Release Comet 0.9.0 (June/July 2025)

[andygrove]

What is the problem the feature request solves?

I would like to start planning the Comet 0.9.0 release. Let's use this issue to co-ordinate on any issues or PRs to resolve for the release.

Issues: https://github.com/apache/datafusion-comet/issues?q=is%3Aissue%20state%3Aopen%20milestone%3A0.9.0

Main issues we want to include in this release:

• feat: Upgrade to official DataFusion 48.0.0 release #1877
• fix: Remove COMET_SHUFFLE_FALLBACK_TO_COLUMNAR hack #1865
• Fix failed Spark SQL tests due to shuffle enabled #231
• [EPIC] Spark SQL test failures when Comet JVM shuffle is used #1254
• feat: support RangePartitioning with native shuffle #1862
• feat: Support hadoop s3a config in native_iceberg_compat #1925
• feat: Change default value of COMET_NATIVE_SCAN_IMPL to auto #1933
• feat: Encapsulate Parquet objects #1920
• perf: Add COMET_RESPECT_PARQUET_FILTER_PUSHDOWN config #1936
• docs: Documentation updates for 0.9.0 release #1981
• Spark SQL test failures in native_iceberg_compat mode #1542
• Run benchmarks to confirm no major regressions

Once the source release is available:

• Publish jars to Maven
• Push new Docker images to Docker Hub

Describe the potential solution

No response

Additional context

No response

[andygrove]

I made a start on a Comet 0.9.0 blog post. Feedback is welcome, as always.

https://docs.google.com/document/d/1l3OnFJfJDuVhQg_jzHOnP2wGnyt8YQcy_YWljCCM7Hs/edit?usp=sharing

[andygrove]

I plan on creating the first release candidate on Monday

[andygrove]

I have started the release vote: https://lists.apache.org/thread/d2c9lmmym78w7pm4o9xf4l76p2dl1k34

[andygrove]

Blog post PR: apache/datafusion-site#78

[andygrove]

The release has been completed, except for uploading Docker images