Hibernate ORM 6.0.0.Alpha9 released

Posted by    |      

Hibernate ORM 6.0 Alpha9 has just been released.

The main design goal for 6.0 is to improve even more Hibernate’s through-put performance. High-load performance testing showed Hibernate’s approach of reading values from ResultSet by name to be its most limiting factor in scaling throughput. At its most basic, 6.0 is all about changing from its old strategy of read-by-name to read-by-position. But that simple goal has a lot of ramifications.

With Alpha9, the APIs are really starting to stabilize which is great.

A lot of work has been done in the Alpha8 and Alpha9 releases to focus on fixing test failures related to all the changes and getting the full testsuite back up and running. We are currently down to 17% failure rate, though at least some of those failures will be simple, small things we have not implemented yet. For example, we recently fixed a bunch of errors related to support for non-id generated attributes (@CreationTimestamp, etc) because the code that reads back the generated values had not yet been converted to use read-by-position. So long story, we are progressing nicely toward full test coverage.

Most of the work done was simply "tasks" to integrate the read-by-position changes, so the work goes well beyond the Jira changelog. This Alpha9 release includes:

  • [HHH-1615] - Support for entity alias in the group by clause

  • [HHH-14495] - Use table name, rather than entity name, as base for per-entity implicit sequence naming

  • [HHH-14745] - Add FunctionContributor contract

  • Support for (non-id) value generation + documentation improvements

  • Support for composite-valued query parameters

  • Support for lock handling

  • Support for indexed collection element access through array subscript syntax

  • Support usage of boolean expression in predicate context e.g. where contains(e.field, 'token')

  • Introduction of the FILTER clause for aggregate functions

  • Fix joined subclass hql wrong join type for associations

  • Fix issue with join type for subclass tables

  • Fix issue with sql rendering of null discriminators

  • Fix wrong creation of inner join

  • Fix throw exception when more than one row with the given identifiers

  • Fix SmqBagJoin, SqmSetJoin and SqmListJoin throwing NotYetImplementedException

  • Fix collection batch loading

  • Fix many to many Order column using wrong table

  • Fix AbstractEntityPersister#findSubpart returning the wrong ModelPart when the Entity has an attribute named id that it is not the identifier

  • Fix create a new instance while a corresponding one is loading

  • Fix multi value parameters expansion for SelfRenderingSqmFunction

  • Fix ManyToOneType#isModified() causing unnecessary update execution

  • Fix java.sql.Date attribute assigned wrong BasicType

  • Fix close connection when a transaction is not active and connection handling mode is not equal to AFTER_STATEMENT

  • Fix StatelessSession#uniqueResult() not clearing the temporary PersistenceContext

  • Fix ScrollableResultsImpl not initializing non lazy collections

hibernate-community-dialects artifact

One change in particular that is worth a discussion is the decision to move certain Dialects out of the main hibernate-core artifact into a separate hibernate-community-dialects artifact. These Dialects are ones that are not routinely tested and so are moved to this new artifact in order to set proper expectations with regard to using them.

This follows our established guidelines about what it means for a Dialect to be considered supported. All supported Dialects exist in the hibernate-core artifact. The process to add a new supported Dialect essentially means someone taking responsibility for setting up CI runs and agreeing to fix any reported issues. Contact us on chat or the hibernate-dev mailing list if you wish to champion a Dialect for support.

Gradle Enterprise

I want to give a huge thanks to the Gradle team for getting Hibernate set up on Gradle Enterprise hosting and giving us lots of help to make huge improvements to the build. Especially Nelson Osacky and Etienne Studer - without you guys we never would have figured all of this out.

I highly recommend having a look at the Gradle Enterprise offering for organizations wanting to get the most productivity, performance and reproducibility out of their builds. See https://gradle.com/ for details

More information

Also check out the release page. Join BinTray and watch the release repository for notifications of releases

To get in touch, use the usual channels as discussed on https://hibernate.org/community/


Back to top