LinSimon-901101 opened a new pull request, #6381: URL: https://github.com/apache/datafusion-comet/pull/6381
## Which issue does this PR close? Closes #6378. ## Rationale for this change The cache pruning predicates added in #6262 do not exercise integer min/max boundaries. An integer upper bound that is one too small can therefore pass all three writer cases. ## What changes are included in this PR? Add `n = 5` to the shared predicate list. This value equals both bounds of its fixture batch, so the existing assertions check IDs 20–23 and exactly one four-row batch across native Arrow, Spark columnar, and row input. ## How are these changes tested? - Rebuilt the native library on main at `4ca04b155` and ran `CometInMemoryCacheSuite`, `CometInMemoryCacheKryoSuite`, and `CometInMemoryCachePruningSuite` on Spark 4.1.3 / JDK 21: 61 passed, 0 failed, 0 canceled. Spotless and Scalastyle passed. - Fault injection on #6262's head: reducing the integer upper bound by one leaves the original three writer tests passing, while adding `n = 5` makes all three fail with missing IDs 20–23. The serializer was restored before the successful normal test runs. -- 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]
