Use immutable XContentText.EncodedBytes record for encoded Text

We actually can't use the native java ByteBuffer class here because it's
a mutable class (since it keeps track of how many bytes have been
consumed), which creates concurrency issues.

Author: jordan-powers

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

@@ -13,8 +13,8 @@
 import java.nio.charset.StandardCharsets;
 
 /**
- * Both {@link String} and {@link ByteBuffer} representation of the text. Starts with one of those, and if
- * the other is requests, caches the other one in a local reference so no additional conversion will be needed.
+ * Both {@link String} and {@link EncodedBytes} representation of the text. Starts with one of those, and if
+ * the other is requested, caches the other one in a local reference so no additional conversion will be needed.
  */
 public final class Text implements XContentString, Comparable<Text>, ToXContentFragment {
 
@@ -31,7 +31,7 @@ public static Text[] convertFromStringArray(String[] strings) {
         return texts;
     }
 
-    private ByteBuffer bytes;
+    private EncodedBytes bytes;
     private String text;
     private int hash;
     private int stringLength = -1;
@@ -40,15 +40,15 @@ public static Text[] convertFromStringArray(String[] strings) {
      * Construct a Text from a UTF-8 encoded ByteBuffer. Since no string length is specified, {@link #stringLength()}
      * will perform a string conversion to measure the string length.
      */
-    public Text(ByteBuffer bytes) {
+    public Text(EncodedBytes bytes) {
         this.bytes = bytes;
     }
 
     /**
      * Construct a Text from a UTF-8 encoded ByteBuffer and an explicit string length. Used to avoid string conversion
      * in {@link #stringLength()}.
      */
-    public Text(ByteBuffer bytes, int stringLength) {
+    public Text(EncodedBytes bytes, int stringLength) {
         this.bytes = bytes;
         this.stringLength = stringLength;
     }
@@ -58,16 +58,18 @@ public Text(String text) {
     }
 
     /**
-     * Whether a {@link ByteBuffer} view of the data is already materialized.
+     * Whether an {@link EncodedBytes} view of the data is already materialized.
      */
     public boolean hasBytes() {
         return bytes != null;
     }
 
     @Override
-    public ByteBuffer bytes() {
+    public EncodedBytes bytes() {
         if (bytes == null) {
-            bytes = StandardCharsets.UTF_8.encode(text);
+            var byteBuff = StandardCharsets.UTF_8.encode(text);
+            assert byteBuff.hasArray();
+            bytes = new EncodedBytes(byteBuff.array(), byteBuff.arrayOffset() + byteBuff.position(), byteBuff.remaining());
         }
         return bytes;
     }
@@ -82,7 +84,8 @@ public boolean hasString() {
     @Override
     public String string() {
         if (text == null) {
-            text = StandardCharsets.UTF_8.decode(bytes).toString();
+            var byteBuff = ByteBuffer.wrap(bytes.bytes(), bytes.offset(), bytes.length());
+            text = StandardCharsets.UTF_8.decode(byteBuff).toString();
         }
         return text;
     }
@@ -131,8 +134,7 @@ public XContentBuilder toXContent(XContentBuilder builder, Params params) throws
         } else {
             // TODO: TextBytesOptimization we can use a buffer here to convert it? maybe add a
             // request to jackson to support InputStream as well?
-            assert bytes.hasArray();
-            return builder.utf8Value(bytes.array(), bytes.arrayOffset() + bytes.position(), bytes.remaining());
+            return builder.utf8Value(bytes.bytes(), bytes.offset(), bytes.length());
         }
     }
 }

