

    87  */

    88 function unicode_check() {

    89   list($GLOBALS['multibyte']) = _unicode_check();

    90 }

    91 

    92 /**

    93  * Perform checks about Unicode support in PHP, and set the right settings if

    94  * needed.

    95  *

    96  * Because Drupal needs to be able to handle text in various encodings, we do

    97  * not support mbstring function overloading. HTTP input/output conversion must

    98  * be disabled for similar reasons.

    99  *

   100  * @param $errors

   101  *   Whether to report any fatal errors with form_set_error().

   102  */

   103 function _unicode_check() {

   104   // Ensure translations don't break during installation.

   105   $t = get_t();

   106 

   107   // Check for mbstring extension

   108   if (!function_exists('mb_strlen')) {

   109     return array(UNICODE_SINGLEBYTE, $t('Operations on Unicode strings are emulated on a best-effort basis. Install the <a href="@url">PHP mbstring extension</a> for improved Unicode support.', array('@url' => 'http://www.php.net/mbstring')));

   110   }

   111 

   112   // Check mbstring configuration

   113   if (ini_get('mbstring.func_overload') != 0) {

   114     return array(UNICODE_ERROR, $t('Multibyte string function overloading in PHP is active and must be disabled. Check the php.ini <em>mbstring.func_overload</em> setting. Please refer to the <a href="@url">PHP mbstring documentation</a> for more information.', array('@url' => 'http://www.php.net/mbstring')));

   115   }

   116   if (ini_get('mbstring.encoding_translation') != 0) {

   117     return array(UNICODE_ERROR, $t('Multibyte string input conversion in PHP is active and must be disabled. Check the php.ini <em>mbstring.encoding_translation</em> setting. Please refer to the <a href="@url">PHP mbstring documentation</a> for more information.', array('@url' => 'http://www.php.net/mbstring')));

   118   }

   119   // mbstring.http_input and mbstring.http_output are deprecated and empty by

   120   // default in PHP 5.6.

   121   if (version_compare(PHP_VERSION, '5.6.0') == -1) {

   122     if (ini_get('mbstring.http_input') != 'pass') {

   123       return array(UNICODE_ERROR, $t('Multibyte string input conversion in PHP is active and must be disabled. Check the php.ini <em>mbstring.http_input</em> setting. Please refer to the <a href="@url">PHP mbstring documentation</a> for more information.', array('@url' => 'http://www.php.net/mbstring')));

   124     }

   125     if (ini_get('mbstring.http_output') != 'pass') {

   126       return array(UNICODE_ERROR, $t('Multibyte string output conversion in PHP is active and must be disabled. Check the php.ini <em>mbstring.http_output</em> setting. Please refer to the <a href="@url">PHP mbstring documentation</a> for more information.', array('@url' => 'http://www.php.net/mbstring')));

   127     }

   128   }

   129 

   130   // Set appropriate configuration

   131   mb_internal_encoding('utf-8');

   132   mb_language('uni');

   133   return array(UNICODE_MULTIBYTE, '');

   134 }

   135 

   136 /**

   137  * Returns Unicode library status and errors.

   138  */

   139 function unicode_requirements() {

   140   // Ensure translations don't break during installation.

   141   $t = get_t();

   142 

   143   $libraries = array(

   144     UNICODE_SINGLEBYTE => $t('Standard PHP'),

   145     UNICODE_MULTIBYTE => $t('PHP Mbstring Extension'),

   146     UNICODE_ERROR => $t('Error'),

   147   );

   148   $severities = array(

   149     UNICODE_SINGLEBYTE => REQUIREMENT_WARNING,

   150     UNICODE_MULTIBYTE => REQUIREMENT_OK,

   151     UNICODE_ERROR => REQUIREMENT_ERROR,

   152   );

   153   list($library, $description) = _unicode_check();

   154 

   155   $requirements['unicode'] = array(

   156     'title' => $t('Unicode library'),

   157     'value' => $libraries[$library],

   158   );

   159   if ($description) {

   160     $requirements['unicode']['description'] = $description;

   161   }

   162 

   163   $requirements['unicode']['severity'] = $severities[$library];

   164 

   165   return $requirements;

   166 }

   167 

   168 /**

   169  * Prepares a new XML parser.

