diff options
| author | David M <davidmj@users.sourceforge.net> | 2006-10-09 17:44:06 +0000 |
|---|---|---|
| committer | David M <davidmj@users.sourceforge.net> | 2006-10-09 17:44:06 +0000 |
| commit | 42dcd7929e7433ecdac023baa870c0bba62cf728 (patch) | |
| tree | 02f9c935589e3d551089a134c91b8648dbd63299 /phpBB/includes/utf/utf_tools.php | |
| parent | 0b2aaae28bb341a38864eda320e135826a0568e4 (diff) | |
| download | forums-42dcd7929e7433ecdac023baa870c0bba62cf728.tar forums-42dcd7929e7433ecdac023baa870c0bba62cf728.tar.gz forums-42dcd7929e7433ecdac023baa870c0bba62cf728.tar.bz2 forums-42dcd7929e7433ecdac023baa870c0bba62cf728.tar.xz forums-42dcd7929e7433ecdac023baa870c0bba62cf728.zip | |
hopefully no bugs...
- faster encoding and way faster decoding (hopefully)
- waaaaay faster strtoupper strtolower transforms
- more powerful utf8_substr
- removed two (now) useless functions
P.S.
decode and upper/lower all had bugs before :P
git-svn-id: file:///svn/phpbb/trunk@6471 89ea8834-ac86-4346-8a33-228a782c2dd0
Diffstat (limited to 'phpBB/includes/utf/utf_tools.php')
| -rw-r--r-- | phpBB/includes/utf/utf_tools.php | 538 |
1 files changed, 246 insertions, 292 deletions
diff --git a/phpBB/includes/utf/utf_tools.php b/phpBB/includes/utf/utf_tools.php index 8f88cc391f..1bcd92e75f 100644 --- a/phpBB/includes/utf/utf_tools.php +++ b/phpBB/includes/utf/utf_tools.php @@ -50,12 +50,11 @@ if (!extension_loaded('xml')) } else if ($num < 0xC0) { - $out .= "\xC2$letter"; + $out .= "\xC2" . $letter; } else { - $chr = chr($num - 64); - $out .= "\xC3$chr"; + $out .= "\xC3" . chr($num - 64); } } return $out; @@ -64,63 +63,41 @@ if (!extension_loaded('xml')) /** * Implementation of PHP's native utf8_decode for people without XML support * - * @author GetID3() * @param string $string UTF-8 encoded data * @return string ISO-8859-1 encoded data */ - function utf8_decode($string) + function utf8_decode($str) { - $newcharstring = ''; - $offset = 0; - $stringlength = strlen($string); - - while ($offset < $stringlength) + $pos = 0; + $len = strlen($str); + $ret = ''; + + while ($pos < $len) { - $ord = ord($string{$offset}); - if (($ord | 0x07) == 0xF7) - { - // 11110bbb 10bbbbbb 10bbbbbb 10bbbbbb - $charval = (($ord & 0x07) << 18) & - ((ord($string{($offset + 1)}) & 0x3F) << 12) & - ((ord($string{($offset + 2)}) & 0x3F) << 6) & - (ord($string{($offset + 3)}) & 0x3F); - $offset += 4; - } - else if (($ord | 0x0F) == 0xEF) + $ord = ord($str[$pos]) & 0xF0; + if ($ord === 0xC0 || $ord === 0xD0) { - // 1110bbbb 10bbbbbb 10bbbbbb - $charval = (($ord & 0x0F) << 12) & - ((ord($string{($offset + 1)}) & 0x3F) << 6) & - (ord($string{($offset + 2)}) & 0x3F); - $offset += 3; + $charval = ((ord($str[$pos]) & 0x1F) << 6) | (ord($str[$pos + 1]) & 0x3F); + $pos += 2; + $ret .= (($charval < 256) ? chr($charval) : '?'); } - else if (($ord | 0x1F) == 0xDF) + else if ($ord === 0xE0) { - // 110bbbbb 10bbbbbb - $charval = ((ord($string{($offset + 0)}) & 0x1F) << 6) & - (ord($string{($offset + 1)}) & 0x3F); - $offset += 2; + $ret .= '?'; + $pos += 3; } - else if (($ord | 0x7F) == 0x7F) + else if ($ord === 0xF0) { - // 0bbbbbbb - $charval = $ord; - $offset += 1; + $ret .= '?'; + $pos += 4; } else { - // error? throw some kind of warning here? - $charval = false; - $offset += 1; - } - - if ($charval !== false) - { - $newcharstring .= (($charval < 256) ? chr($charval) : '?'); + $ret .