TeeSinkTokenFilter: a sink never clears attributes that only its own filter chain adds (SinkTokenStream.incrementToken() calls restoreState() without clearAttributes()), so a keyword marker or StemmerOverrideFilter before a stemmer on a sink silently leaves every token after the first protected word unstemmed

Finding LFV-651Area: analysis / TeeSinkTokenFilter (SinkTokenStream replay through AttributeSource.restoreState without clearAttributes(): attributes that only the sink's chain adds)Priority: lowImpact: silent wrong resultsNovelty: new

Finding page for LFV-651

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

What it shows

You index one field from a TeeSinkTokenFilter and a second field from one of its sinks, wrapped in a keyword marker and a stemmer (for example SetKeywordMarkerFilter protecting 'quick', then PorterStemFilter), and expect the sink field to get what the same chain gives on the tokenizer: [the, quick, fox, jump] for 'The quick foxes jumped'. Instead it gets [the, quick, foxes, jumped]: the sink replays each token with restoreState() but never calls clearAttributes(), and the KeywordAttribute that only the sink's chain added is not in the tee's captured states, so once 'quick' is marked as a keyword every later token stays marked and is not stemmed. Nothing is thrown; a query for stemmed:fox finds nothing.

The test shows that a TeeSinkTokenFilter sink never clears attributes that only its own chain adds, so SetKeywordMarkerFilter's keyword flag for "quick" sticks to every later token and PorterStemFilter stops stemming.

Expected vs actual

Expected: the sink chain on "quick foxes" gives [quick, fox], as the same chain on the tokenizer does. Actual (-ea and -da): [quick, foxes].

Files

• src/org/apache/lucene/analysis/sinks/TestTeeSinkSinkOnlyAttribute.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=$CP:$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.sinks.TestTeeSinkSinkOnlyAttribute

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).

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.