Added strtonum for locale-independent number parsing

pull/379/head
Alex Astashyn 2016-12-03 19:05:09 -05:00
parent bc28942101
commit 6b78b5c2be
3 changed files with 483 additions and 269 deletions

View File

@ -9060,66 +9060,226 @@ basic_json_parser_66:
return result;
}
/*!
@brief parse floating point number
This function (and its overloads) serves to select the most approprate
standard floating point number parsing function based on the type
supplied via the first parameter. Set this to @a
static_cast<number_float_t*>(nullptr).
@param[in] type the @ref number_float_t in use
@param[in,out] endptr recieves a pointer to the first character after
the number
@return the floating point number
*/
long double str_to_float_t(long double* /* type */, char** endptr) const
{
return std::strtold(reinterpret_cast<typename string_t::const_pointer>(m_start), endptr);
}
/*!
@brief parse floating point number
@brief parse string into a built-in arithmetic type as if
the current locale is POSIX.
This function (and its overloads) serves to select the most approprate
standard floating point number parsing function based on the type
supplied via the first parameter. Set this to @a
static_cast<number_float_t*>(nullptr).
e.g. const auto x = static_cast<float>("123.4");
@param[in] type the @ref number_float_t in use
@param[in,out] endptr recieves a pointer to the first character after
the number
@return the floating point number
throw if can't parse the entire string as a number,
or if the destination type is integral and the value
is outside of the type's range.
*/
double str_to_float_t(double* /* type */, char** endptr) const
struct strtonum
{
return std::strtod(reinterpret_cast<typename string_t::const_pointer>(m_start), endptr);
public:
strtonum(const char* start, const char* end)
: m_start(start), m_end(end)
{}
template<typename T,
typename = typename std::enable_if<
std::is_arithmetic<T>::value>::type >
operator T() const
{
T val{0};
if(parse(val, std::is_integral<T>())) {
return val;
}
/*!
@brief parse floating point number
This function (and its overloads) serves to select the most approprate
standard floating point number parsing function based on the type
supplied via the first parameter. Set this to @a
static_cast<number_float_t*>(nullptr).
@param[in] type the @ref number_float_t in use
@param[in,out] endptr recieves a pointer to the first character after
the number
@return the floating point number
*/
float str_to_float_t(float* /* type */, char** endptr) const
{
return std::strtof(reinterpret_cast<typename string_t::const_pointer>(m_start), endptr);
throw std::invalid_argument(
std::string()
+ "Can't parse '"
+ std::string(m_start, m_end)
+ "' as type "
+ typeid(T).name());
}
/// return true iff token matches ^[+-]\d+$
bool is_integral() const
{
const char* p = m_start;
if(!p) {
return false;
}
if(*p == '-' or *p == '+') {
++p;
}
if(p == m_end) {
return false;
}
while(p < m_end and *p >= '0' and *p <= '9') {
++p;
}
return p == m_end;
}
private:
const char* const m_start = nullptr;
const char* const m_end = nullptr;
static void strtof(float& f,
const char* str,
char** endptr)
{
f = std::strtof(str, endptr);
}
static void strtof(double& f,
const char* str,
char** endptr)
{
f = std::strtod(str, endptr);
}
static void strtof(long double& f,
const char* str,
char** endptr)
{
f = std::strtold(str, endptr);
}
template<typename T>
bool parse(T& value, /*is_integral=*/std::false_type) const
{
const char* data = m_start;
const size_t len = static_cast<size_t>(m_end - m_start);
const char decimal_point_char =
std::use_facet< std::numpunct<char> >(
std::locale()).decimal_point();
// replace decimal separator with locale-specific
// version, if necessary; data will be repointed
// to either buf or tempstr containing the fixed
// string.
