Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
84.69% |
83 / 98 |
|
28.57% |
2 / 7 |
CRAP | |
0.00% |
0 / 1 |
| IcuCollation | |
84.69% |
83 / 98 |
|
28.57% |
2 / 7 |
36.91 | |
0.00% |
0 / 1 |
| __construct | |
64.29% |
9 / 14 |
|
0.00% |
0 / 1 |
4.73 | |||
| getSortKey | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| getFirstLetter | |
79.17% |
19 / 24 |
|
0.00% |
0 / 1 |
8.58 | |||
| getPrimarySortKey | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| getFirstLetterData | |
100.00% |
15 / 15 |
|
100.00% |
1 / 1 |
2 | |||
| fetchFirstLetterData | |
92.31% |
36 / 39 |
|
0.00% |
0 / 1 |
13.08 | |||
| isCjk | |
75.00% |
3 / 4 |
|
0.00% |
0 / 1 |
4.25 | |||
| 1 | <?php |
| 2 | /** |
| 3 | * This program is free software; you can redistribute it and/or modify |
| 4 | * it under the terms of the GNU General Public License as published by |
| 5 | * the Free Software Foundation; either version 2 of the License, or |
| 6 | * (at your option) any later version. |
| 7 | * |
| 8 | * This program is distributed in the hope that it will be useful, |
| 9 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 10 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 11 | * GNU General Public License for more details. |
| 12 | * |
| 13 | * You should have received a copy of the GNU General Public License along |
| 14 | * with this program; if not, write to the Free Software Foundation, Inc., |
| 15 | * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. |
| 16 | * http://www.gnu.org/copyleft/gpl.html |
| 17 | * |
| 18 | * @file |
| 19 | */ |
| 20 | |
| 21 | use MediaWiki\Language\Language; |
| 22 | use MediaWiki\Languages\LanguageFactory; |
| 23 | use MediaWiki\MediaWikiServices; |
| 24 | use Wikimedia\ArrayUtils\ArrayUtils; |
| 25 | |
| 26 | /** |
| 27 | * @since 1.16.3 |
| 28 | */ |
| 29 | class IcuCollation extends Collation { |
| 30 | private const FIRST_LETTER_VERSION = 4; |
| 31 | |
| 32 | /** @var Collator */ |
| 33 | private $primaryCollator; |
| 34 | |
| 35 | /** @var Collator */ |
| 36 | private $mainCollator; |
| 37 | |
| 38 | /** @var string */ |
| 39 | private $locale; |
| 40 | |
| 41 | /** @var Language */ |
| 42 | protected $digitTransformLanguage; |
| 43 | |
| 44 | /** @var bool */ |
| 45 | private $useNumericCollation = false; |
| 46 | |
| 47 | /** @var array */ |
| 48 | private $firstLetterData; |
| 49 | |
| 50 | /** |
| 51 | * Unified CJK blocks. |
| 52 | * |
| 53 | * The same definition of a CJK block must be used for both Collation and |
| 54 | * generateCollationData.php. These blocks are omitted from the first |
| 55 | * letter data, as an optimisation measure and because the default UCA table |
| 56 | * is pretty useless for sorting Chinese text anyway. Japanese and Korean |
