mirror of
https://github.com/LadybirdBrowser/ladybird.git
synced 2025-04-20 11:36:10 +00:00
LibJS: Use exact mathematical values for Intl.DurationFormat
We can't use doubles due to precision loss for extremely large values.
This commit is contained in:
Timothy Flynn
Tim Flynn
parent
912e38c5fb
commit
3c64e4595a
Notes:
github-actions[bot]
2024-12-04 13:02:26 +00:00
Author: https://github.com/trflynn89 Commit: https://github.com/LadybirdBrowser/ladybird/commit/3c64e4595a4 Pull-request: https://github.com/LadybirdBrowser/ladybird/pull/2731 Reviewed-by: https://github.com/shannonbooth
3 changed files with 57 additions and 30 deletions
|
|
@ -12,6 +12,7 @@
|
|||
#include <LibJS/Runtime/Intl/DurationFormat.h>
|
||||
#include <LibJS/Runtime/Intl/ListFormat.h>
|
||||
#include <LibJS/Runtime/Intl/ListFormatConstructor.h>
|
||||
#include <LibJS/Runtime/Intl/MathematicalValue.h>
|
||||
#include <LibJS/Runtime/Intl/NumberFormatConstructor.h>
|
||||
#include <LibJS/Runtime/Intl/PluralRules.h>
|
||||
#include <LibJS/Runtime/Intl/PluralRulesConstructor.h>
|
||||
|
|
@ -398,10 +399,10 @@ ThrowCompletionOr<DurationUnitOptions> get_duration_unit_options(VM& vm, Duratio
|
|||
}
|
||||
|
||||
// 1.1.7 ComputeFractionalDigits ( durationFormat, duration ), https://tc39.es/proposal-intl-duration-format/#sec-computefractionaldigits
|
||||
double compute_fractional_digits(DurationFormat const& duration_format, DurationRecord const& duration)
|
||||
Crypto::BigFraction compute_fractional_digits(DurationFormat const& duration_format, DurationRecord const& duration)
|
||||
{
|
||||
// 1. Let result be 0.
|
||||
double result = 0;
|
||||
Crypto::BigFraction result;
|
||||
|
||||
// 2. Let exponent be 3.
|
||||
double exponent = 3;
|
||||
|
|
@ -417,13 +418,14 @@ double compute_fractional_digits(DurationFormat const& duration_format, Duration
|
|||
VERIFY(first_is_one_of(duration_instances_component.unit, DurationFormat::Unit::Milliseconds, DurationFormat::Unit::Microseconds, DurationFormat::Unit::Nanoseconds));
|
||||
|
||||
// ii. Let value be the value of duration's field whose name is the Value Field value of the current row.
|
||||
auto value = duration.*duration_instances_component.value_slot;
|
||||
|
||||
// iii. Set value to value / 10**exponent.
|
||||
value = value / pow(10, exponent);
|
||||
Crypto::BigFraction value {
|
||||
Crypto::SignedBigInteger { duration.*duration_instances_component.value_slot },
|
||||
Crypto::UnsignedBigInteger { pow(10, exponent) }
|
||||
};
|
||||
|
||||
// iv. Set result to result + value.
|
||||
result += value;
|
||||
result = result + value;
|
||||
|
||||
// v. Set exponent to exponent + 3.
|
||||
exponent += 3;
|
||||
|
|
@ -457,7 +459,7 @@ bool next_unit_fractional(DurationFormat const& duration_format, DurationFormat:
|
|||
}
|
||||
|
||||
// 1.1.9 FormatNumericHours ( durationFormat, hoursValue, signDisplayed ), https://tc39.es/proposal-intl-duration-format/#sec-formatnumerichours
|
||||
Vector<DurationFormatPart> format_numeric_hours(VM& vm, DurationFormat const& duration_format, double hours_value, bool sign_displayed)
|
||||
Vector<DurationFormatPart> format_numeric_hours(VM& vm, DurationFormat const& duration_format, MathematicalValue const& hours_value, bool sign_displayed)
|
||||
{
|
||||
auto& realm = *vm.current_realm();
|
||||
|
||||
|
|
@ -498,7 +500,7 @@ Vector<DurationFormatPart> format_numeric_hours(VM& vm, DurationFormat const& du
|
|||
auto number_format = construct_number_format(vm, duration_format, number_format_options);
|
||||
|
||||
// 11. Let hoursParts be PartitionNumberPattern(nf, hoursValue).
