Hunspell spell-checker: CHECKCOMPOUNDTRIPLE misses an x|xx triple letter (one letter before the boundary, two after) at the second and later boundaries of a compound, so words such as operabareeel (opera + bare|eel) are accepted (absolute break position compared with the tail's relative length)

Finding LFV-373Area: analysis / hunspell (Hunspell.spell compound checking)Priority: lowImpact: silent wrong resultsNovelty: new

Finding page for LFV-373

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

What it shows

You load a Hunspell dictionary with CHECKCOMPOUNDTRIPLE, which hunspell(5) defines as 'Forbid compounding, if compound word contains triple repeating letters (e.g. foo|ox or xo|oof)', and expect Hunspell.spell to reject a compound with a triple letter at any boundary, as C Hunspell 1.3.2 and 1.7.2 do and as Lucene's own checkcompoundtriple fixture requires for bareeel (bare|eel). Instead Lucene rejects bareeel but accepts operabareeel, barebareeel and foobareeel (opera, bare or foo + bare|eel), and with the public Swedish dictionary (sv_SE) it accepts sommarmusikradiooomskuren while it rejects musikradiooomskuren. The cause: Hunspell.mayBreakIntoCompounds receives an absolute break position (the tail's offset in the word plus the relative break) but tests 'breakPos < length - 1' against the tail's relative length, so at the second and later boundaries the x|xx half of the check is skipped whenever the absolute break position is at least the tail length minus 1, while the xx|x half ('breakPos > offset + 1') is offset-aware and still works.

The test shows that Hunspell's CHECKCOMPOUNDTRIPLE misses an x|xx triple letter at the second boundary of a compound: with opera/A bare/A eel/A, operabareeel (opera + bare|eel) is accepted.

Expected vs actual

Expected spell("operabareeel") to be false (bare|eel has "eee", and Lucene's own fixture rejects bareeel); actual: true (expected:<false> but was:<true>), with -ea and -da.

Files

Run it

# 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: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.hunspell.TestCheckCompoundTripleLaterBoundary

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: hunspell

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.