[Resolved] splitting a word into letters (encoding unknown)

neerav

While making animated graphics through mobile phone, I take a string and make it look like it's being typed out. To do this, I take a substring from 0 to i and create the graphic with that substring. At the end, put all the graphics together and I have the animation.

		for ($i=1;$i<=strlen($text);$i++) {
	$frame++;  //advance frame number
	$result = create_frame("$image_url".urlencode(substr($text,0,$i)),$filename."-$frame");
	if (!$result) return array("",false); else $results[] = $result;
}

This works perfectly fine for most people. For me, it even works with characters with umlauts, accents, and the like. However, a Danish user complained that before each "special" character, there appeared an extra character. It seems to always be an A with a tilde over it. See the slow-motion sample attached.

I figure this has to do with the way I am taking the substrings. If I take two characters at a time, this works on the Danish mobile, but only displays every other character for most mobiles.

So how do I determine whether I need to take one or two characters at a time? Or is there a better way to take the substrings?

konsu

Hello, where does you php code get the string to animate from? Is it a web request? Does the request carry the encoding header? But php does not have unicode support. so it is all useless. : -) I would try to write something up in a system that supports unicode to find where the cause is. I think the cause is that your string is unicode and the PHP's function substr is not aware about it.

neerav

The string comes from the mobile user through a text field in a form.

If php doesn't support unicode then how does this work from the danish phone in the first place? I think I am missing some fundamental understanding, but I don't know what! 😉

The only relevent header available is the Accept encoding. Usually that it just *, which doesn't help me in determining what the request encoding was. 🙁

Any other ideas?

konsu

the reason why it works in some languages and not the other is probably because the encoding is utf-8 and not utf-16, so it works for the characters that use just 8 bits and fails for characters that use longer byte sequences. as i said, the only way i can think of to verify this is to implement this in, say, python that supports unicode.

neerav

Ok. Even though php isn't multibyte compatible, isn't there a way to test the string for utf-8 or utf-16? Then I can instruct php to ignore or throw an error if the string is utf-16.

Thank you for your assistance.

konsu

this might help: http://www.php.net/manual/en/ref.mbstring.php

also, there are plenty of resources on the web about utf-8, so if the above is not feasible, it is still possible to implement a string splitting function that knows how to deal with utf-8.

neerav

Perfect! I've been making the wrong searches in the manual. Haven't seen this yet.

It looks like this set of functions will do all I need and more. Now to find out if the installation of php on my server has enabled it!

Konsu, thank you.