= $str[$pos]; + ++$pos; } } - - return $newcharstring; + return $ret; } } @@ -308,7 +285,6 @@ else */ function utf8_strpos($str, $needle, $offset = null) { - // native if (is_null($offset)) { $ar = explode($needle, $str); @@ -338,95 +314,113 @@ else } $UTF8_UPPER_TO_LOWER = array( - 0x0041=>0x0061, 0x03A6=>0x03C6, 0x0162=>0x0163, 0x00C5=>0x00E5, 0x0042=>0x0062, - 0x0139=>0x013A, 0x00C1=>0x00E1, 0x0141=>0x0142, 0x038E=>0x03CD, 0x0100=>0x0101, - 0x0490=>0x0491, 0x0394=>0x03B4, 0x015A=>0x015B, 0x0044=>0x0064, 0x0393=>0x03B3, - 0x00D4=>0x00F4, 0x042A=>0x044A, 0x0419=>0x0439, 0x0112=>0x0113, 0x041C=>0x043C, - 0x015E=>0x015F, 0x0143=>0x0144, 0x00CE=>0x00EE, 0x040E=>0x045E, 0x042F=>0x044F, - 0x039A=>0x03BA, 0x0154=>0x0155, 0x0049=>0x0069, 0x0053=>0x0073, 0x1E1E=>0x1E1F, - 0x0134=>0x0135, 0x0427=>0x0447, 0x03A0=>0x03C0, 0x0418=>0x0438, 0x00D3=>0x00F3, - 0x0420=>0x0440, 0x0404=>0x0454, 0x0415=>0x0435, 0x0429=>0x0449, 0x014A=>0x014B, - 0x0411=>0x0431, 0x0409=>0x0459, 0x1E02=>0x1E03, 0x00D6=>0x00F6, 0x00D9=>0x00F9, - 0x004E=>0x006E, 0x0401=>0x0451, 0x03A4=>0x03C4, 0x0423=>0x0443, 0x015C=>0x015D, - 0x0403=>0x0453, 0x03A8=>0x03C8, 0x0158=>0x0159, 0x0047=>0x0067, 0x00C4=>0x00E4, - 0x0386=>0x03AC, 0x0389=>0x03AE, 0x0166=>0x0167, 0x039E=>0x03BE, 0x0164=>0x0165, - 0x0116=>0x0117, 0x0108=>0x0109, 0x0056=>0x0076, 0x00DE=>0x00FE, 0x0156=>0x0157, - 0x00DA=>0x00FA, 0x1E60=>0x1E61, 0x1E82=>0x1E83, 0x00C2=>0x00E2, 0x0118=>0x0119, - 0x0145=>0x0146, 0x0050=>0x0070, 0x0150=>0x0151, 0x042E=>0x044E, 0x0128=>0x0129, - 0x03A7=>0x03C7, 0x013D=>0x013E, 0x0422=>0x0442, 0x005A=>0x007A, 0x0428=>0x0448, - 0x03A1=>0x03C1, 0x1E80=>0x1E81, 0x016C=>0x016D, 0x00D5=>0x00F5, 0x0055=>0x0075, - 0x0176=>0x0177, 0x00DC=>0x00FC, 0x1E56=>0x1E57, 0x03A3=>0x03C3, 0x041A=>0x043A, - 0x004D=>0x006D, 0x016A=>0x016B, 0x0170=>0x0171, 0x0424=>0x0444, 0x00CC=>0x00EC, - 0x0168=>0x0169, 0x039F=>0x03BF, 0x004B=>0x006B, 0x00D2=>0x00F2, 0x00C0=>0x00E0, - 0x0414=>0x0434, 0x03A9=>0x03C9, 0x1E6A=>0x1E6B, 0x00C3=>0x00E3, 0x042D=>0x044D, - 0x0416=>0x0436, 0x01A0=>0x01A1, 0x010C=>0x010D, 0x011C=>0x011D, 0x00D0=>0x00F0, - 0x013B=>0x013C, 0x040F=>0x045F, 0x040A=>0x045A, 0x00C8=>0x00E8, 0x03A5=>0x03C5, - 0x0046=>0x0066, 0x00DD=>0x00FD, 0x0043=>0x0063, 0x021A=>0x021B, 0x00CA=>0x00EA, - 0x0399=>0x03B9, 0x0179=>0x017A, 0x00CF=>0x00EF, 0x01AF=>0x01B0, 0x0045=>0x0065, - 0x039B=>0x03BB, 0x0398=>0x03B8, 0x039C=>0x03BC, 0x040C=>0x045C, 0x041F=>0x043F, - 0x042C=>0x044C, 0x00DE=>0x00FE, 0x00D0=>0x00F0, 0x1EF2=>0x1EF3, 0x0048=>0x0068, - 0x00CB=>0x00EB, 0x0110=>0x0111, 0x0413=>0x0433, 0x012E=>0x012F, 0x00C6=>0x00E6, - 0x0058=>0x0078, 0x0160=>0x0161, 0x016E=>0x016F, 0x0391=>0x03B1, 0x0407=>0x0457, - 0x0172=>0x0173, 0x0178=>0x00FF, 0x004F=>0x006F, 0x041B=>0x043B, 0x0395=>0x03B5, - 0x0425=>0x0445, 0x0120=>0x0121, 0x017D=>0x017E, 0x017B=>0x017C, 0x0396=>0x03B6, - 0x0392=>0x03B2, 0x0388=>0x03AD, 0x1E84=>0x1E85, 0x0174=>0x0175, 0x0051=>0x0071, - 0x0417=>0x0437, 0x1E0A=>0x1E0B, 0x0147=>0x0148, 0x0104=>0x0105, 0x0408=>0x0458, - 0x014C=>0x014D, 0x00CD=>0x00ED, 0x0059=>0x0079, 0x010A=>0x010B, 0x038F=>0x03CE, - 0x0052=>0x0072, 0x0410=>0x0430, 0x0405=>0x0455, 0x0402=>0x0452, 0x0126=>0x0127, - 0x0136=>0x0137, 0x012A=>0x012B, 0x038A=>0x03AF, 0x042B=>0x044B, 0x004C=>0x006C, - 0x0397=>0x03B7, 0x0124=>0x0125, 0x0218=>0x0219, 0x00DB=>0x00FB, 0x011E=>0x011F, - 0x041E=>0x043E, 0x1E40=>0x1E41, 0x039D=>0x03BD, 0x0106=>0x0107, 0x03AB=>0x03CB, - 0x0426=>0x0446, 0x00DE=>0x00FE, 0x00C7=>0x00E7, 0x03AA=>0x03CA, 0x0421=>0x0441, - 0x0412=>0x0432, 0x010E=>0x010F, 0x00D8=>0x00F8, 0x0057=>0x0077, 0x011A=>0x011B, - 0x0054=>0x0074, 0x004A=>0x006A, 0x040B=>0x045B, 0x0406=>0x0456, 0x0102=>0x0103, - 0x039B=>0x03BB, 0x00D1=>0x00F1, 0x041D=>0x043D, 0x038C=>0x03CC, 0x00C9=>0x00E9, - 0x00D0=>0x00F0, 0x0407=>0x0457, 