395 lines
13 KiB
C++
395 lines
13 KiB
C++
/*
|
|
* Copyright (c) 2022, Tim Flynn <trflynn89@pm.me>
|
|
*
|
|
* SPDX-License-Identifier: BSD-2-Clause
|
|
*/
|
|
|
|
#include "../LibUnicode/GeneratorUtil.h" // FIXME: Move this somewhere common.
|
|
#include <AK/Format.h>
|
|
#include <AK/HashMap.h>
|
|
#include <AK/SourceGenerator.h>
|
|
#include <AK/String.h>
|
|
#include <AK/StringBuilder.h>
|
|
#include <AK/Vector.h>
|
|
#include <LibCore/ArgsParser.h>
|
|
#include <LibCore/File.h>
|
|
|
|
namespace {
|
|
|
|
struct DateTime {
|
|
u16 year { 0 };
|
|
Optional<u8> month;
|
|
Optional<u8> day;
|
|
|
|
Optional<u8> last_weekday;
|
|
Optional<u8> after_weekday;
|
|
|
|
Optional<u8> hour;
|
|
Optional<u8> minute;
|
|
Optional<u8> second;
|
|
};
|
|
|
|
struct TimeZoneOffset {
|
|
i64 offset { 0 };
|
|
Optional<DateTime> until;
|
|
};
|
|
|
|
struct TimeZoneData {
|
|
HashMap<String, Vector<TimeZoneOffset>> time_zones;
|
|
Vector<String> time_zone_names;
|
|
Vector<Alias> time_zone_aliases;
|
|
};
|
|
|
|
}
|
|
|
|
template<>
|
|
struct AK::Formatter<DateTime> : Formatter<FormatString> {
|
|
ErrorOr<void> format(FormatBuilder& builder, DateTime const& date_time)
|
|
{
|
|
return Formatter<FormatString>::format(builder,
|
|
"{{ {}, {}, {}, {}, {}, {}, {}, {} }}",
|
|
date_time.year,
|
|
date_time.month.value_or(1),
|
|
date_time.day.value_or(1),
|
|
date_time.last_weekday.value_or(0),
|
|
date_time.after_weekday.value_or(0),
|
|
date_time.hour.value_or(0),
|
|
date_time.minute.value_or(0),
|
|
date_time.second.value_or(0));
|
|
}
|
|
};
|
|
|
|
template<>
|
|
struct AK::Formatter<TimeZoneOffset> : Formatter<FormatString> {
|
|
ErrorOr<void> format(FormatBuilder& builder, TimeZoneOffset const& time_zone_offset)
|
|
{
|
|
return Formatter<FormatString>::format(builder,
|
|
"{{ {}, {}, {} }}",
|
|
time_zone_offset.offset,
|
|
time_zone_offset.until.value_or({}),
|
|
time_zone_offset.until.has_value());
|
|
}
|
|
};
|
|
|
|
static Optional<DateTime> parse_date_time(Span<StringView const> segments)
|
|
{
|
|
constexpr auto months = Array { "Jan"sv, "Feb"sv, "Mar"sv, "Apr"sv, "May"sv, "Jun"sv, "Jul"sv, "Aug"sv, "Sep"sv, "Oct"sv, "Nov"sv, "Dec"sv };
|
|
constexpr auto weekdays = Array { "Sun"sv, "Mon"sv, "Tue"sv, "Wed"sv, "Thu"sv, "Fri"sv, "Sat"sv };
|
|
|
|
auto comment_index = find_index(segments.begin(), segments.end(), "#"sv);
|
|
if (comment_index != segments.size())
|
|
segments = segments.slice(0, comment_index);
|
|
if (segments.is_empty())
|
|
return {};
|
|
|
|
DateTime date_time {};
|
|
date_time.year = segments[0].to_uint().value();
|
|
|
|
if (segments.size() > 1)
|
|
date_time.month = find_index(months.begin(), months.end(), segments[1]) + 1;
|
|
|
|
if (segments.size() > 2) {
|
|
if (segments[2].starts_with("last"sv)) {
|
|
auto weekday = segments[2].substring_view("last"sv.length());
|
|
date_time.last_weekday = find_index(weekdays.begin(), weekdays.end(), weekday);
|
|
} else if (auto index = segments[2].find(">="sv); index.has_value()) {
|
|
auto weekday = segments[2].substring_view(0, *index);
|
|
date_time.after_weekday = find_index(weekdays.begin(), weekdays.end(), weekday);
|
|
|
|
auto day = segments[2].substring_view(*index + ">="sv.length());
|
|
date_time.day = day.to_uint().value();
|
|
} else {
|
|
date_time.day = segments[2].to_uint().value();
|
|
}
|
|
}
|
|
|
|
if (segments.size() > 3) {
|
|
// FIXME: Some times end with a letter, e.g. "2:00u" and "2:00s". Figure out what this means and handle it.
