rdtr commented on issue #12892: URL: https://github.com/apache/gluten/issues/12892#issuecomment-5967042444
@LuciferYang @baibaichen I can on this. Before I start, I'd like to check the approach. A PR that adds the Spark 3.4 and 3.5 jobs can't run them through `/ansi-test` before it's merged, so I can't find out what fails beforehand. My plan: 1. Add Spark 3.4 and 3.5 jobs (backends-velox and spark-ut) to `velox_backend_ansi.yml`, mirroring the 4.0 jobs with `spark.gluten.sql.ansiFallback.enabled=false`. No test changes. 2. After it's merged, run `/ansi-test` to see what fails on 3.4 and 3.5, then fix the failures or exclude them with a reason in follow-up PRs. That includes enabling the ANSI-on cast suite. Step 1 may expose a number of new failures in this workflow. It only runs on demand, so it doesn't affect the regular CI, and the 4.0 and 4.1 jobs already have failures being worked on. Would that be OK, or would you prefer the new jobs to land together with the exclusions? -- 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]