   170  *

   171  * This is a wrapper around xml_parser_create() which extracts the encoding

   172  * from the XML data first and sets the output encoding to UTF-8. This function

   173  * should be used instead of xml_parser_create(), because PHP 4's XML parser

   174  * doesn't check the input encoding itself. "Starting from PHP 5, the input

   175  * encoding is automatically detected, so that the encoding parameter specifies

   176  * only the output encoding."

   177  *

   178  * This is also where unsupported encodings will be converted. Callers should

   179  * take this into account: $data might have been changed after the call.

   180  *

   181  * @param $data

   182  *   The XML data which will be parsed later.

   183  *

   184  * @return

   185  *   An XML parser object or FALSE on error.

   186  *

   187  * @ingroup php_wrappers

   188  */

   189 function drupal_xml_parser_create(&$data) {

   190   // Default XML encoding is UTF-8

   191   $encoding = 'utf-8';

   192   $bom = FALSE;

   193 

   194   // Check for UTF-8 byte order mark (PHP5's XML parser doesn't handle it).

   195   if (!strncmp($data, "\xEF\xBB\xBF", 3)) {

   196     $bom = TRUE;

   197     $data = substr($data, 3);

   198   }

   199 

   200   // Check for an encoding declaration in the XML prolog if no BOM was found.

   201   if (!$bom && preg_match('/^<\?xml[^>]+encoding="(.+?)"/', $data, $match)) {

   202     $encoding = $match[1];

   203   }

   204 

   205   // Unsupported encodings are converted here into UTF-8.

   206   $php_supported = array('utf-8', 'iso-8859-1', 'us-ascii');

   207   if (!in_array(strtolower($encoding), $php_supported)) {

   208     $out = drupal_convert_to_utf8($data, $encoding);

   209     if ($out !== FALSE) {

   210       $encoding = 'utf-8';

   211       $data = preg_replace('/^(<\?xml[^>]+encoding)="(.+?)"/', '\\1="utf-8"', $out);

   212     }

   213     else {

   214       watchdog('php', 'Could not convert XML encoding %s to UTF-8.', array('%s' => $encoding), WATCHDOG_WARNING);

   215       return FALSE;

   216     }

   217   }

   218 

   219   $xml_parser = xml_parser_create($encoding);

   220   xml_parser_set_option($xml_parser, XML_OPTION_TARGET_ENCODING, 'utf-8');

   221   return $xml_parser;

   222 }

   223 

   224 /**

   225  * Converts data to UTF-8.

   226  *

   227  * Requires the iconv, GNU recode or mbstring PHP extension.

   228  *

   229  * @param $data

   230  *   The data to be converted.

   231  * @param $encoding

   232  *   The encoding that the data is in.

   233  *

   234  * @return

   235  *   Converted data or FALSE.

   236  */

   237 function drupal_convert_to_utf8($data, $encoding) {

   238   if (function_exists('iconv')) {

   239     $out = @iconv($encoding, 'utf-8', $data);

   240   }

   241   elseif (function_exists('mb_convert_encoding')) {

   242     $out = @mb_convert_encoding($data, 'utf-8', $encoding);

   243   }

   244   elseif (function_exists('recode_string')) {

   245     $out = @recode_string($encoding . '..utf-8', $data);

   246   }

   247   else {

   248     watchdog('php', 'Unsupported encoding %s. Please install iconv, GNU recode or mbstring for PHP.', array('%s' => $encoding), WATCHDOG_ERROR);

   249     return FALSE;

   250   }

   251 

   252   return $out;

   253 }

   254 

   255 /**

   256  * Truncates a UTF-8-encoded string safely to a number of bytes.

