languages/Language.php

   1 <?php
   2 /**
   3  * Internationalisation code.
   4  * See https://www.mediawiki.org/wiki/Special:MyLanguage/Localisation for more information.
   5  *
   6  * This program is free software; you can redistribute it and/or modify
   7  * it under the terms of the GNU General Public License as published by
   8  * the Free Software Foundation; either version 2 of the License, or
   9  * (at your option) any later version.
  10  *
  11  * This program is distributed in the hope that it will be useful,
  12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  14  * GNU General Public License for more details.
  15  *
  16  * You should have received a copy of the GNU General Public License along
  17  * with this program; if not, write to the Free Software Foundation, Inc.,
  18  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  19  * http://www.gnu.org/copyleft/gpl.html
  20  *
  21  * @file
  22  * @ingroup Language
  23  */
  24
  25 /**
  26  * @defgroup Language Language
  27  */
  28
  29 use CLDRPluralRuleParser\Evaluator;
  30 use MediaWiki\MediaWikiServices;
  31 use Wikimedia\Assert\Assert;
  32
  33 /**
  34  * Internationalisation code
  35  * @ingroup Language
  36  */
  37 class Language {
  38         /**
  39          * Return autonyms in fetchLanguageName(s).
  40          * @since 1.32
  41          */
  42         const AS_AUTONYMS = null;
  43
  44         /**
  45          * Return all known languages in fetchLanguageName(s).
  46          * @since 1.32
  47          */
  48         const ALL = 'all';
  49
  50         /**
  51          * Return in fetchLanguageName(s) only the languages for which we have at
  52          * least some localisation.
  53          * @since 1.32
  54          */
  55         const SUPPORTED = 'mwfile';
  56
  57         /**
  58          * @var LanguageConverter
  59          */
  60         public $mConverter;
  61
  62         public $mVariants, $mCode, $mLoaded = false;
  63         public $mMagicExtensions = [];
  64         private $mHtmlCode = null, $mParentLanguage = false;
  65
  66         public $dateFormatStrings = [];
  67         public $mExtendedSpecialPageAliases;
  68
  69         /** @var array|null */
  70         protected $namespaceNames;
  71         protected $mNamespaceIds, $namespaceAliases;
  72
  73         /**
  74          * ReplacementArray object caches
  75          */
  76         public $transformData = [];
  77
  78         /**
  79          * @var LocalisationCache
  80          */
  81         public static $dataCache;
  82
  83         public static $mLangObjCache = [];
  84
  85         /**
  86          * Return a fallback chain for messages in getFallbacksFor
  87          * @since 1.32
  88          */
  89         const MESSAGES_FALLBACKS = 0;
  90
  91         /**
  92          * Return a strict fallback chain in getFallbacksFor
  93          * @since 1.32
  94          */
  95         const STRICT_FALLBACKS = 1;
  96
  97         public static $mWeekdayMsgs = [
  98                 'sunday', 'monday', 'tuesday', 'wednesday', 'thursday',
  99                 'friday', 'saturday'
 100         ];
 101
 102         public static $mWeekdayAbbrevMsgs = [
 103                 'sun', 'mon', 'tue', 'wed', 'thu', 'fri', 'sat'
 104         ];
 105
 106         public static $mMonthMsgs = [
 107                 'january', 'february', 'march', 'april', 'may_long', 'june',
 108                 'july', 'august', 'september', 'october', 'november',
 109                 'december'
 110         ];
 111         public static $mMonthGenMsgs = [
 112                 'january-gen', 'february-gen', 'march-gen', 'april-gen', 'may-gen', 'june-gen',
 113                 'july-gen', 'august-gen', 'september-gen', 'october-gen', 'november-gen',
 114                 'december-gen'
 115         ];
 116         public static $mMonthAbbrevMsgs = [
 117                 'jan', 'feb', 'mar', 'apr', 'may', 'jun', 'jul', 'aug',
 118                 'sep', 'oct', 'nov', 'dec'
 119         ];
 120
 121         public static $mIranianCalendarMonthMsgs = [
 122                 'iranian-calendar-m1', 'iranian-calendar-m2', 'iranian-calendar-m3',
 123                 'iranian-calendar-m4', 'iranian-calendar-m5', 'iranian-calendar-m6',
 124                 'iranian-calendar-m7', 'iranian-calendar-m8', 'iranian-calendar-m9',
 125                 'iranian-calendar-m10', 'iranian-calendar-m11', 'iranian-calendar-m12'
 126         ];
 127
 128         public static $mHebrewCalendarMonthMsgs = [
 129                 'hebrew-calendar-m1', 'hebrew-calendar-m2', 'hebrew-calendar-m3',
 130                 'hebrew-calendar-m4', 'hebrew-calendar-m5', 'hebrew-calendar-m6',
 131                 'hebrew-calendar-m7', 'hebrew-calendar-m8', 'hebrew-calendar-m9',
 132                 'hebrew-calendar-m10', 'hebrew-calendar-m11', 'hebrew-calendar-m12',
 133                 'hebrew-calendar-m6a', 'hebrew-calendar-m6b'
 134         ];
 135
 136         public static $mHebrewCalendarMonthGenMsgs = [
 137                 'hebrew-calendar-m1-gen', 'hebrew-calendar-m2-gen', 'hebrew-calendar-m3-gen',
 138                 'hebrew-calendar-m4-gen', 'hebrew-calendar-m5-gen', 'hebrew-calendar-m6-gen',
 139                 'hebrew-calendar-m7-gen', 'hebrew-calendar-m8-gen', 'hebrew-calendar-m9-gen',
 140                 'hebrew-calendar-m10-gen', 'hebrew-calendar-m11-gen', 'hebrew-calendar-m12-gen',
 141                 'hebrew-calendar-m6a-gen', 'hebrew-calendar-m6b-gen'
 142         ];
 143
 144         public static $mHijriCalendarMonthMsgs = [
 145                 'hijri-calendar-m1', 'hijri-calendar-m2', 'hijri-calendar-m3',
 146                 'hijri-calendar-m4', 'hijri-calendar-m5', 'hijri-calendar-m6',
 147                 'hijri-calendar-m7', 'hijri-calendar-m8', 'hijri-calendar-m9',
 148                 'hijri-calendar-m10', 'hijri-calendar-m11', 'hijri-calendar-m12'
 149         ];
 150
 151         /**
 152          * @since 1.20
 153          * @var array
 154          */
 155         public static $durationIntervals = [
 156                 'millennia' => 31556952000,
 157                 'centuries' => 3155695200,
 158                 'decades' => 315569520,
 159                 'years' => 31556952, // 86400 * ( 365 + ( 24 * 3 + 25 ) / 400 )
 160                 'weeks' => 604800,
 161                 'days' => 86400,
 162                 'hours' => 3600,
 163                 'minutes' => 60,
 164                 'seconds' => 1,
 165         ];
 166
 167         /**
 168          * Cache for language fallbacks.
 169          * @see Language::getFallbacksIncludingSiteLanguage
 170          * @since 1.21
 171          * @var array
 172          */
 173         private static $fallbackLanguageCache = [];
 174
 175         /**
 176          * Cache for grammar rules data
 177          * @var MapCacheLRU|null
 178          */
 179         private static $grammarTransformations;
 180
 181         /**
 182          * Cache for language names
 183          * @var HashBagOStuff|null
 184          */
 185         private static $languageNameCache;
 186
 187         /**
 188          * Unicode directional formatting characters, for embedBidi()
 189          */
 190         private static $lre = "\u{202A}"; // U+202A LEFT-TO-RIGHT EMBEDDING
 191         private static $rle = "\u{202B}"; // U+202B RIGHT-TO-LEFT EMBEDDING
 192         private static $pdf = "\u{202C}"; // U+202C POP DIRECTIONAL FORMATTING
 193
 194         /**
 195          * Directionality test regex for embedBidi(). Matches the first strong directionality codepoint:
 196          * - in group 1 if it is LTR
 197          * - in group 2 if it is RTL
 198          * Does not match if there is no strong directionality codepoint.
 199          *
 200          * The form is '/(?:([strong ltr codepoint])|([strong rtl codepoint]))/u' .
 201          *
 202          * Generated by UnicodeJS (see tools/strongDir) from the UCD; see
 203          * https://phabricator.wikimedia.org/diffusion/GUJS/ .
 204          */
 205         // @codeCoverageIgnoreStart
 206         // phpcs:ignore Generic.Files.LineLength
 207         private static $strongDirRegex = '/(?:([\x{41}-\x{5a}\x{61}-\x{7a}\x{aa}\x{b5}\x{ba}\x{c0}-\x{d6}\x{d8}-\x{f6}\x{f8}-\x{2b8}\x{2bb}-\x{2c1}\x{2d0}\x{2d1}\x{2e0}-\x{2e4}\x{2ee}\x{370}-\x{373}\x{376}\x{377}\x{37a}-\x{37d}\x{37f}\x{386}\x{388}-\x{38a}\x{38c}\x{38e}-\x{3a1}\x{3a3}-\x{3f5}\x{3f7}-\x{482}\x{48a}-\x{52f}\x{531}-\x{556}\x{559}-\x{55f}\x{561}-\x{587}\x{589}\x{903}-\x{939}\x{93b}\x{93d}-\x{940}\x{949}-\x{94c}\x{94e}-\x{950}\x{958}-\x{961}\x{964}-\x{980}\x{982}\x{983}\x{985}-\x{98c}\x{98f}\x{990}\x{993}-\x{9a8}\x{9aa}-\x{9b0}\x{9b2}\x{9b6}-\x{9b9}\x{9bd}-\x{9c0}\x{9c7}\x{9c8}\x{9cb}\x{9cc}\x{9ce}\x{9d7}\x{9dc}\x{9dd}\x{9df}-\x{9e1}\x{9e6}-\x{9f1}\x{9f4}-\x{9fa}\x{a03}\x{a05}-\x{a0a}\x{a0f}\x{a10}\x{a13}-\x{a28}\x{a2a}-\x{a30}\x{a32}\x{a33}\x{a35}\x{a36}\x{a38}\x{a39}\x{a3e}-\x{a40}\x{a59}-\x{a5c}\x{a5e}\x{a66}-\x{a6f}\x{a72}-\x{a74}\x{a83}\x{a85}-\x{a8d}\x{a8f}-\x{a91}\x{a93}-\x{aa8}\x{aaa}-\x{ab0}\x{ab2}\x{ab3}\x{ab5}-\x{ab9}\x{abd}-\x{ac0}\x{ac9}\x{acb}\x{acc}\x{ad0}\x{ae0}\x{ae1}\x{ae6}-\x{af0}\x{af9}\x{b02}\x{b03}\x{b05}-\x{b0c}\x{b0f}\x{b10}\x{b13}-\x{b28}\x{b2a}-\x{b30}\x{b32}\x{b33}\x{b35}-\x{b39}\x{b3d}\x{b3e}\x{b40}\x{b47}\x{b48}\x{b4b}\x{b4c}\x{b57}\x{b5c}\x{b5d}\x{b5f}-\x{b61}\x{b66}-\x{b77}\x{b83}\x{b85}-\x{b8a}\x{b8e}-\x{b90}\x{b92}-\x{b95}\x{b99}\x{b9a}\x{b9c}\x{b9e}\x{b9f}\x{ba3}\x{ba4}\x{ba8}-\x{baa}\x{bae}-\x{bb9}\x{bbe}\x{bbf}\x{bc1}\x{bc2}\x{bc6}-\x{bc8}\x{bca}-\x{bcc}\x{bd0}\x{bd7}\x{be6}-\x{bf2}\x{c01}-\x{c03}\x{c05}-\x{c0c}\x{c0e}-\x{c10}\x{c12}-\x{c28}\x{c2a}-\x{c39}\x{c3d}\x{c41}-\x{c44}\x{c58}-\x{c5a}\x{c60}\x{c61}\x{c66}-\x{c6f}\x{c7f}\x{c82}\x{c83}\x{c85}-\x{c8c}\x{c8e}-\x{c90}\x{c92}-\x{ca8}\x{caa}-\x{cb3}\x{cb5}-\x{cb9}\x{cbd}-\x{cc4}\x{cc6}-\x{cc8}\x{cca}\x{ccb}\x{cd5}\x{cd6}\x{cde}\x{ce0}\x{ce1}\x{ce6}-\x{cef}\x{cf1}\x{cf2}\x{d02}\x{d03}\x{d05}-\x{d0c}\x{d0e}-\x{d10}\x{d12}-\x{d3a}\x{d3d}-\x{d40}\x{d46}-\x{d48}\x{d4a}-\x{d4c}\x{d4e}\x{d57}\x{d5f}-\x{d61}\x{d66}-\x{d75}\x{d79}-\x{d7f}\x{d82}\x{d83}\x{d85}-\x{d96}\x{d9a}-\x{db1}\x{db3}-\x{dbb}\x{dbd}\x{dc0}-\x{dc6}\x{dcf}-\x{dd1}\x{dd8}-\x{ddf}\x{de6}-\x{def}\x{df2}-\x{df4}\x{e01}-\x{e30}\x{e32}\x{e33}\x{e40}-\x{e46}\x{e4f}-\x{e5b}\x{e81}\x{e82}\x{e84}\x{e87}\x{e88}\x{e8a}\x{e8d}\x{e94}-\x{e97}\x{e99}-\x{e9f}\x{ea1}-\x{ea3}\x{ea5}\x{ea7}\x{eaa}\x{eab}\x{ead}-\x{eb0}\x{eb2}\x{eb3}\x{ebd}\x{ec0}-\x{ec4}\x{ec6}\x{ed0}-\x{ed9}\x{edc}-\x{edf}\x{f00}-\x{f17}\x{f1a}-\x{f34}\x{f36}\x{f38}\x{f3e}-\x{f47}\x{f49}-\x{f6c}\x{f7f}\x{f85}\x{f88}-\x{f8c}\x{fbe}-\x{fc5}\x{fc7}-\x{fcc}\x{fce}-\x{fda}\x{1000}-\x{102c}\x{1031}\x{1038}\x{103b}\x{103c}\x{103f}-\x{1057}\x{105a}-\x{105d}\x{1061}-\x{1070}\x{1075}-\x{1081}\x{1083}\x{1084}\x{1087}-\x{108c}\x{108e}-\x{109c}\x{109e}-\x{10c5}\x{10c7}\x{10cd}\x{10d0}-\x{1248}\x{124a}-\x{124d}\x{1250}-\x{1256}\x{1258}\x{125a}-\x{125d}\x{1260}-\x{1288}\x{128a}-\x{128d}\x{1290}-\x{12b0}\x{12b2}-\x{12b5}\x{12b8}-\x{12be}\x{12c0}\x{12c2}-\x{12c5}\x{12c8}-\x{12d6}\x{12d8}-\x{1310}\x{1312}-\x{1315}\x{1318}-\x{135a}\x{1360}-\x{137c}\x{1380}-\x{138f}\x{13a0}-\x{13f5}\x{13f8}-\x{13fd}\x{1401}-\x{167f}\x{1681}-\x{169a}\x{16a0}-\x{16f8}\x{1700}-\x{170c}\x{170e}-\x{1711}\x{1720}-\x{1731}\x{1735}\x{1736}\x{1740}-\x{1751}\x{1760}-\x{176c}\x{176e}-\x{1770}\x{1780}-\x{17b3}\x{17b6}\x{17be}-\x{17c5}\x{17c7}\x{17c8}\x{17d4}-\x{17da}\x{17dc}\x{17e0}-\x{17e9}\x{1810}-\x{1819}\x{1820}-\x{1877}\x{1880}-\x{18a8}\x{18aa}\x{18b0}-\x{18f5}\x{1900}-\x{191e}\x{1923}-\x{1926}\x{1929}-\x{192b}\x{1930}\x{1931}\x{1933}-\x{1938}\x{1946}-\x{196d}\x{1970}-\x{1974}\x{1980}-\x{19ab}\x{19b0}-\x{19c9}\x{19d0}-\x{19da}\x{1a00}-\x{1a16}\x{1a19}\x{1a1a}\x{1a1e}-\x{1a55}\x{1a57}\x{1a61}\x{1a63}\x{1a64}\x{1a6d}-\x{1a72}\x{1a80}-\x{1a89}\x{1a90}-\x{1a99}\x{1aa0}-\x{1aad}\x{1b04}-\x{1b33}\x{1b35}\x{1b3b}\x{1b3d}-\x{1b41}\x{1b43}-\x{1b4b}\x{1b50}-\x{1b6a}\x{1b74}-\x{1b7c}\x{1b82}-\x{1ba1}\x{1ba6}\x{1ba7}\x{1baa}\x{1bae}-\x{1be5}\x{1be7}\x{1bea}-\x{1bec}\x{1bee}\x{1bf2}\x{1bf3}\x{1bfc}-\x{1c2b}\x{1c34}\x{1c35}\x{1c3b}-\x{1c49}\x{1c4d}-\x{1c7f}\x{1cc0}-\x{1cc7}\x{1cd3}\x{1ce1}\x{1ce9}-\x{1cec}\x{1cee}-\x{1cf3}\x{1cf5}\x{1cf6}\x{1d00}-\x{1dbf}\x{1e00}-\x{1f15}\x{1f18}-\x{1f1d}\x{1f20}-\x{1f45}\x{1f48}-\x{1f4d}\x{1f50}-\x{1f57}\x{1f59}\x{1f5b}\x{1f5d}\x{1f5f}-\x{1f7d}\x{1f80}-\x{1fb4}\x{1fb6}-\x{1fbc}\x{1fbe}\x{1fc2}-\x{1fc4}\x{1fc6}-\x{1fcc}\x{1fd0}-\x{1fd3}\x{1fd6}-\x{1fdb}\x{1fe0}-\x{1fec}\x{1ff2}-\x{1ff4}\x{1ff6}-\x{1ffc}\x{200e}\x{2071}\x{207f}\x{2090}-\x{209c}\x{2102}\x{2107}\x{210a}-\x{2113}\x{2115}\x{2119}-\x{211d}\x{2124}\x{2126}\x{2128}\x{212a}-\x{212d}\x{212f}-\x{2139}\x{213c}-\x{213f}\x{2145}-\x{2149}\x{214e}\x{214f}\x{2160}-\x{2188}\x{2336}-\x{237a}\x{2395}\x{249c}-\x{24e9}\x{26ac}\x{2800}-\x{28ff}\x{2c00}-\x{2c2e}\x{2c30}-\x{2c5e}\x{2c60}-\x{2ce4}\x{2ceb}-\x{2cee}\x{2cf2}\x{2cf3}\x{2d00}-\x{2d25}\x{2d27}\x{2d2d}\x{2d30}-\x{2d67}\x{2d6f}\x{2d70}\x{2d80}-\x{2d96}\x{2da0}-\x{2da6}\x{2da8}-\x{2dae}\x{2db0}-\x{2db6}\x{2db8}-\x{2dbe}\x{2dc0}-\x{2dc6}\x{2dc8}-\x{2dce}\x{2dd0}-\x{2dd6}\x{2dd8}-\x{2dde}\x{3005}-\x{3007}\x{3021}-\x{3029}\x{302e}\x{302f}\x{3031}-\x{3035}\x{3038}-\x{303c}\x{3041}-\x{3096}\x{309d}-\x{309f}\x{30a1}-\x{30fa}\x{30fc}-\x{30ff}\x{3105}-\x{312d}\x{3131}-\x{318e}\x{3190}-\x{31ba}\x{31f0}-\x{321c}\x{3220}-\x{324f}\x{3260}-\x{327b}\x{327f}-\x{32b0}\x{32c0}-\x{32cb}\x{32d0}-\x{32fe}\x{3300}-\x{3376}\x{337b}-\x{33dd}\x{33e0}-\x{33fe}\x{3400}-\x{4db5}\x{4e00}-\x{9fd5}\x{a000}-\x{a48c}\x{a4d0}-\x{a60c}\x{a610}-\x{a62b}\x{a640}-\x{a66e}\x{a680}-\x{a69d}\x{a6a0}-\x{a6ef}\x{a6f2}-\x{a6f7}\x{a722}-\x{a787}\x{a789}-\x{a7ad}\x{a7b0}-\x{a7b7}\x{a7f7}-\x{a801}\x{a803}-\x{a805}\x{a807}-\x{a80a}\x{a80c}-\x{a824}\x{a827}\x{a830}-\x{a837}\x{a840}-\x{a873}\x{a880}-\x{a8c3}\x{a8ce}-\x{a8d9}\x{a8f2}-\x{a8fd}\x{a900}-\x{a925}\x{a92e}-\x{a946}\x{a952}\x{a953}\x{a95f}-\x{a97c}\x{a983}-\x{a9b2}\x{a9b4}\x{a9b5}\x{a9ba}\x{a9bb}\x{a9bd}-\x{a9cd}\x{a9cf}-\x{a9d9}\x{a9de}-\x{a9e4}\x{a9e6}-\x{a9fe}\x{aa00}-\x{aa28}\x{aa2f}\x{aa30}\x{aa33}\x{aa34}\x{aa40}-\x{aa42}\x{aa44}-\x{aa4b}\x{aa4d}\x{aa50}-\x{aa59}\x{aa5c}-\x{aa7b}\x{aa7d}-\x{aaaf}\x{aab1}\x{aab5}\x{aab6}\x{aab9}-\x{aabd}\x{aac0}\x{aac2}\x{aadb}-\x{aaeb}\x{aaee}-\x{aaf5}\x{ab01}-\x{ab06}\x{ab09}-\x{ab0e}\x{ab11}-\x{ab16}\x{ab20}-\x{ab26}\x{ab28}-\x{ab2e}\x{ab30}-\x{ab65}\x{ab70}-\x{abe4}\x{abe6}\x{abe7}\x{abe9}-\x{abec}\x{abf0}-\x{abf9}\x{ac00}-\x{d7a3}\x{d7b0}-\x{d7c6}\x{d7cb}-\x{d7fb}\x{e000}-\x{fa6d}\x{fa70}-\x{fad9}\x{fb00}-\x{fb06}\x{fb13}-\x{fb17}\x{ff21}-\x{ff3a}\x{ff41}-\x{ff5a}\x{ff66}-\x{ffbe}\x{ffc2}-\x{ffc7}\x{ffca}-\x{ffcf}\x{ffd2}-\x{ffd7}\x{ffda}-\x{ffdc}\x{10000}-\x{1000b}\x{1000d}-\x{10026}\x{10028}-\x{1003a}\x{1003c}\x{1003d}\x{1003f}-\x{1004d}\x{10050}-\x{1005d}\x{10080}-\x{100fa}\x{10100}\x{10102}\x{10107}-\x{10133}\x{10137}-\x{1013f}\x{101d0}-\x{101fc}\x{10280}-\x{1029c}\x{102a0}-\x{102d0}\x{10300}-\x{10323}\x{10330}-\x{1034a}\x{10350}-\x{10375}\x{10380}-\x{1039d}\x{1039f}-\x{103c3}\x{103c8}-\x{103d5}\x{10400}-\x{1049d}\x{104a0}-\x{104a9}\x{10500}-\x{10527}\x{10530}-\x{10563}\x{1056f}\x{10600}-\x{10736}\x{10740}-\x{10755}\x{10760}-\x{10767}\x{11000}\x{11002}-\x{11037}\x{11047}-\x{1104d}\x{11066}-\x{1106f}\x{11082}-\x{110b2}\x{110b7}\x{110b8}\x{110bb}-\x{110c1}\x{110d0}-\x{110e8}\x{110f0}-\x{110f9}\x{11103}-\x{11126}\x{1112c}\x{11136}-\x{11143}\x{11150}-\x{11172}\x{11174}-\x{11176}\x{11182}-\x{111b5}\x{111bf}-\x{111c9}\x{111cd}\x{111d0}-\x{111df}\x{111e1}-\x{111f4}\x{11200}-\x{11211}\x{11213}-\x{1122e}\x{11232}\x{11233}\x{11235}\x{11238}-\x{1123d}\x{11280}-\x{11286}\x{11288}\x{1128a}-\x{1128d}\x{1128f}-\x{1129d}\x{1129f}-\x{112a9}\x{112b0}-\x{112de}\x{112e0}-\x{112e2}\x{112f0}-\x{112f9}\x{11302}\x{11303}\x{11305}-\x{1130c}\x{1130f}\x{11310}\x{11313}-\x{11328}\x{1132a}-\x{11330}\x{11332}\x{11333}\x{11335}-\x{11339}\x{1133d}-\x{1133f}\x{11341}-\x{11344}\x{11347}\x{11348}\x{1134b}-\x{1134d}\x{11350}\x{11357}\x{1135d}-\x{11363}\x{11480}-\x{114b2}\x{114b9}\x{114bb}-\x{114be}\x{114c1}\x{114c4}-\x{114c7}\x{114d0}-\x{114d9}\x{11580}-\x{115b1}\x{115b8}-\x{115bb}\x{115be}\x{115c1}-\x{115db}\x{11600}-\x{11632}\x{1163b}\x{1163c}\x{1163e}\x{11641}-\x{11644}\x{11650}-\x{11659}\x{11680}-\x{116aa}\x{116ac}\x{116ae}\x{116af}\x{116b6}\x{116c0}-\x{116c9}\x{11700}-\x{11719}\x{11720}\x{11721}\x{11726}\x{11730}-\x{1173f}\x{118a0}-\x{118f2}\x{118ff}\x{11ac0}-\x{11af8}\x{12000}-\x{12399}\x{12400}-\x{1246e}\x{12470}-\x{12474}\x{12480}-\x{12543}\x{13000}-\x{1342e}\x{14400}-\x{14646}\x{16800}-\x{16a38}\x{16a40}-\x{16a5e}\x{16a60}-\x{16a69}\x{16a6e}\x{16a6f}\x{16ad0}-\x{16aed}\x{16af5}\x{16b00}-\x{16b2f}\x{16b37}-\x{16b45}\x{16b50}-\x{16b59}\x{16b5b}-\x{16b61}\x{16b63}-\x{16b77}\x{16b7d}-\x{16b8f}\x{16f00}-\x{16f44}\x{16f50}-\x{16f7e}\x{16f93}-\x{16f9f}\x{1b000}\x{1b001}\x{1bc00}-\x{1bc6a}\x{1bc70}-\x{1bc7c}\x{1bc80}-\x{1bc88}\x{1bc90}-\x{1bc99}\x{1bc9c}\x{1bc9f}\x{1d000}-\x{1d0f5}\x{1d100}-\x{1d126}\x{1d129}-\x{1d166}\x{1d16a}-\x{1d172}\x{1d183}\x{1d184}\x{1d18c}-\x{1d1a9}\x{1d1ae}-\x{1d1e8}\x{1d360}-\x{1d371}\x{1d400}-\x{1d454}\x{1d456}-\x{1d49c}\x{1d49e}\x{1d49f}\x{1d4a2}\x{1d4a5}\x{1d4a6}\x{1d4a9}-\x{1d4ac}\x{1d4ae}-\x{1d4b9}\x{1d4bb}\x{1d4bd}-\x{1d4c3}\x{1d4c5}-\x{1d505}\x{1d507}-\x{1d50a}\x{1d50d}-\x{1d514}\x{1d516}-\x{1d51c}\x{1d51e}-\x{1d539}\x{1d53b}-\x{1d53e}\x{1d540}-\x{1d544}\x{1d546}\x{1d54a}-\x{1d550}\x{1d552}-\x{1d6a5}\x{1d6a8}-\x{1d6da}\x{1d6dc}-\x{1d714}\x{1d716}-\x{1d74e}\x{1d750}-\x{1d788}\x{1d78a}-\x{1d7c2}\x{1d7c4}-\x{1d7cb}\x{1d800}-\x{1d9ff}\x{1da37}-\x{1da3a}\x{1da6d}-\x{1da74}\x{1da76}-\x{1da83}\x{1da85}-\x{1da8b}\x{1f110}-\x{1f12e}\x{1f130}-\x{1f169}\x{1f170}-\x{1f19a}\x{1f1e6}-\x{1f202}\x{1f210}-\x{1f23a}\x{1f240}-\x{1f248}\x{1f250}\x{1f251}\x{20000}-\x{2a6d6}\x{2a700}-\x{2b734}\x{2b740}-\x{2b81d}\x{2b820}-\x{2cea1}\x{2f800}-\x{2fa1d}\x{f0000}-\x{ffffd}\x{100000}-\x{10fffd}])|([\x{590}\x{5be}\x{5c0}\x{5c3}\x{5c6}\x{5c8}-\x{5ff}\x{7c0}-\x{7ea}\x{7f4}\x{7f5}\x{7fa}-\x{815}\x{81a}\x{824}\x{828}\x{82e}-\x{858}\x{85c}-\x{89f}\x{200f}\x{fb1d}\x{fb1f}-\x{fb28}\x{fb2a}-\x{fb4f}\x{10800}-\x{1091e}\x{10920}-\x{10a00}\x{10a04}\x{10a07}-\x{10a0b}\x{10a10}-\x{10a37}\x{10a3b}-\x{10a3e}\x{10a40}-\x{10ae4}\x{10ae7}-\x{10b38}\x{10b40}-\x{10e5f}\x{10e7f}-\x{10fff}\x{1e800}-\x{1e8cf}\x{1e8d7}-\x{1edff}\x{1ef00}-\x{1efff}\x{608}\x{60b}\x{60d}\x{61b}-\x{64a}\x{66d}-\x{66f}\x{671}-\x{6d5}\x{6e5}\x{6e6}\x{6ee}\x{6ef}\x{6fa}-\x{710}\x{712}-\x{72f}\x{74b}-\x{7a5}\x{7b1}-\x{7bf}\x{8a0}-\x{8e2}\x{fb50}-\x{fd3d}\x{fd40}-\x{fdcf}\x{fdf0}-\x{fdfc}\x{fdfe}\x{fdff}\x{fe70}-\x{fefe}\x{1ee00}-\x{1eeef}\x{1eef2}-\x{1eeff}]))/u';
 208         // @codeCoverageIgnoreEnd
 209
 210         /**
 211          * Get a cached or new language object for a given language code
 212          * @param string $code
 213          * @throws MWException
 214          * @return Language
 215          */
 216         static function factory( $code ) {
 217                 global $wgDummyLanguageCodes, $wgLangObjCacheSize;
 218
 219                 if ( isset( $wgDummyLanguageCodes[$code] ) ) {
 220                         $code = $wgDummyLanguageCodes[$code];
 221                 }
 222
 223                 // get the language object to process
 224                 $langObj = self::$mLangObjCache[$code] ?? self::newFromCode( $code );
 225
 226                 // merge the language object in to get it up front in the cache
 227                 self::$mLangObjCache = array_merge( [ $code => $langObj ], self::$mLangObjCache );
 228                 // get rid of the oldest ones in case we have an overflow
 229                 self::$mLangObjCache = array_slice( self::$mLangObjCache, 0, $wgLangObjCacheSize, true );
 230
 231                 return $langObj;
 232         }
 233
 234         /**
 235          * Create a language object for a given language code
 236          * @param string $code
 237          * @param bool $fallback Whether we're going through language fallback chain
 238          * @throws MWException
 239          * @return Language
 240          */
 241         protected static function newFromCode( $code, $fallback = false ) {
 242                 if ( !self::isValidCode( $code ) ) {
 243                         throw new MWException( "Invalid language code \"$code\"" );
 244                 }
 245
 246                 if ( !self::isValidBuiltInCode( $code ) ) {
 247                         // It's not possible to customise this code with class files, so
 248                         // just return a Language object. This is to support uselang= hacks.
