International Characters and JSON

[Shawazi]

I have a string with some international characters. The information containing these characters is pulled onto the page using AJAX. Here is the weird set of circumstances:

If I load the information using transport.reponseText (where transport is the AJAX object) everything loads fine.

If I want to use json_encode there is an issue. It will cut off the string and stop my JSON output right there.

This does not happen with this code:

$response['html'] = $description."<br />";
echo json_encode($response);

But if I truncate longer words it will:

function truncate($txt, $max, $sub) {
	$txtRay = explode(" ", $txt);
	$new_txt = '';
	foreach ($txtRay as $key => $word) {
		$len = strlen($word);
		if ($len > $max) {
		    $start = mb_substr($word, 0, ($max-$sub));
		$end   = mb_substr($word, ($len-$sub));

		$word  = $start."...".$end;
	}
	$new_txt .= ' '.$word;
}
$new_txt = trim($new_txt);
return $new_txt;
}
$response['html'] = truncate($description, 35, 6)."<br /><br />";
echo json_encode($response);

As you can see I am using mb_substr to be extra safe. However when I print out the json_encoded string AFTER running it through that function, it cuts off shortly after the letter.

[Shawazi]

So basically it would only mess up if I json_encode something after using that substr. But actually my coworker just read this post and I changed the truncation to:

function truncate($txt, $max, $sub) {
	$txtRay = explode(" ", $txt);
	$new_txt = '';
	foreach ($txtRay as $key => $word) {
		$len = strlen($word);
		if ($len > $max) {

		$word = mb_convert_encoding($word, "ASCII", "UTF-8"); //new line

	    $start = mb_substr($word, 0, ($max-$sub));
		$end   = mb_substr($word, ($len-$sub));

		$word  = $start."...".$end;
	}
	$new_txt .= ' '.$word;
}
$new_txt = trim($new_txt);
return $new_txt;
}

and it worked fine. Can anyone explain why this would have happened?