   257  *

   258  * If the end position is in the middle of a UTF-8 sequence, it scans backwards

   259  * until the beginning of the byte sequence.

   260  *

   261  * Use this function whenever you want to chop off a string at an unsure

   262  * location. On the other hand, if you're sure that you're splitting on a

   263  * character boundary (e.g. after using strpos() or similar), you can safely

   264  * use substr() instead.

   265  *

   266  * @param $string

   267  *   The string to truncate.

   268  * @param $len

   269  *   An upper limit on the returned string length.

   270  *

   271  * @return

   272  *   The truncated string.

   273  */

   274 function drupal_truncate_bytes($string, $len) {

   275   if (strlen($string) <= $len) {

   276     return $string;

   277   }

   278   if ((ord($string[$len]) < 0x80) || (ord($string[$len]) >= 0xC0)) {

   279     return substr($string, 0, $len);

   280   }

   281   // Scan backwards to beginning of the byte sequence.

   282   while (--$len >= 0 && ord($string[$len]) >= 0x80 && ord($string[$len]) < 0xC0);

   283 

   284   return substr($string, 0, $len);

   285 }

   286 

   287 /**

   288  * Truncates a UTF-8-encoded string safely to a number of characters.

   289  *

   290  * @param $string

   291  *   The string to truncate.

   292  * @param $max_length

   293  *   An upper limit on the returned string length, including trailing ellipsis

   294  *   if $add_ellipsis is TRUE.

   295  * @param $wordsafe

   296  *   If TRUE, attempt to truncate on a word boundary. Word boundaries are

   297  *   spaces, punctuation, and Unicode characters used as word boundaries in

   298  *   non-Latin languages; see PREG_CLASS_UNICODE_WORD_BOUNDARY for more

   299  *   information. If a word boundary cannot be found that would make the length

   300  *   of the returned string fall within length guidelines (see parameters

   301  *   $max_length and $min_wordsafe_length), word boundaries are ignored.

   302  * @param $add_ellipsis

   303  *   If TRUE, add t('...') to the end of the truncated string (defaults to

   304  *   FALSE). The string length will still fall within $max_length.

   305  * @param $min_wordsafe_length

   306  *   If $wordsafe is TRUE, the minimum acceptable length for truncation (before

   307  *   adding an ellipsis, if $add_ellipsis is TRUE). Has no effect if $wordsafe

   308  *   is FALSE. This can be used to prevent having a very short resulting string

   309  *   that will not be understandable. For instance, if you are truncating the

   310  *   string "See myverylongurlexample.com for more information" to a word-safe

   311  *   return length of 20, the only available word boundary within 20 characters

   312  *   is after the word "See", which wouldn't leave a very informative string. If

   313  *   you had set $min_wordsafe_length to 10, though, the function would realise

   314  *   that "See" alone is too short, and would then just truncate ignoring word

   315  *   boundaries, giving you "See myverylongurl..." (assuming you had set

   316  *   $add_ellipses to TRUE).

   317  *

   318  * @return string

   319  *   The truncated string.

   320  */

   321 function truncate_utf8($string, $max_length, $wordsafe = FALSE, $add_ellipsis = FALSE, $min_wordsafe_length = 1) {

   322   $ellipsis = '';

   323   $max_length = max($max_length, 0);

   324   $min_wordsafe_length = max($min_wordsafe_length, 0);

   325 

   326   if (drupal_strlen($string) <= $max_length) {

   327     // No truncation needed, so don't add ellipsis, just return.

   328     return $string;

   329   }

   330 

   331   if ($add_ellipsis) {

   332     // Truncate ellipsis in case $max_length is small.

   333     $ellipsis = drupal_substr(t('...'), 0, $max_length);

   334     $max_length -= drupal_strlen($ellipsis);

   335     $max_length = max($max_length, 0);

   336   }

   337 

   338   if ($max_length <= $min_wordsafe_length) {

   339     // Do not attempt word-safe if lengths are bad.

