LibUnicode: Implement grammar validators for Unicode TR-35

ECMA-402 requires validating user input against the EBNF grammar for
Unicode locales described in TR-35: https://www.unicode.org/reports/tr35

This commit adds validators for that grammar, as well as other helper to
e.g. canonicalize a locale string.

Tests/LibUnicode/TestUnicodeLocale.cpp

@@ -0,0 +1,120 @@
+/*
+ * Copyright (c) 2021, Tim Flynn <trflynn89@pm.me>
+ *
+ * SPDX-License-Identifier: BSD-2-Clause
+ */
+
+#include <LibTest/TestCase.h>
+
+#include <LibUnicode/Locale.h>
+
+TEST_CASE(is_unicode_language_subtag)
+{
+    EXPECT(Unicode::is_unicode_language_subtag("aa"sv));
+    EXPECT(Unicode::is_unicode_language_subtag("aaa"sv));
+    EXPECT(Unicode::is_unicode_language_subtag("aaaaa"sv));
+    EXPECT(Unicode::is_unicode_language_subtag("aaaaaa"sv));
+    EXPECT(Unicode::is_unicode_language_subtag("aaaaaaa"sv));
+    EXPECT(Unicode::is_unicode_language_subtag("aaaaaaaa"sv));
+
+    EXPECT(!Unicode::is_unicode_language_subtag(""sv));
+    EXPECT(!Unicode::is_unicode_language_subtag("a"sv));
+    EXPECT(!Unicode::is_unicode_language_subtag("aaaa"sv));
+    EXPECT(!Unicode::is_unicode_language_subtag("aaaaaaaaa"sv));
+    EXPECT(!Unicode::is_unicode_language_subtag("123"sv));
+}
+
+TEST_CASE(is_unicode_script_subtag)
+{
+    EXPECT(Unicode::is_unicode_script_subtag("aaaa"sv));
+
+    EXPECT(!Unicode::is_unicode_script_subtag(""sv));
+    EXPECT(!Unicode::is_unicode_script_subtag("a"sv));
+    EXPECT(!Unicode::is_unicode_script_subtag("aa"sv));
+    EXPECT(!Unicode::is_unicode_script_subtag("aaa"sv));
+    EXPECT(!Unicode::is_unicode_script_subtag("aaaaa"sv));
+    EXPECT(!Unicode::is_unicode_script_subtag("1234"sv));
+}
+
+TEST_CASE(is_unicode_region_subtag)
+{
+    EXPECT(Unicode::is_unicode_region_subtag("aa"sv));
+    EXPECT(Unicode::is_unicode_region_subtag("123"sv));
+
+    EXPECT(!Unicode::is_unicode_region_subtag(""sv));
+    EXPECT(!Unicode::is_unicode_region_subtag("a"sv));
+    EXPECT(!Unicode::is_unicode_region_subtag("aaa"sv));
+    EXPECT(!Unicode::is_unicode_region_subtag("12"sv));
+    EXPECT(!Unicode::is_unicode_region_subtag("12a"sv));
+}
+
+TEST_CASE(is_unicode_variant_subtag)
+{
+    EXPECT(Unicode::is_unicode_variant_subtag("aaaaa"sv));
+    EXPECT(Unicode::is_unicode_variant_subtag("aaaaaa"sv));
+    EXPECT(Unicode::is_unicode_variant_subtag("aaaaaaa"sv));
+    EXPECT(Unicode::is_unicode_variant_subtag("aaaaaaaa"sv));
+
+    EXPECT(Unicode::is_unicode_variant_subtag("1aaa"sv));
+    EXPECT(Unicode::is_unicode_variant_subtag("12aa"sv));