| 57 | * blocks are not included here, because they are smaller and more useful. |
| 58 | */ |
| 59 | private const CJK_BLOCKS = [ |
| 60 | [ 0x2E80, 0x2EFF ], // CJK Radicals Supplement |
| 61 | [ 0x2F00, 0x2FDF ], // Kangxi Radicals |
| 62 | [ 0x2FF0, 0x2FFF ], // Ideographic Description Characters |
| 63 | [ 0x3000, 0x303F ], // CJK Symbols and Punctuation |
| 64 | [ 0x31C0, 0x31EF ], // CJK Strokes |
| 65 | [ 0x3200, 0x32FF ], // Enclosed CJK Letters and Months |
| 66 | [ 0x3300, 0x33FF ], // CJK Compatibility |
| 67 | [ 0x3400, 0x4DBF ], // CJK Unified Ideographs Extension A |
| 68 | [ 0x4E00, 0x9FFF ], // CJK Unified Ideographs |
| 69 | [ 0xF900, 0xFAFF ], // CJK Compatibility Ideographs |
| 70 | [ 0xFE30, 0xFE4F ], // CJK Compatibility Forms |
| 71 | [ 0x20000, 0x2A6DF ], // CJK Unified Ideographs Extension B |
| 72 | [ 0x2A700, 0x2B73F ], // CJK Unified Ideographs Extension C |
| 73 | [ 0x2B740, 0x2B81F ], // CJK Unified Ideographs Extension D |
| 74 | [ 0x2F800, 0x2FA1F ], // CJK Compatibility Ideographs Supplement |
| 75 | ]; |
| 76 | |
| 77 | /** |
| 78 | * Additional characters (or character groups) to be considered separate |
| 79 | * letters for given languages, or to be removed from the list of such |
| 80 | * letters (denoted by keys starting with '-'). |
| 81 | * |
| 82 | * These are additions to (or subtractions from) the data stored in the |
| 83 | * first-letters-root.php data file (which among others includes full basic |
| 84 | * Latin, Cyrillic and Greek alphabets). |
| 85 | * |
| 86 | * "Separate letter" is a letter that would have a separate heading/section |
| 87 | * for it in a dictionary or a phone book in this language. This data isn't |
| 88 | * used for sorting (the ICU library handles that), only for deciding which |
| 89 | * characters (or character groups) to use as headings. |
| 90 | * |
| 91 | * Initially generated based on the primary level of Unicode collation |
| 92 | * tailorings available at http://developer.mimer.com/charts/tailorings.htm , |
| 93 | * later modified. |
| 94 | * |
| 95 | * Empty arrays are intended; this signifies that the data for the language is |
| 96 | * available and that there are, in fact, no additional letters to consider. |
| 97 | */ |
| 98 | private const TAILORING_FIRST_LETTERS = [ |
| 99 | 'af' => [], |
| 100 | 'am' => [], |
| 101 | 'ar' => [], |
| 102 | 'as' => [ "\u{0982}", "\u{0981}", "\u{0983}", "\u{09CE}", "ক্ষ " ], |
| 103 | 'ast' => [ "Ch", "Ll", "Ñ" ], // not in libicu |
| 104 | 'az' => [ "Ç", "Ə", "Ğ", "İ", "Ö", "Ş", "Ü" ], |
| 105 | 'be' => [ "Ё" ], |
| 106 | 'be-tarask' => [ "Ё" ], |
| 107 | 'bg' => [], |
| 108 | 'bn' => [ 'ং', 'ঃ', 'ঁ' ], |
| 109 | 'bn@collation=traditional' => [ |
