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(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>>>{
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")", u8"\U00010437
"); // UTF16 surrogate pair escape.
178 Compare("\"\xE2\x82\xAC\xF0\x9D\x84\x9E\"", u8"\u20ac\U0001d11e
"); // UTF8
180 R"("LoneLeading=\ud801, LoneTrailing=\udc01, LeadingLeadingTrailing=\ud801\ud801\udc37")",
181 u8"LoneLeading
=\ufffd, LoneTrailing
=\ufffd, "
182 u8"LeadingLeadingTrailing
=\ufffd\U00010437
"); // Invalid unicode.
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), {{{{}}}}},
190 Compare("\r[\n\t] ", {});
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;
199 handleAllErrors(E.takeError(), [S, Msg](const llvm::ErrorInfoBase &E) {
200 EXPECT_THAT(E.message(), testing::HasSubstr(std::string(Msg))) << S;
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
"({
222 ExpectErr("Invalid UTF-8 sequence", "\"\xC0\x80\""); // WTF-8 null
225 // Direct tests of isUTF8 and fixUTF8. Internal uses are also tested elsewhere.
226 TEST(JSONTest
, UTF8
) {
227 for (const char *Valid
: {
228 "this is ASCII text",
229 "thïs tëxt häs BMP chäräctërs",
232 EXPECT_TRUE(isUTF8(Valid
)) << Valid
;
233 EXPECT_EQ(fixUTF8(Valid
), Valid
);
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
);
249 TEST(JSONTest
, Inspection
) {
250 llvm::Expected
<Value
> Doc
= parse(R
"(
256 "array
": [null, true, 3.14, "hello
", [1,2,3], {"time
": "arrow
"}],
257 "object
": {"fruit
": "banana
"}
262 Object
*O
= Doc
->getAsObject();
265 EXPECT_FALSE(O
->getNull("missing"));
266 EXPECT_FALSE(O
->getNull("boolean"));
267 EXPECT_TRUE(O
->getNull("null"));
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"}}));
277 Array
*A
= O
->getArray("array");
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));
285 for (Value
&E
: *A
) {
287 ASSERT_TRUE(E
.getAsObject());
288 EXPECT_EQ(E
.getAsObject()->getString("time"),
289 llvm::Optional
<llvm::StringRef
>("arrow"));
291 EXPECT_FALSE(E
.getAsObject());
295 // Verify special integer handling - we try to preserve exact int64 values.
296 TEST(JSONTest
, Integers
) {
301 llvm::Optional
<int64_t> AsInt
;
302 llvm::Optional
<double> AsNumber
;
305 "Non-integer. Stored as double, not convertible.",
313 "Integer, not exact double. Stored as int64, convertible.",
314 int64_t{0x4000000000000001},
315 "4611686018427387905",
316 int64_t{0x4000000000000001},
317 double{0x4000000000000000},
321 "Negative integer, not exact double. Stored as int64, convertible.",
322 int64_t{-0x4000000000000001},
323 "-4611686018427387905",
324 int64_t{-0x4000000000000001},
325 double{-0x4000000000000000},
329 // https://developercommunity.visualstudio.com/content/problem/1093399/incorrect-result-when-printing-6917529027641081856.html
330 #if !defined(_MSC_VER) || _MSC_VER < 1926
332 "Dynamically exact integer. Stored as double, convertible.",
333 double{0x6000000000000000},
334 "6.9175290276410819e+18",
335 int64_t{0x6000000000000000},
336 double{0x6000000000000000},
341 "Dynamically integer, >64 bits. Stored as double, not convertible.",
342 1.5 * double{0x8000000000000000},
343 "1.3835058055282164e+19",
345 1.5 * double{0x8000000000000000},
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
;
359 // Verify uint64_t type.
360 TEST(JSONTest
, U64Integers
) {
361 Value Val
= uint64_t{3100100100};
362 uint64_t Var
= 3100100100;
365 Val
= uint64_t{std::numeric_limits
<uint64_t>::max()};
366 Var
= std::numeric_limits
<uint64_t>::max();
369 // Test the parse() part.
