Merge pull request #13709 from vespa-engine/bratseth/surrogate-aware-gram-splitting

Surrogate aware gram splitting

1 files changed, 2 insertions, 2 deletions

diff --git a/indexinglanguage/src/main/java/com/yahoo/vespa/indexinglanguage/expressions/NGramExpression.java b/indexinglanguage/src/main/java/com/yahoo/vespa/indexinglanguage/expressions/NGramExpression.java
index d91338e3d3f..adf3e4ecaaa 100644
--- a/indexinglanguage/src/main/java/com/yahoo/vespa/indexinglanguage/expressions/NGramExpression.java
+++ b/indexinglanguage/src/main/java/com/yahoo/vespa/indexinglanguage/expressions/NGramExpression.java
@@ -65,10 +65,10 @@ public final class NGramExpression extends Expression {
 
             // annotate gram as a word term
             String gramString = gram.extractFrom(input.getString());
-            typedSpan(gram.getStart(), gram.getLength(), TokenType.ALPHABETIC, spanList).
+            typedSpan(gram.getStart(), gram.getCodePointCount(), TokenType.ALPHABETIC, spanList).
                     annotate(LinguisticsAnnotator.lowerCaseTermAnnotation(gramString, gramString));
 
-            lastPosition = gram.getStart() + gram.getLength();
+            lastPosition = gram.getStart() + gram.getCodePointCount();
         }
         // handle punctuation at the end
         if (lastPosition < input.toString().length()) {