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          * @suppress PhanTypeSuspiciousNonTraversableForeach
 471          */
 472         function __destruct() {
 473                 foreach ( $this as $name => $value ) {
 474                         unset( $this->$name );
 475                 }
 476         }
 477
 478         /**
 479          * Hook which will be called if this is the content language.
 480          * Descendants can use this to register hook functions or modify globals
 481          */
 482         function initContLang() {
 483         }
 484
 485         /**
 486          * @return array
 487          * @since 1.19
 488          */
 489         public function getFallbackLanguages() {
 490                 return self::getFallbacksFor( $this->mCode );
 491         }
 492
 493         /**
 494          * Exports $wgBookstoreListEn
 495          * @return array
 496          */
 497         public function getBookstoreList() {
 498                 return self::$dataCache->getItem( $this->mCode, 'bookstoreList' );
 499         }
 500
 501         /**
 502          * Returns an array of localised namespaces indexed by their numbers. If the namespace is not
 503          * available in localised form, it will be included in English.
 504          *
 505          * @return string[] List of localized namespace names, indexed by numeric namespace ID.
 506          */
 507         public function getNamespaces() {
 508                 if ( is_null( $this->namespaceNames ) ) {
 509                         global $wgMetaNamespace, $wgMetaNamespaceTalk, $wgExtraNamespaces;
 510
 511                         $validNamespaces = MediaWikiServices::getInstance()->getNamespaceInfo()->
 512                                 getCanonicalNamespaces();
 513
 514                         $this->namespaceNames = $wgExtraNamespaces +
 515                                 self::$dataCache->getItem( $this->mCode, 'namespaceNames' );
 516                         $this->namespaceNames += $validNamespaces;
 517
 518                         $this->namespaceNames[NS_PROJECT] = $wgMetaNamespace;
 519                         if ( $wgMetaNamespaceTalk ) {
 520                                 $this->namespaceNames[NS_PROJECT_TALK] = $wgMetaNamespaceTalk;
 521                         } else {
 522                                 $talk = $this->namespaceNames[NS_PROJECT_TALK];
 523                                 $this->namespaceNames[NS_PROJECT_TALK] =
 524                                         $this->fixVariableInNamespace( $talk );
 525                         }
 526
 527                         # Sometimes a language will be localised but not actually exist on this wiki.
 528                         foreach ( $this->namespaceNames as $key => $text ) {
 529                                 if ( !isset( $validNamespaces[$key] ) ) {
 530                                         unset( $this->namespaceNames[$key] );
 531                                 }
 532                         }
 533
 534                         # The above mixing may leave namespaces out of canonical order.
 535                         # Re-order by namespace ID number...
 536                         ksort( $this->namespaceNames );
 537
 538                         Hooks::run( 'LanguageGetNamespaces', [ &$this->namespaceNames ] );
 539                 }
 540
 541                 return $this->namespaceNames;
 542         }
 543
 544         /**
 545          * Arbitrarily set all of the namespace names at once. Mainly used for testing
 546          * @param array $namespaces Array of namespaces (id => name)
 547          */
 548         public function setNamespaces( array $namespaces ) {
 549                 $this->namespaceNames = $namespaces;
 550                 $this->mNamespaceIds = null;
 551         }
 552
 553         /**
 554          * Resets all of the namespace caches. Mainly used for testing
 555          */
 556         public function resetNamespaces() {
 557                 $this->namespaceNames = null;
 558                 $this->mNamespaceIds = null;
 559                 $this->namespaceAliases = null;
 560         }
 561
 562         /**
 563          * A convenience function that returns getNamespaces() with spaces instead of underscores
 564          * in values. Useful for producing output to be displayed e.g. in `<select>` forms.