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

@@ -12,6 +12,34 @@
 import java.nio.ByteBuffer;
 
 public interface XContentString {
+    record EncodedBytes(byte[] bytes, int offset, int length) implements Comparable<EncodedBytes> {
+        public EncodedBytes(byte[] bytes) {
+            this(bytes, 0, bytes.length);
+        }
+
+        @Override
+        public int compareTo(EncodedBytes o) {
+            return ByteBuffer.wrap(bytes, offset, length).compareTo(ByteBuffer.wrap(o.bytes, o.offset, o.length));
+        }
+
+        @Override
+        public boolean equals(Object o) {
+            if (this == o) {
+                return true;
+            }
+            if (o == null || getClass() != o.getClass()) {
+                return false;
+            }
+
+            return this.compareTo((EncodedBytes) o) == 0;
+        }
+
+        @Override
+        public int hashCode() {
+            return ByteBuffer.wrap(bytes, offset, length).hashCode();
+        }
+    }
+
     /**
      * Returns a {@link String} view of the data.
      */
@@ -20,7 +48,7 @@ public interface XContentString {
     /**
      * Returns a UTF8-encoded {@link ByteBuffer} view of the data.
      */
-    ByteBuffer bytes();
+    EncodedBytes bytes();
 
     /**
      * Returns the number of characters in the represented string.

server/src/main/java/org/elasticsearch/common/io/stream/StreamInput.java

@@ -29,6 +29,7 @@
 import org.elasticsearch.core.Nullable;
 import org.elasticsearch.core.TimeValue;
 import org.elasticsearch.xcontent.Text;
+import org.elasticsearch.xcontent.XContentString;
 
 import java.io.EOFException;
 import java.io.FilterInputStream;
@@ -395,8 +396,8 @@ public Text readOptionalText() throws IOException {
         if (length > 0) {
             readBytes(bytes, 0, length);
         }
-        var byteBuff = ByteBuffer.wrap(bytes);
-        return new Text(byteBuff);
+        var encoded = new XContentString.EncodedBytes(bytes);
+        return new Text(encoded);
     }
 
     public Text readText() throws IOException {
@@ -406,8 +407,8 @@ public Text readText() throws IOException {
         if (length > 0) {
             readBytes(bytes, 0, length);
         }
-        var byteBuff = ByteBuffer.wrap(bytes);
-        return new Text(byteBuff);
+        var encoded = new XContentString.EncodedBytes(bytes);
+        return new Text(encoded);
     }
 
     @Nullable

server/src/main/java/org/elasticsearch/common/io/stream/StreamOutput.java

@@ -419,7 +419,8 @@ public void writeText(Text text) throws IOException {
             writeInt(spare.length());
             write(spare.bytes(), 0, spare.length());
         } else {
-            BytesReference bytes = BytesReference.fromByteBuffer(text.bytes());
+            var encoded = text.bytes();
+            BytesReference bytes = new BytesArray(encoded.bytes(), encoded.offset(), encoded.length());
             writeInt(bytes.length());
             bytes.writeTo(this);
         }

server/src/main/java/org/elasticsearch/search/suggest/term/TermSuggester.java

@@ -20,9 +20,9 @@
 import org.elasticsearch.search.suggest.SuggestionSearchContext.SuggestionContext;
 import org.elasticsearch.search.suggest.phrase.DirectCandidateGenerator;
 import org.elasticsearch.xcontent.Text;
+import org.elasticsearch.xcontent.XContentString;
 
 import java.io.IOException;
-import java.nio.ByteBuffer;
 import java.util.ArrayList;
 import java.util.List;
 
@@ -48,7 +48,8 @@ public TermSuggestion innerExecute(String name, TermSuggestionContext suggestion
                 suggestion.getDirectSpellCheckerSettings().suggestMode()
             );
             var termBytes = token.term.bytes();
-            Text key = new Text(ByteBuffer.wrap(termBytes.bytes, termBytes.offset, termBytes.length));
+            var termEncoded = new XContentString.EncodedBytes(termBytes.bytes, termBytes.offset, termBytes.length);
+            Text key = new Text(termEncoded);
             TermSuggestion.Entry resultEntry = new TermSuggestion.Entry(key, token.startOffset, token.endOffset - token.startOffset);
             for (SuggestWord suggestWord : suggestedWords) {
                 Text word = new Text(suggestWord.string);
@@ -98,7 +99,8 @@ protected TermSuggestion emptySuggestion(String name, TermSuggestionContext sugg
         List<Token> tokens = queryTerms(suggestion, spare);
         for (Token token : tokens) {
             var termBytes = token.term.bytes();
-            Text key = new Text(ByteBuffer.wrap(termBytes.bytes, termBytes.offset, termBytes.length));
+            var termEncoded = new XContentString.EncodedBytes(termBytes.bytes, termBytes.offset, termBytes.length);
+            Text key = new Text(termEncoded);
             TermSuggestion.Entry resultEntry = new TermSuggestion.Entry(key, token.startOffset, token.endOffset - token.startOffset);
             termSuggestion.addTerm(resultEntry);
         }