|
|
auto time_segments = segments[3].split_view(':');
|
|
|
|
date_time.hour = time_segments[0].to_int().value();
|
|
date_time.minute = time_segments.size() > 1 ? time_segments[1].substring_view(0, 2).to_uint().value() : 0;
|
|
date_time.second = time_segments.size() > 2 ? time_segments[2].substring_view(0, 2).to_uint().value() : 0;
|
|
}
|
|
|
|
return date_time;
|
|
}
|
|
|
|
static i64 parse_time_offset(StringView segment)
|
|
{
|
|
auto segments = segment.split_view(':');
|
|
|
|
i64 hours = segments[0].to_int().value();
|
|
i64 minutes = segments.size() > 1 ? segments[1].to_uint().value() : 0;
|
|
i64 seconds = segments.size() > 2 ? segments[2].to_uint().value() : 0;
|
|
|
|
i64 sign = ((hours < 0) || (segments[0] == "-0"sv)) ? -1 : 1;
|
|
return (hours * 3600) + sign * ((minutes * 60) + seconds);
|
|
}
|
|
|
|
static Vector<TimeZoneOffset>& parse_zone(StringView zone_line, TimeZoneData& time_zone_data)
|
|
{
|
|
auto segments = zone_line.split_view_if([](char ch) { return (ch == '\t') || (ch == ' '); });
|
|
|
|
// "Zone" NAME STDOFF RULES FORMAT [UNTIL]
|
|
VERIFY(segments[0] == "Zone"sv);
|
|
auto name = segments[1];
|
|
|
|
TimeZoneOffset time_zone {};
|
|
time_zone.offset = parse_time_offset(segments[2]);
|
|
|
|
if (segments.size() > 5)
|
|
time_zone.until = parse_date_time(segments.span().slice(5));
|
|
|
|
auto& time_zones = time_zone_data.time_zones.ensure(name);
|
|
time_zones.append(move(time_zone));
|
|
|
|
if (!time_zone_data.time_zone_names.contains_slow(name))
|
|
time_zone_data.time_zone_names.append(name);
|
|
|
|
return time_zones;
|
|
}
|
|
|
|
static void parse_zone_continuation(StringView zone_line, Vector<TimeZoneOffset>& time_zones)
|
|
{
|
|
auto segments = zone_line.split_view_if([](char ch) { return (ch == '\t') || (ch == ' '); });
|
|
|
|
// STDOFF RULES FORMAT [UNTIL]
|
|
TimeZoneOffset time_zone {};
|
|
time_zone.offset = parse_time_offset(segments[0]);
|
|
|
|
if (segments.size() > 3)
|
|
time_zone.until = parse_date_time(segments.span().slice(3));
|
|
|
|
time_zones.append(move(time_zone));
|
|
}
|
|
|
|
static void parse_link(StringView link_line, TimeZoneData& time_zone_data)
|
|
{
|
|
auto segments = link_line.split_view_if([](char ch) { return (ch == '\t') || (ch == ' '); });
|
|
|
|
// Link TARGET LINK-NAME
|
|
VERIFY(segments[0] == "Link"sv);
|
|
auto target = segments[1];
|
|
auto alias = segments[2];
|
|
|
|
time_zone_data.time_zone_aliases.append({ target, alias });
|
|
}
|
|
|
|
static ErrorOr<void> parse_time_zones(StringView time_zone_path, TimeZoneData& time_zone_data)
|
|
{
|
|
// For reference, the man page for `zic` has the best documentation of the TZDB file format.