 565          *
 566          * @return array
 567          */
 568         public function getFormattedNamespaces() {
 569                 $ns = $this->getNamespaces();
 570                 foreach ( $ns as $k => $v ) {
 571                         $ns[$k] = strtr( $v, '_', ' ' );
 572                 }
 573                 return $ns;
 574         }
 575
 576         /**
 577          * Get a namespace value by key
 578          *
 579          * <code>
 580          * $mw_ns = $lang->getNsText( NS_MEDIAWIKI );
 581          * echo $mw_ns; // prints 'MediaWiki'
 582          * </code>
 583          *
 584          * @param int $index The array key of the namespace to return
 585          * @return string|bool String if the namespace value exists, otherwise false
 586          */
 587         public function getNsText( $index ) {
 588                 $ns = $this->getNamespaces();
 589                 return $ns[$index] ?? false;
 590         }
 591
 592         /**
 593          * A convenience function that returns the same thing as
 594          * getNsText() except with '_' changed to ' ', useful for
 595          * producing output.
 596          *
 597          * <code>
 598          * $mw_ns = $lang->getFormattedNsText( NS_MEDIAWIKI_TALK );
 599          * echo $mw_ns; // prints 'MediaWiki talk'
 600          * </code>
 601          *
 602          * @param int $index The array key of the namespace to return
 603          * @return string Namespace name without underscores (empty string if namespace does not exist)
 604          */
 605         public function getFormattedNsText( $index ) {
 606                 $ns = $this->getNsText( $index );
 607                 return strtr( $ns, '_', ' ' );
 608         }
 609
 610         /**
 611          * Returns gender-dependent namespace alias if available.
 612          * See https://www.mediawiki.org/wiki/Manual:$wgExtraGenderNamespaces
 613          * @param int $index Namespace index
 614          * @param string $gender Gender key (male, female... )
 615          * @return string
 616          * @since 1.18
 617          */
 618         public function getGenderNsText( $index, $gender ) {
 619                 global $wgExtraGenderNamespaces;
 620
 621                 $ns = $wgExtraGenderNamespaces +
 622                         (array)self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 623
 624                 return $ns[$index][$gender] ?? $this->getNsText( $index );
 625         }
 626
 627         /**
 628          * Whether this language uses gender-dependent namespace aliases.
 629          * See https://www.mediawiki.org/wiki/Manual:$wgExtraGenderNamespaces
 630          * @return bool
 631          * @since 1.18
 632          */
 633         public function needsGenderDistinction() {
 634                 global $wgExtraGenderNamespaces, $wgExtraNamespaces;
 635                 if ( count( $wgExtraGenderNamespaces ) > 0 ) {
 636                         // $wgExtraGenderNamespaces overrides everything
 637                         return true;
 638                 } elseif ( isset( $wgExtraNamespaces[NS_USER] ) && isset( $wgExtraNamespaces[NS_USER_TALK] ) ) {
 639                         /// @todo There may be other gender namespace than NS_USER & NS_USER_TALK in the future
 640                         // $wgExtraNamespaces overrides any gender aliases specified in i18n files
 641                         return false;
 642                 } else {
 643                         // Check what is in i18n files
 644                         $aliases = self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 645                         return count( $aliases ) > 0;
 646                 }
 647         }
 648
 649         /**
 650          * Get a namespace key by value, case insensitive.
 651          * Only matches namespace names for the current language, not the
 652          * canonical ones defined in Namespace.php.
 653          *
 654          * @param string $text
 655          * @return int|bool An integer if $text is a valid value otherwise false
 656          */
 657         function getLocalNsIndex( $text ) {
 658                 $lctext = $this->lc( $text );
 659                 $ids = $this->getNamespaceIds();
 660                 return $ids[$lctext] ?? false;
 661         }
 662
 663         /**
 664          * @return array
 665          */
 666         public function getNamespaceAliases() {
 667                 if ( is_null( $this->namespaceAliases ) ) {
 668                         $aliases = self::$dataCache->getItem( $this->mCode, 'namespaceAliases' );
 669                         if ( !$aliases ) {
 670                                 $aliases = [];
 671                         } else {
 672                                 foreach ( $aliases as $name => $index ) {
 673                                         if ( $index === NS_PROJECT_TALK ) {
 674                                                 unset( $aliases[$name] );
 675                                                 $name = $this->fixVariableInNamespace( $name );
 676                                                 $aliases[$name] = $index;
 677                                         }
 678                                 }
 679                         }
 680
 681                         global $wgExtraGenderNamespaces;
 682                         $genders = $wgExtraGenderNamespaces +
 683                                 (array)self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 684                         foreach ( $genders as $index => $forms ) {
 685                                 foreach ( $forms as $alias ) {
 686                                         $aliases[$alias] = $index;
 687                                 }
 688                         }
 689
 690                         # Also add converted namespace names as aliases, to avoid confusion.
