1 //===-- JSONTest.cpp - JSON unit tests --------------------------*- C++ -*-===//
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
7 //===----------------------------------------------------------------------===//
9 #include "llvm/Support/JSON.h"
10 #include "llvm/Support/raw_ostream.h"
11 #include "llvm/Testing/Support/Error.h"
13 #include "gmock/gmock.h"
14 #include "gtest/gtest.h"
21 std::string
s(const Value
&E
) { return llvm::formatv("{0}", E
).str(); }
22 std::string
sp(const Value
&E
) { return llvm::formatv("{0:2}", E
).str(); }
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}}));
33 EXPECT_EQ(R
"("��
")", s("\xC0\x80"));
34 EXPECT_EQ(R
"({"��
":0})", s(Object
{{"\xC0\x80", 0}}));
36 EXPECT_DEATH(s("\xC0\x80"), "Invalid UTF-8");
37 EXPECT_DEATH(s(Object
{{"\xC0\x80", 0}}), "Invalid UTF-8");
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(std::optional
<double>()));
50 EXPECT_EQ("2.5", s(std::optional
<double>(2.5)));
51 EXPECT_EQ("[[2.5,null]]", s(std::vector
<std::vector
<std::optional
<double>>>{
52 {2.5, std::nullopt
}}));
55 TEST(JSONTest
, StringOwnership
) {
57 Value Alias
= static_cast<const char *>(X
);
59 EXPECT_EQ(R
"("Hallo
")", s(Alias
));
61 std::string Y
= "Hello";
64 EXPECT_EQ(R
"("Hello
")", s(Copy
));
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));
74 TEST(JSONTest
, Escaping
) {
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.
84 std::string TestString
= R
"("\u0000\u0008\u000c\r\n\tS
\"\\)"
87 EXPECT_EQ(TestString, s(Test));
89 EXPECT_EQ(R"({"object keys are\nescaped":true})",
90 s(Object{{"object keys are
\nescaped
", true}}));
93 TEST(JSONTest, PrettyPrinting) {
94 const char Str[] = R"({
110 EXPECT_EQ(Str, sp(Object{
111 {"empty_object
", Object{}},
113 {"full_array
", {1, nullptr}},
118 {"property
", "value
"},
124 TEST(JSONTest, Array) {
127 A.emplace(++A.begin(), 0);
129 A.insert(++++A.begin(), 99);
131 EXPECT_EQ(A.size(), 6u);
132 EXPECT_EQ(R"([1,0,99,2,3,4])", s(std::move(A)));
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);
140 auto D = O.find("d
");
141 EXPECT_NE(D, O.end());
142 auto E = O.find("e
");
143 EXPECT_EQ(E, O.end());
147 EXPECT_EQ(O.size(), 2u);
148 EXPECT_EQ(R"({"a":1,"c":3})", s(std::move(O)));
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));
158 handleAllErrors(E.takeError(), [S](const llvm::ErrorInfoBase &E) {
159 FAIL() << "Failed to parse JSON
>>> " << S << " <<<: " << E.message();
164 Compare(R"(true)", true);
165 Compare(R"(false)", false);
166 Compare(R"(null
)", nullptr);
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());
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")", // UTF-16 surrogate pair escape.
178 /*U+10437*/ "\xf0\x90\x90\xb7");
179 Compare("\"\xE2\x82\xAC\xF0\x9D\x84\x9E\"", // UTF-8
180 /*U+20AC U+1D11E*/ "\xe2\x82\xac\xf0\x9d\x84\x9e");
183 R"("LoneLeading=\ud801, LoneTrailing=\udc01, LeadLeadTrail=\ud801\ud801\udc37")",
184 "LoneLeading
=" /*U+FFFD*/ "\xef\xbf\xbd, "
185 "LoneTrailing
=" /*U+FFFD*/ "\xef\xbf\xbd, "
186 "LeadLeadTrail
=" /*U+FFFD U+10437*/ "\xef\xbf\xbd\xf0\x90\x90\xb7");
188 Compare(R"({"":0,"":0})", Object{{"", 0}});
189 Compare(R"({"obj":{},"arr":[]})", Object{{"obj
", Object{}}, {"arr
", {}}});
190 Compare(R"({"\n":{"\u0000":[[[[]]]]}})",
191 Object{{"\n", Object{
192 {llvm::StringRef("\
0", 1), {{{{}}}}},
194 Compare("\r[\n\t] ", {});
197 TEST(JSONTest, ParseErrors) {
198 auto ExpectErr = [](llvm::StringRef Msg, llvm::StringRef S) {
199 if (auto E = parse(S)) {
200 // Compare both string forms and with operator==, in case we have bugs.
