mirror of
https://github.com/fmtlib/fmt.git
synced 2024-12-25 15:21:54 +00:00
stop -> handler and swap args
This commit is contained in:
parent
f90d33ca12
commit
b1f7cca89e
@ -468,9 +468,9 @@ inline round_direction get_round_direction(uint64_t divisor, uint64_t remainder,
|
||||
}
|
||||
|
||||
// Generates output using Grisu2 digit-gen algorithm.
|
||||
template <typename Stop>
|
||||
template <typename Handler>
|
||||
int grisu2_gen_digits(char* buf, fp value, uint64_t error, int& exp,
|
||||
Stop stop) {
|
||||
Handler handler) {
|
||||
fp one(1ull << -value.e, value.e);
|
||||
// The integral part of scaled value (p1 in Grisu) = value / one. It cannot be
|
||||
// zero because it contains a product of two 64-bit numbers with MSB set (due
|
||||
@ -482,8 +482,8 @@ int grisu2_gen_digits(char* buf, fp value, uint64_t error, int& exp,
|
||||
uint64_t fractional = value.f & (one.f - 1);
|
||||
exp = count_digits(integral); // kappa in Grisu.
|
||||
int size = 0;
|
||||
if (stop.init(buf, size, value.f, data::POWERS_OF_10_64[exp] << -one.e,
|
||||
error, exp)) {
|
||||
if (handler.on_start(buf, size, data::POWERS_OF_10_64[exp] << -one.e, value.f,
|
||||
error, exp)) {
|
||||
return size;
|
||||
}
|
||||
// Generate digits for the integral part. This can produce up to 10 digits.
|
||||
@ -539,8 +539,8 @@ int grisu2_gen_digits(char* buf, fp value, uint64_t error, int& exp,
|
||||
--exp;
|
||||
uint64_t remainder =
|
||||
(static_cast<uint64_t>(integral) << -one.e) + fractional;
|
||||
if (stop(buf, size, remainder, data::POWERS_OF_10_64[exp] << -one.e,
|
||||
error, exp, true)) {
|
||||
if (handler(buf, size, data::POWERS_OF_10_64[exp] << -one.e, remainder,
|
||||
error, exp, true)) {
|
||||
return size;
|
||||
}
|
||||
} while (exp > 0);
|
||||
@ -552,12 +552,12 @@ int grisu2_gen_digits(char* buf, fp value, uint64_t error, int& exp,
|
||||
buf[size++] = static_cast<char>('0' + digit);
|
||||
fractional &= one.f - 1;
|
||||
--exp;
|
||||
if (stop(buf, size, fractional, one.f, error, exp, false)) return size;
|
||||
if (handler(buf, size, one.f, fractional, error, exp, false)) return size;
|
||||
}
|
||||
}
|
||||
|
||||
// Stopping condition for the fixed precision.
|
||||
struct fixed_stop {
|
||||
// The fixed precision digit handler.
|
||||
struct fixed_handler {
|
||||
int precision;
|
||||
int exp10;
|
||||
bool fixed;
|
||||
@ -566,8 +566,8 @@ struct fixed_stop {
|
||||
return full_exp <= 0 && -full_exp >= precision;
|
||||
}
|
||||
|
||||
bool init(char* buf, int& size, uint64_t remainder, uint64_t divisor,
|
||||
uint64_t error, int& exp) {
|
||||
bool on_start(char* buf, int& size, uint64_t divisor, uint64_t remainder,
|
||||
uint64_t error, int& exp) {
|
||||
if (!fixed) return false;
|
||||
int full_exp = exp + exp10;
|
||||
if (full_exp >= 0) precision += full_exp;
|
||||
@ -586,7 +586,7 @@ struct fixed_stop {
|
||||
}
|
||||
|
||||
// TODO: test
|
||||
bool operator()(char* buf, int& size, uint64_t remainder, uint64_t divisor,
|
||||
bool operator()(char* buf, int& size, uint64_t divisor, uint64_t remainder,
|
||||
uint64_t error, int& exp, bool integral) const {
|
||||
assert(remainder < divisor);
|
||||
if (size != precision && !enough_precision(exp + exp10)) return false;
|
||||
@ -623,13 +623,15 @@ struct fixed_stop {
|
||||
}
|
||||
};
|
||||
|
||||
// Stopping condition for the shortest representation.
|
||||
struct shortest_stop {
|
||||
// The shortest representation digit handler.
|
||||
struct shortest_handler {
|
||||
fp diff; // wp_w in Grisu.
|
||||
|
||||
bool init(char*, int&, uint64_t, uint64_t, uint64_t, int&) { return false; }
|
||||
bool on_start(char*, int&, uint64_t, uint64_t, uint64_t, int&) {
|
||||
return false;
|
||||
}
|
||||
|
||||
bool operator()(char* buf, int& size, uint64_t remainder, uint64_t divisor,
|
||||
bool operator()(char* buf, int& size, uint64_t divisor, uint64_t remainder,
|
||||
uint64_t error, int& exp, bool integral) {
|
||||
if (remainder > error) return false;
|
||||
uint64_t d = integral ? diff.f : diff.f * data::POWERS_OF_10_64[-exp];
|
||||
@ -668,8 +670,9 @@ grisu2_format(Double value, buffer& buf, int precision, bool fixed, int& exp) {
|
||||
auto cached_pow = get_cached_power(
|
||||
min_exp - (fp_value.e + fp::significand_size), cached_exp10);
|
||||
fp_value = fp_value * cached_pow;
|
||||
int size = grisu2_gen_digits(buf.data(), fp_value, 1, exp,
|
||||
fixed_stop{precision, -cached_exp10, fixed});
|
||||
int size =
|
||||
grisu2_gen_digits(buf.data(), fp_value, 1, exp,
|
||||
fixed_handler{precision, -cached_exp10, fixed});
|
||||
if (size < 0) return false;
|
||||
buf.resize(to_unsigned(size));
|
||||
} else {
|
||||
@ -687,7 +690,7 @@ grisu2_format(Double value, buffer& buf, int precision, bool fixed, int& exp) {
|
||||
lower = lower * cached_pow; // \tilde{M}^- in Grisu.
|
||||
++lower.f; // \tilde{M}^- + 1 ulp -> M^-_{\uparrow}.
|
||||
int size = grisu2_gen_digits(buf.data(), upper, upper.f - lower.f, exp,
|
||||
shortest_stop{upper - fp_value});
|
||||
shortest_handler{upper - fp_value});
|
||||
if (size < 0) return false;
|
||||
buf.resize(to_unsigned(size));
|
||||
}
|
||||
|
Loading…
Reference in New Issue
Block a user