Question

Why does this:

<?
var_dump ( json_decode('{"unicode":"\u30d7\u30ec\u30b9\u30ad\u30c3\u30c8"}') );
?>

give me:

object(stdClass)#1 (1) {
  ["unicode"]=>
  string(18) "ãã¬ã¹ã­ãã"
}

when according to the docs it should be “プレスキット”

This happens to me on one machine running PHP 5.1.6 with this module installed as per the instructions on http://www.aurore.net/projects/php-json/ and on a machine running PHP 5.2.11

Answer 1

You need to let the browser know to interpret the data as UTF-8 by sending an HTTP header prior to outputting the actual JSON like so:

header(‘Content-Type: application/json; charset=utf-8’);

Note that headers must be sent before any other output. You should probably read up on the HTTP spec and learn about how headers are used. Here’s a link to get you started with sending HTTP headers in PHP.

---

UPDATE

In an instance like this where the correct headers have been sent, it’s important to make sure the locale settings on your testing machine allow for UTF-8 characters. If the default locale on your system doesn’t support UTF-8 and you don’t have one installed that does, you’ll never see the correct output. You can read a little more about this here