371 const char *Str
= "4611686018427387905";
372 llvm::Expected
<Value
> Doc
= parse(Str
);
375 EXPECT_EQ(Doc
->getAsInteger(), int64_t{4611686018427387905});
376 EXPECT_EQ(Doc
->getAsUINT64(), uint64_t{4611686018427387905});
380 const char *Str
= "-78278238238328222";
381 llvm::Expected
<Value
> Doc
= parse(Str
);
384 EXPECT_EQ(Doc
->getAsInteger(), int64_t{-78278238238328222});
385 EXPECT_EQ(Doc
->getAsUINT64(), llvm::None
);
388 // Test with the largest 64 signed int.
390 const char *Str
= "9223372036854775807";
391 llvm::Expected
<Value
> Doc
= parse(Str
);
394 EXPECT_EQ(Doc
->getAsInteger(), int64_t{9223372036854775807});
395 EXPECT_EQ(Doc
->getAsUINT64(), uint64_t{9223372036854775807});
398 // Test with the largest 64 unsigned int.
400 const char *Str
= "18446744073709551615";
401 llvm::Expected
<Value
> Doc
= parse(Str
);
404 EXPECT_EQ(Doc
->getAsInteger(), None
);
405 EXPECT_EQ(Doc
->getAsUINT64(), uint64_t{18446744073709551615u});
408 // Test with a number that is too big for 64 bits.
410 const char *Str
= "184467440737095516150";
411 llvm::Expected
<Value
> Doc
= parse(Str
);
414 EXPECT_EQ(Doc
->getAsInteger(), None
);
415 EXPECT_EQ(Doc
->getAsUINT64(), None
);
416 // The number was parsed as a double.
417 EXPECT_TRUE(!!Doc
->getAsNumber());
420 // Test with a negative number that is too small for 64 bits.
422 const char *Str
= "-18446744073709551615";
423 llvm::Expected
<Value
> Doc
= parse(Str
);
426 EXPECT_EQ(Doc
->getAsInteger(), None
);
427 EXPECT_EQ(Doc
->getAsUINT64(), None
);
428 // The number was parsed as a double.
429 EXPECT_TRUE(!!Doc
->getAsNumber());
431 // Test with a large number that is malformed.
433 const char *Str
= "184467440737095516150.12.12";
434 llvm::Expected
<Value
> Doc
= parse(Str
);
436 EXPECT_EQ("[1:27, byte=27]: Invalid JSON value (number?)",
437 llvm::toString(Doc
.takeError()));
441 // Sample struct with typical JSON-mapping rules.
442 struct CustomStruct
{
443 CustomStruct() : B(false) {}
444 CustomStruct(std::string S
, llvm::Optional
<int> I
, bool B
)
445 : S(S
), I(I
), B(B
) {}
447 llvm::Optional
<int> I
;
450 inline bool operator==(const CustomStruct
&L
, const CustomStruct
&R
) {
451 return L
.S
== R
.S
&& L
.I
== R
.I
&& L
.B
== R
.B
;
453 inline llvm::raw_ostream
&operator<<(llvm::raw_ostream
&OS
,
454 const CustomStruct
&S
) {
455 return OS
<< "(" << S
.S
<< ", " << (S
.I
? std::to_string(*S
.I
) : "None")
456 << ", " << S
.B
<< ")";
458 bool fromJSON(const Value
&E
, CustomStruct
&R
, Path P
) {
459 ObjectMapper
O(E
, P
);
460 return O
&& O
.map("str", R
.S
) && O
.map("int", R
.I
) &&
461 O
.mapOptional("bool", R
.B
);
464 static std::string
errorContext(const Value
&V
, const Path::Root
&R
) {
466 llvm::raw_string_ostream
OS(Context
);
467 R
.printErrorContext(V
, OS
);
471 TEST(JSONTest
, Deserialize
) {
472 std::map
<std::string
, std::vector
<CustomStruct
>> R
;
473 CustomStruct ExpectedStruct
= {"foo", 42, true};
474 std::map
<std::string
, std::vector
<CustomStruct
>> Expected
;
475 Value J
= Object
{{"foo", Array
{
480 {"unknown", "ignored"},
482 Object
{{"str", "bar"}},
485 CustomStruct("foo", 42, true),
486 CustomStruct("bar", llvm::None
, false),
488 Path::Root
Root("CustomStruct");
489 ASSERT_TRUE(fromJSON(J
, R
, Root
));
490 EXPECT_EQ(R
, Expected
);
492 (*J
.getAsObject()->getArray("foo"))[0] = 123;
493 ASSERT_FALSE(fromJSON(J
, R
, Root
));
494 EXPECT_EQ("expected object at CustomStruct.foo[0]",
495 toString(Root
.getError()));
496 const char *ExpectedDump
= R
"({
498 /* error: expected object */
503 EXPECT_EQ(ExpectedDump
, errorContext(J
, Root
));
506 EXPECT_FALSE(fromJSON(nullptr, V
, Root
));
507 EXPECT_EQ("expected object when parsing CustomStruct",
508 toString(Root
.getError()));
510 EXPECT_FALSE(fromJSON(Object
{}, V
, Root
));
511 EXPECT_EQ("missing value at CustomStruct.str", toString(Root
.getError()));
513 EXPECT_FALSE(fromJSON(Object
{{"str", 1}}, V
, Root
));
514 EXPECT_EQ("expected string at CustomStruct.str", toString(Root
.getError()));
516 // Optional<T> must parse as the correct type if present.
