1
Fork 0
mirror of https://github.com/RGBCube/serenity synced 2025-07-25 22:07:35 +00:00

LibJS: Start fleshing out an ISO 8601 parser for Temporal

This is the start of a parser for the ISO 8601 grammar used in the
Temporal spec:
https://tc39.es/proposal-temporal/#sec-temporal-iso8601grammar

We will, on purpose, not use a generic ISO 8601 parser from AK or
similar for two reasons:

- Many AOs make specific assumptions about which productions exist and
  access them directly, even when they're part of a larger production.
- The spec says "The grammar deviates from the standard given in ISO
  8601 in the following ways:" and then lists 17 of such deviations.
  Making that work with a general purpose parser is not worth it.

The public API is not being used anywhere yet, but will be in the next
couple of commits. Likewise, the Production enum will be populated with
all the productions accessed directly (e.g. TemporalDateString).

Many thanks to Ali for showing me how to improve my initial approach
full of macros with a nice RAII helper - it's much nicer :^)

Co-Authored-By: Ali Mohammad Pur <mpfard@serenityos.org>
This commit is contained in:
Linus Groh 2021-11-19 18:04:50 +00:00
parent dd76ba2fe1
commit de23f0b68c
5 changed files with 634 additions and 26 deletions

View file

@ -161,6 +161,7 @@ set(SOURCES
Runtime/Temporal/Instant.cpp
Runtime/Temporal/InstantConstructor.cpp
Runtime/Temporal/InstantPrototype.cpp
Runtime/Temporal/ISO8601.cpp
Runtime/Temporal/Now.cpp
Runtime/Temporal/PlainDate.cpp
Runtime/Temporal/PlainDateConstructor.cpp