|
||||
auto hours_parts = partition_number_pattern(number_format, MathematicalValue { hours_value });
|
||||
auto hours_parts = partition_number_pattern(number_format, hours_value);
|
||||
|
||||
// 12. For each Record { [[Type]], [[Value]] } part of hoursParts, do
|
||||
result.ensure_capacity(hours_parts.size());
|
||||
|
|
@ -513,7 +515,7 @@ Vector<DurationFormatPart> format_numeric_hours(VM& vm, DurationFormat const& du
|
|||
}
|
||||
|
||||
// 1.1.10 FormatNumericMinutes ( durationFormat, minutesValue, hoursDisplayed, signDisplayed ), https://tc39.es/proposal-intl-duration-format/#sec-formatnumericminutes
|
||||
Vector<DurationFormatPart> format_numeric_minutes(VM& vm, DurationFormat const& duration_format, double minutes_value, bool hours_displayed, bool sign_displayed)
|
||||
Vector<DurationFormatPart> format_numeric_minutes(VM& vm, DurationFormat const& duration_format, MathematicalValue const& minutes_value, bool hours_displayed, bool sign_displayed)
|
||||
{
|
||||
auto& realm = *vm.current_realm();
|
||||
|
||||
|
|
@ -563,7 +565,7 @@ Vector<DurationFormatPart> format_numeric_minutes(VM& vm, DurationFormat const&
|
|||
auto number_format = construct_number_format(vm, duration_format, number_format_options);
|
||||
|
||||
// 12. Let minutesParts be PartitionNumberPattern(nf, minutesValue).
|
||||
auto minutes_parts = partition_number_pattern(number_format, MathematicalValue { minutes_value });
|
||||
auto minutes_parts = partition_number_pattern(number_format, minutes_value);
|
||||
|
||||
// 13. For each Record { [[Type]], [[Value]] } part of minutesParts, do
|
||||
result.ensure_capacity(result.size() + minutes_parts.size());
|
||||
|
|
@ -578,7 +580,7 @@ Vector<DurationFormatPart> format_numeric_minutes(VM& vm, DurationFormat const&
|
|||
}
|
||||
|
||||
// 1.1.11 FormatNumericSeconds ( durationFormat, secondsValue, minutesDisplayed, signDisplayed ), https://tc39.es/proposal-intl-duration-format/#sec-formatnumericseconds
|
||||
Vector<DurationFormatPart> format_numeric_seconds(VM& vm, DurationFormat const& duration_format, double seconds_value, bool minutes_displayed, bool sign_displayed)
|
||||
Vector<DurationFormatPart> format_numeric_seconds(VM& vm, DurationFormat const& duration_format, MathematicalValue const& seconds_value, bool minutes_displayed, bool sign_displayed)
|
||||
{
|
||||
auto& realm = *vm.current_realm();
|
||||
|
||||
|
|
@ -657,7 +659,7 @@ Vector<DurationFormatPart> format_numeric_seconds(VM& vm, DurationFormat const&
|
|||
auto number_format = construct_number_format(vm, duration_format, number_format_options);
|
||||
|
||||
// 17. Let secondsParts be PartitionNumberPattern(nf, secondsValue).
|
||||
auto seconds_parts = partition_number_pattern(number_format, MathematicalValue { seconds_value });
|
||||
auto seconds_parts = partition_number_pattern(number_format, seconds_value);
|
||||
|
||||
// 18. For each Record { [[Type]], [[Value]] } part of secondsParts, do
|
||||
result.ensure_capacity(result.size() + seconds_parts.size());
|
||||
|
|
@ -693,12 +695,12 @@ Vector<DurationFormatPart> format_numeric_units(VM& vm, DurationFormat const& du
|
|||
auto minutes_display = duration_format.minutes_display();
|
||||
|
||||
// 7. Let secondsValue be duration.[[Seconds]].