| 110 | 'ং', 'ঃ', 'ঁ', 'ক্', 'খ্', 'গ্', 'ঘ্', 'ঙ্', 'চ্', 'ছ্', 'জ্', 'ঝ্', |
| 111 | 'ঞ্', 'ট্', 'ঠ্', 'ড্', 'ঢ্', 'ণ্', 'ৎ', 'থ্', 'দ্', 'ধ্', 'ন্', 'প্', |
| 112 | 'ফ্', 'ব্', 'ভ্', 'ম্', 'য্', 'র্', 'ৰ্', 'ল্', 'ৱ্', 'শ্', 'ষ্', 'স্', 'হ্' |
| 113 | ], |
| 114 | 'bo' => [], |
| 115 | 'br' => [ "Ch", "C'h" ], |
| 116 | 'bs' => [ "Č", "Ć", "Dž", "Đ", "Lj", "Nj", "Š", "Ž" ], |
| 117 | 'bs-Cyrl' => [], |
| 118 | 'ca' => [], |
| 119 | 'chr' => [], |
| 120 | 'co' => [], // not in libicu |
| 121 | 'cs' => [ "Č", "Ch", "Ř", "Š", "Ž" ], |
| 122 | 'cy' => [ "Ch", "Dd", "Ff", "Ng", "Ll", "Ph", "Rh", "Th" ], |
| 123 | 'da' => [ "Æ", "Ø", "Å" ], |
| 124 | 'de' => [], |
| 125 | 'de-AT@collation=phonebook' => [ 'ä', 'ö', 'ü', 'ß' ], |
| 126 | 'dsb' => [ "Č", "Ć", "Dź", "Ě", "Ch", "Ł", "Ń", "Ŕ", "Š", "Ś", "Ž", "Ź" ], |
| 127 | 'ee' => [ "Dz", "Ɖ", "Ɛ", "Ƒ", "Gb", "Ɣ", "Kp", "Ny", "Ŋ", "Ɔ", "Ts", "Ʋ" ], |
| 128 | 'el' => [], |
| 129 | 'en' => [], |
| 130 | 'eo' => [ "Ĉ", "Ĝ", "Ĥ", "Ĵ", "Ŝ", "Ŭ" ], |
| 131 | 'es' => [ "Ñ" ], |
| 132 | 'et' => [ "Š", "Ž", "Õ", "Ä", "Ö", "Ü" ], |
| 133 | 'eu' => [ "Ñ" ], // not in libicu |
| 134 | 'fa' => [ |
| 135 | // RTL, let's put each letter on a new line |
| 136 | "آ", |
| 137 | "ء", |
| 138 | "ه", |
| 139 | "ا", |
| 140 | "و" |
| 141 | ], |
| 142 | 'fi' => [ "Å", "Ä", "Ö" ], |
| 143 | 'fil' => [ "Ñ", "Ng" ], |
| 144 | 'fo' => [ "Á", "Ð", "Í", "Ó", "Ú", "Ý", "Æ", "Ø", "Å" ], |
| 145 | 'fr' => [], |
| 146 | 'fr-CA' => [], // fr-CA sorts accents slightly different from fr. |
| 147 | 'fur' => [ "À", "Á", "Â", "È", "Ì", "Ò", "Ù" ], // not in libicu |
| 148 | 'fy' => [], // not in libicu |
| 149 | 'ga' => [], |
| 150 | 'gd' => [], // not in libicu |
| 151 | 'gl' => [ "Ch", "Ll", "Ñ" ], |
| 152 | 'gu' => [ "\u{0A82}", "\u{0A83}", "\u{0A81}", "\u{0AB3}" ], |
| 153 | 'ha' => [ 'Ɓ', 'Ɗ', 'Ƙ', 'Sh', 'Ts', 'Ƴ' ], |
| 154 | 'haw' => [ 'ʻ' ], |
| 155 | 'he' => [], |
| 156 | 'hi' => [ "\u{0902}", "\u{0903}" ], |
| 157 | 'hr' => [ "Č", "Ć", "Dž", "Đ", "Lj", "Nj", "Š", "Ž" ], |
| 158 | 'hsb' => [ "Č", "Dź", "Ě", "Ch", "Ł", "Ń", "Ř", "Š", "Ć", "Ž" ], |
| 159 | 'hu' => [ "Cs", "Dz", "Dzs", "Gy", "Ly", "Ny", "Ö", "Sz", "Ty", "Ü", "Zs" ], |
| 160 | 'hy' => [ "և" ], |
| 161 | 'id' => [], |
| 162 | 'ig' => [ "Ch", "Gb", "Gh", "Gw", "Ị", "Kp", "Kw", "Ṅ", "Nw", "Ny", "Ọ", "Sh", "Ụ" ], |
| 163 | 'is' => [ "Á", "Ð", "É", "Í", "Ó", "Ú", "Ý", "Þ", "Æ", "Ö", "Å" ], |
| 164 | 'it' => [], |
| 165 | 'ka' => [], |
| 166 | 'kk' => [ "Ё", "Ү", "І" ], |
| 167 | 'kl' => [ "Æ", "Ø", "Å" ], |
| 168 | 'km' => [ |
| 169 | "រ", "ឫ", "ឬ", "ល", "ឭ", "ឮ", "\u{17BB}\u{17C6}", |
| 170 | "\u{17C6}", "\u{17B6}\u{17C6}", "\u{17C7}", |
| 171 | "\u{17B7}\u{17C7}", "\u{17BB}\u{17C7}", |
| 172 | "\u{17C1}\u{17C7}", "\u{17C4}\u{17C7}", |
