Re: [rdf4j-dev] RDF4J 6

[Piotr Sowiński <piotr@xxxxxxxxxxxxx>]

Hi,

I'm very much in favor of moving on to RDF4J 6 right after 5.3.

For RDF4J 6 my first wish is to upgrade to as new JDK version as possible (preferably 25), to allow us to use the many new Java features to further improve performance.

I would also suggest to consider if any changes to public APIs in the query engine are needed (around BindingSets) to support processing queries with lower overhead (e.g., raw value IDs), or with batches of bindings (vectorized).

We could also reconsider mutability of some classes in the engine (e.g., values, statements, binding sets). Moving to immutable records today could already offer small performance improvements. But more importantly, this would be a first step to supporting Valhalla value classes, when they finally are released.

I think some time ago there was some work in the engine to support iterations with a specific order, but this is still experimental – this could be taken further, perhaps?

On Mon, 15 Dec 2025 at 22:22, Håvard Ottestad via rdf4j-dev <rdf4j-dev@xxxxxxxxxxx> wrote:

Hi everyone!

I’m wondering what people think about having 5.3.0 be the last minor release of RDF4J and move to 6.0.0 on the develop branch after that?

Anything we want to remove in 6.0.0 needs to be deprecated in 5.3.0. I’m going to deprecate the entire Solr sail for removal because Solr is very far behind on updates and will hold us back from updating to the latest Lucene and Jakarta EE versions.

I’ve release 5.2.2 today and plan to publish a milestone build of 5.3.0 as soon as I can.

Cheers,
Håvard




_______________________________________________
rdf4j-dev mailing list
rdf4j-dev@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/rdf4j-dev

--

Piotr Sowiński

CTO & Co-Founder @ NeverBlink
https://neverblink.eu