Allow exposing CBOR simple values as VALUE_EMBEDDED_OBJECT (#590)

iifawzi · web-flow · commit 33a2294a4094 · 2025-05-10T19:40:25.000-07:00
diff --git a/cbor/src/main/java/com/fasterxml/jackson/dataformat/cbor/CBORParser.java b/cbor/src/main/java/com/fasterxml/jackson/dataformat/cbor/CBORParser.java
@@ -60,7 +60,21 @@ public enum Feature implements FormatFeature
          *
          * @since 2.20
          */
-        READ_UNDEFINED_AS_EMBEDDED_OBJECT(false)
+        READ_UNDEFINED_AS_EMBEDDED_OBJECT(false),
+
+        /**
+         * Feature that determines how a CBOR "simple value" of major type 7 is exposed by parser.
+         * <p>
+         * When enabled, the parser returns {@link JsonToken#VALUE_EMBEDDED_OBJECT} with
+         * an embedded value of type {@link CBORSimpleValue}, allowing the caller to distinguish
+         * these values from actual {@link JsonToken#VALUE_NUMBER_INT}s.
+         * When disabled, simple values are returned as {@link JsonToken#VALUE_NUMBER_INT}.
+         *<p>
+         * The default value is {@code false} for backwards compatibility (with versions prior to 2.20).
+         *
+         * @since 2.20
+         */
+        READ_SIMPLE_VALUE_AS_EMBEDDED_OBJECT(false)
         ;
 
         private final boolean _defaultState;
@@ -363,6 +377,14 @@ public int getFirstTag() {
      */
     protected TagList _tagValues = new TagList();
 
+    /**
+     * When major type 7 value is encountered and exposed as {@link JsonToken#VALUE_EMBEDDED_OBJECT},
+     * the value will be stored here.
+     *
+     * @since 2.20
+     */
+    protected CBORSimpleValue _simpleValue;
+    
     /**
      * Flag that indicates that the current token has not yet
      * been fully processed, and needs to be finished for
@@ -824,9 +846,9 @@ public JsonToken nextToken() throws IOException
             _skipIncomplete();
         }
         _tokenInputTotal = _currInputProcessed + _inputPtr;
-        // also: clear any data retained so far
-        _numTypesValid = NR_UNKNOWN;
-        _binaryValue = null;
+
+        // also: clear any data retained for previous token
+        clearRetainedValues();
 
         // First: need to keep track of lengths of defined-length Arrays and
         // Objects (to materialize END_ARRAY/END_OBJECT as necessary);
@@ -1453,12 +1475,12 @@ public boolean nextFieldName(SerializableString str) throws IOException
     {
         // Two parsing modes; can only succeed if expecting field name, so handle that first:
         if (_streamReadContext.inObject() && _currToken != JsonToken.FIELD_NAME) {
-            _numTypesValid = NR_UNKNOWN;
             if (_tokenIncomplete) {
                 _skipIncomplete();
             }
             _tokenInputTotal = _currInputProcessed + _inputPtr;
-            _binaryValue = null;
+            // need to clear retained values for previous token
+            clearRetainedValues();
             _tagValues.clear();
             // completed the whole Object?
             if (!_streamReadContext.expectMoreValues()) {
@@ -1506,19 +1528,19 @@ public boolean nextFieldName(SerializableString str) throws IOException
             }
         }
         // otherwise just fall back to default handling; should occur rarely
-        return (nextToken() == JsonToken.FIELD_NAME) && str.getValue().equals(getCurrentName());
+        return (nextToken() == JsonToken.FIELD_NAME) && str.getValue().equals(currentName());
     }
 