View file

@ -16,6 +16,7 @@
#include <LibJS/Runtime/Temporal/AbstractOperations.h>
#include <LibJS/Runtime/Temporal/Calendar.h>
#include <LibJS/Runtime/Temporal/Duration.h>
#include <LibJS/Runtime/Temporal/ISO8601.h>
#include <LibJS/Runtime/Temporal/PlainDate.h>
#include <LibJS/Runtime/Temporal/PlainDateTime.h>
#include <LibJS/Runtime/Temporal/PlainTime.h>
@ -1052,22 +1053,21 @@ BigInt* round_number_to_increment(GlobalObject& global_object, BigInt const& x,
}
// 13.34 ParseISODateTime ( isoString ), https://tc39.es/proposal-temporal/#sec-temporal-parseisodatetime
ThrowCompletionOr<ISODateTime> parse_iso_date_time(GlobalObject& global_object, [[maybe_unused]] String const& iso_string)
ThrowCompletionOr<ISODateTime> parse_iso_date_time(GlobalObject& global_object, ParseResult const& parse_result)
{
auto& vm = global_object.vm();
// 1. Assert: Type(isoString) is String.
// 2. Let year, month, day, hour, minute, second, fraction, and calendar be the parts of isoString produced respectively by the DateYear, DateMonth, DateDay, TimeHour, TimeMinute, TimeSecond, TimeFractionalPart, and CalendarName productions, or undefined if not present.
Optional<StringView> year_part;
Optional<StringView> month_part;
Optional<StringView> day_part;
Optional<StringView> hour_part;
Optional<StringView> minute_part;
Optional<StringView> second_part;
Optional<StringView> fraction_part;
Optional<StringView> calendar_part;
return vm.throw_completion<InternalError>(global_object, ErrorType::NotImplemented, "ParseISODateTime");
auto year_part = parse_result.date_year;
auto month_part = parse_result.date_month;
auto day_part = parse_result.date_day;
auto hour_part = parse_result.time_hour;
auto minute_part = parse_result.time_minute;
auto second_part = parse_result.time_second;
auto fraction_part = parse_result.time_fractional_part;
auto calendar_part = parse_result.calendar_name;
// 3. Let year be the part of isoString produced by the DateYear production.
// 4. If the first code unit of year is 0x2212 (MINUS SIGN), replace it with the code unit 0x002D (HYPHEN-MINUS).
@ -1075,10 +1075,10 @@ ThrowCompletionOr<ISODateTime> parse_iso_date_time(GlobalObject& global_object,
if (year_part.has_value() && year_part->starts_with("\xE2\x88\x92"sv))
normalized_year = String::formatted("-{}", year_part->substring_view(3));
else
normalized_year = year_part.value_or("");
normalized_year = year_part.value_or("0");
// 5. Set year to ! ToIntegerOrInfinity(year).
i32 year = MUST(Value(js_string(vm, normalized_year)).to_integer_or_infinity(global_object));
auto year = *normalized_year.to_int<i32>();
u8 month;
// 6. If month is undefined, then
@ -1168,7 +1168,7 @@ ThrowCompletionOr<TemporalInstant> parse_temporal_instant_string(GlobalObject& g
// TODO
// 3. Let result be ! ParseISODateTime(isoString).
auto result = MUST(parse_iso_date_time(global_object, iso_string));
auto result = MUST(parse_iso_date_time(global_object, {}));
// 4. Let timeZoneResult be ? ParseTemporalTimeZoneString(isoString).
auto time_zone_result = TRY(parse_temporal_time_zone_string(global_object, iso_string));
@ -1199,7 +1199,7 @@ ThrowCompletionOr<TemporalZonedDateTime> parse_temporal_zoned_date_time_string(G
// TODO
// 3. Let result be ! ParseISODateTime(isoString).
auto result = MUST(parse_iso_date_time(global_object, iso_string));
auto result = MUST(parse_iso_date_time(global_object, {}));
// 4. Let timeZoneResult be ? ParseTemporalTimeZoneString(isoString).
auto time_zone_result = TRY(parse_temporal_time_zone_string(global_object, iso_string));
@ -1240,7 +1240,7 @@ ThrowCompletionOr<String> parse_temporal_calendar_string(GlobalObject& global_ob
}
// 13.38 ParseTemporalDateString ( isoString ), https://tc39.es/proposal-temporal/#sec-temporal-parsetemporaldatestring
ThrowCompletionOr<TemporalDate> parse_temporal_date_string(GlobalObject& global_object, String const& iso_string)
ThrowCompletionOr<TemporalDate> parse_temporal_date_string(GlobalObject& global_object, [[maybe_unused]] String const& iso_string)
{
// 1. Assert: Type(isoString) is String.
@ -1249,14 +1249,14 @@ ThrowCompletionOr<TemporalDate> parse_temporal_date_string(GlobalObject& global_
// TODO
// 3. Let result be ? ParseISODateTime(isoString).
auto result = TRY(parse_iso_date_time(global_object, iso_string));
auto result = TRY(parse_iso_date_time(global_object, {}));
// 4. Return the Record { [[Year]]: result.[[Year]], [[Month]]: result.[[Month]], [[Day]]: result.[[Day]], [[Calendar]]: result.[[Calendar]] }.