| 173 | ], |
| 174 | 'kn' => [ "\u{0C81}", "\u{0C83}", "\u{0CF1}", "\u{0CF2}" ], |
| 175 | 'kok' => [ "\u{0902}", "\u{0903}", "ळ", "क्ष" ], |
| 176 | 'ku' => [ "Ç", "Ê", "Î", "Ş", "Û" ], // not in libicu |
| 177 | 'ky' => [ "Ё" ], |
| 178 | 'la' => [], // not in libicu |
| 179 | 'lb' => [], |
| 180 | 'lkt' => [ 'Č', 'Ǧ', 'Ȟ', 'Š', 'Ž' ], |
| 181 | 'ln' => [ 'Ɛ' ], |
| 182 | 'lo' => [], |
| 183 | 'lt' => [ "Č", "Š", "Ž" ], |
| 184 | 'lv' => [ "Č", "Ģ", "Ķ", "Ļ", "Ņ", "Š", "Ž" ], |
| 185 | 'mk' => [ "Ѓ", "Ќ" ], |
| 186 | 'ml' => [], |
| 187 | 'mn' => [], |
| 188 | 'mo' => [ "Ă", "Â", "Î", "Ș", "Ț" ], // not in libicu |
| 189 | 'mr' => [ "\u{0902}", "\u{0903}", "ळ", "क्ष", "ज्ञ" ], |
| 190 | 'ms' => [], |
| 191 | 'mt' => [ "Ċ", "Ġ", "Għ", "Ħ", "Ż" ], |
| 192 | 'nb' => [ "Æ", "Ø", "Å" ], |
| 193 | 'ne' => [], |
| 194 | 'nl' => [], |
| 195 | 'nn' => [ "Æ", "Ø", "Å" ], |
| 196 | 'no' => [ "Æ", "Ø", "Å" ], // not in libicu. You should probably use nb or nn instead. |
| 197 | 'oc' => [], // not in libicu |
| 198 | 'om' => [ 'Ch', 'Dh', 'Kh', 'Ny', 'Ph', 'Sh' ], |
| 199 | 'or' => [ "\u{0B01}", "\u{0B02}", "\u{0B03}", "କ୍ଷ" ], |
| 200 | 'pa' => [ "\u{0A4D}" ], |
| 201 | 'pl' => [ "Ą", "Ć", "Ę", "Ł", "Ń", "Ó", "Ś", "Ź", "Ż" ], |
| 202 | 'pt' => [], |
| 203 | 'rm' => [], // not in libicu |
| 204 | 'ro' => [ "Ă", "Â", "Î", "Ș", "Ț" ], |
| 205 | 'ru' => [], |
| 206 | 'rup' => [ "Ă", "Â", "Î", "Ľ", "Ń", "Ș", "Ț" ], // not in libicu |
| 207 | 'sco' => [], |
| 208 | 'se' => [ |
| 209 | 'Á', 'Č', 'Ʒ', 'Ǯ', 'Đ', 'Ǧ', 'Ǥ', 'Ǩ', 'Ŋ', |
| 210 | 'Š', 'Ŧ', 'Ž', 'Ø', 'Æ', 'Ȧ', 'Ä', 'Ö' |
| 211 | ], |
| 212 | 'si' => [ "\u{0D82}", "\u{0D83}", "\u{0DA4}" ], |
| 213 | 'sk' => [ "Ä", "Č", "Ch", "Ô", "Š", "Ž" ], |
| 214 | 'sl' => [ "Č", "Š", "Ž" ], |
| 215 | 'smn' => [ "Â", "Á", "Č", "Đ", "Ŋ", "Š", "Ŧ", "Ž", "Æ", "Ø", "Å", "Ä", "Ö" ], |
| 216 | 'sq' => [ "Ç", "Dh", "Ë", "Gj", "Ll", "Nj", "Rr", "Sh", "Th", "Xh", "Zh" ], |
| 217 | 'sr' => [], |
| 218 | 'sr-Latn' => [ "Č", "Ć", "Dž", "Đ", "Lj", "Nj", "Š", "Ž" ], |
| 219 | 'sv' => [ "Å", "Ä", "Ö" ], |
| 220 | 'sv@collation=standard' => [ "Å", "Ä", "Ö" ], |
| 221 | 'sw' => [], |
| 222 | 'ta' => [ |
| 223 | "\u{0B82}", "ஃ", "க்ஷ", "க்", "ங்", "ச்", "ஞ்", "ட்", "ண்", "த்", "ந்", |
| 224 | "ப்", "ம்", "ய்", "ர்", "ல்", "வ்", "ழ்", "ள்", "ற்", "ன்", "ஜ்", "ஶ்", "ஷ்", |
| 225 | "ஸ்", "ஹ்", "க்ஷ்" |
| 226 | ], |
| 227 | 'te' => [ "\u{0C01}", "\u{0C02}", "\u{0C03}" ], |
| 228 | 'th' => [ "ฯ", "\u{0E46}", "\u{0E4D}", "\u{0E3A}" ], |
| 229 | 'tk' => [ "Ç", "Ä", "Ž", "Ň", "Ö", "Ş", "Ü", "Ý" ], |
| 230 | 'tl' => [ "Ñ", "Ng" ], // not in libicu |
| 231 | 'to' => [ "Ng", "ʻ" ], |
| 232 | 'tr' => [ "Ç", "Ğ", "İ", "Ö", "Ş", "Ü" ], |
| 233 | '-tr' => [ "ı" ], |
