1 //===-- JSONTest.cpp - JSON unit tests --------------------------*- C++ -*-===// 2 // 3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. 4 // See https://llvm.org/LICENSE.txt for license information. 5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception 6 // 7 //===----------------------------------------------------------------------===// 8 9 #include "llvm/Support/JSON.h" 10 #include "llvm/Support/raw_ostream.h" 11 12 #include "gmock/gmock.h" 13 #include "gtest/gtest.h" 14 15 namespace llvm { 16 namespace json { 17 18 namespace { 19 20 std::string s(const Value &E) { return llvm::formatv("{0}", E).str(); } 21 std::string sp(const Value &E) { return llvm::formatv("{0:2}", E).str(); } 22 23 TEST(JSONTest, Types) { 24 EXPECT_EQ("true", s(true)); 25 EXPECT_EQ("null", s(nullptr)); 26 EXPECT_EQ("2.5", s(2.5)); 27 EXPECT_EQ(R"("foo")", s("foo")); 28 EXPECT_EQ("[1,2,3]", s({1, 2, 3})); 29 EXPECT_EQ(R"({"x":10,"y":20})", s(Object{{"x", 10}, {"y", 20}})); 30 31 #ifdef NDEBUG 32 EXPECT_EQ(R"("��")", s("\xC0\x80")); 33 EXPECT_EQ(R"({"��":0})", s(Object{{"\xC0\x80", 0}})); 34 #else 35 EXPECT_DEATH(s("\xC0\x80"), "Invalid UTF-8"); 36 EXPECT_DEATH(s(Object{{"\xC0\x80", 0}}), "Invalid UTF-8"); 37 #endif 38 } 39 40 TEST(JSONTest, Constructors) { 41 // Lots of edge cases around empty and singleton init lists. 42 EXPECT_EQ("[[[3]]]", s({{{3}}})); 43 EXPECT_EQ("[[[]]]", s({{{}}})); 44 EXPECT_EQ("[[{}]]", s({{Object{}}})); 45 EXPECT_EQ(R"({"A":{"B":{}}})", s(Object{{"A", Object{{"B", Object{}}}}})); 46 EXPECT_EQ(R"({"A":{"B":{"X":"Y"}}})", 47 s(Object{{"A", Object{{"B", Object{{"X", "Y"}}}}}})); 48 EXPECT_EQ("null", s(llvm::Optional<double>())); 49 EXPECT_EQ("2.5", s(llvm::Optional<double>(2.5))); 50 EXPECT_EQ("[[2.5,null]]", s(std::vector<std::vector<llvm::Optional<double>>>{ 51 {2.5, llvm::None}})); 52 } 53 54 TEST(JSONTest, StringOwnership) { 55 char X[] = "Hello"; 56 Value Alias = static_cast<const char *>(X); 57 X[1] = 'a'; 58 EXPECT_EQ(R"("Hallo")", s(Alias)); 59 60 std::string Y = "Hello"; 61 Value Copy = Y; 62 Y[1] = 'a'; 63 EXPECT_EQ(R"("Hello")", s(Copy)); 64 } 65 66 TEST(JSONTest, CanonicalOutput) { 67 // Objects are sorted (but arrays aren't)! 68 EXPECT_EQ(R"({"a":1,"b":2,"c":3})", s(Object{{"a", 1}, {"c", 3}, {"b", 2}})); 69 EXPECT_EQ(R"(["a","c","b"])", s({"a", "c", "b"})); 70 EXPECT_EQ("3", s(3.0)); 71 } 72 73 TEST(JSONTest, Escaping) { 74 std::string Test = { 75 0, // Strings may contain nulls. 76 '\b', '\f', // Have mnemonics, but we escape numerically. 77 '\r', '\n', '\t', // Escaped with mnemonics. 78 'S', '\"', '\\', // Printable ASCII characters. 79 '\x7f', // Delete is not escaped. 80 '\xce', '\x94', // Non-ASCII UTF-8 is not escaped. 