Hi Hossein:

On Thu, 19 Jun 2008, Hossein Rafighi wrote:
>    - character_set_database: latin1

This is not good, you should have obtained utf8 there.  See below.

>    - collation_database: latin1_swedish_ci
>    - collation_server: latin1_swedish_ci

Ditto here.

> Yes. That is correct: --default-character-set="utf8"

This is not really necessary, the MySQL server can run with default
Latin-1 settings, only the Invenio database must run in UTF-8.

Here is the output I obtain:

   $ echo "SHOW CREATE DATABASE cdsinvenio" | /opt/cds-invenio/bin/dbexec
   Database        Create Database
   cdsinvenio CREATE DATABASE `cdsinvenio` /*!40100 DEFAULT CHARACTER
   SET utf8 */

   $ echo "SHOW CREATE TABLE idxWORD09F" | /opt/cds-invenio/bin/dbexec
   Table   Create Table
   idxWORD09F CREATE TABLE `idxWORD09F` (\n `id` mediumint(9) unsigned
   NOT NULL auto_increment,\n `term` varchar(50) default NULL,\n
   `hitlist` longblob,\n PRIMARY KEY (`id`),\n UNIQUE KEY `term`
   (`term`)\n) ENGINE=MyISAM AUTO_INCREMENT=38790 DEFAULT CHARSET=utf8

Do you get the same, or do you obtain latin1 there as I suspect?

If the latter, did you create your database with an explicit UTF-8
charset definition, as hinted by inveniocfg?

  CREATE DATABASE cdsinvenio DEFAULT CHARACTER SET utf8;

If not, you will have to fix this.

> * MySQLdb version: 1.2.1_p2

This should be okay, e.g. on a Gentoo test box I'm using MySQLdb
1.2.1_p2 with MySQL 5.0.54-log.  FWIW, on our SL4 boxes we use the
latest MySQLdb 1.2.2 with MySQL 4.1.x.  You may perhaps want to upgrade
your MySQLdb, although it should not be really required.

Best regards
-- 
Tibor Simko ** CERN Document Server ** <http://cds.cern.ch/>

Reply via email to