 249                         $lang = new Language;
 250                         $lang->mCode = $code;
 251                         return $lang;
 252                 }
 253
 254                 // Check if there is a language class for the code
 255                 $class = self::classFromCode( $code, $fallback );
 256                 // LanguageCode does not inherit Language
 257                 if ( class_exists( $class ) && is_a( $class, 'Language', true ) ) {
 258                         $lang = new $class;
 259                         return $lang;
 260                 }
 261
 262                 // Keep trying the fallback list until we find an existing class
 263                 $fallbacks = self::getFallbacksFor( $code );
 264                 foreach ( $fallbacks as $fallbackCode ) {
 265                         if ( !self::isValidBuiltInCode( $fallbackCode ) ) {
 266                                 throw new MWException( "Invalid fallback '$fallbackCode' in fallback sequence for '$code'" );
 267                         }
 268
 269                         $class = self::classFromCode( $fallbackCode );
 270                         if ( class_exists( $class ) ) {
 271                                 $lang = new $class;
 272                                 $lang->mCode = $code;
 273                                 return $lang;
 274                         }
 275                 }
 276
 277                 throw new MWException( "Invalid fallback sequence for language '$code'" );
 278         }
 279
 280         /**
 281          * Intended for tests that may change configuration in a way that invalidates caches.
 282          *
 283          * @since 1.32
 284          */
 285         public static function clearCaches() {
 286                 if ( !defined( 'MW_PHPUNIT_TEST' ) ) {
 287                         throw new MWException( __METHOD__ . ' must not be used outside tests' );
 288                 }
 289                 self::$dataCache = null;
 290                 // Reinitialize $dataCache, since it's expected to always be available
 291                 self::getLocalisationCache();
 292                 self::$mLangObjCache = [];
 293                 self::$fallbackLanguageCache = [];
 294                 self::$grammarTransformations = null;
 295                 self::$languageNameCache = null;
 296         }
 297
 298         /**
 299          * Checks whether any localisation is available for that language tag
 300          * in MediaWiki (MessagesXx.php exists).
 301          *
 302          * @param string $code Language tag (in lower case)
 303          * @return bool Whether language is supported
 304          * @since 1.21
 305          */
 306         public static function isSupportedLanguage( $code ) {
 307                 if ( !self::isValidBuiltInCode( $code ) ) {
 308                         return false;
 309                 }
 310
 311                 if ( $code === 'qqq' ) {
 312                         return false;
 313                 }
 314
 315                 return is_readable( self::getMessagesFileName( $code ) ) ||
 316                         is_readable( self::getJsonMessagesFileName( $code ) );
 317         }
 318
 319         /**
 320          * Returns true if a language code string is a well-formed language tag
 321          * according to RFC 5646.
 322          * This function only checks well-formedness; it doesn't check that
 323          * language, script or variant codes actually exist in the repositories.
 324          *
 325          * Based on regexes by Mark Davis of the Unicode Consortium:
 326          * https://www.unicode.org/repos/cldr/trunk/tools/java/org/unicode/cldr/util/data/langtagRegex.txt
 327          *
 328          * @param string $code
 329          * @param bool $lenient Whether to allow '_' as separator. The default is only '-'.
 330          *
 331          * @return bool
 332          * @since 1.21
 333          */
 334         public static function isWellFormedLanguageTag( $code, $lenient = false ) {
 335                 $alpha = '[a-z]';
 336                 $digit = '[0-9]';
 337                 $alphanum = '[a-z0-9]';
 338                 $x = 'x'; # private use singleton
 339                 $singleton = '[a-wy-z]'; # other singleton
 340                 $s = $lenient ? '[-_]' : '-';
 341
 342                 $language = "$alpha{2,8}|$alpha{2,3}$s$alpha{3}";
 343                 $script = "$alpha{4}"; # ISO 15924
 344                 $region = "(?:$alpha{2}|$digit{3})"; # ISO 3166-1 alpha-2 or UN M.49
 345                 $variant = "(?:$alphanum{5,8}|$digit$alphanum{3})";
 346                 $extension = "$singleton(?:$s$alphanum{2,8})+";
 347                 $privateUse = "$x(?:$s$alphanum{1,8})+";
 348
 349                 # Define certain grandfathered codes, since otherwise the regex is pretty useless.
 350                 # Since these are limited, this is safe even later changes to the registry --
 351                 # the only oddity is that it might change the type of the tag, and thus
 352                 # the results from the capturing groups.
 353                 # https://www.iana.org/assignments/language-subtag-registry
 354
 355                 $grandfathered = "en{$s}GB{$s}oed"
 356                         . "|i{$s}(?:ami|bnn|default|enochian|hak|klingon|lux|mingo|navajo|pwn|tao|tay|tsu)"
 357                         . "|no{$s}(?:bok|nyn)"
 358                         . "|sgn{$s}(?:BE{$s}(?:fr|nl)|CH{$s}de)"
 359                         . "|zh{$s}min{$s}nan";
 360
 361                 $variantList = "$variant(?:$s$variant)*";
 362                 $extensionList = "$extension(?:$s$extension)*";
 363
 364                 $langtag = "(?:($language)"
 365                         . "(?:$s$script)?"
 366                         . "(?:$s$region)?"
 367                         . "(?:$s$variantList)?"
 368                         . "(?:$s$extensionList)?"
 369                         . "(?:$s$privateUse)?)";
 370
 371                 # The final breakdown, with capturing groups for each of these components
 372                 # The variants, extensions, grandfathered, and private-use may have interior '-'
 373
 374                 $root = "^(?:$langtag|$privateUse|$grandfathered)$";
 375
 376                 return (bool)preg_match( "/$root/", strtolower( $code ) );
 377         }
 378
 379         /**
 380          * Returns true if a language code string is of a valid form, whether or
 381          * not it exists. This includes codes which are used solely for
 382          * customisation via the MediaWiki namespace.
 383          *
 384          * @param string $code
 385          *
 386          * @return bool
 387          */
 388         public static function isValidCode( $code ) {
 389                 static $cache = [];
 390                 Assert::parameterType( 'string', $code, '$code' );
 391                 if ( !isset( $cache[$code] ) ) {
 392                         // People think language codes are html safe, so enforce it.
 393                         // Ideally we should only allow a-zA-Z0-9-
 394                         // but, .+ and other chars are often used for {{int:}} hacks
 395                         // see bugs T39564, T39587, T38938
 396                         $cache[$code] =
 397                                 // Protect against path traversal
 398                                 strcspn( $code, ":/\\\000&<>'\"" ) === strlen( $code )
 399                                 && !preg_match( MediaWikiTitleCodec::getTitleInvalidRegex(), $code );
 400                 }
 401                 return $cache[$code];
 402         }
 403
 404         /**
 405          * Returns true if a language code is of a valid form for the purposes of
 406          * internal customisation of MediaWiki, via Messages*.php or *.json.
 407          *
 408          * @param string $code
 409          *
 410          * @since 1.18
 411          * @return bool
 412          */
 413         public static function isValidBuiltInCode( $code ) {
 414                 Assert::parameterType( 'string', $code, '$code' );
 415
 416                 return (bool)preg_match( '/^[a-z0-9-]{2,}$/', $code );
 417         }
 418
 419         /**
 420          * Returns true if a language code is an IETF tag known to MediaWiki.
 421          *
 422          * @param string $tag
 423          *
 424          * @since 1.21
 425          * @return bool
 426          */
 427         public static function isKnownLanguageTag( $tag ) {
 428                 // Quick escape for invalid input to avoid exceptions down the line
 429                 // when code tries to process tags which are not valid at all.
 430                 if ( !self::isValidBuiltInCode( $tag ) ) {
 431                         return false;
 432                 }
 433
 434                 if ( isset( MediaWiki\Languages\Data\Names::$names[$tag] )
 435                         || self::fetchLanguageName( $tag, $tag ) !== ''
 436                 ) {
 437                         return true;
 438                 }
 439
 440                 return false;
 441         }
 442
 443         /**
 444          * Get the LocalisationCache instance
 445          *
 446          * @return LocalisationCache
 447          */
 448         public static function getLocalisationCache() {
 449                 if ( is_null( self::$dataCache ) ) {
 450                         global $wgLocalisationCacheConf;
 451                         $class = $wgLocalisationCacheConf['class'];
 452                         self::$dataCache = new $class( $wgLocalisationCacheConf );
 453                 }
 454                 return self::$dataCache;
 455         }
 456
 457         function __construct() {
 458                 $this->mConverter = new FakeConverter( $this );
 459                 // Set the code to the name of the descendant
 460                 if ( static::class === 'Language' ) {
 461                         $this->mCode = 'en';
 462                 } else {
 463                         $this->mCode = str_replace( '_', '-', strtolower( substr( static::class, 8 ) ) );
 464                 }
 465                 self::getLocalisationCache();
 466         }
 467
 468         /**
 469          * Reduce memory usage
 470          */
 471         function __destruct() {
 472                 foreach ( $this as $name => $value ) {
 473                         unset( $this->$name );
 474                 }
 475         }
 476
 477         /**
 478          * Hook which will be called if this is the content language.
 479          * Descendants can use this to register hook functions or modify globals
 480          */
 481         function initContLang() {
 482         }
 483
 484         /**
 485          * @return array
 486          * @since 1.19
 487          */
 488         public function getFallbackLanguages() {
 489                 return self::getFallbacksFor( $this->mCode );
 490         }
 491
 492         /**
 493          * Exports $wgBookstoreListEn
 494          * @return array
 495          */
 496         public function getBookstoreList() {
 497                 return self::$dataCache->getItem( $this->mCode, 'bookstoreList' );
 498         }
 499
 500         /**
 501          * Returns an array of localised namespaces indexed by their numbers. If the namespace is not
 502          * available in localised form, it will be included in English.
 503          *
 504          * @return string[] List of localized namespace names, indexed by numeric namespace ID.
 505          */
 506         public function getNamespaces() {
 507                 if ( is_null( $this->namespaceNames ) ) {
 508                         global $wgMetaNamespace, $wgMetaNamespaceTalk, $wgExtraNamespaces;
 509
 510                         $validNamespaces = MediaWikiServices::getInstance()->getNamespaceInfo()->
 511                                 getCanonicalNamespaces();
 512
 513                         $this->namespaceNames = $wgExtraNamespaces +
 514                                 self::$dataCache->getItem( $this->mCode, 'namespaceNames' );
 515                         $this->namespaceNames += $validNamespaces;
 516
 517                         $this->namespaceNames[NS_PROJECT] = $wgMetaNamespace;
 518                         if ( $wgMetaNamespaceTalk ) {
 519                                 $this->namespaceNames[NS_PROJECT_TALK] = $wgMetaNamespaceTalk;
 520                         } else {
 521                                 $talk = $this->namespaceNames[NS_PROJECT_TALK];
 522                                 $this->namespaceNames[NS_PROJECT_TALK] =
 523                                         $this->fixVariableInNamespace( $talk );
 524                         }
 525
 526                         # Sometimes a language will be localised but not actually exist on this wiki.
 527                         foreach ( $this->namespaceNames as $key => $text ) {
 528                                 if ( !isset( $validNamespaces[$key] ) ) {
 529                                         unset( $this->namespaceNames[$key] );
 530                                 }
 531                         }
 532
 533                         # The above mixing may leave namespaces out of canonical order.
 534                         # Re-order by namespace ID number...
 535                         ksort( $this->namespaceNames );
 536
 537                         Hooks::run( 'LanguageGetNamespaces', [ &$this->namespaceNames ] );
 538                 }
 539
 540                 return $this->namespaceNames;
 541         }
 542
 543         /**
 544          * Arbitrarily set all of the namespace names at once. Mainly used for testing
 545          * @param array $namespaces Array of namespaces (id => name)
 546          */
 547         public function setNamespaces( array $namespaces ) {
 548                 $this->namespaceNames = $namespaces;
 549                 $this->mNamespaceIds = null;
 550         }
 551
 552         /**
 553          * Resets all of the namespace caches. Mainly used for testing
 554          */
 555         public function resetNamespaces() {
 556                 $this->namespaceNames = null;
 557                 $this->mNamespaceIds = null;
 558                 $this->namespaceAliases = null;
 559         }
 560
 561         /**
 562          * A convenience function that returns getNamespaces() with spaces instead of underscores
 563          * in values. Useful for producing output to be displayed e.g. in `<select>` forms.
 564          *
 565          * @return array
 566          */
 567         public function getFormattedNamespaces() {
 568                 $ns = $this->getNamespaces();
 569                 foreach ( $ns as $k => $v ) {
 570                         $ns[$k] = strtr( $v, '_', ' ' );
 571                 }
 572                 return $ns;
 573         }
 574
 575         /**
 576          * Get a namespace value by key
 577          *
 578          * <code>
 579          * $mw_ns = $lang->getNsText( NS_MEDIAWIKI );
 580          * echo $mw_ns; // prints 'MediaWiki'
 581          * </code>
 582          *
 583          * @param int $index The array key of the namespace to return
 584          * @return string|bool String if the namespace value exists, otherwise false
 585          */
 586         public function getNsText( $index ) {
 587                 $ns = $this->getNamespaces();
 588                 return $ns[$index] ?? false;
 589         }
 590
 591         /**
 592          * A convenience function that returns the same thing as
 593          * getNsText() except with '_' changed to ' ', useful for
 594          * producing output.
 595          *
 596          * <code>
 597          * $mw_ns = $lang->getFormattedNsText( NS_MEDIAWIKI_TALK );
 598          * echo $mw_ns; // prints 'MediaWiki talk'
 599          * </code>
 600          *
 601          * @param int $index The array key of the namespace to return
 602          * @return string Namespace name without underscores (empty string if namespace does not exist)
 603          */
 604         public function getFormattedNsText( $index ) {
 605                 $ns = $this->getNsText( $index );
 606                 return strtr( $ns, '_', ' ' );
 607         }
 608
 609         /**
 610          * Returns gender-dependent namespace alias if available.
 611          * See https://www.mediawiki.org/wiki/Manual:$wgExtraGenderNamespaces
 612          * @param int $index Namespace index
 613          * @param string $gender Gender key (male, female... )
 614          * @return string
 615          * @since 1.18
 616          */
 617         public function getGenderNsText( $index, $gender ) {
 618                 global $wgExtraGenderNamespaces;
 619
 620                 $ns = $wgExtraGenderNamespaces +
 621                         (array)self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 622
 623                 return $ns[$index][$gender] ?? $this->getNsText( $index );
 624         }
 625
 626         /**
 627          * Whether this language uses gender-dependent namespace aliases.
 628          * See https://www.mediawiki.org/wiki/Manual:$wgExtraGenderNamespaces
 629          * @return bool
 630          * @since 1.18
 631          */
 632         public function needsGenderDistinction() {
 633                 global $wgExtraGenderNamespaces, $wgExtraNamespaces;
 634                 if ( count( $wgExtraGenderNamespaces ) > 0 ) {
 635                         // $wgExtraGenderNamespaces overrides everything
 636                         return true;
 637                 } elseif ( isset( $wgExtraNamespaces[NS_USER] ) && isset( $wgExtraNamespaces[NS_USER_TALK] ) ) {
 638                         /// @todo There may be other gender namespace than NS_USER & NS_USER_TALK in the future
 639                         // $wgExtraNamespaces overrides any gender aliases specified in i18n files
 640                         return false;
 641                 } else {
 642                         // Check what is in i18n files
 643                         $aliases = self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 644                         return count( $aliases ) > 0;
 645                 }
 646         }
 647
 648         /**
 649          * Get a namespace key by value, case insensitive.
 650          * Only matches namespace names for the current language, not the
 651          * canonical ones defined in Namespace.php.
 652          *
 653          * @param string $text
 654          * @return int|bool An integer if $text is a valid value otherwise false
 655          */
 656         function getLocalNsIndex( $text ) {
 657                 $lctext = $this->lc( $text );
 658                 $ids = $this->getNamespaceIds();
 659                 return $ids[$lctext] ?? false;
 660         }
 661
 662         /**
 663          * @return array
 664          */
 665         public function getNamespaceAliases() {
 666                 if ( is_null( $this->namespaceAliases ) ) {
 667                         $aliases = self::$dataCache->getItem( $this->mCode, 'namespaceAliases' );
 668                         if ( !$aliases ) {
 669                                 $aliases = [];
 670                         } else {
 671                                 foreach ( $aliases as $name => $index ) {
 672                                         if ( $index === NS_PROJECT_TALK ) {
 673                                                 unset( $aliases[$name] );
 674                                                 $name = $this->fixVariableInNamespace( $name );
 675                                                 $aliases[$name] = $index;
 676                                         }
 677                                 }
 678                         }
 679
 680                         global $wgExtraGenderNamespaces;
 681                         $genders = $wgExtraGenderNamespaces +
 682                                 (array)self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 683                         foreach ( $genders as $index => $forms ) {
 684                                 foreach ( $forms as $alias ) {
 685                                         $aliases[$alias] = $index;
 686                                 }
 687                         }
 688
 689                         # Also add converted namespace names as aliases, to avoid confusion.
 690                         $convertedNames = [];
 691                         foreach ( $this->getVariants() as $variant ) {
 692                                 if ( $variant === $this->mCode ) {
 693                                         continue;
 694                                 }
 695                                 foreach ( $this->getNamespaces() as $ns => $_ ) {
 696                                         $convertedNames[$this->getConverter()->convertNamespace( $ns, $variant )] = $ns;
 697                                 }
 698                         }
 699
 700                         $this->namespaceAliases = $aliases + $convertedNames;
 701
 702                         # Filter out aliases to namespaces that don't exist, e.g. from extensions
 703                         # that aren't loaded here but are included in the l10n cache.
 704                         # (array_intersect preserves keys from its first argument)
 705                         $this->namespaceAliases = array_intersect(
 706                                 $this->namespaceAliases,
 707                                 array_keys( $this->getNamespaces() )
 708                         );
 709                 }
 710
 711                 return $this->namespaceAliases;
 712         }
 713
 714         /**
 715          * @return array
 716          */
 717         public function getNamespaceIds() {
 718                 if ( is_null( $this->mNamespaceIds ) ) {
 719                         global $wgNamespaceAliases;
 720                         # Put namespace names and aliases into a hashtable.
 721                         # If this is too slow, then we should arrange it so that it is done
 722                         # before caching. The catch is that at pre-cache time, the above
 723                         # class-specific fixup hasn't been done.
 724                         $this->mNamespaceIds = [];
 725                         foreach ( $this->getNamespaces() as $index => $name ) {
 726                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 727                         }
 728                         foreach ( $this->getNamespaceAliases() as $name => $index ) {
 729                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 730                         }
 731                         if ( $wgNamespaceAliases ) {
 732                                 foreach ( $wgNamespaceAliases as $name => $index ) {
 733                                         $this->mNamespaceIds[$this->lc( $name )] = $index;
 734                                 }
 735                         }
 736                 }
 737                 return $this->mNamespaceIds;
 738         }
 739
 740         /**
 741          * Get a namespace key by value, case insensitive.  Canonical namespace
 742          * names override custom ones defined for the current language.
 743          *
 744          * @param string $text
 745          * @return int|bool An integer if $text is a valid value otherwise false
 746          */
 747         public function getNsIndex( $text ) {
 748                 $lctext = $this->lc( $text );
 749                 $ns = MediaWikiServices::getInstance()->getNamespaceInfo()->
 750                         getCanonicalIndex( $lctext );
 751                 if ( $ns !== null ) {
 752                         return $ns;
 753                 }
 754                 $ids = $this->getNamespaceIds();
 755                 return $ids[$lctext] ?? false;
 756         }
 757
 758         /**
 759          * short names for language variants used for language conversion links.
 760          *
 761          * @param string $code
 762          * @param bool $usemsg Use the "variantname-xyz" message if it exists
 763          * @return string
 764          */
 765         public function getVariantname( $code, $usemsg = true ) {
 766                 $msg = "variantname-$code";
 767                 if ( $usemsg && wfMessage( $msg )->exists() ) {
 768                         return $this->getMessageFromDB( $msg );
 769                 }
 770                 $name = self::fetchLanguageName( $code );
 771                 if ( $name ) {
 772                         return $name; # if it's defined as a language name, show that
 773                 } else {
 774                         # otherwise, output the language code
 775                         return $code;
 776                 }
 777         }
 778
 779         /**
 780          * @return string[]|bool List of date format preference keys, or false if disabled.
 781          */
 782         public function getDatePreferences() {
 783                 return self::$dataCache->getItem( $this->mCode, 'datePreferences' );
 784         }
 785
 786         /**
 787          * @return array
 788          */
 789         function getDateFormats() {
 790                 return self::$dataCache->getItem( $this->mCode, 'dateFormats' );
 791         }
 792
 793         /**
 794          * @return array|string
 795          */
 796         public function getDefaultDateFormat() {
 797                 $df = self::$dataCache->getItem( $this->mCode, 'defaultDateFormat' );
 798                 if ( $df === 'dmy or mdy' ) {
 799                         global $wgAmericanDates;
 800                         return $wgAmericanDates ? 'mdy' : 'dmy';
 801                 } else {
 802                         return $df;
 803                 }
 804         }
 805
 806         /**
 807          * @return array
 808          */
 809         public function getDatePreferenceMigrationMap() {
 810                 return self::$dataCache->getItem( $this->mCode, 'datePreferenceMigrationMap' );
 811         }
 812
 813         /**
 814          * @return array
 815          */
 816         public function getExtraUserToggles() {
 817                 return (array)self::$dataCache->getItem( $this->mCode, 'extraUserToggles' );
 818         }
 819
 820         /**
 821          * @param string $tog
 822          * @return string
 823          */
 824         function getUserToggle( $tog ) {
 825                 return $this->getMessageFromDB( "tog-$tog" );
 826         }
 827
 828         /**
 829          * Get an array of language names, indexed by code.