81 }; 82 83 std::string TestString = R"("\u0000\u0008\u000c\r\n\tS\"\\)" 84 "\x7f\xCE\x94\""; 85 86 EXPECT_EQ(TestString, s(Test)); 87 88 EXPECT_EQ(R"({"object keys are\nescaped":true})", 89 s(Object{{"object keys are\nescaped", true}})); 90 } 91 92 TEST(JSONTest, PrettyPrinting) { 93 const char Str[] = R"({ 94 "empty_array": [], 95 "empty_object": {}, 96 "full_array": [ 97 1, 98 null 99 ], 100 "full_object": { 101 "nested_array": [ 102 { 103 "property": "value" 104 } 105 ] 106 } 107 })"; 108 109 EXPECT_EQ(Str, sp(Object{ 110 {"empty_object", Object{}}, 111 {"empty_array", {}}, 112 {"full_array", {1, nullptr}}, 113 {"full_object", 114 Object{ 115 {"nested_array", 116 {Object{ 117 {"property", "value"}, 118 }}}, 119 }}, 120 })); 121 } 122 123 TEST(JSONTest, Array) { 124 Array A{1, 2}; 125 A.emplace_back(3); 126 A.emplace(++A.begin(), 0); 127 A.push_back(4); 128 A.insert(++++A.begin(), 99); 129 130 EXPECT_EQ(A.size(), 6u); 131 EXPECT_EQ(R"([1,0,99,2,3,4])", s(std::move(A))); 132 } 133 134 TEST(JSONTest, Object) { 135 Object O{{"a", 1}, {"b", 2}, {"c", 3}}; 136 EXPECT_TRUE(O.try_emplace("d", 4).second); 137 EXPECT_FALSE(O.try_emplace("a", 4).second); 138 139 auto D = O.find("d"); 140 EXPECT_FALSE(D == O.end()); 141 auto E = O.find("e"); 142 EXPECT_TRUE(E == O.end()); 143 144 O.erase("b"); 145 O.erase(D); 146 EXPECT_EQ(O.size(), 2u); 147 EXPECT_EQ(R"({"a":1,"c":3})", s(std::move(O))); 148 } 149 150 TEST(JSONTest, Parse) { 151 auto Compare = [](llvm::StringRef S, Value Expected) { 152 if (auto E = parse(S)) { 153 // Compare both string forms and with operator==, in case we have bugs. 154 EXPECT_EQ(*E, Expected); 155 EXPECT_EQ(sp(*E), sp(Expected)); 156 } else { 157 handleAllErrors(E.takeError(), [S](const llvm::ErrorInfoBase &E) { 158 FAIL() << "Failed to parse JSON >>> " << S << " <<<: " << E.message(); 159 }); 160 } 161 }; 162 163 Compare(R"(true)", true); 164 Compare(R"(false)", false); 165 Compare(R"(null)", nullptr); 166 167 Compare(R"(42)", 42); 168 Compare(R"(2.5)", 2.5); 169 Compare(R"(2e50)", 2e50); 170 Compare(R"(1.2e3456789)", std::numeric_limits<double>::infinity()); 171 172 Compare(R"("foo")", "foo"); 173 Compare(R"("\"\\\b\f\n\r\t")", "\"\\\b\f\n\r\t"); 174 Compare(R"("\u0000")", llvm::StringRef("\0", 1)); 175 Compare("\"\x7f\"", "\x7f"); 176 Compare(R"("\ud801\udc37")", u8"\U00010437"); // UTF16 surrogate pair escape. 177 Compare("\"\xE2\x82\xAC\xF0\x9D\x84\x9E\"", u8"\u20ac\U0001d11e"); // UTF8 178 Compare( 179 R"("LoneLeading=\ud801, LoneTrailing=\udc01, LeadingLeadingTrailing=\ud801\ud801\udc37")", 180 u8"LoneLeading=\ufffd, LoneTrailing=\ufffd, " 181 u8"LeadingLeadingTrailing=\ufffd\U00010437"); // Invalid unicode. 