lsh1215 commented on PR #16300: URL: https://github.com/apache/lucene/pull/16300#issuecomment-4913034229
Hi @dweiss, thanks for filing this issue. You reviewed the earlier attempt in #15118, so I thought you might want to take a look at this one. It takes a different approach from the range-capping in #15118. Instead of shrinking the randomized inputs, it goes after the memory pressure itself. Two changes: - Disable query caching during the equivalence check, so cached doc-ids don't stick around across iterations. - Use a bitset comparison on the no-score path instead of the materialized `TopDocs`. The original stress shape stays the same. I reproduced the OOM under a 16MB heap and it passes after the change. Let me know what you think when you get a chance. -- 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]