0x0122=>0x0123, + "\x41" => "\x61", "\x42" => "\x62", "\x43" => "\x63", "\x44" => "\x64", + "\x45" => "\x65", "\x46" => "\x66", "\x47" => "\x67", "\x48" => "\x68", + "\x49" => "\x69", "\x4A" => "\x6A", "\x4B" => "\x6B", "\x4C" => "\x6C", + "\x4D" => "\x6D", "\x4E" => "\x6E", "\x4F" => "\x6F", "\x50" => "\x70", + "\x51" => "\x71", "\x52" => "\x72", "\x53" => "\x73", "\x54" => "\x74", + "\x55" => "\x75", "\x56" => "\x76", "\x57" => "\x77", "\x58" => "\x78", + "\x59" => "\x79", "\x5A" => "\x7A", "\xC3\x80" => "\xC3\xA0", "\xC3\x81" => "\xC3\xA1", + "\xC3\x82" => "\xC3\xA2", "\xC3\x83" => "\xC3\xA3", "\xC3\x84" => "\xC3\xA4", "\xC3\x85" => "\xC3\xA5", + "\xC3\x86" => "\xC3\xA6", "\xC3\x87" => "\xC3\xA7", "\xC3\x88" => "\xC3\xA8", "\xC3\x89" => "\xC3\xA9", + "\xC3\x8A" => "\xC3\xAA", "\xC3\x8B" => "\xC3\xAB", "\xC3\x8C" => "\xC3\xAC", "\xC3\x8D" => "\xC3\xAD", + "\xC3\x8E" => "\xC3\xAE", "\xC3\x8F" => "\xC3\xAF", "\xC3\x90" => "\xC3\xB0", "\xC3\x91" => "\xC3\xB1", + "\xC3\x92" => "\xC3\xB2", "\xC3\x93" => "\xC3\xB3", "\xC3\x94" => "\xC3\xB4", "\xC3\x95" => "\xC3\xB5", + "\xC3\x96" => "\xC3\xB6", "\xC3\x98" => "\xC3\xB8", "\xC3\x99" => "\xC3\xB9", "\xC3\x9A" => "\xC3\xBA", + "\xC3\x9B" => "\xC3\xBB", "\xC3\x9C" => "\xC3\xBC", "\xC3\x9D" => "\xC3\xBD", "\xC3\x9E" => "\xC3\xBE", + "\xC4\x80" => "\xC4\x81", "\xC4\x82" => "\xC4\x83", "\xC4\x84" => "\xC4\x85", "\xC4\x86" => "\xC4\x87", + "\xC4\x88" => "\xC4\x89", "\xC4\x8A" => "\xC4\x8B", "\xC4\x8C" => "\xC4\x8D", "\xC4\x8E" => "\xC4\x8F", + "\xC4\x90" => "\xC4\x91", "\xC4\x92" => "\xC4\x93", "\xC4\x96" => "\xC4\x97", "\xC4\x98" => "\xC4\x99", + "\xC4\x9A" => "\xC4\x9B", "\xC4\x9C" => "\xC4\x9D", "\xC4\x9E" => "\xC4\x9F", "\xC4\xA0" => "\xC4\xA1", + "\xC4\xA2" => "\xC4\xA3", "\xC4\xA4" => "\xC4\xA5", "\xC4\xA6" => "\xC4\xA7", "\xC4\xA8" => "\xC4\xA9", + "\xC4\xAA" => "\xC4\xAB", "\xC4\xAE" => "\xC4\xAF", "\xC4\xB4" => "\xC4\xB5", "\xC4\xB6" => "\xC4\xB7", + "\xC4\xB9" => "\xC4\xBA", "\xC4\xBB" => "\xC4\xBC", "\xC4\xBD" => "\xC4\xBE", "\xC5\x81" => "\xC5\x82", + "\xC5\x83" => "\xC5\x84", "\xC5\x85" => "\xC5\x86", "\xC5\x87" => "\xC5\x88", "\xC5\x8A" => "\xC5\x8B", + "\xC5\x8C" => "\xC5\x8D", "\xC5\x90" => "\xC5\x91", "\xC5\x94" => "\xC5\x95", "\xC5\x96" => "\xC5\x97", + "\xC5\x98" => "\xC5\x99", "\xC5\x9A" => "\xC5\x9B", "\xC5\x9C" => "\xC5\x9D", "\xC5\x9E" => "\xC5\x9F", + "\xC5\xA0" => "\xC5\xA1", "\xC5\xA2" => "\xC5\xA3", "\xC5\xA4" => "\xC5\xA5", "\xC5\xA6" => "\xC5\xA7", + "\xC5\xA8" => "\xC5\xA9", "\xC5\xAA" => "\xC5\xAB", "\xC5\xAC" => "\xC5\xAD", "\xC5\xAE" => "\xC5\xAF", + "\xC5\xB0" => "\xC5\xB1", "\xC5\xB2" => "\xC5\xB3", "\xC5\xB4" => "\xC5\xB5", "\xC5\xB6" => "\xC5\xB7", + "\xC5\xB8" => "\xC3\xBF", "\xC5\xB9" => "\xC5\xBA", "\xC5\xBB" => "\xC5\xBC", "\xC5\xBD" => "\xC5\xBE", + "\xC6\xA0" => "\xC6\xA1", "\xC6\xAF" => "\xC6\xB0", "\xC8\x98" => "\xC8\x99", "\xC8\x9A" => "\xC8\x9B", + "\xCE\x86" => "\xCE\xAC", "\xCE\x88" => "\xCE\xAD", "\xCE\x89" => "\xCE\xAE", "\xCE\x8A" => "\xCE\xAF", + "\xCE\x8C" => "\xCF\x8C", "\xCE\x8E" => "\xCF\x8D", "\xCE\x8F" => "\xCF\x8E", "\xCE\x91" => "\xCE\xB1", + "\xCE\x92" => "\xCE\xB2", "\xCE\x93" => "\xCE\xB3", "\xCE\x94" => "\xCE\xB4", "\xCE\x95" => "\xCE\xB5", + "\xCE\x96" => "\xCE\xB6", "\xCE\x97" => "\xCE\xB7", "\xCE\x98" => "\xCE\xB8", "\xCE\x99" => "\xCE\xB9", + "\xCE\x9A" => "\xCE\xBA", "\xCE\x9B" => "\xCE\xBB", "\xCE\x9C" => "\xCE\xBC", "\xCE\x9D" => "\xCE\xBD", + "\xCE\x9E" => "\xCE\xBE", "\xCE\x9F" => "\xCE\xBF", "\xCE\xA0" => "\xCF\x80", "\xCE\xA1" => "\xCF\x81", + "\xCE\xA3" => "\xCF\x83", "\xCE\xA4" => "\xCF\x84", "\xCE\xA5" => "\xCF\x85", "\xCE\xA6" => "\xCF\x86", + "\xCE\xA7" => "\xCF\x87", "\xCE\xA8" => "\xCF\x88", "\xCE\xA9" => "\xCF\x89", "\xCE\xAA" => "\xCF\x8A", + "\xCE\xAB" => "\xCF\x8B", "