return TemporalDate { .year = result.year, .month = result.month, .day = result.day, .calendar = move(result.calendar) };
}
// 13.39 ParseTemporalDateTimeString ( isoString ), https://tc39.es/proposal-temporal/#sec-temporal-parsetemporaldatetimestring
ThrowCompletionOr<ISODateTime> parse_temporal_date_time_string(GlobalObject& global_object, String const& iso_string)
ThrowCompletionOr<ISODateTime> parse_temporal_date_time_string(GlobalObject& global_object, [[maybe_unused]] String const& iso_string)
{
// 1. Assert: Type(isoString) is String.
@ -1265,7 +1265,7 @@ ThrowCompletionOr<ISODateTime> parse_temporal_date_time_string(GlobalObject& glo
// TODO
// 3. Let result be ? ParseISODateTime(isoString).
auto result = TRY(parse_iso_date_time(global_object, iso_string));
auto result = TRY(parse_iso_date_time(global_object, {}));
// 4. Return result.
return result;
@ -1280,7 +1280,7 @@ ThrowCompletionOr<TemporalDuration> parse_temporal_duration_string(GlobalObject&
}
// 13.41 ParseTemporalMonthDayString ( isoString ), https://tc39.es/proposal-temporal/#sec-temporal-parsetemporalmonthdaystring
ThrowCompletionOr<TemporalMonthDay> parse_temporal_month_day_string(GlobalObject& global_object, String const& iso_string)
ThrowCompletionOr<TemporalMonthDay> parse_temporal_month_day_string(GlobalObject& global_object, [[maybe_unused]] String const& iso_string)
{
// 1. Assert: Type(isoString) is String.
@ -1289,7 +1289,7 @@ ThrowCompletionOr<TemporalMonthDay> parse_temporal_month_day_string(GlobalObject
// TODO
// 3. Let result be ? ParseISODateTime(isoString).
auto result = TRY(parse_iso_date_time(global_object, iso_string));
auto result = TRY(parse_iso_date_time(global_object, {}));
// 4. Let year be result.[[Year]].
Optional<i32> year = result.year;
@ -1303,7 +1303,7 @@ ThrowCompletionOr<TemporalMonthDay> parse_temporal_month_day_string(GlobalObject
}
// 13.42 ParseTemporalRelativeToString ( isoString ), https://tc39.es/proposal-temporal/#sec-temporal-parsetemporalrelativetostring
ThrowCompletionOr<TemporalZonedDateTime> parse_temporal_relative_to_string(GlobalObject& global_object, String const& iso_string)
ThrowCompletionOr<TemporalZonedDateTime> parse_temporal_relative_to_string(GlobalObject& global_object, [[maybe_unused]] String const& iso_string)
{
// 1. Assert: Type(isoString) is String.
@ -1312,7 +1312,7 @@ ThrowCompletionOr<TemporalZonedDateTime> parse_temporal_relative_to_string(Globa
// TODO
// 3. Let result be ! ParseISODateTime(isoString).
auto result = MUST(parse_iso_date_time(global_object, iso_string));
auto result = MUST(parse_iso_date_time(global_object, {}));
bool z;
Optional<String> offset;
@ -1345,7 +1345,7 @@ ThrowCompletionOr<TemporalTime> parse_temporal_time_string(GlobalObject& global_
// TODO
// 3. Let result be ? ParseISODateTime(isoString).
auto result = TRY(parse_iso_date_time(global_object, iso_string));
auto result = TRY(parse_iso_date_time(global_object, {}));
// 4. Return the Record { [[Hour]]: result.[[Hour]], [[Minute]]: result.[[Minute]], [[Second]]: result.[[Second]], [[Millisecond]]: result.[[Millisecond]], [[Microsecond]]: result.[[Microsecond]], [[Nanosecond]]: result.[[Nanosecond]], [[Calendar]]: result.[[Calendar]] }.
return TemporalTime { .hour = result.hour, .minute = result.minute, .second = result.second, .millisecond = result.millisecond, .microsecond = result.microsecond, .nanosecond = result.nanosecond, .calendar = move(result.calendar) };
@ -1444,7 +1444,7 @@ ThrowCompletionOr<TemporalTimeZone> parse_temporal_time_zone_string(GlobalObject
}
// 13.45 ParseTemporalYearMonthString ( isoString ), https://tc39.es/proposal-temporal/#sec-temporal-parsetemporalyearmonthstring
ThrowCompletionOr<TemporalYearMonth> parse_temporal_year_month_string(GlobalObject& global_object, String const& iso_string)
ThrowCompletionOr<TemporalYearMonth> parse_temporal_year_month_string(GlobalObject& global_object, [[maybe_unused]] String const& iso_string)
{
// 1. Assert: Type(isoString) is String.
@ -1453,7 +1453,7 @@ ThrowCompletionOr<TemporalYearMonth> parse_temporal_year_month_string(GlobalObje
// TODO
// 3. Let result be ? ParseISODateTime(isoString).
auto result = TRY(parse_iso_date_time(global_object, iso_string));
auto result = TRY(parse_iso_date_time(global_object, {}));
// 4. Return the Record { [[Year]]: result.[[Year]], [[Month]]: result.[[Month]], [[Day]]: result.[[Day]], [[Calendar]]: result.[[Calendar]] }.
return TemporalYearMonth { .year = result.year, .month = result.month, .day = result.day, .calendar = move(result.calendar) };