|
||||
auto seconds_value = duration.seconds;
|
||||
Crypto::BigFraction seconds_value { duration.seconds };
|
||||
|
||||
// 8. If duration.[[Milliseconds]] is not 0 or duration.[[Microseconds]] is not 0 or duration.[[Nanoseconds]] is not 0, then
|
||||
if (duration.milliseconds != 0 || duration.microseconds != 0 || duration.nanoseconds != 0) {
|
||||
// a. Set secondsValue to secondsValue + ComputeFractionalDigits(durationFormat, duration).
|
||||
seconds_value += compute_fractional_digits(duration_format, duration);
|
||||
seconds_value = seconds_value + compute_fractional_digits(duration_format, duration);
|
||||
}
|
||||
|
||||
// 9. Let secondsDisplay be durationFormat.[[SecondsDisplay]].
|
||||
|
|
@ -720,7 +722,7 @@ Vector<DurationFormatPart> format_numeric_units(VM& vm, DurationFormat const& du
|
|||
// a. Let secondsFormatted be true.
|
||||
// 13. Else,
|
||||
// a. Let secondsFormatted be false.
|
||||
auto seconds_formatted = seconds_value != 0 || seconds_display == DurationFormat::Display::Always;
|
||||
auto seconds_formatted = !seconds_value.is_zero() || seconds_display == DurationFormat::Display::Always;
|
||||
|
||||
// 14. Let minutesFormatted be false.
|
||||
auto minutes_formatted = false;
|
||||
|
|
@ -741,17 +743,19 @@ Vector<DurationFormatPart> format_numeric_units(VM& vm, DurationFormat const& du
|
|||
|
||||
// 16. If hoursFormatted is true, then
|
||||
if (hours_formatted) {
|
||||
MathematicalValue hours_mv { hours_value };
|
||||
|
||||
// a. If signDisplayed is true, then
|
||||
if (sign_displayed) {
|
||||
// i. If hoursValue is 0 and DurationSign(duration) is -1, then
|
||||
if (hours_value == 0 && duration_sign(duration) == -1) {
|
||||
// 1. Set hoursValue to NEGATIVE-ZERO.
|
||||
hours_value = -0.0;
|
||||
hours_mv = MathematicalValue { MathematicalValue::Symbol::NegativeZero };
|
||||
}
|
||||
}
|
||||
|
||||
// b. Let hoursParts be FormatNumericHours(durationFormat, hoursValue, signDisplayed).
|
||||
auto hours_parts = format_numeric_hours(vm, duration_format, hours_value, sign_displayed);
|
||||
auto hours_parts = format_numeric_hours(vm, duration_format, hours_mv, sign_displayed);
|
||||
|
||||
// b. Set numericPartsList to the list-concatenation of numericPartsList and hoursParts.
|
||||
numeric_parts_list.extend(move(hours_parts));
|
||||
|
|
@ -762,17 +766,19 @@ Vector<DurationFormatPart> format_numeric_units(VM& vm, DurationFormat const& du
|
|||
|
||||
// 17. If minutesFormatted is true, then
|
||||
if (minutes_formatted) {
|
||||
MathematicalValue minutes_mv { minutes_value };
|
||||
|
||||
// a. If signDisplayed is true, then
|
||||
if (sign_displayed) {
|
||||
// i. If minutesValue is 0 and DurationSign(duration) is -1, then
|
||||
if (minutes_value == 0 && duration_sign(duration) == -1) {
|
||||
// 1. Set minutesValue to NEGATIVE-ZERO.
|
||||
minutes_value = -0.0;
|
||||
minutes_mv = MathematicalValue { MathematicalValue::Symbol::NegativeZero };
|
||||
}
|
||||
}
|
||||
|
||||
// b. Let minutesParts be FormatNumericMinutes(durationFormat, minutesValue, hoursFormatted, signDisplayed).
|
||||
auto minutes_parts = format_numeric_minutes(vm, duration_format, minutes_value, hours_formatted, sign_displayed);
|
||||
auto minutes_parts = format_numeric_minutes(vm, duration_format, minutes_mv, hours_formatted, sign_displayed);
|
||||
|
||||
// c. Set numericPartsList to the list-concatenation of numericPartsList and minutesParts.
