Topic: MySQL wants out of the GPL…

[Steffen]

www.devside.net wrote:

It looks like MySQL has no more plans of providing the updated binaries [and source] to the ‘Community Server’ edition of the MySQL Server.

Currently the newest release version is 5.0.30, while Downloads only lists 5.0.27. The situation is the same under the changelog of the Manual.

MySQL is the copyright holder on the software, and is not bound to the license of any previous release, so this is all perfectly valid.

While this will not effect the Linux crowd much, this will be a problem for WAMP distributions as MySQL has always been difficult to compile on Windows platform.

Dropping the ‘Community Server’ edition in exchange of only providing to payed customers will also backfire.

A little more info can be found in this blog entry… http://www.mysqlperformanceblog.com/2006/12/29/where-to-get-recent-mysql-version/

[Brian]

I guess it may be time to investigate the advantages of PostgreSQL.

If they are going to charge, why not use the free version of MsSQL, or Postgre, or just never update.... Few web sites need any of the advantages that version 5 holds over version 4, such as stored procedures.

I have noticed also that they are really updating their web site to be more commercially appealing. The last thing they have left to do as I see it, is to drop the Google Ads, then you know where they are headed.

Sad, but what some Linux Open Source users who champion free software don't realize, if there is demand and not a real alternative, it will not always be free.

[Brian]

Some additional insight may be found here:

http://news.zdnet.com/2100-3513_22-6147312.html

[tdonovan]

Kaj Arnö (MySQL community manager) tries to reduce the community's concern in his blog: MySQL Community Server recap.

In any case, there are signs that MySQL will not be so quick to drop the community edition.
That's what they say anyway. I hope it is what they actually do.

-tom-

For what it's worth - the MySQL Windows source for versions 5.0.30 and 5.0.32 are now at ftp://ftp.mysql.com/pub/mysql/src/

Both versions build OK** with Microsoft VC++ 2005 Express.
The source contains a VC++ 7.1 mysql.sln file which the VC++ 2005 Conversion Wizard converts without problems,
so perhaps it is no longer too difficult to compile MySQL on Windows.

** One small source change is necessary to build with VC++ 2005. In file extra\yassl\taocrypt\include\integer.hpp remove or comment out lines 65-69 to prevent the definition of the two symbols SSE2_INTRINSICS_AVAILABLE and TAOCRYPT_MM_MALLOC_AVAILABLE. This change will have a small (extremely small!) performance impact on SSL connections.

[tdonovan]

One more little thing you should be aware of: Bug 21635 is not fixed in MySQL 5.0.32, contrary to the notes in the bug.

If you test it with mysql_client_test.exe, this program will fail on the very last test - test_bug21635.

Looking at the bug, I cannot imagine why anyone would care!
The bug is if you SELECT MAX(fld) FROM myTbl, the metadata for MAX(fld) says its original-table-name was "myTbl" instead of being empty.
The specs say that calculated fields should never have an original table name in their metadata. Ditto for original-field-name.

I commented out line 15737 in tests\mysql_client_test.c so I could run the MySQL test suite without error.