From 28f14bd68f30872417a6c96dc4be1e7195c24f5d Mon Sep 17 00:00:00 2001 From: miloyip Date: Sun, 3 May 2015 09:51:15 +0800 Subject: [PATCH] Add parsing of URI fragment representation of JSON pointer --- include/rapidjson/pointer.h | 44 ++++++++- test/unittest/pointertest.cpp | 178 ++++++++++++++++++++++++++++++++++ 2 files changed, 220 insertions(+), 2 deletions(-) diff --git a/include/rapidjson/pointer.h b/include/rapidjson/pointer.h index c0b2253..4086751 100644 --- a/include/rapidjson/pointer.h +++ b/include/rapidjson/pointer.h @@ -25,7 +25,9 @@ enum PointerParseErrorCode { kPointerParseErrorNone = 0, kPointerParseErrorTokenMustBeginWithSolidus, - kPointerParseErrorInvalidEscape + kPointerParseErrorInvalidEscape, + kPointerParseErrorInvalidPercentEncoding, + kPointerParseErrorCharacterMustPercentEncode }; template @@ -363,6 +365,12 @@ public: } private: + //! Parse a JSON String or its URI fragment representation into tokens. + /*! + \param source Either a JSON Pointer string, or its URI fragment representation. Not need to be null terminated. + \param length Length of the source string. + \note Source cannot be JSON String Representation of JSON Pointer, e.g. In "/\u0000", \u0000 will not be unescaped. + */ void Parse(const Ch* source, size_t length) { // Create own allocator if user did not supply. if (!allocator_) @@ -380,7 +388,14 @@ private: size_t i = 0; - if (length != 0 && source[i] != '/') { + // Detect if it is a URI fragment + bool uriFragment = false; + if (source[i] == '#') { + uriFragment = true; + i++; + } + + if (i != length && source[i] != '/') { parseErrorCode_ = kPointerParseErrorTokenMustBeginWithSolidus; goto error; } @@ -395,6 +410,31 @@ private: while (i < length && source[i] != '/') { Ch c = source[i++]; + + if (uriFragment) { + // Decoding percent-encoding for URI fragment + if (c == '%') { + c = 0; + for (int j = 0; j < 2; j++) { + c <<= 4; + Ch h = source[i]; + if (h >= '0' && h <= '9') c += h - '0'; + else if (h >= 'A' && h <= 'F') c += h - 'A' + 10; + else if (h >= 'a' && h <= 'f') c += h - 'a' + 10; + else { + parseErrorCode_ = kPointerParseErrorInvalidPercentEncoding; + goto error; + } + i++; + } + } + else if (!((c >= '0' && c <= '9') || (c >= 'A' && c <='Z') || (c >= 'a' && c <= 'z') || c == '-' || c == '.' || c == '_' || c =='~')) { + // RFC 3986 2.3 Unreserved Characters + i--; + parseErrorCode_ = kPointerParseErrorCharacterMustPercentEncode; + goto error; + } + } // Escaping "~0" -> '~', "~1" -> '/' if (c == '~') { diff --git a/test/unittest/pointertest.cpp b/test/unittest/pointertest.cpp index 75ca5a3..32f659b 100644 --- a/test/unittest/pointertest.cpp +++ b/test/unittest/pointertest.cpp @@ -165,6 +165,184 @@ TEST(Pointer, Parse) { } } +TEST(Pointer, Parse_URIFragment) { + { + Pointer p("#"); + EXPECT_TRUE(p.IsValid()); + EXPECT_EQ(0u, p.GetTokenCount()); + } + + { + Pointer p("#/foo"); + EXPECT_TRUE(p.IsValid()); + EXPECT_EQ(1u, p.GetTokenCount()); + EXPECT_EQ(3u, p.GetTokens()[0].length); + EXPECT_STREQ("foo", p.GetTokens()[0].name); + } + + { + Pointer p("#/foo/0"); + EXPECT_TRUE(p.IsValid()); + EXPECT_EQ(2u, p.GetTokenCount()); + EXPECT_EQ(3u, p.GetTokens()[0].length); + EXPECT_STREQ("foo", p.GetTokens()[0].name); + EXPECT_EQ(1u, p.GetTokens()[1].length); + EXPECT_STREQ("0", p.GetTokens()[1].name); + EXPECT_EQ(0u, p.GetTokens()[1].index); + } + + { + // Unescape ~1 + Pointer p("#/a~1b"); + EXPECT_TRUE(p.IsValid()); + EXPECT_EQ(1u, p.GetTokenCount()); + EXPECT_EQ(3u, p.GetTokens()[0].length); + EXPECT_STREQ("a/b", p.GetTokens()[0].name); + } + + { + // Unescape ~0 + Pointer p("#/m~0n"); + EXPECT_TRUE(p.IsValid()); + EXPECT_EQ(1u, p.GetTokenCount()); + EXPECT_EQ(3u, p.GetTokens()[0].length); + EXPECT_STREQ("m~n", p.GetTokens()[0].name); + } + + { + // empty