std::string tempstr;
std::array<char, 64> buf;
do {
if(decimal_point_char == '.') {
break; // don't need to convert
}
const size_t ds_pos = static_cast<size_t>(
std::find(m_start, m_end, '.') - m_start );
if(ds_pos == len) {
break; // no decimal separator
}
// copy the data into the local buffer or
// tempstr, if buffer is too small;
// replace decimal separator, and update
// data to point to the modified bytes
if(len + 1 < buf.size()) {
std::copy(m_start, m_end, buf.data());
buf[len] = 0;
buf[ds_pos] = decimal_point_char;
data = buf.data();
} else {
tempstr.assign(m_start, m_end);
tempstr[ds_pos] = decimal_point_char;
data = tempstr.c_str();
}
} while(0);
char* endptr = nullptr;
value = 0;
strtof(value, data, &endptr);
// note that reading past the end is OK, the data may be,
// for example, "123.", where the parsed token only contains "123",
// but strtod will read the dot as well.
const bool ok = endptr >= data + len
and len > 0;
if(ok and value == 0.0 and *data == '-') {
// some implementations forget to negate the zero
value = -0.0;
}
if(!ok) {
std::cerr << "data:" << data
<< " token:" << std::string(m_start, len)
<< " value:" << value
<< " len:" << len
<< " parsed_len:" << (endptr - data) << std::endl;
}
return ok;
}
template<typename T>
bool parse(T& value, /*is_integral=*/std::true_type) const
{
const char* beg = m_start;
const char* const end = m_end;
if(beg == end) {
return false;
}
const bool is_negative = (*beg == '-');
// json numbers can't start with '+' but
// this code is not json-specific
if(is_negative or *beg == '+') {
++beg; // skip the leading sign
}
bool valid = beg < end // must have some digits;
and ( T(-1) < 0 // type must be signed
or !is_negative); // if value is negative
value = 0;
while(beg < end and valid) {
const uint8_t c = static_cast<uint8_t>(*beg - '0');
const T upd_value = value * 10 + c;
valid &= value <= std::numeric_limits<T>::max() / 10
and value <= upd_value // did not overflow
and c < 10; // char was digit
value = upd_value;
++beg;
}
if(is_negative) {
value = 0 - value;
}
if(!valid) {
std::cerr << " token:" << std::string(m_start, m_end)
<< std::endl;
}
return valid;
}
};
/*!
@brief return number value for number tokens
@ -9127,112 +9287,58 @@ basic_json_parser_66:
number type (either integer, unsigned integer or floating point),
which is passed back to the caller via the result parameter.
This function parses the integer component up to the radix point or
exponent while collecting information about the 'floating point
representation', which it stores in the result parameter. If there is
no radix point or exponent, and the number can fit into a @ref
number_integer_t or @ref number_unsigned_t then it sets the result
parameter accordingly.
integral numbers that don't fit into the the range of the respective
type are parsed as number_float_t
If the number is a floating point number the number is then parsed
using @a std:strtod (or @a std:strtof or @a std::strtold).
floating-point values do not satisfy std::isfinite predicate
are converted to value_t::null
@param[out] result @ref basic_json object to receive the number, or
NAN if the conversion read past the current token. The latter case
needs to be treated by the caller function.
throws if the entire string [m_start .. m_cursor) cannot be
interpreted as a number
@param[out] result @ref basic_json object to receive the number.
*/
void get_number(basic_json& result) const
{
assert(m_start != nullptr);
assert(m_start < m_cursor);
const lexer::lexer_char_t* curptr = m_start;
strtonum num(reinterpret_cast<const char*>(m_start),
reinterpret_cast<const char*>(m_cursor));
// accumulate the integer conversion result (unsigned for now)
number_unsigned_t value = 0;
const bool is_negative = *m_start == '-';
// maximum absolute value of the relevant integer type
number_unsigned_t max;
// temporarily store the type to avoid unecessary bitfield access
value_t type;
// look for sign
if (*curptr == '-')
try {
if(not num.is_integral()) {
;
}
else if(is_negative)
{
type = value_t::number_integer;
max = static_cast<uint64_t>((std::numeric_limits<number_integer_t>::max)()) + 1;
curptr++;
result.m_type = value_t::number_integer;
result.m_value = static_cast<number_integer_t>(num);
return;
}
else
{
type = value_t::number_unsigned;
max = static_cast<uint64_t>((std::numeric_limits<number_unsigned_t>::max)());
result.m_type = value_t::number_unsigned;
result.m_value = static_cast<number_unsigned_t>(num);
return;
}
} catch (std::invalid_argument&) {
; // overflow - will parse as double
}
// count the significant figures
for (; curptr < m_cursor; curptr++)
{
// quickly skip tests if a digit
if (*curptr < '0' || *curptr > '9')
{
if (*curptr == '.')