     @Override
     public String nextFieldName() throws IOException
     {
         if (_streamReadContext.inObject() && _currToken != JsonToken.FIELD_NAME) {
-            _numTypesValid = NR_UNKNOWN;
             if (_tokenIncomplete) {
                 _skipIncomplete();
             }
             _tokenInputTotal = _currInputProcessed + _inputPtr;
-            _binaryValue = null;
+            // need to clear retained values for previous token
+            clearRetainedValues();
             _tagValues.clear();
             // completed the whole Object?
             if (!_streamReadContext.expectMoreValues()) {
@@ -1843,7 +1865,10 @@ public Object getEmbeddedObject() throws IOException
         if (_tokenIncomplete) {
             _finishToken();
         }
-        if (_currToken == JsonToken.VALUE_EMBEDDED_OBJECT ) {
+        if (_currToken == JsonToken.VALUE_EMBEDDED_OBJECT) {
+            if (_simpleValue != null) {
+                return _simpleValue;
+            }
             return _binaryValue;
         }
         return null;
@@ -1933,11 +1958,11 @@ private final byte[] _getBinaryFromString(Base64Variant variant) throws IOExcept
     /**
      * Checking whether the current token represents an `undefined` value (0xF7).
      * <p>
-     * This method allows distinguishing between real {@code null} and `undefined`,
+     * This method allows distinguishing between real {@code null} and {@code undefined},
      * even if {@link CBORParser.Feature#READ_UNDEFINED_AS_EMBEDDED_OBJECT} is disabled
      * and the token is reported as {@link JsonToken#VALUE_NULL}.
      *
-     * @return {@code true} if current token is an `undefined`, {@code false} otherwise
+     * @return {@code true} if current token is an {@code undefined}, {@code false} otherwise
      *
      * @since 2.20
      */
@@ -3713,38 +3738,50 @@ protected JsonToken _decodeUndefinedValue() {
      * Helper method that deals with details of decoding unallocated "simple values"
      * and exposing them as expected token.
      * <p>
-     * As of Jackson 2.12, simple values are exposed as
-     * {@link JsonToken#VALUE_NUMBER_INT}s,
-     * but in later versions this is planned to be changed to separate value type.
+     * Starting with Jackson 2.20, this behavior can be changed by enabling the
+     * {@link CBORParser.Feature#READ_SIMPLE_VALUE_AS_EMBEDDED_OBJECT}
+     * feature, in which case simple values are returned as {@link JsonToken#VALUE_EMBEDDED_OBJECT} with an
+     * embedded {@link CBORSimpleValue} instance.
      *
      * @since 2.12
      */
     public JsonToken _decodeSimpleValue(int lowBits, int ch) throws IOException {
         if (lowBits > 24) {
             _invalidToken(ch);
         }
+        final boolean simpleAsEmbedded = Feature.READ_SIMPLE_VALUE_AS_EMBEDDED_OBJECT.enabledIn(_formatFeatures);
         if (lowBits < 24) {
-            _numberInt = lowBits;
+            if (simpleAsEmbedded) {
+                _simpleValue = new CBORSimpleValue(lowBits);
+            } else {
+                _numberInt = lowBits;
+            }
         } else { // need another byte
             if (_inputPtr >= _inputEnd) {
                 loadMoreGuaranteed();
             }
-            _numberInt = _inputBuffer[_inputPtr++] & 0xFF;
+
             // As per CBOR spec, values below 32 not allowed to avoid
             // confusion (as well as guarantee uniqueness of encoding)
-            if (_numberInt < 32) {
+            int value = _inputBuffer[_inputPtr++] & 0xFF;
+            if (value < 32) {
                 throw _constructError("Invalid second byte for simple value: 0x"
-                        +Integer.toHexString(_numberInt)+" (only values 0x20 - 0xFF allowed)");
+                        +Integer.toHexString(value)+" (only values 0x20 - 0xFF allowed)");
+            }
+
+            if (simpleAsEmbedded) {
+                _simpleValue = new CBORSimpleValue(value);
+            } else {
+                _numberInt = value;
             }
         }
 
-        // 25-Nov-2020, tatu: Although ideally we should report these
-        //    as `JsonToken.VALUE_EMBEDDED_OBJECT`, due to late addition
-        //    of handling in 2.12, simple value in 2.12 will be reported
-        //    as simple ints.
+        if (simpleAsEmbedded) {
+            return JsonToken.VALUE_EMBEDDED_OBJECT;
+        }
 
         _numTypesValid = NR_INT;
-        return (JsonToken.VALUE_NUMBER_INT);
+        return JsonToken.VALUE_NUMBER_INT;
     }
 