View file

@ -13,6 +13,7 @@
#include <LibJS/Forward.h>
#include <LibJS/Runtime/Completion.h>
#include <LibJS/Runtime/GlobalObject.h>
#include <LibJS/Runtime/Temporal/ISO8601.h>
namespace JS::Temporal {
@ -126,7 +127,7 @@ double sign(double);
double constrain_to_range(double x, double minimum, double maximum);
i64 round_number_to_increment(double, u64 increment, StringView rounding_mode);
BigInt* round_number_to_increment(GlobalObject&, BigInt const&, u64 increment, StringView rounding_mode);
ThrowCompletionOr<ISODateTime> parse_iso_date_time(GlobalObject&, String const& iso_string);
ThrowCompletionOr<ISODateTime> parse_iso_date_time(GlobalObject&, ParseResult const& parse_result);
ThrowCompletionOr<TemporalInstant> parse_temporal_instant_string(GlobalObject&, String const& iso_string);
ThrowCompletionOr<TemporalZonedDateTime> parse_temporal_zoned_date_time_string(GlobalObject&, String const& iso_string);
ThrowCompletionOr<String> parse_temporal_calendar_string(GlobalObject&, String const& iso_string);

View file

@ -0,0 +1,489 @@
/*
* Copyright (c) 2021, Linus Groh <linusg@serenityos.org>
*
* SPDX-License-Identifier: BSD-2-Clause
*/
#include <AK/CharacterTypes.h>
#include <LibJS/Runtime/Temporal/ISO8601.h>
namespace JS::Temporal {
namespace Detail {
// https://tc39.es/proposal-temporal/#prod-DecimalDigit
bool ISO8601Parser::parse_decimal_digit()
{
// DecimalDigit : one of
// 0 1 2 3 4 5 6 7 8 9
if (m_state.lexer.next_is(is_ascii_digit)) {
m_state.lexer.consume();
return true;
}
return false;
}
// https://tc39.es/proposal-temporal/#prod-NonZeroDigit
bool ISO8601Parser::parse_non_zero_digit()
{
// NonZeroDigit : one of
// 1 2 3 4 5 6 7 8 9
if (m_state.lexer.next_is(is_ascii_digit) && !m_state.lexer.next_is('0')) {
m_state.lexer.consume();
return true;
}
return false;
}
// https://tc39.es/proposal-temporal/#prod-ASCIISign
bool ISO8601Parser::parse_ascii_sign()
{
// ASCIISign : one of
// + -
return m_state.lexer.consume_specific('+')
|| m_state.lexer.consume_specific('-');
}
// https://tc39.es/proposal-temporal/#prod-Sign
bool ISO8601Parser::parse_sign()
{
// Sign :
// ASCIISign
// U+2212
StateTransaction transaction { *this };
auto success = parse_ascii_sign()
|| m_state.lexer.consume_specific("\xE2\x88\x92"sv);
if (!success)
return false;
m_state.parse_result.sign = transaction.parsed_string_view();
transaction.commit();
return true;
}
// https://tc39.es/proposal-temporal/#prod-Hour
bool ISO8601Parser::parse_hour()
{
// Hour :
// 0 DecimalDigit
// 1 DecimalDigit
// 20
// 21
// 22
// 23
StateTransaction transaction { *this };
if (m_state.lexer.consume_specific('0') || m_state.lexer.consume_specific('1')) {
if (!parse_decimal_digit())
return false;
} else {
auto success = m_state.lexer.consume_specific("20"sv)
|| m_state.lexer.consume_specific("21"sv)
|| m_state.lexer.consume_specific("22"sv)
|| m_state.lexer.consume_specific("23"sv);
if (!success)
return false;
}
transaction.commit();
return true;
}
// https://tc39.es/proposal-temporal/#prod-MinuteSecond
bool ISO8601Parser::parse_minute_second()
{
// MinuteSecond :
// 0 DecimalDigit
// 1 DecimalDigit
// 2 DecimalDigit
// 3 DecimalDigit
// 4 DecimalDigit