   340     $wordsafe = FALSE;

   341   }

   342 

   343   if ($wordsafe) {

   344     $matches = array();

   345     // Find the last word boundary, if there is one within $min_wordsafe_length

   346     // to $max_length characters. preg_match() is always greedy, so it will

   347     // find the longest string possible.

   348     $found = preg_match('/^(.{' . $min_wordsafe_length . ',' . $max_length . '})[' . PREG_CLASS_UNICODE_WORD_BOUNDARY . ']/u', $string, $matches);

   349     if ($found) {

   350       $string = $matches[1];

   351     }

   352     else {

   353       $string = drupal_substr($string, 0, $max_length);

   354     }

   355   }

   356   else {

   357     $string = drupal_substr($string, 0, $max_length);

   358   }

   359 

   360   if ($add_ellipsis) {

   361     $string .= $ellipsis;

   362   }

   363 

   364   return $string;

   365 }

   366 

   367 /**

   368  * Encodes MIME/HTTP header values that contain incorrectly encoded characters.

   369  *

   370  * For example, mime_header_encode('tést.txt') returns "=?UTF-8?B?dMOpc3QudHh0?=".

   371  *

   372  * See http://www.rfc-editor.org/rfc/rfc2047.txt for more information.

   373  *

   374  * Notes:

   375  * - Only encode strings that contain non-ASCII characters.

   376  * - We progressively cut-off a chunk with truncate_utf8(). This is to ensure

   377  *   each chunk starts and ends on a character boundary.

   378  * - Using \n as the chunk separator may cause problems on some systems and may

   379  *   have to be changed to \r\n or \r.

   380  *

   381  * @param $string

   382  *   The header to encode.

   383  *

   384  * @return string

   385  *   The mime-encoded header.

   386  *

   387  * @see mime_header_decode()

   388  */

   389 function mime_header_encode($string) {

   390   if (preg_match('/[^\x20-\x7E]/', $string)) {

   391     $chunk_size = 47; // floor((75 - strlen("=?UTF-8?B??=")) * 0.75);

   392     $len = strlen($string);

   393     $output = '';

   394     while ($len > 0) {

   395       $chunk = drupal_truncate_bytes($string, $chunk_size);

   396       $output .= ' =?UTF-8?B?' . base64_encode($chunk) . "?=\n";

   397       $c = strlen($chunk);

   398       $string = substr($string, $c);

   399       $len -= $c;

   400     }

   401     return trim($output);

   402   }

   403   return $string;

   404 }

   405 

   406 /**

   407  * Decodes MIME/HTTP encoded header values.

   408  *

   409  * @param $header

   410  *   The header to decode.

   411  *

   412  * @return string

   413  *   The mime-decoded header.

   414  *

   415  * @see mime_header_encode()

   416  */

   417 function mime_header_decode($header) {

   418   // First step: encoded chunks followed by other encoded chunks (need to collapse whitespace)

   419   $header = preg_replace_callback('/=\?([^?]+)\?(Q|B)\?([^?]+|\?(?!=))\?=\s+(?==\?)/', '_mime_header_decode', $header);

   420   // Second step: remaining chunks (do not collapse whitespace)

   421   return preg_replace_callback('/=\?([^?]+)\?(Q|B)\?([^?]+|\?(?!=))\?=/', '_mime_header_decode', $header);

   422 }

   423 

   424 /**

   425  * Decodes encoded header data passed from mime_header_decode().

   426  *

   427  * Callback for preg_replace_callback() within mime_header_decode().

   428  *

   429  * @param $matches

   430  *   The array of matches from preg_replace_callback().

   431  *

   432  * @return string

   433  *   The mime-decoded string.

