2pk03 opened a new pull request, #638: URL: https://github.com/apache/incubator-wayang/pull/638
Port object-file IO onto a safe codec ------------------------------------ - add `ObjectFileSerializationMode` + helper to switch between legacy Java serialization and a JSON-based format, with unit tests covering both paths - rework Java/Flink/Spark object-file sources/sinks (and Flink’s output format) to use the shared serializer, log deprecation warnings for the legacy mode, and stop calling raw `ObjectInputStream` - update Spark’s SequenceFile readers/writers to chunk records via RDD transformations and share logic with the new serializer Motivation: the legacy path deserializes untrusted SequenceFile payloads via `ObjectInputStream.readObject`, letting an attacker ship a gadget chain that runs arbitrary bytecode inside the Wayang JVM. That RCE can be used to execute system commands, exfiltrate data, or tamper with jobs, so we move to JSON by default and require explicit opt-in for the old codec. ========================== IMPORTANT ========================= Before we merge test the patch, ideally in a setup with source and sink platforms. Compilation and tests when through in my setup. ``` mvn -pl wayang-commons/wayang-basic -Dtest=org.apache.wayang.basic.operators.ObjectFileSerializationTest test [INFO] Building Wayang Basic 1.1.1-SNAPSHOT [WARNING] 2 problems were encountered while building the effective model for org.apache.yetus:audience-annotations:jar:0.5.0 during dependency collection step for project (use -X to see details) [INFO] Tests run: 2, Failures: 0, Errors: 0, Skipped: 0 [INFO] [INFO] ------------------------------------------------------------------------ [INFO] BUILD SUCCESS [INFO] ------------------------------------------------------------------------ [INFO] Total time: 2.683 s [INFO] Finished at: 2025-12-08T19:47:45+01:00 [INFO] ------------------------------------------------------------------------ ``` -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
