mirror of
https://github.com/nlohmann/json.git
synced 2024-11-27 16:49:02 +08:00
🚨 fix warnings #2615
This commit is contained in:
parent
b83fe5dbf2
commit
ffdeb77468
@ -490,51 +490,49 @@ inline int find_largest_pow10(const std::uint32_t n, std::uint32_t& pow10)
|
||||
return 10;
|
||||
}
|
||||
// LCOV_EXCL_STOP
|
||||
else if (n >= 100000000)
|
||||
if (n >= 100000000)
|
||||
{
|
||||
pow10 = 100000000;
|
||||
return 9;
|
||||
}
|
||||
else if (n >= 10000000)
|
||||
if (n >= 10000000)
|
||||
{
|
||||
pow10 = 10000000;
|
||||
return 8;
|
||||
}
|
||||
else if (n >= 1000000)
|
||||
if (n >= 1000000)
|
||||
{
|
||||
pow10 = 1000000;
|
||||
return 7;
|
||||
}
|
||||
else if (n >= 100000)
|
||||
if (n >= 100000)
|
||||
{
|
||||
pow10 = 100000;
|
||||
return 6;
|
||||
}
|
||||
else if (n >= 10000)
|
||||
if (n >= 10000)
|
||||
{
|
||||
pow10 = 10000;
|
||||
return 5;
|
||||
}
|
||||
else if (n >= 1000)
|
||||
if (n >= 1000)
|
||||
{
|
||||
pow10 = 1000;
|
||||
return 4;
|
||||
}
|
||||
else if (n >= 100)
|
||||
if (n >= 100)
|
||||
{
|
||||
pow10 = 100;
|
||||
return 3;
|
||||
}
|
||||
else if (n >= 10)
|
||||
if (n >= 10)
|
||||
{
|
||||
pow10 = 10;
|
||||
return 2;
|
||||
}
|
||||
else
|
||||
{
|
||||
pow10 = 1;
|
||||
return 1;
|
||||
}
|
||||
|
||||
pow10 = 1;
|
||||
return 1;
|
||||
}
|
||||
|
||||
inline void grisu2_round(char* buf, int len, std::uint64_t dist, std::uint64_t delta,
|
||||
|
@ -135,10 +135,8 @@ class iterator_input_adapter
|
||||
std::advance(current, 1);
|
||||
return result;
|
||||
}
|
||||
else
|
||||
{
|
||||
return std::char_traits<char_type>::eof();
|
||||
}
|
||||
|
||||
return std::char_traits<char_type>::eof();
|
||||
}
|
||||
|
||||
private:
|
||||
|
@ -393,62 +393,62 @@ class parser
|
||||
parse_error::create(101, m_lexer.get_position(),
|
||||
exception_message(token_type::end_array, "array")));
|
||||
}
|
||||
else // object
|
||||
|
||||
// states.back() is false -> object
|
||||
|
||||
// comma -> next value
|
||||
if (get_token() == token_type::value_separator)
|
||||
{
|
||||
// comma -> next value
|
||||
if (get_token() == token_type::value_separator)
|
||||
// parse key
|
||||
if (JSON_HEDLEY_UNLIKELY(get_token() != token_type::value_string))
|
||||
{
|
||||
// parse key
|
||||
if (JSON_HEDLEY_UNLIKELY(get_token() != token_type::value_string))
|
||||
{
|
||||
return sax->parse_error(m_lexer.get_position(),
|
||||
m_lexer.get_token_string(),
|
||||
parse_error::create(101, m_lexer.get_position(),
|
||||
exception_message(token_type::value_string, "object key")));
|
||||
}
|
||||
|
||||
if (JSON_HEDLEY_UNLIKELY(!sax->key(m_lexer.get_string())))
|
||||
{
|
||||
return false;
|
||||
}
|
||||
|
||||
// parse separator (:)
|
||||
if (JSON_HEDLEY_UNLIKELY(get_token() != token_type::name_separator))
|
||||
{
|
||||
return sax->parse_error(m_lexer.get_position(),
|
||||
m_lexer.get_token_string(),
|
||||
parse_error::create(101, m_lexer.get_position(),
|
||||
exception_message(token_type::name_separator, "object separator")));
|
||||
}
|
||||
|
||||
// parse values
|
||||
get_token();
|
||||
continue;
|
||||
return sax->parse_error(m_lexer.get_position(),
|
||||
m_lexer.get_token_string(),
|
||||
parse_error::create(101, m_lexer.get_position(),
|
||||
exception_message(token_type::value_string, "object key")));
|
||||
}
|
||||
|
||||
// closing }
|
||||
if (JSON_HEDLEY_LIKELY(last_token == token_type::end_object))
|
||||
if (JSON_HEDLEY_UNLIKELY(!sax->key(m_lexer.get_string())))
|
||||
{
|
||||
if (JSON_HEDLEY_UNLIKELY(!sax->end_object()))
|
||||
{
|
||||
return false;
|
||||
}
|
||||
|
||||
// We are done with this object. Before we can parse a
|
||||
// new value, we need to evaluate the new state first.