   434  *

   435  * @see mime_header_decode()

   436  */

   437 function _mime_header_decode($matches) {

   438   // Regexp groups:

   439   // 1: Character set name

   440   // 2: Escaping method (Q or B)

   441   // 3: Encoded data

   442   $data = ($matches[2] == 'B') ? base64_decode($matches[3]) : str_replace('_', ' ', quoted_printable_decode($matches[3]));

   443   if (strtolower($matches[1]) != 'utf-8') {

   444     $data = drupal_convert_to_utf8($data, $matches[1]);

   445   }

   446   return $data;

   447 }

   448 

   449 /**

   450  * Decodes all HTML entities (including numerical ones) to regular UTF-8 bytes.

   451  *

   452  * Double-escaped entities will only be decoded once ("<" becomes "<"

   453  * , not "<"). Be careful when using this function, as decode_entities can

   454  * revert previous sanitization efforts (&lt;script&gt; will become <script>).

   455  *

   456  * @param $text

   457  *   The text to decode entities in.

   458  *

   459  * @return

   460  *   The input $text, with all HTML entities decoded once.

   461  */

   462 function decode_entities($text) {

   463   return html_entity_decode($text, ENT_QUOTES, 'UTF-8');

   464 }

   465 

   466 /**

   467  * Counts the number of characters in a UTF-8 string.

   468  *

   469  * This is less than or equal to the byte count.

   470  *

   471  * @param $text

   472  *   The string to run the operation on.

   473  *

   474  * @return integer

   475  *   The length of the string.

   476  *

   477  * @ingroup php_wrappers

   478  */

   479 function drupal_strlen($text) {

   480   global $multibyte;

   481   if ($multibyte == UNICODE_MULTIBYTE) {

   482     return mb_strlen($text);

   483   }

   484   else {

   485     // Do not count UTF-8 continuation bytes.

   486     return strlen(preg_replace("/[\x80-\xBF]/", '', $text));

   487   }

   488 }

   489 

   490 /**

   491  * Uppercase a UTF-8 string.

   492  *

   493  * @param $text

   494  *   The string to run the operation on.

   495  *

   496  * @return string

   497  *   The string in uppercase.

   498  *

   499  * @ingroup php_wrappers

   500  */

   501 function drupal_strtoupper($text) {

   502   global $multibyte;

   503   if ($multibyte == UNICODE_MULTIBYTE) {

   504     return mb_strtoupper($text);

   505   }

   506   else {

   507     // Use C-locale for ASCII-only uppercase

   508     $text = strtoupper($text);

   509     // Case flip Latin-1 accented letters

   510     $text = preg_replace_callback('/\xC3[\xA0-\xB6\xB8-\xBE]/', '_unicode_caseflip', $text);

   511     return $text;

   512   }

   513 }

   514 

   515 /**

   516  * Lowercase a UTF-8 string.

   517  *

   518  * @param $text

   519  *   The string to run the operation on.

   520  *

   521  * @return string

   522  *   The string in lowercase.

   523  *

   524  * @ingroup php_wrappers

   525  */

   526 function drupal_strtolower($text) {

   527   global $multibyte;

   528   if ($multibyte == UNICODE_MULTIBYTE) {

   529     return mb_strtolower($text);

   530   }

   531   else {

   532     // Use C-locale for ASCII-only lowercase

   533     $text = strtolower($text);

   534     // Case flip Latin-1 accented letters

   535     $text = preg_replace_callback('/\xC3[\x80-\x96\x98-\x9E]/', '_unicode_caseflip', $text);

   536     return $text;

   537   }

   538 }

   539 

   540 /**

   541  * Flips U+C0-U+DE to U+E0-U+FD and back.

   542  *

   543  * @param $matches

   544  *   An array of matches.

   545  *

   546  * @return array

   547  *   The Latin-1 version of the array of matches.

   548  *

   549  * @see drupal_strtolower()

   550  */

   551 function _unicode_caseflip($matches) {

   552   return $matches[0][0] . chr(ord($matches[0][1]) ^ 32);

   553 }

   554 

   555 /**

   556  * Capitalizes the first letter of a UTF-8 string.

   557  *

   558  * @param $text

   559  *   The string to convert.

   560  *

   561  * @return

   562  *   The string with the first letter as uppercase.