// 5 DecimalDigit
StateTransaction transaction { *this };
auto success = m_state.lexer.consume_specific('0')
|| m_state.lexer.consume_specific('1')
|| m_state.lexer.consume_specific('2')
|| m_state.lexer.consume_specific('3')
|| m_state.lexer.consume_specific('4')
|| m_state.lexer.consume_specific('5');
if (!success)
return false;
if (!parse_decimal_digit())
return false;
transaction.commit();
return true;
}
// https://tc39.es/proposal-temporal/#prod-DecimalSeparator
bool ISO8601Parser::parse_decimal_separator()
{
// DecimalSeparator : one of
// . ,
return m_state.lexer.consume_specific('.')
|| m_state.lexer.consume_specific(',');
}
// https://tc39.es/proposal-temporal/#prod-DateTimeSeparator
bool ISO8601Parser::parse_date_time_separator()
{
// DateTimeSeparator :
// <SP>
// T
// t
return m_state.lexer.consume_specific(' ')
|| m_state.lexer.consume_specific('T')
|| m_state.lexer.consume_specific('t');
}
// https://tc39.es/proposal-temporal/#prod-DateYear
bool ISO8601Parser::parse_date_year()
{
// DateFourDigitYear :
// DecimalDigit DecimalDigit DecimalDigit DecimalDigit
// DateExtendedYear :
// Sign DecimalDigit DecimalDigit DecimalDigit DecimalDigit DecimalDigit DecimalDigit
// DateYear :
// DateFourDigitYear
// DateExtendedYear
StateTransaction transaction { *this };
if (parse_sign()) {
for (size_t i = 0; i < 6; ++i) {
if (!parse_decimal_digit())
return false;
}
} else {
for (size_t i = 0; i < 4; ++i) {
if (!parse_decimal_digit())
return false;
}
}
m_state.parse_result.date_year = transaction.parsed_string_view();
transaction.commit();
return true;
}
// https://tc39.es/proposal-temporal/#prod-DateMonth
bool ISO8601Parser::parse_date_month()
{
// DateMonth :
// 0 NonZeroDigit
// 10
// 11
// 12
StateTransaction transaction { *this };
if (m_state.lexer.consume_specific('0')) {
if (!parse_non_zero_digit())
return false;
} else {
auto success = m_state.lexer.consume_specific("10"sv)
|| m_state.lexer.consume_specific("11"sv)
|| m_state.lexer.consume_specific("12"sv);
if (!success)
return false;
}
m_state.parse_result.date_month = transaction.parsed_string_view();
transaction.commit();
return true;
}
// https://tc39.es/proposal-temporal/#prod-DateDay
bool ISO8601Parser::parse_date_day()
{
// DateDay :
// 0 NonZeroDigit
// 1 DecimalDigit
// 2 DecimalDigit
// 30
// 31
StateTransaction transaction { *this };
if (m_state.lexer.consume_specific('0')) {
if (!parse_non_zero_digit())
return false;
} else if (m_state.lexer.consume_specific('1') || m_state.lexer.consume_specific('2')) {
if (!parse_decimal_digit())
return false;
} else {
auto success = m_state.lexer.consume_specific("30"sv)
|| m_state.lexer.consume_specific("31"sv);
if (!success)
return false;
}
m_state.parse_result.date_day = transaction.parsed_string_view();
transaction.commit();
return true;
}
// https://tc39.es/proposal-temporal/#prod-Date
bool ISO8601Parser::parse_date()
{
// Date :
// DateYear - DateMonth - DateDay
// DateYear DateMonth DateDay
StateTransaction transaction { *this };
if (!parse_date_year())
return false;
auto with_dashes = m_state.lexer.consume_specific('-');
if (!parse_date_month())
return false;
if (with_dashes && !m_state.lexer.consume_specific('-'))