 830          * @param null|string $inLanguage Code of language in which to return the names
 831          *              Use self::AS_AUTONYMS for autonyms (native names)
 832          * @param string $include One of:
 833          *              self::ALL all available languages
 834          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 835          *              self::SUPPORTED only if the language is in 'mw' *and* has a message file
 836          * @return array Language code => language name (sorted by key)
 837          * @since 1.20
 838          */
 839         public static function fetchLanguageNames( $inLanguage = self::AS_AUTONYMS, $include = 'mw' ) {
 840                 $cacheKey = $inLanguage === self::AS_AUTONYMS ? 'null' : $inLanguage;
 841                 $cacheKey .= ":$include";
 842                 if ( self::$languageNameCache === null ) {
 843                         self::$languageNameCache = new HashBagOStuff( [ 'maxKeys' => 20 ] );
 844                 }
 845
 846                 $ret = self::$languageNameCache->get( $cacheKey );
 847                 if ( !$ret ) {
 848                         $ret = self::fetchLanguageNamesUncached( $inLanguage, $include );
 849                         self::$languageNameCache->set( $cacheKey, $ret );
 850                 }
 851                 return $ret;
 852         }
 853
 854         /**
 855          * Uncached helper for fetchLanguageNames
 856          * @param null|string $inLanguage Code of language in which to return the names
 857          *              Use self::AS_AUTONYMS for autonyms (native names)
 858          * @param string $include One of:
 859          *              self::ALL all available languages
 860          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 861          *              self::SUPPORTED only if the language is in 'mw' *and* has a message file
 862          * @return array Language code => language name (sorted by key)
 863          */
 864         private static function fetchLanguageNamesUncached(
 865                 $inLanguage = self::AS_AUTONYMS,
 866                 $include = 'mw'
 867         ) {
 868                 global $wgExtraLanguageNames, $wgUsePigLatinVariant;
 869
 870                 // If passed an invalid language code to use, fallback to en
 871                 if ( $inLanguage !== self::AS_AUTONYMS && !self::isValidCode( $inLanguage ) ) {
 872                         $inLanguage = 'en';
 873                 }
 874
 875                 $names = [];
 876
 877                 if ( $inLanguage ) {
 878                         # TODO: also include when $inLanguage is null, when this code is more efficient
 879                         Hooks::run( 'LanguageGetTranslatedLanguageNames', [ &$names, $inLanguage ] );
 880                 }
 881
 882                 $mwNames = $wgExtraLanguageNames + MediaWiki\Languages\Data\Names::$names;
 883                 if ( $wgUsePigLatinVariant ) {
 884                         // Pig Latin (for variant development)
 885                         $mwNames['en-x-piglatin'] = 'Igpay Atinlay';
 886                 }
 887
 888                 foreach ( $mwNames as $mwCode => $mwName ) {
 889                         # - Prefer own MediaWiki native name when not using the hook
 890                         # - For other names just add if not added through the hook
 891                         if ( $mwCode === $inLanguage || !isset( $names[$mwCode] ) ) {
 892                                 $names[$mwCode] = $mwName;
 893                         }
 894                 }
 895
 896                 if ( $include === self::ALL ) {
 897                         ksort( $names );
 898                         return $names;
 899                 }
 900
 901                 $returnMw = [];
 902                 $coreCodes = array_keys( $mwNames );
 903                 foreach ( $coreCodes as $coreCode ) {
 904                         $returnMw[$coreCode] = $names[$coreCode];
 905                 }
 906
 907                 if ( $include === self::SUPPORTED ) {
 908                         $namesMwFile = [];
 909                         # We do this using a foreach over the codes instead of a directory
 910                         # loop so that messages files in extensions will work correctly.
 911                         foreach ( $returnMw as $code => $value ) {
 912                                 if ( is_readable( self::getMessagesFileName( $code ) )
 913                                         || is_readable( self::getJsonMessagesFileName( $code ) )
 914                                 ) {
 915                                         $namesMwFile[$code] = $names[$code];
 916                                 }
 917                         }
 918
 919                         ksort( $namesMwFile );
 920                         return $namesMwFile;
 921                 }
 922
 923                 ksort( $returnMw );
 924                 # 'mw' option; default if it's not one of the other two options (all/mwfile)
 925                 return $returnMw;
 926         }
 927
 928         /**
 929          * @param string $code The code of the language for which to get the name
 930          * @param null|string $inLanguage Code of language in which to return the name
 931          *   (SELF::AS_AUTONYMS for autonyms)
 932          * @param string $include See fetchLanguageNames()
 933          * @return string Language name or empty
 934          * @since 1.20
 935          */
 936         public static function fetchLanguageName(
 937                 $code,
 938                 $inLanguage = self::AS_AUTONYMS,
 939                 $include = self::ALL
 940         ) {
 941                 $code = strtolower( $code );
 942                 $array = self::fetchLanguageNames( $inLanguage, $include );
 943                 return !array_key_exists( $code, $array ) ? '' : $array[$code];
 944         }
 945
 946         /**
 947          * Get a message from the MediaWiki namespace.
 948          *
 949          * @param string $msg Message name
 950          * @return string
 951          */
 952         public function getMessageFromDB( $msg ) {
 953                 return $this->msg( $msg )->text();
 954         }
 955
 956         /**
 957          * Get message object in this language. Only for use inside this class.
 958          *
 959          * @param string $msg Message name
 960          * @return Message
 961          */
 962         protected function msg( $msg ) {
 963                 return wfMessage( $msg )->inLanguage( $this );
 964         }
 965
 966         /**
 967          * @param string $key
 968          * @return string
 969          */
 970         public function getMonthName( $key ) {
 971                 return $this->getMessageFromDB( self::$mMonthMsgs[$key - 1] );
 972         }
 973
 974         /**
 975          * @return array
 976          */
 977         public function getMonthNamesArray() {
 978                 $monthNames = [ '' ];
 979                 for ( $i = 1; $i < 13; $i++ ) {
 980                         $monthNames[] = $this->getMonthName( $i );
 981                 }
 982                 return $monthNames;
 983         }
 984
 985         /**
 986          * @param string $key
 987          * @return string
 988          */
 989         public function getMonthNameGen( $key ) {
 990                 return $this->getMessageFromDB( self::$mMonthGenMsgs[$key - 1] );
 991         }
 992
 993         /**
 994          * @param string $key
 995          * @return string
 996          */
 997         public function getMonthAbbreviation( $key ) {
 998                 return $this->getMessageFromDB( self::$mMonthAbbrevMsgs[$key - 1] );
 999         }
1000
1001         /**
1002          * @return array
1003          */
1004         public function getMonthAbbreviationsArray() {
1005                 $monthNames = [ '' ];
1006                 for ( $i = 1; $i < 13; $i++ ) {
1007                         $monthNames[] = $this->getMonthAbbreviation( $i );
1008                 }
1009                 return $monthNames;
1010         }
1011
1012         /**
1013          * @param string $key
1014          * @return string
1015          */
1016         public function getWeekdayName( $key ) {
1017                 return $this->getMessageFromDB( self::$mWeekdayMsgs[$key - 1] );
1018         }
1019
1020         /**
1021          * @param string $key
1022          * @return string
1023          */
1024         function getWeekdayAbbreviation( $key ) {
1025                 return $this->getMessageFromDB( self::$mWeekdayAbbrevMsgs[$key - 1] );
1026         }
1027
1028         /**
1029          * @param string $key
1030          * @return string
1031          */
1032         function getIranianCalendarMonthName( $key ) {
1033                 return $this->getMessageFromDB( self::$mIranianCalendarMonthMsgs[$key - 1] );
1034         }
1035
1036         /**
1037          * @param string $key
1038          * @return string
1039          */
1040         function getHebrewCalendarMonthName( $key ) {
1041                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthMsgs[$key - 1] );
1042         }
1043
1044         /**
1045          * @param string $key
1046          * @return string
1047          */
1048         function getHebrewCalendarMonthNameGen( $key ) {
1049                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthGenMsgs[$key - 1] );
1050         }
1051
1052         /**
1053          * @param string $key
1054          * @return string
1055          */
1056         function getHijriCalendarMonthName( $key ) {
1057                 return $this->getMessageFromDB( self::$mHijriCalendarMonthMsgs[$key - 1] );
1058         }
1059
1060         /**
1061          * Pass through result from $dateTimeObj->format()
1062          * @param DateTime|bool|null &$dateTimeObj
1063          * @param string $ts
1064          * @param DateTimeZone|bool|null $zone
1065          * @param string $code
1066          * @return string
1067          */
1068         private static function dateTimeObjFormat( &$dateTimeObj, $ts, $zone, $code ) {
1069                 if ( !$dateTimeObj ) {
1070                         $dateTimeObj = DateTime::createFromFormat(
1071                                 'YmdHis', $ts, $zone ?: new DateTimeZone( 'UTC' )
1072                         );
1073                 }
1074                 return $dateTimeObj->format( $code );
1075         }
1076
1077         /**
1078          * This is a workalike of PHP's date() function, but with better
1079          * internationalisation, a reduced set of format characters, and a better
1080          * escaping format.
1081          *
1082          * Supported format characters are dDjlNwzWFmMntLoYyaAgGhHiscrUeIOPTZ. See
1083          * the PHP manual for definitions. There are a number of extensions, which
1084          * start with "x":
1085          *
1086          *    xn   Do not translate digits of the next numeric format character
1087          *    xN   Toggle raw digit (xn) flag, stays set until explicitly unset
1088          *    xr   Use roman numerals for the next numeric format character
1089          *    xh   Use hebrew numerals for the next numeric format character
1090          *    xx   Literal x
1091          *    xg   Genitive month name
1092          *
1093          *    xij  j (day number) in Iranian calendar
1094          *    xiF  F (month name) in Iranian calendar
1095          *    xin  n (month number) in Iranian calendar
1096          *    xiy  y (two digit year) in Iranian calendar
1097          *    xiY  Y (full year) in Iranian calendar
1098          *    xit  t (days in month) in Iranian calendar
1099          *    xiz  z (day of the year) in Iranian calendar
1100          *
1101          *    xjj  j (day number) in Hebrew calendar
1102          *    xjF  F (month name) in Hebrew calendar
1103          *    xjt  t (days in month) in Hebrew calendar
1104          *    xjx  xg (genitive month name) in Hebrew calendar
1105          *    xjn  n (month number) in Hebrew calendar
1106          *    xjY  Y (full year) in Hebrew calendar
1107          *
1108          *    xmj  j (day number) in Hijri calendar
1109          *    xmF  F (month name) in Hijri calendar
1110          *    xmn  n (month number) in Hijri calendar
1111          *    xmY  Y (full year) in Hijri calendar
1112          *
1113          *    xkY  Y (full year) in Thai solar calendar. Months and days are
1114          *                       identical to the Gregorian calendar
1115          *    xoY  Y (full year) in Minguo calendar or Juche year.
1116          *                       Months and days are identical to the
1117          *                       Gregorian calendar
1118          *    xtY  Y (full year) in Japanese nengo. Months and days are
1119          *                       identical to the Gregorian calendar
1120          *
1121          * Characters enclosed in double quotes will be considered literal (with
1122          * the quotes themselves removed). Unmatched quotes will be considered
1123          * literal quotes. Example:
1124          *
1125          * "The month is" F       => The month is January
1126          * i's"                   => 20'11"
1127          *
1128          * Backslash escaping is also supported.
1129          *
1130          * Input timestamp is assumed to be pre-normalized to the desired local
1131          * time zone, if any. Note that the format characters crUeIOPTZ will assume
1132          * $ts is UTC if $zone is not given.
1133          *
1134          * @param string $format
1135          * @param string $ts 14-character timestamp
1136          *      YYYYMMDDHHMMSS
1137          *      01234567890123
1138          * @param DateTimeZone|null $zone Timezone of $ts
1139          * @param int &$ttl The amount of time (in seconds) the output may be cached for.
1140          * Only makes sense if $ts is the current time.
1141          * @todo handling of "o" format character for Iranian, Hebrew, Hijri & Thai?
1142          *
1143          * @throws MWException
1144          * @return string
1145          */
1146         public function sprintfDate( $format, $ts, DateTimeZone $zone = null, &$ttl = 'unused' ) {
1147                 $s = '';
1148                 $raw = false;
1149                 $roman = false;
1150                 $hebrewNum = false;
1151                 $dateTimeObj = false;
1152                 $rawToggle = false;
1153                 $iranian = false;
1154                 $hebrew = false;
1155                 $hijri = false;
1156                 $thai = false;
1157                 $minguo = false;
1158                 $tenno = false;
1159
1160                 $usedSecond = false;
1161                 $usedMinute = false;
1162                 $usedHour = false;
1163                 $usedAMPM = false;
1164                 $usedDay = false;
1165                 $usedWeek = false;
1166                 $usedMonth = false;
1167                 $usedYear = false;
1168                 $usedISOYear = false;
1169                 $usedIsLeapYear = false;
1170
1171                 $usedHebrewMonth = false;
1172                 $usedIranianMonth = false;
1173                 $usedHijriMonth = false;
1174                 $usedHebrewYear = false;
1175                 $usedIranianYear = false;
1176                 $usedHijriYear = false;
1177                 $usedTennoYear = false;
1178
1179                 if ( strlen( $ts ) !== 14 ) {
1180                         throw new MWException( __METHOD__ . ": The timestamp $ts should have 14 characters" );
1181                 }
1182
1183                 if ( !ctype_digit( $ts ) ) {
1184                         throw new MWException( __METHOD__ . ": The timestamp $ts should be a number" );
1185                 }
1186
1187                 $formatLength = strlen( $format );
1188                 for ( $p = 0; $p < $formatLength; $p++ ) {
1189                         $num = false;
1190                         $code = $format[$p];
1191                         if ( $code == 'x' && $p < $formatLength - 1 ) {
1192                                 $code .= $format[++$p];
1193                         }
1194
1195                         if ( ( $code === 'xi'
1196                                         || $code === 'xj'
1197                                         || $code === 'xk'
1198                                         || $code === 'xm'
1199                                         || $code === 'xo'
1200                                         || $code === 'xt' )
1201                                 && $p < $formatLength - 1 ) {
1202                                 $code .= $format[++$p];
1203                         }
1204
1205                         switch ( $code ) {
1206                                 case 'xx':
1207                                         $s .= 'x';
1208                                         break;
1209                                 case 'xn':
1210                                         $raw = true;
1211                                         break;
1212                                 case 'xN':
1213                                         $rawToggle = !$rawToggle;
1214                                         break;
1215                                 case 'xr':
1216                                         $roman = true;
1217                                         break;
1218                                 case 'xh':
1219                                         $hebrewNum = true;
1220                                         break;
1221                                 case 'xg':
1222                                         $usedMonth = true;
1223                                         $s .= $this->getMonthNameGen( substr( $ts, 4, 2 ) );
1224                                         break;
1225                                 case 'xjx':
1226                                         $usedHebrewMonth = true;
1227                                         if ( !$hebrew ) {
1228                                                 $hebrew = self::tsToHebrew( $ts );
1229                                         }
1230                                         $s .= $this->getHebrewCalendarMonthNameGen( $hebrew[1] );
1231                                         break;
1232                                 case 'd':
1233                                         $usedDay = true;
1234                                         $num = substr( $ts, 6, 2 );
1235                                         break;
1236                                 case 'D':
1237                                         $usedDay = true;
1238                                         $s .= $this->getWeekdayAbbreviation(
1239                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1240                                         );
1241                                         break;
1242                                 case 'j':
1243                                         $usedDay = true;
1244                                         $num = intval( substr( $ts, 6, 2 ) );
1245                                         break;
1246                                 case 'xij':
1247                                         $usedDay = true;
1248                                         if ( !$iranian ) {
1249                                                 $iranian = self::tsToIranian( $ts );
1250                                         }
1251                                         $num = $iranian[2];
1252                                         break;
1253                                 case 'xmj':
1254                                         $usedDay = true;
1255                                         if ( !$hijri ) {
1256                                                 $hijri = self::tsToHijri( $ts );
1257                                         }
1258                                         $num = $hijri[2];
1259                                         break;
1260                                 case 'xjj':
1261                                         $usedDay = true;
1262                                         if ( !$hebrew ) {
1263                                                 $hebrew = self::tsToHebrew( $ts );
1264                                         }
1265                                         $num = $hebrew[2];
1266                                         break;
1267                                 case 'l':
1268                                         $usedDay = true;
1269                                         $s .= $this->getWeekdayName(
1270                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1271                                         );
1272                                         break;
1273                                 case 'F':
1274                                         $usedMonth = true;
1275                                         $s .= $this->getMonthName( substr( $ts, 4, 2 ) );
1276                                         break;
1277                                 case 'xiF':
1278                                         $usedIranianMonth = true;
1279                                         if ( !$iranian ) {
1280                                                 $iranian = self::tsToIranian( $ts );
1281                                         }
1282                                         $s .= $this->getIranianCalendarMonthName( $iranian[1] );
1283                                         break;
1284                                 case 'xmF':
1285                                         $usedHijriMonth = true;
1286                                         if ( !$hijri ) {
1287                                                 $hijri = self::tsToHijri( $ts );
1288                                         }
1289                                         $s .= $this->getHijriCalendarMonthName( $hijri[1] );
1290                                         break;
1291                                 case 'xjF':
1292                                         $usedHebrewMonth = true;
1293                                         if ( !$hebrew ) {
1294                                                 $hebrew = self::tsToHebrew( $ts );
1295                                         }
1296                                         $s .= $this->getHebrewCalendarMonthName( $hebrew[1] );
1297                                         break;
1298                                 case 'm':
1299                                         $usedMonth = true;
1300                                         $num = substr( $ts, 4, 2 );
1301                                         break;
1302                                 case 'M':
1303                                         $usedMonth = true;
1304                                         $s .= $this->getMonthAbbreviation( substr( $ts, 4, 2 ) );
1305                                         break;
1306                                 case 'n':
1307                                         $usedMonth = true;
1308                                         $num = intval( substr( $ts, 4, 2 ) );
1309                                         break;
1310                                 case 'xin':
1311                                         $usedIranianMonth = true;
1312                                         if ( !$iranian ) {
1313                                                 $iranian = self::tsToIranian( $ts );
1314                                         }
1315                                         $num = $iranian[1];
1316                                         break;
1317                                 case 'xmn':
1318                                         $usedHijriMonth = true;
1319                                         if ( !$hijri ) {
1320                                                 $hijri = self::tsToHijri( $ts );
1321                                         }
1322                                         $num = $hijri[1];
1323                                         break;
1324                                 case 'xjn':
1325                                         $usedHebrewMonth = true;
1326                                         if ( !$hebrew ) {
1327                                                 $hebrew = self::tsToHebrew( $ts );
1328                                         }
1329                                         $num = $hebrew[1];
1330                                         break;
1331                                 case 'xjt':
1332                                         $usedHebrewMonth = true;
1333                                         if ( !$hebrew ) {
1334                                                 $hebrew = self::tsToHebrew( $ts );
1335                                         }
1336                                         $num = $hebrew[3];
1337                                         break;
1338                                 case 'Y':
1339                                         $usedYear = true;
1340                                         $num = substr( $ts, 0, 4 );
1341                                         break;
1342                                 case 'xiY':
1343                                         $usedIranianYear = true;
1344                                         if ( !$iranian ) {
1345                                                 $iranian = self::tsToIranian( $ts );
1346                                         }
1347                                         $num = $iranian[0];
1348                                         break;
1349                                 case 'xmY':
1350                                         $usedHijriYear = true;
1351                                         if ( !$hijri ) {
1352                                                 $hijri = self::tsToHijri( $ts );
1353                                         }
1354                                         $num = $hijri[0];
1355                                         break;
1356                                 case 'xjY':
1357                                         $usedHebrewYear = true;
1358                                         if ( !$hebrew ) {
1359                                                 $hebrew = self::tsToHebrew( $ts );
1360                                         }
1361                                         $num = $hebrew[0];
1362                                         break;
1363                                 case 'xkY':
1364                                         $usedYear = true;
1365                                         if ( !$thai ) {
1366                                                 $thai = self::tsToYear( $ts, 'thai' );
1367                                         }
1368                                         $num = $thai[0];
1369                                         break;
1370                                 case 'xoY':
1371                                         $usedYear = true;
1372                                         if ( !$minguo ) {
1373                                                 $minguo = self::tsToYear( $ts, 'minguo' );
1374                                         }
1375                                         $num = $minguo[0];
1376                                         break;
1377                                 case 'xtY':
1378                                         $usedTennoYear = true;
1379                                         if ( !$tenno ) {
1380                                                 $tenno = self::tsToYear( $ts, 'tenno' );
1381                                         }
1382                                         $num = $tenno[0];
1383                                         break;
1384                                 case 'y':
1385                                         $usedYear = true;
1386                                         $num = substr( $ts, 2, 2 );
1387                                         break;
1388                                 case 'xiy':
1389                                         $usedIranianYear = true;
1390                                         if ( !$iranian ) {
1391                                                 $iranian = self::tsToIranian( $ts );
1392                                         }
1393                                         $num = substr( $iranian[0], -2 );
1394                                         break;
1395                                 case 'xit':
1396                                         $usedIranianYear = true;
1397                                         if ( !$iranian ) {
1398                                                 $iranian = self::tsToIranian( $ts );
1399                                         }
1400                                         $num = self::$IRANIAN_DAYS[$iranian[1] - 1];
1401                                         break;
1402                                 case 'xiz':
1403                                         $usedIranianYear = true;
1404                                         if ( !$iranian ) {
1405                                                 $iranian = self::tsToIranian( $ts );
1406                                         }
1407                                         $num = $iranian[3];
1408                                         break;
1409                                 case 'a':
1410                                         $usedAMPM = true;
1411                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'am' : 'pm';
1412                                         break;
1413                                 case 'A':
1414                                         $usedAMPM = true;
1415                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'AM' : 'PM';
1416                                         break;
1417                                 case 'g':
1418                                         $usedHour = true;
1419                                         $h = substr( $ts, 8, 2 );
1420                                         $num = $h % 12 ?: 12;
1421                                         break;
1422                                 case 'G':
1423                                         $usedHour = true;
1424                                         $num = intval( substr( $ts, 8, 2 ) );
1425                                         break;
1426                                 case 'h':
1427                                         $usedHour = true;
1428                                         $h = substr( $ts, 8, 2 );
1429                                         $num = sprintf( '%02d', $h % 12 ?: 12 );
1430                                         break;
1431                                 case 'H':
1432                                         $usedHour = true;
1433                                         $num = substr( $ts, 8, 2 );
1434                                         break;
1435                                 case 'i':
1436                                         $usedMinute = true;
1437                                         $num = substr( $ts, 10, 2 );
1438                                         break;
1439                                 case 's':
1440                                         $usedSecond = true;
1441                                         $num = substr( $ts, 12, 2 );
1442                                         break;
1443                                 case 'c':
1444                                 case 'r':
1445                                         $usedSecond = true;
1446                                         // fall through
1447                                 case 'e':
1448                                 case 'O':
1449                                 case 'P':
1450                                 case 'T':
1451                                         $s .= self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1452                                         break;
1453                                 case 'w':
1454                                 case 'N':
1455                                 case 'z':
1456                                         $usedDay = true;
1457                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1458                                         break;
1459                                 case 'W':
1460                                         $usedWeek = true;
1461                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1462                                         break;
1463                                 case 't':
1464                                         $usedMonth = true;
1465                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1466                                         break;
1467                                 case 'L':
1468                                         $usedIsLeapYear = true;
1469                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1470                                         break;
1471                                 case 'o':
1472                                         $usedISOYear = true;
1473                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1474                                         break;
1475                                 case 'U':
1476                                         $usedSecond = true;
1477                                         // fall through
1478                                 case 'I':
1479                                 case 'Z':
1480                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1481                                         break;
1482                                 case '\\':
1483                                         # Backslash escaping
1484                                         if ( $p < $formatLength - 1 ) {
1485                                                 $s .= $format[++$p];
1486                                         } else {
1487                                                 $s .= '\\';
1488                                         }
1489                                         break;
1490                                 case '"':
1491                                         # Quoted literal
1492                                         if ( $p < $formatLength - 1 ) {
1493                                                 $endQuote = strpos( $format, '"', $p + 1 );
1494                                                 if ( $endQuote === false ) {
1495                                                         # No terminating quote, assume literal "
1496                                                         $s .= '"';
1497                                                 } else {
1498                                                         $s .= substr( $format, $p + 1, $endQuote - $p - 1 );
1499                                                         $p = $endQuote;
1500                                                 }
1501                                         } else {
1502                                                 # Quote at end of string, assume literal "
1503                                                 $s .= '"';
1504                                         }
1505                                         break;
1506                                 default:
1507                                         $s .= $format[$p];
1508                         }
1509                         if ( $num !== false ) {
1510                                 if ( $rawToggle || $raw ) {
1511                                         $s .= $num;
1512                                         $raw = false;
1513                                 } elseif ( $roman ) {
1514                                         $s .= self::romanNumeral( $num );
1515                                         $roman = false;
1516                                 } elseif ( $hebrewNum ) {
1517                                         $s .= self::hebrewNumeral( $num );
1518                                         $hebrewNum = false;
1519                                 } else {
1520                                         $s .= $this->formatNum( $num, true );
1521                                 }
1522                         }
1523                 }
1524
1525                 if ( $ttl === 'unused' ) {
1526                         // No need to calculate the TTL, the caller wont use it anyway.
1527                 } elseif ( $usedSecond ) {
1528                         $ttl = 1;
1529                 } elseif ( $usedMinute ) {
1530                         $ttl = 60 - substr( $ts, 12, 2 );
1531                 } elseif ( $usedHour ) {
1532                         $ttl = 3600 - substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1533                 } elseif ( $usedAMPM ) {
1534                         $ttl = 43200 - ( substr( $ts, 8, 2 ) % 12 ) * 3600 -
1535                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1536                 } elseif (
1537                         $usedDay ||
1538                         $usedHebrewMonth ||
1539                         $usedIranianMonth ||
1540                         $usedHijriMonth ||
1541                         $usedHebrewYear ||
1542                         $usedIranianYear ||
1543                         $usedHijriYear ||
1544                         $usedTennoYear
1545                 ) {
1546                         // @todo Someone who understands the non-Gregorian calendars
1547                         // should write proper logic for them so that they don't need purged every day.
