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 #include "llvm/Testing/Support/Error.h"
12 
13 #include "gmock/gmock.h"
14 #include "gtest/gtest.h"
15 
16 namespace llvm {
17 namespace json {
18 
19 namespace {
20 
s(const Value & E)21 std::string s(const Value &E) { return llvm::formatv("{0}", E).str(); }
sp(const Value & E)22 std::string sp(const Value &E) { return llvm::formatv("{0:2}", E).str(); }
23 
TEST(JSONTest,Types)24 TEST(JSONTest, Types) {
25   EXPECT_EQ("true", s(true));
26   EXPECT_EQ("null", s(nullptr));
27   EXPECT_EQ("2.5", s(2.5));
28   EXPECT_EQ(R"("foo")", s("foo"));
29   EXPECT_EQ("[1,2,3]", s({1, 2, 3}));
30   EXPECT_EQ(R"({"x":10,"y":20})", s(Object{{"x", 10}, {"y", 20}}));
31 
32 #ifdef NDEBUG
33   EXPECT_EQ(R"("��")", s("\xC0\x80"));
34   EXPECT_EQ(R"({"��":0})", s(Object{{"\xC0\x80", 0}}));
35 #else
36   EXPECT_DEATH(s("\xC0\x80"), "Invalid UTF-8");
37   EXPECT_DEATH(s(Object{{"\xC0\x80", 0}}), "Invalid UTF-8");
38 #endif
39 }
40 
TEST(JSONTest,Constructors)41 TEST(JSONTest, Constructors) {
42   // Lots of edge cases around empty and singleton init lists.
43   EXPECT_EQ("[[[3]]]", s({{{3}}}));
44   EXPECT_EQ("[[[]]]", s({{{}}}));
45   EXPECT_EQ("[[{}]]", s({{Object{}}}));
46   EXPECT_EQ(R"({"A":{"B":{}}})", s(Object{{"A", Object{{"B", Object{}}}}}));
47   EXPECT_EQ(R"({"A":{"B":{"X":"Y"}}})",
48             s(Object{{"A", Object{{"B", Object{{"X", "Y"}}}}}}));
49   EXPECT_EQ("null", s(llvm::Optional<double>()));
50   EXPECT_EQ("2.5", s(llvm::Optional<double>(2.5)));
51   EXPECT_EQ("[[2.5,null]]", s(std::vector<std::vector<llvm::Optional<double>>>{
52                                  {2.5, llvm::None}}));
53 }
54 
TEST(JSONTest,StringOwnership)55 TEST(JSONTest, StringOwnership) {
56   char X[] = "Hello";
57   Value Alias = static_cast<const char *>(X);
58   X[1] = 'a';
59   EXPECT_EQ(R"("Hallo")", s(Alias));
60 
61   std::string Y = "Hello";
62   Value Copy = Y;
63   Y[1] = 'a';
64   EXPECT_EQ(R"("Hello")", s(Copy));
65 }
66 
TEST(JSONTest,CanonicalOutput)67 TEST(JSONTest, CanonicalOutput) {
68   // Objects are sorted (but arrays aren't)!
69   EXPECT_EQ(R"({"a":1,"b":2,"c":3})", s(Object{{"a", 1}, {"c", 3}, {"b", 2}}));
70   EXPECT_EQ(R"(["a","c","b"])", s({"a", "c", "b"}));
71   EXPECT_EQ("3", s(3.0));
72 }
73 
TEST(JSONTest,Escaping)74 TEST(JSONTest, Escaping) {
75   std::string Test = {
76       0,                    // Strings may contain nulls.
77       '\b',   '\f',         // Have mnemonics, but we escape numerically.
78       '\r',   '\n',   '\t', // Escaped with mnemonics.
79       'S',    '\"',   '\\', // Printable ASCII characters.
80       '\x7f',               // Delete is not escaped.
81       '\xce', '\x94',       // Non-ASCII UTF-8 is not escaped.
