Merge pull request #584 from ludocode/trailing-commas

Added optional support for trailing commas

doc/dom.md

@@ -116,6 +116,7 @@ Parse flags                   | Meaning
 `kParseStopWhenDoneFlag`      | After parsing a complete JSON root from stream, stop further processing the rest of stream. When this flag is used, parser will not generate `kParseErrorDocumentRootNotSingular` error. Using this flag for parsing multiple JSONs in the same stream.
 `kParseFullPrecisionFlag`     | Parse number in full precision (slower). If this flag is not set, the normal precision (faster) is used. Normal precision has maximum 3 [ULP](http://en.wikipedia.org/wiki/Unit_in_the_last_place) error.
 `kParseCommentsFlag`          | Allow one-line `// ...` and multi-line `/* ... */` comments (relaxed JSON syntax).
+`kParseTrailingCommasFlag`    | Allow trailing commas at the end of objects and arrays (relaxed JSON syntax).
 
 By using a non-type template parameter, instead of a function parameter, C++ compiler can generate code which is optimized for specified combinations, improving speed, and reducing code size (if only using a single specialization). The downside is the flags needed to be determined in compile-time.
 

include/rapidjson/reader.h

@@ -149,6 +149,7 @@ enum ParseFlag {
     kParseFullPrecisionFlag = 16,   //!< Parse number in full precision (but slower).
     kParseCommentsFlag = 32,        //!< Allow one-line (//) and multi-line (/**/) comments.
     kParseNumbersAsStringsFlag = 64,    //!< Parse all numbers (ints/doubles) as strings.
+    kParseTrailingCommasFlag = 128, //!< Allow trailing commas at the end of objects and arrays.
     kParseDefaultFlags = RAPIDJSON_PARSE_DEFAULT_FLAGS  //!< Default parse flags. Can be customized by defining RAPIDJSON_PARSE_DEFAULT_FLAGS
 };
 
@@ -636,6 +637,15 @@ private:
                     RAPIDJSON_PARSE_ERROR(kParseErrorObjectMissCommaOrCurlyBracket, is.Tell());
                     break;
             }
+
+            if (parseFlags & kParseTrailingCommasFlag) {
+                if (is.Peek() == '}') {
+                    if (RAPIDJSON_UNLIKELY(!handler.EndObject(memberCount)))
+                        RAPIDJSON_PARSE_ERROR(kParseErrorTermination, is.Tell());
+                    is.Take();
+                    return;
+                }
+            }
         }
     }
 
@@ -676,6 +686,15 @@ private:
             }
             else
                 RAPIDJSON_PARSE_ERROR(kParseErrorArrayMissCommaOrSquareBracket, is.Tell());
+
+            if (parseFlags & kParseTrailingCommasFlag) {
+                if (is.Peek() == ']') {
+                    if (RAPIDJSON_UNLIKELY(!handler.EndArray(elementCount)))
+                        RAPIDJSON_PARSE_ERROR(kParseErrorTermination, is.Tell());
+                    is.Take();
+                    return;
+                }
+            }
         }
     }
 
@@ -1522,7 +1541,7 @@ private:
                 IterativeParsingErrorState,         // Left bracket
                 IterativeParsingErrorState,         // Right bracket
                 IterativeParsingErrorState,         // Left curly bracket
-                IterativeParsingErrorState,         // Right curly bracket
+                IterativeParsingObjectFinishState,  // Right curly bracket
                 IterativeParsingErrorState,         // Comma
                 IterativeParsingErrorState,         // Colon
                 IterativeParsingMemberKeyState,     // String
@@ -1568,7 +1587,7 @@ private:
             // ElementDelimiter
             {
                 IterativeParsingArrayInitialState,      // Left bracket(push Element state)
-                IterativeParsingErrorState,             // Right bracket
+                IterativeParsingArrayFinishState,       // Right bracket
                 IterativeParsingObjectInitialState,     // Left curly bracket(push Element state)
                 IterativeParsingErrorState,             // Right curly bracket
                 IterativeParsingErrorState,             // Comma
@@ -1670,6 +1689,11 @@ private:
 