+    EXPECT(Unicode::is_unicode_variant_subtag("123a"sv));
+    EXPECT(Unicode::is_unicode_variant_subtag("1234"sv));
+
+    EXPECT(!Unicode::is_unicode_variant_subtag(""sv));
+    EXPECT(!Unicode::is_unicode_variant_subtag("a"sv));
+    EXPECT(!Unicode::is_unicode_variant_subtag("aa"sv));
+    EXPECT(!Unicode::is_unicode_variant_subtag("aaa"sv));
+    EXPECT(!Unicode::is_unicode_variant_subtag("aaaa"sv));
+    EXPECT(!Unicode::is_unicode_variant_subtag("aaaaaaaaa"sv));
+    EXPECT(!Unicode::is_unicode_variant_subtag("a234"sv));
+}
+
+TEST_CASE(parse_unicode_locale_id)
+{
+    auto fail = [](StringView locale) {
+        auto locale_id = Unicode::parse_unicode_locale_id(locale);
+        EXPECT(!locale_id.has_value());
+    };
+    auto pass = [](StringView locale, Optional<StringView> expected_language, Optional<StringView> expected_script, Optional<StringView> expected_region, Vector<StringView> expected_variants) {
+        auto locale_id = Unicode::parse_unicode_locale_id(locale);
+        VERIFY(locale_id.has_value());
+
+        EXPECT_EQ(locale_id->language_id.language, expected_language);
+        EXPECT_EQ(locale_id->language_id.script, expected_script);
+        EXPECT_EQ(locale_id->language_id.region, expected_region);
+        EXPECT_EQ(locale_id->language_id.variants, expected_variants);
+    };
+
+    fail("a"sv);
+    fail("1234"sv);
+    fail("aaa-"sv);
+    fail("aaa-cc-"sv);
+    fail("aaa-bbbb-cc-"sv);
+    fail("aaa-bbbb-cc-123"sv);
+
+    pass("aaa"sv, "aaa"sv, {}, {}, {});
+    pass("aaa-bbbb"sv, "aaa"sv, "bbbb"sv, {}, {});
+    pass("aaa-cc"sv, "aaa"sv, {}, "cc"sv, {});
+    pass("aaa-bbbb-cc"sv, "aaa"sv, "bbbb"sv, "cc"sv, {});
+    pass("aaa-bbbb-cc-1234"sv, "aaa"sv, "bbbb"sv, "cc"sv, { "1234"sv });
+    pass("aaa-bbbb-cc-1234-5678"sv, "aaa"sv, "bbbb"sv, "cc"sv, { "1234"sv, "5678"sv });
+}
+
+TEST_CASE(canonicalize_unicode_locale_id)
+{
+    auto test = [](StringView locale, StringView expected_canonical_locale) {
+        auto locale_id = Unicode::parse_unicode_locale_id(locale);
+        VERIFY(locale_id.has_value());
+
+        auto canonical_locale = Unicode::canonicalize_unicode_locale_id(*locale_id);
+        EXPECT_EQ(canonical_locale, expected_canonical_locale);
+    };
+
+    test("aaa"sv, "aaa"sv);
+    test("AaA"sv, "aaa"sv);
+    test("aaa-bbbb"sv, "aaa-Bbbb"sv);
+    test("aaa-cc"sv, "aaa-CC"sv);
+    test("aaa-bBBB-cC"sv, "aaa-Bbbb-CC"sv);
+    test("aaa-bbbb-cc-1234"sv, "aaa-Bbbb-CC-1234"sv);
+    test("aaa-bbbb-cc-ABCDE"sv, "aaa-Bbbb-CC-abcde"sv);
+}