82   };
83 
84   std::string TestString = R"("\u0000\u0008\u000c\r\n\tS\"\\)"
85                            "\x7f\xCE\x94\"";
86 
87   EXPECT_EQ(TestString, s(Test));
88 
89   EXPECT_EQ(R"({"object keys are\nescaped":true})",
90             s(Object{{"object keys are\nescaped", true}}));
91 }
92 
TEST(JSONTest,PrettyPrinting)93 TEST(JSONTest, PrettyPrinting) {
94   const char Str[] = R"({
95   "empty_array": [],
96   "empty_object": {},
97   "full_array": [
98     1,
99     null
100   ],
101   "full_object": {
102     "nested_array": [
103       {
104         "property": "value"
105       }
106     ]
107   }
108 })";
109 
110   EXPECT_EQ(Str, sp(Object{
111                      {"empty_object", Object{}},
112                      {"empty_array", {}},
113                      {"full_array", {1, nullptr}},
114                      {"full_object",
115                       Object{
116                           {"nested_array",
117                            {Object{
118                                {"property", "value"},
119                            }}},
120                       }},
121                  }));
122 }
123 
TEST(JSONTest,Array)124 TEST(JSONTest, Array) {
125   Array A{1, 2};
126   A.emplace_back(3);
127   A.emplace(++A.begin(), 0);
128   A.push_back(4);
129   A.insert(++++A.begin(), 99);
130 
131   EXPECT_EQ(A.size(), 6u);
132   EXPECT_EQ(R"([1,0,99,2,3,4])", s(std::move(A)));
133 }
134 
TEST(JSONTest,Object)135 TEST(JSONTest, Object) {
136   Object O{{"a", 1}, {"b", 2}, {"c", 3}};
137   EXPECT_TRUE(O.try_emplace("d", 4).second);
138   EXPECT_FALSE(O.try_emplace("a", 4).second);
139 
140   auto D = O.find("d");
141   EXPECT_FALSE(D == O.end());
142   auto E = O.find("e");
143   EXPECT_TRUE(E == O.end());
144 
145   O.erase("b");
146   O.erase(D);
147   EXPECT_EQ(O.size(), 2u);
148   EXPECT_EQ(R"({"a":1,"c":3})", s(std::move(O)));
149 }
150 
TEST(JSONTest,Parse)151 TEST(JSONTest, Parse) {
152   auto Compare = [](llvm::StringRef S, Value Expected) {
153     if (auto E = parse(S)) {
154       // Compare both string forms and with operator==, in case we have bugs.
155       EXPECT_EQ(*E, Expected);
156       EXPECT_EQ(sp(*E), sp(Expected));
157     } else {
158       handleAllErrors(E.takeError(), [S](const llvm::ErrorInfoBase &E) {
159         FAIL() << "Failed to parse JSON >>> " << S << " <<<: " << E.message();
160       });
161     }
162   };
163 
164   Compare(R"(true)", true);
165   Compare(R"(false)", false);
166   Compare(R"(null)", nullptr);
167 
168   Compare(R"(42)", 42);
169   Compare(R"(2.5)", 2.5);
170   Compare(R"(2e50)", 2e50);
171   Compare(R"(1.2e3456789)", std::numeric_limits<double>::infinity());
172 
173   Compare(R"("foo")", "foo");
174   Compare(R"("\"\\\b\f\n\r\t")", "\"\\\b\f\n\r\t");
175   Compare(R"("\u0000")", llvm::StringRef("\0", 1));
176   Compare("\"\x7f\"", "\x7f");
177   Compare(R"("\ud801\udc37")", u8"\U00010437"); // UTF16 surrogate pair escape.
178   Compare("\"\xE2\x82\xAC\xF0\x9D\x84\x9E\"", u8"\u20ac\U0001d11e"); // UTF8
179   Compare(
180       R"("LoneLeading=\ud801, LoneTrailing=\udc01, LeadingLeadingTrailing=\ud801\ud801\udc37")",
181       u8"LoneLeading=\ufffd, LoneTrailing=\ufffd, "
182       u8"LeadingLeadingTrailing=\ufffd\U00010437"); // Invalid unicode.