{
// don't count '.' but change to float
type = value_t::number_float;
continue;
}
// assume exponent (if not then will fail parse): change to
// float, stop counting and record exponent details
type = value_t::number_float;
break;
}
// skip if definitely not an integer
if (type != value_t::number_float)
{
// multiply last value by ten and add the new digit
auto temp = value * 10 + *curptr - '0';
// test for overflow
if (temp < value || temp > max)
{
// overflow
type = value_t::number_float;
}
else
{
// no overflow - save it
value = temp;
}
}
}
// save the value (if not a float)
if (type == value_t::number_unsigned)
{
result.m_value.number_unsigned = value;
}
else if (type == value_t::number_integer)
{
result.m_value.number_integer = -static_cast<number_integer_t>(value);
}
else
{
// parse with strtod
result.m_value.number_float = str_to_float_t(static_cast<number_float_t*>(nullptr), NULL);
result.m_type = value_t::number_float;
result.m_value = static_cast<number_float_t>(num);
// replace infinity and NAN by null
if (not std::isfinite(result.m_value.number_float))
{
type = value_t::null;
result.m_type = value_t::null;
result.m_value = basic_json::json_value();
}
}
// save the type
result.m_type = type;
}
private:
/// optional input stream
std::istream* m_stream = nullptr;

View File

@ -8209,66 +8209,226 @@ class basic_json
return result;
}
/*!
@brief parse floating point number
This function (and its overloads) serves to select the most approprate
standard floating point number parsing function based on the type
supplied via the first parameter. Set this to @a
static_cast<number_float_t*>(nullptr).
@param[in] type the @ref number_float_t in use
@param[in,out] endptr recieves a pointer to the first character after
the number
@return the floating point number
*/
long double str_to_float_t(long double* /* type */, char** endptr) const
{
return std::strtold(reinterpret_cast<typename string_t::const_pointer>(m_start), endptr);
}
/*!
@brief parse floating point number
@brief parse string into a built-in arithmetic type as if
the current locale is POSIX.
This function (and its overloads) serves to select the most approprate
standard floating point number parsing function based on the type
supplied via the first parameter. Set this to @a
static_cast<number_float_t*>(nullptr).
e.g. const auto x = static_cast<float>("123.4");
@param[in] type the @ref number_float_t in use
@param[in,out] endptr recieves a pointer to the first character after
the number
@return the floating point number
throw if can't parse the entire string as a number,
or if the destination type is integral and the value
is outside of the type's range.
*/
double str_to_float_t(double* /* type */, char** endptr) const
struct strtonum
{
return std::strtod(reinterpret_cast<typename string_t::const_pointer>(m_start), endptr);
public:
strtonum(const char* start, const char* end)
: m_start(start), m_end(end)
{}
template<typename T,
typename = typename std::enable_if<
std::is_arithmetic<T>::value>::type >
operator T() const
{
T val{0};
if(parse(val, std::is_integral<T>())) {
return val;
}
/*!
@brief parse floating point number
This function (and its overloads) serves to select the most approprate
standard floating point number parsing function based on the type
supplied via the first parameter. Set this to @a
static_cast<number_float_t*>(nullptr).