 691                         $convertedNames = [];
 692                         foreach ( $this->getVariants() as $variant ) {
 693                                 if ( $variant === $this->mCode ) {
 694                                         continue;
 695                                 }
 696                                 foreach ( $this->getNamespaces() as $ns => $_ ) {
 697                                         $convertedNames[$this->getConverter()->convertNamespace( $ns, $variant )] = $ns;
 698                                 }
 699                         }
 700
 701                         $this->namespaceAliases = $aliases + $convertedNames;
 702
 703                         # Filter out aliases to namespaces that don't exist, e.g. from extensions
 704                         # that aren't loaded here but are included in the l10n cache.
 705                         # (array_intersect preserves keys from its first argument)
 706                         $this->namespaceAliases = array_intersect(
 707                                 $this->namespaceAliases,
 708                                 array_keys( $this->getNamespaces() )
 709                         );
 710                 }
 711
 712                 return $this->namespaceAliases;
 713         }
 714
 715         /**
 716          * @return array
 717          */
 718         public function getNamespaceIds() {
 719                 if ( is_null( $this->mNamespaceIds ) ) {
 720                         global $wgNamespaceAliases;
 721                         # Put namespace names and aliases into a hashtable.
 722                         # If this is too slow, then we should arrange it so that it is done
 723                         # before caching. The catch is that at pre-cache time, the above
 724                         # class-specific fixup hasn't been done.
 725                         $this->mNamespaceIds = [];
 726                         foreach ( $this->getNamespaces() as $index => $name ) {
 727                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 728                         }
 729                         foreach ( $this->getNamespaceAliases() as $name => $index ) {
 730                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 731                         }
 732                         if ( $wgNamespaceAliases ) {
 733                                 foreach ( $wgNamespaceAliases as $name => $index ) {
 734                                         $this->mNamespaceIds[$this->lc( $name )] = $index;
 735                                 }
 736                         }
 737                 }
 738                 return $this->mNamespaceIds;
 739         }
 740
 741         /**
 742          * Get a namespace key by value, case insensitive.  Canonical namespace
 743          * names override custom ones defined for the current language.
 744          *
 745          * @param string $text
 746          * @return int|bool An integer if $text is a valid value otherwise false
 747          */
 748         public function getNsIndex( $text ) {
 749                 $lctext = $this->lc( $text );
 750                 $ns = MediaWikiServices::getInstance()->getNamespaceInfo()->
 751                         getCanonicalIndex( $lctext );
 752                 if ( $ns !== null ) {
 753                         return $ns;
 754                 }
 755                 $ids = $this->getNamespaceIds();
 756                 return $ids[$lctext] ?? false;
 757         }
 758
 759         /**
 760          * short names for language variants used for language conversion links.
 761          *
 762          * @param string $code
 763          * @param bool $usemsg Use the "variantname-xyz" message if it exists
 764          * @return string
 765          */
 766         public function getVariantname( $code, $usemsg = true ) {
 767                 $msg = "variantname-$code";
 768                 if ( $usemsg && wfMessage( $msg )->exists() ) {
 769                         return $this->getMessageFromDB( $msg );
 770                 }
 771                 $name = self::fetchLanguageName( $code );
 772                 if ( $name ) {
 773                         return $name; # if it's defined as a language name, show that
 774                 } else {
 775                         # otherwise, output the language code
 776                         return $code;
 777                 }
 778         }
 779
 780         /**
 781          * @return string[]|bool List of date format preference keys, or false if disabled.
