mundaym opened a new pull request #29407: URL: https://github.com/apache/spark/pull/29407
In order to produce consistent results from SizeEstimator the tests override some system properties that are used during SizeEstimator initialization. However there were several places where either the compressed references property wasn't set or the system properties were set but the SizeEstimator not re-initialized. This caused failures when running the tests with a large heap build of OpenJ9 because it does not use compressed references unlike most environments. ### What changes were proposed in this pull request? Initialize SizeEstimator class explicitly in the tests where required to avoid relying on a particular environment. ### Why are the changes needed? Test failures can be seen when compressed references are disabled (e.g. using an OpenJ9 large heap build or Hotspot with a large heap). ### Does this PR introduce _any_ user-facing change? No ### How was this patch tested? Tests run on machine running OpenJ9 large heap build. ---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