@param[in] type the @ref number_float_t in use
@param[in,out] endptr recieves a pointer to the first character after
the number
@return the floating point number
*/
float str_to_float_t(float* /* type */, char** endptr) const
{
return std::strtof(reinterpret_cast<typename string_t::const_pointer>(m_start), endptr);
throw std::invalid_argument(
std::string()
+ "Can't parse '"
+ std::string(m_start, m_end)
+ "' as type "
+ typeid(T).name());
}
/// return true iff token matches ^[+-]\d+$
bool is_integral() const
{
const char* p = m_start;
if(!p) {
return false;
}
if(*p == '-' or *p == '+') {
++p;
}
if(p == m_end) {
return false;
}
while(p < m_end and *p >= '0' and *p <= '9') {
++p;
}
return p == m_end;
}
private:
const char* const m_start = nullptr;
const char* const m_end = nullptr;
static void strtof(float& f,
const char* str,
char** endptr)
{
f = std::strtof(str, endptr);
}
static void strtof(double& f,
const char* str,
char** endptr)
{
f = std::strtod(str, endptr);
}
static void strtof(long double& f,
const char* str,
char** endptr)
{
f = std::strtold(str, endptr);
}
template<typename T>
bool parse(T& value, /*is_integral=*/std::false_type) const
{
const char* data = m_start;
const size_t len = static_cast<size_t>(m_end - m_start);
const char decimal_point_char =
std::use_facet< std::numpunct<char> >(
std::locale()).decimal_point();
// replace decimal separator with locale-specific
// version, if necessary; data will be repointed
// to either buf or tempstr containing the fixed
// string.
std::string tempstr;
std::array<char, 64> buf;
do {
if(decimal_point_char == '.') {
break; // don't need to convert
}
const size_t ds_pos = static_cast<size_t>(
std::find(m_start, m_end, '.') - m_start );
if(ds_pos == len) {
break; // no decimal separator
}
// copy the data into the local buffer or
// tempstr, if buffer is too small;
// replace decimal separator, and update
// data to point to the modified bytes
if(len + 1 < buf.size()) {
std::copy(m_start, m_end, buf.data());
buf[len] = 0;
buf[ds_pos] = decimal_point_char;
data = buf.data();
} else {
tempstr.assign(m_start, m_end);
tempstr[ds_pos] = decimal_point_char;
data = tempstr.c_str();
}
} while(0);
char* endptr = nullptr;
value = 0;
strtof(value, data, &endptr);
// note that reading past the end is OK, the data may be,
// for example, "123.", where the parsed token only contains "123",
// but strtod will read the dot as well.
const bool ok = endptr >= data + len
and len > 0;
if(ok and value == 0.0 and *data == '-') {
// some implementations forget to negate the zero
value = -0.0;
}
if(!ok) {
std::cerr << "data:" << data
<< " token:" << std::string(m_start, len)
<< " value:" << value
<< " len:" << len
<< " parsed_len:" << (endptr - data) << std::endl;
}
return ok;
}
template<typename T>
bool parse(T& value, /*is_integral=*/std::true_type) const
{
const char* beg = m_start;
const char* const end = m_end;
if(beg == end) {
return false;
}
const bool is_negative = (*beg == '-');
// json numbers can't start with '+' but
// this code is not json-specific
if(is_negative or *beg == '+') {
++beg; // skip the leading sign
}
bool valid = beg < end // must have some digits;
and ( T(-1) < 0 // type must be signed
or !is_negative); // if value is negative
value = 0;
while(beg < end and valid) {
const uint8_t c = static_cast<uint8_t>(*beg - '0');
const T upd_value = value * 10 + c;
valid &= value <= std::numeric_limits<T>::max() / 10
and value <= upd_value // did not overflow
and c < 10; // char was digit
value = upd_value;
++beg;
}
if(is_negative) {
value = 0 - value;
}
if(!valid) {
std::cerr << " token:" << std::string(m_start, m_end)
<< std::endl;
}
return valid;
}
};
/*!
@brief return number value for number tokens
@ -8276,112 +8436,58 @@ class basic_json
number type (either integer, unsigned integer or floating point),
which is passed back to the caller via the result parameter.