183 
184   Compare(R"({"":0,"":0})", Object{{"", 0}});
185   Compare(R"({"obj":{},"arr":[]})", Object{{"obj", Object{}}, {"arr", {}}});
186   Compare(R"({"\n":{"\u0000":[[[[]]]]}})",
187           Object{{"\n", Object{
188                             {llvm::StringRef("\0", 1), {{{{}}}}},
189                         }}});
190   Compare("\r[\n\t] ", {});
191 }
192 
TEST(JSONTest,ParseErrors)193 TEST(JSONTest, ParseErrors) {
194   auto ExpectErr = [](llvm::StringRef Msg, llvm::StringRef S) {
195     if (auto E = parse(S)) {
196       // Compare both string forms and with operator==, in case we have bugs.
197       FAIL() << "Parsed JSON >>> " << S << " <<< but wanted error: " << Msg;
198     } else {
199       handleAllErrors(E.takeError(), [S, Msg](const llvm::ErrorInfoBase &E) {
200         EXPECT_THAT(E.message(), testing::HasSubstr(std::string(Msg))) << S;
201       });
202     }
203   };
204   ExpectErr("Unexpected EOF", "");
205   ExpectErr("Unexpected EOF", "[");
206   ExpectErr("Text after end of document", "[][]");
207   ExpectErr("Invalid JSON value (false?)", "fuzzy");
208   ExpectErr("Expected , or ]", "[2?]");
209   ExpectErr("Expected object key", "{a:2}");
210   ExpectErr("Expected : after object key", R"({"a",2})");
211   ExpectErr("Expected , or } after object property", R"({"a":2 "b":3})");
212   ExpectErr("Invalid JSON value", R"([&%!])");
213   ExpectErr("Invalid JSON value (number?)", "1e1.0");
214   ExpectErr("Unterminated string", R"("abc\"def)");
215   ExpectErr("Control character in string", "\"abc\ndef\"");
216   ExpectErr("Invalid escape sequence", R"("\030")");
217   ExpectErr("Invalid \\u escape sequence", R"("\usuck")");
218   ExpectErr("[3:3, byte=19]", R"({
219   "valid": 1,
220   invalid: 2
221 })");
222   ExpectErr("Invalid UTF-8 sequence", "\"\xC0\x80\""); // WTF-8 null
223 }
224 
225 // Direct tests of isUTF8 and fixUTF8. Internal uses are also tested elsewhere.
TEST(JSONTest,UTF8)226 TEST(JSONTest, UTF8) {
227   for (const char *Valid : {
228            "this is ASCII text",
229            "thïs tëxt häs BMP chäräctërs",
230            "����L���� C��������",
231        }) {
232     EXPECT_TRUE(isUTF8(Valid)) << Valid;
233     EXPECT_EQ(fixUTF8(Valid), Valid);
234   }
235   for (auto Invalid : std::vector<std::pair<const char *, const char *>>{
236            {"lone trailing \x81\x82 bytes", "lone trailing �� bytes"},
237            {"missing trailing \xD0 bytes", "missing trailing � bytes"},
238            {"truncated character \xD0", "truncated character �"},
239            {"not \xC1\x80 the \xE0\x9f\xBF shortest \xF0\x83\x83\x83 encoding",
240             "not �� the ��� shortest ���� encoding"},
241            {"too \xF9\x80\x80\x80\x80 long", "too ����� long"},
242            {"surrogate \xED\xA0\x80 invalid \xF4\x90\x80\x80",
243             "surrogate ��� invalid ����"}}) {
244     EXPECT_FALSE(isUTF8(Invalid.first)) << Invalid.first;
245     EXPECT_EQ(fixUTF8(Invalid.first), Invalid.second);
246   }
247 }
248 
TEST(JSONTest,Inspection)249 TEST(JSONTest, Inspection) {
250   llvm::Expected<Value> Doc = parse(R"(
251     {
252       "null": null,
253       "boolean": false,
254       "number": 2.78,
255       "string": "json",
256       "array": [null, true, 3.14, "hello", [1,2,3], {"time": "arrow"}],
257       "object": {"fruit": "banana"}
258     }
259   )");
260   EXPECT_TRUE(!!Doc);
261 
262   Object *O = Doc->getAsObject();
263   ASSERT_TRUE(O);
264 