|
||||
// By setting skip_to_state_evaluation to false, we
|
||||
// are effectively jumping to the beginning of this if.
|
||||
JSON_ASSERT(!states.empty());
|
||||
states.pop_back();
|
||||
skip_to_state_evaluation = true;
|
||||
continue;
|
||||
return false;
|
||||
}
|
||||
|
||||
return sax->parse_error(m_lexer.get_position(),
|
||||
m_lexer.get_token_string(),
|
||||
parse_error::create(101, m_lexer.get_position(),
|
||||
exception_message(token_type::end_object, "object")));
|
||||
// parse separator (:)
|
||||
if (JSON_HEDLEY_UNLIKELY(get_token() != token_type::name_separator))
|
||||
{
|
||||
return sax->parse_error(m_lexer.get_position(),
|
||||
m_lexer.get_token_string(),
|
||||
parse_error::create(101, m_lexer.get_position(),
|
||||
exception_message(token_type::name_separator, "object separator")));
|
||||
}
|
||||
|
||||
// parse values
|
||||
get_token();
|
||||
continue;
|
||||
}
|
||||
|
||||
// closing }
|
||||
if (JSON_HEDLEY_LIKELY(last_token == token_type::end_object))
|
||||
{
|
||||
if (JSON_HEDLEY_UNLIKELY(!sax->end_object()))
|
||||
{
|
||||
return false;
|
||||
}
|
||||
|
||||
// We are done with this object. Before we can parse a
|
||||
// new value, we need to evaluate the new state first.
|
||||
// By setting skip_to_state_evaluation to false, we
|
||||
// are effectively jumping to the beginning of this if.
|
||||
JSON_ASSERT(!states.empty());
|
||||
states.pop_back();
|
||||
skip_to_state_evaluation = true;
|
||||
continue;
|
||||
}
|
||||
|
||||
return sax->parse_error(m_lexer.get_position(),
|
||||
m_lexer.get_token_string(),
|
||||
parse_error::create(101, m_lexer.get_position(),
|
||||
exception_message(token_type::end_object, "object")));
|
||||
}
|
||||
}
|
||||
|
||||
|
@ -39,7 +39,7 @@ template<typename IteratorType> class iteration_proxy_value
|
||||
/// a string representation of the array index
|
||||
mutable string_type array_index_str = "0";
|
||||
/// an empty string (to return a reference for primitive values)
|
||||
const string_type empty_str = "";
|
||||
const string_type empty_str;
|
||||
|
||||
public:
|
||||
explicit iteration_proxy_value(IteratorType it) noexcept : anchor(it) {}
|
||||
|
@ -3944,7 +3944,7 @@ template<typename IteratorType> class iteration_proxy_value
|
||||
/// a string representation of the array index
|
||||
mutable string_type array_index_str = "0";
|
||||
/// an empty string (to return a reference for primitive values)
|
||||
const string_type empty_str = "";
|
||||
const string_type empty_str;
|
||||
|
||||
public:
|
||||
explicit iteration_proxy_value(IteratorType it) noexcept : anchor(it) {}
|
||||
@ -4947,10 +4947,8 @@ class iterator_input_adapter
|
||||
std::advance(current, 1);
|
||||
return result;
|
||||
}
|
||||
else
|
||||
{
|
||||
return std::char_traits<char_type>::eof();
|
||||
}
|
||||
|
||||
return std::char_traits<char_type>::eof();
|
||||
}
|
||||
|
||||
private:
|
||||
@ -10618,62 +10616,62 @@ class parser
|
||||
parse_error::create(101, m_lexer.get_position(),
|
||||
exception_message(token_type::end_array, "array")));
|
||||
}
|
||||
else // object
|
||||
|
||||
// states.back() is false -> object
|
||||
|
||||
// comma -> next value
|
||||
if (get_token() == token_type::value_separator)
|
||||
{
|
||||
// comma -> next value
|
||||
if (get_token() == token_type::value_separator)
|
||||
// parse key
|
||||
if (JSON_HEDLEY_UNLIKELY(get_token() != token_type::value_string))
|
||||
{
|
||||
// parse key
|
||||
if (JSON_HEDLEY_UNLIKELY(get_token() != token_type::value_string))
|
||||
{
|
||||
return sax->parse_error(m_lexer.get_position(),
|
||||
m_lexer.get_token_string(),
|
||||
parse_error::create(101, m_lexer.get_position(),
|
||||
exception_message(token_type::value_string, "object key")));
|
||||
}
|
||||
|
||||
if (JSON_HEDLEY_UNLIKELY(!sax->key(m_lexer.get_string())))
|
||||
{
|
||||
return false;
|
||||
}
|
||||
|
||||
// parse separator (:)
|
||||
if (JSON_HEDLEY_UNLIKELY(get_token() != token_type::name_separator))
|
||||
{
|
||||
return sax->parse_error(m_lexer.get_position(),
|
||||
m_lexer.get_token_string(),
|
||||
parse_error::create(101, m_lexer.get_position(),
|
||||
exception_message(token_type::name_separator, "object separator")));
|
||||
}
|
||||
|
||||
// parse values
|
||||
get_token();
|
||||
continue;
|
||||
return sax->parse_error(m_lexer.get_position(),
|
||||
m_lexer.get_token_string(),
|
||||
parse_error::create(101, m_lexer.get_position(),
|
||||
exception_message(token_type::value_string, "object key")));
|
||||
}
|
||||
|
||||
// closing }
|
||||
if (JSON_HEDLEY_LIKELY(last_token == token_type::end_object))
|
||||
if (JSON_HEDLEY_UNLIKELY(!sax->key(m_lexer.get_string())))
|
||||
{
|
||||
if (JSON_HEDLEY_UNLIKELY(!sax->end_object()))
|
||||
{
|
||||
return false;
|
||||
}
|
||||
|
||||
// We are done with this object. Before we can parse a
|
||||
// new value, we need to evaluate the new state first.
