list cons(element e)
{
list l = malloc(sizeof(item));
l->value.value = BuildValue(e);
l->value.type = e.type;
l->next = NULL;
l->root = l;
return l;
}
element MinValue(list l)
{
element x = { BuildValue(head(l)), head(l).type };
for (size_t i = 0; tail(l) != NULL; ++i)
{
if (cmp(x, (head(tail(l)))) > 0)
AssignElement(&x, head(tail(l)));
l = tail(l);
}
return x;
}
element MaxValue(list l)
{
element x = { BuildValue(head(l)), head(l).type };
while (empty(tail(l)) == false)
{
if (cmp(x, (head(tail(l)))) < 0)
AssignElement(&x, head(tail(l)));
l = tail(l);
}
return x;
}
TEST(SliceContainerTest, FromSlice) {
std::shared_ptr<Builder> builder = BuildValue("\"this is a string of 20 bytes\"");
Slice slice = builder->slice();
SliceContainer sb(slice);
ASSERT_EQ(sb.byteSize(), slice.byteSize());
ASSERT_EQ(0, memcmp(sb.data(), slice.begin(), slice.byteSize()));
ASSERT_EQ(29UL, sb.byteSize());
ASSERT_TRUE(sb.slice().isString());
ASSERT_EQ("this is a string of 20 bytes", sb.slice().copyString());
ASSERT_NE(sb.data(), slice.begin());
}
TEST(SliceContainerTest, FromUInt8Longer) {
std::shared_ptr<Builder> builder = BuildValue("[\"the eagle has landed\",\"test\",\"qux\"]");
uint8_t const* begin = builder->start();
Slice slice = builder->slice();
SliceContainer sb(begin, slice.byteSize());
ASSERT_EQ(sb.byteSize(), slice.byteSize());
ASSERT_EQ(0, memcmp(sb.data(), builder->start(), slice.byteSize()));
ASSERT_TRUE(sb.slice().isArray());
ASSERT_NE(sb.data(), begin);
}
TEST(SliceContainerTest, FromChar) {
std::shared_ptr<Builder> builder = BuildValue("null");
uint8_t const* begin = builder->start();
Slice slice = builder->slice();
SliceContainer sb(reinterpret_cast<char const*>(begin), slice.byteSize());
ASSERT_EQ(sb.byteSize(), slice.byteSize());
ASSERT_EQ(0, memcmp(sb.data(), builder->start(), slice.byteSize()));
ASSERT_EQ(1UL, sb.byteSize());
ASSERT_TRUE(sb.slice().isNull());
ASSERT_NE(sb.data(), begin);
}
Value* JSONReader::JsonToValue(const std::string& json, bool check_root,
bool allow_trailing_comma)
{
// 输入必须是UTF-8编码.
if(!IsStringUTF8(json.c_str()))
{
error_code_ = JSON_UNSUPPORTED_ENCODING;
return NULL;
}
// 从UTF8到wstring的转换会移除空字节(好事).
std::wstring json_wide(UTF8ToWide(json));
start_pos_ = json_wide.c_str();
// 当输入的JSON字符串开头有UTF-8的Byte-Order-Mark(0xEF, 0xBB, 0xBF),
// UTF8ToWide()函数会把它转换成BOM(U+FEFF). 为防止JSONReader::BuildValue()
// 函数把它当成非法字符而返回NULL, 如果存在Unicode的BOM则先跳过.
if(!json_wide.empty() && start_pos_[0]==0xFEFF)
{
++start_pos_;
}
json_pos_ = start_pos_;
allow_trailing_comma_ = allow_trailing_comma;
stack_depth_ = 0;
error_code_ = JSON_NO_ERROR;
scoped_ptr<Value> root(BuildValue(check_root));
if(root.get())
{
if(ParseToken().type == Token::END_OF_INPUT)
{
return root.release();
}
else
{
SetErrorCode(JSON_UNEXPECTED_DATA_AFTER_ROOT, json_pos_);
}
}
// "语法错误".