 782          */
 783         public function getDatePreferences() {
 784                 return self::$dataCache->getItem( $this->mCode, 'datePreferences' );
 785         }
 786
 787         /**
 788          * @return array
 789          */
 790         function getDateFormats() {
 791                 return self::$dataCache->getItem( $this->mCode, 'dateFormats' );
 792         }
 793
 794         /**
 795          * @return array|string
 796          */
 797         public function getDefaultDateFormat() {
 798                 $df = self::$dataCache->getItem( $this->mCode, 'defaultDateFormat' );
 799                 if ( $df === 'dmy or mdy' ) {
 800                         global $wgAmericanDates;
 801                         return $wgAmericanDates ? 'mdy' : 'dmy';
 802                 } else {
 803                         return $df;
 804                 }
 805         }
 806
 807         /**
 808          * @return array
 809          */
 810         public function getDatePreferenceMigrationMap() {
 811                 return self::$dataCache->getItem( $this->mCode, 'datePreferenceMigrationMap' );
 812         }
 813
 814         /**
 815          * @return array
 816          */
 817         public function getExtraUserToggles() {
 818                 wfDeprecated( __METHOD__, '1.34' );
 819                 return [];
 820         }
 821
 822         /**
 823          * @param string $tog
 824          * @return string
 825          */
 826         function getUserToggle( $tog ) {
 827                 return $this->getMessageFromDB( "tog-$tog" );
 828         }
 829
 830         /**
 831          * Get an array of language names, indexed by code.
 832          * @param null|string $inLanguage Code of language in which to return the names
 833          *              Use self::AS_AUTONYMS for autonyms (native names)
 834          * @param string $include One of:
 835          *              self::ALL all available languages
 836          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 837          *              self::SUPPORTED only if the language is in 'mw' *and* has a message file
 838          * @return array Language code => language name (sorted by key)
 839          * @since 1.20
 840          */
 841         public static function fetchLanguageNames( $inLanguage = self::AS_AUTONYMS, $include = 'mw' ) {
 842                 $cacheKey = $inLanguage === self::AS_AUTONYMS ? 'null' : $inLanguage;
 843                 $cacheKey .= ":$include";
 844                 if ( self::$languageNameCache === null ) {
 845                         self::$languageNameCache = new HashBagOStuff( [ 'maxKeys' => 20 ] );
 846                 }
 847
 848                 $ret = self::$languageNameCache->get( $cacheKey );
 849                 if ( !$ret ) {
 850                         $ret = self::fetchLanguageNamesUncached( $inLanguage, $include );
 851                         self::$languageNameCache->set( $cacheKey, $ret );
 852                 }
 853                 return $ret;
 854         }
 855
 856         /**
 857          * Uncached helper for fetchLanguageNames
 858          * @param null|string $inLanguage Code of language in which to return the names
 859          *              Use self::AS_AUTONYMS for autonyms (native names)
 860          * @param string $include One of:
 861          *              self::ALL all available languages
 862          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 863          *              self::SUPPORTED only if the language is in 'mw' *and* has a message file
 864          * @return array Language code => language name (sorted by key)
 865          */
 866         private static function fetchLanguageNamesUncached(
 867                 $inLanguage = self::AS_AUTONYMS,
 868                 $include = 'mw'
 869         ) {
 870                 global $wgExtraLanguageNames, $wgUsePigLatinVariant;
 871
 872                 // If passed an invalid language code to use, fallback to en
