It _finally_ occurred to me to ask why we have the restriction that the destination of a copyField must have stored=false. I understand what currently happens when that’s the case, you get repeats.
What I wondered is why we can’t detect that a field is the destination of a copyField and _not_ pull the stored values out of it during atomic updates? Or do we run afoul of things in tlog retrieval or RTG? Is this a silly idea or should I raise a JIRA? --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