201 FAIL() << "Parsed JSON
>>> " << S << " <<< but wanted error
: " << Msg;
203 handleAllErrors(E.takeError(), [S, Msg](const llvm::ErrorInfoBase &E) {
204 EXPECT_THAT(E.message(), testing::HasSubstr(std::string(Msg))) << S;
208 ExpectErr("Unexpected EOF
", "");
209 ExpectErr("Unexpected EOF
", "[");
210 ExpectErr("Text after end of document
", "[][]");
211 ExpectErr("Invalid JSON
value (false?)", "fuzzy
");
212 ExpectErr("Expected
, or ]", "[2?]");
213 ExpectErr("Expected object key
", "{a
:2}");
214 ExpectErr("Expected
: after object key
", R"({"a",2})");
215 ExpectErr("Expected
, or } after object property
", R"({"a":2 "b":3})");
216 ExpectErr("Invalid JSON value
", R"([&%!])");
217 ExpectErr("Invalid JSON
value (number
?)", "1e1
.0
");
218 ExpectErr("Unterminated string
", R"("abc\"def)");
219 ExpectErr("Control character in string", "\"abc\ndef\"");
220 ExpectErr("Invalid escape sequence", R
"("\030")");
221 ExpectErr("Invalid \\u escape sequence", R
"("\usuck
")");
222 ExpectErr("[3:3, byte=19]", R
"({
226 ExpectErr("Invalid UTF-8 sequence", "\"\xC0\x80\""); // WTF-8 null
229 // Direct tests of isUTF8 and fixUTF8. Internal uses are also tested elsewhere.
230 TEST(JSONTest
, UTF8
) {
231 for (const char *Valid
: {
232 "this is ASCII text",
233 "thïs tëxt häs BMP chäräctërs",
236 EXPECT_TRUE(isUTF8(Valid
)) << Valid
;
237 EXPECT_EQ(fixUTF8(Valid
), Valid
);
239 for (auto Invalid
: std::vector
<std::pair
<const char *, const char *>>{
240 {"lone trailing \x81\x82 bytes", "lone trailing �� bytes"},
241 {"missing trailing \xD0 bytes", "missing trailing � bytes"},
242 {"truncated character \xD0", "truncated character �"},
243 {"not \xC1\x80 the \xE0\x9f\xBF shortest \xF0\x83\x83\x83 encoding",
244 "not �� the ��� shortest ���� encoding"},
245 {"too \xF9\x80\x80\x80\x80 long", "too ����� long"},
246 {"surrogate \xED\xA0\x80 invalid \xF4\x90\x80\x80",
247 "surrogate ��� invalid ����"}}) {
248 EXPECT_FALSE(isUTF8(Invalid
.first
)) << Invalid
.first
;
249 EXPECT_EQ(fixUTF8(Invalid
.first
), Invalid
.second
);
253 TEST(JSONTest
, Inspection
) {
254 llvm::Expected
<Value
> Doc
= parse(R
"(
260 "array
": [null, true, 3.14, "hello
", [1,2,3], {"time
": "arrow
"}],
261 "object
": {"fruit
": "banana
"}
266 Object
*O
= Doc
->getAsObject();
269 EXPECT_FALSE(O
->getNull("missing"));
270 EXPECT_FALSE(O
->getNull("boolean"));
271 EXPECT_TRUE(O
->getNull("null"));
273 EXPECT_EQ(O
->getNumber("number"), std::optional
<double>(2.78));
274 EXPECT_FALSE(O
->getInteger("number"));
275 EXPECT_EQ(O
->getString("string"), std::optional
<llvm::StringRef
>("json"));
276 ASSERT_FALSE(O
->getObject("missing"));
277 ASSERT_FALSE(O
->getObject("array"));
278 ASSERT_TRUE(O
->getObject("object"));
279 EXPECT_EQ(*O
->getObject("object"), (Object
{{"fruit", "banana"}}));
281 Array
*A
= O
->getArray("array");
283 EXPECT_EQ((*A
)[1].getAsBoolean(), std::optional
<bool>(true));
284 ASSERT_TRUE((*A
)[4].getAsArray());
285 EXPECT_EQ(*(*A
)[4].getAsArray(), (Array
{1, 2, 3}));
286 EXPECT_EQ((*(*A
)[4].getAsArray())[1].getAsInteger(),
287 std::optional
<int64_t>(2));
289 for (Value
&E
: *A
) {
291 ASSERT_TRUE(E
.getAsObject());
292 EXPECT_EQ(E
.getAsObject()->getString("time"),
293 std::optional
<llvm::StringRef
>("arrow"));
295 EXPECT_FALSE(E
.getAsObject());
299 // Verify special integer handling - we try to preserve exact int64 values.