         case IterativeParsingObjectFinishState:
         {
+            // Transit from delimiter is only allowed when trailing commas are enabled
+            if (!(parseFlags & kParseTrailingCommasFlag) && src == IterativeParsingMemberDelimiterState) {
+                RAPIDJSON_PARSE_ERROR_NORETURN(kParseErrorObjectMissName, is.Tell());
+                return IterativeParsingErrorState;
+            }
             // Get member count.
             SizeType c = *stack_.template Pop<SizeType>(1);
             // If the object is not empty, count the last member.
@@ -1695,6 +1719,11 @@ private:
 
         case IterativeParsingArrayFinishState:
         {
+            // Transit from delimiter is only allowed when trailing commas are enabled
+            if (!(parseFlags & kParseTrailingCommasFlag) && src == IterativeParsingElementDelimiterState) {
+                RAPIDJSON_PARSE_ERROR_NORETURN(kParseErrorValueInvalid, is.Tell());
+                return IterativeParsingErrorState;
+            }
             // Get element count.
             SizeType c = *stack_.template Pop<SizeType>(1);
             // If the array is not empty, count the last element.
@@ -1754,6 +1783,9 @@ private:
         case IterativeParsingMemberDelimiterState:  RAPIDJSON_PARSE_ERROR(kParseErrorObjectMissName, is.Tell()); return;
         case IterativeParsingMemberKeyState:        RAPIDJSON_PARSE_ERROR(kParseErrorObjectMissColon, is.Tell()); return;
         case IterativeParsingMemberValueState:      RAPIDJSON_PARSE_ERROR(kParseErrorObjectMissCommaOrCurlyBracket, is.Tell()); return;
+        case IterativeParsingKeyValueDelimiterState:
+        case IterativeParsingArrayInitialState:
+        case IterativeParsingElementDelimiterState: RAPIDJSON_PARSE_ERROR(kParseErrorValueInvalid, is.Tell()); return;
         case IterativeParsingElementState:          RAPIDJSON_PARSE_ERROR(kParseErrorArrayMissCommaOrSquareBracket, is.Tell()); return;
         default:                                    RAPIDJSON_PARSE_ERROR(kParseErrorUnspecificSyntaxError, is.Tell()); return;
         }       

test/unittest/readertest.cpp

@@ -778,6 +778,10 @@ TEST(Reader, ParseArray_Error) {
     TEST_ARRAY_ERROR(kParseErrorArrayMissCommaOrSquareBracket, "[1}", 2);
     TEST_ARRAY_ERROR(kParseErrorArrayMissCommaOrSquareBracket, "[1 2]", 3);
 
+    // Array cannot have a trailing comma (without kParseTrailingCommasFlag);
+    // a value must follow a comma
+    TEST_ARRAY_ERROR(kParseErrorValueInvalid, "[1,]", 3);
+
 #undef TEST_ARRAY_ERROR
 }
 
@@ -978,6 +982,10 @@ TEST(Reader, ParseObject_Error) {
     // Must be a comma or '}' after an object member
     TEST_ERROR(kParseErrorObjectMissCommaOrCurlyBracket, "{\"a\":1]", 6);
 
