上次提取了utf-8转GBK,以下内容是从phpcms v9中提取的函数源码，去除去其它的转码代码，只留下GBK转UTF-8部门的源码，对收罗更有用哦
共有两个文件，一个PHP函数源码文件，一个转编码库文件，
<?php /** * gbk转utf8 * @param $gbstr */ function gbk_to_utf8($gbstr) { global $CODETABLE; if(empty($CODETABLE)) { $filename = 'encoding/gb-unicode.table'; $fp = fopen($filename, 'rb'); while($l = fgets($fp,15)) { $CODETABLE[hexdec(substr($l, 0, 6))] = substr($l, 7, 6); } fclose($fp); } $ret = ''; $utf8 = ''; while($gbstr) { if(ord(substr($gbstr, 0, 1)) > 0x80) { $thisW = substr($gbstr, 0, 2); $gbstr = substr($gbstr, 2, strlen($gbstr)); $utf8 = ''; @$utf8 = unicode_to_utf8(hexdec($CODETABLE[hexdec(bin2hex($thisW)) - 0x8080])); if($utf8 != '') { for($i = 0; $i < strlen($utf8); $i += 3) $ret .= chr(substr($utf8, $i, 3)); } } else { $ret .= substr($gbstr, 0, 1); $gbstr = substr($gbstr, 1, strlen($gbstr)); } } return $ret; } /** * unicode转utf8 * @param $c */ function unicode_to_utf8($c) { $str = ''; if($c < 0x80) { $str .= $c; } elseif($c < 0x800) { $str .= (0xC0 | $c >> 6); $str .= (0x80 | $c & 0x3F); } elseif($c < 0x10000) { $str .= (0xE0 | $c >> 12); $str .= (0x80 | $c >> 6 & 0x3F); $str .= (0x80 | $c & 0x3F); } elseif($c < 0x200000) { $str .= (0xF0 | $c >> 18); $str .= (0x80 | $c >> 12 & 0x3F); $str .= (0x80 | $c >> 6 & 0x3F); $str .= (0x80 | $c & 0x3F); } return $str; } //以下是测试内容 //test.txt 为GBK编码文档 header('Content-type: text/html; charset=utf-8'); $str = file_get_contents('test.txt'); echo gbk_to_utf8($str); ?>又是另一个转码?post=198
end
附件下载/演示源码：
【 gbk_to_utf8.rar 】 35.38KB