if(error_code_ == 0)
{
SetErrorCode(JSON_SYNTAX_ERROR, json_pos_);
}
return NULL;
}
TEST(SliceContainerTest, MoveConstruct) {
std::shared_ptr<Builder> builder = BuildValue("\"this is a string of 20 bytes\"");
Slice slice = builder->slice();
SliceContainer sb(slice.begin(), slice.byteSize());
SliceContainer sb2(std::move(sb));
ASSERT_TRUE(sb.slice().isNone()); // must be empty now
ASSERT_EQ(1UL, sb.slice().byteSize());
ASSERT_EQ(29UL, sb2.byteSize());
ASSERT_TRUE(sb2.slice().isString());
ASSERT_EQ("this is a string of 20 bytes", sb2.slice().copyString());
ASSERT_NE(sb.slice().begin(), sb2.slice().begin());
}
TEST(SliceContainerTest, FromSliceLonger) {
std::string s("[-1");
for (size_t i = 0; i < 2000; ++i) {
s.push_back(',');
s.append(std::to_string(i));
}
s.push_back(']');
std::shared_ptr<Builder> builder = BuildValue(s);
Slice slice = builder->slice();
SliceContainer sb(slice);
ASSERT_EQ(sb.byteSize(), slice.byteSize());
ASSERT_EQ(0, memcmp(sb.data(), builder->start(), slice.byteSize()));
ASSERT_TRUE(sb.slice().isArray());
ASSERT_NE(sb.data(), slice.begin());
}
TEST(SliceContainerTest, SizeLengthByteSize) {
std::shared_ptr<Builder> builder = BuildValue("\"this is a string of 20 bytes\"");
Slice slice = builder->slice();
SliceContainer sb(slice.begin(), slice.byteSize());
ASSERT_TRUE(sb.slice().isString());
ASSERT_EQ(29UL, sb.size());
ASSERT_EQ(29UL, sb.length());
ASSERT_EQ(29UL, sb.byteSize());
ASSERT_EQ(sb.data(), sb.begin());
Slice empty;
sb = SliceContainer(empty);
ASSERT_EQ(1UL, sb.size());
ASSERT_EQ(1UL, sb.length());
ASSERT_EQ(1UL, sb.byteSize());
ASSERT_EQ(sb.data(), sb.begin());
}
Value* JSONReader::BuildValue(bool is_root)
{
++stack_depth_;
if(stack_depth_ > kStackLimit)
{
SetErrorCode(JSON_TOO_MUCH_NESTING, json_pos_);
return NULL;
}
Token token = ParseToken();
// 根token必须是数组或者对象.
if(is_root && token.type!=Token::OBJECT_BEGIN &&
token.type!=Token::ARRAY_BEGIN)
{
SetErrorCode(JSON_BAD_ROOT_ELEMENT_TYPE, json_pos_);
return NULL;
}
scoped_ptr<Value> node;
switch(token.type)
{
case Token::END_OF_INPUT:
case Token::INVALID_TOKEN:
return NULL;
case Token::NULL_TOKEN:
node.reset(Value::CreateNullValue());
break;
case Token::BOOL_TRUE:
node.reset(Value::CreateBooleanValue(true));
break;
case Token::BOOL_FALSE:
node.reset(Value::CreateBooleanValue(false));
break;
case Token::NUMBER:
node.reset(DecodeNumber(token));
if(!node.get())
{
return NULL;
}
break;
case Token::STRING:
node.reset(DecodeString(token));
if(!node.get())
{
return NULL;
}
break;
case Token::ARRAY_BEGIN:
{
json_pos_ += token.length;
token = ParseToken();
node.reset(new ListValue());
while(token.type != Token::ARRAY_END)
{
Value* array_node = BuildValue(false);
if(!array_node)
{
return NULL;
}
static_cast<ListValue*>(node.get())->Append(array_node);
// list数据后面应该是逗号, 否则list结束.
token = ParseToken();
if(token.type == Token::LIST_SEPARATOR)
{
json_pos_ += token.length;
token = ParseToken();
// 按照JSON RFC结尾的逗号是不合法的, 但是有些人希望能宽松一些,
// 所以做一些相应处理.
if(token.type == Token::ARRAY_END)
{
if(!allow_trailing_comma_)
{
SetErrorCode(JSON_TRAILING_COMMA, json_pos_);
return NULL;
}
// 结尾有逗号, 停止继续解析Array.
break;
}
}
else if(token.type != Token::ARRAY_END)
{
// 非预期数据, 直接返回.
return NULL;
}
}
if(token.type != Token::ARRAY_END)
{
return NULL;
}
break;
}
case Token::OBJECT_BEGIN:
{
json_pos_ += token.length;
token = ParseToken();
node.reset(new DictionaryValue());
while(token.type != Token::OBJECT_END)
{
if(token.type != Token::STRING)
{
SetErrorCode(JSON_UNQUOTED_DICTIONARY_KEY, json_pos_);
return NULL;
}
scoped_ptr<Value> dict_key_value(DecodeString(token));
if(!dict_key_value.get())
{
return NULL;
}
// key转换成wstring.
std::string dict_key;
bool success = dict_key_value->GetAsString(&dict_key);
DCHECK(success);
json_pos_ += token.length;
token = ParseToken();
if(token.type != Token::OBJECT_PAIR_SEPARATOR)
{
return NULL;
}
json_pos_ += token.length;
token = ParseToken();
Value* dict_value = BuildValue(false);
if(!dict_value)
{
return NULL;
}
static_cast<DictionaryValue*>(node.get())->SetWithoutPathExpansion(
dict_key, dict_value);
// key/value后面应该是逗号, 否则对象结束.
token = ParseToken();
if(token.type == Token::LIST_SEPARATOR)
{
json_pos_ += token.length;
token = ParseToken();
// 按照JSON RFC结尾的逗号是不合法的, 但是有些人希望能宽松一些,
// 所以做一些相应处理.
if(token.type == Token::OBJECT_END)
{
if(!allow_trailing_comma_)
{
SetErrorCode(JSON_TRAILING_COMMA, json_pos_);
return NULL;
}
// 结尾有逗号, 停止继续解析Array.
break;
}
}
else if(token.type != Token::OBJECT_END)
{
// 非预期数据, 直接返回.
return NULL;
}
}
if(token.type != Token::OBJECT_END)
{
return NULL;
}
break;
}
default:
// 非数据token.
return NULL;
}
json_pos_ += token.length;
--stack_depth_;
return node.release();
}