1548                         $ttl = 86400 - substr( $ts, 8, 2 ) * 3600 -
1549                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1550                 } else {
1551                         $possibleTtls = [];
1552                         $timeRemainingInDay = 86400 - substr( $ts, 8, 2 ) * 3600 -
1553                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1554                         if ( $usedWeek ) {
1555                                 $possibleTtls[] =
1556                                         ( 7 - self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400 +
1557                                         $timeRemainingInDay;
1558                         } elseif ( $usedISOYear ) {
1559                                 // December 28th falls on the last ISO week of the year, every year.
1560                                 // The last ISO week of a year can be 52 or 53.
1561                                 $lastWeekOfISOYear = DateTime::createFromFormat(
1562                                         'Ymd',
1563                                         substr( $ts, 0, 4 ) . '1228',
1564                                         $zone ?: new DateTimeZone( 'UTC' )
1565                                 )->format( 'W' );
1566                                 $currentISOWeek = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'W' );
1567                                 $weeksRemaining = $lastWeekOfISOYear - $currentISOWeek;
1568                                 $timeRemainingInWeek =
1569                                         ( 7 - self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400
1570                                         + $timeRemainingInDay;
1571                                 $possibleTtls[] = $weeksRemaining * 604800 + $timeRemainingInWeek;
1572                         }
1573
1574                         if ( $usedMonth ) {
1575                                 $possibleTtls[] =
1576                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 't' ) -
1577                                                 substr( $ts, 6, 2 ) ) * 86400
1578                                         + $timeRemainingInDay;
1579                         } elseif ( $usedYear ) {
1580                                 $possibleTtls[] =
1581                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1582                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1583                                         + $timeRemainingInDay;
1584                         } elseif ( $usedIsLeapYear ) {
1585                                 $year = substr( $ts, 0, 4 );
1586                                 $timeRemainingInYear =
1587                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1588                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1589                                         + $timeRemainingInDay;
1590                                 $mod = $year % 4;
1591                                 if ( $mod || ( !( $year % 100 ) && $year % 400 ) ) {
1592                                         // this isn't a leap year. see when the next one starts
1593                                         $nextCandidate = $year - $mod + 4;
1594                                         if ( $nextCandidate % 100 || !( $nextCandidate % 400 ) ) {
1595                                                 $possibleTtls[] = ( $nextCandidate - $year - 1 ) * 365 * 86400 +
1596                                                         $timeRemainingInYear;
1597                                         } else {
1598                                                 $possibleTtls[] = ( $nextCandidate - $year + 3 ) * 365 * 86400 +
1599                                                         $timeRemainingInYear;
1600                                         }
1601                                 } else {
1602                                         // this is a leap year, so the next year isn't
1603                                         $possibleTtls[] = $timeRemainingInYear;
1604                                 }
1605                         }
1606
1607                         if ( $possibleTtls ) {
1608                                 $ttl = min( $possibleTtls );
1609                         }
1610                 }
1611
1612                 return $s;
1613         }
1614
1615         private static $GREG_DAYS = [ 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 ];
1616         private static $IRANIAN_DAYS = [ 31, 31, 31, 31, 31, 31, 30, 30, 30, 30, 30, 29 ];
1617
1618         /**
1619          * Algorithm by Roozbeh Pournader and Mohammad Toossi to convert
1620          * Gregorian dates to Iranian dates. Originally written in C, it
1621          * is released under the terms of GNU Lesser General Public
1622          * License. Conversion to PHP was performed by Niklas Laxström.
1623          *
1624          * Link: http://www.farsiweb.info/jalali/jalali.c
1625          *
1626          * @param string $ts
1627          *
1628          * @return int[]
1629          */
1630         private static function tsToIranian( $ts ) {
1631                 $gy = substr( $ts, 0, 4 ) - 1600;
1632                 $gm = substr( $ts, 4, 2 ) - 1;
1633                 $gd = substr( $ts, 6, 2 ) - 1;
1634
1635                 # Days passed from the beginning (including leap years)
1636                 $gDayNo = 365 * $gy
1637                         + floor( ( $gy + 3 ) / 4 )
1638                         - floor( ( $gy + 99 ) / 100 )
1639                         + floor( ( $gy + 399 ) / 400 );
1640
1641                 // Add days of the past months of this year
1642                 for ( $i = 0; $i < $gm; $i++ ) {
1643                         $gDayNo += self::$GREG_DAYS[$i];
1644                 }
1645
1646                 // Leap years
1647                 if ( $gm > 1 && ( ( $gy % 4 === 0 && $gy % 100 !== 0 || ( $gy % 400 == 0 ) ) ) ) {
1648                         $gDayNo++;
1649                 }
1650
1651                 // Days passed in current month
1652                 $gDayNo += (int)$gd;
1653
1654                 $jDayNo = $gDayNo - 79;
1655
1656                 $jNp = floor( $jDayNo / 12053 );
1657                 $jDayNo %= 12053;
1658
1659                 $jy = 979 + 33 * $jNp + 4 * floor( $jDayNo / 1461 );
1660                 $jDayNo %= 1461;
1661
1662                 if ( $jDayNo >= 366 ) {
1663                         $jy += floor( ( $jDayNo - 1 ) / 365 );
1664                         $jDayNo = floor( ( $jDayNo - 1 ) % 365 );
1665                 }
1666
1667                 $jz = $jDayNo;
1668
1669                 for ( $i = 0; $i < 11 && $jDayNo >= self::$IRANIAN_DAYS[$i]; $i++ ) {
1670                         $jDayNo -= self::$IRANIAN_DAYS[$i];
1671                 }
1672
1673                 $jm = $i + 1;
1674                 $jd = $jDayNo + 1;
1675
1676                 return [ $jy, $jm, $jd, $jz ];
1677         }
1678
1679         /**
1680          * Converting Gregorian dates to Hijri dates.
1681          *
1682          * Based on a PHP-Nuke block by Sharjeel which is released under GNU/GPL license
1683          *
1684          * @see https://phpnuke.org/modules.php?name=News&file=article&sid=8234&mode=thread&order=0&thold=0
1685          *
1686          * @param string $ts
1687          *
1688          * @return int[]
1689          */
1690         private static function tsToHijri( $ts ) {
1691                 $year = substr( $ts, 0, 4 );
1692                 $month = substr( $ts, 4, 2 );
1693                 $day = substr( $ts, 6, 2 );
1694
1695                 $zyr = $year;
1696                 $zd = $day;
1697                 $zm = $month;
1698                 $zy = $zyr;
1699
1700                 if (
1701                         ( $zy > 1582 ) || ( ( $zy == 1582 ) && ( $zm > 10 ) ) ||
1702                         ( ( $zy == 1582 ) && ( $zm == 10 ) && ( $zd > 14 ) )
1703                 ) {
1704                         $zjd = (int)( ( 1461 * ( $zy + 4800 + (int)( ( $zm - 14 ) / 12 ) ) ) / 4 ) +
1705                                         (int)( ( 367 * ( $zm - 2 - 12 * ( (int)( ( $zm - 14 ) / 12 ) ) ) ) / 12 ) -
1706                                         (int)( ( 3 * (int)( ( ( $zy + 4900 + (int)( ( $zm - 14 ) / 12 ) ) / 100 ) ) ) / 4 ) +
1707                                         $zd - 32075;
1708                 } else {
1709                         $zjd = 367 * $zy - (int)( ( 7 * ( $zy + 5001 + (int)( ( $zm - 9 ) / 7 ) ) ) / 4 ) +
1710                                                                 (int)( ( 275 * $zm ) / 9 ) + $zd + 1729777;
1711                 }
1712
1713                 $zl = $zjd - 1948440 + 10632;
1714                 $zn = (int)( ( $zl - 1 ) / 10631 );
1715                 $zl = $zl - 10631 * $zn + 354;
1716                 $zj = ( (int)( ( 10985 - $zl ) / 5316 ) ) * ( (int)( ( 50 * $zl ) / 17719 ) ) +
1717                         ( (int)( $zl / 5670 ) ) * ( (int)( ( 43 * $zl ) / 15238 ) );
1718                 $zl = $zl - ( (int)( ( 30 - $zj ) / 15 ) ) * ( (int)( ( 17719 * $zj ) / 50 ) ) -
1719                         ( (int)( $zj / 16 ) ) * ( (int)( ( 15238 * $zj ) / 43 ) ) + 29;
1720                 $zm = (int)( ( 24 * $zl ) / 709 );
1721                 $zd = $zl - (int)( ( 709 * $zm ) / 24 );
1722                 $zy = 30 * $zn + $zj - 30;
1723
1724                 return [ $zy, $zm, $zd ];
1725         }
1726
1727         /**
1728          * Converting Gregorian dates to Hebrew dates.
1729          *
1730          * Based on a JavaScript code by Abu Mami and Yisrael Hersch
1731          * (abu-mami@kaluach.net, http://www.kaluach.net), who permitted
1732          * to translate the relevant functions into PHP and release them under
1733          * GNU GPL.
1734          *
1735          * The months are counted from Tishrei = 1. In a leap year, Adar I is 13
1736          * and Adar II is 14. In a non-leap year, Adar is 6.
1737          *
1738          * @param string $ts
1739          *
1740          * @return int[]
1741          */
1742         private static function tsToHebrew( $ts ) {
1743                 # Parse date
1744                 $year = substr( $ts, 0, 4 );
1745                 $month = substr( $ts, 4, 2 );
1746                 $day = substr( $ts, 6, 2 );
1747
1748                 # Calculate Hebrew year
1749                 $hebrewYear = $year + 3760;
1750
1751                 # Month number when September = 1, August = 12
1752                 $month += 4;
1753                 if ( $month > 12 ) {
1754                         # Next year
1755                         $month -= 12;
1756                         $year++;
1757                         $hebrewYear++;
1758                 }
1759
1760                 # Calculate day of year from 1 September
1761                 $dayOfYear = $day;
1762                 for ( $i = 1; $i < $month; $i++ ) {
1763                         if ( $i == 6 ) {
1764                                 # February
1765                                 $dayOfYear += 28;
1766                                 # Check if the year is leap
1767                                 if ( $year % 400 == 0 || ( $year % 4 == 0 && $year % 100 > 0 ) ) {
1768                                         $dayOfYear++;
1769                                 }
1770                         } elseif ( $i == 8 || $i == 10 || $i == 1 || $i == 3 ) {
1771                                 $dayOfYear += 30;
1772                         } else {
1773                                 $dayOfYear += 31;
1774                         }
1775                 }
1776
1777                 # Calculate the start of the Hebrew year
1778                 $start = self::hebrewYearStart( $hebrewYear );
1779
1780                 # Calculate next year's start
1781                 if ( $dayOfYear <= $start ) {
1782                         # Day is before the start of the year - it is the previous year
1783                         # Next year's start
1784                         $nextStart = $start;
1785                         # Previous year
1786                         $year--;
1787                         $hebrewYear--;
1788                         # Add days since previous year's 1 September
1789                         $dayOfYear += 365;
1790                         if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1791                                 # Leap year
1792                                 $dayOfYear++;
1793                         }
1794                         # Start of the new (previous) year
1795                         $start = self::hebrewYearStart( $hebrewYear );
1796                 } else {
1797                         # Next year's start
1798                         $nextStart = self::hebrewYearStart( $hebrewYear + 1 );
1799                 }
1800
1801                 # Calculate Hebrew day of year
1802                 $hebrewDayOfYear = $dayOfYear - $start;
1803
1804                 # Difference between year's days
1805                 $diff = $nextStart - $start;
1806                 # Add 12 (or 13 for leap years) days to ignore the difference between
1807                 # Hebrew and Gregorian year (353 at least vs. 365/6) - now the
1808                 # difference is only about the year type
1809                 if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1810                         $diff += 13;
1811                 } else {
1812                         $diff += 12;
1813                 }
1814
1815                 # Check the year pattern, and is leap year
1816                 # 0 means an incomplete year, 1 means a regular year, 2 means a complete year
1817                 # This is mod 30, to work on both leap years (which add 30 days of Adar I)
1818                 # and non-leap years
1819                 $yearPattern = $diff % 30;
1820                 # Check if leap year
1821                 $isLeap = $diff >= 30;
1822
1823                 # Calculate day in the month from number of day in the Hebrew year
1824                 # Don't check Adar - if the day is not in Adar, we will stop before;
1825                 # if it is in Adar, we will use it to check if it is Adar I or Adar II
1826                 $hebrewDay = $hebrewDayOfYear;
1827                 $hebrewMonth = 1;
1828                 $days = 0;
1829                 while ( $hebrewMonth <= 12 ) {
1830                         # Calculate days in this month
1831                         if ( $isLeap && $hebrewMonth == 6 ) {
1832                                 # Leap year - has Adar I, with 30 days, and Adar II, with 29 days
1833                                 $days = 30;
1834                                 if ( $hebrewDay <= $days ) {
1835                                         # Day in Adar I
1836                                         $hebrewMonth = 13;
1837                                 } else {
1838                                         # Subtract the days of Adar I
1839                                         $hebrewDay -= $days;
1840                                         # Try Adar II
1841                                         $days = 29;
1842                                         if ( $hebrewDay <= $days ) {
1843                                                 # Day in Adar II
1844                                                 $hebrewMonth = 14;
1845                                         }
1846                                 }
1847                         } elseif ( $hebrewMonth == 2 && $yearPattern == 2 ) {
1848                                 # Cheshvan in a complete year (otherwise as the rule below)
1849                                 $days = 30;
1850                         } elseif ( $hebrewMonth == 3 && $yearPattern == 0 ) {
1851                                 # Kislev in an incomplete year (otherwise as the rule below)
1852                                 $days = 29;
1853                         } else {
1854                                 # Odd months have 30 days, even have 29
1855                                 $days = 30 - ( $hebrewMonth - 1 ) % 2;
1856                         }
1857                         if ( $hebrewDay <= $days ) {
1858                                 # In the current month
1859                                 break;
1860                         } else {
1861                                 # Subtract the days of the current month
1862                                 $hebrewDay -= $days;
1863                                 # Try in the next month
1864                                 $hebrewMonth++;
1865                         }
1866                 }
1867
1868                 return [ $hebrewYear, $hebrewMonth, $hebrewDay, $days ];
1869         }
1870
1871         /**
1872          * This calculates the Hebrew year start, as days since 1 September.
1873          * Based on Carl Friedrich Gauss algorithm for finding Easter date.
1874          * Used for Hebrew date.
1875          *
1876          * @param int $year
1877          *
1878          * @return string
1879          */
1880         private static function hebrewYearStart( $year ) {
1881                 $a = intval( ( 12 * ( $year - 1 ) + 17 ) % 19 );
1882                 $b = intval( ( $year - 1 ) % 4 );
1883                 $m = 32.044093161144 + 1.5542417966212 * $a + $b / 4.0 - 0.0031777940220923 * ( $year - 1 );
1884                 if ( $m < 0 ) {
1885                         $m--;
1886                 }
1887                 $Mar = intval( $m );
1888                 if ( $m < 0 ) {
1889                         $m++;
1890                 }
1891                 $m -= $Mar;
1892
1893                 $c = intval( ( $Mar + 3 * ( $year - 1 ) + 5 * $b + 5 ) % 7 );
1894                 if ( $c == 0 && $a > 11 && $m >= 0.89772376543210 ) {
1895                         $Mar++;
1896                 } elseif ( $c == 1 && $a > 6 && $m >= 0.63287037037037 ) {
1897                         $Mar += 2;
1898                 } elseif ( $c == 2 || $c == 4 || $c == 6 ) {
1899                         $Mar++;
1900                 }
1901
1902                 $Mar += intval( ( $year - 3761 ) / 100 ) - intval( ( $year - 3761 ) / 400 ) - 24;
1903                 return $Mar;
1904         }
1905
1906         /**
1907          * Algorithm to convert Gregorian dates to Thai solar dates,
1908          * Minguo dates or Minguo dates.
1909          *
1910          * Link: https://en.wikipedia.org/wiki/Thai_solar_calendar
1911          *       https://en.wikipedia.org/wiki/Minguo_calendar
1912          *       https://en.wikipedia.org/wiki/Japanese_era_name
1913          *
1914          * @param string $ts 14-character timestamp
1915          * @param string $cName Calender name
1916          * @return array Converted year, month, day
1917          */
1918         private static function tsToYear( $ts, $cName ) {
1919                 $gy = substr( $ts, 0, 4 );
1920                 $gm = substr( $ts, 4, 2 );
1921                 $gd = substr( $ts, 6, 2 );
1922
1923                 if ( !strcmp( $cName, 'thai' ) ) {
1924                         # Thai solar dates
1925                         # Add 543 years to the Gregorian calendar
1926                         # Months and days are identical
1927                         $gy_offset = $gy + 543;
1928                         # fix for dates between 1912 and 1941
1929                         # https://en.wikipedia.org/?oldid=836596673#New_year
1930                         if ( $gy >= 1912 && $gy <= 1940 ) {
1931                                 if ( $gm <= 3 ) {
1932                                         $gy_offset--;
1933                                 }
1934                                 $gm = ( $gm - 3 ) % 12;
1935                         }
1936                 } elseif ( ( !strcmp( $cName, 'minguo' ) ) || !strcmp( $cName, 'juche' ) ) {
1937                         # Minguo dates
1938                         # Deduct 1911 years from the Gregorian calendar
1939                         # Months and days are identical
1940                         $gy_offset = $gy - 1911;
1941                 } elseif ( !strcmp( $cName, 'tenno' ) ) {
1942                         # Nengō dates up to Meiji period
1943                         # Deduct years from the Gregorian calendar
1944                         # depending on the nengo periods
1945                         # Months and days are identical
1946                         if ( ( $gy < 1912 )
1947                                 || ( ( $gy == 1912 ) && ( $gm < 7 ) )
1948                                 || ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd < 31 ) )
1949                         ) {
1950                                 # Meiji period
1951                                 $gy_gannen = $gy - 1868 + 1;
1952                                 $gy_offset = $gy_gannen;
1953                                 if ( $gy_gannen == 1 ) {
1954                                         $gy_offset = '元';
1955                                 }
1956                                 $gy_offset = '明治' . $gy_offset;
1957                         } elseif (
1958                                 ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd == 31 ) ) ||
1959                                 ( ( $gy == 1912 ) && ( $gm >= 8 ) ) ||
1960                                 ( ( $gy > 1912 ) && ( $gy < 1926 ) ) ||
1961                                 ( ( $gy == 1926 ) && ( $gm < 12 ) ) ||
1962                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd < 26 ) )
1963                         ) {
1964                                 # Taishō period
1965                                 $gy_gannen = $gy - 1912 + 1;
1966                                 $gy_offset = $gy_gannen;
1967                                 if ( $gy_gannen == 1 ) {
1968                                         $gy_offset = '元';
1969                                 }
1970                                 $gy_offset = '大正' . $gy_offset;
1971                         } elseif (
1972                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd >= 26 ) ) ||
1973                                 ( ( $gy > 1926 ) && ( $gy < 1989 ) ) ||
1974                                 ( ( $gy == 1989 ) && ( $gm == 1 ) && ( $gd < 8 ) )
1975                         ) {
1976                                 # Shōwa period
1977                                 $gy_gannen = $gy - 1926 + 1;
1978                                 $gy_offset = $gy_gannen;
1979                                 if ( $gy_gannen == 1 ) {
1980                                         $gy_offset = '元';
1981                                 }
1982                                 $gy_offset = '昭和' . $gy_offset;
1983                         } elseif (
1984                                 ( ( $gy == 1989 ) && ( $gm == 1 ) && ( $gd >= 8 ) ) ||
1985                                 ( ( $gy > 1989 ) && ( $gy < 2019 ) ) ||
1986                                 ( ( $gy == 2019 ) && ( $gm < 5 ) )
1987                         ) {
1988                                 # Heisei period
1989                                 $gy_gannen = $gy - 1989 + 1;
1990                                 $gy_offset = $gy_gannen;
1991                                 if ( $gy_gannen == 1 ) {
1992                                         $gy_offset = '元';
1993                                 }
1994                                 $gy_offset = '平成' . $gy_offset;
1995                         } else {
1996                                 # Reiwa period
1997                                 $gy_gannen = $gy - 2019 + 1;
1998                                 $gy_offset = $gy_gannen;
1999                                 if ( $gy_gannen == 1 ) {
2000                                         $gy_offset = '元';
2001                                 }
2002                                 $gy_offset = '令和' . $gy_offset;
2003                         }
2004                 } else {
2005                         $gy_offset = $gy;
2006                 }
2007
2008                 return [ $gy_offset, $gm, $gd ];
2009         }
2010
2011         /**
2012          * Gets directionality of the first strongly directional codepoint, for embedBidi()
2013          *
2014          * This is the rule the BIDI algorithm uses to determine the directionality of
2015          * paragraphs ( https://www.unicode.org/reports/tr9/#The_Paragraph_Level ) and
2016          * FSI isolates ( https://www.unicode.org/reports/tr9/#Explicit_Directional_Isolates ).
2017          *
2018          * TODO: Does not handle BIDI control characters inside the text.
2019          * TODO: Does not handle unallocated characters.
2020          *
2021          * @param string $text Text to test
2022          * @return null|string Directionality ('ltr' or 'rtl') or null
2023          */
2024         private static function strongDirFromContent( $text = '' ) {
2025                 if ( !preg_match( self::$strongDirRegex, $text, $matches ) ) {
2026                         return null;
2027                 }
2028                 if ( $matches[1] === '' ) {
2029                         return 'rtl';
2030                 }
2031                 return 'ltr';
2032         }
2033
2034         /**
2035          * Roman number formatting up to 10000
2036          *
2037          * @param int $num
2038          *
2039          * @return string
2040          */
2041         static function romanNumeral( $num ) {
2042                 static $table = [
2043                         [ '', 'I', 'II', 'III', 'IV', 'V', 'VI', 'VII', 'VIII', 'IX', 'X' ],
2044                         [ '', 'X', 'XX', 'XXX', 'XL', 'L', 'LX', 'LXX', 'LXXX', 'XC', 'C' ],
2045                         [ '', 'C', 'CC', 'CCC', 'CD', 'D', 'DC', 'DCC', 'DCCC', 'CM', 'M' ],
2046                         [ '', 'M', 'MM', 'MMM', 'MMMM', 'MMMMM', 'MMMMMM', 'MMMMMMM',
2047                                 'MMMMMMMM', 'MMMMMMMMM', 'MMMMMMMMMM' ]
2048                 ];
2049
2050                 $num = intval( $num );
2051                 if ( $num > 10000 || $num <= 0 ) {
2052                         return $num;
2053                 }
2054
2055                 $s = '';
2056                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
2057                         if ( $num >= $pow10 ) {
2058                                 $s .= $table[$i][(int)floor( $num / $pow10 )];
2059                         }
2060                         $num = $num % $pow10;
2061                 }
2062                 return $s;
2063         }
2064
2065         /**
2066          * Hebrew Gematria number formatting up to 9999
2067          *
2068          * @param int $num
2069          *
2070          * @return string
2071          */
2072         static function hebrewNumeral( $num ) {
2073                 static $table = [
2074                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ],
2075                         [ '', 'י', 'כ', 'ל', 'מ', 'נ', 'ס', 'ע', 'פ', 'צ', 'ק' ],
2076                         [ '',
2077                                 [ 'ק' ],
2078                                 [ 'ר' ],
2079                                 [ 'ש' ],
2080                                 [ 'ת' ],
2081                                 [ 'ת', 'ק' ],
2082                                 [ 'ת', 'ר' ],
2083                                 [ 'ת', 'ש' ],
2084                                 [ 'ת', 'ת' ],
2085                                 [ 'ת', 'ת', 'ק' ],
2086                                 [ 'ת', 'ת', 'ר' ],
2087                         ],
2088                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ]
2089                 ];
2090
2091                 $num = intval( $num );
2092                 if ( $num > 9999 || $num <= 0 ) {
2093                         return $num;
2094                 }
2095
2096                 // Round thousands have special notations
2097                 if ( $num === 1000 ) {
2098                         return "א' אלף";
2099                 } elseif ( $num % 1000 === 0 ) {
2100                         return $table[0][$num / 1000] . "' אלפים";
2101                 }
2102
2103                 $letters = [];
2104
2105                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
2106                         if ( $num >= $pow10 ) {
2107                                 if ( $num === 15 || $num === 16 ) {
2108                                         $letters[] = $table[0][9];
2109                                         $letters[] = $table[0][$num - 9];
2110                                         $num = 0;
2111                                 } else {
2112                                         $letters = array_merge(
2113                                                 $letters,
2114                                                 (array)$table[$i][intval( $num / $pow10 )]
2115                                         );
2116
2117                                         if ( $pow10 === 1000 ) {
2118                                                 $letters[] = "'";
2119                                         }
2120                                 }
2121                         }
2122
2123                         $num = $num % $pow10;
2124                 }
2125
2126                 $preTransformLength = count( $letters );
2127                 if ( $preTransformLength === 1 ) {
2128                         // Add geresh (single quote) to one-letter numbers
2129                         $letters[] = "'";
2130                 } else {
2131                         $lastIndex = $preTransformLength - 1;
2132                         $letters[$lastIndex] = str_replace(
2133                                 [ 'כ', 'מ', 'נ', 'פ', 'צ' ],
2134                                 [ 'ך', 'ם', 'ן', 'ף', 'ץ' ],
2135                                 $letters[$lastIndex]
2136                         );
2137
2138                         // Add gershayim (double quote) to multiple-letter numbers,
2139                         // but exclude numbers with only one letter after the thousands
2140                         // (1001-1009, 1020, 1030, 2001-2009, etc.)
2141                         if ( $letters[1] === "'" && $preTransformLength === 3 ) {
2142                                 $letters[] = "'";
2143                         } else {
2144                                 array_splice( $letters, -1, 0, '"' );
2145                         }
2146                 }
2147
2148                 return implode( $letters );
2149         }
2150
2151         /**
2152          * Used by date() and time() to adjust the time output.
