unittests/Support/JSONTest.cpp

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