charliec05 opened a new pull request, #58401: URL: https://github.com/apache/spark/pull/58401
### What changes were proposed in this pull request? Restore each local dependency's URI fragment after Kubernetes uploads it to the configured remote file system. The existing fragment restoration for archives is generalized to JARs, files, and Python files while retaining the archive-specific behavior of removing the fragment before the physical upload. A regression test covers local JAR, file, and Python-file aliases and checks both the uploaded physical name and the preserved fragment. ### Why are the changes needed? Spark uses URI fragments in dependency paths as application-visible aliases, for example `--files /tmp/query-random.sql#query.sql`. The Kubernetes submission path kept the uploaded physical filename but dropped `#query.sql`, so the driver and executors could not resolve the dependency by its requested alias. ### Does this PR introduce _any_ user-facing change? Yes. Local dependencies uploaded during Kubernetes submission now retain their `#alias` fragments in `spark.jars`, `spark.files`, and `spark.submit.pyFiles`. ### How was this patch tested? - Verified the new regression test fails before the fix and passes after it. - `build/sbt -Pkubernetes "kubernetes/Test/testOnly org.apache.spark.deploy.k8s.features.BasicDriverFeatureStepSuite"` (25 tests passed) - `dev/lint-scala` Generated-by: OpenAI Codex (GPT-5) -- 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]