return false;
if (!parse_date_day())
return false;
transaction.commit();
return true;
}
// https://tc39.es/proposal-temporal/#prod-TimeHour
bool ISO8601Parser::parse_time_hour()
{
// TimeHour :
// Hour
StateTransaction transaction { *this };
if (!parse_hour())
return false;
m_state.parse_result.time_hour = transaction.parsed_string_view();
transaction.commit();
return true;
}
// https://tc39.es/proposal-temporal/#prod-TimeMinute
bool ISO8601Parser::parse_time_minute()
{
// TimeMinute :
// MinuteSecond
StateTransaction transaction { *this };
if (!parse_minute_second())
return false;
m_state.parse_result.time_minute = transaction.parsed_string_view();
transaction.commit();
return true;
}
// https://tc39.es/proposal-temporal/#prod-TimeSecond
bool ISO8601Parser::parse_time_second()
{
// TimeSecond :
// MinuteSecond
// 60
StateTransaction transaction { *this };
auto success = parse_minute_second()
|| m_state.lexer.consume_specific("60"sv);
if (!success)
return false;
m_state.parse_result.time_second = transaction.parsed_string_view();
transaction.commit();
return true;
}
// https://tc39.es/proposal-temporal/#prod-FractionalPart
bool ISO8601Parser::parse_fractional_part()
{
// FractionalPart :
// DecimalDigit DecimalDigit[opt] DecimalDigit[opt] DecimalDigit[opt] DecimalDigit[opt] DecimalDigit[opt] DecimalDigit[opt] DecimalDigit[opt] DecimalDigit[opt]
if (!parse_decimal_digit())
return false;
for (size_t i = 0; i < 8; ++i) {
if (!parse_decimal_digit())
break;
}
return true;
}
// https://tc39.es/proposal-temporal/#prod-TimeFractionalPart
bool ISO8601Parser::parse_time_fractional_part()
{
// TimeFractionalPart :
// FractionalPart
StateTransaction transaction { *this };
if (!parse_fractional_part())
return false;
m_state.parse_result.time_fractional_part = transaction.parsed_string_view();
transaction.commit();
return true;
}
// https://tc39.es/proposal-temporal/#prod-Fraction
bool ISO8601Parser::parse_fraction()
{
// Fraction :
// DecimalSeparator TimeFractionalPart
StateTransaction transaction { *this };
if (!parse_decimal_separator())
return false;
if (!parse_time_fractional_part())
return false;
transaction.commit();
return true;
}
// https://tc39.es/proposal-temporal/#prod-TimeFraction
bool ISO8601Parser::parse_time_fraction()
{
// TimeFraction :
// Fraction
return parse_fraction();
}
// https://tc39.es/proposal-temporal/#prod-TimeZoneOffsetRequired
bool ISO8601Parser::parse_time_zone_offset_required()
{
// TimeZoneOffsetRequired :
// TimeZoneUTCOffset TimeZoneBracketedAnnotation[opt]
return false;
}
// https://tc39.es/proposal-temporal/#prod-TimeZoneNameRequired
bool ISO8601Parser::parse_time_zone_name_required()
{
// TimeZoneNameRequired :
// TimeZoneUTCOffset[opt] TimeZoneBracketedAnnotation
return false;
}
// https://tc39.es/proposal-temporal/#prod-TimeZone
bool ISO8601Parser::parse_time_zone()
{
// TimeZone :
// TimeZoneOffsetRequired
// TimeZoneNameRequired
return parse_time_zone_offset_required()
|| parse_time_zone_name_required();
}
// https://tc39.es/proposal-temporal/#prod-CalendarName
bool ISO8601Parser::parse_calendar_name()
{
// CalChar :
// Alpha
// DecimalDigit
