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

Author: rnetuka

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

@@ -202,6 +202,8 @@ public enum Feature {
          */
         WRITE_BIGDECIMAL_AS_PLAIN(false),
 
+        WRITE_UTF8_SURROGATES(false),
+
         // // Schema/Validity support features
 
         /**

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

@@ -74,6 +74,8 @@ public enum StreamWriteFeature
      */
     WRITE_BIGDECIMAL_AS_PLAIN(JsonGenerator.Feature.WRITE_BIGDECIMAL_AS_PLAIN),
 
+    WRITE_UTF8_SURROGATES(JsonGenerator.Feature.WRITE_UTF8_SURROGATES),
+
     // // Schema/Validity support features
 
     /**

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 writeSurrogates = Feature.WRITE_UTF8_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 (writeSurrogates && 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 writeSurrogates = Feature.WRITE_UTF8_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 (writeSurrogates && 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/core/json/GeneratorFeaturesTest.java

@@ -28,6 +28,7 @@ void configDefaults() throws IOException
 
         assertFalse(g.isEnabled(JsonGenerator.Feature.WRITE_BIGDECIMAL_AS_PLAIN));
         assertFalse(g.isEnabled(StreamWriteFeature.WRITE_BIGDECIMAL_AS_PLAIN));
+        assertFalse(g.isEnabled(StreamWriteFeature.WRITE_UTF8_SURROGATES));
 
         assertTrue(g.canOmitFields());
         assertFalse(g.canWriteBinaryNatively());

src/test/java/com/fasterxml/jackson/core/json/StringGenerationTest.java

@@ -6,6 +6,7 @@
 
 import java.util.Random;
 
+import static com.fasterxml.jackson.core.JsonGenerator.Feature;
 import static org.junit.jupiter.api.Assertions.*;
 
 import org.junit.jupiter.api.Test;
@@ -86,6 +87,17 @@ void longerRandomMultiChunk() throws Exception
         }
     }
 
+    @Test
+    public void testWritingSurrogatePairs() throws IOException {
+        ByteArrayOutputStream stream = new ByteArrayOutputStream();
+        JsonGenerator generator = FACTORY.createGenerator(stream, JsonEncoding.UTF8).enable(Feature.WRITE_UTF8_SURROGATES);
+        String string = "システム\uD867\uDE3D"; // システム𩸽
+        generator.writeString(string);
+        generator.flush();
+        generator.close();
+        assertEquals("\"" + string + "\"", stream.toString());
+    }
+
     /*
     /**********************************************************
     /* Internal methods