|
|
auto file = TRY(Core::File::open(time_zone_path, Core::OpenMode::ReadOnly));
|
|
Vector<TimeZoneOffset>* last_parsed_zone = nullptr;
|
|
|
|
while (file->can_read_line()) {
|
|
auto line = file->read_line();
|
|
if (line.is_empty() || line.trim_whitespace(TrimMode::Left).starts_with('#'))
|
|
continue;
|
|
|
|
if (line.starts_with("Zone"sv)) {
|
|
last_parsed_zone = &parse_zone(line, time_zone_data);
|
|
} else if (line.starts_with('\t')) {
|
|
VERIFY(last_parsed_zone != nullptr);
|
|
parse_zone_continuation(line, *last_parsed_zone);
|
|
} else {
|
|
last_parsed_zone = nullptr;
|
|
|
|
if (line.starts_with("Link"sv))
|
|
parse_link(line, time_zone_data);
|
|
}
|
|
}
|
|
|
|
return {};
|
|
}
|
|
|
|
static String format_identifier(StringView owner, String identifier)
|
|
{
|
|
constexpr auto gmt_time_zones = Array { "Etc/GMT"sv, "GMT"sv };
|
|
|
|
for (auto gmt_time_zone : gmt_time_zones) {
|
|
if (identifier.starts_with(gmt_time_zone)) {
|
|
auto offset = identifier.substring_view(gmt_time_zone.length());
|
|
|
|
if (offset.starts_with('+'))
|
|
identifier = String::formatted("{}_Ahead_{}", gmt_time_zone, offset.substring_view(1));
|
|
else if (offset.starts_with('-'))
|
|
identifier = String::formatted("{}_Behind_{}", gmt_time_zone, offset.substring_view(1));
|
|
}
|
|
}
|
|
|
|
identifier = identifier.replace("-"sv, "_"sv, true);
|
|
identifier = identifier.replace("/"sv, "_"sv, true);
|
|
|
|
if (all_of(identifier, is_ascii_digit))
|
|
return String::formatted("{}_{}", owner[0], identifier);
|
|
if (is_ascii_lower_alpha(identifier[0]))
|
|
return String::formatted("{:c}{}", to_ascii_uppercase(identifier[0]), identifier.substring_view(1));
|
|
return identifier;
|
|
}
|
|
|
|
static void generate_time_zone_data_header(Core::File& file, TimeZoneData& time_zone_data)
|
|
{
|
|
StringBuilder builder;
|
|
SourceGenerator generator { builder };
|
|
|
|
generator.append(R"~~~(
|
|
#pragma once
|
|
|
|
#include <AK/Types.h>
|
|
|
|
namespace TimeZone {
|
|
)~~~");
|
|
|
|
generate_enum(generator, format_identifier, "TimeZone"sv, {}, time_zone_data.time_zone_names, time_zone_data.time_zone_aliases);
|
|
|
|
generator.append(R"~~~(
|
|
}
|
|
)~~~");
|
|
|
|
VERIFY(file.write(generator.as_string_view()));
|
|
}
|
|
|
|
static void generate_time_zone_data_implementation(Core::File& file, TimeZoneData& time_zone_data)
|
|
{
|
|
StringBuilder builder;
|
|
SourceGenerator generator { builder };
|
|
|
|
generator.append(R"~~~(
|
|
#include <AK/Array.h>
|
|
#include <AK/BinarySearch.h>
|
|
#include <AK/Optional.h>
|
|
#include <AK/Span.h>
|
|
#include <AK/StringView.h>
|
|
#include <AK/Time.h>
|
|
#include <LibTimeZone/TimeZone.h>
|
|
#include <LibTimeZone/TimeZoneData.h>
|
|
|
|
namespace TimeZone {
|
|
|
|
struct DateTime {
|
|
AK::Time time_since_epoch() const
|
|
{
|
|
// FIXME: This implementation does not take last_weekday or after_weekday into account.