Userland/Libraries/LibUnicode/CMakeLists.txt

@@ -3,6 +3,7 @@ include(unicode_data.cmake)
 SET(SOURCES
     ${UNICODE_DATA_SOURCES}
     CharacterTypes.cpp
+    Locale.cpp
 )
 
 serenity_lib(LibUnicode unicode)

Userland/Libraries/LibUnicode/Locale.cpp

@@ -0,0 +1,152 @@
+/*
+ * Copyright (c) 2021, Tim Flynn <trflynn89@pm.me>
+ *
+ * SPDX-License-Identifier: BSD-2-Clause
+ */
+
+#include <AK/AllOf.h>
+#include <AK/CharacterTypes.h>
+#include <AK/GenericLexer.h>
+#include <AK/QuickSort.h>
+#include <AK/StringBuilder.h>
+#include <LibUnicode/Locale.h>
+
+namespace Unicode {
+
+bool is_unicode_language_subtag(StringView subtag)
+{
+    // unicode_language_subtag = alpha{2,3} | alpha{5,8}
+    if ((subtag.length() < 2) || (subtag.length() == 4) || (subtag.length() > 8))
+        return false;
+    return all_of(subtag, is_ascii_alpha);
+}
+
+bool is_unicode_script_subtag(StringView subtag)
+{
+    // unicode_script_subtag = alpha{4}
+    if (subtag.length() != 4)
+        return false;
+    return all_of(subtag, is_ascii_alpha);
+}
+
+bool is_unicode_region_subtag(StringView subtag)
+{
+    // unicode_region_subtag = (alpha{2} | digit{3})
+    if (subtag.length() == 2)
+        return all_of(subtag, is_ascii_alpha);
+    if (subtag.length() == 3)
+        return all_of(subtag, is_ascii_digit);
+    return false;
+}
+
+bool is_unicode_variant_subtag(StringView subtag)
+{
+    // unicode_variant_subtag = (alphanum{5,8} | digit alphanum{3})
+    if ((subtag.length() >= 5) && (subtag.length() <= 8))
+        return all_of(subtag, is_ascii_alphanumeric);
+    if (subtag.length() == 4)
+        return is_ascii_digit(subtag[0]) && all_of(subtag.substring_view(1), is_ascii_alphanumeric);
+    return false;
+}
+
+Optional<LanguageID> parse_unicode_language_id(StringView language)
+{
+    // https://unicode.org/reports/tr35/#Unicode_language_identifier
+    //
+    // unicode_language_id = "root"
+    //     OR
+    // unicode_language_id = ((unicode_language_subtag (sep unicode_script_subtag)?) | unicode_script_subtag)
+    //                       (sep unicode_region_subtag)?
+    //                       (sep unicode_variant_subtag)*
+    LanguageID language_id {};
+
+    if (language == "root"sv) {
+        language_id.is_root = true;
+        return language_id;
+    }
+
+    auto segments = language.split_view_if(is_any_of("-_"sv), true); // keep_empty=true to ensure valid data follows a separator.
+    size_t index = 0;
+
+    if (segments.size() == index)
+        return {};
+
+    if (is_unicode_language_subtag(segments[index])) {
+        language_id.language = segments[index];
+        if (segments.size() == ++index)
+            return language_id;
+    }
+
+    if (is_unicode_script_subtag(segments[index])) {
+        language_id.script = segments[index];
+        if (segments.size() == ++index)
+            return language_id;
+    } else if (!language_id.language.has_value()) {
+        return {};
+    }
+
+    if (is_unicode_region_subtag(segments[index])) {
+        language_id.region = segments[index];
+        if (segments.size() == ++index)
+            return language_id;
+    }
+
+    while (index < segments.size()) {
+        if (!is_unicode_variant_subtag(segments[index]))
+            return {};
+        language_id.variants.append(segments[index++]);
+    }
+
+    return language_id;
+}
+
+Optional<LocaleID> parse_unicode_locale_id(StringView locale)
+{
+    LocaleID locale_id {};
+
+    // https://unicode.org/reports/tr35/#Unicode_locale_identifier
+    //
+    // unicode_locale_id = unicode_language_id
+    //                     extensions*
+    //                     pu_extensions?
+    auto language_id = parse_unicode_language_id(locale);
+    if (!language_id.has_value())
+        return {};
+
+    // FIXME: Handle extensions and pu_extensions.
+    return LocaleID { language_id.release_value() };
+}
+
+Optional<String> canonicalize_unicode_locale_id(LocaleID& locale_id)
+{
+    // https://unicode.org/reports/tr35/#Canonical_Unicode_Locale_Identifiers
+    StringBuilder builder;
+
+    if (!locale_id.language_id.language.has_value())
+        return {};
+
+    builder.append(locale_id.language_id.language->to_lowercase_string());
+
+    if (locale_id.language_id.script.has_value()) {
+        builder.append('-');
+        builder.append(locale_id.language_id.script->to_titlecase_string());
+    }
+
+    if (locale_id.language_id.region.has_value()) {
+        builder.append('-');
+        builder.append(locale_id.language_id.region->to_uppercase_string());
+    }
+
+    quick_sort(locale_id.language_id.variants);
+
+    for (auto const& variant : locale_id.language_id.variants) {
+        builder.append('-');
+        builder.append(variant.to_lowercase_string());
+    }
+
+    // FIXME: Handle extensions and pu_extensions.
+
+    return builder.build();
+}
+
+}

Userland/Libraries/LibUnicode/Locale.h

@@ -0,0 +1,40 @@
+/*
+ * Copyright (c) 2021, Tim Flynn <trflynn89@pm.me>
+ *
+ * SPDX-License-Identifier: BSD-2-Clause
+ */
+
+#pragma once
+
+#include <AK/Optional.h>
+#include <AK/String.h>
+#include <AK/StringView.h>
+#include <AK/Vector.h>
+#include <LibUnicode/Forward.h>
+
+namespace Unicode {
+
+struct LanguageID {
+    bool is_root { false };
+    Optional<StringView> language {};
+    Optional<StringView> script {};
+    Optional<StringView> region {};
+    Vector<StringView> variants {};
+};
+
+struct LocaleID {
+    LanguageID language_id {};
+};
+
+// Note: These methods only verify that the provided strings match the EBNF grammar of the
+// Unicode identifier subtag (i.e. no validation is done that the tags actually exist).
+bool is_unicode_language_subtag(StringView);
+bool is_unicode_script_subtag(StringView);
+bool is_unicode_region_subtag(StringView);
+bool is_unicode_variant_subtag(StringView);
+
+Optional<LanguageID> parse_unicode_language_id(StringView);
+Optional<LocaleID> parse_unicode_locale_id(StringView);
+Optional<String> canonicalize_unicode_locale_id(LocaleID&);
+
+}