|
||||
numeric_parts_list.extend(move(minutes_parts));
|
||||
|
|
@ -784,7 +790,7 @@ Vector<DurationFormatPart> format_numeric_units(VM& vm, DurationFormat const& du
|
|||
// 18. If secondsFormatted is true, then
|
||||
if (seconds_formatted) {
|
||||
// a. Let secondsParts be FormatNumericSeconds(durationFormat, secondsValue, minutesFormatted, signDisplayed).
|
||||
auto seconds_parts = format_numeric_seconds(vm, duration_format, seconds_value, minutes_formatted, sign_displayed);
|
||||
auto seconds_parts = format_numeric_seconds(vm, duration_format, MathematicalValue { seconds_value.to_string(9) }, minutes_formatted, sign_displayed);
|
||||
|
||||
// b. Set numericPartsList to the list-concatenation of numericPartsList and secondsParts.
|
||||
numeric_parts_list.extend(move(seconds_parts));
|
||||
|
|
@ -904,7 +910,7 @@ Vector<DurationFormatPart> partition_duration_format_pattern(VM& vm, DurationFor
|
|||
auto const& duration_instances_component = duration_instances_components[i];
|
||||
|
||||
// a. Let value be the value of duration's field whose name is the Value Field value of the current row.
|
||||
auto value = duration.*duration_instances_component.value_slot;
|
||||
Crypto::BigFraction value { duration.*duration_instances_component.value_slot };
|
||||
|
||||
// b. Let style be the value of durationFormat's internal slot whose name is the Style Slot value of the current row.
|
||||
auto style = (duration_format.*duration_instances_component.get_style_slot)();
|
||||
|
|
@ -941,7 +947,7 @@ Vector<DurationFormatPart> partition_duration_format_pattern(VM& vm, DurationFor
|
|||
// 1. If NextUnitFractional(durationFormat, unit) is true, then
|
||||
if (next_unit_fractional(duration_format, unit)) {
|
||||
// a. Set value to value + ComputeFractionalDigits(durationFormat, duration).
|
||||
value += compute_fractional_digits(duration_format, duration);
|
||||
value = value + compute_fractional_digits(duration_format, duration);
|
||||
|
||||
u8 maximum_fraction_digits = 0;
|
||||
u8 minimum_fraction_digits = 0;
|
||||
|
|
@ -978,7 +984,9 @@ Vector<DurationFormatPart> partition_duration_format_pattern(VM& vm, DurationFor
|
|||
}
|
||||
|
||||
// iii. If value is not 0 or display is "always", then
|
||||
if (value != 0 || display == DurationFormat::Display::Always) {
|
||||
if (!value.is_zero() || display == DurationFormat::Display::Always) {
|
||||
MathematicalValue value_mv { value.to_string(9) };
|
||||
|
||||
// 1. Let numberingSystem be durationFormat.[[NumberingSystem]].
|
||||
auto const& numbering_system = duration_format.numbering_system();
|
||||
|
||||
|
|
@ -991,9 +999,9 @@ Vector<DurationFormatPart> partition_duration_format_pattern(VM& vm, DurationFor
|
|||
sign_displayed = false;
|
||||
|
||||
// b. If value is 0 and DurationSign(duration) is -1, then
|
||||
if (value == 0 && duration_sign(duration) == -1) {
|
||||
if (value.is_zero() && duration_sign(duration) == -1) {
|
||||
// i. Set value to NEGATIVE-ZERO.
|
||||
value = -0.0;
|
||||
value_mv = MathematicalValue { MathematicalValue::Symbol::NegativeZero };
|
||||
}
|
||||
}
|
||||
// 4. Else,
|
||||
|
|
@ -1019,7 +1027,7 @@ Vector<DurationFormatPart> partition_duration_format_pattern(VM& vm, DurationFor
|
|||
auto number_format = construct_number_format(vm, duration_format, number_format_options);
|
||||
|
||||
// 10. Let parts be PartitionNumberPattern(nf, value).
|
||||
auto parts = partition_number_pattern(number_format, MathematicalValue { value });
|
||||
auto parts = partition_number_pattern(number_format, value_mv);
|
||||
|
||||
// 11. Let list be a new empty List.
