[Summary view]

Differences Between: [Versions 310 and 311] [Versions 311 and 400] [Versions 311 and 401] [Versions 311 and 402] [Versions 311 and 403] [Versions 39 and 311]

   1  <?php
   2  
   3  namespace PhpOffice\PhpSpreadsheet\Shared;
   4  
   5  use PhpOffice\PhpSpreadsheet\Exception as PhpSpreadsheetException;
   6  
   7  class CodePage
   8  {
   9      public const DEFAULT_CODE_PAGE = 'CP1252';
  10  
  11      private static $pageArray = [
  12          0 => 'CP1252', //    CodePage is not always correctly set when the xls file was saved by Apple's Numbers program
  13          367 => 'ASCII', //    ASCII
  14          437 => 'CP437', //    OEM US
  15          //720 => 'notsupported', //    OEM Arabic
  16          737 => 'CP737', //    OEM Greek
  17          775 => 'CP775', //    OEM Baltic
  18          850 => 'CP850', //    OEM Latin I
  19          852 => 'CP852', //    OEM Latin II (Central European)
  20          855 => 'CP855', //    OEM Cyrillic
  21          857 => 'CP857', //    OEM Turkish
  22          858 => 'CP858', //    OEM Multilingual Latin I with Euro
  23          860 => 'CP860', //    OEM Portugese
  24          861 => 'CP861', //    OEM Icelandic
  25          862 => 'CP862', //    OEM Hebrew
  26          863 => 'CP863', //    OEM Canadian (French)
  27          864 => 'CP864', //    OEM Arabic
  28          865 => 'CP865', //    OEM Nordic
  29          866 => 'CP866', //    OEM Cyrillic (Russian)
  30          869 => 'CP869', //    OEM Greek (Modern)
  31          874 => 'CP874', //    ANSI Thai
  32          932 => 'CP932', //    ANSI Japanese Shift-JIS
  33          936 => 'CP936', //    ANSI Chinese Simplified GBK
  34          949 => 'CP949', //    ANSI Korean (Wansung)
  35          950 => 'CP950', //    ANSI Chinese Traditional BIG5
  36          1200 => 'UTF-16LE', //    UTF-16 (BIFF8)
  37          1250 => 'CP1250', //    ANSI Latin II (Central European)
  38          1251 => 'CP1251', //    ANSI Cyrillic
  39          1252 => 'CP1252', //    ANSI Latin I (BIFF4-BIFF7)
  40          1253 => 'CP1253', //    ANSI Greek
  41          1254 => 'CP1254', //    ANSI Turkish
  42          1255 => 'CP1255', //    ANSI Hebrew
  43          1256 => 'CP1256', //    ANSI Arabic
  44          1257 => 'CP1257', //    ANSI Baltic
  45          1258 => 'CP1258', //    ANSI Vietnamese
  46          1361 => 'CP1361', //    ANSI Korean (Johab)
  47          10000 => 'MAC', //    Apple Roman
  48          10001 => 'CP932', //    Macintosh Japanese
  49          10002 => 'CP950', //    Macintosh Chinese Traditional
  50          10003 => 'CP1361', //    Macintosh Korean
  51          10004 => 'MACARABIC', //    Apple Arabic
  52          10005 => 'MACHEBREW', //    Apple Hebrew
  53          10006 => 'MACGREEK', //    Macintosh Greek
  54          10007 => 'MACCYRILLIC', //    Macintosh Cyrillic
  55          10008 => 'CP936', //    Macintosh - Simplified Chinese (GB 2312)
  56          10010 => 'MACROMANIA', //    Macintosh Romania
  57          10017 => 'MACUKRAINE', //    Macintosh Ukraine
  58          10021 => 'MACTHAI', //    Macintosh Thai
  59          10029 => 'MACCENTRALEUROPE', //    Macintosh Central Europe
  60          10079 => 'MACICELAND', //    Macintosh Icelandic
  61          10081 => 'MACTURKISH', //    Macintosh Turkish
  62          10082 => 'MACCROATIAN', //    Macintosh Croatian
  63          21010 => 'UTF-16LE', //    UTF-16 (BIFF8) This isn't correct, but some Excel writer libraries erroneously use Codepage 21010 for UTF-16LE
  64          32768 => 'MAC', //    Apple Roman
  65          //32769 => 'unsupported', //    ANSI Latin I (BIFF2-BIFF3)
  66          65000 => 'UTF-7', //    Unicode (UTF-7)
  67          65001 => 'UTF-8', //    Unicode (UTF-8)
  68      ];
  69  
  70      public static function validate(string $codePage): bool
  71      {
  72          return in_array($codePage, self::$pageArray, true);
  73      }
  74  
  75      /**
  76       * Convert Microsoft Code Page Identifier to Code Page Name which iconv
  77       * and mbstring understands.
  78       *
  79       * @param int $codePage Microsoft Code Page Indentifier
  80       *
  81       * @return string Code Page Name
  82       */
  83      public static function numberToName(int $codePage): string
  84      {
  85          if (array_key_exists($codePage, self::$pageArray)) {
  86              return self::$pageArray[$codePage];
  87          }
  88          if ($codePage == 720 || $codePage == 32769) {
  89              throw new PhpSpreadsheetException("Code page $codePage not supported."); //    OEM Arabic
  90          }
  91  
  92          throw new PhpSpreadsheetException('Unknown codepage: ' . $codePage);
  93      }
  94  
  95      public static function getEncodings(): array
  96      {
  97          return self::$pageArray;
  98      }
  99  }