 873                 if ( $inLanguage !== self::AS_AUTONYMS && !self::isValidCode( $inLanguage ) ) {
 874                         $inLanguage = 'en';
 875                 }
 876
 877                 $names = [];
 878
 879                 if ( $inLanguage ) {
 880                         # TODO: also include when $inLanguage is null, when this code is more efficient
 881                         Hooks::run( 'LanguageGetTranslatedLanguageNames', [ &$names, $inLanguage ] );
 882                 }
 883
 884                 $mwNames = $wgExtraLanguageNames + MediaWiki\Languages\Data\Names::$names;
 885                 if ( $wgUsePigLatinVariant ) {
 886                         // Pig Latin (for variant development)
 887                         $mwNames['en-x-piglatin'] = 'Igpay Atinlay';
 888                 }
 889
 890                 foreach ( $mwNames as $mwCode => $mwName ) {
 891                         # - Prefer own MediaWiki native name when not using the hook
 892                         # - For other names just add if not added through the hook
 893                         if ( $mwCode === $inLanguage || !isset( $names[$mwCode] ) ) {
 894                                 $names[$mwCode] = $mwName;
 895                         }
 896                 }
 897
 898                 if ( $include === self::ALL ) {
 899                         ksort( $names );
 900                         return $names;
 901                 }
 902
 903                 $returnMw = [];
 904                 $coreCodes = array_keys( $mwNames );
 905                 foreach ( $coreCodes as $coreCode ) {
 906                         $returnMw[$coreCode] = $names[$coreCode];
 907                 }
 908
 909                 if ( $include === self::SUPPORTED ) {
 910                         $namesMwFile = [];
 911                         # We do this using a foreach over the codes instead of a directory
 912                         # loop so that messages files in extensions will work correctly.
 913                         foreach ( $returnMw as $code => $value ) {
 914                                 if ( is_readable( self::getMessagesFileName( $code ) )
 915                                         || is_readable( self::getJsonMessagesFileName( $code ) )
 916                                 ) {
 917                                         $namesMwFile[$code] = $names[$code];
 918                                 }
 919                         }
 920
 921                         ksort( $namesMwFile );
 922                         return $namesMwFile;
 923                 }
 924
 925                 ksort( $returnMw );
 926                 # 'mw' option; default if it's not one of the other two options (all/mwfile)
 927                 return $returnMw;
 928         }
 929
 930         /**
 931          * @param string $code The code of the language for which to get the name
 932          * @param null|string $inLanguage Code of language in which to return the name
 933          *   (SELF::AS_AUTONYMS for autonyms)
 934          * @param string $include See fetchLanguageNames()
 935          * @return string Language name or empty
 936          * @since 1.20
 937          */
 938         public static function fetchLanguageName(
 939                 $code,
 940                 $inLanguage = self::AS_AUTONYMS,
 941                 $include = self::ALL
 942         ) {
 943                 $code = strtolower( $code );
 944                 $array = self::fetchLanguageNames( $inLanguage, $include );
 945                 return !array_key_exists( $code, $array ) ? '' : $array[$code];
 946         }
 947
 948         /**
 949          * Get a message from the MediaWiki namespace.
 950          *
 951          * @param string $msg Message name
 952          * @return string
 953          */
 954         public function getMessageFromDB( $msg ) {
 955                 return $this->msg( $msg )->text();
 956         }
 957
 958         /**
 959          * Get message object in this language. Only for use inside this class.
