beenull/ladybird
1
0
Fork 0
mirror of https://github.com/LadybirdBrowser/ladybird.git synced 2024-11-21 23:20:20 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 6
ladybird/Libraries/LibUnicode/DateTimeFormat.cpp

1010 lines
32 KiB
C++
Raw Permalink Normal View History

LibUnicode: Parse and generate date, time, and date-time format patterns
2021-11-27 19:54:48 +00:00
/*
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
* Copyright (c) 2021-2024, Tim Flynn <trflynn89@serenityos.org>
LibUnicode: Parse and generate date, time, and date-time format patterns
2021-11-27 19:54:48 +00:00
*
* SPDX-License-Identifier: BSD-2-Clause
*/
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
#include <AK/AllOf.h>
LibUnicode: Add method to combine two format pattern skeletons The fields of the generated elements must be in the same order as the table here: https://unicode.org/reports/tr35/tr35-dates.html#Date_Field_Symbol_Table Further, only one field from each group of fields is allowed.
2021-12-09 00:52:48 +00:00
#include <AK/Array.h>
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
#include <AK/GenericLexer.h>
LibUnicode: Add method to combine two format pattern skeletons The fields of the generated elements must be in the same order as the table here: https://unicode.org/reports/tr35/tr35-dates.html#Date_Field_Symbol_Table Further, only one field from each group of fields is allowed.
2021-12-09 00:52:48 +00:00
#include <AK/StringBuilder.h>
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
#include <AK/TypeCasts.h>
LibUnicode+Everywhere: Merge LibLocale back into LibUnicode LibLocale was split off from LibUnicode a couple years ago to reduce the number of applications on SerenityOS that depend on CLDR data. Now that we use ICU, both LibUnicode and LibLocale are actually linking in this data. And since vcpkg gives us static libraries, both libraries are over 30MB in size. This patch reverts the separation and merges LibLocale into LibUnicode again. We now have just one library that includes the ICU data. Further, this will let LibUnicode share the locale cache that previously would only exist in LibLocale.
2024-06-23 13:14:27 +00:00
#include <LibUnicode/DateTimeFormat.h>
#include <LibUnicode/ICU.h>
#include <LibUnicode/Locale.h>
#include <LibUnicode/NumberFormat.h>
#include <LibUnicode/PartitionRange.h>
LibUnicode: Implement TR-35's localized GMT offset formatting This adds an API to use LibTimeZone to convert a time zone such as "America/New_York" to a GMT offset string like "GMT-5" (short form) or "GMT-05:00" (long form).
2022-01-11 16:35:50 +00:00
#include <stdlib.h>
LibUnicode: Parse and generate date, time, and date-time format patterns
2021-11-27 19:54:48 +00:00
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
#include <unicode/calendar.h>
#include <unicode/datefmt.h>
#include <unicode/dtitvfmt.h>
#include <unicode/dtptngen.h>
#include <unicode/gregocal.h>
#include <unicode/smpdtfmt.h>
#include <unicode/timezone.h>
LibJS+LibLocale: Replace calendar weekday information with ICU
2024-06-12 21:00:45 +00:00
#include <unicode/ucal.h>
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
LibUnicode+Everywhere: Merge LibLocale back into LibUnicode LibLocale was split off from LibUnicode a couple years ago to reduce the number of applications on SerenityOS that depend on CLDR data. Now that we use ICU, both LibUnicode and LibLocale are actually linking in this data. And since vcpkg gives us static libraries, both libraries are over 30MB in size. This patch reverts the separation and merges LibLocale into LibUnicode again. We now have just one library that includes the ICU data. Further, this will let LibUnicode share the locale cache that previously would only exist in LibLocale.
2024-06-23 13:14:27 +00:00
namespace Unicode {
LibUnicode: Parse and generate date, time, and date-time format patterns
2021-11-27 19:54:48 +00:00
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
DateTimeStyle date_time_style_from_string(StringView style)
{
if (style == "full"sv)
return DateTimeStyle::Full;
if (style == "long"sv)
return DateTimeStyle::Long;
if (style == "medium"sv)
return DateTimeStyle::Medium;
if (style == "short"sv)
return DateTimeStyle::Short;
VERIFY_NOT_REACHED();
}
StringView date_time_style_to_string(DateTimeStyle style)
{
switch (style) {
case DateTimeStyle::Full:
return "full"sv;
case DateTimeStyle::Long:
return "long"sv;
case DateTimeStyle::Medium:
return "medium"sv;
case DateTimeStyle::Short:
return "short"sv;
}
VERIFY_NOT_REACHED();
}
static constexpr icu::DateFormat::EStyle icu_date_time_style(DateTimeStyle style)
{
switch (style) {
case DateTimeStyle::Full:
return icu::DateFormat::EStyle::kFull;
case DateTimeStyle::Long:
return icu::DateFormat::EStyle::kLong;
case DateTimeStyle::Medium:
return icu::DateFormat::EStyle::kMedium;
case DateTimeStyle::Short:
return icu::DateFormat::EStyle::kShort;
}
VERIFY_NOT_REACHED();
}
LibUnicode: Parse and generate regional hour cycles Unlike most data in the CLDR, hour cycles are not stored on a per-locale basis. Instead, they are keyed by a string that is usually a region, but sometimes is a locale. Therefore, given a locale, to determine the hour cycles for that locale, we: 1. Check if the locale itself is assigned hour cycles. 2. If the locale has a region, check if that region is assigned hour cycles. 3. Otherwise, maximize that locale, and if the maximized locale has a region, check if that region is assigned hour cycles. 4. If the above all fail, fallback to the "001" region. Further, each locale's default hour cycle is the first assigned hour cycle.
2021-11-28 01:57:21 +00:00
HourCycle hour_cycle_from_string(StringView hour_cycle)
{
if (hour_cycle == "h11"sv)
Userland: Move files destined for LibLocale to the Locale namespace
2022-09-02 16:01:10 +00:00
return HourCycle::H11;
LibLocale: Remove "else" after "if" statements that always return Mostly just to keep clangd quiet.
2023-01-27 15:27:33 +00:00
if (hour_cycle == "h12"sv)
Userland: Move files destined for LibLocale to the Locale namespace
2022-09-02 16:01:10 +00:00
return HourCycle::H12;
LibLocale: Remove "else" after "if" statements that always return Mostly just to keep clangd quiet.
2023-01-27 15:27:33 +00:00
if (hour_cycle == "h23"sv)
Userland: Move files destined for LibLocale to the Locale namespace
2022-09-02 16:01:10 +00:00
return HourCycle::H23;
LibLocale: Remove "else" after "if" statements that always return Mostly just to keep clangd quiet.
2023-01-27 15:27:33 +00:00
if (hour_cycle == "h24"sv)
Userland: Move files destined for LibLocale to the Locale namespace
2022-09-02 16:01:10 +00:00
return HourCycle::H24;
LibUnicode: Parse and generate regional hour cycles Unlike most data in the CLDR, hour cycles are not stored on a per-locale basis. Instead, they are keyed by a string that is usually a region, but sometimes is a locale. Therefore, given a locale, to determine the hour cycles for that locale, we: 1. Check if the locale itself is assigned hour cycles. 2. If the locale has a region, check if that region is assigned hour cycles. 3. Otherwise, maximize that locale, and if the maximized locale has a region, check if that region is assigned hour cycles. 4. If the above all fail, fallback to the "001" region. Further, each locale's default hour cycle is the first assigned hour cycle.
