inc/misc/utf8.php

   1 <?php
   2 /*
   3     This program is free software; you can redistribute it and/or modify
   4     it under the terms of the Revised BSD License.
   5
   6     This program is distributed in the hope that it will be useful,
   7     but WITHOUT ANY WARRANTY; without even the implied warranty of
   8     MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
   9     Revised BSD License for more details.
  10
  11     Copyright 2004-2019 iDB Support - https://idb.osdn.jp/support/category.php?act=view&id=1
  12     Copyright 2004-2019 Game Maker 2k - https://idb.osdn.jp/support/category.php?act=view&id=2
  13
  14     $FileInfo: utf8.php - Last Update: 08/02/2019 SVN 905 - Author: cooldude2k $
  15 */
  16 // UTF8 helper functions
  17 // author: Scott Michael Reynen "scott@randomchaos.com"
  18 // url: http://www.randomchaos.com/document.php?source=php_and_unicode
  19 // utf8_substr by frank at jkelloggs dot dk
  20 $File3Name = basename($_SERVER['SCRIPT_NAME']);
  21 if ($File3Name=="utf8.php"||$File3Name=="/utf8.php") {
  22         require('index.php');
  23         exit(); }
  24
  25 function utf8_strlen($str) {
  26   return strlen(utf8_decode($_GET['text'])); }
  27 function pre_strlen($str) {
  28 global $chkcharset;
  29 if($chkcharset=="UTF-8") {
  30 if(!defined('UTF8_NOMBSTRING')&&function_exists('mb_strlen')) {
  31 return mb_strlen($str,'utf-8'); }
  32 else { return utf8_strlen($str); } }
  33 if($chkcharset!="UTF-8") { return strlen($str); } }
  34
  35 // utf8_substr by frank at jkelloggs dot dk
  36 // http://us3.php.net/manual/en/function.substr.php#55107
  37 function utf8_substr($str,$start)
  38 {
  39    preg_match_all("/./su", $str, $ar);
  40    if(func_num_args() >= 3) {
  41        $end = func_get_arg(2);
  42        return join("",array_slice($ar[0],$start,$end));
  43    } else {
  44        return join("",array_slice($ar[0],$start));
  45    }
  46 }
  47
  48 function pre_substr($string,$start,$length) {
  49    global $chkcharset;
  50    if($chkcharset=="UTF-8") {
  51       if(!defined('UTF8_NOMBSTRING')&&function_exists('mb_substr')) {
  52       return mb_substr($string,$start,$length,'utf-8'); }
  53    else { return utf8_substr($string,$start,$length); } }
  54    if($chkcharset!="UTF-8") { return substr($string,$start,$length); } }
  55       if(isset($_GET['text'])) {
  56    echo pre_substr($_GET['text'],0,6); }
  57
  58 // author: Scott Michael Reynen "scott@randomchaos.com"
  59 // url: http://www.randomchaos.com/document.php?source=php_and_unicode
  60 function utf8_strpos($haystack, $needle,$offset=0) {
  61   if(!defined('UTF8_NOMBSTRING')&&function_exists('mb_strpos')) {
  62   return mb_strpos($haystack,$needle,$offset,'utf-8'); }
  63   $haystack = utf8_to_unicode($haystack);
  64   $needle   = utf8_to_unicode($needle);
  65   $position = $offset;
  66   $found = false;
  67   while( (! $found ) && ( $position < count( $haystack ) ) ) {
  68     if ( $needle[0] == $haystack[$position] ) {
  69       for ($i = 1; $i < count( $needle ); $i++ ) {
  70         if ( $needle[$i] != $haystack[ $position + $i ] ) break;
  71       } // for
  72       if ( $i == count( $needle ) ) {
  73         $found = true;
  74         $position--;
  75       } // if
  76     } // if
  77     $position++;
  78   } // while
  79   return ( $found == true ) ? $position : false;
  80 } // strpos_unicode
  81
  82 // author: Scott Michael Reynen "scott@randomchaos.com"
  83 // url: http://www.randomchaos.com/document.php?source=php_and_unicode
  84 function utf8_to_unicode( $str ) {
  85   $unicode = array();
  86   $values = array();
  87   $lookingFor = 1;
  88   for ($i = 0; $i < strlen( $str ); $i++ ) {
  89     $thisValue = ord( $str[ $i ] );
  90     if ( $thisValue < 128 ) $unicode[] = $thisValue;
  91     else {
  92       if ( count( $values ) == 0 ) $lookingFor = ( $thisValue < 224 ) ? 2 : 3;
  93       $values[] = $thisValue;
  94       if ( count( $values ) == $lookingFor ) {
  95   $number = ( $lookingFor == 3 ) ?
  96     ( ( $values[0] % 16 ) * 4096 ) + ( ( $values[1] % 64 ) * 64 ) + ( $values[2] % 64 ):
  97         ( ( $values[0] % 32 ) * 64 ) + ( $values[1] % 64 );
  98   $unicode[] = $number;
  99   $values = array();
 100   $lookingFor = 1;
 101       } // if
 102     } // if
 103   } // for
 104   return $unicode;
 105 } // utf8_to_unicode
 106
 107 // author: Scott Michael Reynen "scott@randomchaos.com"
 108 // url: http://www.randomchaos.com/document.php?source=php_and_unicode
 109 function unicode_to_utf8( $str ) {
 110   $utf8 = '';
 111   foreach( $str as $unicode ) {
 112     if ( $unicode < 128 ) {
 113       $utf8.= chr( $unicode );
 114     } elseif ( $unicode < 2048 ) {
 115       $utf8.= chr( 192 +  ( ( $unicode - ( $unicode % 64 ) ) / 64 ) );
 116       $utf8.= chr( 128 + ( $unicode % 64 ) );
 117     } else {
 118       $utf8.= chr( 224 + ( ( $unicode - ( $unicode % 4096 ) ) / 4096 ) );
 119       $utf8.= chr( 128 + ( ( ( $unicode % 4096 ) - ( $unicode % 64 ) ) / 64 ) );
 120       $utf8.= chr( 128 + ( $unicode % 64 ) );
 121     } // if
 122   } // foreach
 123   return $utf8;
 124 } // unicode_to_utf8
 125 ?>