|
||||
Vector<DurationFormatPart> list;
|
||||
|
|
|
|||
|
|
@ -9,6 +9,7 @@
|
|||
|
||||
#include <AK/Array.h>
|
||||
#include <AK/String.h>
|
||||
#include <LibCrypto/BigFraction/BigFraction.h>
|
||||
#include <LibJS/Runtime/Intl/AbstractOperations.h>
|
||||
#include <LibJS/Runtime/Object.h>
|
||||
#include <LibUnicode/Locale.h>
|
||||
|
|
@ -263,11 +264,11 @@ struct DurationFormatPart {
|
|||
ThrowCompletionOr<DurationRecord> to_duration_record(VM&, Value input);
|
||||
i8 duration_sign(DurationRecord const&);
|
||||
ThrowCompletionOr<DurationUnitOptions> get_duration_unit_options(VM&, DurationFormat::Unit unit, Object const& options, DurationFormat::Style base_style, ReadonlySpan<StringView> styles_list, DurationFormat::ValueStyle digital_base, Optional<DurationFormat::ValueStyle> previous_style, bool two_digit_hours);
|
||||
double compute_fractional_digits(DurationFormat const&, DurationRecord const&);
|
||||
Crypto::BigFraction compute_fractional_digits(DurationFormat const&, DurationRecord const&);
|
||||
bool next_unit_fractional(DurationFormat const&, DurationFormat::Unit unit);
|
||||
Vector<DurationFormatPart> format_numeric_hours(VM&, DurationFormat const&, double hours_value, bool sign_displayed);
|
||||
Vector<DurationFormatPart> format_numeric_minutes(VM&, DurationFormat const&, double minutes_value, bool hours_displayed, bool sign_displayed);
|
||||
Vector<DurationFormatPart> format_numeric_seconds(VM&, DurationFormat const&, double seconds_value, bool minutes_displayed, bool sign_displayed);
|
||||
Vector<DurationFormatPart> format_numeric_hours(VM&, DurationFormat const&, MathematicalValue const& hours_value, bool sign_displayed);
|
||||
Vector<DurationFormatPart> format_numeric_minutes(VM&, DurationFormat const&, MathematicalValue const& minutes_value, bool hours_displayed, bool sign_displayed);
|
||||
Vector<DurationFormatPart> format_numeric_seconds(VM&, DurationFormat const&, MathematicalValue const& seconds_value, bool minutes_displayed, bool sign_displayed);
|
||||
Vector<DurationFormatPart> format_numeric_units(VM&, DurationFormat const&, DurationRecord const&, DurationFormat::Unit first_numeric_unit, bool sign_displayed);
|
||||
Vector<DurationFormatPart> list_format_parts(VM&, DurationFormat const&, Vector<Vector<DurationFormatPart>>& partitioned_parts_list);
|
||||
Vector<DurationFormatPart> partition_duration_format_pattern(VM&, DurationFormat const&, DurationRecord const&);
|
||||
|
|
|
|||
|
|
@ -145,6 +145,24 @@ describe("correct behavior", () => {
|
|||
const de = new Intl.DurationFormat("de", { style: "digital" });
|
||||
expect(de.format(duration)).toBe("123456:456789:789123");
|
||||
});
|
||||
|
||||
test("precise mathematical values", () => {
|
||||
const en = new Intl.DurationFormat("en", { style: "digital" });
|
||||
|
||||
let duration = {
|
||||
seconds: 10000000,
|
||||
nanoseconds: 1,
|
||||
};
|
||||
expect(en.format(duration)).toBe("0:00:10000000.000000001");
|
||||
|
||||
duration = {
|
||||
seconds: 1,
|
||||
milliseconds: 2,
|
||||
microseconds: 3,
|
||||
nanoseconds: Number.MAX_SAFE_INTEGER,
|
||||
};
|
||||
expect(en.format(duration)).toBe("0:00:9007200.256743991");
|
||||
});
|
||||
});
|
||||
|
||||
describe("errors", () => {
|
||||
|
|
|
|||
Loading…
Add table
Reference in a new issue