server/src/test/java/org/elasticsearch/common/xcontent/BaseXContentTestCase.java

@@ -34,6 +34,7 @@
 import org.elasticsearch.xcontent.XContentParser;
 import org.elasticsearch.xcontent.XContentParser.Token;
 import org.elasticsearch.xcontent.XContentParserConfiguration;
+import org.elasticsearch.xcontent.XContentString;
 import org.elasticsearch.xcontent.XContentType;
 import org.hamcrest.Matcher;
 import org.hamcrest.Matchers;
@@ -377,15 +378,18 @@ public void testText() throws Exception {
         assertResult("{'text':''}", () -> builder().startObject().field("text", new Text("")).endObject());
         assertResult("{'text':'foo bar'}", () -> builder().startObject().field("text", new Text("foo bar")).endObject());
 
-        final var random = ByteBuffer.wrap(randomBytes());
-        XContentBuilder builder = builder().startObject().field("text", new Text(random)).endObject();
+        final var random = randomBytes();
+        XContentBuilder builder = builder().startObject().field("text", new Text(new XContentString.EncodedBytes(random))).endObject();
 
         try (XContentParser parser = createParser(xcontentType().xContent(), BytesReference.bytes(builder))) {
             assertSame(parser.nextToken(), Token.START_OBJECT);
             assertSame(parser.nextToken(), Token.FIELD_NAME);
             assertEquals(parser.currentName(), "text");
             assertTrue(parser.nextToken().isValue());
-            assertThat(new BytesRef(parser.charBuffer()).utf8ToString(), equalTo(StandardCharsets.UTF_8.decode(random).toString()));
+            assertThat(
+                new BytesRef(parser.charBuffer()).utf8ToString(),
+                equalTo(StandardCharsets.UTF_8.decode(ByteBuffer.wrap(random)).toString())
+            );
             assertSame(parser.nextToken(), Token.END_OBJECT);
             assertNull(parser.nextToken());
         }
@@ -593,7 +597,7 @@ public void testObjects() throws Exception {
         objects.put("{'objects':['a','b','c']}", new Object[] { "a", "b", "c" });
         objects.put(
             "{'objects':['a','b','c']}",
-            new Object[] { new Text("a"), new Text(StandardCharsets.UTF_8.encode("b")), new Text("c") }
+            new Object[] { new Text("a"), new Text(new XContentString.EncodedBytes("b".getBytes(StandardCharsets.UTF_8))), new Text("c") }
         );
         objects.put("{'objects':null}", null);
         objects.put("{'objects':[null,null,null]}", new Object[] { null, null, null });
@@ -640,7 +644,7 @@ public void testObject() throws Exception {
         object.put("{'object':1}", (short) 1);
         object.put("{'object':'string'}", "string");
         object.put("{'object':'a'}", new Text("a"));
-        object.put("{'object':'b'}", new Text(StandardCharsets.UTF_8.encode("b")));
+        object.put("{'object':'b'}", new Text(new XContentString.EncodedBytes("b".getBytes(StandardCharsets.UTF_8))));
         object.put("{'object':null}", null);
         object.put("{'object':'OPEN'}", IndexMetadata.State.OPEN);
         object.put("{'object':'NM'}", DistanceUnit.NAUTICALMILES);