   563  *

   564  * @ingroup php_wrappers

   565  */

   566 function drupal_ucfirst($text) {

   567   // Note: no mbstring equivalent!

   568   return drupal_strtoupper(drupal_substr($text, 0, 1)) . drupal_substr($text, 1);

   569 }

   570 

   571 /**

   572  * Cuts off a piece of a string based on character indices and counts.

   573  *

   574  * Follows the same behavior as PHP's own substr() function. Note that for

   575  * cutting off a string at a known character/substring location, the usage of

   576  * PHP's normal strpos/substr is safe and much faster.

   577  *

   578  * @param $text

   579  *   The input string.

   580  * @param $start

   581  *   The position at which to start reading.

   582  * @param $length

   583  *   The number of characters to read.

   584  *

   585  * @return

   586  *   The shortened string.

   587  *

   588  * @ingroup php_wrappers

   589  */

   590 function drupal_substr($text, $start, $length = NULL) {

   591   global $multibyte;

   592   if ($multibyte == UNICODE_MULTIBYTE) {

   593     return $length === NULL ? mb_substr($text, $start) : mb_substr($text, $start, $length);

   594   }

   595   else {

   596     $strlen = strlen($text);

   597     // Find the starting byte offset.

   598     $bytes = 0;

   599     if ($start > 0) {

   600       // Count all the continuation bytes from the start until we have found

   601       // $start characters or the end of the string.

   602       $bytes = -1; $chars = -1;

   603       while ($bytes < $strlen - 1 && $chars < $start) {

   604         $bytes++;

   605         $c = ord($text[$bytes]);

   606         if ($c < 0x80 || $c >= 0xC0) {

   607           $chars++;

   608         }

   609       }

   610     }

   611     elseif ($start < 0) {

   612       // Count all the continuation bytes from the end until we have found

   613       // abs($start) characters.

   614       $start = abs($start);

   615       $bytes = $strlen; $chars = 0;

   616       while ($bytes > 0 && $chars < $start) {

   617         $bytes--;

   618         $c = ord($text[$bytes]);

   619         if ($c < 0x80 || $c >= 0xC0) {

   620           $chars++;

   621         }

   622       }

   623     }

   624     $istart = $bytes;

   625 

   626     // Find the ending byte offset.

   627     if ($length === NULL) {

   628       $iend = $strlen;

   629     }

   630     elseif ($length > 0) {

   631       // Count all the continuation bytes from the starting index until we have

   632       // found $length characters or reached the end of the string, then

   633       // backtrace one byte.

   634       $iend = $istart - 1;

   635       $chars = -1;

   636       $last_real = FALSE;

   637       while ($iend < $strlen - 1 && $chars < $length) {

   638         $iend++;

   639         $c = ord($text[$iend]);

   640         $last_real = FALSE;

   641         if ($c < 0x80 || $c >= 0xC0) {

   642           $chars++;

   643           $last_real = TRUE;

   644         }

   645       }

   646       // Backtrace one byte if the last character we found was a real character

   647       // and we don't need it.

   648       if ($last_real && $chars >= $length) {

   649         $iend--;

   650       }

   651     }

   652     elseif ($length < 0) {

   653       // Count all the continuation bytes from the end until we have found

   654       // abs($start) characters, then backtrace one byte.

   655       $length = abs($length);

   656       $iend = $strlen; $chars = 0;

   657       while ($iend > 0 && $chars < $length) {

   658         $iend--;

   659         $c = ord($text[$iend]);

   660         if ($c < 0x80 || $c >= 0xC0) {

   661           $chars++;

   662         }

   663       }

   664       // Backtrace one byte if we are not at the beginning of the string.

   665       if ($iend > 0) {

   666         $iend--;

   667       }

   668     }

   669     else {

   670       // $length == 0, return an empty string.

   671       return '';

   672     }

   673 

   674     return substr($text, $istart, max(0, $iend - $istart + 1));

   675   }

   676 }