javanna commented on code in PR #13542:
URL: https://github.com/apache/lucene/pull/13542#discussion_r1690241865
##########
lucene/core/src/java/org/apache/lucene/search/PointRangeQuery.java:
##########
@@ -362,6 +362,9 @@ public long cost() {
final IntersectVisitor visitor = getIntersectVisitor(result);
long cost = -1;
+ // maybe allowing calling get multiple times, or clone the scorer
to avoid duplication
+ // across multiple
+ // slices that point to the same segment
Review Comment:
I think what I said above is debatable. I initially thought this was mostly
a question of performance and optimizations. As I understand now, it impacts
functionality as well in that existing weight implementations may behave as if
they went multiple times through the same segment. Strange that there are no
failures other than `TestSortRandom` so far, but that is a good example of a
possible real situation that we'd ideally address without asking users to
modify their code.
I am focusing on this now with the goal of proposing some kind of solution
and understanding what breaking changes we'd have to make.
--
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]