// CalendarNameComponent :
// CalChar CalChar CalChar CalChar[opt] CalChar[opt] CalChar[opt] CalChar[opt] CalChar[opt]
// CalendarNameTail :
// CalendarNameComponent
// CalendarNameComponent - CalendarNameTail
// CalendarName :
// CalendarNameTail
auto parse_calendar_name_component = [&] {
for (size_t i = 0; i < 8; ++i) {
if (!m_state.lexer.next_is(is_ascii_alphanumeric))
return i > 2;
m_state.lexer.consume();
}
return true;
};
StateTransaction transaction { *this };
do {
if (!parse_calendar_name_component())
return false;
} while (m_state.lexer.consume_specific('-'));
m_state.parse_result.calendar_name = transaction.parsed_string_view();
transaction.commit();
return true;
}
// https://tc39.es/proposal-temporal/#prod-Calendar
bool ISO8601Parser::parse_calendar()
{
// Calendar :
// [u-ca= CalendarName ]
StateTransaction transaction { *this };
if (!m_state.lexer.consume_specific("[u-ca="sv))
return false;
if (!parse_calendar_name())
return false;
if (!m_state.lexer.consume_specific(']'))
return false;
transaction.commit();
return true;
}
// https://tc39.es/proposal-temporal/#prod-TimeSpec
bool ISO8601Parser::parse_time_spec()
{
// TimeSpec :
// TimeHour
// TimeHour : TimeMinute
// TimeHour TimeMinute
// TimeHour : TimeMinute : TimeSecond TimeFraction[opt]
// TimeHour TimeMinute TimeSecond TimeFraction[opt]
StateTransaction transaction { *this };
if (!parse_time_hour())
return false;
if (m_state.lexer.consume_specific(':')) {
if (!parse_time_minute())
return false;
if (m_state.lexer.consume_specific(':')) {
if (!parse_time_second())
return false;
(void)parse_time_fraction();
}
} else if (parse_time_minute()) {
if (parse_time_second())
(void)parse_time_fraction();
}
transaction.commit();
return true;
}
// https://tc39.es/proposal-temporal/#prod-TimeSpecSeparator
bool ISO8601Parser::parse_time_spec_separator()
{
// TimeSpecSeparator :
// DateTimeSeparator TimeSpec
StateTransaction transaction { *this };
if (!parse_date_time_separator())
return false;
if (!parse_time_spec())
return false;
transaction.commit();
return true;
}
// https://tc39.es/proposal-temporal/#prod-DateTime
bool ISO8601Parser::parse_date_time()
{
// DateTime :
// Date TimeSpecSeparator[opt] TimeZone[opt]
if (!parse_date())
return false;
(void)parse_time_spec_separator();
(void)parse_time_zone();
return true;
}
// https://tc39.es/proposal-temporal/#prod-CalendarDateTime
bool ISO8601Parser::parse_calendar_date_time()
{
// CalendarDateTime :
// DateTime Calendar[opt]
if (!parse_date_time())
return false;
(void)parse_calendar();
return true;
}
}
#define JS_ENUMERATE_ISO8601_PRODUCTION_PARSERS
Optional<ParseResult> parse_iso8601(Production production, StringView input)
{
auto parser = Detail::ISO8601Parser { input };
switch (production) {
#define __JS_ENUMERATE(ProductionName, parse_production) \
case Production::ProductionName: \
if (!parser.parse_production()) \
return {}; \
break;
JS_ENUMERATE_ISO8601_PRODUCTION_PARSERS
#undef __JS_ENUMERATE
default:
VERIFY_NOT_REACHED();
}
// If we parsed successfully but didn't reach the end, the string doesn't match the given production.
if (!parser.lexer().is_eof())
return {};
return parser.parse_result();
}
}

