gsmiller commented on PR #848: URL: https://github.com/apache/lucene/pull/848#issuecomment-1112187182
@gautamworah96 thanks for looking! The mismatch in the original Jira was a slightly different check, which is the one now set to a delta of `0`. The fix proposed here is to ensure the floats are summed in the exact same order, which leads to identical results. The mismatch originated because the floats were being summed in different orders. Instead of chasing a delta that will work for "all" cases (which may be difficult since the test is random), I propose fixing it in this way. There's still a different check that does require a delta because it's difficult to ensure the test case can match the same ordering used internally by the faceting implementation. That's the one that you might be seeing that's still set to 1? I think this is acceptable because it's only 3 floats being summed in this case and the "drift" will be much smaller. -- 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: issues-unsubscr...@lucene.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org For additional commands, e-mail: issues-h...@lucene.apache.org
