JapaneseCompletionFilter expands a reading into every combination of romaji spellings with no bound, so 24-32 kana (for example 32 x "ン", or 32 x "し" in QUERY mode) throw OutOfMemoryError inside one incrementToken() call
Finding LFV-346Area: analysis / kuromoji (JapaneseCompletionFilter and JapaneseCompletionAnalyzer romanization, KatakanaRomanizer)Priority: mediumImpact: crashNovelty: new
AI-assisted finding, confirmed by automated checks, not yet triaged by a human Lucene committer. This site files nothing upstream.
What it shows
You analyze a short Japanese string with JapaneseCompletionAnalyzer or JapaneseCompletionFilter, for example 32 x "ン" when indexing or 32 x "し" as an auto-complete query, and expect the surface token plus a bounded set of romanized spellings. Instead KatakanaRomanizer.romanize builds every combination of the alternative spellings of each ambiguous kana (2^n for シ ジ チ ツ ヅ フ ヲ ン, 3^n for ヂ; for ン most are duplicates: 2^n tokens but only n+1 distinct strings), eagerly and with no limit, and the filter buffers all of them before it returns the first token. So 20 x "ン" already gives 1,048,577 tokens at one position, 24-32 kana throw OutOfMemoryError, a LimitTokenCountFilter placed after the filter cannot help, and an IndexWriter that analyzes such a field closes.
The test shows that JapaneseCompletionAnalyzer expands every combination of romaji spellings eagerly, so "ン" x 8 gives 2^8 = 256 romanized tokens at one position, nearly all duplicates (the same 2^n growth exhausts the heap at 24-32 kana).
Expected vs actual
Expected: each token once, 10 tokens (the surface form and n x 8 .. n x 16). Actual (-ea and -da): 257 tokens (expected:<10> but was:<257>).
Files
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=$CP:$LUCENE/lucene/analysis/common/build/libs/lucene-analysis-common-$V.jar
CP=$CP:$LUCENE/lucene/analysis/kuromoji/build/libs/lucene-analysis-kuromoji-$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.analysis.ja.TestJapaneseCompletionDuplicateRomajiThe module(s) analysis-kuromoji also need their third-party runtime jars; ./gradlew -p lucene/<module> dependencies lists them.
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: japanese-completion-fifo
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.