public function compare($oldText, $newText)
{
$this->content = array();
$this->wordIndices = array();
$this->oldWords = $this->ConvertHtmlToListOfWords(str_split_utf8($oldText));
$this->newWords = $this->ConvertHtmlToListOfWords(str_split_utf8($newText));
$this->wordIndices = $this->IndexNewWords($this->newWords);
$operations = $this->Operations();
//print_r($this->wordIndices);
//print_r($this->newWords);
foreach ($operations as $item) {
$this->PerformOperation($item);
}
return implode('', $this->content);
}
$split = 1;
}
$key = NULL;
for ($j = 0; $j < $split; $j++, $i++) {
$key .= $str[$i];
}
array_push($array, $key);
}
return $array;
}
error_reporting(0);
if ($_POST['submit']) {
$arr = file($_FILES['file']['tmp_name']);
$bword = $_POST['bword'];
foreach ($arr as $line) {
$lineArray = str_split_utf8(trim($line));
sort($lineArray);
$result[$line] = implode('', $lineArray);
}
$result = array_unique($result);
echo implode('<br>', array_keys($result));
die;
}
?>
<form name="word" action="" method="post" enctype="multipart/form-data">
<label for="file">A文件:</label>
<input type="file" name="file" id="file" />
<br />
<input type="submit" name="submit" value="提交查看结果" />
function str_split_any($str, $encode = NULL)
{
empty($encode) && ($encode = mb_detect_encoding($str, 'ASCII,UTF-8,UCS-2,UCS-4,UTF-16,UTF-32'));
$result = array();
$encode = encoding_aliases($encode);
switch ($encode) {
case 'UTF-8':
case 'UTF-8-Mobile#DOCOMO':
case 'UTF-8-Mobile#KDDI-A':
case 'UTF-8-Mobile#KDDI-B':
case 'UTF-8-Mobile#SOFTBANK':
$result = str_split_utf8($str);
break;
case 'UTF-16':
case 'UTF-16LE':
case 'UTF-16BE':
case 'UCS-2':
case 'UCS-2LE':
case 'UCS-2BE':
$result = str_split_utf16($str);
break;
case 'UTF-32':
case 'UTF-32LE':
case 'UTF-32BE':
case 'UCS-4':
case 'UCS-4LE':
case 'UCS-4BE':
$result = str_split_utf32($str);
break;
case 'ASCII':
//单字节
//单字节
case '8bit':
//拉丁文区
//拉丁文区
case 'ISO-8859-1':
case 'ISO-8859-2':
case 'ISO-8859-3':
case 'ISO-8859-4':
case 'ISO-8859-5':
case 'ISO-8859-6':
case 'ISO-8859-7':
case 'ISO-8859-8':
case 'ISO-8859-9':
case 'ISO-8859-10':
case 'ISO-8859-13':
case 'ISO-8859-14':
case 'ISO-8859-15':
case 'ISO-8859-16':
case 'Windows-1251':
case 'Windows-1252':
case 'Windows-1254':
case 'CP850':
case 'CP866':
case 'ArmSCII-8':
//单字节区
//单字节区
case 'pass':
case 'auto':
case 'wchar':
case 'byte2be':
case 'byte2le':
case 'byte4be':
case 'byte4le':
case 'BASE64':
case 'UUENCODE':
case 'HTML-ENTITIES':
case 'Quoted-Printable':
case '7bit':
$result = str_split($str);
break;
default:
$result = str_split_multibyte($str, $encode);
break;
}
return $result;
}
private function dividePatternStr($str)
{
$strArray1 = explode(" ", $str);
foreach ($strArray1 as $str) {
unset($strArray2);
$strArray2 = str_split_utf8($str);
$resultStr = implode("[^,]*", $strArray2);
$result[] = "/,[^,]*" . $resultStr . "[^,]*,/i";
}
return $result;
}