\xD0\x81" => "\xD1\x91", "\xD0\x82" => "\xD1\x92", "\xD0\x83" => "\xD1\x93", + "\xD0\x84" => "\xD1\x94", "\xD0\x85" => "\xD1\x95", "\xD0\x86" => "\xD1\x96", "\xD0\x87" => "\xD1\x97", + "\xD0\x88" => "\xD1\x98", "\xD0\x89" => "\xD1\x99", "\xD0\x8A" => "\xD1\x9A", "\xD0\x8B" => "\xD1\x9B", + "\xD0\x8C" => "\xD1\x9C", "\xD0\x8E" => "\xD1\x9E", "\xD0\x8F" => "\xD1\x9F", "\xD0\x90" => "\xD0\xB0", + "\xD0\x91" => "\xD0\xB1", "\xD0\x92" => "\xD0\xB2", "\xD0\x93" => "\xD0\xB3", "\xD0\x94" => "\xD0\xB4", + "\xD0\x95" => "\xD0\xB5", "\xD0\x96" => "\xD0\xB6", "\xD0\x97" => "\xD0\xB7", "\xD0\x98" => "\xD0\xB8", + "\xD0\x99" => "\xD0\xB9", "\xD0\x9A" => "\xD0\xBA", "\xD0\x9B" => "\xD0\xBB", "\xD0\x9C" => "\xD0\xBC", + "\xD0\x9D" => "\xD0\xBD", "\xD0\x9E" => "\xD0\xBE", "\xD0\x9F" => "\xD0\xBF", "\xD0\xA0" => "\xD1\x80", + "\xD0\xA1" => "\xD1\x81", "\xD0\xA2" => "\xD1\x82", "\xD0\xA3" => "\xD1\x83", "\xD0\xA4" => "\xD1\x84", + "\xD0\xA5" => "\xD1\x85", "\xD0\xA6" => "\xD1\x86", "\xD0\xA7" => "\xD1\x87", "\xD0\xA8" => "\xD1\x88", + "\xD0\xA9" => "\xD1\x89", "\xD0\xAA" => "\xD1\x8A", "\xD0\xAB" => "\xD1\x8B", "\xD0\xAC" => "\xD1\x8C", + "\xD0\xAD" => "\xD1\x8D", "\xD0\xAE" => "\xD1\x8E", "\xD0\xAF" => "\xD1\x8F", "\xD2\x90" => "\xD2\x91", + "\xE1\xB8\x82" => "\xE1\xB8\x83", "\xE1\xB8\x8A" => "\xE1\xB8\x8B", "\xE1\xB8\x9E" => "\xE1\xB8\x9F", "\xE1\xB9\x80" => "\xE1\xB9\x81", + "\xE1\xB9\x96" => "\xE1\xB9\x97", "\xE1\xB9\xA0" => "\xE1\xB9\xA1", "\xE1\xB9\xAA" => "\xE1\xB9\xAB", "\xE1\xBA\x80" => "\xE1\xBA\x81", + "\xE1\xBA\x82" => "\xE1\xBA\x83", "\xE1\xBA\x84" => "\xE1\xBA\x85", "\xE1\xBB\xB2" => "\xE1\xBB\xB3" ); $UTF8_LOWER_TO_UPPER = array( - 0x0061=>0x0041, 0x03C6=>0x03A6, 0x0163=>0x0162, 0x00E5=>0x00C5, 0x0062=>0x0042, - 0x013A=>0x0139, 0x00E1=>0x00C1, 0x0142=>0x0141, 0x03CD=>0x038E, 0x0101=>0x0100, - 0x0491=>0x0490, 0x03B4=>0x0394, 0x015B=>0x015A, 0x0064=>0x0044, 0x03B3=>0x0393, - 0x00F4=>0x00D4, 0x044A=>0x042A, 0x0439=>0x0419, 0x0113=>0x0112, 0x043C=>0x041C, - 0x015F=>0x015E, 0x0144=>0x0143, 0x00EE=>0x00CE, 0x045E=>0x040E, 0x044F=>0x042F, - 0x03BA=>0x039A, 0x0155=>0x0154, 0x0069=>0x0049, 0x0073=>0x0053, 0x1E1F=>0x1E1E, - 0x0135=>0x0134, 0x0447=>0x0427, 0x03C0=>0x03A0, 0x0438=>0x0418, 0x00F3=>0x00D3, - 0x0440=>0x0420, 0x0454=>0x0404, 0x0435=>0x0415, 0x0449=>0x0429, 0x014B=>0x014A, - 0x0431=>0x0411, 0x0459=>0x0409, 0x1E03=>0x1E02, 0x00F6=>0x00D6, 0x00F9=>0x00D9, - 0x006E=>0x004E, 0x0451=>0x0401, 0x03C4=>0x03A4, 0x0443=>0x0423, 0x015D=>0x015C, - 0x0453=>0x0403, 0x03C8=>0x03A8, 0x0159=>0x0158, 0x0067=>0x0047, 0x00E4=>0x00C4, - 0x03AC=>0x0386, 0x03AE=>0x0389, 0x0167=>0x0166, 0x03BE=>0x039E, 0x0165=>0x0164, - 0x0117=>0x0116, 0x0109=>0x0108, 0x0076=>0x0056, 0x00FE=>0x00DE, 0x0157=>0x0156, - 0x00FA=>0x00DA, 0x1E61=>0x1E60, 0x1E83=>0x1E82, 0x00E2=>0x00C2, 0x0119=>0x0118, - 0x0146=>0x0145, 0x0070=>0x0050, 0x0151=>0x0150, 0x044E=>0x042E, 0x0129=>0x0128, - 0x03C7=>0x03A7, 0x013E=>0x013D, 0x0442=>0x0422, 0x007A=>0x005A, 0x0448=>0x0428, - 0x03C1=>0x03A1, 0x1E81=>0x1E80, 0x016D=>0x016C, 0x00F5=>0x00D5, 0x0075=>0x0055, - 0x0177=>0x0176, 0x00FC=>0x00DC, 0x1E57=>0x1E56, 0x03C3=>0x03A3, 0x043A=>0x041A, - 0x006D=>0x004D, 0x016B=>0x016A, 0x0171=>0x0170, 0x0444=>0x0424, 0x00EC=>0x00CC, - 0x0169=>0x0168, 0x03BF=>0x039F, 0x006B=>0x004B, 0x00F2=>0x00D2, 0x00E0=>0x00C0, - 0x0434=>0x0414, 0x03C9=>0x03A9, 0x1E6B=>0x1E6A, 0x00E3=>0x00C3, 0x044D=>0x042D, - 0x0436=>0x0416, 0x01A1=>0x01A0, 0x010D=>0x010C, 0x011D=>0x011C, 0x00F0=>0x00D0, - 0x013C=>0x013B, 0x045F=>0x040F, 0x045A=>0x040A, 0x00E8=>0x00C8, 0x03C5=>0x03A5, - 0x0066=>0x0046, 0x00FD=>0x00DD, 0x0063=>0x0043, 0x021B=>0x021A, 0x00EA=>0x00CA, - 0x03B9=>0x0399, 