name + Pointer p("#/"); + EXPECT_TRUE(p.IsValid()); + EXPECT_EQ(1u, p.GetTokenCount()); + EXPECT_EQ(0u, p.GetTokens()[0].length); + EXPECT_STREQ("", p.GetTokens()[0].name); + } + + { + // empty and non-empty name + Pointer p("#//a"); + EXPECT_TRUE(p.IsValid()); + EXPECT_EQ(2u, p.GetTokenCount()); + EXPECT_EQ(0u, p.GetTokens()[0].length); + EXPECT_STREQ("", p.GetTokens()[0].name); + EXPECT_EQ(1u, p.GetTokens()[1].length); + EXPECT_STREQ("a", p.GetTokens()[1].name); + } + + { + // Null characters + Pointer p("#/%00%00"); + EXPECT_TRUE(p.IsValid()); + EXPECT_EQ(1u, p.GetTokenCount()); + EXPECT_EQ(2u, p.GetTokens()[0].length); + EXPECT_EQ('\0', p.GetTokens()[0].name[0]); + EXPECT_EQ('\0', p.GetTokens()[0].name[1]); + EXPECT_EQ('\0', p.GetTokens()[0].name[2]); + } + + { + // Percentage Escapes + EXPECT_STREQ("c%d", Pointer("#/c%25d").GetTokens()[0].name); + EXPECT_STREQ("e^f", Pointer("#/e%5Ef").GetTokens()[0].name); + EXPECT_STREQ("g|h", Pointer("#/g%7Ch").GetTokens()[0].name); + EXPECT_STREQ("i\\j", Pointer("#/i%5Cj").GetTokens()[0].name); + EXPECT_STREQ("k\"l", Pointer("#/k%22l").GetTokens()[0].name); + EXPECT_STREQ(" ", Pointer("#/%20").GetTokens()[0].name); + } + + { + // Valid index + Pointer p("#/123"); + EXPECT_TRUE(p.IsValid()); + EXPECT_EQ(1u, p.GetTokenCount()); + EXPECT_STREQ("123", p.GetTokens()[0].name); + EXPECT_EQ(123u, p.GetTokens()[0].index); + } + + { + // Invalid index (with leading zero) + Pointer p("#/01"); + EXPECT_TRUE(p.IsValid()); + EXPECT_EQ(1u, p.GetTokenCount()); + EXPECT_STREQ("01", p.GetTokens()[0].name); + EXPECT_EQ(kPointerInvalidIndex, p.GetTokens()[0].index); + } + + if (sizeof(SizeType) == 4) { + // Invalid index (overflow) + Pointer p("#/4294967296"); + EXPECT_TRUE(p.IsValid()); + EXPECT_EQ(1u, p.GetTokenCount()); + EXPECT_STREQ("4294967296", p.GetTokens()[0].name); + EXPECT_EQ(kPointerInvalidIndex, p.GetTokens()[0].index); + } + + { + // kPointerParseErrorTokenMustBeginWithSolidus + Pointer p("# "); + EXPECT_FALSE(p.IsValid()); + EXPECT_EQ(kPointerParseErrorTokenMustBeginWithSolidus, p.GetParseErrorCode()); + EXPECT_EQ(1u, p.GetParseErrorOffset()); + } + + { + // kPointerParseErrorInvalidEscape + Pointer p("#/~"); + EXPECT_FALSE(p.IsValid()); + EXPECT_EQ(kPointerParseErrorInvalidEscape, p.GetParseErrorCode()); + EXPECT_EQ(3u, p.GetParseErrorOffset()); + } + + { + // kPointerParseErrorInvalidEscape + Pointer p("#/~2"); + EXPECT_FALSE(p.IsValid()); + EXPECT_EQ(kPointerParseErrorInvalidEscape, p.GetParseErrorCode()); + EXPECT_EQ(3u, p.GetParseErrorOffset()); + } + + { + // kPointerParseErrorInvalidPercentEncoding + Pointer p("#/%"); + EXPECT_FALSE(p.IsValid()); + EXPECT_EQ(kPointerParseErrorInvalidPercentEncoding, p.GetParseErrorCode()); + EXPECT_EQ(3u, p.GetParseErrorOffset()); + } + + { + // kPointerParseErrorInvalidPercentEncoding + Pointer p("#/%g0"); + EXPECT_FALSE(p.IsValid()); + EXPECT_EQ(kPointerParseErrorInvalidPercentEncoding, p.GetParseErrorCode()); + EXPECT_EQ(3u, p.GetParseErrorOffset()); + } + + { + // kPointerParseErrorInvalidPercentEncoding + Pointer p("#/%0g"); + EXPECT_FALSE(p.IsValid()); + EXPECT_EQ(kPointerParseErrorInvalidPercentEncoding, p.GetParseErrorCode()); + EXPECT_EQ(4u, p.GetParseErrorOffset()); + } + + { + // kPointerParseErrorCharacterMustPercentEncode + Pointer p("#/ "); + EXPECT_FALSE(p.IsValid()); + EXPECT_EQ(kPointerParseErrorCharacterMustPercentEncode, p.GetParseErrorCode()); + EXPECT_EQ(2u, p.GetParseErrorOffset()); + } + + { + // kPointerParseErrorCharacterMustPercentEncode + Pointer p("#/\\"); + EXPECT_FALSE(p.IsValid()); + EXPECT_EQ(kPointerParseErrorCharacterMustPercentEncode, p.GetParseErrorCode()); + EXPECT_EQ(2u, p.GetParseErrorOffset()); + } + +} + TEST(Pointer, Stringify) { // Test by roundtrip const char* sources[] = {