300 TEST(JSONTest
, Integers
) {
305 std::optional
<int64_t> AsInt
;
306 std::optional
<double> AsNumber
;
309 "Non-integer. Stored as double, not convertible.",
317 "Integer, not exact double. Stored as int64, convertible.",
318 int64_t{0x4000000000000001},
319 "4611686018427387905",
320 int64_t{0x4000000000000001},
321 double{0x4000000000000000},
325 "Negative integer, not exact double. Stored as int64, convertible.",
326 int64_t{-0x4000000000000001},
327 "-4611686018427387905",
328 int64_t{-0x4000000000000001},
329 double{-0x4000000000000000},
333 // https://developercommunity.visualstudio.com/content/problem/1093399/incorrect-result-when-printing-6917529027641081856.html
334 #if !defined(_MSC_VER) || _MSC_VER < 1926
336 "Dynamically exact integer. Stored as double, convertible.",
337 double{0x6000000000000000},
338 "6.9175290276410819e+18",
339 int64_t{0x6000000000000000},
340 double{0x6000000000000000},
345 "Dynamically integer, >64 bits. Stored as double, not convertible.",
346 1.5 * double{0x8000000000000000},
347 "1.3835058055282164e+19",
349 1.5 * double{0x8000000000000000},
352 for (const auto &T
: TestCases
) {
353 EXPECT_EQ(T
.Str
, s(T
.Val
)) << T
.Desc
;
354 llvm::Expected
<Value
> Doc
= parse(T
.Str
);
355 EXPECT_TRUE(!!Doc
) << T
.Desc
;
356 EXPECT_EQ(Doc
->getAsInteger(), T
.AsInt
) << T
.Desc
;
357 EXPECT_EQ(Doc
->getAsNumber(), T
.AsNumber
) << T
.Desc
;
358 EXPECT_EQ(T
.Val
, *Doc
) << T
.Desc
;
359 EXPECT_EQ(T
.Str
, s(*Doc
)) << T
.Desc
;
363 // Verify uint64_t type.
364 TEST(JSONTest
, U64Integers
) {
365 Value Val
= uint64_t{3100100100};
366 uint64_t Var
= 3100100100;
369 Val
= uint64_t{std::numeric_limits
<uint64_t>::max()};
370 Var
= std::numeric_limits
<uint64_t>::max();
373 // Test the parse() part.
375 const char *Str
= "4611686018427387905";
376 llvm::Expected
<Value
> Doc
= parse(Str
);
379 EXPECT_EQ(Doc
->getAsInteger(), int64_t{4611686018427387905});
380 EXPECT_EQ(Doc
->getAsUINT64(), uint64_t{4611686018427387905});
384 const char *Str
= "-78278238238328222";
385 llvm::Expected
<Value
> Doc
= parse(Str
);
388 EXPECT_EQ(Doc
->getAsInteger(), int64_t{-78278238238328222});
389 EXPECT_EQ(Doc
->getAsUINT64(), std::nullopt
);
392 // Test with the largest 64 signed int.
394 const char *Str
= "9223372036854775807";
395 llvm::Expected
<Value
> Doc
= parse(Str
);
398 EXPECT_EQ(Doc
->getAsInteger(), int64_t{9223372036854775807});
399 EXPECT_EQ(Doc
->getAsUINT64(), uint64_t{9223372036854775807});
402 // Test with the largest 64 unsigned int.
404 const char *Str
= "18446744073709551615";
405 llvm::Expected
<Value
> Doc
= parse(Str
);
408 EXPECT_EQ(Doc
->getAsInteger(), std::nullopt
);
409 EXPECT_EQ(Doc
->getAsUINT64(), uint64_t{18446744073709551615u});
412 // Test with a number that is too big for 64 bits.
