Overall I find the achivement with this release quite impressive. 5.6 is the beginning of a new way of *thinking the server*, I think. What we have is:
1) Lots of new features – in particular I note
* Security enhancements (but maybe not everything here is as it should be – read Kolbe Kegel’s comment)
* Memcached plugin for InnoDB
* Multithreaded replication on slaves
* Performance_Schema enhancements
.. and more
2) Bug fixes.
It has been a problem with MySQL for many years that only a fraction of reported and verified bugs got fixed in reasonable time. And those not fixed within 3-6 months had a tendency to be forgotten. This has been very annnoying.
With this release a lot of bugs many years old are fixed – including a handful of those I reported. This is highly appreciated.
3) Code refactoring.
I am not a C-developer myself but follow various discussions in mailing lists etc. The *lack of modularity* in MySQL code is often claimed to be a problem. It is actually difficult to do a fix and ensure that the fix does not have unwanted ‘side-effects’ (and even though I am not a C-developer, I am able to read code basically. When I have opened some source files I often found too many ‘fingerprints’ of individual coders there – such as non-consistent indentation and commenting styles etc.). From the discussions I am referring to, I understand that there has also been improvements in this respect with 5.6.
A problem – or side-effect – resulting from that is that this is then making life more difficult for ‘forks’ (it would not be simple for MariaDB to shift the MySQL code base from 5.5 to 5.6 for instance), but still it is a good development – even if part of the story that from Oracle’s side there is a desire to make life harder for ‘forks’.
So overall I would say KUDOS to the MySQL team for this release. A release of this quality eradicates fears (as expressed by various people at the time Oracle aquired Sun) that Oracle would ‘park MySQL at the sideline’ and not continue it seriously.
What I fear however is that Oracle is going to push the 5.6 tree to a premature GA release (BTW: did we ever see this before with MySQL?). One particular reason for this fear is that I noted in bugs.mysql.com that some bugfixes were committed to MySQL 5.7.0. Nothing is known about 5.7.0 at this point of time (no release notes available), but it indicates to me that 5.6 is coming close to a feature-freeze and a GA release.
It should be GA-quality and not only named as a ‘GA’.