2021-11-28 01:57:21 +00:00
VERIFY_NOT_REACHED();
}
StringView hour_cycle_to_string(HourCycle hour_cycle)
{
switch (hour_cycle) {
case HourCycle::H11:
return "h11"sv;
case HourCycle::H12:
return "h12"sv;
case HourCycle::H23:
return "h23"sv;
case HourCycle::H24:
return "h24"sv;
}
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
VERIFY_NOT_REACHED();
}
LibJS+LibLocale: Replace preferred hour cycle lookups with ICU
2024-06-12 20:16:49 +00:00
Optional<HourCycle> default_hour_cycle(StringView locale)
{
UErrorCode status = U_ZERO_ERROR;
auto locale_data = LocaleData::for_locale(locale);
if (!locale_data.has_value())
return {};
auto hour_cycle = locale_data->date_time_pattern_generator().getDefaultHourCycle(status);
if (icu_failure(status))
return {};
switch (hour_cycle) {
case UDAT_HOUR_CYCLE_11:
return HourCycle::H11;
case UDAT_HOUR_CYCLE_12:
return HourCycle::H12;
case UDAT_HOUR_CYCLE_23:
return HourCycle::H23;
case UDAT_HOUR_CYCLE_24:
return HourCycle::H24;
}
VERIFY_NOT_REACHED();
}
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
static constexpr char icu_hour_cycle(Optional<HourCycle> const& hour_cycle, Optional<bool> const& hour12)
{
if (hour12.has_value())
return *hour12 ? 'h' : 'H';
if (!hour_cycle.has_value())
return 'j';
switch (*hour_cycle) {
case HourCycle::H11:
return 'K';
case HourCycle::H12:
return 'h';
case HourCycle::H23:
return 'H';
case HourCycle::H24:
return 'k';
}
VERIFY_NOT_REACHED();
LibUnicode: Parse and generate regional hour cycles Unlike most data in the CLDR, hour cycles are not stored on a per-locale basis. Instead, they are keyed by a string that is usually a region, but sometimes is a locale. Therefore, given a locale, to determine the hour cycles for that locale, we: 1. Check if the locale itself is assigned hour cycles. 2. If the locale has a region, check if that region is assigned hour cycles. 3. Otherwise, maximize that locale, and if the maximized locale has a region, check if that region is assigned hour cycles. 4. If the above all fail, fallback to the "001" region. Further, each locale's default hour cycle is the first assigned hour cycle.
2021-11-28 01:57:21 +00:00
}
LibUnicode: Parse and generate date, time, and date-time format patterns
2021-11-27 19:54:48 +00:00
CalendarPatternStyle calendar_pattern_style_from_string(StringView style)
{
if (style == "narrow"sv)
return CalendarPatternStyle::Narrow;
if (style == "short"sv)
return CalendarPatternStyle::Short;
if (style == "long"sv)
return CalendarPatternStyle::Long;
if (style == "numeric"sv)
return CalendarPatternStyle::Numeric;
if (style == "2-digit"sv)
return CalendarPatternStyle::TwoDigit;
LibUnicode: Add minimal support for generic & offset-based time zones ECMA-402 now supports short-offset, long-offset, short-generic, and long-generic time zone name formatting. For example, in the en-US locale the America/Eastern time zone would be formatted as: short-offset: GMT-5 long-offset: GMT-05:00 short-generic: ET long-generic: Eastern Time We currently only support the UTC time zone, however. Therefore, this very minimal implementation does not consider GMT offset or generic display names. Instead, the CLDR defines specific strings for UTC.
2022-01-02 19:23:24 +00:00
if (style == "shortOffset"sv)
return CalendarPatternStyle::ShortOffset;
if (style == "longOffset"sv)
return CalendarPatternStyle::LongOffset;
if (style == "shortGeneric"sv)
return CalendarPatternStyle::ShortGeneric;
if (style == "longGeneric"sv)
return CalendarPatternStyle::LongGeneric;
LibUnicode: Parse and generate date, time, and date-time format patterns
2021-11-27 19:54:48 +00:00
VERIFY_NOT_REACHED();
}
StringView calendar_pattern_style_to_string(CalendarPatternStyle style)
{
switch (style) {
case CalendarPatternStyle::Narrow:
return "narrow"sv;
case CalendarPatternStyle::Short:
return "short"sv;
case CalendarPatternStyle::Long:
return "long"sv;
case CalendarPatternStyle::Numeric:
LibUnicode: Fix copy-paste error in calendar_pattern_style_to_string The string returned must be lowercase.
2021-11-30 15:39:56 +00:00
return "numeric"sv;
LibUnicode: Parse and generate date, time, and date-time format patterns
2021-11-27 19:54:48 +00:00
case CalendarPatternStyle::TwoDigit:
return "2-digit"sv;
LibUnicode: Add minimal support for generic & offset-based time zones ECMA-402 now supports short-offset, long-offset, short-generic, and long-generic time zone name formatting. For example, in the en-US locale the America/Eastern time zone would be formatted as: short-offset: GMT-5 long-offset: GMT-05:00 short-generic: ET long-generic: Eastern Time We currently only support the UTC time zone, however. Therefore, this very minimal implementation does not consider GMT offset or generic display names. Instead, the CLDR defines specific strings for UTC.
2022-01-02 19:23:24 +00:00
case CalendarPatternStyle::ShortOffset:
return "shortOffset"sv;
case CalendarPatternStyle::LongOffset:
return "longOffset"sv;
case CalendarPatternStyle::ShortGeneric:
return "shortGeneric"sv;
case CalendarPatternStyle::LongGeneric:
return "longGeneric"sv;
LibUnicode: Parse and generate date, time, and date-time format patterns
2021-11-27 19:54:48 +00:00
}
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
VERIFY_NOT_REACHED();
}
// https://unicode.org/reports/tr35/tr35-dates.html#Date_Field_Symbol_Table
String CalendarPattern::to_pattern() const
{
// What we refer to as Narrow, Short, and Long, TR-35 refers to as Narrow, Abbreviated, and Wide.