414 const char *Str
= "184467440737095516150";
415 llvm::Expected
<Value
> Doc
= parse(Str
);
418 EXPECT_EQ(Doc
->getAsInteger(), std::nullopt
);
419 EXPECT_EQ(Doc
->getAsUINT64(), std::nullopt
);
420 // The number was parsed as a double.
421 EXPECT_TRUE(!!Doc
->getAsNumber());
424 // Test with a negative number that is too small for 64 bits.
426 const char *Str
= "-18446744073709551615";
427 llvm::Expected
<Value
> Doc
= parse(Str
);
430 EXPECT_EQ(Doc
->getAsInteger(), std::nullopt
);
431 EXPECT_EQ(Doc
->getAsUINT64(), std::nullopt
);
432 // The number was parsed as a double.
433 EXPECT_TRUE(!!Doc
->getAsNumber());
435 // Test with a large number that is malformed.
437 const char *Str
= "184467440737095516150.12.12";
438 llvm::Expected
<Value
> Doc
= parse(Str
);
440 EXPECT_EQ("[1:27, byte=27]: Invalid JSON value (number?)",
441 llvm::toString(Doc
.takeError()));
445 template <typename T
> void checkCppIntegers() {
450 Val
= T
{std::numeric_limits
<T
>::max()};
451 Var
= std::numeric_limits
<T
>::max();
454 Val
= T
{std::numeric_limits
<T
>::min()};
455 Var
= std::numeric_limits
<T
>::min();
459 // Test that underlying C++ integer types behave as expected.
460 TEST(JSONTest
, CppIntegers
) {
461 checkCppIntegers
<char>();
462 checkCppIntegers
<signed char>();
463 checkCppIntegers
<unsigned char>();
465 checkCppIntegers
<short>();
466 checkCppIntegers
<unsigned short>();
468 checkCppIntegers
<int>();
469 checkCppIntegers
<unsigned int>();
471 checkCppIntegers
<long>();
472 checkCppIntegers
<unsigned long>();
474 checkCppIntegers
<long long>();
475 checkCppIntegers
<unsigned long long>();
478 // Sample struct with typical JSON-mapping rules.
479 struct CustomStruct
{
480 CustomStruct() : B(false) {}
481 CustomStruct(std::string S
, std::optional
<int> I
, bool B
)
482 : S(S
), I(I
), B(B
) {}
484 std::optional
<int> I
;
487 inline bool operator==(const CustomStruct
&L
, const CustomStruct
&R
) {
488 return L
.S
== R
.S
&& L
.I
== R
.I
&& L
.B
== R
.B
;
490 inline llvm::raw_ostream
&operator<<(llvm::raw_ostream
&OS
,
491 const CustomStruct
&S
) {
492 return OS
<< "(" << S
.S
<< ", " << (S
.I
? std::to_string(*S
.I
) : "None")
493 << ", " << S
.B
<< ")";
495 bool fromJSON(const Value
&E
, CustomStruct
&R
, Path P
) {
496 ObjectMapper
O(E
, P
);
497 return O
&& O
.map("str", R
.S
) && O
.map("int", R
.I
) &&
498 O
.mapOptional("bool", R
.B
);
501 static std::string
errorContext(const Value
&V
, const Path::Root
&R
) {
503 llvm::raw_string_ostream
OS(Context
);
504 R
.printErrorContext(V
, OS
);
508 TEST(JSONTest
, Deserialize
) {
509 std::map
<std::string
, std::vector
<CustomStruct
>> R
;
510 CustomStruct ExpectedStruct
= {"foo", 42, true};
511 std::map
<std::string
, std::vector
<CustomStruct
>> Expected
;
512 Value J
= Object
{{"foo", Array
{
517 {"unknown", "ignored"},
519 Object
{{"str", "bar"}},
522 CustomStruct("foo", 42, true),
523 CustomStruct("bar", std::nullopt
, false),
525 Path::Root
Root("CustomStruct");
526 ASSERT_TRUE(fromJSON(J
, R
, Root
));
527 EXPECT_EQ(R
, Expected
);
529 (*J
.getAsObject()->getArray("foo"))[0] = 123;
530 ASSERT_FALSE(fromJSON(J
, R
, Root
));
531 EXPECT_EQ("expected object at CustomStruct.foo[0]",
532 toString(Root
.getError()));
533 const char *ExpectedDump
= R
"({
535 /* error: expected object */
540 EXPECT_EQ(ExpectedDump
, errorContext(J
, Root
));
543 EXPECT_FALSE(fromJSON(nullptr, V
, Root
));
544 EXPECT_EQ("expected object when parsing CustomStruct",
545 toString(Root
.getError()));
547 EXPECT_FALSE(fromJSON(Object
{}, V
, Root
));
548 EXPECT_EQ("missing value at CustomStruct.str", toString(Root
.getError()));
550 EXPECT_FALSE(fromJSON(Object
{{"str", 1}}, V
, Root
));
551 EXPECT_EQ("expected string at CustomStruct.str", toString(Root
.getError()));
553 // std::optional<T> must parse as the correct type if present.