2153          *
2154          * @param string $ts The time in date('YmdHis') format
2155          * @param mixed $tz Adjust the time by this amount (default false, mean we
2156          *   get user timecorrection setting)
2157          * @return int
2158          */
2159         public function userAdjust( $ts, $tz = false ) {
2160                 global $wgUser, $wgLocalTZoffset;
2161
2162                 if ( $tz === false ) {
2163                         $tz = $wgUser->getOption( 'timecorrection' );
2164                 }
2165
2166                 $data = explode( '|', $tz, 3 );
2167
2168                 if ( $data[0] == 'ZoneInfo' ) {
2169                         try {
2170                                 $userTZ = new DateTimeZone( $data[2] );
2171                                 $date = new DateTime( $ts, new DateTimeZone( 'UTC' ) );
2172                                 $date->setTimezone( $userTZ );
2173                                 return $date->format( 'YmdHis' );
2174                         } catch ( Exception $e ) {
2175                                 // Unrecognized timezone, default to 'Offset' with the stored offset.
2176                                 $data[0] = 'Offset';
2177                         }
2178                 }
2179
2180                 if ( $data[0] == 'System' || $tz == '' ) {
2181                         # Global offset in minutes.
2182                         $minDiff = $wgLocalTZoffset;
2183                 } elseif ( $data[0] == 'Offset' ) {
2184                         $minDiff = intval( $data[1] );
2185                 } else {
2186                         $data = explode( ':', $tz );
2187                         if ( count( $data ) == 2 ) {
2188                                 $data[0] = intval( $data[0] );
2189                                 $data[1] = intval( $data[1] );
2190                                 $minDiff = abs( $data[0] ) * 60 + $data[1];
2191                                 if ( $data[0] < 0 ) {
2192                                         $minDiff = -$minDiff;
2193                                 }
2194                         } else {
2195                                 $minDiff = intval( $data[0] ) * 60;
2196                         }
2197                 }
2198
2199                 # No difference ? Return time unchanged
2200                 if ( $minDiff == 0 ) {
2201                         return $ts;
2202                 }
2203
2204                 Wikimedia\suppressWarnings(); // E_STRICT system time bitching
2205                 # Generate an adjusted date; take advantage of the fact that mktime
2206                 # will normalize out-of-range values so we don't have to split $minDiff
2207                 # into hours and minutes.
2208                 $t = mktime( (
2209                         (int)substr( $ts, 8, 2 ) ), # Hours
2210                         (int)substr( $ts, 10, 2 ) + $minDiff, # Minutes
2211                         (int)substr( $ts, 12, 2 ), # Seconds
2212                         (int)substr( $ts, 4, 2 ), # Month
2213                         (int)substr( $ts, 6, 2 ), # Day
2214                         (int)substr( $ts, 0, 4 ) ); # Year
2215
2216                 $date = date( 'YmdHis', $t );
2217                 Wikimedia\restoreWarnings();
2218
2219                 return $date;
2220         }
2221
2222         /**
2223          * This is meant to be used by time(), date(), and timeanddate() to get
2224          * the date preference they're supposed to use, it should be used in
2225          * all children.
2226          *
2227          *     function timeanddate([...], $format = true) {
2228          *       $datePreference = $this->dateFormat($format);
2229          *       [...]
2230          *     }
2231          *
2232          * @param int|string|bool $usePrefs If true, the user's preference is used
2233          *   if false, the site/language default is used
2234          *   if int/string, assumed to be a format.
2235          * @return string
2236          */
2237         function dateFormat( $usePrefs = true ) {
2238                 global $wgUser;
2239
2240                 if ( is_bool( $usePrefs ) ) {
2241                         if ( $usePrefs ) {
2242                                 $datePreference = $wgUser->getDatePreference();
2243                         } else {
2244                                 $datePreference = (string)User::getDefaultOption( 'date' );
2245                         }
2246                 } else {
2247                         $datePreference = (string)$usePrefs;
2248                 }
2249
2250                 // return int
2251                 if ( $datePreference == '' ) {
2252                         return 'default';
2253                 }
2254
2255                 return $datePreference;
2256         }
2257
2258         /**
2259          * Get a format string for a given type and preference
2260          * @param string $type May be 'date', 'time', 'both', or 'pretty'.
2261          * @param string $pref The format name as it appears in Messages*.php under
2262          *  $datePreferences.
2263          *
2264          * @since 1.22 New type 'pretty' that provides a more readable timestamp format
2265          *
2266          * @return string
2267          */
2268         function getDateFormatString( $type, $pref ) {
2269                 $wasDefault = false;
2270                 if ( $pref == 'default' ) {
2271                         $wasDefault = true;
2272                         $pref = $this->getDefaultDateFormat();
2273                 }
2274
2275                 if ( !isset( $this->dateFormatStrings[$type][$pref] ) ) {
2276                         $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2277
2278                         if ( $type === 'pretty' && $df === null ) {
2279                                 $df = $this->getDateFormatString( 'date', $pref );
2280                         }
2281
2282                         if ( !$wasDefault && $df === null ) {
2283                                 $pref = $this->getDefaultDateFormat();
2284                                 $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2285                         }
2286
2287                         $this->dateFormatStrings[$type][$pref] = $df;
2288                 }
2289                 return $this->dateFormatStrings[$type][$pref];
2290         }
2291
2292         /**
2293          * @param string $ts The time format which needs to be turned into a
2294          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2295          * @param bool $adj Whether to adjust the time output according to the
2296          *   user configured offset ($timecorrection)
2297          * @param mixed $format True to use user's date format preference
2298          * @param string|bool $timecorrection The time offset as returned by
2299          *   validateTimeZone() in Special:Preferences
2300          * @return string
2301          */
2302         public function date( $ts, $adj = false, $format = true, $timecorrection = false ) {
2303                 $ts = wfTimestamp( TS_MW, $ts );
2304                 if ( $adj ) {
2305                         $ts = $this->userAdjust( $ts, $timecorrection );
2306                 }
2307                 $df = $this->getDateFormatString( 'date', $this->dateFormat( $format ) );
2308                 return $this->sprintfDate( $df, $ts );
2309         }
2310
2311         /**
2312          * @param string $ts The time format which needs to be turned into a
2313          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2314          * @param bool $adj Whether to adjust the time output according to the
2315          *   user configured offset ($timecorrection)
2316          * @param mixed $format True to use user's date format preference
2317          * @param string|bool $timecorrection The time offset as returned by
2318          *   validateTimeZone() in Special:Preferences
2319          * @return string
2320          */
2321         public function time( $ts, $adj = false, $format = true, $timecorrection = false ) {
2322                 $ts = wfTimestamp( TS_MW, $ts );
2323                 if ( $adj ) {
2324                         $ts = $this->userAdjust( $ts, $timecorrection );
2325                 }
2326                 $df = $this->getDateFormatString( 'time', $this->dateFormat( $format ) );
2327                 return $this->sprintfDate( $df, $ts );
2328         }
2329
2330         /**
2331          * @param string $ts The time format which needs to be turned into a
2332          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2333          * @param bool $adj Whether to adjust the time output according to the
2334          *   user configured offset ($timecorrection)
2335          * @param mixed $format What format to return, if it's false output the
2336          *   default one (default true)
2337          * @param string|bool $timecorrection The time offset as returned by
2338          *   validateTimeZone() in Special:Preferences
2339          * @return string
2340          */
2341         public function timeanddate( $ts, $adj = false, $format = true, $timecorrection = false ) {
2342                 $ts = wfTimestamp( TS_MW, $ts );
2343                 if ( $adj ) {
2344                         $ts = $this->userAdjust( $ts, $timecorrection );
2345                 }
2346                 $df = $this->getDateFormatString( 'both', $this->dateFormat( $format ) );
2347                 return $this->sprintfDate( $df, $ts );
2348         }
2349
2350         /**
2351          * Takes a number of seconds and turns it into a text using values such as hours and minutes.
2352          *
2353          * @since 1.20
2354          *
2355          * @param int $seconds The amount of seconds.
2356          * @param array $chosenIntervals The intervals to enable.
2357          *
2358          * @return string
2359          */
2360         public function formatDuration( $seconds, array $chosenIntervals = [] ) {
2361                 $intervals = $this->getDurationIntervals( $seconds, $chosenIntervals );
2362
2363                 $segments = [];
2364
2365                 foreach ( $intervals as $intervalName => $intervalValue ) {
2366                         // Messages: duration-seconds, duration-minutes, duration-hours, duration-days, duration-weeks,
2367                         // duration-years, duration-decades, duration-centuries, duration-millennia
2368                         $message = wfMessage( 'duration-' . $intervalName )->numParams( $intervalValue );
2369                         $segments[] = $message->inLanguage( $this )->escaped();
2370                 }
2371
2372                 return $this->listToText( $segments );
2373         }
2374
2375         /**
2376          * Takes a number of seconds and returns an array with a set of corresponding intervals.
2377          * For example 65 will be turned into [ minutes => 1, seconds => 5 ].
2378          *
2379          * @since 1.20
2380          *
2381          * @param int $seconds The amount of seconds.
2382          * @param array $chosenIntervals The intervals to enable.
2383          *
2384          * @return array
2385          */
2386         public function getDurationIntervals( $seconds, array $chosenIntervals = [] ) {
2387                 if ( empty( $chosenIntervals ) ) {
2388                         $chosenIntervals = [
2389                                 'millennia',
2390                                 'centuries',
2391                                 'decades',
2392                                 'years',
2393                                 'days',
2394                                 'hours',
2395                                 'minutes',
2396                                 'seconds'
2397                         ];
2398                 }
2399
2400                 $intervals = array_intersect_key( self::$durationIntervals, array_flip( $chosenIntervals ) );
2401                 $sortedNames = array_keys( $intervals );
2402                 $smallestInterval = array_pop( $sortedNames );
2403
2404                 $segments = [];
2405
2406                 foreach ( $intervals as $name => $length ) {
2407                         $value = floor( $seconds / $length );
2408
2409                         if ( $value > 0 || ( $name == $smallestInterval && empty( $segments ) ) ) {
2410                                 $seconds -= $value * $length;
2411                                 $segments[$name] = $value;
2412                         }
2413                 }
2414
2415                 return $segments;
2416         }
2417
2418         /**
2419          * Internal helper function for userDate(), userTime() and userTimeAndDate()
2420          *
2421          * @param string $type Can be 'date', 'time' or 'both'
2422          * @param string $ts The time format which needs to be turned into a
2423          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2424          * @param User $user User object used to get preferences for timezone and format
2425          * @param array $options Array, can contain the following keys:
2426          *   - 'timecorrection': time correction, can have the following values:
2427          *     - true: use user's preference
2428          *     - false: don't use time correction
2429          *     - int: value of time correction in minutes
2430          *   - 'format': format to use, can have the following values:
2431          *     - true: use user's preference
2432          *     - false: use default preference
2433          *     - string: format to use
2434          * @since 1.19
2435          * @return string
2436          */
2437         private function internalUserTimeAndDate( $type, $ts, User $user, array $options ) {
2438                 $ts = wfTimestamp( TS_MW, $ts );
2439                 $options += [ 'timecorrection' => true, 'format' => true ];
2440                 if ( $options['timecorrection'] !== false ) {
2441                         if ( $options['timecorrection'] === true ) {
2442                                 $offset = $user->getOption( 'timecorrection' );
2443                         } else {
2444                                 $offset = $options['timecorrection'];
2445                         }
2446                         $ts = $this->userAdjust( $ts, $offset );
2447                 }
2448                 if ( $options['format'] === true ) {
2449                         $format = $user->getDatePreference();
2450                 } else {
2451                         $format = $options['format'];
2452                 }
2453                 $df = $this->getDateFormatString( $type, $this->dateFormat( $format ) );
2454                 return $this->sprintfDate( $df, $ts );
2455         }
2456
2457         /**
2458          * Get the formatted date for the given timestamp and formatted for
2459          * the given user.
2460          *
2461          * @param mixed $ts Mixed: the time format which needs to be turned into a
2462          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2463          * @param User $user User object used to get preferences for timezone and format
2464          * @param array $options Array, can contain the following keys:
2465          *   - 'timecorrection': time correction, can have the following values:
2466          *     - true: use user's preference
2467          *     - false: don't use time correction
2468          *     - int: value of time correction in minutes
2469          *   - 'format': format to use, can have the following values:
2470          *     - true: use user's preference
2471          *     - false: use default preference
2472          *     - string: format to use
2473          * @since 1.19
2474          * @return string
2475          */
2476         public function userDate( $ts, User $user, array $options = [] ) {
2477                 return $this->internalUserTimeAndDate( 'date', $ts, $user, $options );
2478         }
2479
2480         /**
2481          * Get the formatted time for the given timestamp and formatted for
2482          * the given user.
2483          *
2484          * @param mixed $ts The time format which needs to be turned into a
2485          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2486          * @param User $user User object used to get preferences for timezone and format
2487          * @param array $options Array, can contain the following keys:
2488          *   - 'timecorrection': time correction, can have the following values:
2489          *     - true: use user's preference
2490          *     - false: don't use time correction
2491          *     - int: value of time correction in minutes
2492          *   - 'format': format to use, can have the following values:
2493          *     - true: use user's preference
2494          *     - false: use default preference
2495          *     - string: format to use
2496          * @since 1.19
2497          * @return string
2498          */
2499         public function userTime( $ts, User $user, array $options = [] ) {
2500                 return $this->internalUserTimeAndDate( 'time', $ts, $user, $options );
2501         }
2502
2503         /**
2504          * Get the formatted date and time for the given timestamp and formatted for
2505          * the given user.
2506          *
2507          * @param mixed $ts The time format which needs to be turned into a
2508          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2509          * @param User $user User object used to get preferences for timezone and format
2510          * @param array $options Array, can contain the following keys:
2511          *   - 'timecorrection': time correction, can have the following values:
2512          *     - true: use user's preference
2513          *     - false: don't use time correction
2514          *     - int: value of time correction in minutes
2515          *   - 'format': format to use, can have the following values:
2516          *     - true: use user's preference
2517          *     - false: use default preference
2518          *     - string: format to use
2519          * @since 1.19
2520          * @return string
2521          */
2522         public function userTimeAndDate( $ts, User $user, array $options = [] ) {
2523                 return $this->internalUserTimeAndDate( 'both', $ts, $user, $options );
2524         }
2525
2526         /**
2527          * Get the timestamp in a human-friendly relative format, e.g., "3 days ago".
2528          *
2529          * Determine the difference between the timestamp and the current time, and
2530          * generate a readable timestamp by returning "<N> <units> ago", where the
2531          * largest possible unit is used.
2532          *
2533          * @since 1.26 (Prior to 1.26 method existed but was not meant to be used directly)
2534          *
2535          * @param MWTimestamp $time
2536          * @param MWTimestamp|null $relativeTo The base timestamp to compare to (defaults to now)
2537          * @param User|null $user User the timestamp is being generated for
2538          *  (or null to use main context's user)
2539          * @return string Formatted timestamp
2540          */
2541         public function getHumanTimestamp(
2542                 MWTimestamp $time, MWTimestamp $relativeTo = null, User $user = null
2543         ) {
2544                 if ( $relativeTo === null ) {
2545                         $relativeTo = new MWTimestamp();
2546                 }
2547                 if ( $user === null ) {
2548                         $user = RequestContext::getMain()->getUser();
2549                 }
2550
2551                 // Adjust for the user's timezone.
2552                 $offsetThis = $time->offsetForUser( $user );
2553                 $offsetRel = $relativeTo->offsetForUser( $user );
2554
2555                 $ts = '';
2556                 if ( Hooks::run( 'GetHumanTimestamp', [ &$ts, $time, $relativeTo, $user, $this ] ) ) {
2557                         $ts = $this->getHumanTimestampInternal( $time, $relativeTo, $user );
2558                 }
2559
2560                 // Reset the timezone on the objects.
2561                 $time->timestamp->sub( $offsetThis );
2562                 $relativeTo->timestamp->sub( $offsetRel );
2563
2564                 return $ts;
2565         }
2566
2567         /**
2568          * Convert an MWTimestamp into a pretty human-readable timestamp using
2569          * the given user preferences and relative base time.
2570          *
2571          * @see Language::getHumanTimestamp
2572          * @param MWTimestamp $ts Timestamp to prettify
2573          * @param MWTimestamp $relativeTo Base timestamp
2574          * @param User $user User preferences to use
2575          * @return string Human timestamp
2576          * @since 1.26
2577          */
2578         private function getHumanTimestampInternal(
2579                 MWTimestamp $ts, MWTimestamp $relativeTo, User $user
2580         ) {
2581                 $diff = $ts->diff( $relativeTo );
2582                 $diffDay = (bool)( (int)$ts->timestamp->format( 'w' ) -
2583                         (int)$relativeTo->timestamp->format( 'w' ) );
2584                 $days = $diff->days ?: (int)$diffDay;
2585                 if ( $diff->invert || $days > 5
2586                         && $ts->timestamp->format( 'Y' ) !== $relativeTo->timestamp->format( 'Y' )
2587                 ) {
2588                         // Timestamps are in different years: use full timestamp
2589                         // Also do full timestamp for future dates
2590                         /**
2591                          * @todo FIXME: Add better handling of future timestamps.
2592                          */
2593                         $format = $this->getDateFormatString( 'both', $user->getDatePreference() ?: 'default' );
2594                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2595                 } elseif ( $days > 5 ) {
2596                         // Timestamps are in same year,  but more than 5 days ago: show day and month only.
2597                         $format = $this->getDateFormatString( 'pretty', $user->getDatePreference() ?: 'default' );
2598                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2599                 } elseif ( $days > 1 ) {
2600                         // Timestamp within the past week: show the day of the week and time
2601                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2602                         $weekday = self::$mWeekdayMsgs[$ts->timestamp->format( 'w' )];
2603                         // Messages:
2604                         // sunday-at, monday-at, tuesday-at, wednesday-at, thursday-at, friday-at, saturday-at
2605                         $ts = wfMessage( "$weekday-at" )
2606                                 ->inLanguage( $this )
2607                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2608                                 ->text();
2609                 } elseif ( $days == 1 ) {
2610                         // Timestamp was yesterday: say 'yesterday' and the time.
2611                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2612                         $ts = wfMessage( 'yesterday-at' )
2613                                 ->inLanguage( $this )
2614                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2615                                 ->text();
2616                 } elseif ( $diff->h > 1 || $diff->h == 1 && $diff->i > 30 ) {
2617                         // Timestamp was today, but more than 90 minutes ago: say 'today' and the time.
2618                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2619                         $ts = wfMessage( 'today-at' )
2620                                 ->inLanguage( $this )
2621                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2622                                 ->text();
2623
2624                 // From here on in, the timestamp was soon enough ago so that we can simply say
2625                 // XX units ago, e.g., "2 hours ago" or "5 minutes ago"
2626                 } elseif ( $diff->h == 1 ) {
2627                         // Less than 90 minutes, but more than an hour ago.
2628                         $ts = wfMessage( 'hours-ago' )->inLanguage( $this )->numParams( 1 )->text();
2629                 } elseif ( $diff->i >= 1 ) {
2630                         // A few minutes ago.
2631                         $ts = wfMessage( 'minutes-ago' )->inLanguage( $this )->numParams( $diff->i )->text();
2632                 } elseif ( $diff->s >= 30 ) {
2633                         // Less than a minute, but more than 30 sec ago.
2634                         $ts = wfMessage( 'seconds-ago' )->inLanguage( $this )->numParams( $diff->s )->text();
2635                 } else {
2636                         // Less than 30 seconds ago.
2637                         $ts = wfMessage( 'just-now' )->text();
2638                 }
2639
2640                 return $ts;
2641         }
2642
2643         /**
2644          * @param string $key
2645          * @return string|null
2646          */
2647         public function getMessage( $key ) {
2648                 return self::$dataCache->getSubitem( $this->mCode, 'messages', $key );
2649         }
2650
2651         /**
2652          * @return array
2653          */
2654         function getAllMessages() {
2655                 return self::$dataCache->getItem( $this->mCode, 'messages' );
2656         }
2657
2658         /**
2659          * @param string $in
2660          * @param string $out
2661          * @param string $string
2662          * @return string
2663          */
2664         public function iconv( $in, $out, $string ) {
2665                 # Even with //IGNORE iconv can whine about illegal characters in
2666                 # *input* string. We just ignore those too.
2667                 # REF: https://bugs.php.net/bug.php?id=37166
2668                 # REF: https://phabricator.wikimedia.org/T18885
2669                 Wikimedia\suppressWarnings();
2670                 $text = iconv( $in, $out . '//IGNORE', $string );
2671                 Wikimedia\restoreWarnings();
2672                 return $text;
2673         }
2674
2675         // callback functions for ucwords(), ucwordbreaks()
2676
2677         /**
2678          * @param array $matches
2679          * @return mixed|string
2680          */
2681         function ucwordbreaksCallbackAscii( $matches ) {
2682                 return $this->ucfirst( $matches[1] );
2683         }
2684
2685         /**
2686          * @param array $matches
2687          * @return string
2688          */
2689         function ucwordbreaksCallbackMB( $matches ) {
2690                 return mb_strtoupper( $matches[0] );
2691         }
2692
2693         /**
2694          * @param array $matches
2695          * @return string
2696          */
2697         function ucwordsCallbackMB( $matches ) {
2698                 return mb_strtoupper( $matches[0] );
2699         }
2700
2701         /**
2702          * Make a string's first character uppercase
2703          *
2704          * @param string $str
2705          *
2706          * @return string
2707          */
2708         public function ucfirst( $str ) {
2709                 $o = ord( $str );
2710                 if ( $o < 96 ) { // if already uppercase...
2711                         return $str;
2712                 } elseif ( $o < 128 ) {
2713                         return ucfirst( $str ); // use PHP's ucfirst()
2714                 } else {
2715                         // fall back to more complex logic in case of multibyte strings
2716                         return $this->uc( $str, true );
2717                 }
2718         }
2719
2720         /**
2721          * Convert a string to uppercase
2722          *
2723          * @param string $str
2724          * @param bool $first
2725          *
2726          * @return string
2727          */
2728         public function uc( $str, $first = false ) {
2729                 if ( $first ) {
2730                         if ( $this->isMultibyte( $str ) ) {
2731                                 return $this->mbUpperChar( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2732                         } else {
2733                                 return ucfirst( $str );
2734                         }
2735                 } else {
2736                         return $this->isMultibyte( $str ) ? mb_strtoupper( $str ) : strtoupper( $str );
2737                 }
2738         }
2739
2740         /**
2741          * Convert character to uppercase, allowing overrides of the default mb_upper
2742          * behaviour, which is buggy in many ways. Having a conversion table can be
2743          * useful during transitions between PHP versions where unicode changes happen.
2744          * This can make some resources unreachable on-wiki, see discussion at T219279.
2745          * Providing such a conversion table can allow to manage the transition period.
2746          *
2747          * @since 1.34
2748          *
2749          * @param string $char
2750          *
2751          * @return string
2752          */
2753         protected function mbUpperChar( $char ) {
2754                 global $wgOverrideUcfirstCharacters;
2755                 if ( array_key_exists( $char, $wgOverrideUcfirstCharacters ) ) {
2756                         return $wgOverrideUcfirstCharacters[$char];
2757                 } else {
2758                         return mb_strtoupper( $char );
2759                 }
2760         }
2761
2762         /**
2763          * @param string $str
2764          * @return mixed|string
2765          */
2766         function lcfirst( $str ) {
2767                 $o = ord( $str );
2768                 if ( !$o ) {
2769                         return strval( $str );
2770                 } elseif ( $o >= 128 ) {
2771                         return $this->lc( $str, true );
2772                 } elseif ( $o > 96 ) {
2773                         return $str;
2774                 } else {
2775                         $str[0] = strtolower( $str[0] );
2776                         return $str;
2777                 }
2778         }
2779
2780         /**
2781          * @param string $str
2782          * @param bool $first
2783          * @return mixed|string
2784          */
2785         function lc( $str, $first = false ) {
2786                 if ( $first ) {
2787                         if ( $this->isMultibyte( $str ) ) {
2788                                 return mb_strtolower( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2789                         } else {
2790                                 return strtolower( substr( $str, 0, 1 ) ) . substr( $str, 1 );
2791                         }
2792                 } else {
2793                         return $this->isMultibyte( $str ) ? mb_strtolower( $str ) : strtolower( $str );
2794                 }
2795         }
2796
2797         /**
2798          * @param string $str
2799          * @return bool
2800          */
2801         function isMultibyte( $str ) {
2802                 return strlen( $str ) !== mb_strlen( $str );
2803         }
2804
2805         /**
2806          * @param string $str
2807          * @return mixed|string
2808          */
2809         function ucwords( $str ) {
2810                 if ( $this->isMultibyte( $str ) ) {
2811                         $str = $this->lc( $str );
2812
2813                         // regexp to find first letter in each word (i.e. after each space)
2814                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)| ([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2815
2816                         // function to use to capitalize a single char
2817                         return preg_replace_callback(
2818                                 $replaceRegexp,
2819                                 [ $this, 'ucwordsCallbackMB' ],
2820                                 $str
2821                         );
2822                 } else {
2823                         return ucwords( strtolower( $str ) );
2824                 }
2825         }
2826
2827         /**
2828          * capitalize words at word breaks
2829          *
2830          * @param string $str
2831          * @return mixed
2832          */
2833         function ucwordbreaks( $str ) {
2834                 if ( $this->isMultibyte( $str ) ) {
2835                         $str = $this->lc( $str );
2836
2837                         // since \b doesn't work for UTF-8, we explicitely define word break chars
2838                         $breaks = "[ \-\(\)\}\{\.,\?!]";
2839
2840                         // find first letter after word break
2841                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)|" .
2842                                 "$breaks([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2843
2844                         return preg_replace_callback(
2845                                 $replaceRegexp,
2846                                 [ $this, 'ucwordbreaksCallbackMB' ],
2847                                 $str
2848                         );
2849                 } else {
2850                         return preg_replace_callback(
2851                                 '/\b([\w\x80-\xff]+)\b/',
2852                                 [ $this, 'ucwordbreaksCallbackAscii' ],
2853                                 $str
2854                         );
2855                 }
2856         }
2857
2858         /**
2859          * Return a case-folded representation of $s
2860          *
2861          * This is a representation such that caseFold($s1)==caseFold($s2) if $s1
2862          * and $s2 are the same except for the case of their characters. It is not
2863          * necessary for the value returned to make sense when displayed.