|
||||
// By setting skip_to_state_evaluation to false, we
|
||||
// are effectively jumping to the beginning of this if.
|
||||
JSON_ASSERT(!states.empty());
|
||||
states.pop_back();
|
||||
skip_to_state_evaluation = true;
|
||||
continue;
|
||||
return false;
|
||||
}
|
||||
|
||||
return sax->parse_error(m_lexer.get_position(),
|
||||
m_lexer.get_token_string(),
|
||||
parse_error::create(101, m_lexer.get_position(),
|
||||
exception_message(token_type::end_object, "object")));
|
||||
// parse separator (:)
|
||||
if (JSON_HEDLEY_UNLIKELY(get_token() != token_type::name_separator))
|
||||
{
|
||||
return sax->parse_error(m_lexer.get_position(),
|
||||
m_lexer.get_token_string(),
|
||||
parse_error::create(101, m_lexer.get_position(),
|
||||
exception_message(token_type::name_separator, "object separator")));
|
||||
}
|
||||
|
||||
// parse values
|
||||
get_token();
|
||||
continue;
|
||||
}
|
||||
|
||||
// closing }
|
||||
if (JSON_HEDLEY_LIKELY(last_token == token_type::end_object))
|
||||
{
|
||||
if (JSON_HEDLEY_UNLIKELY(!sax->end_object()))
|
||||
{
|
||||
return false;
|
||||
}
|
||||
|
||||
// We are done with this object. Before we can parse a
|
||||
// new value, we need to evaluate the new state first.
|
||||
// By setting skip_to_state_evaluation to false, we
|
||||
// are effectively jumping to the beginning of this if.
|
||||
JSON_ASSERT(!states.empty());
|
||||
states.pop_back();
|
||||
skip_to_state_evaluation = true;
|
||||
continue;
|
||||
}
|
||||
|
||||
return sax->parse_error(m_lexer.get_position(),
|
||||
m_lexer.get_token_string(),
|
||||
parse_error::create(101, m_lexer.get_position(),
|
||||
exception_message(token_type::end_object, "object")));
|
||||
}
|
||||
}
|
||||
|
||||
@ -14948,51 +14946,49 @@ inline int find_largest_pow10(const std::uint32_t n, std::uint32_t& pow10)
|
||||
return 10;
|
||||
}
|
||||
// LCOV_EXCL_STOP
|
||||
else if (n >= 100000000)
|
||||
if (n >= 100000000)
|
||||
{
|
||||
pow10 = 100000000;
|
||||
return 9;
|
||||
}
|
||||
else if (n >= 10000000)
|
||||
if (n >= 10000000)
|
||||
{
|
||||
pow10 = 10000000;
|
||||
return 8;
|
||||
}
|
||||
else if (n >= 1000000)
|
||||
if (n >= 1000000)
|
||||
{
|
||||
pow10 = 1000000;
|
||||
return 7;
|
||||
}
|
||||
else if (n >= 100000)
|
||||
if (n >= 100000)
|
||||
{
|
||||
pow10 = 100000;
|
||||
return 6;
|
||||
}
|
||||
else if (n >= 10000)
|
||||
if (n >= 10000)
|
||||
{
|
||||
pow10 = 10000;
|
||||
return 5;
|
||||
}
|
||||
else if (n >= 1000)
|
||||
if (n >= 1000)
|
||||
{
|
||||
pow10 = 1000;
|
||||
return 4;
|
||||
}
|
||||
else if (n >= 100)
|
||||
if (n >= 100)
|
||||
{
|
||||
pow10 = 100;
|
||||
return 3;
|
||||
}
|
||||
else if (n >= 10)
|
||||
if (n >= 10)
|
||||
{
|
||||
pow10 = 10;
|
||||
return 2;
|
||||
}
|
||||
else
|
||||
{
|
||||
pow10 = 1;
|
||||
return 1;
|
||||
}
|
||||
|
||||
pow10 = 1;
|
||||
return 1;
|
||||
}
|
||||
|
||||
inline void grisu2_round(char* buf, int len, std::uint64_t dist, std::uint64_t delta,
|
||||
|
Loading…
Reference in New Issue
Block a user