™Milestone number 3 of the upcoming 6.0.0 release of RDF4J is now available for download.
RDF4J 6.0.0 is a major release of the RDF4J framework. This third milestone focuses on hardening the RDF 1.2 and SPARQL 1.2 work from the previous milestones, expanding store support, and fixing several release-readiness issues found during verification.
Please note that RDF4J 6.0.0-M3 requires Java 25.
This is the final planned milestone build before RDF4J 6.0.0. Once M3 has been verified, we will start planning the final release. Please try this build and let us know about regressions, upgrade problems, missing documentation, or any other surprises by opening a GitHub issue or starting a GitHub discussion.
Some of the highlights covered in this third milestone:
RDF4J 6.0.0-M3 expands RDF 1.2 and SPARQL 1.2 release coverage across the model, query, protocol, parser, writer, and store layers (PR-5886). This includes additional coverage for base-direction-aware literals, triple terms, query result serialization, HTTP transaction serialization, and SPARQL 1.2 query/update compliance.
The LMDBStore now supports RDF 1.2 triple terms and literals with base direction (GH-3723, PR-5859). NativeStore received transaction-status and concurrency fixes while preserving its WAL-based recovery work (GH-5876, PR-5877).
The RDF 1.2 conversion helper names have been corrected (GH-5865, PR-5866), and literal-only comparison helpers no longer apply triple-term comparison logic where it does not belong (GH-5849, PR-5851).
RDF/XML now has native RDF 1.2 support in the parser, writer, and pretty writer (GH-5878, PR-5882). Turtle and N3 writers can now emit directive-style output where applications need it (GH-5852, PR-5854).
The generated SPARQL parser has also been tuned to reduce expensive LOOKAHEAD bounds and clean up generated parser code (GH-5856, PR-5857, PR-5885).
SPARQL LATERAL support has landed for parsing, algebra, evaluation, rendering, and optimizer handling (GH-4315, PR-5875).
FedX now has initial RDF 1.2 support for triple-term query patterns (GH-5870, PR-5871), and its connection interface contract has been clarified so exceptions are handled consistently by upstream callers (GH-5828, PR-5862).
LMDB sketch-based planning is now disabled by default while the estimator and planning work is
being hardened for the final release
(GH-5868,
PR-5869). A MemoryStore regression in
query evaluation involving BIND has also been fixed by reverting the standard optimizer
filter pre-pass
(GH-5860,
PR-5864).
XML parsing defaults have been hardened, including transaction XML parsing, RDF/XML, TriX, and SPARQL XML result parsing. Workbench server selection now rejects unsafe server paths and dot-segment escapes (GH-5824, PR-5825).
The Apache HttpComponents 5 backend now makes request construction overridable, allowing subclasses to cancel long-running HTTP requests such as SPARQL queries (GH-5879, PR-5881).
This milestone is intended to shake out the final release issues. If you test RDF4J 6.0.0-M3 and find a bug, an upgrade concern, missing documentation, or behavior that needs discussion, please open a GitHub issue or start a GitHub discussion.
Eclipse RDF4J™ is a powerful Java framework for processing and handling RDF data. This includes creating, parsing, scalable storage, reasoning and querying with RDF and Linked Data. It offers an easy-to-use API that can be connected to all leading RDF database solutions. It allows you to connect with SPARQL endpoints and create applications that leverage the power of linked data and Semantic Web.