2864          *
2865          * Do *not* perform any other normalisation in this function. If a caller
2866          * uses this function when it should be using a more general normalisation
2867          * function, then fix the caller.
2868          *
2869          * @param string $s
2870          *
2871          * @return string
2872          */
2873         function caseFold( $s ) {
2874                 return $this->uc( $s );
2875         }
2876
2877         /**
2878          * TODO: $s is not always a string per T218883
2879          * @param string $s
2880          * @return string
2881          */
2882         function checkTitleEncoding( $s ) {
2883                 if ( is_array( $s ) ) {
2884                         throw new MWException( 'Given array to checkTitleEncoding.' );
2885                 }
2886                 if ( StringUtils::isUtf8( $s ) ) {
2887                         return $s;
2888                 }
2889
2890                 return $this->iconv( $this->fallback8bitEncoding(), 'utf-8', $s );
2891         }
2892
2893         /**
2894          * @return string
2895          */
2896         function fallback8bitEncoding() {
2897                 return self::$dataCache->getItem( $this->mCode, 'fallback8bitEncoding' );
2898         }
2899
2900         /**
2901          * Most writing systems use whitespace to break up words.
2902          * Some languages such as Chinese don't conventionally do this,
2903          * which requires special handling when breaking up words for
2904          * searching etc.
2905          *
2906          * @return bool
2907          */
2908         function hasWordBreaks() {
2909                 return true;
2910         }
2911
2912         /**
2913          * Some languages such as Chinese require word segmentation,
2914          * Specify such segmentation when overridden in derived class.
2915          *
2916          * @param string $string
2917          * @return string
2918          */
2919         function segmentByWord( $string ) {
2920                 return $string;
2921         }
2922
2923         /**
2924          * Some languages have special punctuation need to be normalized.
2925          * Make such changes here.
2926          *
2927          * @param string $string
2928          * @return string
2929          */
2930         function normalizeForSearch( $string ) {
2931                 return self::convertDoubleWidth( $string );
2932         }
2933
2934         /**
2935          * convert double-width roman characters to single-width.
2936          * range: ff00-ff5f ~= 0020-007f
2937          *
2938          * @param string $string
2939          *
2940          * @return string
2941          */
2942         protected static function convertDoubleWidth( $string ) {
2943                 static $full = null;
2944                 static $half = null;
2945
2946                 if ( $full === null ) {
2947                         $fullWidth = "０１２３４５６７８９ＡＢＣＤＥＦＧＨＩＪＫＬＭＮＯＰＱＲＳＴＵＶＷＸＹＺａｂｃｄｅｆｇｈｉｊｋｌｍｎｏｐｑｒｓｔｕｖｗｘｙｚ";
2948                         $halfWidth = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz";
2949                         $full = str_split( $fullWidth, 3 );
2950                         $half = str_split( $halfWidth );
2951                 }
2952
2953                 $string = str_replace( $full, $half, $string );
2954                 return $string;
2955         }
2956
2957         /**
2958          * @param string $string
2959          * @param string $pattern
2960          * @return string
2961          */
2962         protected static function insertSpace( $string, $pattern ) {
2963                 $string = preg_replace( $pattern, " $1 ", $string );
2964                 $string = preg_replace( '/ +/', ' ', $string );
2965                 return $string;
2966         }
2967
2968         /**
2969          * @param array $termsArray
2970          * @return array
2971          */
2972         function convertForSearchResult( $termsArray ) {
2973                 # some languages, e.g. Chinese, need to do a conversion
2974                 # in order for search results to be displayed correctly
2975                 return $termsArray;
2976         }
2977
2978         /**
2979          * Get the first character of a string.
2980          *
2981          * @param string $s
2982          * @return string
2983          */
2984         function firstChar( $s ) {
2985                 $matches = [];
2986                 preg_match(
2987                         '/^([\x00-\x7f]|[\xc0-\xdf][\x80-\xbf]|' .
2988                                 '[\xe0-\xef][\x80-\xbf]{2}|[\xf0-\xf7][\x80-\xbf]{3})/',
2989                         $s,
2990                         $matches
2991                 );
2992
2993                 if ( isset( $matches[1] ) ) {
2994                         if ( strlen( $matches[1] ) != 3 ) {
2995                                 return $matches[1];
2996                         }
2997
2998                         // Break down Hangul syllables to grab the first jamo
2999                         $code = UtfNormal\Utils::utf8ToCodepoint( $matches[1] );
3000                         if ( $code < 0xac00 || 0xd7a4 <= $code ) {
3001                                 return $matches[1];
3002                         } elseif ( $code < 0xb098 ) {
3003                                 return "\u{3131}";
3004                         } elseif ( $code < 0xb2e4 ) {
3005                                 return "\u{3134}";
3006                         } elseif ( $code < 0xb77c ) {
3007                                 return "\u{3137}";
3008                         } elseif ( $code < 0xb9c8 ) {
3009                                 return "\u{3139}";
3010                         } elseif ( $code < 0xbc14 ) {
3011                                 return "\u{3141}";
3012                         } elseif ( $code < 0xc0ac ) {
3013                                 return "\u{3142}";
3014                         } elseif ( $code < 0xc544 ) {
3015                                 return "\u{3145}";
3016                         } elseif ( $code < 0xc790 ) {
3017                                 return "\u{3147}";
3018                         } elseif ( $code < 0xcc28 ) {
3019                                 return "\u{3148}";
3020                         } elseif ( $code < 0xce74 ) {
3021                                 return "\u{314A}";
3022                         } elseif ( $code < 0xd0c0 ) {
3023                                 return "\u{314B}";
3024                         } elseif ( $code < 0xd30c ) {
3025                                 return "\u{314C}";
3026                         } elseif ( $code < 0xd558 ) {
3027                                 return "\u{314D}";
3028                         } else {
3029                                 return "\u{314E}";
3030                         }
3031                 } else {
3032                         return '';
3033                 }
3034         }
3035
3036         /**
3037          * Convert a UTF-8 string to normal form C. In Malayalam and Arabic, this
3038          * also cleans up certain backwards-compatible sequences, converting them
3039          * to the modern Unicode equivalent.
3040          *
3041          * This is language-specific for performance reasons only.
3042          *
3043          * @param string $s
3044          *
3045          * @return string
3046          */
3047         public function normalize( $s ) {
3048                 global $wgAllUnicodeFixes;
3049                 $s = UtfNormal\Validator::cleanUp( $s );
3050                 if ( $wgAllUnicodeFixes ) {
3051                         $s = $this->transformUsingPairFile( 'normalize-ar.php', $s );
3052                         $s = $this->transformUsingPairFile( 'normalize-ml.php', $s );
3053                 }
3054
3055                 return $s;
3056         }
3057
3058         /**
3059          * Transform a string using serialized data stored in the given file (which
3060          * must be in the serialized subdirectory of $IP). The file contains pairs
3061          * mapping source characters to destination characters.
3062          *
3063          * The data is cached in process memory. This will go faster if you have the
3064          * FastStringSearch extension.
3065          *
3066          * @param string $file
3067          * @param string $string
3068          *
3069          * @throws MWException
3070          * @return string
3071          */
3072         protected function transformUsingPairFile( $file, $string ) {
3073                 if ( !isset( $this->transformData[$file] ) ) {
3074                         global $IP;
3075                         $data = require "$IP/languages/data/{$file}";
3076                         $this->transformData[$file] = new ReplacementArray( $data );
3077                 }
3078                 return $this->transformData[$file]->replace( $string );
3079         }
3080
3081         /**
3082          * For right-to-left language support
3083          *
3084          * @return bool
3085          */
3086         function isRTL() {
3087                 return self::$dataCache->getItem( $this->mCode, 'rtl' );
3088         }
3089
3090         /**
3091          * Return the correct HTML 'dir' attribute value for this language.
3092          * @return string
3093          */
3094         function getDir() {
3095                 return $this->isRTL() ? 'rtl' : 'ltr';
3096         }
3097
3098         /**
3099          * Return 'left' or 'right' as appropriate alignment for line-start
3100          * for this language's text direction.
3101          *
3102          * Should be equivalent to CSS3 'start' text-align value....
3103          *
3104          * @return string
3105          */
3106         function alignStart() {
3107                 return $this->isRTL() ? 'right' : 'left';
3108         }
3109
3110         /**
3111          * Return 'right' or 'left' as appropriate alignment for line-end
3112          * for this language's text direction.
3113          *
3114          * Should be equivalent to CSS3 'end' text-align value....
3115          *
3116          * @return string
3117          */
3118         function alignEnd() {
3119                 return $this->isRTL() ? 'left' : 'right';
3120         }
3121
3122         /**
3123          * A hidden direction mark (LRM or RLM), depending on the language direction.
3124          * Unlike getDirMark(), this function returns the character as an HTML entity.
3125          * This function should be used when the output is guaranteed to be HTML,
3126          * because it makes the output HTML source code more readable. When
3127          * the output is plain text or can be escaped, getDirMark() should be used.
3128          *
3129          * @param bool $opposite Get the direction mark opposite to your language
3130          * @return string
3131          * @since 1.20
3132          */
3133         function getDirMarkEntity( $opposite = false ) {
3134                 if ( $opposite ) {
3135                         return $this->isRTL() ? '&lrm;' : '&rlm;';
3136                 }
3137                 return $this->isRTL() ? '&rlm;' : '&lrm;';
3138         }
3139
3140         /**
3141          * A hidden direction mark (LRM or RLM), depending on the language direction.
3142          * This function produces them as invisible Unicode characters and
3143          * the output may be hard to read and debug, so it should only be used
3144          * when the output is plain text or can be escaped. When the output is
3145          * HTML, use getDirMarkEntity() instead.
3146          *
3147          * @param bool $opposite Get the direction mark opposite to your language
3148          * @return string
3149          */
3150         function getDirMark( $opposite = false ) {
3151                 $lrm = "\u{200E}"; # LEFT-TO-RIGHT MARK, commonly abbreviated LRM
3152                 $rlm = "\u{200F}"; # RIGHT-TO-LEFT MARK, commonly abbreviated RLM
3153                 if ( $opposite ) {
3154                         return $this->isRTL() ? $lrm : $rlm;
3155                 }
3156                 return $this->isRTL() ? $rlm : $lrm;
3157         }
3158
3159         /**
3160          * @return array
3161          */
3162         function capitalizeAllNouns() {
3163                 return self::$dataCache->getItem( $this->mCode, 'capitalizeAllNouns' );
3164         }
3165
3166         /**
3167          * An arrow, depending on the language direction.
3168          *
3169          * @param string $direction The direction of the arrow: forwards (default),
3170          *   backwards, left, right, up, down.
3171          * @return string
3172          */
3173         function getArrow( $direction = 'forwards' ) {
3174                 switch ( $direction ) {
3175                         case 'forwards':
3176                                 return $this->isRTL() ? '←' : '→';
3177                         case 'backwards':
3178                                 return $this->isRTL() ? '→' : '←';
3179                         case 'left':
3180                                 return '←';
3181                         case 'right':
3182                                 return '→';
3183                         case 'up':
3184                                 return '↑';
3185                         case 'down':
3186                                 return '↓';
3187                 }
3188         }
3189
3190         /**
3191          * To allow "foo[[bar]]" to extend the link over the whole word "foobar"
3192          *
3193          * @return bool
3194          */
3195         function linkPrefixExtension() {
3196                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixExtension' );
3197         }
3198
3199         /**
3200          * Get all magic words from cache.
3201          * @return array
3202          */
3203         function getMagicWords() {
3204                 return self::$dataCache->getItem( $this->mCode, 'magicWords' );
3205         }
3206
3207         /**
3208          * Fill a MagicWord object with data from here
3209          *
3210          * @param MagicWord $mw
3211          */
3212         function getMagic( $mw ) {
3213                 $rawEntry = $this->mMagicExtensions[$mw->mId] ??
3214                         self::$dataCache->getSubitem( $this->mCode, 'magicWords', $mw->mId );
3215
3216                 if ( !is_array( $rawEntry ) ) {
3217                         wfWarn( "\"$rawEntry\" is not a valid magic word for \"$mw->mId\"" );
3218                 } else {
3219                         $mw->mCaseSensitive = $rawEntry[0];
3220                         $mw->mSynonyms = array_slice( $rawEntry, 1 );
3221                 }
3222         }
3223
3224         /**
3225          * Add magic words to the extension array
3226          *
3227          * @param array $newWords
3228          */
3229         function addMagicWordsByLang( $newWords ) {
3230                 $fallbackChain = $this->getFallbackLanguages();
3231                 $fallbackChain = array_reverse( $fallbackChain );
3232                 foreach ( $fallbackChain as $code ) {
3233                         if ( isset( $newWords[$code] ) ) {
3234                                 $this->mMagicExtensions = $newWords[$code] + $this->mMagicExtensions;
3235                         }
3236                 }
3237         }
3238
3239         /**
3240          * Get special page names, as an associative array
3241          *   canonical name => array of valid names, including aliases
3242          * @return array
3243          */
3244         function getSpecialPageAliases() {
3245                 // Cache aliases because it may be slow to load them
3246                 if ( is_null( $this->mExtendedSpecialPageAliases ) ) {
3247                         // Initialise array
3248                         $this->mExtendedSpecialPageAliases =
3249                                 self::$dataCache->getItem( $this->mCode, 'specialPageAliases' );
3250                 }
3251
3252                 return $this->mExtendedSpecialPageAliases;
3253         }
3254
3255         /**
3256          * Italic is unsuitable for some languages
3257          *
3258          * @param string $text The text to be emphasized.
3259          * @return string
3260          */
3261         function emphasize( $text ) {
3262                 return "<em>$text</em>";
3263         }
3264
3265         /**
3266          * Normally we output all numbers in plain en_US style, that is
3267          * 293,291.235 for twohundredninetythreethousand-twohundredninetyone
3268          * point twohundredthirtyfive. However this is not suitable for all
3269          * languages, some such as Bengali (bn) want ২,৯৩,২৯১.২৩৫ and others such as
3270          * Icelandic just want to use commas instead of dots, and dots instead
3271          * of commas like "293.291,235".
3272          *
3273          * An example of this function being called:
3274          * <code>
3275          * wfMessage( 'message' )->numParams( $num )->text()
3276          * </code>
3277          *
3278          * See $separatorTransformTable on MessageIs.php for
3279          * the , => . and . => , implementation.
3280          *
3281          * @todo check if it's viable to use localeconv() for the decimal separator thing.
3282          * @param int|float $number The string to be formatted, should be an integer
3283          *   or a floating point number.
3284          * @param bool $nocommafy Set to true for special numbers like dates
3285          * @return string
3286          */
3287         public function formatNum( $number, $nocommafy = false ) {
3288                 global $wgTranslateNumerals;
3289                 if ( !$nocommafy ) {
3290                         $number = $this->commafy( $number );
3291                         $s = $this->separatorTransformTable();
3292                         if ( $s ) {
3293                                 $number = strtr( $number, $s );
3294                         }
3295                 }
3296
3297                 if ( $wgTranslateNumerals ) {
3298                         $s = $this->digitTransformTable();
3299                         if ( $s ) {
3300                                 $number = strtr( $number, $s );
3301                         }
3302                 }
3303
3304                 return (string)$number;
3305         }
3306
3307         /**
3308          * Front-end for non-commafied formatNum
3309          *
3310          * @param int|float $number The string to be formatted, should be an integer
3311          *        or a floating point number.
3312          * @since 1.21
3313          * @return string
3314          */
3315         public function formatNumNoSeparators( $number ) {
3316                 return $this->formatNum( $number, true );
3317         }
3318
3319         /**
3320          * @param string $number
3321          * @return string
3322          */
3323         public function parseFormattedNumber( $number ) {
3324                 $s = $this->digitTransformTable();
3325                 if ( $s ) {
3326                         // eliminate empty array values such as ''. (T66347)
3327                         $s = array_filter( $s );
3328                         $number = strtr( $number, array_flip( $s ) );
3329                 }
3330
3331                 $s = $this->separatorTransformTable();
3332                 if ( $s ) {
3333                         // eliminate empty array values such as ''. (T66347)
3334                         $s = array_filter( $s );
3335                         $number = strtr( $number, array_flip( $s ) );
3336                 }
3337
3338                 $number = strtr( $number, [ ',' => '' ] );
3339                 return $number;
3340         }
3341
3342         /**
3343          * Adds commas to a given number
3344          * @since 1.19
3345          * @param mixed $number
3346          * @return string
3347          */
3348         function commafy( $number ) {
3349                 $digitGroupingPattern = $this->digitGroupingPattern();
3350                 $minimumGroupingDigits = $this->minimumGroupingDigits();
3351                 if ( $number === null ) {
3352                         return '';
3353                 }
3354
3355                 if ( !$digitGroupingPattern || $digitGroupingPattern === "###,###,###" ) {
3356                         // Default grouping is at thousands, use the same for ###,###,### pattern too.
3357                         // In some languages it's conventional not to insert a thousands separator
3358                         // in numbers that are four digits long (1000-9999).
3359                         if ( $minimumGroupingDigits ) {
3360                                 // Number of '#' characters after last comma in the grouping pattern.
3361                                 // The pattern is hardcoded here, but this would vary for different patterns.
3362                                 $primaryGroupingSize = 3;
3363                                 // Maximum length of a number to suppress digit grouping for.
3364                                 $maximumLength = $minimumGroupingDigits + $primaryGroupingSize - 1;
3365                                 if ( preg_match( '/^\-?\d{1,' . $maximumLength . '}(\.\d+)?$/', $number ) ) {
3366                                         return $number;
3367                                 }
3368                         }
3369                         return strrev( (string)preg_replace( '/(\d{3})(?=\d)(?!\d*\.)/', '$1,', strrev( $number ) ) );
3370                 } else {
3371                         // Ref: http://cldr.unicode.org/translation/number-patterns
3372                         $sign = "";
3373                         if ( intval( $number ) < 0 ) {
3374                                 // For negative numbers apply the algorithm like positive number and add sign.
3375                                 $sign = "-";
3376                                 $number = substr( $number, 1 );
3377                         }
3378                         $integerPart = [];
3379                         $decimalPart = [];
3380                         $numMatches = preg_match_all( "/(#+)/", $digitGroupingPattern, $matches );
3381                         preg_match( "/\d+/", $number, $integerPart );
3382                         preg_match( "/\.\d*/", $number, $decimalPart );
3383                         $groupedNumber = ( count( $decimalPart ) > 0 ) ? $decimalPart[0] : "";
3384                         if ( $groupedNumber === $number ) {
3385                                 // the string does not have any number part. Eg: .12345
3386                                 return $sign . $groupedNumber;
3387                         }
3388                         $start = $end = ( $integerPart ) ? strlen( $integerPart[0] ) : 0;
3389                         while ( $start > 0 ) {
3390                                 $match = $matches[0][$numMatches - 1];
3391                                 $matchLen = strlen( $match );
3392                                 $start = $end - $matchLen;
3393                                 if ( $start < 0 ) {
3394                                         $start = 0;
3395                                 }
3396                                 $groupedNumber = substr( $number, $start, $end - $start ) . $groupedNumber;
3397                                 $end = $start;
3398                                 if ( $numMatches > 1 ) {
3399                                         // use the last pattern for the rest of the number
3400                                         $numMatches--;
3401                                 }
3402                                 if ( $start > 0 ) {
3403                                         $groupedNumber = "," . $groupedNumber;
3404                                 }
3405                         }
3406                         return $sign . $groupedNumber;
3407                 }
3408         }
3409
3410         /**
3411          * @return string
3412          */
3413         function digitGroupingPattern() {
3414                 return self::$dataCache->getItem( $this->mCode, 'digitGroupingPattern' );
3415         }
3416
3417         /**
3418          * @return array
3419          */
3420         function digitTransformTable() {
3421                 return self::$dataCache->getItem( $this->mCode, 'digitTransformTable' );
3422         }
3423
3424         /**
3425          * @return array
3426          */
3427         function separatorTransformTable() {
3428                 return self::$dataCache->getItem( $this->mCode, 'separatorTransformTable' );
3429         }
3430
3431         /**
3432          * @return int|null
3433          */
3434         function minimumGroupingDigits() {
3435                 return self::$dataCache->getItem( $this->mCode, 'minimumGroupingDigits' );
3436         }
3437
3438         /**
3439          * Take a list of strings and build a locale-friendly comma-separated
3440          * list, using the local comma-separator message.
3441          * The last two strings are chained with an "and".
3442          *
3443          * @param string[] $list
3444          * @return string
3445          */
3446         public function listToText( array $list ) {
3447                 $itemCount = count( $list );
3448                 if ( $itemCount < 1 ) {
3449                         return '';
3450                 }
3451                 $text = array_pop( $list );
3452                 if ( $itemCount > 1 ) {
3453                         $and = $this->msg( 'and' )->escaped();
3454                         $space = $this->msg( 'word-separator' )->escaped();
3455                         $comma = '';
3456                         if ( $itemCount > 2 ) {
3457                                 $comma = $this->msg( 'comma-separator' )->escaped();
3458                         }
3459                         $text = implode( $comma, $list ) . $and . $space . $text;
3460                 }
3461                 return $text;
3462         }
3463
3464         /**
3465          * Take a list of strings and build a locale-friendly comma-separated
3466          * list, using the local comma-separator message.
3467          * @param string[] $list Array of strings to put in a comma list
3468          * @return string
3469          */
3470         function commaList( array $list ) {
3471                 return implode(
3472                         wfMessage( 'comma-separator' )->inLanguage( $this )->escaped(),
3473                         $list
3474                 );
3475         }
3476
3477         /**
3478          * Take a list of strings and build a locale-friendly semicolon-separated
3479          * list, using the local semicolon-separator message.
3480          * @param string[] $list Array of strings to put in a semicolon list
3481          * @return string
3482          */
3483         function semicolonList( array $list ) {
3484                 return implode(
3485                         wfMessage( 'semicolon-separator' )->inLanguage( $this )->escaped(),
3486                         $list
3487                 );
3488         }
3489
3490         /**
3491          * Same as commaList, but separate it with the pipe instead.
3492          * @param string[] $list Array of strings to put in a pipe list
3493          * @return string
3494          */
3495         function pipeList( array $list ) {
3496                 return implode(
3497                         wfMessage( 'pipe-separator' )->inLanguage( $this )->escaped(),
3498                         $list
3499                 );
3500         }
3501
3502         /**
3503          * Truncate a string to a specified length in bytes, appending an optional
3504          * string (e.g. for ellipsis)
3505          *
3506          * If $length is negative, the string will be truncated from the beginning
3507          *
3508          * @since 1.31
3509          *
3510          * @param string $string String to truncate
3511          * @param int $length Maximum length in bytes
3512          * @param string $ellipsis String to append to the end of truncated text
3513          * @param bool $adjustLength Subtract length of ellipsis from $length
3514          *
3515          * @return string
3516          */
3517         function truncateForDatabase( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3518                 return $this->truncateInternal(
3519                         $string, $length, $ellipsis, $adjustLength, 'strlen', 'substr'
3520                 );
3521         }
3522
3523         /**
3524          * Truncate a string to a specified number of characters, appending an optional
3525          * string (e.g. for ellipsis).
3526          *
3527          * This provides multibyte version of truncateForDatabase() method of this class,
3528          * suitable for truncation based on number of characters, instead of number of bytes.
3529          *
3530          * If $length is negative, the string will be truncated from the beginning.
3531          *
3532          * @since 1.31
3533          *
3534          * @param string $string String to truncate
3535          * @param int $length Maximum number of characters
3536          * @param string $ellipsis String to append to the end of truncated text
3537          * @param bool $adjustLength Subtract length of ellipsis from $length
3538          *
3539          * @return string
3540          */
3541         function truncateForVisual( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3542                 // Passing encoding to mb_strlen and mb_substr is optional.
3543                 // Encoding defaults to mb_internal_encoding(), which is set to UTF-8 in Setup.php, so
3544                 // explicit specification of encoding is skipped.
3545                 // Note: Both multibyte methods are callables invoked in truncateInternal.
3546                 return $this->truncateInternal(
3547                         $string, $length, $ellipsis, $adjustLength, 'mb_strlen', 'mb_substr'
3548                 );
3549         }
3550
3551         /**
3552          * Internal method used for truncation. This method abstracts text truncation into
3553          * one common method, allowing users to provide length measurement function and
3554          * function for finding substring.
3555          *
3556          * For usages, see truncateForDatabase and truncateForVisual.
3557          *
3558          * @param string $string String to truncate
3559          * @param int $length Maximum length of final text
3560          * @param string $ellipsis String to append to the end of truncated text
3561          * @param bool $adjustLength Subtract length of ellipsis from $length
3562          * @param callable $measureLength Callable function used for determining the length of text
3563          * @param callable $getSubstring Callable function used for getting the substrings
3564          *
3565          * @return string
3566          */
3567         private function truncateInternal(
3568                 $string, $length, $ellipsis, $adjustLength, callable $measureLength, callable $getSubstring
3569         ) {
3570                 # Check if there is no need to truncate
3571                 if ( $measureLength( $string ) <= abs( $length ) ) {
3572                         return $string; // no need to truncate
3573                 }
3574
3575                 # Use the localized ellipsis character
3576                 if ( $ellipsis == '...' ) {
3577                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3578                 }
3579                 if ( $length == 0 ) {
3580                         return $ellipsis; // convention
3581                 }
3582
3583                 $stringOriginal = $string;
3584                 # If ellipsis length is >= $length then we can't apply $adjustLength
3585                 if ( $adjustLength && $measureLength( $ellipsis ) >= abs( $length ) ) {
3586                         $string = $ellipsis; // this can be slightly unexpected
3587                 # Otherwise, truncate and add ellipsis...
3588                 } else {
3589                         $ellipsisLength = $adjustLength ? $measureLength( $ellipsis ) : 0;
3590                         if ( $length > 0 ) {
3591                                 $length -= $ellipsisLength;
3592                                 $string = $getSubstring( $string, 0, $length ); // xyz...
3593                                 $string = $this->removeBadCharLast( $string );
3594                                 $string = rtrim( $string ) . $ellipsis;
3595                         } else {
3596                                 $length += $ellipsisLength;
3597                                 $string = $getSubstring( $string, $length ); // ...xyz
3598                                 $string = $this->removeBadCharFirst( $string );
3599                                 $string = $ellipsis . ltrim( $string );
3600                         }
3601                 }
3602
3603                 # Do not truncate if the ellipsis makes the string longer/equal (T24181).