| 234 | 'tt' => [ "Ә", "Ө", "Ү", "Җ", "Ң", "Һ" ], // not in libicu |
| 235 | 'uk' => [ "Ґ", "Ь" ], |
| 236 | 'uz' => [ "Ch", "G'", "Ng", "O'", "Sh" ], // not in libicu |
| 237 | 'vi' => [ "Ă", "Â", "Đ", "Ê", "Ô", "Ơ", "Ư" ], |
| 238 | 'vo' => [ "Ä", "Ö", "Ü" ], |
| 239 | 'yi' => [ |
| 240 | "\u{05D1}\u{05BF}", "\u{05DB}\u{05BC}", "\u{05E4}\u{05BC}", |
| 241 | "\u{05E9}\u{05C2}", "\u{05EA}\u{05BC}" |
| 242 | ], |
| 243 | 'yo' => [ "Ẹ", "Gb", "Ọ", "Ṣ" ], |
| 244 | 'zu' => [], |
| 245 | ]; |
| 246 | |
| 247 | /** |
| 248 | * @param LanguageFactory $languageFactory |
| 249 | * @param string $locale |
| 250 | */ |
| 251 | public function __construct( |
| 252 | LanguageFactory $languageFactory, |
| 253 | $locale |
| 254 | ) { |
| 255 | $this->locale = $locale; |
| 256 | // Drop everything after the '@' in locale's name |
| 257 | $localeParts = explode( '@', $locale ); |
| 258 | $this->digitTransformLanguage = $languageFactory->getLanguage( $locale === 'root' ? 'en' : $localeParts[0] ); |
| 259 | |
| 260 | $mainCollator = Collator::create( $locale ); |
| 261 | if ( !$mainCollator ) { |
| 262 | throw new InvalidArgumentException( "Invalid ICU locale specified for collation: $locale" ); |
| 263 | } |
| 264 | $this->mainCollator = $mainCollator; |
| 265 | |
| 266 | // @phan-suppress-next-line PhanPossiblyNullTypeMismatchProperty successed before, no null check needed |
| 267 | $this->primaryCollator = Collator::create( $locale ); |
| 268 | $this->primaryCollator->setStrength( Collator::PRIMARY ); |
| 269 | |
| 270 | // If the special suffix for numeric collation is present, turn on numeric collation. |
| 271 | if ( str_ends_with( $locale, '-u-kn' ) ) { |
| 272 | $this->useNumericCollation = true; |
| 273 | // Strip off the special suffix so it doesn't trip up fetchFirstLetterData(). |
| 274 | $this->locale = substr( $this->locale, 0, -5 ); |
| 275 | $this->mainCollator->setAttribute( Collator::NUMERIC_COLLATION, Collator::ON ); |
| 276 | $this->primaryCollator->setAttribute( Collator::NUMERIC_COLLATION, Collator::ON ); |
| 277 | } |
| 278 | } |
| 279 | |
| 280 | /** @inheritDoc */ |
| 281 | public function getSortKey( $string ) { |
| 282 | return $this->mainCollator->getSortKey( $string ); |
| 283 | } |
| 284 | |
| 285 | /** @inheritDoc */ |
| 286 | public function getFirstLetter( $string ) { |
| 287 | $string = strval( $string ); |
| 288 | if ( $string === '' ) { |
| 289 | return ''; |
| 290 | } |
| 291 | |
| 292 | $firstChar = mb_substr( $string, 0, 1, 'UTF-8' ); |
| 293 | |
| 294 | // If the first character is a CJK character, just return that character. |
| 295 | if ( ord( $firstChar ) > 0x7f && self::isCjk( mb_ord( $firstChar ) ) ) { |
| 296 | return $firstChar; |
| 297 | } |
| 298 | |
