Use unsigned comparison in UTF8Bytes#compareTo

Author: jordan-powers

libs/x-content/src/main/java/org/elasticsearch/xcontent/XContentString.java

@@ -10,6 +10,7 @@
 package org.elasticsearch.xcontent;
 
 import java.nio.ByteBuffer;
+import java.util.Arrays;
 
 public interface XContentString {
     record UTF8Bytes(byte[] bytes, int offset, int length) implements Comparable<UTF8Bytes> {
@@ -23,7 +24,7 @@ public int compareTo(UTF8Bytes o) {
                 return 0;
             }
 
-            return ByteBuffer.wrap(bytes, offset, length).compareTo(ByteBuffer.wrap(o.bytes, o.offset, o.length));
+            return Arrays.compareUnsigned(bytes, offset, offset + length, o.bytes, o.offset, o.offset + o.length);
         }
 
         @Override

libs/x-content/src/test/java/org/elasticsearch/xcontent/TextTests.java

@@ -12,6 +12,7 @@
 import org.elasticsearch.test.ESTestCase;
 
 import java.nio.charset.StandardCharsets;
+import java.util.Arrays;
 
 public class TextTests extends ESTestCase {
     public void testConvertToBytes() {
@@ -136,7 +137,8 @@ public void testCompareTo() {
         byte[] encodedArr2 = value2.getBytes(StandardCharsets.UTF_8);
         var encoded2 = new XContentString.UTF8Bytes(encodedArr2);
 
-        int compSign = (int) Math.signum(encoded1.compareTo(encoded2));
+        // String.compareTo() wasn't handling surrogate pairs very well, so here we compare the full 32-bit codepoints
+        int compSign = (int) Math.signum(Arrays.compare(value1.codePoints().toArray(), value2.codePoints().toArray()));
 
         {
             var text1 = new Text(value1);