3604                 # This check is *not* redundant if $adjustLength, due to the single case where
3605                 # LEN($ellipsis) > ABS($limit arg); $stringOriginal could be shorter than $string.
3606                 if ( $measureLength( $string ) < $measureLength( $stringOriginal ) ) {
3607                         return $string;
3608                 } else {
3609                         return $stringOriginal;
3610                 }
3611         }
3612
3613         /**
3614          * Remove bytes that represent an incomplete Unicode character
3615          * at the end of string (e.g. bytes of the char are missing)
3616          *
3617          * @param string $string
3618          * @return string
3619          */
3620         protected function removeBadCharLast( $string ) {
3621                 if ( $string != '' ) {
3622                         $char = ord( $string[strlen( $string ) - 1] );
3623                         $m = [];
3624                         if ( $char >= 0xc0 ) {
3625                                 # We got the first byte only of a multibyte char; remove it.
3626                                 $string = substr( $string, 0, -1 );
3627                         } elseif ( $char >= 0x80 &&
3628                                 // Use the /s modifier (PCRE_DOTALL) so (.*) also matches newlines
3629                                 preg_match( '/^(.*)(?:[\xe0-\xef][\x80-\xbf]|' .
3630                                         '[\xf0-\xf7][\x80-\xbf]{1,2})$/s', $string, $m )
3631                         ) {
3632                                 # We chopped in the middle of a character; remove it
3633                                 $string = $m[1];
3634                         }
3635                 }
3636                 return $string;
3637         }
3638
3639         /**
3640          * Remove bytes that represent an incomplete Unicode character
3641          * at the start of string (e.g. bytes of the char are missing)
3642          *
3643          * @param string $string
3644          * @return string
3645          */
3646         protected function removeBadCharFirst( $string ) {
3647                 if ( $string != '' ) {
3648                         $char = ord( $string[0] );
3649                         if ( $char >= 0x80 && $char < 0xc0 ) {
3650                                 # We chopped in the middle of a character; remove the whole thing
3651                                 $string = preg_replace( '/^[\x80-\xbf]+/', '', $string );
3652                         }
3653                 }
3654                 return $string;
3655         }
3656
3657         /**
3658          * Truncate a string of valid HTML to a specified length in bytes,
3659          * appending an optional string (e.g. for ellipses), and return valid HTML
3660          *
3661          * This is only intended for styled/linked text, such as HTML with
3662          * tags like <span> and <a>, were the tags are self-contained (valid HTML).
3663          * Also, this will not detect things like "display:none" CSS.
3664          *
3665          * Note: since 1.18 you do not need to leave extra room in $length for ellipses.
3666          *
3667          * @param string $text HTML string to truncate
3668          * @param int $length (zero/positive) Maximum length (including ellipses)
3669          * @param string $ellipsis String to append to the truncated text
3670          * @return string
3671          */
3672         function truncateHtml( $text, $length, $ellipsis = '...' ) {
3673                 # Use the localized ellipsis character
3674                 if ( $ellipsis == '...' ) {
3675                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3676                 }
3677                 # Check if there is clearly no need to truncate
3678                 if ( $length <= 0 ) {
3679                         return $ellipsis; // no text shown, nothing to format (convention)
3680                 } elseif ( strlen( $text ) <= $length ) {
3681                         return $text; // string short enough even *with* HTML (short-circuit)
3682                 }
3683
3684                 $dispLen = 0; // innerHTML legth so far
3685                 $testingEllipsis = false; // checking if ellipses will make string longer/equal?
3686                 $tagType = 0; // 0-open, 1-close
3687                 $bracketState = 0; // 1-tag start, 2-tag name, 0-neither
3688                 $entityState = 0; // 0-not entity, 1-entity
3689                 $tag = $ret = ''; // accumulated tag name, accumulated result string
3690                 $openTags = []; // open tag stack
3691                 $maybeState = null; // possible truncation state
3692
3693                 $textLen = strlen( $text );
3694                 $neLength = max( 0, $length - strlen( $ellipsis ) ); // non-ellipsis len if truncated
3695                 for ( $pos = 0; true; ++$pos ) {
3696                         # Consider truncation once the display length has reached the maximim.
3697                         # We check if $dispLen > 0 to grab tags for the $neLength = 0 case.
3698                         # Check that we're not in the middle of a bracket/entity...
3699                         if ( $dispLen && $dispLen >= $neLength && $bracketState == 0 && !$entityState ) {
3700                                 if ( !$testingEllipsis ) {
3701                                         $testingEllipsis = true;
3702                                         # Save where we are; we will truncate here unless there turn out to
3703                                         # be so few remaining characters that truncation is not necessary.
3704                                         if ( !$maybeState ) { // already saved? ($neLength = 0 case)
3705                                                 $maybeState = [ $ret, $openTags ]; // save state
3706                                         }
3707                                 } elseif ( $dispLen > $length && $dispLen > strlen( $ellipsis ) ) {
3708                                         # String in fact does need truncation, the truncation point was OK.
3709                                         // @phan-suppress-next-line PhanTypeInvalidExpressionArrayDestructuring
3710                                         list( $ret, $openTags ) = $maybeState; // reload state
3711                                         $ret = $this->removeBadCharLast( $ret ); // multi-byte char fix
3712                                         $ret .= $ellipsis; // add ellipsis
3713                                         break;
3714                                 }
3715                         }
3716                         if ( $pos >= $textLen ) {
3717                                 break; // extra iteration just for above checks
3718                         }
3719
3720                         # Read the next char...
3721                         $ch = $text[$pos];
3722                         $lastCh = $pos ? $text[$pos - 1] : '';
3723                         $ret .= $ch; // add to result string
3724                         if ( $ch == '<' ) {
3725                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags ); // for bad HTML
3726                                 $entityState = 0; // for bad HTML
3727                                 $bracketState = 1; // tag started (checking for backslash)
3728                         } elseif ( $ch == '>' ) {
3729                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags );
3730                                 $entityState = 0; // for bad HTML
3731                                 $bracketState = 0; // out of brackets
3732                         } elseif ( $bracketState == 1 ) {
3733                                 if ( $ch == '/' ) {
3734                                         $tagType = 1; // close tag (e.g. "</span>")
3735                                 } else {
3736                                         $tagType = 0; // open tag (e.g. "<span>")
3737                                         $tag .= $ch;
3738                                 }
3739                                 $bracketState = 2; // building tag name
3740                         } elseif ( $bracketState == 2 ) {
3741                                 if ( $ch != ' ' ) {
3742                                         $tag .= $ch;
3743                                 } else {
3744                                         // Name found (e.g. "<a href=..."), add on tag attributes...
3745                                         $pos += $this->truncate_skip( $ret, $text, "<>", $pos + 1 );
3746                                 }
3747                         } elseif ( $bracketState == 0 ) {
3748                                 if ( $entityState ) {
3749                                         if ( $ch == ';' ) {
3750                                                 $entityState = 0;
3751                                                 $dispLen++; // entity is one displayed char
3752                                         }
3753                                 } else {
3754                                         if ( $neLength == 0 && !$maybeState ) {
3755                                                 // Save state without $ch. We want to *hit* the first
3756                                                 // display char (to get tags) but not *use* it if truncating.
3757                                                 $maybeState = [ substr( $ret, 0, -1 ), $openTags ];
3758                                         }
3759                                         if ( $ch == '&' ) {
3760                                                 $entityState = 1; // entity found, (e.g. "&#160;")
3761                                         } else {
3762                                                 $dispLen++; // this char is displayed
3763                                                 // Add the next $max display text chars after this in one swoop...
3764                                                 $max = ( $testingEllipsis ? $length : $neLength ) - $dispLen;
3765                                                 $skipped = $this->truncate_skip( $ret, $text, "<>&", $pos + 1, $max );
3766                                                 $dispLen += $skipped;
3767                                                 $pos += $skipped;
3768                                         }
3769                                 }
3770                         }
3771                 }
3772                 // Close the last tag if left unclosed by bad HTML
3773                 $this->truncate_endBracket( $tag, $text[$textLen - 1], $tagType, $openTags );
3774                 while ( count( $openTags ) > 0 ) {
3775                         $ret .= '</' . array_pop( $openTags ) . '>'; // close open tags
3776                 }
3777                 return $ret;
3778         }
3779
3780         /**
3781          * truncateHtml() helper function
3782          * like strcspn() but adds the skipped chars to $ret
3783          *
3784          * @param string $ret
3785          * @param string $text
3786          * @param string $search
3787          * @param int $start
3788          * @param null|int $len
3789          * @return int
3790          */
3791         private function truncate_skip( &$ret, $text, $search, $start, $len = null ) {
3792                 if ( $len === null ) {
3793                         $len = -1; // -1 means "no limit" for strcspn
3794                 } elseif ( $len < 0 ) {
3795                         $len = 0; // sanity
3796                 }
3797                 $skipCount = 0;
3798                 if ( $start < strlen( $text ) ) {
3799                         $skipCount = strcspn( $text, $search, $start, $len );
3800                         $ret .= substr( $text, $start, $skipCount );
3801                 }
3802                 return $skipCount;
3803         }
3804
3805         /**
3806          * truncateHtml() helper function
3807          * (a) push or pop $tag from $openTags as needed
3808          * (b) clear $tag value
3809          * @param string &$tag Current HTML tag name we are looking at
3810          * @param int $tagType (0-open tag, 1-close tag)
3811          * @param string $lastCh Character before the '>' that ended this tag
3812          * @param array &$openTags Open tag stack (not accounting for $tag)
3813          */
3814         private function truncate_endBracket( &$tag, $tagType, $lastCh, &$openTags ) {
3815                 $tag = ltrim( $tag );
3816                 if ( $tag != '' ) {
3817                         if ( $tagType == 0 && $lastCh != '/' ) {
3818                                 $openTags[] = $tag; // tag opened (didn't close itself)
3819                         } elseif ( $tagType == 1 ) {
3820                                 if ( $openTags && $tag == $openTags[count( $openTags ) - 1] ) {
3821                                         array_pop( $openTags ); // tag closed
3822                                 }
3823                         }
3824                         $tag = '';
3825                 }
3826         }
3827
3828         /**
3829          * Grammatical transformations, needed for inflected languages
3830          * Invoked by putting {{grammar:case|word}} in a message
3831          *
3832          * @param string $word
3833          * @param string $case
3834          * @return string
3835          */
3836         function convertGrammar( $word, $case ) {
3837                 global $wgGrammarForms;
3838                 if ( isset( $wgGrammarForms[$this->getCode()][$case][$word] ) ) {
3839                         return $wgGrammarForms[$this->getCode()][$case][$word];
3840                 }
3841
3842                 $grammarTransformations = $this->getGrammarTransformations();
3843
3844                 if ( isset( $grammarTransformations[$case] ) ) {
3845                         $forms = $grammarTransformations[$case];
3846
3847                         // Some names of grammar rules are aliases for other rules.
3848                         // In such cases the value is a string rather than object,
3849                         // so load the actual rules.
3850                         if ( is_string( $forms ) ) {
3851                                 $forms = $grammarTransformations[$forms];
3852                         }
3853
3854                         foreach ( array_values( $forms ) as $rule ) {
3855                                 $form = $rule[0];
3856
3857                                 if ( $form === '@metadata' ) {
3858                                         continue;
3859                                 }
3860
3861                                 $replacement = $rule[1];
3862
3863                                 $regex = '/' . addcslashes( $form, '/' ) . '/u';
3864                                 $patternMatches = preg_match( $regex, $word );
3865
3866                                 if ( $patternMatches === false ) {
3867                                         wfLogWarning(
3868                                                 'An error occurred while processing grammar. ' .
3869                                                 "Word: '$word'. Regex: /$form/."
3870                                         );
3871                                 } elseif ( $patternMatches === 1 ) {
3872                                         $word = preg_replace( $regex, $replacement, $word );
3873
3874                                         break;
3875                                 }
3876                         }
3877                 }
3878
3879                 return $word;
3880         }
3881
3882         /**
3883          * Get the grammar forms for the content language
3884          * @return array Array of grammar forms
3885          * @since 1.20
3886          */
3887         function getGrammarForms() {
3888                 global $wgGrammarForms;
3889                 if ( isset( $wgGrammarForms[$this->getCode()] )
3890                         && is_array( $wgGrammarForms[$this->getCode()] )
3891                 ) {
3892                         return $wgGrammarForms[$this->getCode()];
3893                 }
3894
3895                 return [];
3896         }
3897
3898         /**
3899          * Get the grammar transformations data for the language.
3900          * Used like grammar forms, with {{GRAMMAR}} and cases,
3901          * but uses pairs of regexes and replacements instead of code.
3902          *
3903          * @return array[] Array of grammar transformations.
3904          * @throws MWException
3905          * @since 1.28
3906          */
3907         public function getGrammarTransformations() {
3908                 $languageCode = $this->getCode();
3909
3910                 if ( self::$grammarTransformations === null ) {
3911                         self::$grammarTransformations = new MapCacheLRU( 10 );
3912                 }
3913
3914                 if ( self::$grammarTransformations->has( $languageCode ) ) {
3915                         return self::$grammarTransformations->get( $languageCode );
3916                 }
3917
3918                 $data = [];
3919
3920                 $grammarDataFile = __DIR__ . "/data/grammarTransformations/$languageCode.json";
3921                 if ( is_readable( $grammarDataFile ) ) {
3922                         $data = FormatJson::decode(
3923                                 file_get_contents( $grammarDataFile ),
3924                                 true
3925                         );
3926
3927                         if ( $data === null ) {
3928                                 throw new MWException( "Invalid grammar data for \"$languageCode\"." );
3929                         }
3930
3931                         self::$grammarTransformations->set( $languageCode, $data );
3932                 }
3933
3934                 return $data;
3935         }
3936
3937         /**
3938          * Provides an alternative text depending on specified gender.
3939          * Usage {{gender:username|masculine|feminine|unknown}}.
3940          * username is optional, in which case the gender of current user is used,
3941          * but only in (some) interface messages; otherwise default gender is used.
3942          *
3943          * If no forms are given, an empty string is returned. If only one form is
3944          * given, it will be returned unconditionally. These details are implied by
3945          * the caller and cannot be overridden in subclasses.
3946          *
3947          * If three forms are given, the default is to use the third (unknown) form.
3948          * If fewer than three forms are given, the default is to use the first (masculine) form.
3949          * These details can be overridden in subclasses.
3950          *
3951          * @param string $gender
3952          * @param array $forms
3953          *
3954          * @return string
3955          */
3956         function gender( $gender, $forms ) {
3957                 if ( !count( $forms ) ) {
3958                         return '';
3959                 }
3960                 $forms = $this->preConvertPlural( $forms, 2 );
3961                 if ( $gender === 'male' ) {
3962                         return $forms[0];
3963                 }
3964                 if ( $gender === 'female' ) {
3965                         return $forms[1];
3966                 }
3967                 return $forms[2] ?? $forms[0];
3968         }
3969
3970         /**
3971          * Plural form transformations, needed for some languages.
3972          * For example, there are 3 form of plural in Russian and Polish,
3973          * depending on "count mod 10". See [[w:Plural]]
3974          * For English it is pretty simple.
3975          *
3976          * Invoked by putting {{plural:count|wordform1|wordform2}}
3977          * or {{plural:count|wordform1|wordform2|wordform3}}
3978          *
3979          * Example: {{plural:{{NUMBEROFARTICLES}}|article|articles}}
3980          *
3981          * @param int $count Non-localized number
3982          * @param array $forms Different plural forms
3983          * @return string Correct form of plural for $count in this language
3984          */
3985         function convertPlural( $count, $forms ) {
3986                 // Handle explicit n=pluralform cases
3987                 $forms = $this->handleExplicitPluralForms( $count, $forms );
3988                 if ( is_string( $forms ) ) {
3989                         return $forms;
3990                 }
3991                 if ( !count( $forms ) ) {
3992                         return '';
3993                 }
3994
3995                 $pluralForm = $this->getPluralRuleIndexNumber( $count );
3996                 $pluralForm = min( $pluralForm, count( $forms ) - 1 );
3997                 return $forms[$pluralForm];
3998         }
3999
4000         /**
4001          * Handles explicit plural forms for Language::convertPlural()
4002          *
4003          * In {{PLURAL:$1|0=nothing|one|many}}, 0=nothing will be returned if $1 equals zero.
4004          * If an explicitly defined plural form matches the $count, then
4005          * string value returned, otherwise array returned for further consideration
4006          * by CLDR rules or overridden convertPlural().
4007          *
4008          * @since 1.23
4009          *
4010          * @param int $count Non-localized number
4011          * @param array $forms Different plural forms
4012          *
4013          * @return array|string
4014          */
4015         protected function handleExplicitPluralForms( $count, array $forms ) {
4016                 foreach ( $forms as $index => $form ) {
4017                         if ( preg_match( '/\d+=/i', $form ) ) {
4018                                 $pos = strpos( $form, '=' );
4019                                 if ( substr( $form, 0, $pos ) === (string)$count ) {
4020                                         return substr( $form, $pos + 1 );
4021                                 }
4022                                 unset( $forms[$index] );
4023                         }
4024                 }
4025                 return array_values( $forms );
4026         }
4027
4028         /**
4029          * Checks that convertPlural was given an array and pads it to requested
4030          * amount of forms by copying the last one.
4031          *
4032          * @param array $forms
4033          * @param int $count How many forms should there be at least
4034          * @return array Padded array of forms
4035          */
4036         protected function preConvertPlural( /* Array */ $forms, $count ) {
4037                 return array_pad( $forms, $count, end( $forms ) );
4038         }
4039
4040         /**
4041          * Wraps argument with unicode control characters for directionality safety
4042          *
4043          * This solves the problem where directionality-neutral characters at the edge of
4044          * the argument string get interpreted with the wrong directionality from the
4045          * enclosing context, giving renderings that look corrupted like "(Ben_(WMF".
4046          *
4047          * The wrapping is LRE...PDF or RLE...PDF, depending on the detected
4048          * directionality of the argument string, using the BIDI algorithm's own "First
4049          * strong directional codepoint" rule. Essentially, this works round the fact that
4050          * there is no embedding equivalent of U+2068 FSI (isolation with heuristic
4051          * direction inference). The latter is cleaner but still not widely supported.
4052          *
4053          * @param string $text Text to wrap
4054          * @return string Text, wrapped in LRE...PDF or RLE...PDF or nothing
4055          */
4056         public function embedBidi( $text = '' ) {
4057                 $dir = self::strongDirFromContent( $text );
4058                 if ( $dir === 'ltr' ) {
4059                         // Wrap in LEFT-TO-RIGHT EMBEDDING ... POP DIRECTIONAL FORMATTING
4060                         return self::$lre . $text . self::$pdf;
4061                 }
4062                 if ( $dir === 'rtl' ) {
4063                         // Wrap in RIGHT-TO-LEFT EMBEDDING ... POP DIRECTIONAL FORMATTING
4064                         return self::$rle . $text . self::$pdf;
4065                 }
4066                 // No strong directionality: do not wrap
4067                 return $text;
4068         }
4069
4070         /**
4071          * @todo Maybe translate block durations.  Note that this function is somewhat misnamed: it
4072          * deals with translating the *duration* ("1 week", "4 days", etc), not the expiry time
4073          * (which is an absolute timestamp). Please note: do NOT add this blindly, as it is used
4074          * on old expiry lengths recorded in log entries. You'd need to provide the start date to
4075          * match up with it.
4076          *
4077          * @param string $str The validated block duration in English
4078          * @param User|null $user User object to use timezone from or null for $wgUser
4079          * @param int $now Current timestamp, for formatting relative block durations
4080          * @return string Somehow translated block duration
4081          * @see LanguageFi.php for example implementation
4082          */
4083         function translateBlockExpiry( $str, User $user = null, $now = 0 ) {
4084                 $duration = SpecialBlock::getSuggestedDurations( $this );
4085                 foreach ( $duration as $show => $value ) {
4086                         if ( strcmp( $str, $value ) == 0 ) {
4087                                 return htmlspecialchars( trim( $show ) );
4088                         }
4089                 }
4090
4091                 if ( wfIsInfinity( $str ) ) {
4092                         foreach ( $duration as $show => $value ) {
4093                                 if ( wfIsInfinity( $value ) ) {
4094                                         return htmlspecialchars( trim( $show ) );
4095                                 }
4096                         }
4097                 }
4098
4099                 // If all else fails, return a standard duration or timestamp description.
4100                 $time = strtotime( $str, $now );
4101                 if ( $time === false ) { // Unknown format. Return it as-is in case.
4102                         return $str;
4103                 } elseif ( $time !== strtotime( $str, $now + 1 ) ) { // It's a relative timestamp.
4104                         // The result differs based on current time, so the difference
4105                         // is a fixed duration length.
4106                         return $this->formatDuration( $time - $now );
4107                 } else { // It's an absolute timestamp.
4108                         if ( $time === 0 ) {
4109                                 // wfTimestamp() handles 0 as current time instead of epoch.
4110                                 $time = '19700101000000';
4111                         }
4112                         if ( $user ) {
4113                                 return $this->userTimeAndDate( $time, $user );
4114                         }
4115                         return $this->timeanddate( $time );
4116                 }
4117         }
4118
4119         /**
4120          * languages like Chinese need to be segmented in order for the diff
4121          * to be of any use
4122          *
4123          * @param string $text
4124          * @return string
4125          */
4126         public function segmentForDiff( $text ) {
4127                 return $text;
4128         }
4129
4130         /**
4131          * and unsegment to show the result
4132          *
4133          * @param string $text
4134          * @return string
4135          */
4136         public function unsegmentForDiff( $text ) {
4137                 return $text;
4138         }
4139
4140         /**
4141          * Return the LanguageConverter used in the Language
4142          *
4143          * @since 1.19
4144          * @return LanguageConverter
4145          */
4146         public function getConverter() {
4147                 return $this->mConverter;
4148         }
4149
4150         /**
4151          * convert text to a variant
4152          *
4153          * @param string $text text to convert
4154          * @param string|bool $variant variant to convert to, or false to use the user's preferred
4155          *      variant (if logged in), or the project default variant
4156          * @return string the converted string
4157          */
4158         public function autoConvert( $text, $variant = false ) {
4159                 return $this->mConverter->autoConvert( $text, $variant );
4160         }
4161
4162         /**
4163          * convert text to all supported variants
4164          *
4165          * @param string $text
4166          * @return array
4167          */
4168         public function autoConvertToAllVariants( $text ) {
4169                 return $this->mConverter->autoConvertToAllVariants( $text );
4170         }
4171
4172         /**
4173          * convert text to different variants of a language.
4174          *
4175          * @warning Glossary state is maintained between calls. This means
4176          *  if you pass unescaped text to this method it can cause an XSS
4177          *  in later calls to this method, even if the later calls have properly
4178          *  escaped the input. Never feed this method user controlled text that
4179          *  is not properly escaped!
4180          * @param string $text Content that has been already escaped for use in HTML
4181          * @return string HTML
4182          */
4183         public function convert( $text ) {
4184                 return $this->mConverter->convert( $text );
4185         }
4186
4187         /**
4188          * Convert a Title object to a string in the preferred variant
4189          *
4190          * @param Title $title
4191          * @return string
4192          */
4193         public function convertTitle( $title ) {
4194                 return $this->mConverter->convertTitle( $title );
4195         }
4196
4197         /**
4198          * Convert a namespace index to a string in the preferred variant
4199          *
4200          * @param int $ns namespace index (https://www.mediawiki.org/wiki/Manual:Namespace)
4201          * @param string|null $variant variant to convert to, or null to use the user's preferred
4202          *      variant (if logged in), or the project default variant
4203          * @return string a string representation of the namespace
4204          */
4205         public function convertNamespace( $ns, $variant = null ) {
4206                 return $this->mConverter->convertNamespace( $ns, $variant );
4207         }
4208
4209         /**
4210          * Check if this is a language with variants
4211          *
4212          * @return bool
4213          */
4214         public function hasVariants() {
4215                 return count( $this->getVariants() ) > 1;
4216         }
4217
4218         /**
4219          * Strict check if the language has the specific variant.
4220          *
4221          * Compare to LanguageConverter::validateVariant() which does a more
4222          * lenient check and attempts to coerce the given code to a valid one.
4223          *
4224          * @since 1.19
4225          * @param string $variant
4226          * @return bool
4227          */
4228         public function hasVariant( $variant ) {
4229                 return $variant && ( $variant === $this->mConverter->validateVariant( $variant ) );
4230         }
4231
4232         /**
4233          * Perform output conversion on a string, and encode for safe HTML output.
4234          * @param string $text Text to be converted
4235          * @return string
4236          * @todo this should get integrated somewhere sane
4237          */
4238         public function convertHtml( $text ) {
4239                 return htmlspecialchars( $this->convert( $text ) );
4240         }
4241
4242         /**
4243          * @param string $key
4244          * @return string
4245          */
4246         public function convertCategoryKey( $key ) {
4247                 return $this->mConverter->convertCategoryKey( $key );
4248         }
4249
4250         /**
4251          * Get the list of variants supported by this language
4252          * see sample implementation in LanguageZh.php
4253          *
4254          * @return string[] An array of language codes
4255          */
4256         public function getVariants() {
4257                 return $this->mConverter->getVariants();
4258         }
4259
4260         /**
4261          * @return string
4262          */
4263         public function getPreferredVariant() {
4264                 return $this->mConverter->getPreferredVariant();
4265         }
4266
4267         /**
4268          * @return string
4269          */
4270         public function getDefaultVariant() {
4271                 return $this->mConverter->getDefaultVariant();
4272         }
4273
4274         /**
4275          * @return string
4276          */
4277         public function getURLVariant() {
4278                 return $this->mConverter->getURLVariant();
4279         }
4280
4281         /**
4282          * If a language supports multiple variants, it is
4283          * possible that non-existing link in one variant
4284          * actually exists in another variant. this function
4285          * tries to find it. See e.g. LanguageZh.php
4286          * The input parameters may be modified upon return
4287          *
4288          * @param string &$link The name of the link
4289          * @param Title &$nt The title object of the link
4290          * @param bool $ignoreOtherCond To disable other conditions when
4291          *   we need to transclude a template or update a category's link
4292          */
4293         public function findVariantLink( &$link, &$nt, $ignoreOtherCond = false ) {
4294                 $this->mConverter->findVariantLink( $link, $nt, $ignoreOtherCond );
4295         }
4296
4297         /**
4298          * returns language specific options used by User::getPageRenderHash()
4299          * for example, the preferred language variant
4300          *
4301          * @return string
4302          */
4303         function getExtraHashOptions() {
4304                 return $this->mConverter->getExtraHashOptions();
4305         }
4306
4307         /**
4308          * For languages that support multiple variants, the title of an
4309          * article may be displayed differently in different variants. this
4310          * function returns the apporiate title defined in the body of the article.