StringBuilder builder;
if (era.has_value()) {
switch (*era) {
case CalendarPatternStyle::Narrow:
builder.append("GGGGG"sv);
break;
case CalendarPatternStyle::Short:
builder.append("G"sv);
break;
case CalendarPatternStyle::Long:
builder.append("GGGG"sv);
break;
default:
break;
}
}
if (year.has_value()) {
switch (*year) {
case CalendarPatternStyle::Numeric:
builder.append("y"sv);
break;
case CalendarPatternStyle::TwoDigit:
builder.append("yy"sv);
break;
default:
break;
}
}
if (month.has_value()) {
switch (*month) {
case CalendarPatternStyle::Numeric:
builder.append("M"sv);
break;
case CalendarPatternStyle::TwoDigit:
builder.append("MM"sv);
break;
case CalendarPatternStyle::Narrow:
builder.append("MMMMM"sv);
break;
case CalendarPatternStyle::Short:
builder.append("MMM"sv);
break;
case CalendarPatternStyle::Long:
builder.append("MMMM"sv);
break;
default:
break;
}
}
if (weekday.has_value()) {
switch (*weekday) {
case CalendarPatternStyle::Narrow:
builder.append("EEEEE"sv);
break;
case CalendarPatternStyle::Short:
builder.append("E"sv);
break;
case CalendarPatternStyle::Long:
builder.append("EEEE"sv);
break;
default:
break;
}
}
if (day.has_value()) {
switch (*day) {
case CalendarPatternStyle::Numeric:
builder.append("d"sv);
break;
case CalendarPatternStyle::TwoDigit:
builder.append("dd"sv);
break;
default:
break;
}
}
if (day_period.has_value()) {
switch (*day_period) {
case CalendarPatternStyle::Narrow:
builder.append("BBBBB"sv);
break;
case CalendarPatternStyle::Short:
builder.append("B"sv);
break;
case CalendarPatternStyle::Long:
builder.append("BBBB"sv);
break;
default:
break;
}
}
if (hour.has_value()) {
auto hour_cycle_symbol = icu_hour_cycle(hour_cycle, hour12);
switch (*hour) {
case CalendarPatternStyle::Numeric:
builder.append(hour_cycle_symbol);
break;
case CalendarPatternStyle::TwoDigit:
builder.append_repeated(hour_cycle_symbol, 2);
break;
default:
break;
}
}
if (minute.has_value()) {
switch (*minute) {
case CalendarPatternStyle::Numeric:
builder.append("m"sv);
break;
case CalendarPatternStyle::TwoDigit:
builder.append("mm"sv);
break;
default:
break;
}
}
if (second.has_value()) {
switch (*second) {
case CalendarPatternStyle::Numeric:
builder.append("s"sv);
break;
case CalendarPatternStyle::TwoDigit:
builder.append("ss"sv);
break;
default:
break;
}
}
if (fractional_second_digits.has_value()) {
for (u8 i = 0; i < *fractional_second_digits; ++i)
builder.append("S"sv);
}
if (time_zone_name.has_value()) {
switch (*time_zone_name) {
case CalendarPatternStyle::Short:
builder.append("z"sv);
break;
case CalendarPatternStyle::Long:
builder.append("zzzz"sv);
break;
case CalendarPatternStyle::ShortOffset:
builder.append("O"sv);
break;
case CalendarPatternStyle::LongOffset:
builder.append("OOOO"sv);
break;
case CalendarPatternStyle::ShortGeneric:
builder.append("v"sv);
break;
case CalendarPatternStyle::LongGeneric:
builder.append("vvvv"sv);
break;
default:
break;
}
}
return MUST(builder.to_string());
}
// https://unicode.org/reports/tr35/tr35-dates.html#Date_Field_Symbol_Table
CalendarPattern CalendarPattern::create_from_pattern(StringView pattern)
{
GenericLexer lexer { pattern };
CalendarPattern format {};
while (!lexer.is_eof()) {
if (lexer.next_is(is_quote)) {
lexer.consume_quoted_string();
continue;
}
auto starting_char = lexer.peek();
auto segment = lexer.consume_while([&](char ch) { return ch == starting_char; });
// Era
if (all_of(segment, is_any_of("G"sv))) {
if (segment.length() <= 3)
format.era = CalendarPatternStyle::Short;
else if (segment.length() == 4)
format.era = CalendarPatternStyle::Long;
else
format.era = CalendarPatternStyle::Narrow;
}
// Year
else if (all_of(segment, is_any_of("yYuUr"sv))) {
if (segment.length() == 2)
format.year = CalendarPatternStyle::TwoDigit;
else
format.year = CalendarPatternStyle::Numeric;
}
// Month
else if (all_of(segment, is_any_of("ML"sv))) {
if (segment.length() == 1)
format.month = CalendarPatternStyle::Numeric;
else if (segment.length() == 2)
format.month = CalendarPatternStyle::TwoDigit;
else if (segment.length() == 3)
format.month = CalendarPatternStyle::Short;
else if (segment.length() == 4)
format.month = CalendarPatternStyle::Long;
else if (segment.length() == 5)
format.month = CalendarPatternStyle::Narrow;
}
// Weekday
else if (all_of(segment, is_any_of("ecE"sv))) {
if (segment.length() == 4)
format.weekday = CalendarPatternStyle::Long;
else if (segment.length() == 5)
format.weekday = CalendarPatternStyle::Narrow;
else
format.weekday = CalendarPatternStyle::Short;
}
// Day
else if (all_of(segment, is_any_of("d"sv))) {
if (segment.length() == 1)
format.day = CalendarPatternStyle::Numeric;
else
format.day = CalendarPatternStyle::TwoDigit;
} else if (all_of(segment, is_any_of("DFg"sv))) {
format.day = CalendarPatternStyle::Numeric;
}
// Day period
else if (all_of(segment, is_any_of("B"sv))) {
if (segment.length() == 4)
format.day_period = CalendarPatternStyle::Long;
else if (segment.length() == 5)
format.day_period = CalendarPatternStyle::Narrow;
else
format.day_period = CalendarPatternStyle::Short;
}
// Hour
else if (all_of(segment, is_any_of("hHKk"sv))) {
switch (starting_char) {
case 'K':
format.hour_cycle = HourCycle::H11;
break;
case 'h':
format.hour_cycle = HourCycle::H12;
break;
case 'H':
format.hour_cycle = HourCycle::H23;
break;
case 'k':
format.hour_cycle = HourCycle::H24;
break;
}
if (segment.length() == 1)
format.hour = CalendarPatternStyle::Numeric;
else
format.hour = CalendarPatternStyle::TwoDigit;
}
// Minute
else if (all_of(segment, is_any_of("m"sv))) {
if (segment.length() == 1)
format.minute = CalendarPatternStyle::Numeric;
else
format.minute = CalendarPatternStyle::TwoDigit;
}
// Second
else if (all_of(segment, is_any_of("s"sv))) {
if (segment.length() == 1)
format.second = CalendarPatternStyle::Numeric;
else
format.second = CalendarPatternStyle::TwoDigit;
} else if (all_of(segment, is_any_of("S"sv))) {
format.fractional_second_digits = static_cast<u8>(segment.length());
}
// Zone
else if (all_of(segment, is_any_of("zV"sv))) {
if (segment.length() < 4)
format.time_zone_name = CalendarPatternStyle::Short;
else
format.time_zone_name = CalendarPatternStyle::Long;
} else if (all_of(segment, is_any_of("ZOXx"sv))) {
if (segment.length() < 4)
format.time_zone_name = CalendarPatternStyle::ShortOffset;
else
format.time_zone_name = CalendarPatternStyle::LongOffset;
} else if (all_of(segment, is_any_of("v"sv))) {
if (segment.length() < 4)
format.time_zone_name = CalendarPatternStyle::ShortGeneric;
else
format.time_zone_name = CalendarPatternStyle::LongGeneric;
}
}
return format;
LibUnicode: Parse and generate date, time, and date-time format patterns
2021-11-27 19:54:48 +00:00
}
LibLocale+LibJS+ClockSettings: Make date time format APIs infallible These APIs only perform small allocations, and are only used by LibJS and the time zone settings widget. Callers which could only have failed from these APIs are also made to be infallible here.
2023-08-22 20:41:36 +00:00
template<typename T, typename GetRegionalValues>
static T find_regional_values_for_locale(StringView locale, GetRegionalValues&& get_regional_values)
LibUnicode: Parse and generate regional hour cycles Unlike most data in the CLDR, hour cycles are not stored on a per-locale basis. Instead, they are keyed by a string that is usually a region, but sometimes is a locale. Therefore, given a locale, to determine the hour cycles for that locale, we: 1. Check if the locale itself is assigned hour cycles. 2. If the locale has a region, check if that region is assigned hour cycles. 3. Otherwise, maximize that locale, and if the maximized locale has a region, check if that region is assigned hour cycles. 4. If the above all fail, fallback to the "001" region. Further, each locale's default hour cycle is the first assigned hour cycle.
2021-11-28 01:57:21 +00:00
{
LibUnicode: Generate per-region week data This includes: * The minimum number of days in a week for that week to count as the first week of a new year. * The day to be shown as the first day of the week in a calendar. * The start/end days of the weekend. Like the existing hour cycle data, week data is presented per-region in the CLDR, rather than per-locale. The method to add likely subtags to a locale to perform region lookups is the same. The list of regions in the CLDR for hour cycle, minimum days, first day, and weekend days are quite different. So rather than changing the existing HourCycleRegion enum to a generic Region enum, we generate separate enums for each of the week data fields. This allows each lookup into these fields to remain simple array-based index access, without any "jumps" for regions that don't have CLDR data for a field.
