Description:
------------
Japanese katakana characters can be encoded as single-byte especially when using mobile phones. Most of the Japanese mobile sites use Shift-JIS encoding. For security purposes, htmlentities is used in all user inputs. Unfortunately, when single-byte katakana strings encoded in Shift-JIS are passed in htmlentities, garbage characters are returned.

Test script:
---------------
//make sure that the code as well as the browser settings are encoded also in Shift-JIS
$txt = "｡｢｣､･ｦｧｨｩｪｫｬｭｮｯ ｰｱｲｳｴｵｶｷｸｹｺｻｼｽｾｿﾀﾁﾂﾃﾄﾅﾆﾇﾈﾉﾊﾋﾌﾍﾎﾏﾐﾑﾒﾓﾔﾕﾖﾗﾘﾙﾚﾛﾜﾝﾞﾟアバ日本語 ＤＯＵＢＬＥＢＹＴＥ SINGLE BYTE<br>";
echo "Original String: " . $txt;
$new_txt = htmlentities($txt, ENT_QUOTES, "SJIS");
echo "<br>Converted String: " . $new_txt;

Actual result:
--------------
Original String: ｡｢｣､･ｦｧｨｩｪｫｬｭｮｯ ｰｱｲｳｴｵｶｷｸｹｺｻｼｽｾｿﾀﾁﾂﾃﾄﾅﾆﾇﾈﾉﾊﾋﾌﾍﾎﾏﾐﾑﾒﾓﾔﾕﾖﾗﾘﾙﾚﾛﾜﾝﾞﾟアバ日本語 ＤＯＵＢＬＥＢＹＴＥ SINGLE BYTE

Converted String: ¡¢£¤¥¦§¨©ª«¬­®¯ °±²³´µ¶·¸¹º»¼½¾¿ÀÁÂÃÄÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖ×ØÙÚÛÜÝÞßアバ日本語 ＤＯＵＢＬＥＢＹＴＥ SINGLE BYTE<br>