     /*
@@ -4101,4 +4138,11 @@ private void createChildObjectContext(final int len) throws IOException {
         _streamReadContext = _streamReadContext.createChildObjectContext(len);
         _streamReadConstraints.validateNestingDepth(_streamReadContext.getNestingDepth());
     }
+
+    // @since 2.20
+    private void clearRetainedValues() {
+        _numTypesValid = NR_UNKNOWN;
+        _binaryValue = null;
+        _simpleValue = null;
+    }
 }
diff --git a/cbor/src/test/java/com/fasterxml/jackson/dataformat/cbor/parse/SimpleValuesTest.java b/cbor/src/test/java/com/fasterxml/jackson/dataformat/cbor/parse/SimpleValuesTest.java
@@ -30,6 +30,23 @@ public void testTinySimpleValues() throws Exception
         }
     }
 
+    @Test
+    public void testTinySimpleValuesAsEmbeddedObjectWhenEnabled() throws Exception
+    {
+        CBORFactory f = CBORFactory.builder()
+                .enable(CBORParser.Feature.READ_SIMPLE_VALUE_AS_EMBEDDED_OBJECT)
+                .build();
+        // Values 0..19 are unassigned, valid to encounter
+        for (int v = 0; v <= 19; ++v) {
+            byte[] doc = new byte[1];
+            doc[0] = (byte) (CBORConstants.PREFIX_TYPE_MISC + v);
+            try (CBORParser p = cborParser(f, doc)) {
+                assertToken(JsonToken.VALUE_EMBEDDED_OBJECT, p.nextToken());
+                assertEquals(new CBORSimpleValue(v), p.getEmbeddedObject());
+            }
+        }
+    }
+
     @Test
     public void testValidByteLengthMinimalValues() throws Exception {
         // Values 32..255 are unassigned, valid to encounter
@@ -44,6 +61,21 @@ public void testValidByteLengthMinimalValues() throws Exception {
         }
     }
 
+    @Test
+    public void testValidByteLengthMinimalValuesAsEmbeddedObjectWhenEnabled() throws Exception {
+        // Values 32..255 are unassigned, valid to encounter
+        CBORFactory f = CBORFactory.builder()
+                .enable(CBORParser.Feature.READ_SIMPLE_VALUE_AS_EMBEDDED_OBJECT)
+                .build();
+        for (int v = 32; v <= 255; ++v) {
+            byte[] doc = { (byte) (CBORConstants.PREFIX_TYPE_MISC + 24), (byte) v };
+            try (CBORParser p = cborParser(f, doc)) {
+                assertToken(JsonToken.VALUE_EMBEDDED_OBJECT, p.nextToken());
+                assertEquals(new CBORSimpleValue(v), p.getEmbeddedObject());
+            }
+        }
+    }
+
     @Test
     public void testInvalidByteLengthMinimalValues() throws Exception {
         // Values 0..31 are invalid for variant that takes 2 bytes...
diff --git a/release-notes/CREDITS-2.x b/release-notes/CREDITS-2.x
@@ -397,3 +397,7 @@ Fawzi Essam (@iifawzi)
  * Contributed fix for #431: (cbor) Negative `BigInteger` values not encoded/decoded
    correctly
   (2.20.0)
+ * Contributed implementation of #587: (cbor) Allow exposing CBOR Simple values as
+   `JsonToken.VALUE_EMBEDDED_OBJECT` with a feature flag
+  (2.20.0)
+
diff --git a/release-notes/VERSION-2.x b/release-notes/VERSION-2.x
@@ -22,6 +22,9 @@ Active maintainers:
 #431: (cbor) Negative `BigInteger` values not encoded/decoded correctly
  (reported by Brian G)
  (fix contributed by Fawzi E) 
+#587: (cbor) Allow exposing CBOR Simple values as `JsonToken.VALUE_EMBEDDED_OBJECT`
+  with a feature flag
+ (implementation contributed by Fawzi E)   
 - Generate SBOMs [JSTEP-14]
 
 2.19.0 (24-Apr-2025)
