This is the latest release before Hibernate Search 3.3 Final. Please test it before next tuesday as we have change some sensitive parts (for better performance).

This release comes with three major changes:

  • a complete revamp of the queuing algorithm. For complex domain models, you should see a big (as in night and day) performance improvement, in particular if you have many indexed associations (we have more to come by the way but that will be for 3.4).
  • a brand new Infinispan-based Lucene directory provider (I'm sure Sanne will blog about it)
  • a major documentation cleanup: now is the time to read it :)

We have also fixed a few issues and incompatibilities, in particular:

  • with Seam 2.2
  • with JBoss AS 6 (MassIndexer)

Check out the new release on's Maven repository or download the distribution. You can also read the documentation.

If you find an issue, you know the way.


PS: Many thanks to my partners in crime, Hardy and Sanne for this fight against Murphy's law today.

PPS: Many thanks to the Git and GitHub contributors, you saved our day.

10. Dec 2010, 09:53 CET | Link
Is there any chance to port Hibernate Search 3.3 to Hibernate Core 3.3.2? Core in version 3.6 brings a lot of regressions and unsolved problems. Additionally, migrating any complex app from 3.3.2 -> 3.5 or 3.6 is impossible. Lots of regressions, non-existent migration path and guide (Hibernate migration guide page to 3.3 -> 3.5 has not been provided for almost a year now..).

Having Hibernate Search 3.3 for the Core 3.3 branch would be great. The last version I'm capable to run is HS 3.1.1. What a lost...
10. Dec 2010, 11:23 CET | Link

Keeping a backport is not likely from our side, that's simply too much work), but if somebody from the community wants to do it, this is welcome.

By the way what regression are you experiencing between 3.3 and 3.6?

13. Dec 2010, 10:46 CET | Link
From the top of the list, as remembered from an attempt a month ago: silent and surprising exceptions from query result transformations using Hibernate-provided transformers on postgresql and strange exceptions from the HQL queries on postgresql in general (working perfectly on HC 3.3). Unfixable despite a solid amount of time spend on research and google.

And general impression of instability - something small broken here, something small broken there. No information what has been broken (no migration guide available for 3.3 -> 3.5 (3.6)). Minefield :(

It's a pity that the last stable and popular branch of HC - 3.3 sticks with HS 3.1.1... You guys have added a lot of functionality since then :(
13. Dec 2010, 16:27 CET | Link

Frankly, it's going to be hard to improve these stability gotchas if we don't have issues reported (ideally with unit tests). 3.6 comes with JPA 2 support which provides a lot of new features (even better standardized).

I am not saying that we are saints and react to every single JIRA opened but we try and stay afloat and we definitely emphasize both the next version in development (4 atm) and the latest stable release (3.6 atm).