0x017A=>0x0179, 0x00EF=>0x00CF, 0x01B0=>0x01AF, 0x0065=>0x0045, - 0x03BB=>0x039B, 0x03B8=>0x0398, 0x03BC=>0x039C, 0x045C=>0x040C, 0x043F=>0x041F, - 0x044C=>0x042C, 0x00FE=>0x00DE, 0x00F0=>0x00D0, 0x1EF3=>0x1EF2, 0x0068=>0x0048, - 0x00EB=>0x00CB, 0x0111=>0x0110, 0x0433=>0x0413, 0x012F=>0x012E, 0x00E6=>0x00C6, - 0x0078=>0x0058, 0x0161=>0x0160, 0x016F=>0x016E, 0x03B1=>0x0391, 0x0457=>0x0407, - 0x0173=>0x0172, 0x00FF=>0x0178, 0x006F=>0x004F, 0x043B=>0x041B, 0x03B5=>0x0395, - 0x0445=>0x0425, 0x0121=>0x0120, 0x017E=>0x017D, 0x017C=>0x017B, 0x03B6=>0x0396, - 0x03B2=>0x0392, 0x03AD=>0x0388, 0x1E85=>0x1E84, 0x0175=>0x0174, 0x0071=>0x0051, - 0x0437=>0x0417, 0x1E0B=>0x1E0A, 0x0148=>0x0147, 0x0105=>0x0104, 0x0458=>0x0408, - 0x014D=>0x014C, 0x00ED=>0x00CD, 0x0079=>0x0059, 0x010B=>0x010A, 0x03CE=>0x038F, - 0x0072=>0x0052, 0x0430=>0x0410, 0x0455=>0x0405, 0x0452=>0x0402, 0x0127=>0x0126, - 0x0137=>0x0136, 0x012B=>0x012A, 0x03AF=>0x038A, 0x044B=>0x042B, 0x006C=>0x004C, - 0x03B7=>0x0397, 0x0125=>0x0124, 0x0219=>0x0218, 0x00FB=>0x00DB, 0x011F=>0x011E, - 0x043E=>0x041E, 0x1E41=>0x1E40, 0x03BD=>0x039D, 0x0107=>0x0106, 0x03CB=>0x03AB, - 0x0446=>0x0426, 0x00FE=>0x00DE, 0x00E7=>0x00C7, 0x03CA=>0x03AA, 0x0441=>0x0421, - 0x0432=>0x0412, 0x010F=>0x010E, 0x00F8=>0x00D8, 0x0077=>0x0057, 0x011B=>0x011A, - 0x0074=>0x0054, 0x006A=>0x004A, 0x045B=>0x040B, 0x0456=>0x0406, 0x0103=>0x0102, - 0x03BB=>0x039B, 0x00F1=>0x00D1, 0x043D=>0x041D, 0x03CC=>0x038C, 0x00E9=>0x00C9, - 0x00F0=>0x00D0, 0x0457=>0x0407, 0x0123=>0x0122, + "\x61" => "\x41", "\x62" => "\x42", "\x63" => "\x43", "\x64" => "\x44", + "\x65" => "\x45", "\x66" => "\x46", "\x67" => "\x47", "\x68" => "\x48", + "\x69" => "\x49", "\x6A" => "\x4A", "\x6B" => "\x4B", "\x6C" => "\x4C", + "\x6D" => "\x4D", "\x6E" => "\x4E", "\x6F" => "\x4F", "\x70" => "\x50", + "\x71" => "\x51", "\x72" => "\x52", "\x73" => "\x53", "\x74" => "\x54", + "\x75" => "\x55", "\x76" => "\x56", "\x77" => "\x57", "\x78" => "\x58", + "\x79" => "\x59", "\x7A" => "\x5A", "\xC3\xA0" => "\xC3\x80", "\xC3\xA1" => "\xC3\x81", + "\xC3\xA2" => "\xC3\x82", "\xC3\xA3" => "\xC3\x83", "\xC3\xA4" => "\xC3\x84", "\xC3\xA5" => "\xC3\x85", + "\xC3\xA6" => "\xC3\x86", "\xC3\xA7" => "\xC3\x87", "\xC3\xA8" => "\xC3\x88", "\xC3\xA9" => "\xC3\x89", + "\xC3\xAA" => "\xC3\x8A", "\xC3\xAB" => "\xC3\x8B", "\xC3\xAC" => "\xC3\x8C", "\xC3\xAD" => "\xC3\x8D", + "\xC3\xAE" => "\xC3\x8E", "\xC3\xAF" => "\xC3\x8F", "\xC3\xB0" => "\xC3\x90", "\xC3\xB1" => "\xC3\x91", + "\xC3\xB2" => "\xC3\x92", "\xC3\xB3" => "\xC3\x93", "\xC3\xB4" => "\xC3\x94", "\xC3\xB5" => "\xC3\x95", + "\xC3\xB6" => "\xC3\x96", "\xC3\xB8" => "\xC3\x98", "\xC3\xB9" => "\xC3\x99", "\xC3\xBA" => "\xC3\x9A", + "\xC3\xBB" => "\xC3\x9B", "\xC3\xBC" => "\xC3\x9C", "\xC3\xBD" => "\xC3\x9D", "\xC3\xBE" => "\xC3\x9E", + "\xC3\xBF" => "\xC5\xB8", "\xC4\x81" => "\xC4\x80", "\xC4\x83" => "\xC4\x82", "\xC4\x85" => "\xC4\x84", + "\xC4\x87" => "\xC4\x86", "\xC4\x89" => "\xC4\x88", "\xC4\x8B" => "\xC4\x8A", "\xC4\x8D" => "\xC4\x8C", + "\xC4\x8F" => "\xC4\x8E", "\xC4\x91" => "\xC4\x90", "\xC4\x93" => "\xC4\x92", "\xC4\x97" => "\xC4\x96", + "\xC4\x99" => "\xC4\x98", "\xC4\x9B" => "\xC4\x9A", "\xC4\x9D" => "\xC4\x9C", "\xC4\x9F" => "\xC4\x9E", + "\xC4\xA1" => "\xC4\xA0", "\xC4\xA3" => "\xC4\xA2", "\xC4\xA5" => "\xC4\xA4", "\xC4\xA7" => "\xC4\xA6", + "\xC4\xA9" => "\xC4\xA8", "\xC4\xAB" => "\xC4\xAA", "\xC4\xAF" => "\xC4\xAE", "\xC4\xB5" => "\xC4\xB4", + "\xC4\xB7" => "\xC4\xB6", "\xC4\xBA" => "\xC4\xB9", "\xC4\xBC" => "\xC4\xBB", "\xC4\xBE" => "\xC4\xBD", + "\xC5\x82" => "\xC5\x81", "\xC5\x84" => "\xC5\x83", "\xC5\x86" => "\xC5\x85", "\xC5\x88" => "\xC5\x