| 299 | $sortKey = $this->getPrimarySortKey( $string ); |
| 300 | $data = $this->getFirstLetterData(); |
| 301 | $keys = $data['keys']; |
| 302 | $letters = $data['chars']; |
| 303 | |
| 304 | // Do a binary search to find the correct letter to sort under |
| 305 | $min = ArrayUtils::findLowerBound( |
| 306 | static function ( $index ) use ( $keys ) { |
| 307 | return $keys[$index]; |
| 308 | }, |
| 309 | count( $keys ), |
| 310 | 'strcmp', |
| 311 | $sortKey ); |
| 312 | |
| 313 | if ( $min === false ) { |
| 314 | // Before the first letter |
| 315 | return ''; |
| 316 | } |
| 317 | |
| 318 | $sortLetter = $letters[$min]; |
| 319 | |
| 320 | if ( $this->useNumericCollation ) { |
| 321 | // If the sort letter is a number, return '0–9' (or localized equivalent). |
| 322 | // ASCII value of 0 is 48. ASCII value of 9 is 57. |
| 323 | // Note that this also applies to non-Arabic numerals since they are |
| 324 | // mapped to Arabic numeral sort letters. For example, ২ sorts as 2. |
| 325 | if ( ord( $sortLetter ) >= 48 && ord( $sortLetter ) <= 57 ) { |
| 326 | $sortLetter = wfMessage( 'category-header-numerals' )->numParams( 0, 9 )->text(); |
| 327 | } |
| 328 | } |
| 329 | return $sortLetter; |
| 330 | } |
| 331 | |
| 332 | private function getPrimarySortKey( string $string ): string { |
| 333 | return $this->primaryCollator->getSortKey( $string ); |
| 334 | } |
| 335 | |
| 336 | /** |
| 337 | * @since 1.16.3 |
| 338 | * @return array |
| 339 | */ |
| 340 | private function getFirstLetterData() { |
| 341 | if ( $this->firstLetterData === null ) { |
| 342 | $cache = MediaWikiServices::getInstance()->getObjectCacheFactory() |
| 343 | ->getLocalServerInstance( CACHE_ANYTHING ); |
| 344 | $cacheKey = $cache->makeKey( |
| 345 | 'first-letters', |
| 346 | static::class, |
| 347 | $this->locale, |
| 348 | $this->digitTransformLanguage->getCode(), |
| 349 | INTL_ICU_VERSION, |
| 350 | self::FIRST_LETTER_VERSION |
| 351 | ); |
| 352 | $this->firstLetterData = $cache->getWithSetCallback( $cacheKey, $cache::TTL_WEEK, function () { |
| 353 | return $this->fetchFirstLetterData(); |
| 354 | } ); |
| 355 | } |
| 356 | return $this->firstLetterData; |
| 357 | } |
| 358 | |
| 359 | /** |
| 360 | * @return array |
| 361 | */ |
| 362 | private function fetchFirstLetterData() { |
| 363 | // Generate data from serialized data file |
| 364 | if ( isset( self::TAILORING_FIRST_LETTERS[$this->locale] ) ) { |
| 365 | $letters = require __DIR__ . "/data/first-letters-root.php"; |
| 366 | // Append additional characters |
| 367 | $letters = array_merge( $letters, self::TAILORING_FIRST_LETTERS[$this->locale] ); |
| 368 | // Remove unnecessary ones, if any |
| 369 | if ( isset( self::TAILORING_FIRST_LETTERS['-' . $this->locale] ) ) { |