2022-07-06 12:21:32 +00:00
auto has_value = [](auto const& container) {
if constexpr (requires { container.has_value(); })
return container.has_value();
else
return !container.is_empty();
};
LibLocale+LibJS+ClockSettings: Make date time format APIs infallible These APIs only perform small allocations, and are only used by LibJS and the time zone settings widget. Callers which could only have failed from these APIs are also made to be infallible here.
2023-08-22 20:41:36 +00:00
if (auto regional_values = get_regional_values(locale); has_value(regional_values))
LibUnicode: Generate per-region week data This includes: * The minimum number of days in a week for that week to count as the first week of a new year. * The day to be shown as the first day of the week in a calendar. * The start/end days of the weekend. Like the existing hour cycle data, week data is presented per-region in the CLDR, rather than per-locale. The method to add likely subtags to a locale to perform region lookups is the same. The list of regions in the CLDR for hour cycle, minimum days, first day, and weekend days are quite different. So rather than changing the existing HourCycleRegion enum to a generic Region enum, we generate separate enums for each of the week data fields. This allows each lookup into these fields to remain simple array-based index access, without any "jumps" for regions that don't have CLDR data for a field.
2022-07-06 12:21:32 +00:00
return regional_values;
LibUnicode: Dynamically load the generated UnicodeDateTimeFormat symbols
2021-12-16 04:19:17 +00:00
LibUnicode: Generate per-region week data This includes: * The minimum number of days in a week for that week to count as the first week of a new year. * The day to be shown as the first day of the week in a calendar. * The start/end days of the weekend. Like the existing hour cycle data, week data is presented per-region in the CLDR, rather than per-locale. The method to add likely subtags to a locale to perform region lookups is the same. The list of regions in the CLDR for hour cycle, minimum days, first day, and weekend days are quite different. So rather than changing the existing HourCycleRegion enum to a generic Region enum, we generate separate enums for each of the week data fields. This allows each lookup into these fields to remain simple array-based index access, without any "jumps" for regions that don't have CLDR data for a field.
2022-07-06 12:21:32 +00:00
auto return_default_values = [&]() { return get_regional_values("001"sv); };
LibUnicode: Parse and generate regional hour cycles Unlike most data in the CLDR, hour cycles are not stored on a per-locale basis. Instead, they are keyed by a string that is usually a region, but sometimes is a locale. Therefore, given a locale, to determine the hour cycles for that locale, we: 1. Check if the locale itself is assigned hour cycles. 2. If the locale has a region, check if that region is assigned hour cycles. 3. Otherwise, maximize that locale, and if the maximized locale has a region, check if that region is assigned hour cycles. 4. If the above all fail, fallback to the "001" region. Further, each locale's default hour cycle is the first assigned hour cycle.
2021-11-28 01:57:21 +00:00
LibLocale+LibJS: Make locale data APIs infallible These APIs only perform small allocations, and are only used by LibJS. Callers which could only have failed from these APIs are also made to be infallible here.
2023-08-22 19:39:18 +00:00
auto language = parse_unicode_language_id(locale);
LibUnicode: Parse and generate regional hour cycles Unlike most data in the CLDR, hour cycles are not stored on a per-locale basis. Instead, they are keyed by a string that is usually a region, but sometimes is a locale. Therefore, given a locale, to determine the hour cycles for that locale, we: 1. Check if the locale itself is assigned hour cycles. 2. If the locale has a region, check if that region is assigned hour cycles. 3. Otherwise, maximize that locale, and if the maximized locale has a region, check if that region is assigned hour cycles. 4. If the above all fail, fallback to the "001" region. Further, each locale's default hour cycle is the first assigned hour cycle.
2021-11-28 01:57:21 +00:00
if (!language.has_value())
LibUnicode: Generate per-region week data This includes: * The minimum number of days in a week for that week to count as the first week of a new year. * The day to be shown as the first day of the week in a calendar. * The start/end days of the weekend. Like the existing hour cycle data, week data is presented per-region in the CLDR, rather than per-locale. The method to add likely subtags to a locale to perform region lookups is the same. The list of regions in the CLDR for hour cycle, minimum days, first day, and weekend days are quite different. So rather than changing the existing HourCycleRegion enum to a generic Region enum, we generate separate enums for each of the week data fields. This allows each lookup into these fields to remain simple array-based index access, without any "jumps" for regions that don't have CLDR data for a field.
2022-07-06 12:21:32 +00:00
return return_default_values();
LibUnicode: Parse and generate regional hour cycles Unlike most data in the CLDR, hour cycles are not stored on a per-locale basis. Instead, they are keyed by a string that is usually a region, but sometimes is a locale. Therefore, given a locale, to determine the hour cycles for that locale, we: 1. Check if the locale itself is assigned hour cycles. 2. If the locale has a region, check if that region is assigned hour cycles. 3. Otherwise, maximize that locale, and if the maximized locale has a region, check if that region is assigned hour cycles. 4. If the above all fail, fallback to the "001" region. Further, each locale's default hour cycle is the first assigned hour cycle.
2021-11-28 01:57:21 +00:00
LibJS+LibLocale: Replace locale maximization and minimization with ICU
2024-06-08 17:23:26 +00:00
if (!language->region.has_value()) {
if (auto maximized = add_likely_subtags(language->to_string()); maximized.has_value())
language = parse_unicode_language_id(*maximized);
}
LibUnicode: Parse and generate regional hour cycles Unlike most data in the CLDR, hour cycles are not stored on a per-locale basis. Instead, they are keyed by a string that is usually a region, but sometimes is a locale. Therefore, given a locale, to determine the hour cycles for that locale, we: 1. Check if the locale itself is assigned hour cycles. 2. If the locale has a region, check if that region is assigned hour cycles. 3. Otherwise, maximize that locale, and if the maximized locale has a region, check if that region is assigned hour cycles. 4. If the above all fail, fallback to the "001" region. Further, each locale's default hour cycle is the first assigned hour cycle.
2021-11-28 01:57:21 +00:00
if (!language.has_value() || !language->region.has_value())
LibUnicode: Generate per-region week data This includes: * The minimum number of days in a week for that week to count as the first week of a new year. * The day to be shown as the first day of the week in a calendar. * The start/end days of the weekend. Like the existing hour cycle data, week data is presented per-region in the CLDR, rather than per-locale. The method to add likely subtags to a locale to perform region lookups is the same. The list of regions in the CLDR for hour cycle, minimum days, first day, and weekend days are quite different. So rather than changing the existing HourCycleRegion enum to a generic Region enum, we generate separate enums for each of the week data fields. This allows each lookup into these fields to remain simple array-based index access, without any "jumps" for regions that don't have CLDR data for a field.
2022-07-06 12:21:32 +00:00
return return_default_values();
LibUnicode: Parse and generate regional hour cycles Unlike most data in the CLDR, hour cycles are not stored on a per-locale basis. Instead, they are keyed by a string that is usually a region, but sometimes is a locale. Therefore, given a locale, to determine the hour cycles for that locale, we: 1. Check if the locale itself is assigned hour cycles. 2. If the locale has a region, check if that region is assigned hour cycles. 3. Otherwise, maximize that locale, and if the maximized locale has a region, check if that region is assigned hour cycles. 4. If the above all fail, fallback to the "001" region. Further, each locale's default hour cycle is the first assigned hour cycle.
2021-11-28 01:57:21 +00:00
LibLocale+LibJS+ClockSettings: Make date time format APIs infallible These APIs only perform small allocations, and are only used by LibJS and the time zone settings widget. Callers which could only have failed from these APIs are also made to be infallible here.