87", + "\xC5\x8B" => "\xC5\x8A", "\xC5\x8D" => "\xC5\x8C", "\xC5\x91" => "\xC5\x90", "\xC5\x95" => "\xC5\x94", + "\xC5\x97" => "\xC5\x96", "\xC5\x99" => "\xC5\x98", "\xC5\x9B" => "\xC5\x9A", "\xC5\x9D" => "\xC5\x9C", + "\xC5\x9F" => "\xC5\x9E", "\xC5\xA1" => "\xC5\xA0", "\xC5\xA3" => "\xC5\xA2", "\xC5\xA5" => "\xC5\xA4", + "\xC5\xA7" => "\xC5\xA6", "\xC5\xA9" => "\xC5\xA8", "\xC5\xAB" => "\xC5\xAA", "\xC5\xAD" => "\xC5\xAC", + "\xC5\xAF" => "\xC5\xAE", "\xC5\xB1" => "\xC5\xB0", "\xC5\xB3" => "\xC5\xB2", "\xC5\xB5" => "\xC5\xB4", + "\xC5\xB7" => "\xC5\xB6", "\xC5\xBA" => "\xC5\xB9", "\xC5\xBC" => "\xC5\xBB", "\xC5\xBE" => "\xC5\xBD", + "\xC6\xA1" => "\xC6\xA0", "\xC6\xB0" => "\xC6\xAF", "\xC8\x99" => "\xC8\x98", "\xC8\x9B" => "\xC8\x9A", + "\xCE\xAC" => "\xCE\x86", "\xCE\xAD" => "\xCE\x88", "\xCE\xAE" => "\xCE\x89", "\xCE\xAF" => "\xCE\x8A", + "\xCE\xB1" => "\xCE\x91", "\xCE\xB2" => "\xCE\x92", "\xCE\xB3" => "\xCE\x93", "\xCE\xB4" => "\xCE\x94", + "\xCE\xB5" => "\xCE\x95", "\xCE\xB6" => "\xCE\x96", "\xCE\xB7" => "\xCE\x97", "\xCE\xB8" => "\xCE\x98", + "\xCE\xB9" => "\xCE\x99", "\xCE\xBA" => "\xCE\x9A", "\xCE\xBB" => "\xCE\x9B", "\xCE\xBC" => "\xCE\x9C", + "\xCE\xBD" => "\xCE\x9D", "\xCE\xBE" => "\xCE\x9E", "\xCE\xBF" => "\xCE\x9F", "\xCF\x80" => "\xCE\xA0", + "\xCF\x81" => "\xCE\xA1", "\xCF\x83" => "\xCE\xA3", "\xCF\x84" => "\xCE\xA4", "\xCF\x85" => "\xCE\xA5", + "\xCF\x86" => "\xCE\xA6", "\xCF\x87" => "\xCE\xA7", "\xCF\x88" => "\xCE\xA8", "\xCF\x89" => "\xCE\xA9", + "\xCF\x8A" => "\xCE\xAA", "\xCF\x8B" => "\xCE\xAB", "\xCF\x8C" => "\xCE\x8C", "\xCF\x8D" => "\xCE\x8E", + "\xCF\x8E" => "\xCE\x8F", "\xD0\xB0" => "\xD0\x90", "\xD0\xB1" => "\xD0\x91", "\xD0\xB2" => "\xD0\x92", + "\xD0\xB3" => "\xD0\x93", "\xD0\xB4" => "\xD0\x94", "\xD0\xB5" => "\xD0\x95", "\xD0\xB6" => "\xD0\x96", + "\xD0\xB7" => "\xD0\x97", "\xD0\xB8" => "\xD0\x98", "\xD0\xB9" => "\xD0\x99", "\xD0\xBA" => "\xD0\x9A", + "\xD0\xBB" => "\xD0\x9B", "\xD0\xBC" => "\xD0\x9C", "\xD0\xBD" => "\xD0\x9D", "\xD0\xBE" => "\xD0\x9E", + "\xD0\xBF" => "\xD0\x9F", "\xD1\x80" => "\xD0\xA0", "\xD1\x81" => "\xD0\xA1", "\xD1\x82" => "\xD0\xA2", + "\xD1\x83" => "\xD0\xA3", "\xD1\x84" => "\xD0\xA4", "\xD1\x85" => "\xD0\xA5", "\xD1\x86" => "\xD0\xA6", + "\xD1\x87" => "\xD0\xA7", "\xD1\x88" => "\xD0\xA8", "\xD1\x89" => "\xD0\xA9", "\xD1\x8A" => "\xD0\xAA", + "\xD1\x8B" => "\xD0\xAB", "\xD1\x8C" => "\xD0\xAC", "\xD1\x8D" => "\xD0\xAD", "\xD1\x8E" => "\xD0\xAE", + "\xD1\x8F" => "\xD0\xAF", "\xD1\x91" => "\xD0\x81", "\xD1\x92" => "\xD0\x82", "\xD1\x93" => "\xD0\x83", + "\xD1\x94" => "\xD0\x84", "\xD1\x95" => "\xD0\x85", "\xD1\x96" => "\xD0\x86", "\xD1\x97" => "\xD0\x87", + "\xD1\x98" => "\xD0\x88", "\xD1\x99" => "\xD0\x89", "\xD1\x9A" => "\xD0\x8A", "\xD1\x9B" => "\xD0\x8B", + "\xD1\x9C" => "\xD0\x8C", "\xD1\x9E" => "\xD0\x8E", "\xD1\x9F" => "\xD0\x8F", "\xD2\x91" => "\xD2\x90", + "\xE1\xB8\x83" => "\xE1\xB8\x82", "\xE1\xB8\x8B" => "\xE1\xB8\x8A", "\xE1\xB8\x9F" => "\xE1\xB8\x9E", "\xE1\xB9\x81" => "\xE1\xB9\x80", + "\xE1\xB9\x97" => "\xE1\xB9\x96", "\xE1\xB9\xA1" => "\xE1\xB9\xA0", "\xE1\xB9\xAB" => "\xE1\xB9\xAA", "\xE1\xBA\x81" => "\xE1\xBA\x80", + "\xE1\xBA\x83" => "\xE1\xBA\x82", "\xE1\xBA\x85" => "\xE1\xBA\x84", "\xE1\xBB\xB3" => "\xE1\xBB\xB2" ); /** @@ -437,30 +431,14 @@ else * not exist in the Chinese alphabet, for example. See Unicode Standard * Annex #21: Case Mappings * - * @author Andreas Gohr <andi@splitbrain.org> * @param string - * @return mixed either string in lowercase or FALSE is UTF-8 invalid + * @return string string in lowercase */ function utf8_strtolower($string) { global $UTF8_UPPER_TO_LOWER; - $uni = utf8_to_unicode($string); - - if (!