MySQL Bugs: #15714: MySQL crashes - Full-text Search (Boolean Mode)

Bug #15714	MySQL crashes - Full-text Search (Boolean Mode) - Data Encoded in UTF-8
Submitted:	13 Dec 2005 15:24	Modified:	13 Dec 2005 19:16
Reporter:	AmirBehzad Eslami	Email Updates:
Status:	Closed	Impact on me:	None
Category:	MySQL Server	Severity:	S1 (Critical)
Version:	4.1.7	OS:	Windows (Windows XP Pro)
Assigned to:	MySQL Verification Team	CPU Architecture:	Any

Description:
I'm programming a search engine to locate words within Persian articles. These articles are encoded in Unicode (UTF-8), the famous multi-byte character set.

I have a simple table, which consists of 2 columns: `article_id` and `article_text`. There's a full-text index on `article_text`.

Everything works perfect in a normal MATCH(article_title)...AGAINST('"EXACT PHRASE - multibyte"').

But MySQL crashes on Boolean Full-text search, using MATCH(article_title)...AGAINST('"EXACT PHRASE - multibyte"' IN BOOLEAN MODE).

And this error only occures when there is at least one matching row.

P.S.
I'm running MySQL 4.1.7-nt-log under Windows XP Pro.

How to repeat:
-- It is suffice to execute to following query:

CREATE TABLE `mytable` (
  `article_id` int(10) unsigned NOT NULL default '0',
  `article_title` text collate utf8_persian_ci NOT NULL,
  PRIMARY KEY  (`article_id`),
  FULLTEXT KEY `article_title` (`article_title`)
) ENGINE=MyISAM DEFAULT CHARSET=utf8 COLLATE=utf8_persian_ci;

INSERT INTO
`mytable` ( `article_id` , `article_title` ) 
VALUES
('1', 'بحری است بحر عشق، که هیچش کناره نیست'),
('2', 'هفت شهر عشق را عطار گشت'),
('3', 'علت عشق ز علتها جداست'),
('4', 'هر که هوایی نپخت، یا به فراقی نسوخت');

SELECT MATCH(article_title) AGAINST ('"عطار"' IN BOOLEAN MODE) FROM mytable

I was unable to repeat the crash with 4.1.16. Could you please try
with the latest release, your version is quite older.

Thanks in advance.

Dear Miguel,
Thank you for your advice.

You meant that it was a bug which is fixed after MySQL 4.1.7 ?

I tested the query at home under MySQL 5 and - Yes! it works perfect. Thanks Man!

But I'm not sure whether my client would install the latest release or not.
They are not using a dedicated server.

I would really appreciate if you take a brief moment of your time to answer my questions:

-- Is there any patch for such a bug?

-- What is the lowest MySQL version, which perfectly performs the Boolean Full-text search within multi-byte characters?

-- I just visited http://dev.mysql.com/downloads/mysql/4.1.html and noted there is only version 4.1.15 available to download. Where can I find MySQL 4.1.16 which you're using now? Where can I find older versions, such as MySQL 4.1.8?

Thank you in advance

When verifying bug we do on the current source and sometimes we are aware
when the bug was fixed, however in this case I can't say for you how was done
the fix, so I recommend you to download the today latest release and to test
on test environment, anyway for sure there are a lot of fixes from that older
version.

Thanks in advance.

I found it :-D
http://bugs.mysql.com/bug.php?id=8351