 960          *
 961          * @param string $msg Message name
 962          * @return Message
 963          */
 964         protected function msg( $msg ) {
 965                 return wfMessage( $msg )->inLanguage( $this );
 966         }
 967
 968         /**
 969          * @param string $key
 970          * @return string
 971          */
 972         public function getMonthName( $key ) {
 973                 return $this->getMessageFromDB( self::$mMonthMsgs[$key - 1] );
 974         }
 975
 976         /**
 977          * @return array
 978          */
 979         public function getMonthNamesArray() {
 980                 $monthNames = [ '' ];
 981                 for ( $i = 1; $i < 13; $i++ ) {
 982                         $monthNames[] = $this->getMonthName( $i );
 983                 }
 984                 return $monthNames;
 985         }
 986
 987         /**
 988          * @param string $key
 989          * @return string
 990          */
 991         public function getMonthNameGen( $key ) {
 992                 return $this->getMessageFromDB( self::$mMonthGenMsgs[$key - 1] );
 993         }
 994
 995         /**
 996          * @param string $key
 997          * @return string
 998          */
 999         public function getMonthAbbreviation( $key ) {
1000                 return $this->getMessageFromDB( self::$mMonthAbbrevMsgs[$key - 1] );
1001         }
1002
1003         /**
1004          * @return array
1005          */
1006         public function getMonthAbbreviationsArray() {
1007                 $monthNames = [ '' ];
1008                 for ( $i = 1; $i < 13; $i++ ) {
1009                         $monthNames[] = $this->getMonthAbbreviation( $i );
1010                 }
1011                 return $monthNames;
1012         }
1013
1014         /**
1015          * @param string $key
1016          * @return string
1017          */
1018         public function getWeekdayName( $key ) {
1019                 return $this->getMessageFromDB( self::$mWeekdayMsgs[$key - 1] );
1020         }
1021
1022         /**
1023          * @param string $key
1024          * @return string
1025          */
1026         function getWeekdayAbbreviation( $key ) {
1027                 return $this->getMessageFromDB( self::$mWeekdayAbbrevMsgs[$key - 1] );
1028         }
1029
1030         /**
1031          * @param string $key
1032          * @return string
1033          */
1034         function getIranianCalendarMonthName( $key ) {
1035                 return $this->getMessageFromDB( self::$mIranianCalendarMonthMsgs[$key - 1] );
1036         }
1037
1038         /**
1039          * @param string $key
1040          * @return string
1041          */
1042         function getHebrewCalendarMonthName( $key ) {
1043                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthMsgs[$key - 1] );
1044         }
1045
1046         /**
1047          * @param string $key
1048          * @return string
1049          */
1050         function getHebrewCalendarMonthNameGen( $key ) {
1051                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthGenMsgs[$key - 1] );
1052         }
1053
1054         /**
1055          * @param string $key
1056          * @return string
1057          */
1058         function getHijriCalendarMonthName( $key ) {
1059                 return $this->getMessageFromDB( self::$mHijriCalendarMonthMsgs[$key - 1] );
1060         }
1061
1062         /**
1063          * Pass through result from $dateTimeObj->format()
1064          * @param DateTime|bool|null &$dateTimeObj
1065          * @param string $ts
1066          * @param DateTimeZone|bool|null $zone
1067          * @param string $code
1068          * @return string
1069          */
1070         private static function dateTimeObjFormat( &$dateTimeObj, $ts, $zone, $code ) {
1071                 if ( !$dateTimeObj ) {
1072                         $dateTimeObj = DateTime::createFromFormat(
1073                                 'YmdHis', $ts, $zone ?: new DateTimeZone( 'UTC' )
1074                         );
1075                 }
1076                 return $dateTimeObj->format( $code );
1077         }
1078
1079         /**
1080          * This is a workalike of PHP's date() function, but with better
1081          * internationalisation, a reduced set of format characters, and a better
1082          * escaping format.
1083          *
1084          * Supported format characters are dDjlNwzWFmMntLoYyaAgGhHiscrUeIOPTZ. See
1085          * the PHP manual for definitions. There are a number of extensions, which
1086          * start with "x":
1087          *
1088          *    xn   Do not translate digits of the next numeric format character
1089          *    xN   Toggle raw digit (xn) flag, stays set until explicitly unset
1090          *    xr   Use roman numerals for the next numeric format character
1091          *    xh   Use hebrew numerals for the next numeric format character
1092          *    xx   Literal x
1093          *    xg   Genitive month name
1094          *
1095          *    xij  j (day number) in Iranian calendar
1096          *    xiF  F (month name) in Iranian calendar
1097          *    xin  n (month number) in Iranian calendar
1098          *    xiy  y (two digit year) in Iranian calendar
1099          *    xiY  Y (full year) in Iranian calendar
1100          *    xit  t (days in month) in Iranian calendar
1101          *    xiz  z (day of the year) in Iranian calendar
1102          *
1103          *    xjj  j (day number) in Hebrew calendar
1104          *    xjF  F (month name) in Hebrew calendar
1105          *    xjt  t (days in month) in Hebrew calendar
1106          *    xjx  xg (genitive month name) in Hebrew calendar
1107          *    xjn  n (month number) in Hebrew calendar
1108          *    xjY  Y (full year) in Hebrew calendar
1109          *
1110          *    xmj  j (day number) in Hijri calendar
1111          *    xmF  F (month name) in Hijri calendar
1112          *    xmn  n (month number) in Hijri calendar
1113          *    xmY  Y (full year) in Hijri calendar
1114          *
1115          *    xkY  Y (full year) in Thai solar calendar. Months and days are
1116          *                       identical to the Gregorian calendar
1117          *    xoY  Y (full year) in Minguo calendar or Juche year.