| 370 | $letters = array_diff( $letters, self::TAILORING_FIRST_LETTERS['-' . $this->locale] ); |
| 371 | } |
| 372 | // Apply digit transforms |
| 373 | $digits = [ '0', '1', '2', '3', '4', '5', '6', '7', '8', '9' ]; |
| 374 | $letters = array_diff( $letters, $digits ); |
| 375 | foreach ( $digits as $digit ) { |
| 376 | $letters[] = $this->digitTransformLanguage->formatNumNoSeparators( $digit ); |
| 377 | } |
| 378 | } elseif ( $this->locale === 'root' ) { |
| 379 | $letters = require __DIR__ . "/data/first-letters-root.php"; |
| 380 | } else { |
| 381 | throw new RuntimeException( "MediaWiki does not support ICU locale " . |
| 382 | "\"{$this->locale}\"" ); |
| 383 | } |
| 384 | |
| 385 | /* Sort the letters. |
| 386 | * |
| 387 | * It's impossible to have the precompiled data file properly sorted, |
| 388 | * because the sort order changes depending on ICU version. If the |
| 389 | * array is not properly sorted, the binary search will return random |
| 390 | * results. |
| 391 | * |
| 392 | * We also take this opportunity to remove primary collisions. |
| 393 | */ |
| 394 | $letterMap = []; |
| 395 | foreach ( $letters as $letter ) { |
| 396 | $key = $this->getPrimarySortKey( $letter ); |
| 397 | if ( isset( $letterMap[$key] ) ) { |
| 398 | // Primary collision (two characters with the same sort position). |
| 399 | // Keep whichever one sorts first in the main collator. |
| 400 | $comp = $this->mainCollator->compare( $letter, $letterMap[$key] ); |
| 401 | wfDebug( "Primary collision '$letter' '{$letterMap[$key]}' (comparison: $comp)" ); |
| 402 | // If that also has a collision, use codepoint as a tiebreaker. |
| 403 | if ( $comp === 0 ) { |
| 404 | $comp = mb_ord( $letter ) <=> mb_ord( $letterMap[$key] ); |
| 405 | } |
| 406 | if ( $comp < 0 ) { |
| 407 | $letterMap[$key] = $letter; |
| 408 | } |
| 409 | } else { |
| 410 | $letterMap[$key] = $letter; |
| 411 | } |
| 412 | } |
| 413 | ksort( $letterMap, SORT_STRING ); |
| 414 | |
| 415 | /* Remove duplicate prefixes. Basically, if something has a sortkey |
| 416 | * which is a prefix of some other sortkey, then it is an |
| 417 | * expansion and probably should not be considered a section |
| 418 | * header. |
| 419 | * |
| 420 | * For example, 'þ' is sometimes sorted as if it is the letters |
| 421 | * 'th'. Other times it is its own primary element. Another |
| 422 | * example is '₨'. Sometimes it's a currency symbol. Sometimes it |
| 423 | * is an 'R' followed by an 's'. |
| 424 | * |
| 425 | * Additionally, an expanded element should always sort directly |
| 426 | * after its first element due to the way sortkeys work. |
| 427 | * |
| 428 | * UCA sortkey elements are of variable length, but no collation |