265   EXPECT_FALSE(O->getNull("missing"));
266   EXPECT_FALSE(O->getNull("boolean"));
267   EXPECT_TRUE(O->getNull("null"));
268 
269   EXPECT_EQ(O->getNumber("number"), llvm::Optional<double>(2.78));
270   EXPECT_FALSE(O->getInteger("number"));
271   EXPECT_EQ(O->getString("string"), llvm::Optional<llvm::StringRef>("json"));
272   ASSERT_FALSE(O->getObject("missing"));
273   ASSERT_FALSE(O->getObject("array"));
274   ASSERT_TRUE(O->getObject("object"));
275   EXPECT_EQ(*O->getObject("object"), (Object{{"fruit", "banana"}}));
276 
277   Array *A = O->getArray("array");
278   ASSERT_TRUE(A);
279   EXPECT_EQ((*A)[1].getAsBoolean(), llvm::Optional<bool>(true));
280   ASSERT_TRUE((*A)[4].getAsArray());
281   EXPECT_EQ(*(*A)[4].getAsArray(), (Array{1, 2, 3}));
282   EXPECT_EQ((*(*A)[4].getAsArray())[1].getAsInteger(),
283             llvm::Optional<int64_t>(2));
284   int I = 0;
285   for (Value &E : *A) {
286     if (I++ == 5) {
287       ASSERT_TRUE(E.getAsObject());
288       EXPECT_EQ(E.getAsObject()->getString("time"),
289                 llvm::Optional<llvm::StringRef>("arrow"));
290     } else
291       EXPECT_FALSE(E.getAsObject());
292   }
293 }
294 
295 // Verify special integer handling - we try to preserve exact int64 values.
TEST(JSONTest,Integers)296 TEST(JSONTest, Integers) {
297   struct {
298     const char *Desc;
299     Value Val;
300     const char *Str;
301     llvm::Optional<int64_t> AsInt;
302     llvm::Optional<double> AsNumber;
303   } TestCases[] = {
304       {
305           "Non-integer. Stored as double, not convertible.",
306           double{1.5},
307           "1.5",
308           llvm::None,
309           1.5,
310       },
311 
312       {
313           "Integer, not exact double. Stored as int64, convertible.",
314           int64_t{0x4000000000000001},
315           "4611686018427387905",
316           int64_t{0x4000000000000001},
317           double{0x4000000000000000},
318       },
319 
320       {
321           "Negative integer, not exact double. Stored as int64, convertible.",
322           int64_t{-0x4000000000000001},
323           "-4611686018427387905",
324           int64_t{-0x4000000000000001},
325           double{-0x4000000000000000},
326       },
327 
328       // PR46470,
329       // https://developercommunity.visualstudio.com/content/problem/1093399/incorrect-result-when-printing-6917529027641081856.html
330 #if !defined(_MSC_VER) || _MSC_VER < 1926
331       {
332           "Dynamically exact integer. Stored as double, convertible.",
333           double{0x6000000000000000},
334           "6.9175290276410819e+18",
335           int64_t{0x6000000000000000},
336           double{0x6000000000000000},
337       },
338 #endif
339 
340       {
341           "Dynamically integer, >64 bits. Stored as double, not convertible.",
342           1.5 * double{0x8000000000000000},
343           "1.3835058055282164e+19",
344           llvm::None,
345           1.5 * double{0x8000000000000000},
346       },
347   };
348   for (const auto &T : TestCases) {
349     EXPECT_EQ(T.Str, s(T.Val)) << T.Desc;
350     llvm::Expected<Value> Doc = parse(T.Str);
351     EXPECT_TRUE(!!Doc) << T.Desc;
352     EXPECT_EQ(Doc->getAsInteger(), T.AsInt) << T.Desc;
353     EXPECT_EQ(Doc->getAsNumber(), T.AsNumber) << T.Desc;
354     EXPECT_EQ(T.Val, *Doc) << T.Desc;
355     EXPECT_EQ(T.Str, s(*Doc)) << T.Desc;
356   }
357 }
358 
359 // Sample struct with typical JSON-mapping rules.
