dajiaohuang opened a new pull request, #56: URL: https://github.com/apache/datafu/pull/56
## Summary - Add a binary-compatibility test path that builds one release-config DataFu Spark JAR and runs the existing tests against it, excluding main-source output from test compile/runtime classpaths. - Add an explicit latest-patch matrix for supported Spark minor lines (3.0.3, 3.1.3, 3.2.4, 3.3.4, 3.4.2, 3.5.9) while retaining the existing source-compatibility job. - Avoid a Spark 3.0 linkage failure by constructing `AggregateExpression` directly instead of relying on `AggregateFunction.toAggregateExpression` overloads whose binary signatures differ across supported Spark versions. ## Validation - Java 8 release-config `:datafu-spark:assemble` passed. - On the release JAR, `DataFrameOpsTests.dedup2_by_int` and `UdafTests.minKeyValue` passed against Spark 3.0.3. - Full Spark 3.0.3 suite: 25/30 passed; three Hive-table tests could not complete on this Windows host because Hadoop `winutils` is unavailable, and two Python bridge tests could not resolve `pyspark_utils` on Windows. - Spark 3.5.9 targeted test setup stopped before test methods because this Windows host lacks `HADOOP_HOME`/`winutils`. - `bash -n` and `git diff --check` passed. The other minor versions have not been locally run. CI has not been queried. AI assistance: OpenAI Codex (`gpt-6-luna`) was used to prepare this contribution; the commit includes a `Generated-by` provenance line. -- 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]
