pan3793 opened a new pull request, #59338: URL: https://github.com/apache/spark/pull/59338
### What changes were proposed in this pull request? Upgrade `zstd-jni` from 1.5.7-16 to 1.5.7-22 and regenerate `dev/deps/spark-deps-hadoop-3-hive-2.3`. The public API is unchanged; on JDK 22+ the jar now loads the FFM implementation of the native bindings from `META-INF/versions/22`, older JDKs keep JNI. ### Why are the changes needed? Keep the bundled compression library current. ### Does this PR introduce _any_ user-facing change? No. ### How was this patch tested? - `dev/test-dependencies.sh --replace-manifest` (diff limited to the zstd-jni line). - `ZStandardTPCDSDataBenchmark` on JDK 25 (FFM) and JDK 17 (JNI): on par. ### Notes There is a potential perf regression under investigation: level-1 compression is about 6% slower than 1.5.7-16 on TPC-DS SF1 text data, on both JDK 25 and JDK 17, so it is not related to the FFM port. Reported upstream at https://github.com/luben/zstd-jni/issues/427 ### Was this patch authored or co-authored using generative AI tooling? Generated-by: DeepSeek Flash -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
