I'm currently doing some utf8 tests with mysql, during these tests i think i've hit a bug already discussed on this list.
http://marc.theaimsgroup.com/?l=mysql&m=105593058922219&w=2 http://marc.theaimsgroup.com/?t=105577025900001&r=1&w=2
these two bug report and responses suggest that it's because client session does not use utf8 character set.
In my case i'm "sure" both client and server are using utf8. ("status"
command report it)So the problem is :
Using Mysql 4.1.0-alpha (recompiled rpm) with InnoDb Database and charset utf8 (both client session and server) any simple select query return this error : "ERROR 1210: Wrong arguments to ="
Exemple :
mysql> status; -------------- mysql Ver 13.5 Distrib 4.1.0-alpha, for pc-linux (i686)
Connection id: 8 Current database: agatechat Current user: [EMAIL PROTECTED] SSL: Not in use Current pager: stdout Using outfile: '' Server version: 4.1.0-alpha-debug-log Protocol version: 10 Connection: Localhost via UNIX socket Client characterset: utf8 Server characterset: utf8 UNIX socket: /var/lib/mysql/mysql.sock Uptime: 2 days 17 hours 4 min 34 sec
Threads: 1 Questions: 79 Slow queries: 0 Opens: 36 Flush tables: 1 Open tables: 30 Queries per second avg: 0.000
--------------
mysql> describe emprunteurs;
+--------+-------------+-----------+------+-----+------------+-------+
| Field | Type | Collation | Null | Key | Default | Extra |
+--------+-------------+-----------+------+-----+------------+-------+
| num | varchar(6) | utf8 | | PRI | | |
| cb | varchar(20) | utf8 | | | | |
| nom | varchar(25) | utf8 | | | | |
| prenom | varchar(20) | utf8 | | | | |
| pwd | varchar(10) | utf8 | | | | |
| email | varchar(30) | utf8 | | | | |
| groupe | varchar(6) | utf8 | | | | |
| lins | varchar(4) | utf8 | | | | |
+--------+-------------+-----------+------+-----+------------+-------+
8 rows in set (0.00 sec)
mysql> select nom from emprunteurs where nom = 'BOUHASSOUN' LIMIT 1;
ERROR 1210: Wrong arguments to =
Note that using 'SET CHARACTER SET utf8' before doesnt resolve this problem.
Specifiying encoding before string in query works :
mysql> select nom from emprunteurs where nom = _utf8'BOUHASSOUN' LIMIT 1 ; +------------+ | nom | +------------+ | BOUHASSOUN | +------------+ 1 row in set (0.00 sec)
So, what i would like to known is if i've missed something obvious. If not, is this an already known bug, corrected ? if yes where could i find a patch ?
I've tryed current bk tree for 4.1.1 and hit another bug, related to compilation system. Everything compile fine, but the shared library are created with an incorrect name. For exemple : directory libmysql_r/.libs/ contains libmysqlclient_r.14.0.0 it should be libmysqlclient_r.so.14.0.0
But it's not only a filename problem : $ objdump -x libmysqlclient_r.14.0.0 | grep SONAME SONAME libmysqlclient_r.14
So the soname is wrong too, i think ld will not like this so i didn't install
and test it.
I've looked at Makefile.am/config.in/aclocal.m4 change since 4.1.0
(which compile and run fine on same system) but i didn't find anything
looking wrong. But i'm not a libtool expert.
Thanks in advance,
Regards, Gilles Magnier.
PS: Some additional informations about test platform :
utf8 bug reproduced on :
one RH9 and one rawhide, both up to date. MySQL 4.1.0-alpha recompiled from www.mysql.com source rpm. (only configure option changes : utf8 default charset)
current bk 4.1.1 Compile problem found on rawhide box only (not tested in RH9) :
gcc (GCC) 3.3.1 20030915 g++ (GCC) 3.3.1 20030915 ltmain.sh (GNU libtool) 1.5 (1.1220.2.1 2003/04/14 22:48:00) autoconf (GNU Autoconf) 2.57 automake (GNU automake) 1.7.7
-- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]