1118          *                       Months and days are identical to the
1119          *                       Gregorian calendar
1120          *    xtY  Y (full year) in Japanese nengo. Months and days are
1121          *                       identical to the Gregorian calendar
1122          *
1123          * Characters enclosed in double quotes will be considered literal (with
1124          * the quotes themselves removed). Unmatched quotes will be considered
1125          * literal quotes. Example:
1126          *
1127          * "The month is" F       => The month is January
1128          * i's"                   => 20'11"
1129          *
1130          * Backslash escaping is also supported.
1131          *
1132          * Input timestamp is assumed to be pre-normalized to the desired local
1133          * time zone, if any. Note that the format characters crUeIOPTZ will assume
1134          * $ts is UTC if $zone is not given.
1135          *
1136          * @param string $format
1137          * @param string $ts 14-character timestamp
1138          *      YYYYMMDDHHMMSS
1139          *      01234567890123
1140          * @param DateTimeZone|null $zone Timezone of $ts
1141          * @param int &$ttl The amount of time (in seconds) the output may be cached for.
1142          * Only makes sense if $ts is the current time.
1143          * @todo handling of "o" format character for Iranian, Hebrew, Hijri & Thai?
1144          *
1145          * @throws MWException
1146          * @return string
1147          */
1148         public function sprintfDate( $format, $ts, DateTimeZone $zone = null, &$ttl = 'unused' ) {
1149                 $s = '';
1150                 $raw = false;
1151                 $roman = false;
1152                 $hebrewNum = false;
1153                 $dateTimeObj = false;
1154                 $rawToggle = false;
1155                 $iranian = false;
1156                 $hebrew = false;
1157                 $hijri = false;
1158                 $thai = false;
1159                 $minguo = false;
1160                 $tenno = false;
1161
1162                 $usedSecond = false;
1163                 $usedMinute = false;
1164                 $usedHour = false;
1165                 $usedAMPM = false;
1166                 $usedDay = false;
1167                 $usedWeek = false;
1168                 $usedMonth = false;
1169                 $usedYear = false;
1170                 $usedISOYear = false;
1171                 $usedIsLeapYear = false;
1172
1173                 $usedHebrewMonth = false;
1174                 $usedIranianMonth = false;
1175                 $usedHijriMonth = false;
1176                 $usedHebrewYear = false;
1177                 $usedIranianYear = false;
1178                 $usedHijriYear = false;
1179                 $usedTennoYear = false;
1180
1181                 if ( strlen( $ts ) !== 14 ) {
1182                         throw new MWException( __METHOD__ . ": The timestamp $ts should have 14 characters" );
1183                 }
1184
1185                 if ( !ctype_digit( $ts ) ) {
1186                         throw new MWException( __METHOD__ . ": The timestamp $ts should be a number" );
1187                 }
1188
1189                 $formatLength = strlen( $format );
1190                 for ( $p = 0; $p < $formatLength; $p++ ) {
1191                         $num = false;
1192                         $code = $format[$p];
1193                         if ( $code == 'x' && $p < $formatLength - 1 ) {
1194                                 $code .= $format[++$p];
1195                         }
1196
1197                         if ( ( $code === 'xi'
1198                                         || $code === 'xj'
1199                                         || $code === 'xk'
1200                                         || $code === 'xm'
1201                                         || $code === 'xo'
1202                                         || $code === 'xt' )
1203                                 && $p < $formatLength - 1 ) {
1204                                 $code .= $format[++$p];
1205                         }
1206
1207                         switch ( $code ) {
1208                                 case 'xx':
1209                                         $s .= 'x';
1210                                         break;
1211                                 case 'xn':
1212                                         $raw = true;
1213                                         break;
1214                                 case 'xN':
1215                                         $rawToggle = !$rawToggle;
1216                                         break;
1217                                 case 'xr':
1218                                         $roman = true;
1219                                         break;
1220                                 case 'xh':
1221                                         $hebrewNum = true;
1222                                         break;
1223                                 case 'xg':
1224                                         $usedMonth = true;