+    // Object cannot have a trailing comma (without kParseTrailingCommasFlag);
+    // an object member name must follow a comma
+    TEST_ERROR(kParseErrorObjectMissName, "{\"a\":1,}", 7);
+
     // This tests that MemoryStream is checking the length in Peek().
     {
         MemoryStream ms("{\"a\"", 1);
@@ -1119,6 +1127,16 @@ TEST(Reader, IterativeParsing_ErrorHandling) {
     TESTERRORHANDLING("{\"a\": 1", kParseErrorObjectMissCommaOrCurlyBracket, 7u);
     TESTERRORHANDLING("[1 2 3]", kParseErrorArrayMissCommaOrSquareBracket, 3u);
     TESTERRORHANDLING("{\"a: 1", kParseErrorStringMissQuotationMark, 6u);
+    TESTERRORHANDLING("{\"a\":}", kParseErrorValueInvalid, 5u);
+    TESTERRORHANDLING("{\"a\":]", kParseErrorValueInvalid, 5u);
+    TESTERRORHANDLING("[1,2,}", kParseErrorValueInvalid, 5u);
+    TESTERRORHANDLING("[}]", kParseErrorValueInvalid, 1u);
+    TESTERRORHANDLING("[,]", kParseErrorValueInvalid, 1u);
+    TESTERRORHANDLING("[1,,]", kParseErrorValueInvalid, 3u);
+
+    // Trailing commas are not allowed without kParseTrailingCommasFlag
+    TESTERRORHANDLING("{\"a\": 1,}", kParseErrorObjectMissName, 8u);
+    TESTERRORHANDLING("[1,2,3,]", kParseErrorValueInvalid, 7u);
 
     // Any JSON value can be a valid root element in RFC7159.
     TESTERRORHANDLING("\"ab", kParseErrorStringMissQuotationMark, 3u);
@@ -1552,6 +1570,149 @@ TEST(Reader, NumbersAsStrings) {
 	}
 }
 