|
|
return AK::Time::from_timestamp(year, month, day, hour, minute, second, 0);
|
|
}
|
|
|
|
u16 year { 0 };
|
|
u8 month { 1 };
|
|
u8 day { 1 };
|
|
|
|
u8 last_weekday { 0 };
|
|
u8 after_weekday { 0 };
|
|
|
|
u8 hour { 0 };
|
|
u8 minute { 0 };
|
|
u8 second { 0 };
|
|
};
|
|
|
|
struct TimeZoneOffset {
|
|
i64 offset { 0 };
|
|
|
|
DateTime until {};
|
|
bool has_until { false };
|
|
};
|
|
)~~~");
|
|
|
|
auto append_time_zone_offsets = [&](auto const& name, auto const& time_zone_offsets) {
|
|
generator.set("name", name);
|
|
generator.set("size", String::number(time_zone_offsets.size()));
|
|
|
|
generator.append(R"~~~(
|
|
static constexpr Array<TimeZoneOffset, @size@> @name@ { {
|
|
)~~~");
|
|
|
|
for (auto const& time_zone_offset : time_zone_offsets)
|
|
generator.append(String::formatted(" {},\n", time_zone_offset));
|
|
|
|
generator.append("} };\n");
|
|
};
|
|
|
|
generate_mapping(generator, time_zone_data.time_zone_names, "TimeZoneOffset"sv, "s_time_zone_offsets"sv, "s_time_zone_offsets_{}", format_identifier,
|
|
[&](auto const& name, auto const& value) {
|
|
auto const& time_zone_offsets = time_zone_data.time_zones.find(value)->value;
|
|
append_time_zone_offsets(name, time_zone_offsets);
|
|
});
|
|
|
|
auto append_string_conversions = [&](StringView enum_title, StringView enum_snake, auto const& values, auto const& aliases) {
|
|
HashValueMap<String> hashes;
|
|
hashes.ensure_capacity(values.size());
|
|
|
|
auto hash = [](auto const& value) {
|
|
return CaseInsensitiveStringViewTraits::hash(value);
|
|
};
|
|
|
|
for (auto const& value : values)
|
|
hashes.set(hash(value), format_identifier(enum_title, value));
|
|
for (auto const& alias : aliases)
|
|
hashes.set(hash(alias.alias), format_identifier(enum_title, alias.alias));
|
|
|
|
ValueFromStringOptions options {};
|
|
options.sensitivity = CaseSensitivity::CaseInsensitive;
|
|
|
|
generate_value_from_string(generator, "{}_from_string"sv, enum_title, enum_snake, move(hashes), options);
|
|
generate_value_to_string(generator, "{}_to_string"sv, enum_title, enum_snake, format_identifier, values);
|
|
};
|
|
|
|
append_string_conversions("TimeZone"sv, "time_zone"sv, time_zone_data.time_zone_names, time_zone_data.time_zone_aliases);
|
|
|
|
generator.append(R"~~~(
|
|
Optional<i64> get_time_zone_offset(TimeZone time_zone, AK::Time time)
|
|
{
|
|
// FIXME: This implementation completely ignores DST.
|
|
auto const& time_zone_offsets = s_time_zone_offsets[to_underlying(time_zone)];
|
|
|
|
size_t index = 0;
|
|
for (; index < time_zone_offsets.size(); ++index) {
|
|
auto const& time_zone_offset = time_zone_offsets[index];
|
|
|
|
if (!time_zone_offset.has_until || (time_zone_offset.until.time_since_epoch() > time))
|
|
break;
|
|
}
|
|
|
|
VERIFY(index < time_zone_offsets.size());
|
|
return time_zone_offsets[index].offset;
|
|
}
|
|
|
|
}
|
|
)~~~");
|
|
|
|
VERIFY(file.write(generator.as_string_view()));
|
|
}
|
|
|
|
ErrorOr<int> serenity_main(Main::Arguments arguments)
|
|
{
|
|
StringView generated_header_path;
|
|
StringView generated_implementation_path;
|
|
Vector<StringView> time_zone_paths;
|
|
|
|
Core::ArgsParser args_parser;
|
|
args_parser.add_option(generated_header_path, "Path to the time zone data header file to generate", "generated-header-path", 'h', "generated-header-path");
|
|
args_parser.add_option(generated_implementation_path, "Path to the time zone data implementation file to generate", "generated-implementation-path", 'c', "generated-implementation-path");
|
|
args_parser.add_positional_argument(time_zone_paths, "Paths to the time zone database files", "time-zone-paths");
|
|
args_parser.parse(arguments);
|
|
|
|
auto open_file = [&](StringView path) -> ErrorOr<NonnullRefPtr<Core::File>> {
|
|
if (path.is_empty()) {
|
|
args_parser.print_usage(stderr, arguments.argv[0]);
|
|
return Error::from_string_literal("Must provide all command line options"sv);
|
|
}
|
|
|
|
return Core::File::open(path, Core::OpenMode::ReadWrite);
|
|
};
|
|
|
|
auto generated_header_file = TRY(open_file(generated_header_path));
|
|
auto generated_implementation_file = TRY(open_file(generated_implementation_path));
|
|
|
|
TimeZoneData time_zone_data {};
|
|
for (auto time_zone_path : time_zone_paths)
|
|
TRY(parse_time_zones(time_zone_path, time_zone_data));
|
|
|
|
generate_time_zone_data_header(generated_header_file, time_zone_data);
|
|
generate_time_zone_data_implementation(generated_implementation_file, time_zone_data);
|
|
|
|
return 0;
|
|
}
|