182 183 Compare(R"({"":0,"":0})", Object{{"", 0}}); 184 Compare(R"({"obj":{},"arr":[]})", Object{{"obj", Object{}}, {"arr", {}}}); 185 Compare(R"({"\n":{"\u0000":[[[[]]]]}})", 186 Object{{"\n", Object{ 187 {llvm::StringRef("\0", 1), {{{{}}}}}, 188 }}}); 189 Compare("\r[\n\t] ", {}); 190 } 191 192 TEST(JSONTest, ParseErrors) { 193 auto ExpectErr = [](llvm::StringRef Msg, llvm::StringRef S) { 194 if (auto E = parse(S)) { 195 // Compare both string forms and with operator==, in case we have bugs. 196 FAIL() << "Parsed JSON >>> " << S << " <<< but wanted error: " << Msg; 197 } else { 198 handleAllErrors(E.takeError(), [S, Msg](const llvm::ErrorInfoBase &E) { 199 EXPECT_THAT(E.message(), testing::HasSubstr(std::string(Msg))) << S; 200 }); 201 } 202 }; 203 ExpectErr("Unexpected EOF", ""); 204 ExpectErr("Unexpected EOF", "["); 205 ExpectErr("Text after end of document", "[][]"); 206 ExpectErr("Invalid JSON value (false?)", "fuzzy"); 207 ExpectErr("Expected , or ]", "[2?]"); 208 ExpectErr("Expected object key", "{a:2}"); 209 ExpectErr("Expected : after object key", R"({"a",2})"); 210 ExpectErr("Expected , or } after object property", R"({"a":2 "b":3})"); 211 ExpectErr("Invalid JSON value", R"([&%!])"); 212 ExpectErr("Invalid JSON value (number?)", "1e1.0"); 213 ExpectErr("Unterminated string", R"("abc\"def)"); 214 ExpectErr("Control character in string", "\"abc\ndef\""); 215 ExpectErr("Invalid escape sequence", R"("\030")"); 216 ExpectErr("Invalid \\u escape sequence", R"("\usuck")"); 217 ExpectErr("[3:3, byte=19]", R"({ 218 "valid": 1, 219 invalid: 2 220 })"); 221 ExpectErr("Invalid UTF-8 sequence", "\"\xC0\x80\""); // WTF-8 null 222 } 223 224 // Direct tests of isUTF8 and fixUTF8. Internal uses are also tested elsewhere. 225 TEST(JSONTest, UTF8) { 226 for (const char *Valid : { 227 "this is ASCII text", 228 "thïs tëxt häs BMP chäräctërs", 229 "����L���� C��������", 230 }) { 231 EXPECT_TRUE(isUTF8(Valid)) << Valid; 232 EXPECT_EQ(fixUTF8(Valid), Valid); 233 } 234 for (auto Invalid : std::vector<std::pair<const char *, const char *>>{ 235 {"lone trailing \x81\x82 bytes", "lone trailing �� bytes"}, 236 {"missing trailing \xD0 bytes", "missing trailing � bytes"}, 237 {"truncated character \xD0", "truncated character �"}, 238 {"not \xC1\x80 the \xE0\x9f\xBF shortest \xF0\x83\x83\x83 encoding", 239 "not �� the ��� shortest ���� encoding"}, 240 {"too \xF9\x80\x80\x80\x80 long", "too ����� long"}, 241 {"surrogate \xED\xA0\x80 invalid \xF4\x90\x80\x80", 242 "surrogate ��� invalid ����"}}) { 243 EXPECT_FALSE(isUTF8(Invalid.first)) << Invalid.first; 244 EXPECT_EQ(fixUTF8(Invalid.first), Invalid.second); 245 } 246 } 247 248 TEST(JSONTest, Inspection) { 249 llvm::Expected<Value> Doc = parse(R"( 250 { 251 "null": null, 252 "boolean": false, 253 "number": 2.78, 254 "string": "json", 255 "array": [null, true, 3.14, "hello", [1,2,3], {"time": "arrow"}], 256 "object": {"fruit": "banana"} 257 } 258 )"); 