+template <unsigned extraFlags>
+void TestTrailingCommas() {
+    {
+        StringStream s("[1,2,3,]");
+        ParseArrayHandler<3> h;
+        Reader reader;
+        EXPECT_TRUE(reader.Parse<extraFlags|kParseTrailingCommasFlag>(s, h));
+        EXPECT_EQ(5u, h.step_);
+    }
+    {
+        const char* json = "{ \"hello\" : \"world\", \"t\" : true , \"f\" : false,"
+                "\"n\": null, \"i\":123, \"pi\": 3.1416, \"a\":[1, 2, 3],}";
+        StringStream s(json);
+        ParseObjectHandler h;
+        Reader reader;
+        EXPECT_TRUE(reader.Parse<extraFlags|kParseTrailingCommasFlag>(s, h));
+        EXPECT_EQ(20u, h.step_);
+    }
+    {
+        // whitespace around trailing commas
+        const char* json = "{ \"hello\" : \"world\", \"t\" : true , \"f\" : false,"
+                "\"n\": null, \"i\":123, \"pi\": 3.1416, \"a\":[1, 2, 3\n,\n]\n,\n} ";
+        StringStream s(json);
+        ParseObjectHandler h;
+        Reader reader;
+        EXPECT_TRUE(reader.Parse<extraFlags|kParseTrailingCommasFlag>(s, h));
+        EXPECT_EQ(20u, h.step_);
+    }
+    {
+        // comments around trailing commas
+        const char* json = "{ \"hello\" : \"world\", \"t\" : true , \"f\" : false, \"n\": null,"
+                "\"i\":123, \"pi\": 3.1416, \"a\":[1, 2, 3/*test*/,/*test*/]/*test*/,/*test*/}";
+        StringStream s(json);
+        ParseObjectHandler h;
+        Reader reader;
+        EXPECT_TRUE(reader.Parse<extraFlags|kParseTrailingCommasFlag|kParseCommentsFlag>(s, h));
+        EXPECT_EQ(20u, h.step_);
+    }
+}
+
+TEST(Reader, TrailingCommas) {
+    TestTrailingCommas<kParseNoFlags>();
+}
+
+TEST(Reader, TrailingCommasIterative) {
+    TestTrailingCommas<kParseIterativeFlag>();
+}
+
+template <unsigned extraFlags>
+void TestMultipleTrailingCommaErrors() {
+    // only a single trailing comma is allowed.
+    {
+        StringStream s("[1,2,3,,]");
+        ParseArrayHandler<3> h;
+        Reader reader;
+        ParseResult r = reader.Parse<extraFlags|kParseTrailingCommasFlag>(s, h);
+        EXPECT_TRUE(reader.HasParseError());
+        EXPECT_EQ(kParseErrorValueInvalid, r.Code());
+        EXPECT_EQ(7u, r.Offset());
+    }
+    {
+        const char* json = "{ \"hello\" : \"world\", \"t\" : true , \"f\" : false,"
+                "\"n\": null, \"i\":123, \"pi\": 3.1416, \"a\":[1, 2, 3,],,}";
+        StringStream s(json);
+        ParseObjectHandler h;
+        Reader reader;
+        ParseResult r = reader.Parse<extraFlags|kParseTrailingCommasFlag>(s, h);
+        EXPECT_TRUE(reader.HasParseError());
+        EXPECT_EQ(kParseErrorObjectMissName, r.Code());
+        EXPECT_EQ(95u, r.Offset());
+    }
+}
+
+TEST(Reader, MultipleTrailingCommaErrors) {
+    TestMultipleTrailingCommaErrors<kParseNoFlags>();
+}
+
+TEST(Reader, MultipleTrailingCommaErrorsIterative) {
+    TestMultipleTrailingCommaErrors<kParseIterativeFlag>();
+}
+
+template <unsigned extraFlags>
+void TestEmptyExceptForCommaErrors() {
+    // not allowed even with trailing commas enabled; the
+    // trailing comma must follow a value.
+    {
+        StringStream s("[,]");
+        ParseArrayHandler<3> h;
+        Reader reader;
+        ParseResult r = reader.Parse<extraFlags|kParseTrailingCommasFlag>(s, h);
+        EXPECT_TRUE(reader.HasParseError());
+        EXPECT_EQ(kParseErrorValueInvalid, r.Code());
+        EXPECT_EQ(1u, r.Offset());
+    }
+    {
+        StringStream s("{,}");
+        ParseObjectHandler h;
+        Reader reader;
+        ParseResult r = reader.Parse<extraFlags|kParseTrailingCommasFlag>(s, h);
+        EXPECT_TRUE(reader.HasParseError());
+        EXPECT_EQ(kParseErrorObjectMissName, r.Code());
+        EXPECT_EQ(1u, r.Offset());
+    }
+}
+
+TEST(Reader, EmptyExceptForCommaErrors) {
+    TestEmptyExceptForCommaErrors<kParseNoFlags>();
+}
+
+TEST(Reader, EmptyExceptForCommaErrorsIterative) {
+    TestEmptyExceptForCommaErrors<kParseIterativeFlag>();
+}
+
+template <unsigned extraFlags>
+void TestTrailingCommaHandlerTermination() {
+    {
+        HandlerTerminateAtEndArray h;
+        Reader reader;
+        StringStream s("[1,2,3,]");
+        ParseResult r = reader.Parse<extraFlags|kParseTrailingCommasFlag>(s, h);
+        EXPECT_TRUE(reader.HasParseError());
+        EXPECT_EQ(kParseErrorTermination, r.Code());
+        EXPECT_EQ(7u, r.Offset());
+    }
+    {
+        HandlerTerminateAtEndObject h;
+        Reader reader;
+        StringStream s("{\"t\": true, \"f\": false,}");
+        ParseResult r = reader.Parse<extraFlags|kParseTrailingCommasFlag>(s, h);
+        EXPECT_TRUE(reader.HasParseError());
+        EXPECT_EQ(kParseErrorTermination, r.Code());
+        EXPECT_EQ(23u, r.Offset());
+    }
+}
+
+TEST(Reader, TrailingCommaHandlerTermination) {
+    TestTrailingCommaHandlerTermination<kParseNoFlags>();
+}
+
+TEST(Reader, TrailingCommaHandlerTerminationIterative) {
+    TestTrailingCommaHandlerTermination<kParseIterativeFlag>();
+}
+
 #ifdef __GNUC__
 RAPIDJSON_DIAG_POP
 #endif