Moved buffer management into the lexer class and implemented YYFILL so that streams are read incrementally.
This commit is contained in:
parent
5526de1385
commit
754c38e871
1 changed files with 50 additions and 16 deletions
|
@ -3047,11 +3047,20 @@ class basic_json
|
|||
|
||||
/// constructor with a given buffer
|
||||
inline lexer(const string_t& s) noexcept
|
||||
: m_content(reinterpret_cast<const lexer_char_t*>(s.c_str()))
|
||||
: m_buffer(s), m_stream(nullptr)
|
||||
{
|
||||
m_content = reinterpret_cast<const lexer_char_t*>(s.c_str());
|
||||
m_start = m_cursor = m_content;
|
||||
m_limit = m_content + s.size();
|
||||
}
|
||||
inline lexer(std::istream* s) noexcept
|
||||
: m_stream(s)
|
||||
{
|
||||
getline(*m_stream, m_buffer);
|
||||
m_content = reinterpret_cast<const lexer_char_t*>(m_buffer.c_str());
|
||||
m_start = m_cursor = m_content;
|
||||
m_limit = m_content + m_buffer.size();
|
||||
}
|
||||
|
||||
/// default constructor
|
||||
inline lexer() = default;
|
||||
|
@ -3182,6 +3191,8 @@ class basic_json
|
|||
// remember the begin of the token
|
||||
m_start = m_cursor;
|
||||
|
||||
#define YYFILL(n) { size_t offset_marker = m_marker - m_start; yyfill(n); m_marker = m_start + offset_marker; }
|
||||
|
||||
/*!re2c
|
||||
re2c:define:YYCTYPE = lexer_char_t;
|
||||
re2c:define:YYCURSOR = m_cursor;
|
||||
|
@ -3190,7 +3201,6 @@ class basic_json
|
|||
re2c:indent:string = " ";
|
||||
re2c:indent:top = 1;
|
||||
re2c:labelprefix = "basic_json_parser_";
|
||||
re2c:yyfill:enable = 0;
|
||||
|
||||
// whitespace
|
||||
ws = [ \t\n\r]+;
|
||||
|
@ -3240,6 +3250,26 @@ class basic_json
|
|||
// anything else is an error
|
||||
. { return token_type::parse_error; }
|
||||
*/
|
||||
|
||||
}
|
||||
|
||||
void yyfill(int n) noexcept
|
||||
{
|
||||
if (not m_stream or not *m_stream) return;
|
||||
|
||||
ssize_t offset_start = m_start - m_content;
|
||||
ssize_t offset_cursor = m_cursor - m_start;
|
||||
ssize_t offset_limit = m_limit - m_start;
|
||||
|
||||
m_buffer.erase(0, offset_start);
|
||||
std::string line;
|
||||
std::getline(*m_stream, line);
|
||||
m_buffer += line;
|
||||
|
||||
m_content = reinterpret_cast<const lexer_char_t*>(m_buffer.c_str()); //reinterpret_cast<lexer_char_t*>(endptr)
|
||||
m_start = m_content + offset_start;
|
||||
m_cursor = m_start + offset_cursor;
|
||||
m_limit = m_start + offset_limit;
|
||||
}
|
||||
|
||||
/// return string representation of last read token
|
||||
|
@ -3404,14 +3434,20 @@ class basic_json
|
|||
}
|
||||
|
||||
private:
|
||||
/// optional input stream
|
||||
std::istream* m_stream;
|
||||
/// the buffer
|
||||
string_t m_buffer;
|
||||
/// the buffer pointer
|
||||
const lexer_char_t* m_content = nullptr;
|
||||
/// pointer to he beginning of the current symbol
|
||||
/// pointer to the beginning of the current symbol
|
||||
const lexer_char_t* m_start = nullptr;
|
||||
/// pointer to the current symbol
|
||||
const lexer_char_t* m_cursor = nullptr;
|
||||
/// pointer to the end of the buffer
|
||||
const lexer_char_t* m_limit = nullptr;
|
||||
/// YYSTATE
|
||||
int m_state;
|
||||
};
|
||||
|
||||
/*!
|
||||
|
@ -3421,24 +3457,24 @@ class basic_json
|
|||
{
|
||||
public:
|
||||
/// constructor for strings
|
||||
inline parser(const string_t& s) : m_buffer(s), m_lexer(m_buffer)
|
||||
inline parser(const string_t& s) : m_lexer(s)
|
||||
{
|
||||
// read first token
|
||||
get_token();
|
||||
}
|
||||
|
||||
/// a parser reading from an input stream
|
||||
inline parser(std::istream& _is)
|
||||
inline parser(std::istream& _is) : m_lexer(&_is)
|
||||
{
|
||||
while (_is)
|
||||
{
|
||||
string_t input_line;
|
||||
std::getline(_is, input_line);
|
||||
m_buffer += input_line;
|
||||
}
|
||||
// while (_is)
|
||||
// {
|
||||
// string_t input_line;
|
||||
// std::getline(_is, input_line);
|
||||
// m_buffer += input_line;
|
||||
// }
|
||||
|
||||
// initializer lexer
|
||||
m_lexer = lexer(m_buffer);
|
||||
// m_lexer = std::move(lexer(_is));
|
||||
|
||||
// read first token
|
||||
get_token();
|
||||
|
@ -3625,8 +3661,6 @@ class basic_json
|
|||
}
|
||||
|
||||
private:
|
||||
/// the buffer
|
||||
string_t m_buffer;
|
||||
/// the type of the last read token
|
||||
typename lexer::token_type last_token = lexer::token_type::uninitialized;
|
||||
/// the lexer
|
||||
|
|
Loading…
Reference in a new issue