Question

I have a website that accepts user submissions to a forum from different locales, English and Swedish are the current “supported” locales. The most common language on the forums are in Swedish and that is where I encounter the intermittent problem of character encoding.

Could it be that some browsers are sending me ISO 8859 encoded strings but the page is encoded in utf-8 (and should be submitted in that encoding?). My php serverside is guessing the encoding with stuff like mb_detect_encoding but that doesn’t seem to help.

I have this code to “guess” the encoding

if ( mb_detect_encoding($str, 'UTF-8, ISO-8859-1') == 'ISO-8859-1') {
    return mb_convert_encoding($str, 'UTF-8', 'ISO-8859-1');
}
return $str;  

on the submissions. Other encoding options are not an issue for this particular problem.
Any help would be appreciated.

Answer 1

The browser may send data in any character encoding, regardless of the character encoding of your HTML page. It should advertise the used encoding in the Content-Type header. You can use the accept-charset Atrribute on the form to specify which character encodings you want to receive.