2023-08-22 20:41:36 +00:00
if (auto regional_values = get_regional_values(*language->region); has_value(regional_values))
LibUnicode: Generate per-region week data This includes: * The minimum number of days in a week for that week to count as the first week of a new year. * The day to be shown as the first day of the week in a calendar. * The start/end days of the weekend. Like the existing hour cycle data, week data is presented per-region in the CLDR, rather than per-locale. The method to add likely subtags to a locale to perform region lookups is the same. The list of regions in the CLDR for hour cycle, minimum days, first day, and weekend days are quite different. So rather than changing the existing HourCycleRegion enum to a generic Region enum, we generate separate enums for each of the week data fields. This allows each lookup into these fields to remain simple array-based index access, without any "jumps" for regions that don't have CLDR data for a field.
2022-07-06 12:21:32 +00:00
return regional_values;
LibUnicode: Parse and generate regional hour cycles Unlike most data in the CLDR, hour cycles are not stored on a per-locale basis. Instead, they are keyed by a string that is usually a region, but sometimes is a locale. Therefore, given a locale, to determine the hour cycles for that locale, we: 1. Check if the locale itself is assigned hour cycles. 2. If the locale has a region, check if that region is assigned hour cycles. 3. Otherwise, maximize that locale, and if the maximized locale has a region, check if that region is assigned hour cycles. 4. If the above all fail, fallback to the "001" region. Further, each locale's default hour cycle is the first assigned hour cycle.
2021-11-28 01:57:21 +00:00
LibUnicode: Generate per-region week data This includes: * The minimum number of days in a week for that week to count as the first week of a new year. * The day to be shown as the first day of the week in a calendar. * The start/end days of the weekend. Like the existing hour cycle data, week data is presented per-region in the CLDR, rather than per-locale. The method to add likely subtags to a locale to perform region lookups is the same. The list of regions in the CLDR for hour cycle, minimum days, first day, and weekend days are quite different. So rather than changing the existing HourCycleRegion enum to a generic Region enum, we generate separate enums for each of the week data fields. This allows each lookup into these fields to remain simple array-based index access, without any "jumps" for regions that don't have CLDR data for a field.
2022-07-06 12:21:32 +00:00
return return_default_values();
}
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
// ICU does not contain a field enumeration for "literal" partitions. Define a custom field so that we may provide a
// type for those partitions.
static constexpr i32 LITERAL_FIELD = -1;
static constexpr StringView icu_date_time_format_field_to_string(i32 field)
LibUnicode: Add method to combine two format pattern skeletons The fields of the generated elements must be in the same order as the table here: https://unicode.org/reports/tr35/tr35-dates.html#Date_Field_Symbol_Table Further, only one field from each group of fields is allowed.
2021-12-09 00:52:48 +00:00
{
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
switch (field) {
case LITERAL_FIELD:
return "literal"sv;
case UDAT_ERA_FIELD:
return "era"sv;
case UDAT_YEAR_FIELD:
case UDAT_EXTENDED_YEAR_FIELD:
return "year"sv;
case UDAT_YEAR_NAME_FIELD:
return "yearName"sv;
case UDAT_RELATED_YEAR_FIELD:
return "relatedYear"sv;
case UDAT_MONTH_FIELD:
case UDAT_STANDALONE_MONTH_FIELD:
return "month"sv;
case UDAT_DAY_OF_WEEK_FIELD:
case UDAT_DOW_LOCAL_FIELD:
case UDAT_STANDALONE_DAY_FIELD:
return "weekday"sv;
case UDAT_DATE_FIELD:
return "day"sv;
case UDAT_AM_PM_FIELD:
case UDAT_AM_PM_MIDNIGHT_NOON_FIELD:
case UDAT_FLEXIBLE_DAY_PERIOD_FIELD:
return "dayPeriod"sv;
case UDAT_HOUR_OF_DAY1_FIELD:
case UDAT_HOUR_OF_DAY0_FIELD:
case UDAT_HOUR1_FIELD:
case UDAT_HOUR0_FIELD:
return "hour"sv;
case UDAT_MINUTE_FIELD:
return "minute"sv;
case UDAT_SECOND_FIELD:
return "second"sv;
case UDAT_FRACTIONAL_SECOND_FIELD:
return "fractionalSecond"sv;
case UDAT_TIMEZONE_FIELD:
case UDAT_TIMEZONE_RFC_FIELD:
case UDAT_TIMEZONE_GENERIC_FIELD:
case UDAT_TIMEZONE_SPECIAL_FIELD:
case UDAT_TIMEZONE_LOCALIZED_GMT_OFFSET_FIELD:
case UDAT_TIMEZONE_ISO_FIELD:
case UDAT_TIMEZONE_ISO_LOCAL_FIELD:
return "timeZoneName"sv;
default:
return "unknown"sv;
}
}
LibUnicode: Add method to combine two format pattern skeletons The fields of the generated elements must be in the same order as the table here: https://unicode.org/reports/tr35/tr35-dates.html#Date_Field_Symbol_Table Further, only one field from each group of fields is allowed.
2021-12-09 00:52:48 +00:00
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
static bool apply_hour_cycle_to_skeleton(icu::UnicodeString& skeleton, Optional<HourCycle> const& hour_cycle, Optional<bool> const& hour12)
{
auto hour_cycle_symbol = icu_hour_cycle(hour_cycle, hour12);
if (hour_cycle_symbol == 'j')
return false;
bool changed_hour_cycle = false;
bool inside_quote = false;
for (i32 i = 0; i < skeleton.length(); ++i) {
switch (skeleton[i]) {
case '\'':
inside_quote = !inside_quote;
break;
case 'h':
case 'H':
case 'k':
case 'K':
if (!inside_quote && static_cast<char>(skeleton[i]) != hour_cycle_symbol) {
skeleton.setCharAt(i, hour_cycle_symbol);
changed_hour_cycle = true;
}
break;
default:
break;
}
}
return changed_hour_cycle;
}
LibUnicode: Add method to combine two format pattern skeletons The fields of the generated elements must be in the same order as the table here: https://unicode.org/reports/tr35/tr35-dates.html#Date_Field_Symbol_Table Further, only one field from each group of fields is allowed.
2021-12-09 00:52:48 +00:00
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
static void apply_time_zone_to_formatter(icu::SimpleDateFormat& formatter, icu::Locale const& locale, StringView time_zone_identifier)
{
UErrorCode status = U_ZERO_ERROR;
LibUnicode: Add method to combine two format pattern skeletons The fields of the generated elements must be in the same order as the table here: https://unicode.org/reports/tr35/tr35-dates.html#Date_Field_Symbol_Table Further, only one field from each group of fields is allowed.
2021-12-09 00:52:48 +00:00
LibUnicode: Cache all created icu::TimeZone objects This cache works exactly the same as the existing icu::Locale cache.
2024-09-03 15:20:20 +00:00
auto time_zone_data = TimeZoneData::for_time_zone(time_zone_identifier);
LibUnicode: Add method to combine two format pattern skeletons The fields of the generated elements must be in the same order as the table here: https://unicode.org/reports/tr35/tr35-dates.html#Date_Field_Symbol_Table Further, only one field from each group of fields is allowed.
2021-12-09 00:52:48 +00:00
LibUnicode: Cache all created icu::TimeZone objects This cache works exactly the same as the existing icu::Locale cache.
2024-09-03 15:20:20 +00:00
auto* calendar = icu::Calendar::createInstance(time_zone_data->time_zone(), locale, status);
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
VERIFY(icu_success(status));
LibUnicode: Add method to combine two format pattern skeletons The fields of the generated elements must be in the same order as the table here: https://unicode.org/reports/tr35/tr35-dates.html#Date_Field_Symbol_Table Further, only one field from each group of fields is allowed.