| 429 | * element should be a prefix of some other element, so I think |
| 430 | * this is safe. See: |
| 431 | * - https://unicode-org.github.io/icu-docs/design/collation/ICU_collation_design.htm |
| 432 | * - https://icu.unicode.org/design/collation/uca-weight-allocation |
| 433 | * |
| 434 | * Additionally, there is something called primary compression to |
| 435 | * worry about. Basically, if you have two primary elements that |
| 436 | * are more than one byte and both start with the same byte, then |
| 437 | * the first byte is dropped on the second primary. Additionally, |
| 438 | * either \x03 or \xFF may be added to mean that the next primary |
| 439 | * does not start with the first byte of the first primary. |
| 440 | * |
| 441 | * This shouldn't matter much, as the first primary is not |
| 442 | * changed, and that is what we are comparing against. |
| 443 | * |
| 444 | * tl;dr: This makes some assumptions about how ICU implements |
| 445 | * collations. It seems incredibly unlikely these assumptions |
| 446 | * will change, but nonetheless they are assumptions. |
| 447 | */ |
| 448 | |
| 449 | $prev = ''; |
| 450 | $duplicatePrefixes = []; |
| 451 | foreach ( $letterMap as $key => $value ) { |
| 452 | // Due to the fact the array is sorted, we only have |
| 453 | // to compare with the element directly previous |
| 454 | // to the current element (skipping expansions). |
| 455 | // An element "X" will always sort directly |
| 456 | // before "XZ" (Unless we have "XY", but we |
| 457 | // do not update $prev in that case). |
| 458 | if ( $prev !== '' && str_starts_with( $key, $prev ) ) { |
| 459 | $duplicatePrefixes[] = $key; |
| 460 | // If this is an expansion, we don't want to |
| 461 | // compare the next element to this element, |
| 462 | // but to what is currently $prev |
| 463 | continue; |
| 464 | } |
| 465 | $prev = $key; |
| 466 | } |
| 467 | foreach ( $duplicatePrefixes as $badKey ) { |
| 468 | wfDebug( "Removing '{$letterMap[$badKey]}' from first letters." ); |
| 469 | unset( $letterMap[$badKey] ); |
| 470 | // This code assumes that unsetting does not change sort order. |
| 471 | } |
| 472 | return [ |
| 473 | 'chars' => array_values( $letterMap ), |
| 474 | 'keys' => array_keys( $letterMap ), |
| 475 | ]; |
| 476 | } |
| 477 | |
| 478 | /** |
| 479 | * Test if a code point is a CJK (Chinese, Japanese, Korean) character |
| 480 | * @param int $codepoint |
| 481 | * @return bool |
| 482 | * @since 1.16.3 |
| 483 | */ |
| 484 | public static function isCjk( $codepoint ) { |
| 485 | foreach ( self::CJK_BLOCKS as $block ) { |
| 486 | if ( $codepoint >= $block[0] && $codepoint <= $block[1] ) { |
| 487 | return true; |
| 488 | } |
| 489 | } |
| 490 | return false; |
| 491 | } |
| 492 | } |