zsxwing edited a comment on pull request #33784: URL: https://github.com/apache/spark/pull/33784#issuecomment-902073060
@LuciferYang Thank you for submitting the revert PR. @sarutak @dongjoon-hyun As I mentioned in [the previous PR](https://github.com/apache/spark/pull/31517#issuecomment-897405632), adding a popular library to Spark's classpath directly is risky. @LuciferYang 's [suggestion](https://github.com/apache/spark/pull/31517#issuecomment-897430503) to shade Caffeine is a good solution. But it sounds not worth because: - There is no evidence to show performance improvement in practice. The code path using cache is not on any hot path, and the difference between Guava Cache and Caffeine is probably not noticeable in practice. - Extra work to shade the following three libraries (Need to verify this approach is working): ``` caffeine/2.9.2//caffeine-2.9.2.jar checker-qual/3.10.0//checker-qual-3.10.0.jar error_prone_annotations/2.5.1//error_prone_annotations-2.5.1.jar ``` - Increase the maintenance burden when backporting a patch that touches the changed code. Do I miss any benefits provided by Caffeine? -- 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]