2021-12-09 00:52:48 +00:00
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
if (calendar->getDynamicClassID() == icu::GregorianCalendar::getStaticClassID()) {
// https://tc39.es/ecma262/#sec-time-values-and-time-range
// A time value supports a slightly smaller range of -8,640,000,000,000,000 to 8,640,000,000,000,000 milliseconds.
static constexpr double ECMA_262_MINIMUM_TIME = -8.64E15;
auto* gregorian_calendar = static_cast<icu::GregorianCalendar*>(calendar);
gregorian_calendar->setGregorianChange(ECMA_262_MINIMUM_TIME, status);
VERIFY(icu_success(status));
LibUnicode: Add method to combine two format pattern skeletons The fields of the generated elements must be in the same order as the table here: https://unicode.org/reports/tr35/tr35-dates.html#Date_Field_Symbol_Table Further, only one field from each group of fields is allowed.
2021-12-09 00:52:48 +00:00
}
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
formatter.adoptCalendar(calendar);
LibUnicode: Add method to combine two format pattern skeletons The fields of the generated elements must be in the same order as the table here: https://unicode.org/reports/tr35/tr35-dates.html#Date_Field_Symbol_Table Further, only one field from each group of fields is allowed.
2021-12-09 00:52:48 +00:00
}
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
static bool is_formatted_range_actually_a_range(icu::FormattedDateInterval const& formatted)
LibUnicode: Parse and generate date, time, and date-time format patterns
2021-11-27 19:54:48 +00:00
{
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
UErrorCode status = U_ZERO_ERROR;
auto result = formatted.toTempString(status);
if (icu_failure(status))
return false;
icu::ConstrainedFieldPosition position;
position.constrainCategory(UFIELD_CATEGORY_DATE_INTERVAL_SPAN);
auto has_range = static_cast<bool>(formatted.nextPosition(position, status));
if (icu_failure(status))
return false;
return has_range;
LibUnicode: Parse and generate date, time, and date-time format patterns
2021-11-27 19:54:48 +00:00
}
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
class DateTimeFormatImpl : public DateTimeFormat {
public:
DateTimeFormatImpl(icu::Locale& locale, icu::UnicodeString const& pattern, StringView time_zone_identifier, NonnullOwnPtr<icu::SimpleDateFormat> formatter)
: m_locale(locale)
, m_pattern(CalendarPattern::create_from_pattern(icu_string_to_string(pattern)))
, m_formatter(move(formatter))
{
apply_time_zone_to_formatter(*m_formatter, m_locale, time_zone_identifier);
}
LibUnicode: Implement TR-35's localized GMT offset formatting This adds an API to use LibTimeZone to convert a time zone such as "America/New_York" to a GMT offset string like "GMT-5" (short form) or "GMT-05:00" (long form).
2022-01-11 16:35:50 +00:00
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
virtual ~DateTimeFormatImpl() override = default;
LibUnicode: Implement TR-35's localized GMT offset formatting This adds an API to use LibTimeZone to convert a time zone such as "America/New_York" to a GMT offset string like "GMT-5" (short form) or "GMT-05:00" (long form).
2022-01-11 16:35:50 +00:00
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
virtual CalendarPattern const& chosen_pattern() const override { return m_pattern; }
LibUnicode: Implement TR-35's localized GMT offset formatting This adds an API to use LibTimeZone to convert a time zone such as "America/New_York" to a GMT offset string like "GMT-5" (short form) or "GMT-05:00" (long form).
2022-01-11 16:35:50 +00:00
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
virtual String format(double time) const override
{
auto formatted_time = format_impl(time);
if (!formatted_time.has_value())
return {};
LibUnicode: Implement TR-35's localized GMT offset formatting This adds an API to use LibTimeZone to convert a time zone such as "America/New_York" to a GMT offset string like "GMT-5" (short form) or "GMT-05:00" (long form).
2022-01-11 16:35:50 +00:00
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
return icu_string_to_string(*formatted_time);
}
LibUnicode: Implement TR-35's localized GMT offset formatting This adds an API to use LibTimeZone to convert a time zone such as "America/New_York" to a GMT offset string like "GMT-5" (short form) or "GMT-05:00" (long form).
2022-01-11 16:35:50 +00:00
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
virtual Vector<Partition> format_to_parts(double time) const override
{
icu::FieldPositionIterator iterator;
LibUnicode: Implement TR-35's localized GMT offset formatting This adds an API to use LibTimeZone to convert a time zone such as "America/New_York" to a GMT offset string like "GMT-5" (short form) or "GMT-05:00" (long form).
2022-01-11 16:35:50 +00:00
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
auto formatted_time = format_impl(time, &iterator);
if (!formatted_time.has_value())
return {};
LibUnicode: Implement TR-35's localized GMT offset formatting This adds an API to use LibTimeZone to convert a time zone such as "America/New_York" to a GMT offset string like "GMT-5" (short form) or "GMT-05:00" (long form).
2022-01-11 16:35:50 +00:00
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
Vector<Partition> result;
LibUnicode: Implement TR-35's localized GMT offset formatting This adds an API to use LibTimeZone to convert a time zone such as "America/New_York" to a GMT offset string like "GMT-5" (short form) or "GMT-05:00" (long form).
2022-01-11 16:35:50 +00:00
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
auto create_partition = [&](i32 field, i32 begin, i32 end) {
Partition partition;
partition.type = icu_date_time_format_field_to_string(field);
partition.value = icu_string_to_string(formatted_time->tempSubStringBetween(begin, end));
partition.source = "shared"sv;
result.append(move(partition));
};
LibUnicode: Implement TR-35's localized GMT offset formatting This adds an API to use LibTimeZone to convert a time zone such as "America/New_York" to a GMT offset string like "GMT-5" (short form) or "GMT-05:00" (long form).
2022-01-11 16:35:50 +00:00
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
icu::FieldPosition position;
i32 previous_end_index = 0;
while (static_cast<bool>(iterator.next(position))) {
if (previous_end_index < position.getBeginIndex())
create_partition(LITERAL_FIELD, previous_end_index, position.getBeginIndex());
if (position.getField() >= 0)
create_partition(position.getField(), position.getBeginIndex(), position.getEndIndex());
previous_end_index = position.getEndIndex();
LibUnicode: Implement TR-35's localized GMT offset formatting This adds an API to use LibTimeZone to convert a time zone such as "America/New_York" to a GMT offset string like "GMT-5" (short form) or "GMT-05:00" (long form).
2022-01-11 16:35:50 +00:00
}
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
if (previous_end_index < formatted_time->length())
create_partition(LITERAL_FIELD, previous_end_index, formatted_time->length());
return result;
LibUnicode: Implement TR-35's localized GMT offset formatting This adds an API to use LibTimeZone to convert a time zone such as "America/New_York" to a GMT offset string like "GMT-5" (short form) or "GMT-05:00" (long form).
2022-01-11 16:35:50 +00:00
}
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
virtual String format_range(double start, double end) const override
{
UErrorCode status = U_ZERO_ERROR;
LibUnicode: Implement TR-35's localized GMT offset formatting This adds an API to use LibTimeZone to convert a time zone such as "America/New_York" to a GMT offset string like "GMT-5" (short form) or "GMT-05:00" (long form).
2022-01-11 16:35:50 +00:00
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
auto formatted = format_range_impl(start, end);
if (!formatted.has_value())
return {};
LibJS+LibUnicode: Return the appropriate time zone name depending on DST
2022-01-19 19:54:19 +00:00
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
if (!is_formatted_range_actually_a_range(*formatted))
return format(start);
LibUnicode: Implement TR-35's localized GMT offset formatting This adds an API to use LibTimeZone to convert a time zone such as "America/New_York" to a GMT offset string like "GMT-5" (short form) or "GMT-05:00" (long form).