517 EXPECT_FALSE(fromJSON(Object
{{"str", "1"}, {"int", "string"}}, V
, Root
));
518 EXPECT_EQ("expected integer at CustomStruct.int", toString(Root
.getError()));
520 // mapOptional must parse as the correct type if present.
521 EXPECT_FALSE(fromJSON(Object
{{"str", "1"}, {"bool", "string"}}, V
, Root
));
522 EXPECT_EQ("expected boolean at CustomStruct.bool", toString(Root
.getError()));
525 TEST(JSONTest
, ParseDeserialize
) {
526 auto E
= parse
<std::vector
<CustomStruct
>>(R
"json(
527 [{"str
": "foo
", "int": 42}, {"int": 42}]
529 EXPECT_THAT_EXPECTED(E
, FailedWithMessage("missing value at (root)[1].str"));
531 E
= parse
<std::vector
<CustomStruct
>>(R
"json(
532 [{"str
": "foo
", "int": 42}, {"str
": "bar
"}
534 EXPECT_THAT_EXPECTED(
536 FailedWithMessage("[3:2, byte=50]: Expected , or ] after array element"));
538 E
= parse
<std::vector
<CustomStruct
>>(R
"json(
539 [{"str
": "foo
", "int": 42}]
541 EXPECT_THAT_EXPECTED(E
, Succeeded());
542 EXPECT_THAT(*E
, testing::SizeIs(1));
545 TEST(JSONTest
, Stream
) {
546 auto StreamStuff
= [](unsigned Indent
) {
548 llvm::raw_string_ostream
OS(S
);
549 OStream
J(OS
, Indent
);
550 J
.comment("top*/level");
552 J
.attributeArray("foo", [&] {
554 J
.comment("element");
559 J
.rawValue([](raw_ostream
&OS
) { OS
<< "'unverified\nraw value'"; });
561 J
.comment("attribute");
562 J
.attributeBegin("bar");
563 J
.comment("attribute value");
567 J
.attribute("baz", "xyz");
573 R
"(/*top* /level*/{"foo
":[null,/*element*/42.5,[43],'unverified
574 raw value'],/*attribute*/"bar
":/*attribute value*/{},"baz
":"xyz
"})";
575 EXPECT_EQ(Plain
, StreamStuff(0));
576 const char *Pretty
= R
"(/* top* /level */
589 "bar
": /* attribute value */ {},
592 EXPECT_EQ(Pretty
, StreamStuff(2));
595 TEST(JSONTest
, Path
) {
597 Path P
= R
, A
= P
.field("a"), B
= P
.field("b");
599 EXPECT_THAT_ERROR(R
.getError(), FailedWithMessage("oh no when parsing foo"));
600 A
.index(1).field("c").index(2).report("boom");
601 EXPECT_THAT_ERROR(R
.getError(), FailedWithMessage("boom at foo.a[1].c[2]"));
602 B
.field("d").field("e").report("bam");
603 EXPECT_THAT_ERROR(R
.getError(), FailedWithMessage("bam at foo.b.d.e"));
610 {"e", Array
{1, Object
{{"x", "y"}}}},
611 {"f", "a moderately long string: 48 characters in total"},
614 const char *Expected
= R
"({
618 "e
": /* error: bam */ [
622 "f
": "a moderately
long string
: 48 characte
..."
626 EXPECT_EQ(Expected
, errorContext(V
, R
));