554 EXPECT_FALSE(fromJSON(Object
{{"str", "1"}, {"int", "string"}}, V
, Root
));
555 EXPECT_EQ("expected integer at CustomStruct.int", toString(Root
.getError()));
557 // mapOptional must parse as the correct type if present.
558 EXPECT_FALSE(fromJSON(Object
{{"str", "1"}, {"bool", "string"}}, V
, Root
));
559 EXPECT_EQ("expected boolean at CustomStruct.bool", toString(Root
.getError()));
562 TEST(JSONTest
, ParseDeserialize
) {
563 auto E
= parse
<std::vector
<CustomStruct
>>(R
"json(
564 [{"str
": "foo
", "int": 42}, {"int": 42}]
566 EXPECT_THAT_EXPECTED(E
, FailedWithMessage("missing value at (root)[1].str"));
568 E
= parse
<std::vector
<CustomStruct
>>(R
"json(
569 [{"str
": "foo
", "int": 42}, {"str
": "bar
"}
571 EXPECT_THAT_EXPECTED(
573 FailedWithMessage("[3:2, byte=50]: Expected , or ] after array element"));
575 E
= parse
<std::vector
<CustomStruct
>>(R
"json(
576 [{"str
": "foo
", "int": 42}]
578 EXPECT_THAT_EXPECTED(E
, Succeeded());
579 EXPECT_THAT(*E
, testing::SizeIs(1));
582 TEST(JSONTest
, Stream
) {
583 auto StreamStuff
= [](unsigned Indent
) {
585 llvm::raw_string_ostream
OS(S
);
586 OStream
J(OS
, Indent
);
587 J
.comment("top*/level");
589 J
.attributeArray("foo", [&] {
591 J
.comment("element");
596 J
.rawValue([](raw_ostream
&OS
) { OS
<< "'unverified\nraw value'"; });
598 J
.comment("attribute");
599 J
.attributeBegin("bar");
600 J
.comment("attribute value");
604 J
.attribute("baz", "xyz");
610 R
"(/*top* /level*/{"foo
":[null,/*element*/42.5,[43],'unverified
611 raw value'],/*attribute*/"bar
":/*attribute value*/{},"baz
":"xyz
"})";
612 EXPECT_EQ(Plain
, StreamStuff(0));
613 const char *Pretty
= R
"(/* top* /level */
626 "bar
": /* attribute value */ {},
629 EXPECT_EQ(Pretty
, StreamStuff(2));
632 TEST(JSONTest
, Path
) {
634 Path P
= R
, A
= P
.field("a"), B
= P
.field("b");
636 EXPECT_THAT_ERROR(R
.getError(), FailedWithMessage("oh no when parsing foo"));
637 A
.index(1).field("c").index(2).report("boom");
638 EXPECT_THAT_ERROR(R
.getError(), FailedWithMessage("boom at foo.a[1].c[2]"));
639 B
.field("d").field("e").report("bam");
640 EXPECT_THAT_ERROR(R
.getError(), FailedWithMessage("bam at foo.b.d.e"));
647 {"e", Array
{1, Object
{{"x", "y"}}}},
648 {"f", "a moderately long string: 48 characters in total"},
651 const char *Expected
= R
"({
655 "e
": /* error: bam */ [
659 "f
": "a moderately
long string
: 48 characte
..."
663 EXPECT_EQ(Expected
, errorContext(V
, R
));