BKDWriter does not cap its in-heap point buffer at the Java array limit, so with maxMBSortInHeap >= 2048 MB a points merge of more than 2^31 bytes throws NegativeArraySizeException and closes the IndexWriter

Finding LFV-613Area: util / BKD points (BKDWriter in-heap sort buffer sizing: HeapPointWriter and the BKDRadixSelector heap writers, reached through Lucene90PointsWriter's public maxMBSortInHeap constructor)Priority: lowImpact: crashNovelty: new

Finding page for LFV-613

AI-assisted finding, confirmed by automated checks, not yet triaged by a human Lucene committer. This site files nothing upstream.

What it shows

You give a points format a sort budget of 2048 MB or more, for example new Lucene90PointsWriter(state, 512, 2200.0) in a custom codec, and merge segments that hold more than 2 GiB of points of one multi-dimensional field (in the repro 8.4M points of 16 dims x 16 bytes). You expect the merge to succeed, because BKDWriter 'uses up to the specified maxMBSortInHeap heap space for writing', spills to disk beyond it, and its constructor accepts the value. Instead BKDWriter sizes its in-heap buffer as budget / bytesPerDoc points with no cap at the Java array limit, HeapPointWriter computes bytesPerDoc * size in int, the length wraps negative, and forceMerge(1) throws NegativeArraySizeException (-2110967296) and leaves the IndexWriter closed with that tragic exception.

The test shows that BKDWriter accepts maxMBSortInHeap = 2200 MB but does not cap its in-heap buffer at the Java array limit, so the first add() for 8.4M points of 260 bytes overflows int.

Expected vs actual

Expected: the point is buffered or spilled to disk. Actual (-ea and -da): NegativeArraySizeException: -2110967296 from HeapPointWriter.<init>.

Files

• src/org/apache/lucene/util/bkd/TestBKDWriterSortBufferOver2GB.java (JUnit test)

Run it

• Tested on Lucene main e1faeba08bde (2026-11-02): still reproduces.
• Tested on Lucene main 645d4fe14a20 (2026-10-28): reproduced.
• Recorded outcome of run 1 on Lucene main 645d4fe14a20: 1 test(s) run, 1 failed.
• How to read the outcome: a failing test is the finding reproducing (the failure message states expected against actual).
• Seeds: a rerun with the same seed replays the same random choices. Assertions off (-da -dsa -Dtests.asserts=false) is the production-like setting.

# Lucene checkout at the tested commit, built with: ./gradlew --no-daemon assemble   (JDK 25)
LUCENE=/path/to/lucene; V=12.0.0-SNAPSHOT
CP=$LUCENE/lucene/core/build/libs/lucene-core-$V.jar
CP=$CP:$LUCENE/lucene/codecs/build/libs/lucene-codecs-$V.jar
CP=$CP:$LUCENE/lucene/test-framework/build/libs/lucene-test-framework-$V.jar
CP=$CP:randomizedtesting-runner-2.9.2.jar:junit-4.13.2.jar:hamcrest-3.0.jar   # test-framework dependencies (Maven Central)
CP=$CP:junit-platform-console-standalone-6.1.3.jar   # JUnit Platform console launcher (Maven Central)

javac -d classes -cp "$CP" $(find src -name '*.java')

# run 1: assertions on, seed DEADBEEF
java -ea -esa --add-modules jdk.incubator.vector --enable-native-access=ALL-UNNAMED -Dtests.seed=DEADBEEF \
     -cp "classes:$CP" org.junit.platform.console.ConsoleLauncher \
     execute --select-class org.apache.lucene.util.bkd.TestBKDWriterSortBufferOver2GB

Gradle alternative: copy the test into the matching module's src/test/ tree of the checkout and run ./gradlew -p lucene/<module> test --tests <class> -Ptests.seed=<seed> (add -Ptests.asserts=false for a run without assertions).

Formal models and scripts this finding's part of the hunt used: bkd-write-niche-configs

Feedback

Open an issue in lucene-ai-bug-hunt. Please do not open issues or pull requests in apache/lucene about this finding unless a Lucene committer agrees it should be filed.

Snapshot of 2026-11-04. AI-assisted; not yet triaged by a human Lucene committer. Feedback: open an issue in lucene-ai-bug-hunt.