[ https://issues.apache.org/jira/browse/LUCENE-9620?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17417412#comment-17417412 ]
Julie Tibshirani commented on LUCENE-9620: ------------------------------------------ Here's a randomized test failure that just popped up: {code} ./gradlew test --tests TestBooleanQuery.testQueryMatchesCount -Dtests.seed=8F43C2BE8EDF853A -Dtests.slow=true -Dtests.badapples=true -Dtests.locale=lv-LV -Dtests.timezone=Asia/Almaty -Dtests.asserts=true -Dtests.file.encoding=UTF-8 {code} {code} org.apache.lucene.search.TestTermQuery > testQueryMatchesCount FAILED 23:10:18 java.lang.IndexOutOfBoundsException: Index 0 out of bounds for length 0 23:10:18 at __randomizedtesting.SeedInfo.seed([8F43C2BE8EDF853A:7BB4170C66E42620]:0) 23:10:18 at java.base/jdk.internal.util.Preconditions.outOfBounds(Preconditions.java:64) 23:10:18 at java.base/jdk.internal.util.Preconditions.outOfBoundsCheckIndex(Preconditions.java:70) 23:10:18 at java.base/jdk.internal.util.Preconditions.checkIndex(Preconditions.java:248) 23:10:18 at java.base/java.util.Objects.checkIndex(Objects.java:372) 23:10:18 at java.base/java.util.ArrayList.get(ArrayList.java:458) 23:10:18 at java.base/java.util.Collections$UnmodifiableList.get(Collections.java:1308) 23:10:18 at org.apache.lucene.search.TestTermQuery.testQueryMatchesCount(TestTermQuery.java:130) {code} > Add Weight#count(LeafReaderContext) > ----------------------------------- > > Key: LUCENE-9620 > URL: https://issues.apache.org/jira/browse/LUCENE-9620 > Project: Lucene - Core > Issue Type: Improvement > Reporter: Adrien Grand > Priority: Minor > Fix For: 8.10 > > Time Spent: 4h 40m > Remaining Estimate: 0h > > We have IndexSearcher#count today, which tries to optimize counting for > TermQuery and MatchAllDocsQuery, and falls back to BulkScorer + > TotalHitCountCollector otherwise. > I'm considering moving this to Weight instead, where it'd be a better place > to add counting optimizations for other queries, e.g. pure disjunctions over > single-valued fields or range queries on points. The default implementation > could use a BulkScorer+TotalHitCountCollector like IndexSearcher#count does > today. -- This message was sent by Atlassian Jira (v8.3.4#803005) --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org For additional commands, e-mail: issues-h...@lucene.apache.org