360 struct CustomStruct {
CustomStructllvm::json::__anon5e3735e20111::CustomStruct361   CustomStruct() : B(false) {}
CustomStructllvm::json::__anon5e3735e20111::CustomStruct362   CustomStruct(std::string S, llvm::Optional<int> I, bool B)
363       : S(S), I(I), B(B) {}
364   std::string S;
365   llvm::Optional<int> I;
366   bool B;
367 };
operator ==(const CustomStruct & L,const CustomStruct & R)368 inline bool operator==(const CustomStruct &L, const CustomStruct &R) {
369   return L.S == R.S && L.I == R.I && L.B == R.B;
370 }
operator <<(llvm::raw_ostream & OS,const CustomStruct & S)371 inline llvm::raw_ostream &operator<<(llvm::raw_ostream &OS,
372                                      const CustomStruct &S) {
373   return OS << "(" << S.S << ", " << (S.I ? std::to_string(*S.I) : "None")
374             << ", " << S.B << ")";
375 }
fromJSON(const Value & E,CustomStruct & R,Path P)376 bool fromJSON(const Value &E, CustomStruct &R, Path P) {
377   ObjectMapper O(E, P);
378   return O && O.map("str", R.S) && O.map("int", R.I) &&
379          O.mapOptional("bool", R.B);
380 }
381 
errorContext(const Value & V,const Path::Root & R)382 static std::string errorContext(const Value &V, const Path::Root &R) {
383   std::string Context;
384   llvm::raw_string_ostream OS(Context);
385   R.printErrorContext(V, OS);
386   return OS.str();
387 }
388 
TEST(JSONTest,Deserialize)389 TEST(JSONTest, Deserialize) {
390   std::map<std::string, std::vector<CustomStruct>> R;
391   CustomStruct ExpectedStruct = {"foo", 42, true};
392   std::map<std::string, std::vector<CustomStruct>> Expected;
393   Value J = Object{{"foo", Array{
394                                Object{
395                                    {"str", "foo"},
396                                    {"int", 42},
397                                    {"bool", true},
398                                    {"unknown", "ignored"},
399                                },
400                                Object{{"str", "bar"}},
401                            }}};
402   Expected["foo"] = {
403       CustomStruct("foo", 42, true),
404       CustomStruct("bar", llvm::None, false),
405   };
406   Path::Root Root("CustomStruct");
407   ASSERT_TRUE(fromJSON(J, R, Root));
408   EXPECT_EQ(R, Expected);
409 
410   (*J.getAsObject()->getArray("foo"))[0] = 123;
411   ASSERT_FALSE(fromJSON(J, R, Root));
412   EXPECT_EQ("expected object at CustomStruct.foo[0]",
413             toString(Root.getError()));
414   const char *ExpectedDump = R"({
415   "foo": [
416     /* error: expected object */
417     123,
418     { ... }
419   ]
420 })";
421   EXPECT_EQ(ExpectedDump, errorContext(J, Root));
422 
423   CustomStruct V;
424   EXPECT_FALSE(fromJSON(nullptr, V, Root));
425   EXPECT_EQ("expected object when parsing CustomStruct",
426             toString(Root.getError()));
427 
428   EXPECT_FALSE(fromJSON(Object{}, V, Root));
429   EXPECT_EQ("missing value at CustomStruct.str", toString(Root.getError()));
430 
431   EXPECT_FALSE(fromJSON(Object{{"str", 1}}, V, Root));
432   EXPECT_EQ("expected string at CustomStruct.str", toString(Root.getError()));
433 
434   // Optional<T> must parse as the correct type if present.
435   EXPECT_FALSE(fromJSON(Object{{"str", "1"}, {"int", "string"}}, V, Root));
436   EXPECT_EQ("expected integer at CustomStruct.int", toString(Root.getError()));
437 
438   // mapOptional must parse as the correct type if present.