259 EXPECT_TRUE(!!Doc); 260 261 Object *O = Doc->getAsObject(); 262 ASSERT_TRUE(O); 263 264 EXPECT_FALSE(O->getNull("missing")); 265 EXPECT_FALSE(O->getNull("boolean")); 266 EXPECT_TRUE(O->getNull("null")); 267 268 EXPECT_EQ(O->getNumber("number"), llvm::Optional<double>(2.78)); 269 EXPECT_FALSE(O->getInteger("number")); 270 EXPECT_EQ(O->getString("string"), llvm::Optional<llvm::StringRef>("json")); 271 ASSERT_FALSE(O->getObject("missing")); 272 ASSERT_FALSE(O->getObject("array")); 273 ASSERT_TRUE(O->getObject("object")); 274 EXPECT_EQ(*O->getObject("object"), (Object{{"fruit", "banana"}})); 275 276 Array *A = O->getArray("array"); 277 ASSERT_TRUE(A); 278 EXPECT_EQ((*A)[1].getAsBoolean(), llvm::Optional<bool>(true)); 279 ASSERT_TRUE((*A)[4].getAsArray()); 280 EXPECT_EQ(*(*A)[4].getAsArray(), (Array{1, 2, 3})); 281 EXPECT_EQ((*(*A)[4].getAsArray())[1].getAsInteger(), 282 llvm::Optional<int64_t>(2)); 283 int I = 0; 284 for (Value &E : *A) { 285 if (I++ == 5) { 286 ASSERT_TRUE(E.getAsObject()); 287 EXPECT_EQ(E.getAsObject()->getString("time"), 288 llvm::Optional<llvm::StringRef>("arrow")); 289 } else 290 EXPECT_FALSE(E.getAsObject()); 291 } 292 } 293 294 // Verify special integer handling - we try to preserve exact int64 values. 295 TEST(JSONTest, Integers) { 296 struct { 297 const char *Desc; 298 Value Val; 299 const char *Str; 300 llvm::Optional<int64_t> AsInt; 301 llvm::Optional<double> AsNumber; 302 } TestCases[] = { 303 { 304 "Non-integer. Stored as double, not convertible.", 305 double{1.5}, 306 "1.5", 307 llvm::None, 308 1.5, 309 }, 310 311 { 312 "Integer, not exact double. Stored as int64, convertible.", 313 int64_t{0x4000000000000001}, 314 "4611686018427387905", 315 int64_t{0x4000000000000001}, 316 double{0x4000000000000000}, 317 }, 318 319 { 320 "Negative integer, not exact double. Stored as int64, convertible.", 321 int64_t{-0x4000000000000001}, 322 "-4611686018427387905", 323 int64_t{-0x4000000000000001}, 324 double{-0x4000000000000000}, 325 }, 326 327 { 328 "Dynamically exact integer. Stored as double, convertible.", 329 double{0x6000000000000000}, 330 "6.9175290276410819e+18", 331 int64_t{0x6000000000000000}, 332 double{0x6000000000000000}, 333 }, 334 335 { 336 "Dynamically integer, >64 bits. Stored as double, not convertible.", 337 1.5 * double{0x8000000000000000}, 338 "1.3835058055282164e+19", 339 llvm::None, 340 1.5 * double{0x8000000000000000}, 341 }, 342 }; 343 for (const auto &T : TestCases) { 344 EXPECT_EQ(T.Str, s(T.Val)) << T.Desc; 345 llvm::Expected<Value> Doc = parse(T.Str); 346 EXPECT_TRUE(!!Doc) << T.Desc; 347 EXPECT_EQ(Doc->getAsInteger(), T.AsInt) << T.Desc; 348 EXPECT_EQ(Doc->getAsNumber(), T.AsNumber) << T.Desc; 349 EXPECT_EQ(T.Val, *Doc) << T.Desc; 350 EXPECT_EQ(T.Str, s(*Doc)) << T.Desc; 351 } 352 } 353 354 // Sample struct with typical JSON-mapping rules. 