2022-01-11 16:35:50 +00:00
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
auto formatted_time = formatted->toTempString(status);
if (icu_failure(status))
return {};
LibUnicode: Implement TR-35's localized GMT offset formatting This adds an API to use LibTimeZone to convert a time zone such as "America/New_York" to a GMT offset string like "GMT-5" (short form) or "GMT-05:00" (long form).
2022-01-11 16:35:50 +00:00
LibUnicode+LibJS: Normalize spaces in formatted date-time strings ICU 72 began using non-ASCII spaces in some formatted date-time strings. Every major browser has found that this introduced major breakage in web compatibility, as many sites and tools expect ASCII spaces. This patch removes these non-ASCII spaces in the same manner as the major engines. Such behavior is also tested by WPT.
2024-08-01 15:30:17 +00:00
normalize_spaces(formatted_time);
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
return icu_string_to_string(formatted_time);
LibUnicode: Implement TR-35's localized GMT offset formatting This adds an API to use LibTimeZone to convert a time zone such as "America/New_York" to a GMT offset string like "GMT-5" (short form) or "GMT-05:00" (long form).
2022-01-11 16:35:50 +00:00
}
LibUnicode: Fall back to GMT offset when a time zone name is unavailable The following table in TR-35 includes a web of fall back rules when the requested time zone style is unavailable: https://unicode.org/reports/tr35/tr35-dates.html#dfst-zone Conveniently, the subset of styles supported by ECMA-402 (and therefore LibUnicode) all either fall back to GMT offset or to a style that is unsupported but itself falls back to GMT offset.
2022-01-11 17:06:26 +00:00
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
virtual Vector<Partition> format_range_to_parts(double start, double end) const override
{
UErrorCode status = U_ZERO_ERROR;
LibUnicode: Fall back to GMT offset when a time zone name is unavailable The following table in TR-35 includes a web of fall back rules when the requested time zone style is unavailable: https://unicode.org/reports/tr35/tr35-dates.html#dfst-zone Conveniently, the subset of styles supported by ECMA-402 (and therefore LibUnicode) all either fall back to GMT offset or to a style that is unsupported but itself falls back to GMT offset.
2022-01-11 17:06:26 +00:00
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
auto formatted = format_range_impl(start, end);
if (!formatted.has_value())
return {};
LibUnicode: Fall back to GMT offset when a time zone name is unavailable The following table in TR-35 includes a web of fall back rules when the requested time zone style is unavailable: https://unicode.org/reports/tr35/tr35-dates.html#dfst-zone Conveniently, the subset of styles supported by ECMA-402 (and therefore LibUnicode) all either fall back to GMT offset or to a style that is unsupported but itself falls back to GMT offset.
2022-01-11 17:06:26 +00:00
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
if (!is_formatted_range_actually_a_range(*formatted))
return format_to_parts(start);
auto formatted_time = formatted->toTempString(status);
if (icu_failure(status))
return {};
LibUnicode+LibJS: Normalize spaces in formatted date-time strings ICU 72 began using non-ASCII spaces in some formatted date-time strings. Every major browser has found that this introduced major breakage in web compatibility, as many sites and tools expect ASCII spaces. This patch removes these non-ASCII spaces in the same manner as the major engines. Such behavior is also tested by WPT.
2024-08-01 15:30:17 +00:00
normalize_spaces(formatted_time);
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
icu::ConstrainedFieldPosition position;
i32 previous_end_index = 0;
Vector<Partition> result;
LibLocale: Move definition of Range structure to dedicated header We were redefining a Range type in a few formatters. Move it to its own header to avoid ODR violations (and give it a more descriptive name).
2024-06-19 16:46:45 +00:00
Optional<PartitionRange> start_range;
Optional<PartitionRange> end_range;
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
auto create_partition = [&](i32 field, i32 begin, i32 end) {
Partition partition;
partition.type = icu_date_time_format_field_to_string(field);
partition.value = icu_string_to_string(formatted_time.tempSubStringBetween(begin, end));
if (start_range.has_value() && start_range->contains(begin))
partition.source = "startRange"sv;
else if (end_range.has_value() && end_range->contains(begin))
partition.source = "endRange"sv;
else
partition.source = "shared"sv;
result.append(move(partition));
};
while (static_cast<bool>(formatted->nextPosition(position, status)) && icu_success(status)) {
if (previous_end_index < position.getStart())
create_partition(LITERAL_FIELD, previous_end_index, position.getStart());
if (position.getCategory() == UFIELD_CATEGORY_DATE_INTERVAL_SPAN) {
auto& range = position.getField() == 0 ? start_range : end_range;
LibLocale: Move definition of Range structure to dedicated header We were redefining a Range type in a few formatters. Move it to its own header to avoid ODR violations (and give it a more descriptive name).
2024-06-19 16:46:45 +00:00
range = PartitionRange { position.getField(), position.getStart(), position.getLimit() };
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
} else if (position.getCategory() == UFIELD_CATEGORY_DATE) {
create_partition(position.getField(), position.getStart(), position.getLimit());
}
previous_end_index = position.getLimit();
}
if (previous_end_index < formatted_time.length())
create_partition(LITERAL_FIELD, previous_end_index, formatted_time.length());
return result;
}
private:
Optional<icu::UnicodeString> format_impl(double time, icu::FieldPositionIterator* iterator = nullptr) const
{
UErrorCode status = U_ZERO_ERROR;
icu::UnicodeString formatted_time;
m_formatter->format(time, formatted_time, iterator, status);
if (icu_failure(status))
return {};
LibUnicode+LibJS: Normalize spaces in formatted date-time strings ICU 72 began using non-ASCII spaces in some formatted date-time strings. Every major browser has found that this introduced major breakage in web compatibility, as many sites and tools expect ASCII spaces. This patch removes these non-ASCII spaces in the same manner as the major engines. Such behavior is also tested by WPT.
2024-08-01 15:30:17 +00:00
normalize_spaces(formatted_time);
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
return formatted_time;
}
Optional<icu::FormattedDateInterval> format_range_impl(double start, double end) const
{
UErrorCode status = U_ZERO_ERROR;
if (!m_range_formatter) {
icu::UnicodeString pattern;
m_formatter->toPattern(pattern);
auto skeleton = icu::DateTimePatternGenerator::staticGetSkeleton(pattern, status);
if (icu_failure(status))
return {};
auto* formatter = icu::DateIntervalFormat::createInstance(skeleton, m_locale, status);
if (icu_failure(status))
return {};
m_range_formatter = adopt_own(*formatter);
m_range_formatter->setTimeZone(m_formatter->getTimeZone());
}
auto start_calendar = adopt_own(*m_formatter->getCalendar()->clone());
start_calendar->setTime(start, status);
if (icu_failure(status))
return {};
auto end_calendar = adopt_own(*m_formatter->getCalendar()->clone());
end_calendar->setTime(end, status);
if (icu_failure(status))
return {};
auto formatted = m_range_formatter->formatToValue(*start_calendar, *end_calendar, status);
if (icu_failure(status))
return {};
return formatted;
}
LibUnicode+LibJS: Normalize spaces in formatted date-time strings ICU 72 began using non-ASCII spaces in some formatted date-time strings. Every major browser has found that this introduced major breakage in web compatibility, as many sites and tools expect ASCII spaces. This patch removes these non-ASCII spaces in the same manner as the major engines. Such behavior is also tested by WPT.
