andrross commented on PR #12374: URL: https://github.com/apache/lucene/pull/12374#issuecomment-1597890665
+1 to not having two ways to customize slicing. > the approach to override a method to compute slices given a list of segments looks more user-friendly to me than having to compute slices up-front? @jpountz I agree. I think the basic problem here is that the protected method is called from the constructor of the parent class, which really limits the flexibility and can lead to unexpected behavior. The simplest approach is to just compute the slices on every call to `#search`. If we want to be more defensive about possible performance regressions, then lazy computation + caching like you said is the way to go (i.e. memoization). Is there any utility in Lucene like this [memoizing supplier](https://github.com/google/guava/blob/master/guava/src/com/google/common/base/Suppliers.java#L157-L201) in Guava? (I couldn't find one but I'm not super familiar with the code base). Hand rolling the memoization logic inside IndexSearch is an option too, we'll just need to take some care to ensure that it is thread safe. -- 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
