CHAR_LENGTH() returns incorrect value on Japanese UTF8 text

[Gregor Kaplan]

Execute the following:

CREATE TABLE multibyte
(
thing VARCHAR(20) CHARACTER SET utf8
);

INSERT INTO multibyte (thing) VALUES('human');
INSERT INTO multibyte (thing) VALUES('ははは'); #if you can't read this it's "hahaha" in hiragana

SELECT thing, CHAR_LENGTH(thing), LENGTH(thing) FROM multibyte;

Result:
+-----------+--------------------+---------------+
| thing | CHAR_LENGTH(thing) | LENGTH(thing) |
+-----------+--------------------+---------------+
| human | 5 | 5 |
| ははは | 9 | 18 |
+-----------+--------------------+

The return value should be 3, as that is the actual number of UTF8 characters. However, if one were to open the file in a plain text editor, it would appear as "„ÅØ„ÅØ„ÅØ" - which is 9 characters, just not 9 meaningful characters. That and they certainly shouldn't have a length of 18.

So, is this a bug, or is there a correct way to call this that I am simply missing?

Much thanks and appreciation in advance,

Gregor