439   EXPECT_FALSE(fromJSON(Object{{"str", "1"}, {"bool", "string"}}, V, Root));
440   EXPECT_EQ("expected boolean at CustomStruct.bool", toString(Root.getError()));
441 }
442 
TEST(JSONTest,ParseDeserialize)443 TEST(JSONTest, ParseDeserialize) {
444   auto E = parse<std::vector<CustomStruct>>(R"json(
445     [{"str": "foo", "int": 42}, {"int": 42}]
446   )json");
447   EXPECT_THAT_EXPECTED(E, FailedWithMessage("missing value at (root)[1].str"));
448 
449   E = parse<std::vector<CustomStruct>>(R"json(
450     [{"str": "foo", "int": 42}, {"str": "bar"}
451   )json");
452   EXPECT_THAT_EXPECTED(
453       E,
454       FailedWithMessage("[3:2, byte=50]: Expected , or ] after array element"));
455 
456   E = parse<std::vector<CustomStruct>>(R"json(
457     [{"str": "foo", "int": 42}]
458   )json");
459   EXPECT_THAT_EXPECTED(E, Succeeded());
460   EXPECT_THAT(*E, testing::SizeIs(1));
461 }
462 
TEST(JSONTest,Stream)463 TEST(JSONTest, Stream) {
464   auto StreamStuff = [](unsigned Indent) {
465     std::string S;
466     llvm::raw_string_ostream OS(S);
467     OStream J(OS, Indent);
468     J.comment("top*/level");
469     J.object([&] {
470       J.attributeArray("foo", [&] {
471         J.value(nullptr);
472         J.comment("element");
473         J.value(42.5);
474         J.arrayBegin();
475         J.value(43);
476         J.arrayEnd();
477         J.rawValue([](raw_ostream &OS) { OS << "'unverified\nraw value'"; });
478       });
479       J.comment("attribute");
480       J.attributeBegin("bar");
481       J.comment("attribute value");
482       J.objectBegin();
483       J.objectEnd();
484       J.attributeEnd();
485       J.attribute("baz", "xyz");
486     });
487     return OS.str();
488   };
489 
490   const char *Plain =
491       R"(/*top* /level*/{"foo":[null,/*element*/42.5,[43],'unverified
492 raw value'],/*attribute*/"bar":/*attribute value*/{},"baz":"xyz"})";
493   EXPECT_EQ(Plain, StreamStuff(0));
494   const char *Pretty = R"(/* top* /level */
495 {
496   "foo": [
497     null,
498     /* element */
499     42.5,
500     [
501       43
502     ],
503     'unverified
504 raw value'
505   ],
506   /* attribute */
507   "bar": /* attribute value */ {},
508   "baz": "xyz"
509 })";
510   EXPECT_EQ(Pretty, StreamStuff(2));
511 }
512 
TEST(JSONTest,Path)513 TEST(JSONTest, Path) {
514   Path::Root R("foo");
515   Path P = R, A = P.field("a"), B = P.field("b");
516   P.report("oh no");
517   EXPECT_THAT_ERROR(R.getError(), FailedWithMessage("oh no when parsing foo"));
518   A.index(1).field("c").index(2).report("boom");
519   EXPECT_THAT_ERROR(R.getError(), FailedWithMessage("boom at foo.a[1].c[2]"));
520   B.field("d").field("e").report("bam");
521   EXPECT_THAT_ERROR(R.getError(), FailedWithMessage("bam at foo.b.d.e"));
522 
523   Value V = Object{
524       {"a", Array{42}},
525       {"b",
526        Object{{"d",
527                Object{
528                    {"e", Array{1, Object{{"x", "y"}}}},
529                    {"f", "a moderately long string: 48 characters in total"},
530                }}}},
531   };
532   const char *Expected = R"({
533   "a": [ ... ],
534   "b": {
535     "d": {
536       "e": /* error: bam */ [
537         1,
538         { ... }
539       ],
540       "f": "a moderately long string: 48 characte..."
541     }
542   }
543 })";
544   EXPECT_EQ(Expected, errorContext(V, R));
545 }
546 
547 } // namespace
548 } // namespace json
549 } // namespace llvm
550