355 struct CustomStruct { 356 CustomStruct() : B(false) {} 357 CustomStruct(std::string S, llvm::Optional<int> I, bool B) 358 : S(S), I(I), B(B) {} 359 std::string S; 360 llvm::Optional<int> I; 361 bool B; 362 }; 363 inline bool operator==(const CustomStruct &L, const CustomStruct &R) { 364 return L.S == R.S && L.I == R.I && L.B == R.B; 365 } 366 inline llvm::raw_ostream &operator<<(llvm::raw_ostream &OS, 367 const CustomStruct &S) { 368 return OS << "(" << S.S << ", " << (S.I ? std::to_string(*S.I) : "None") 369 << ", " << S.B << ")"; 370 } 371 bool fromJSON(const Value &E, CustomStruct &R) { 372 ObjectMapper O(E); 373 if (!O || !O.map("str", R.S) || !O.map("int", R.I)) 374 return false; 375 O.map("bool", R.B); 376 return true; 377 } 378 379 TEST(JSONTest, Deserialize) { 380 std::map<std::string, std::vector<CustomStruct>> R; 381 CustomStruct ExpectedStruct = {"foo", 42, true}; 382 std::map<std::string, std::vector<CustomStruct>> Expected; 383 Value J = Object{ 384 {"foo", 385 Array{ 386 Object{ 387 {"str", "foo"}, 388 {"int", 42}, 389 {"bool", true}, 390 {"unknown", "ignored"}, 391 }, 392 Object{{"str", "bar"}}, 393 Object{ 394 {"str", "baz"}, {"bool", "string"}, // OK, deserialize ignores. 395 }, 396 }}}; 397 Expected["foo"] = { 398 CustomStruct("foo", 42, true), 399 CustomStruct("bar", llvm::None, false), 400 CustomStruct("baz", llvm::None, false), 401 }; 402 ASSERT_TRUE(fromJSON(J, R)); 403 EXPECT_EQ(R, Expected); 404 405 CustomStruct V; 406 EXPECT_FALSE(fromJSON(nullptr, V)) << "Not an object " << V; 407 EXPECT_FALSE(fromJSON(Object{}, V)) << "Missing required field " << V; 408 EXPECT_FALSE(fromJSON(Object{{"str", 1}}, V)) << "Wrong type " << V; 409 // Optional<T> must parse as the correct type if present. 410 EXPECT_FALSE(fromJSON(Object{{"str", 1}, {"int", "string"}}, V)) 411 << "Wrong type for Optional<T> " << V; 412 } 413 414 TEST(JSONTest, Stream) { 415 auto StreamStuff = [](unsigned Indent) { 416 std::string S; 417 llvm::raw_string_ostream OS(S); 418 OStream J(OS, Indent); 419 J.object([&] { 420 J.attributeArray("foo", [&] { 421 J.value(nullptr); 422 J.value(42.5); 423 J.arrayBegin(); 424 J.value(43); 425 J.arrayEnd(); 426 }); 427 J.attributeBegin("bar"); 428 J.objectBegin(); 429 J.objectEnd(); 430 J.attributeEnd(); 431 J.attribute("baz", "xyz"); 432 }); 433 return OS.str(); 434 }; 435 436 const char *Plain = R"({"foo":[null,42.5,[43]],"bar":{},"baz":"xyz"})"; 437 EXPECT_EQ(Plain, StreamStuff(0)); 438 const char *Pretty = R"({ 439 "foo": [ 440 null, 441 42.5, 442 [ 443 43 444 ] 445 ], 446 "bar": {}, 447 "baz": "xyz" 448 })"; 449 EXPECT_EQ(Pretty, StreamStuff(2)); 450 } 451 452 } // namespace 453 } // namespace json 454 } // namespace llvm 455