$uni) - { - return false; - } - - for ($i = 0, $cnt = sizeof($uni); $i < $cnt; $i++) - { - if (isset($UTF8_UPPER_TO_LOWER[$uni[$i]])) - { - $uni[$i] = $UTF8_UPPER_TO_LOWER[$uni[$i]]; - } - } - - return utf8_from_unicode($uni); + return strtr($string, $UTF8_UPPER_TO_LOWER); } /** @@ -471,106 +449,160 @@ else * not exist in the Chinese alphabet, for example. See Unicode Standard * Annex #21: Case Mappings * - * @author Andreas Gohr <andi@splitbrain.org> * @param string - * @return mixed either string in lowercase or FALSE is UTF-8 invalid + * @return string string in uppercase */ - function utf8_strtoupper($str) + function utf8_strtoupper($string) { global $UTF8_LOWER_TO_UPPER; - $uni = utf8_to_unicode($string); - - if (!$uni) - { - return false; - } - - for ($i = 0, $cnt = sizeof($uni); $i < $cnt; $i++) - { - if (isset($UTF8_LOWER_TO_UPPER[$uni[$i]])) - { - $uni[$i] = $UTF8_LOWER_TO_UPPER[$uni[$i]]; - } - } - - return utf8_from_unicode($uni); + return strtr($string, $UTF8_LOWER_TO_UPPER); } /** * UTF-8 aware alternative to substr * Return part of a string given character offset (and optionally length) - * - * @author Harry Fuecks + * + * Note arguments: comparied to substr - if offset or length are + * not integers, this version will not complain but rather massages them + * into an integer. + * + * Note on returned values: substr documentation states false can be + * returned in some cases (e.g. offset > string length) + * mb_substr never returns false, it will return an empty string instead. + * This adopts the mb_substr approach + * + * Note on implementation: PCRE only supports repetitions of less than + * 65536, in order to accept up to MAXINT values for offset and length, + * we'll repeat a group of 65535 characters when needed. + * + * Note on implementation: calculating the number of characters in the + * string is a relatively expensive operation, so we only carry it out when + * necessary. It isn't necessary for +ve offsets and no specified length + * + * @author Chris Smith<chris@jalakai.co.uk> * @param string * @param integer number of UTF-8 characters offset (from left) * @param integer (optional) length in UTF-8 characters from offset * @return mixed string or FALSE if failure */ - function utf8_substr($str, $offset, $length = null) + function utf8_substr($str, $offset, $length = NULL) { - if ($offset >= 0 && $length >= 0) + // generates E_NOTICE + // for PHP4 objects, but not PHP5 objects + $str = (string) $str; + $offset = (int) $offset; + if (!is_null($length)) { - if ($length === null) - { - $length = '*'; - } - else - { - if (!preg_match('/^[0-9]+$/', $length)) - { - trigger_error('utf8_substr expects parameter 3 to be long', E_USER_WARNING); - return false; - } - - $strlen = strlen(utf8_decode($str)); - if ($offset > $strlen) - { - return ''; - } + $length = (int) $length; + } - if (($offset + $length) > $strlen) - { - $length = '*'; - } - else - { - $length = '{' . $length . '}'; - } - } + // handle trivial cases + if ($length === 0 || ($offset < 0 && $length < 0 && $length < $offset)) + { + return ''; + } - if (!preg_match('/^[0-9]+$/', $offset)) + // normalise negative offsets (we could use a tail + // anchored pattern, but they are horribly slow!) + if ($offset < 0) + { + // see notes + $strlen = utf8_strlen($str); + $offset = $strlen + $offset; + if ($offset < 0) { - trigger_error('utf8_substr expects parameter 2 to be long', E_USER_WARNING); - return false; + $offset = 0; } + } - $pattern = '/^.