2024-08-01 15:30:17 +00:00
// ICU 72 introduced the use of NBSP to separate time fields and day periods. All major browsers have found that
// this significantly breaks web compatibilty, and they all replace these spaces with normal ASCII spaces. See:
//
// https://bugzilla.mozilla.org/show_bug.cgi?id=1806042
// https://bugs.webkit.org/show_bug.cgi?id=252147
// https://issues.chromium.org/issues/40256057
static void normalize_spaces(icu::UnicodeString& string)
{
static char16_t NARROW_NO_BREAK_SPACE = 0x202f;
static char16_t THIN_SPACE = 0x2009;
for (i32 i = 0; i < string.length(); ++i) {
if (string[i] == NARROW_NO_BREAK_SPACE || string[i] == THIN_SPACE)
string.setCharAt(i, ' ');
}
}
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
icu::Locale& m_locale;
CalendarPattern m_pattern;
NonnullOwnPtr<icu::SimpleDateFormat> m_formatter;
mutable OwnPtr<icu::DateIntervalFormat> m_range_formatter;
};
NonnullOwnPtr<DateTimeFormat> DateTimeFormat::create_for_date_and_time_style(
StringView locale,
StringView time_zone_identifier,
Optional<HourCycle> const& hour_cycle,
Optional<bool> const& hour12,
Optional<DateTimeStyle> const& date_style,
Optional<DateTimeStyle> const& time_style)
{
UErrorCode status = U_ZERO_ERROR;
auto locale_data = LocaleData::for_locale(locale);
VERIFY(locale_data.has_value());
auto formatter = adopt_own(*verify_cast<icu::SimpleDateFormat>([&]() {
if (date_style.has_value() && time_style.has_value()) {
return icu::DateFormat::createDateTimeInstance(
icu_date_time_style(*date_style), icu_date_time_style(*time_style), locale_data->locale());
}
if (date_style.has_value()) {
return icu::DateFormat::createDateInstance(
icu_date_time_style(*date_style), locale_data->locale());
}
if (time_style.has_value()) {
return icu::DateFormat::createTimeInstance(
icu_date_time_style(*time_style), locale_data->locale());
}
LibUnicode: Fall back to GMT offset when a time zone name is unavailable The following table in TR-35 includes a web of fall back rules when the requested time zone style is unavailable: https://unicode.org/reports/tr35/tr35-dates.html#dfst-zone Conveniently, the subset of styles supported by ECMA-402 (and therefore LibUnicode) all either fall back to GMT offset or to a style that is unsupported but itself falls back to GMT offset.
2022-01-11 17:06:26 +00:00
VERIFY_NOT_REACHED();
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
}()));
icu::UnicodeString pattern;
formatter->toPattern(pattern);
auto skeleton = icu::DateTimePatternGenerator::staticGetSkeleton(pattern, status);
VERIFY(icu_success(status));
if (apply_hour_cycle_to_skeleton(skeleton, hour_cycle, hour12)) {
pattern = locale_data->date_time_pattern_generator().getBestPattern(skeleton, UDATPG_MATCH_ALL_FIELDS_LENGTH, status);
VERIFY(icu_success(status));
apply_hour_cycle_to_skeleton(pattern, hour_cycle, hour12);
formatter = adopt_own(*new icu::SimpleDateFormat(pattern, locale_data->locale(), status));
VERIFY(icu_success(status));
LibUnicode: Fall back to GMT offset when a time zone name is unavailable The following table in TR-35 includes a web of fall back rules when the requested time zone style is unavailable: https://unicode.org/reports/tr35/tr35-dates.html#dfst-zone Conveniently, the subset of styles supported by ECMA-402 (and therefore LibUnicode) all either fall back to GMT offset or to a style that is unsupported but itself falls back to GMT offset.
2022-01-11 17:06:26 +00:00
}
LibJS+LibLocale: Replace date-time formatting with ICU This uses ICU for the Intl.DateTimeFormat `format` `formatToParts`, `formatRange`, and `formatRangeToParts`. This lets us remove most data from our date-time format generator. All that remains are time zone data and locale week info, which are relied upon still for other interfaces. So they will be removed in a future patch. Note: All of the changes to the test files in this patch are now aligned with other browsers. This includes: * Some very incorrect formatting of Japanese symbols. (Looking at the old results now, it's very obvious they were wrong.) * Old FIXMEs regarding range formatting not including the start/end date when only time fields were requested, but the dates differ. * Day period inconsistencies.
2024-06-12 14:47:20 +00:00
return adopt_own(*new DateTimeFormatImpl(locale_data->locale(), pattern, time_zone_identifier, move(formatter)));
}
NonnullOwnPtr<DateTimeFormat> DateTimeFormat::create_for_pattern_options(
StringView locale,
StringView time_zone_identifier,
CalendarPattern const& options)
{
UErrorCode status = U_ZERO_ERROR;
auto locale_data = LocaleData::for_locale(locale);
VERIFY(locale_data.has_value());
auto skeleton = icu_string(options.to_pattern());
auto pattern = locale_data->date_time_pattern_generator().getBestPattern(skeleton, UDATPG_MATCH_ALL_FIELDS_LENGTH, status);
VERIFY(icu_success(status));
apply_hour_cycle_to_skeleton(pattern, options.hour_cycle, {});
auto formatter = adopt_own(*new icu::SimpleDateFormat(pattern, locale_data->locale(), status));
VERIFY(icu_success(status));
return adopt_own(*new DateTimeFormatImpl(locale_data->locale(), pattern, time_zone_identifier, move(formatter)));
LibUnicode: Implement TR-35's localized GMT offset formatting This adds an API to use LibTimeZone to convert a time zone such as "America/New_York" to a GMT offset string like "GMT-5" (short form) or "GMT-05:00" (long form).
2022-01-11 16:35:50 +00:00
}
LibUnicode: Parse and generate time zone names in long and short form
2021-12-06 20:46:49 +00:00
LibJS+LibLocale: Replace calendar weekday information with ICU
2024-06-12 21:00:45 +00:00
static constexpr Weekday icu_calendar_day_to_weekday(UCalendarDaysOfWeek day)
{
switch (day) {
case UCAL_SUNDAY:
return Weekday::Sunday;
case UCAL_MONDAY:
return Weekday::Monday;
case UCAL_TUESDAY:
return Weekday::Tuesday;
case UCAL_WEDNESDAY:
return Weekday::Wednesday;
case UCAL_THURSDAY:
return Weekday::Thursday;
case UCAL_FRIDAY:
return Weekday::Friday;
case UCAL_SATURDAY:
return Weekday::Saturday;
}
VERIFY_NOT_REACHED();
}
WeekInfo week_info_of_locale(StringView locale)
{
UErrorCode status = U_ZERO_ERROR;
auto locale_data = LocaleData::for_locale(locale);
if (!locale_data.has_value())
return {};
auto calendar = adopt_own_if_nonnull(icu::Calendar::createInstance(locale_data->locale(), status));
if (icu_failure(status))
return {};
WeekInfo week_info;
week_info.minimal_days_in_first_week = calendar->getMinimalDaysInFirstWeek();
if (auto day = calendar->getFirstDayOfWeek(status); icu_success(status))
week_info.first_day_of_week = icu_calendar_day_to_weekday(day);
auto append_if_weekend = [&](auto day) {
auto type = calendar->getDayOfWeekType(day, status);
if (icu_failure(status))
return;
switch (type) {
case UCAL_WEEKEND_ONSET:
case UCAL_WEEKEND_CEASE:
case UCAL_WEEKEND:
week_info.weekend_days.append(icu_calendar_day_to_weekday(day));
break;
default:
break;
}
};
append_if_weekend(UCAL_SUNDAY);
append_if_weekend(UCAL_MONDAY);
append_if_weekend(UCAL_TUESDAY);
append_if_weekend(UCAL_WEDNESDAY);
append_if_weekend(UCAL_THURSDAY);
append_if_weekend(UCAL_FRIDAY);
append_if_weekend(UCAL_SATURDAY);
return week_info;
}
LibUnicode: Parse and generate date, time, and date-time format patterns
2021-11-27 19:54:48 +00:00
}
Reference in a new issue Copy permalink