Hi,

MySQL Connector/Net 1.0.2, a new version of the fully-managed, ADO.Net provider for the popular MySQL database system has been released.

Version 1.0.2 is the first gamma release of the original ByteFX MySQL provider acquired earlier this year. This release is suitable for use with any MySQL version including MySQL-4.1 or MySQL-5.0.

It is now available in source and binary form from the Connector/Net download pages at http://dev.mysql.com/downloads/connector/net/1.0.html and mirror sites (note that not all mirror sites may be up to date at this point of time - if you can't find this version on some mirror, please try again later or choose another download site.)

========================= Attention ====================================== There is a known problem where data packets larger than 16M are not handled properly when the connection is using compression. There is no need to file bug reports on this issue as the test suite already identifies this as a problem.

Illegal Mix of Collations problem

--------------------------------- Many people have been wondering about the illegal mix of collations problem. I will explain the problem here and explain how 1.0.2 handles it. Starting with MySQL 4.1, it is possible to specify what character set SQL will be sent to the server in. Upon connection, the server identifies what character set is the "default" however tables and fields can use whatever character set is desired.

No matter what character set is used for the connection, anytime SQL that uses a different character set (think multi-language projects here) is used, then the user will need to use functions like CONVERT to make sure the SQL is understood by the server properly. The user will also need to specify an appropriate connection character set, such as utf8. This is done by including "charset=xxx" on the connection string.

Now to the reason you are seeing this "bug". Versions 1.0.0 and 1.0.1 of the provider always used utf8 for the connection. This means that literal elements of SQL ( WHERE name = 'this is the literal' ) would be treated as utf8 by the server. When it compared these to latin1 columns, for example, it complained. This is not really a bug. There's just several different approaches to handling the situation and we are trying to handle it in such a way that causes the least headaches for users.

Hopefully 1.0.2 will ease some of the migration pains that so many of you have been posting about. Good luck!

-Reggie

- From the changelog:

11-15-04 - Version 1.0.2-gamma

- Fixed problem with MySqlBinary where string values could not be used to update extended text columns - Fixed Installation directory ignored using custom installation (Bug #6329) - Fixed problem where setting command text leaves the command in a prepared state - Fixed double type handling in MySqlParameter(string parameterName, object value) (Bug #6428) - Fixed Zero date "0000-00-00" is returned wrong when filling Dataset (Bug #6429) - Fixed problem where calling stored procedures might cause an "Illegal mix of collations" problem. - Added charset connection string option - Fixed #HY000Illegal mix of collations (latin1_swedish_ci,IMPLICIT) and (utf8_general_ (Bug #6322) - Added the TableEditor CS and VB sample - Fixed Charset-map for UCS-2 (Bug #6541) - Updated the installer to include the new samples - Fixed Long inserts take very long time (Bug #5453) - Fixed Objects not being disposed (Bug #6649) - Provider is now using character set specified by server as default