This function parses the integer component up to the radix point or
exponent while collecting information about the 'floating point
representation', which it stores in the result parameter. If there is
no radix point or exponent, and the number can fit into a @ref
number_integer_t or @ref number_unsigned_t then it sets the result
parameter accordingly.
integral numbers that don't fit into the the range of the respective
type are parsed as number_float_t
If the number is a floating point number the number is then parsed
using @a std:strtod (or @a std:strtof or @a std::strtold).
floating-point values do not satisfy std::isfinite predicate
are converted to value_t::null
@param[out] result @ref basic_json object to receive the number, or
NAN if the conversion read past the current token. The latter case
needs to be treated by the caller function.
throws if the entire string [m_start .. m_cursor) cannot be
interpreted as a number
@param[out] result @ref basic_json object to receive the number.
*/
void get_number(basic_json& result) const
{
assert(m_start != nullptr);
assert(m_start < m_cursor);
const lexer::lexer_char_t* curptr = m_start;
strtonum num(reinterpret_cast<const char*>(m_start),
reinterpret_cast<const char*>(m_cursor));
// accumulate the integer conversion result (unsigned for now)
number_unsigned_t value = 0;
const bool is_negative = *m_start == '-';
// maximum absolute value of the relevant integer type
number_unsigned_t max;
// temporarily store the type to avoid unecessary bitfield access
value_t type;
// look for sign
if (*curptr == '-')
try {
if(not num.is_integral()) {
;
}
else if(is_negative)
{
type = value_t::number_integer;
max = static_cast<uint64_t>((std::numeric_limits<number_integer_t>::max)()) + 1;
curptr++;
result.m_type = value_t::number_integer;
result.m_value = static_cast<number_integer_t>(num);
return;
}
else
{
type = value_t::number_unsigned;
max = static_cast<uint64_t>((std::numeric_limits<number_unsigned_t>::max)());
result.m_type = value_t::number_unsigned;
result.m_value = static_cast<number_unsigned_t>(num);
return;
}
} catch (std::invalid_argument&) {
; // overflow - will parse as double
}
// count the significant figures
for (; curptr < m_cursor; curptr++)
{
// quickly skip tests if a digit
if (*curptr < '0' || *curptr > '9')
{
if (*curptr == '.')
{
// don't count '.' but change to float
type = value_t::number_float;
continue;
}
// assume exponent (if not then will fail parse): change to
// float, stop counting and record exponent details
type = value_t::number_float;
break;
}
// skip if definitely not an integer
if (type != value_t::number_float)
{
// multiply last value by ten and add the new digit
auto temp = value * 10 + *curptr - '0';
// test for overflow
if (temp < value || temp > max)
{
// overflow
type = value_t::number_float;
}
else
{
// no overflow - save it
value = temp;
}
}
}
// save the value (if not a float)
if (type == value_t::number_unsigned)
{
result.m_value.number_unsigned = value;
}
else if (type == value_t::number_integer)
{
result.m_value.number_integer = -static_cast<number_integer_t>(value);
}
else
{
// parse with strtod
result.m_value.number_float = str_to_float_t(static_cast<number_float_t*>(nullptr), NULL);
result.m_type = value_t::number_float;
result.m_value = static_cast<number_float_t>(num);
// replace infinity and NAN by null
if (not std::isfinite(result.m_value.number_float))
{
type = value_t::null;
result.m_type = value_t::null;
result.m_value = basic_json::json_value();
}
}
// save the type
result.m_type = type;
}
private:
/// optional input stream
std::istream* m_stream = nullptr;

View File

@ -375,7 +375,7 @@ TEST_CASE("regression tests")
};
// change locale to mess with decimal points
std::locale::global(std::locale(std::locale(), new CommaDecimalSeparator));
auto orig_locale = std::locale::global(std::locale(std::locale(), new CommaDecimalSeparator));
CHECK(j1a.dump() == "23.42");
CHECK(j1b.dump() == "23.42");
@ -399,6 +399,8 @@ TEST_CASE("regression tests")
CHECK(j3c.dump() == "10000");
//CHECK(j3b.dump() == "1E04"); // roundtrip error
//CHECK(j3c.dump() == "1e04"); // roundtrip error
std::locale::global(orig_locale);
}
SECTION("issue #233 - Can't use basic_json::iterator as a base iterator for std::move_iterator")