View file

@ -0,0 +1,117 @@
/*
* Copyright (c) 2021, Linus Groh <linusg@serenityos.org>
*
* SPDX-License-Identifier: BSD-2-Clause
*/
#pragma once
#include <AK/GenericLexer.h>
#include <AK/Optional.h>
#include <AK/StringView.h>
#include <AK/Vector.h>
namespace JS::Temporal {
struct ParseResult {
Optional<StringView> sign;
Optional<StringView> date_year;
Optional<StringView> date_month;
Optional<StringView> date_day;
Optional<StringView> time_hour;
Optional<StringView> time_minute;
Optional<StringView> time_second;
Optional<StringView> time_fractional_part;
Optional<StringView> calendar_name;
};
enum class Production {
};
Optional<ParseResult> parse_iso8601(Production, StringView);
namespace Detail {
class ISO8601Parser {
public:
explicit ISO8601Parser(StringView input)
: m_input(input)
, m_state({
.lexer = GenericLexer { input },
.parse_result = {},
})
{
}
[[nodiscard]] GenericLexer const& lexer() const { return m_state.lexer; }
[[nodiscard]] ParseResult const& parse_result() const { return m_state.parse_result; }
[[nodiscard]] bool parse_decimal_digit();
[[nodiscard]] bool parse_non_zero_digit();
[[nodiscard]] bool parse_ascii_sign();
[[nodiscard]] bool parse_sign();
[[nodiscard]] bool parse_hour();
[[nodiscard]] bool parse_minute_second();
[[nodiscard]] bool parse_decimal_separator();
[[nodiscard]] bool parse_date_time_separator();
[[nodiscard]] bool parse_date_year();
[[nodiscard]] bool parse_date_month();
[[nodiscard]] bool parse_date_day();
[[nodiscard]] bool parse_date();
[[nodiscard]] bool parse_time_hour();
[[nodiscard]] bool parse_time_minute();
[[nodiscard]] bool parse_time_second();
[[nodiscard]] bool parse_fractional_part();
[[nodiscard]] bool parse_time_fractional_part();
[[nodiscard]] bool parse_fraction();
[[nodiscard]] bool parse_time_fraction();
[[nodiscard]] bool parse_time_zone_offset_required();
[[nodiscard]] bool parse_time_zone_name_required();
[[nodiscard]] bool parse_time_zone();
[[nodiscard]] bool parse_calendar_name();
[[nodiscard]] bool parse_calendar();
[[nodiscard]] bool parse_time_spec();
[[nodiscard]] bool parse_time_spec_separator();
[[nodiscard]] bool parse_date_time();
[[nodiscard]] bool parse_calendar_date_time();
private:
struct State {
GenericLexer lexer;
ParseResult parse_result;
};
struct StateTransaction {
explicit StateTransaction(ISO8601Parser& parser)
: m_parser(parser)
, m_saved_state(parser.m_state)
, m_start_index(parser.m_state.lexer.tell())
{
}
~StateTransaction()
{
if (!m_commit)
m_parser.m_state = move(m_saved_state);
}
void commit() { m_commit = true; }
StringView parsed_string_view() const
{
return m_parser.m_input.substring_view(m_start_index, m_parser.m_state.lexer.tell() - m_start_index);
}
private:
ISO8601Parser& m_parser;
State m_saved_state;
size_t m_start_index { 0 };
bool m_commit { false };
};
StringView m_input;
State m_state;
};
}
}