msfroh commented on PR #16222: URL: https://github.com/apache/lucene/pull/16222#issuecomment-4674164687
> Thanks @txwei! I think this is a pretty obvious win over any minor regression we'd get from a less accurate cost estimate. Oof... the earlier issue showed that we can take a fairly significant performance hit when the automaton matches few terms, since we lose caching. I would not call it a minor regression. If we can hook into the automaton execution and put a bound on its work during the planning phase (i.e. limit the number of visited states), I think we could satisfy both cases. I can look into what our options are there. -- 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]
