Bug #2052 escaping charactor in 2byte charactor-encodings.
Submitted: 8 Dec 2003 19:46 Modified: 28 Mar 2014 13:40
Reporter: Yoshikazu Hashimoto Email Updates:
Status: Not a Bug Impact on me:
None 
Category:Connector / J Severity:S3 (Non-critical)
Version:3.0.9 OS:Windows (Windows, Java)
Assigned to: Alexander Soklakov CPU Architecture:Any

[8 Dec 2003 19:46] Yoshikazu Hashimoto 
Description:
Version 3.0.9-stable includes the [Fix for BUG#879, double-escaping of '\' when charset is SJIS or GBK and '\' appears in non-escaped input].
I need the driver to be fixed same problem when I use other 2byte charactor encoding converters, ie. MS932 or Windows-31J.

How to repeat:
Set the database url to be following.

jdbc:mysql://host/database?user=root&useUnicode=true&characterEncoding=MS932

Suggested fix:
To incorporate same logic as SJIS.
[8 Dec 2003 20:36] Yoshikazu Hashimoto 
FYI, "MS932" and "Windows-31J" is the name of the same encoding converter that implements the Japanese localized MS-Windows's native charactor encoding.

Since the "SJIS" does not equal to MS-Windows's native encoding, driver cannot properly handle a few charactors that was inserted from the native applications (i.e. MS-Access) through the mysql-ODBC connector if we use "SJIS", so we MUST set the driver to use MS932 (or Windows-31J) rather than SJIS.

This is the reason why I cannot be satisfied by the previous bug fix for "SJIS".
[23 Dec 2003 23:32] Yoshikazu Hashimoto 
Dear Sir,

PLEASE HOLD THIS REPORT!

I informed at a mailing list that the event might be due to inadequate sequence of the setup procedure for MySQL Server. I will confirm and inform.
[2 Feb 2004 18:09] Yoshikazu Hashimoto 
Please close this report.

Each database seems to have a charset property that is copied from the system valiable when the database is built. If this is correct, my mistake was, that I set the sjis charset to system valiable after I built the database.

Thanks a lot.
[28 Mar 2014 13:40] Alexander Soklakov 
Closing as Not a Bug