Apache OpenNLP 3.0.0-M3
This release focuses on security hardening, new NLP capabilities, and dependency maintenance.
Security Fixes
Three security issues are addressed in this release (also backported to 2.5.9).
XXE in DictionaryEntryPersistor (OPENNLP-1819)
The DictionaryEntryPersistor previously used a SAXParserFactory that did not enable secure processing or disable DTD handling, leaving external entity resolution active. A malicious dictionary file could exploit this for local file disclosure or SSRF before any dictionary entry was processed.
The parsing path is now aligned with the project's existing XmlUtil helper, which properly sets FEATURE_SECURE_PROCESSING and disallow-doctype-decl.
Arbitrary Class Instantiation in ExtensionLoader (OPENNLP-1820)
ExtensionLoader.instantiateExtension() performed its isAssignableFrom type check after Class.forName() had already executed the target class's static initializer, allowing a crafted model archive to trigger the static initializer of any class on the classpath.
The fix introduces a package-prefix allowlist consulted before Class.forName() is invoked:
- Classes under
opennlp.*remain permitted by default. - Other packages must be opted in via
ExtensionLoader.registerAllowedPackage(String)or theOPENNLP_EXT_ALLOWED_PACKAGESsystem property (comma-separated list).
OOM via Unbounded Array Allocation in AbstractModelReader (OPENNLP-1821)
getOutcomes(), getOutcomePatterns(), and getPredicates() read attacker-controlled 32-bit count fields from binary model streams and passed them directly to array allocations. A crafted .bin file could trigger an immediate OutOfMemoryError and crash the JVM.
Each count is now bounded (default 10,000,000, configurable via -DOPENNLP_MAX_ENTRIES=<n>), with negative or oversized values failing fast via IllegalArgumentException.
⚠️ For all three issues, users who cannot upgrade immediately should restrict input (dictionary and model files) to trusted sources only.
New Features & Improvements
- Roberta-based model support via ONNX — OPENNLP-1518 (#998)
- Byte Pair Encoding (BPE) tokenization — OPENNLP-1220 (#1011)
Parse.createFromTokens()convenience method for tokenized input — OPENNLP-53 (#1012)- Thread-safe ME classes by eliminating shared mutable instance state — OPENNLP-1816 (#1003)
What's Changed
- Apache OpenNLP 3.0.0-M2 by @mawiesne in #996
- OPENNLP-1518: Roberta-based Models - Add support for utilization via Onxx by @rzo1 in #998
- OPENNLP-1817: Update log4j2 to 2.25.4 by @dependabot[bot] in #1001
- Regenerated NOTICE file after dependency changes by @github-actions[bot] in #1009
- OPENNLP-1818: Update zlibsvm-core to 3.0.0 by @dependabot[bot] in #1000
- Regenerated NOTICE file after dependency changes by @github-actions[bot] in #1010
- OPENNLP-53: Add Parse.createFromTokens() for convenient tokenized input by @mawiesne in #1012
- OPENNLP-1220: Add support for Byte Pair Encoding (BPE) by @mawiesne in #1011
- Bump com.ruleoftech:markdown-page-generator-plugin from 2.4.2 to 2.4.3 by @dependabot[bot] in #1016
- Bump peter-evans/create-pull-request from 8.1.0 to 8.1.1 by @dependabot[bot] in #1014
- Bump actions/cache from 5.0.4 to 5.0.5 by @dependabot[bot] in #1017
- OPENNLP-1819: Align DictionaryEntryPersistor XML parsing with XmlUtil by @rzo1 in #1019
- OPENNLP-1816: Make ME classes thread-safe by eliminating shared mutable instance state by @krickert in #1003
- OPENNLP-1822: Update ONNX runtime to 1.25.0 by @dependabot[bot] in #1024
- Regenerated NOTICE file after dependency changes by @github-actions[bot] in #1025
- OPENNLP-1821: Prevent OutOfMemory Due To Huge Array Allocation by @subbudvk in #1022
- OPENNLP-1820: Restrict ExtensionLoader to allowlisted package prefixes by @subbudvk in #1021
New Contributors
Full Changelog: https://issues.apache.org/jira/secure/ReleaseNote.jspa?projectId=12311215&version=12356813