{' . $offset . '}(.' . $length . ')/us'; + $op = ''; + $lp = ''; - preg_match($pattern, $str, $matches); + // establish a pattern for offset, a + // non-captured group equal in length to offset + if ($offset > 0) + { + $ox = (int) ($offset / 65535); + $oy = $offset % 65535; - if (isset($matches[1])) + if ($ox) { - return $matches[1]; + $op = '(?:.{65535}){' . $ox . '}'; } - return false; + $op = '^(?:' . $op . '.{' . $oy . '})'; + } + else + { + // offset == 0; just anchor the pattern + $op = '^'; + } + + // establish a pattern for length + if (is_null($length)) + { + // the rest of the string + $lp = '(.*)$'; } else { - // Handle negatives using different, slower technique - // From: http://www.php.net/manual/en/function.substr.php#44838 - preg_match_all('/./u', $str, $ar); + if (!isset($strlen)) + { + // see notes + $strlen = utf8_strlen($str); + } - if ($length !== null) + // another trivial case + if ($offset > $strlen) { - return join('', array_slice($ar[0], $offset, $length)); + return ''; } - else + + if ($length > 0) + { + // reduce any length that would + // go passed the end of the string + $length = min($strlen - $offset, $length); + + $lx = (int) ($length / 65535); + $ly = $length % 65535; + + // negative length requires a captured group + // of length characters + if ($lx) + { + $lp = '(?:.{65535}){' . $lx . '}'; + } + $lp = '(' . $lp . '.{'. $ly . '})'; + } + else if ($length < 0) { - return join('', array_slice($ar[0], $offset)); + if ($length < ($offset - $strlen)) + { + return ''; + } + + $lx = (int)((-$length) / 65535); + $ly = (-$length) % 65535; + + // negative length requires ... capture everything + // except a group of -length characters + // anchored at the tail-end of the string + if ($lx) + { + $lp = '(?:.{65535}){' . $lx . '}'; + } + $lp = '(.*)(?:' . $lp . '.{' . $ly . '})$'; } } + + if (!preg_match('#' . $op . $lp . '#us', $str, $match)) + { + return ''; + } + + return $match[1]; } /** @@ -854,84 +886,6 @@ function utf8_decode_ncr_callback($m) } /** -* Takes an UTF-8 string and returns an array of ints representing the -* Unicode characters. -* -* @param string UTF-8 encoded string -* @return array array of UNICODE code points -*/ -function utf8_to_unicode($string) -{ - $unicode = array(); - $offset = 0; - $stringlength = strlen($string); - - while ($offset < $stringlength) - { - $ord = ord($string{$offset}); - if (($ord | 0x07) == 0xF7) - { - // 11110bbb 10bbbbbb 10bbbbbb 10bbbbbb - $charval = (($ord & 0x07) << 18) & - ((ord($string{($offset + 1)}) & 0x3F) << 12) & - ((ord($string{($offset + 2)}) & 0x3F) << 6) & - (ord($string{($offset + 3)}) & 0x3F); - $offset += 4; - } - else if (($ord | 0x0F) == 0xEF) - { - // 1110bbbb 10bbbbbb 10bbbbbb - $charval = (($ord & 0x0F) << 12) & - ((ord($string{($offset + 1)}) & 0x3F) << 6) & - (ord($string{($offset + 2)}) & 0x3F); - $offset += 3; - } - else if (($ord | 0x1F) == 0xDF) - { - // 110bbbbb 10bbbbbb - $charval = (($ord & 0x1F) << 6) & - (ord($string{($offset + 1)}) & 0x3F); - $offset += 2; - } - else if (($ord | 0x7F) == 0x7F) - { - // 0bbbbbbb - $charval = $ord; - $offset += 1; - } - else - { - // error? throw some kind of warning here? - $charval = false; - $offset += 1; - } - if ($charval !== false) - { - $unicode[] = $charval; - } - } - return $unicode; -} - -/** -* Takes an array of ints representing the Unicode characters and returns -* a UTF-8 string. -* -* @param array $array array of unicode code points representing a string -* @return string UTF-8 character string -*/ -function utf8_from_unicode($array) -{ - $str = ''; - foreach ($array as $value) - { - $str .= utf8_chr($value); - } - return $str; -} - - -/** * Takes an array of ints representing the Unicode characters and returns * a UTF-8 string. * |