s1monw commented on pull request #1623: URL: https://github.com/apache/lucene-solr/pull/1623#issuecomment-674108890
> I'm not familiar with the pooled reader stuff, but it seems reasonable to me. I trust the tests (especially w/ added test coverage) far more than I trust my knowledge of that part of the code. @msfroh we maintain a pool of readers and their corresponding updates in the IndexWriter in order to reuse them in NRT readers, for applying deletes and DV updates or as merge sources. The key here is that the core reader is immutable and we maintain a reference to it even if DV or delete generations change. For this change I borrow the reader from the pool the the time being and once the merge is done I pull a reference to the freshly created segment ie. it's core reader with DV and delete generations being `0` That reader is then used to open the directory reader on top of the merged and remaining segments ---------------------------------------------------------------- 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: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org For additional commands, e-mail: issues-h...@lucene.apache.org
