Write 4-byte characters (surrogate pairs) instead of escapes

Author: rnetuka

src/main/java/com/fasterxml/jackson/core/JsonGenerator.java

@@ -275,7 +275,18 @@ public enum Feature {
          *
          * @since 2.17
          */
-        ESCAPE_FORWARD_SLASHES(false);
+        ESCAPE_FORWARD_SLASHES(false),
+
+        /**
+         * Feature that specifies how 4-byte characters should be handled in {@link JsonGenerator}. If enabled,
+         * 4-byte characters made by surrogate pairs are combined and flushed as a single character encoded in UTF-8.
+         * If disabled, each pair is written as UTF-16 escape.
+         * <p>
+         * Feature is disabled by default
+         *
+         * @since 2.18
+         */
+        COMBINE_UNICODE_SURROGATES(false);
 
         private final boolean _defaultState;
         private final int _mask;

src/main/java/com/fasterxml/jackson/core/json/UTF8JsonGenerator.java

@@ -3,6 +3,7 @@
 import java.io.*;
 import java.math.BigDecimal;
 import java.math.BigInteger;
+import java.nio.charset.StandardCharsets;
 
 import com.fasterxml.jackson.core.*;
 import com.fasterxml.jackson.core.io.CharTypes;
@@ -659,6 +660,10 @@ public void writeUTF8String(byte[] text, int offset, int len) throws IOException
         _outputBuffer[_outputTail++] = _quoteChar;
     }
 
+    private boolean isSurrogatePair(char ch) {
+        return (ch & 0xD800) == 0xD800;
+    }
+
     /*
     /**********************************************************
     /* Output method implementations, unprocessed ("raw")
@@ -1489,6 +1494,8 @@ private final void _writeStringSegment2(final char[] cbuf, int offset, final int
         final byte[] outputBuffer = _outputBuffer;
         final int[] escCodes = _outputEscapes;
 
+        boolean combineSurrogates = Feature.COMBINE_UNICODE_SURROGATES.enabledIn(_features);
+
         while (offset < end) {
             int ch = cbuf[offset++];
             if (ch <= 0x7F) {
@@ -1510,7 +1517,14 @@ private final void _writeStringSegment2(final char[] cbuf, int offset, final int
                 outputBuffer[outputPtr++] = (byte) (0xc0 | (ch >> 6));
                 outputBuffer[outputPtr++] = (byte) (0x80 | (ch & 0x3f));
             } else {
-                outputPtr = _outputMultiByteChar(ch, outputPtr);
+                // multibyte character
+                if (combineSurrogates && isSurrogatePair((char) ch) && offset < end) {
+                    char highSurrogate = (char) ch;
+                    char lowSurrogate = cbuf[offset++];
+                    outputPtr = _outputSurrogatePair(highSurrogate, lowSurrogate, outputPtr);
+                } else {
+                    outputPtr = _outputMultiByteChar(ch, outputPtr);
+                }
             }
         }
         _outputTail = outputPtr;
@@ -1527,6 +1541,8 @@ private final void _writeStringSegment2(final String text, int offset, final int
         final byte[] outputBuffer = _outputBuffer;
         final int[] escCodes = _outputEscapes;
 
+        boolean combineSurrogates = Feature.COMBINE_UNICODE_SURROGATES.enabledIn(_features);
+
         while (offset < end) {
             int ch = text.charAt(offset++);
             if (ch <= 0x7F) {
@@ -1548,7 +1564,14 @@ private final void _writeStringSegment2(final String text, int offset, final int
                 outputBuffer[outputPtr++] = (byte) (0xc0 | (ch >> 6));
                 outputBuffer[outputPtr++] = (byte) (0x80 | (ch & 0x3f));
             } else {
-                outputPtr = _outputMultiByteChar(ch, outputPtr);
+                // multibyte character
+                if (combineSurrogates && isSurrogatePair((char) ch) && offset < end) {
+                    char highSurrogate = (char) ch;
+                    char lowSurrogate = text.charAt(offset++);
+                    outputPtr = _outputSurrogatePair(highSurrogate, lowSurrogate, outputPtr);
+                } else {
+                    outputPtr = _outputMultiByteChar(ch, outputPtr);
+                }
             }
         }
         _outputTail = outputPtr;
@@ -2133,6 +2156,13 @@ protected final void _outputSurrogates(int surr1, int surr2) throws IOException
         bbuf[_outputTail++] = (byte) (0x80 | (c & 0x3f));
     }
 
+    private int _outputSurrogatePair(char highSurrogate, char lowSurrogate, int outputPtr) {
+        String s = String.valueOf(highSurrogate) + lowSurrogate;
+        byte[] bytes = s.getBytes(StandardCharsets.UTF_8);
+        System.arraycopy(bytes, 0, _outputBuffer, outputPtr, bytes.length);
+        return outputPtr + bytes.length;
+    }
+
     /**
      *
      * @param ch

src/test/java/com/fasterxml/jackson/failing/Surrogate223Test.java renamed to src/test/java/com/fasterxml/jackson/core/json/Surrogate223Test.java

@@ -1,4 +1,4 @@
-package com.fasterxml.jackson.failing;
+package com.fasterxml.jackson.core.json;
 
 import java.io.ByteArrayOutputStream;
 import java.io.StringWriter;
@@ -8,6 +8,7 @@
 
 import org.junit.jupiter.api.Test;
 
+import static com.fasterxml.jackson.core.JsonGenerator.Feature;
 import static org.junit.jupiter.api.Assertions.assertEquals;
 
 class Surrogate223Test extends JUnit5TestBase
@@ -27,7 +28,7 @@ void surrogatesByteBacked() throws Exception
 //        assertFalse(JSON_F.isEnabled(JsonGenerator.Feature.ESCAPE_UTF8_SURROGATES));
 
         out = new ByteArrayOutputStream();
-        g = JSON_F.createGenerator(out);
+        g = JSON_F.createGenerator(out).enable(Feature.COMBINE_UNICODE_SURROGATES);
         g.writeStartArray();
         g.writeString(toQuote);
         g.writeEndArray();
@@ -43,8 +44,7 @@ void surrogatesByteBacked() throws Exception
 
         // but may revert back to original behavior
         out = new ByteArrayOutputStream();
-        g = JSON_F.createGenerator(out);
-//        g.enable(JsonGenerator.Feature.ESCAPE_UTF8_SURROGATES);
+        g = JSON_F.createGenerator(out).disable(Feature.COMBINE_UNICODE_SURROGATES);
         g.writeStartArray();
         g.writeString(toQuote);
         g.writeEndArray();