4311          *
4312          * @return string
4313          */
4314         public function getParsedTitle() {
4315                 return $this->mConverter->getParsedTitle();
4316         }
4317
4318         /**
4319          * Refresh the cache of conversion tables when
4320          * MediaWiki:Conversiontable* is updated.
4321          *
4322          * @param Title $title The Title of the page being updated
4323          */
4324         public function updateConversionTable( Title $title ) {
4325                 $this->mConverter->updateConversionTable( $title );
4326         }
4327
4328         /**
4329          * A regular expression to match legal word-trailing characters
4330          * which should be merged onto a link of the form [[foo]]bar.
4331          *
4332          * @return string
4333          */
4334         public function linkTrail() {
4335                 return self::$dataCache->getItem( $this->mCode, 'linkTrail' );
4336         }
4337
4338         /**
4339          * A regular expression character set to match legal word-prefixing
4340          * characters which should be merged onto a link of the form foo[[bar]].
4341          *
4342          * @return string
4343          */
4344         public function linkPrefixCharset() {
4345                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixCharset' );
4346         }
4347
4348         /**
4349          * Get the "parent" language which has a converter to convert a "compatible" language
4350          * (in another variant) to this language (eg. zh for zh-cn, but not en for en-gb).
4351          *
4352          * @return Language|null
4353          * @since 1.22
4354          */
4355         public function getParentLanguage() {
4356                 if ( $this->mParentLanguage !== false ) {
4357                         return $this->mParentLanguage;
4358                 }
4359
4360                 $code = explode( '-', $this->getCode() )[0];
4361                 if ( !in_array( $code, LanguageConverter::$languagesWithVariants ) ) {
4362                         $this->mParentLanguage = null;
4363                         return null;
4364                 }
4365                 $lang = self::factory( $code );
4366                 if ( !$lang->hasVariant( $this->getCode() ) ) {
4367                         $this->mParentLanguage = null;
4368                         return null;
4369                 }
4370
4371                 $this->mParentLanguage = $lang;
4372                 return $lang;
4373         }
4374
4375         /**
4376          * Compare with an other language object
4377          *
4378          * @since 1.28
4379          * @param Language $lang
4380          * @return bool
4381          */
4382         public function equals( Language $lang ) {
4383                 return $lang === $this || $lang->getCode() === $this->mCode;
4384         }
4385
4386         /**
4387          * Get the internal language code for this language object
4388          *
4389          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4390          * htmlspecialchars() or similar
4391          *
4392          * @return string
4393          */
4394         public function getCode() {
4395                 return $this->mCode;
4396         }
4397
4398         /**
4399          * Get the code in BCP 47 format which we can use
4400          * inside of html lang="" tags.
4401          *
4402          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4403          * htmlspecialchars() or similar.
4404          *
4405          * @since 1.19
4406          * @return string
4407          */
4408         public function getHtmlCode() {
4409                 if ( is_null( $this->mHtmlCode ) ) {
4410                         $this->mHtmlCode = LanguageCode::bcp47( $this->getCode() );
4411                 }
4412                 return $this->mHtmlCode;
4413         }
4414
4415         /**
4416          * Get the language code from a file name. Inverse of getFileName()
4417          * @param string $filename $prefix . $languageCode . $suffix
4418          * @param string $prefix Prefix before the language code
4419          * @param string $suffix Suffix after the language code
4420          * @return string Language code, or false if $prefix or $suffix isn't found
4421          */
4422         public static function getCodeFromFileName( $filename, $prefix = 'Language', $suffix = '.php' ) {
4423                 $m = null;
4424                 preg_match( '/' . preg_quote( $prefix, '/' ) . '([A-Z][a-z_]+)' .
4425                         preg_quote( $suffix, '/' ) . '/', $filename, $m );
4426                 if ( !count( $m ) ) {
4427                         return false;
4428                 }
4429                 return str_replace( '_', '-', strtolower( $m[1] ) );
4430         }
4431
4432         /**
4433          * @param string $code
4434          * @param bool $fallback Whether we're going through language fallback chain
4435          * @return string Name of the language class
4436          */
4437         public static function classFromCode( $code, $fallback = true ) {
4438                 if ( $fallback && $code == 'en' ) {
4439                         return 'Language';
4440                 } else {
4441                         return 'Language' . str_replace( '-', '_', ucfirst( $code ) );
4442                 }
4443         }
4444
4445         /**
4446          * Get the name of a file for a certain language code
4447          * @param string $prefix Prepend this to the filename
4448          * @param string $code Language code
4449          * @param string $suffix Append this to the filename
4450          * @throws MWException
4451          * @return string $prefix . $mangledCode . $suffix
4452          */
4453         public static function getFileName( $prefix, $code, $suffix = '.php' ) {
4454                 if ( !self::isValidBuiltInCode( $code ) ) {
4455                         throw new MWException( "Invalid language code \"$code\"" );
4456                 }
4457
4458                 return $prefix . str_replace( '-', '_', ucfirst( $code ) ) . $suffix;
4459         }
4460
4461         /**
4462          * @param string $code
4463          * @return string
4464          */
4465         public static function getMessagesFileName( $code ) {
4466                 global $IP;
4467                 $file = self::getFileName( "$IP/languages/messages/Messages", $code, '.php' );
4468                 Hooks::run( 'Language::getMessagesFileName', [ $code, &$file ] );
4469                 return $file;
4470         }
4471
4472         /**
4473          * @param string $code
4474          * @return string
4475          * @throws MWException
4476          * @since 1.23
4477          */
4478         public static function getJsonMessagesFileName( $code ) {
4479                 global $IP;
4480
4481                 if ( !self::isValidBuiltInCode( $code ) ) {
4482                         throw new MWException( "Invalid language code \"$code\"" );
4483                 }
4484
4485                 return "$IP/languages/i18n/$code.json";
4486         }
4487
4488         /**
4489          * Get the first fallback for a given language.
4490          *
4491          * @param string $code
4492          *
4493          * @return bool|string
4494          */
4495         public static function getFallbackFor( $code ) {
4496                 $fallbacks = self::getFallbacksFor( $code );
4497                 if ( $fallbacks ) {
4498                         return $fallbacks[0];
4499                 }
4500                 return false;
4501         }
4502
4503         /**
4504          * Get the ordered list of fallback languages.
4505          *
4506          * @since 1.19
4507          * @param string $code Language code
4508          * @param int $mode Fallback mode, either MESSAGES_FALLBACKS (which always falls back to 'en'),
4509          * or STRICT_FALLBACKS (whic honly falls back to 'en' when explicitly defined)
4510          * @throws MWException
4511          * @return array List of language codes
4512          */
4513         public static function getFallbacksFor( $code, $mode = self::MESSAGES_FALLBACKS ) {
4514                 if ( $code === 'en' || !self::isValidBuiltInCode( $code ) ) {
4515                         return [];
4516                 }
4517                 switch ( $mode ) {
4518                         case self::MESSAGES_FALLBACKS:
4519                                 // For unknown languages, fallbackSequence returns an empty array,
4520                                 // hardcode fallback to 'en' in that case as English messages are
4521                                 // always defined.
4522                                 return self::getLocalisationCache()->getItem( $code, 'fallbackSequence' ) ?: [ 'en' ];
4523                         case self::STRICT_FALLBACKS:
4524                                 // Use this mode when you don't want to fallback to English unless
4525                                 // explicitly defined, for example when you have language-variant icons
4526                                 // and an international language-independent fallback.
4527                                 return self::getLocalisationCache()->getItem( $code, 'originalFallbackSequence' );
4528                         default:
4529                                 throw new MWException( "Invalid fallback mode \"$mode\"" );
4530                 }
4531         }
4532
4533         /**
4534          * Get the ordered list of fallback languages, ending with the fallback
4535          * language chain for the site language.
4536          *
4537          * @since 1.22
4538          * @param string $code Language code
4539          * @return array Array( fallbacks, site fallbacks )
4540          */
4541         public static function getFallbacksIncludingSiteLanguage( $code ) {
4542                 global $wgLanguageCode;
4543
4544                 // Usually, we will only store a tiny number of fallback chains, so we
4545                 // keep them in static memory.
4546                 $cacheKey = "{$code}-{$wgLanguageCode}";
4547
4548                 if ( !array_key_exists( $cacheKey, self::$fallbackLanguageCache ) ) {
4549                         $fallbacks = self::getFallbacksFor( $code );
4550
4551                         // Append the site's fallback chain, including the site language itself
4552                         $siteFallbacks = self::getFallbacksFor( $wgLanguageCode );
4553                         array_unshift( $siteFallbacks, $wgLanguageCode );
4554
4555                         // Eliminate any languages already included in the chain
4556                         $siteFallbacks = array_diff( $siteFallbacks, $fallbacks );
4557
4558                         self::$fallbackLanguageCache[$cacheKey] = [ $fallbacks, $siteFallbacks ];
4559                 }
4560                 return self::$fallbackLanguageCache[$cacheKey];
4561         }
4562
4563         /**
4564          * Get all messages for a given language
4565          * WARNING: this may take a long time. If you just need all message *keys*
4566          * but need the *contents* of only a few messages, consider using getMessageKeysFor().
4567          *
4568          * @param string $code
4569          *
4570          * @return array
4571          */
4572         public static function getMessagesFor( $code ) {
4573                 return self::getLocalisationCache()->getItem( $code, 'messages' );
4574         }
4575
4576         /**
4577          * Get a message for a given language
4578          *
4579          * @param string $key
4580          * @param string $code
4581          *
4582          * @return string
4583          */
4584         public static function getMessageFor( $key, $code ) {
4585                 return self::getLocalisationCache()->getSubitem( $code, 'messages', $key );
4586         }
4587
4588         /**
4589          * Get all message keys for a given language. This is a faster alternative to
4590          * array_keys( Language::getMessagesFor( $code ) )
4591          *
4592          * @since 1.19
4593          * @param string $code Language code
4594          * @return array Array of message keys (strings)
4595          */
4596         public static function getMessageKeysFor( $code ) {
4597                 return self::getLocalisationCache()->getSubitemList( $code, 'messages' );
4598         }
4599
4600         /**
4601          * @param string $talk
4602          * @return string
4603          */
4604         function fixVariableInNamespace( $talk ) {
4605                 if ( strpos( $talk, '$1' ) === false ) {
4606                         return $talk;
4607                 }
4608
4609                 global $wgMetaNamespace;
4610                 $talk = str_replace( '$1', $wgMetaNamespace, $talk );
4611
4612                 # Allow grammar transformations
4613                 # Allowing full message-style parsing would make simple requests
4614                 # such as action=raw much more expensive than they need to be.
4615                 # This will hopefully cover most cases.
4616                 $talk = preg_replace_callback( '/{{grammar:(.*?)\|(.*?)}}/i',
4617                         [ $this, 'replaceGrammarInNamespace' ], $talk );
4618                 return str_replace( ' ', '_', $talk );
4619         }
4620
4621         /**
4622          * @param string $m
4623          * @return string
4624          */
4625         function replaceGrammarInNamespace( $m ) {
4626                 return $this->convertGrammar( trim( $m[2] ), trim( $m[1] ) );
4627         }
4628
4629         /**
4630          * Decode an expiry (block, protection, etc) which has come from the DB
4631          *
4632          * @param string $expiry Database expiry String
4633          * @param bool|int $format True to process using language functions, or TS_ constant
4634          *     to return the expiry in a given timestamp
4635          * @param string $infinity If $format is not true, use this string for infinite expiry
4636          * @return string
4637          * @since 1.18
4638          */
4639         public function formatExpiry( $expiry, $format = true, $infinity = 'infinity' ) {
4640                 static $dbInfinity;
4641                 if ( $dbInfinity === null ) {
4642                         $dbInfinity = wfGetDB( DB_REPLICA )->getInfinity();
4643                 }
4644
4645                 if ( $expiry == '' || $expiry === 'infinity' || $expiry == $dbInfinity ) {
4646                         return $format === true
4647                                 ? $this->getMessageFromDB( 'infiniteblock' )
4648                                 : $infinity;
4649                 } else {
4650                         return $format === true
4651                                 ? $this->timeanddate( $expiry, /* User preference timezone */ true )
4652                                 : wfTimestamp( $format, $expiry );
4653                 }
4654         }
4655
4656         /**
4657          * Formats a time given in seconds into a string representation of that time.
4658          *
4659          * @param int|float $seconds
4660          * @param array $format An optional argument that formats the returned string in different ways:
4661          *   If $format['avoid'] === 'avoidseconds': don't show seconds if $seconds >= 1 hour,
4662          *   If $format['avoid'] === 'avoidminutes': don't show seconds/minutes if $seconds > 48 hours,
4663          *   If $format['noabbrevs'] is true: use 'seconds' and friends instead of 'seconds-abbrev'
4664          *     and friends.
4665          * @note For backwards compatibility, $format may also be one of the strings 'avoidseconds'
4666          *     or 'avoidminutes'.
4667          * @return string
4668          */
4669         function formatTimePeriod( $seconds, $format = [] ) {
4670                 if ( !is_array( $format ) ) {
4671                         $format = [ 'avoid' => $format ]; // For backwards compatibility
4672                 }
4673                 if ( !isset( $format['avoid'] ) ) {
4674                         $format['avoid'] = false;
4675                 }
4676                 if ( !isset( $format['noabbrevs'] ) ) {
4677                         $format['noabbrevs'] = false;
4678                 }
4679                 $secondsMsg = wfMessage(
4680                         $format['noabbrevs'] ? 'seconds' : 'seconds-abbrev' )->inLanguage( $this );
4681                 $minutesMsg = wfMessage(
4682                         $format['noabbrevs'] ? 'minutes' : 'minutes-abbrev' )->inLanguage( $this );
4683                 $hoursMsg = wfMessage(
4684                         $format['noabbrevs'] ? 'hours' : 'hours-abbrev' )->inLanguage( $this );
4685                 $daysMsg = wfMessage(
4686                         $format['noabbrevs'] ? 'days' : 'days-abbrev' )->inLanguage( $this );
4687
4688                 if ( round( $seconds * 10 ) < 100 ) {
4689                         $s = $this->formatNum( sprintf( "%.1f", round( $seconds * 10 ) / 10 ) );
4690                         $s = $secondsMsg->params( $s )->text();
4691                 } elseif ( round( $seconds ) < 60 ) {
4692                         $s = $this->formatNum( round( $seconds ) );
4693                         $s = $secondsMsg->params( $s )->text();
4694                 } elseif ( round( $seconds ) < 3600 ) {
4695                         $minutes = floor( $seconds / 60 );
4696                         $secondsPart = round( fmod( $seconds, 60 ) );
4697                         if ( $secondsPart == 60 ) {
4698                                 $secondsPart = 0;
4699                                 $minutes++;
4700                         }
4701                         $s = $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4702                         $s .= ' ';
4703                         $s .= $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4704                 } elseif ( round( $seconds ) <= 2 * 86400 ) {
4705                         $hours = floor( $seconds / 3600 );
4706                         $minutes = floor( ( $seconds - $hours * 3600 ) / 60 );
4707                         $secondsPart = round( $seconds - $hours * 3600 - $minutes * 60 );
4708                         if ( $secondsPart == 60 ) {
4709                                 $secondsPart = 0;
4710                                 $minutes++;
4711                         }
4712                         if ( $minutes == 60 ) {
4713                                 $minutes = 0;
4714                                 $hours++;
4715                         }
4716                         $s = $hoursMsg->params( $this->formatNum( $hours ) )->text();
4717                         $s .= ' ';
4718                         $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4719                         if ( !in_array( $format['avoid'], [ 'avoidseconds', 'avoidminutes' ] ) ) {
4720                                 $s .= ' ' . $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4721                         }
4722                 } else {
4723                         $days = floor( $seconds / 86400 );
4724                         if ( $format['avoid'] === 'avoidminutes' ) {
4725                                 $hours = round( ( $seconds - $days * 86400 ) / 3600 );
4726                                 if ( $hours == 24 ) {
4727                                         $hours = 0;
4728                                         $days++;
4729                                 }
4730                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4731                                 $s .= ' ';
4732                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4733                         } elseif ( $format['avoid'] === 'avoidseconds' ) {
4734                                 $hours = floor( ( $seconds - $days * 86400 ) / 3600 );
4735                                 $minutes = round( ( $seconds - $days * 86400 - $hours * 3600 ) / 60 );
4736                                 if ( $minutes == 60 ) {
4737                                         $minutes = 0;
4738                                         $hours++;
4739                                 }
4740                                 if ( $hours == 24 ) {
4741                                         $hours = 0;
4742                                         $days++;
4743                                 }
4744                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4745                                 $s .= ' ';
4746                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4747                                 $s .= ' ';
4748                                 $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4749                         } else {
4750                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4751                                 $s .= ' ';
4752                                 $s .= $this->formatTimePeriod( $seconds - $days * 86400, $format );
4753                         }
4754                 }
4755                 return $s;
4756         }
4757
4758         /**
4759          * Format a bitrate for output, using an appropriate
4760          * unit (bps, kbps, Mbps, Gbps, Tbps, Pbps, Ebps, Zbps or Ybps) according to
4761          *   the magnitude in question.
4762          *
4763          * This use base 1000. For base 1024 use formatSize(), for another base
4764          * see formatComputingNumbers().
4765          *
4766          * @param int $bps
4767          * @return string
4768          */
4769         function formatBitrate( $bps ) {
4770                 return $this->formatComputingNumbers( $bps, 1000, "bitrate-$1bits" );
4771         }
4772
4773         /**
4774          * @param int $size Size of the unit
4775          * @param int $boundary Size boundary (1000, or 1024 in most cases)
4776          * @param string $messageKey Message key to be uesd
4777          * @return string
4778          */
4779         function formatComputingNumbers( $size, $boundary, $messageKey ) {
4780                 if ( $size <= 0 ) {
4781                         return str_replace( '$1', $this->formatNum( $size ),
4782                                 $this->getMessageFromDB( str_replace( '$1', '', $messageKey ) )
4783                         );
4784                 }
4785                 $sizes = [ '', 'kilo', 'mega', 'giga', 'tera', 'peta', 'exa', 'zeta', 'yotta' ];
4786                 $index = 0;
4787
4788                 $maxIndex = count( $sizes ) - 1;
4789                 while ( $size >= $boundary && $index < $maxIndex ) {
4790                         $index++;
4791                         $size /= $boundary;
4792                 }
4793
4794                 // For small sizes no decimal places necessary
4795                 $round = 0;
4796                 if ( $index > 1 ) {
4797                         // For MB and bigger two decimal places are smarter
4798                         $round = 2;
4799                 }
4800                 $msg = str_replace( '$1', $sizes[$index], $messageKey );
4801
4802                 $size = round( $size, $round );
4803                 $text = $this->getMessageFromDB( $msg );
4804                 return str_replace( '$1', $this->formatNum( $size ), $text );
4805         }
4806
4807         /**
4808          * Format a size in bytes for output, using an appropriate
4809          * unit (B, KB, MB, GB, TB, PB, EB, ZB or YB) according to the magnitude in question
4810          *
4811          * This method use base 1024. For base 1000 use formatBitrate(), for
4812          * another base see formatComputingNumbers()
4813          *
4814          * @param int $size Size to format
4815          * @return string Plain text (not HTML)
4816          */
4817         function formatSize( $size ) {
4818                 return $this->formatComputingNumbers( $size, 1024, "size-$1bytes" );
4819         }
4820
4821         /**
4822          * Make a list item, used by various special pages
4823          *
4824          * @param string $page Page link
4825          * @param string $details HTML safe text between brackets
4826          * @param bool $oppositedm Add the direction mark opposite to your
4827          *   language, to display text properly
4828          * @return string HTML escaped
4829          */
4830         function specialList( $page, $details, $oppositedm = true ) {
4831                 if ( !$details ) {
4832                         return $page;
4833                 }
4834
4835                 $dirmark = ( $oppositedm ? $this->getDirMark( true ) : '' ) . $this->getDirMark();
4836                 return $page .
4837                         $dirmark .
4838                         $this->msg( 'word-separator' )->escaped() .
4839                         $this->msg( 'parentheses' )->rawParams( $details )->escaped();
4840         }
4841
4842         /**
4843          * Generate (prev x| next x) (20|50|100...) type links for paging
4844          *
4845          * @param Title $title Title object to link
4846          * @param int $offset
4847          * @param int $limit
4848          * @param array $query Optional URL query parameter string
4849          * @param bool $atend Optional param for specified if this is the last page
4850          * @return string
4851          * @deprecated since 1.33, use SpecialPage::viewPrevNext()
4852          *  instead.
4853          */
4854         public function viewPrevNext( Title $title, $offset, $limit,
4855                 array $query = [], $atend = false
4856         ) {
4857                 // @todo FIXME: Why on earth this needs one message for the text and another one for tooltip?
4858
4859                 # Make 'previous' link
4860                 $prev = wfMessage( 'prevn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4861                 if ( $offset > 0 ) {
4862                         $plink = $this->numLink( $title, max( $offset - $limit, 0 ), $limit,
4863                                 $query, $prev, 'prevn-title', 'mw-prevlink' );
4864                 } else {
4865                         $plink = htmlspecialchars( $prev );
4866                 }
4867
4868                 # Make 'next' link
4869                 $next = wfMessage( 'nextn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4870                 if ( $atend ) {
4871                         $nlink = htmlspecialchars( $next );
4872                 } else {
4873                         $nlink = $this->numLink( $title, $offset + $limit, $limit,
4874                                 $query, $next, 'nextn-title', 'mw-nextlink' );
4875                 }
4876
4877                 # Make links to set number of items per page
4878                 $numLinks = [];
4879                 foreach ( [ 20, 50, 100, 250, 500 ] as $num ) {
4880                         $numLinks[] = $this->numLink( $title, $offset, $num,
4881                                 $query, $this->formatNum( $num ), 'shown-title', 'mw-numlink' );
4882                 }
4883
4884                 return wfMessage( 'viewprevnext' )->inLanguage( $this )->title( $title
4885                         )->rawParams( $plink, $nlink, $this->pipeList( $numLinks ) )->escaped();
4886         }
4887
4888         /**
4889          * Helper function for viewPrevNext() that generates links
4890          *
4891          * @param Title $title Title object to link
4892          * @param int $offset
4893          * @param int $limit
4894          * @param array $query Extra query parameters
4895          * @param string $link Text to use for the link; will be escaped
4896          * @param string $tooltipMsg Name of the message to use as tooltip
4897          * @param string $class Value of the "class" attribute of the link
4898          * @return string HTML fragment
4899          */
4900         private function numLink( Title $title, $offset, $limit, array $query, $link,
4901                 $tooltipMsg, $class
4902         ) {
4903                 $query = [ 'limit' => $limit, 'offset' => $offset ] + $query;
4904                 $tooltip = wfMessage( $tooltipMsg )->inLanguage( $this )->title( $title )
4905                         ->numParams( $limit )->text();
4906
4907                 return Html::element( 'a', [ 'href' => $title->getLocalURL( $query ),
4908                         'title' => $tooltip, 'class' => $class ], $link );
4909         }
4910
4911         /**
4912          * Get the conversion rule title, if any.
4913          *
4914          * @return string
4915          */
4916         public function getConvRuleTitle() {
4917                 return $this->mConverter->getConvRuleTitle();
4918         }
4919
4920         /**
4921          * Get the compiled plural rules for the language
4922          * @since 1.20
4923          * @return array Associative array with plural form, and plural rule as key-value pairs
4924          */
4925         public function getCompiledPluralRules() {
4926                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'compiledPluralRules' );
4927                 $fallbacks = self::getFallbacksFor( $this->mCode );
4928                 if ( !$pluralRules ) {
4929                         foreach ( $fallbacks as $fallbackCode ) {
4930                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'compiledPluralRules' );
4931                                 if ( $pluralRules ) {
4932                                         break;
4933                                 }
4934                         }
4935                 }
4936                 return $pluralRules;
4937         }
4938
4939         /**
4940          * Get the plural rules for the language
4941          * @since 1.20
4942          * @return array Associative array with plural form number and plural rule as key-value pairs
4943          */
4944         public function getPluralRules() {
4945                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRules' );
4946                 $fallbacks = self::getFallbacksFor( $this->mCode );
4947                 if ( !$pluralRules ) {
4948                         foreach ( $fallbacks as $fallbackCode ) {
4949                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRules' );
4950                                 if ( $pluralRules ) {
4951                                         break;
4952                                 }
4953                         }
4954                 }
4955                 return $pluralRules;
4956         }
4957
4958         /**
4959          * Get the plural rule types for the language
4960          * @since 1.22
4961          * @return array Associative array with plural form number and plural rule type as key-value pairs
4962          */
4963         public function getPluralRuleTypes() {
4964                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRuleTypes' );
4965                 $fallbacks = self::getFallbacksFor( $this->mCode );
4966                 if ( !$pluralRuleTypes ) {
4967                         foreach ( $fallbacks as $fallbackCode ) {
4968                                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRuleTypes' );
4969                                 if ( $pluralRuleTypes ) {
4970                                         break;
4971                                 }
4972                         }
4973                 }
4974                 return $pluralRuleTypes;
4975         }
4976
4977         /**
4978          * Find the index number of the plural rule appropriate for the given number
4979          * @param int $number
4980          * @return int The index number of the plural rule
4981          */
4982         public function getPluralRuleIndexNumber( $number ) {
4983                 $pluralRules = $this->getCompiledPluralRules();
4984                 $form = Evaluator::evaluateCompiled( $number, $pluralRules );
4985                 return $form;
4986         }
4987
4988         /**
4989          * Find the plural rule type appropriate for the given number
4990          * For example, if the language is set to Arabic, getPluralType(5) should
4991          * return 'few'.
4992          * @since 1.22
4993          * @param int $number
4994          * @return string The name of the plural rule type, e.g. one, two, few, many
4995          */
4996         public function getPluralRuleType( $number ) {
4997                 $index = $this->getPluralRuleIndexNumber( $number );
4998                 $pluralRuleTypes = $this->getPluralRuleTypes();
4999                 return $pluralRuleTypes[$index] ?? 'other';
5000         }
5001 }