1225                                         $s .= $this->getMonthNameGen( substr( $ts, 4, 2 ) );
1226                                         break;
1227                                 case 'xjx':
1228                                         $usedHebrewMonth = true;
1229                                         if ( !$hebrew ) {
1230                                                 $hebrew = self::tsToHebrew( $ts );
1231                                         }
1232                                         $s .= $this->getHebrewCalendarMonthNameGen( $hebrew[1] );
1233                                         break;
1234                                 case 'd':
1235                                         $usedDay = true;
1236                                         $num = substr( $ts, 6, 2 );
1237                                         break;
1238                                 case 'D':
1239                                         $usedDay = true;
1240                                         $s .= $this->getWeekdayAbbreviation(
1241                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1242                                         );
1243                                         break;
1244                                 case 'j':
1245                                         $usedDay = true;
1246                                         $num = intval( substr( $ts, 6, 2 ) );
1247                                         break;
1248                                 case 'xij':
1249                                         $usedDay = true;
1250                                         if ( !$iranian ) {
1251                                                 $iranian = self::tsToIranian( $ts );
1252                                         }
1253                                         $num = $iranian[2];
1254                                         break;
1255                                 case 'xmj':
1256                                         $usedDay = true;
1257                                         if ( !$hijri ) {
1258                                                 $hijri = self::tsToHijri( $ts );
1259                                         }
1260                                         $num = $hijri[2];
1261                                         break;
1262                                 case 'xjj':
1263                                         $usedDay = true;
1264                                         if ( !$hebrew ) {
1265                                                 $hebrew = self::tsToHebrew( $ts );
1266                                         }
1267                                         $num = $hebrew[2];
1268                                         break;
1269                                 case 'l':
1270                                         $usedDay = true;
1271                                         $s .= $this->getWeekdayName(
1272                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1273                                         );
1274                                         break;
1275                                 case 'F':
1276                                         $usedMonth = true;
1277                                         $s .= $this->getMonthName( substr( $ts, 4, 2 ) );
1278                                         break;
1279                                 case 'xiF':
1280                                         $usedIranianMonth = true;
1281                                         if ( !$iranian ) {
1282                                                 $iranian = self::tsToIranian( $ts );
1283                                         }
1284                                         $s .= $this->getIranianCalendarMonthName( $iranian[1] );
1285                                         break;
1286                                 case 'xmF':
1287                                         $usedHijriMonth = true;
1288                                         if ( !$hijri ) {
1289                                                 $hijri = self::tsToHijri( $ts );
1290                                         }
1291                                         $s .= $this->getHijriCalendarMonthName( $hijri[1] );
1292                                         break;
1293                                 case 'xjF':
1294                                         $usedHebrewMonth = true;
1295                                         if ( !$hebrew ) {
1296                                                 $hebrew = self::tsToHebrew( $ts );
1297                                         }
1298                                         $s .= $this->getHebrewCalendarMonthName( $hebrew[1] );
1299                                         break;
1300                                 case 'm':
1301                                         $usedMonth = true;
1302                                         $num = substr( $ts, 4, 2 );
1303                                         break;
1304                                 case 'M':
1305                                         $usedMonth = true;
1306                                         $s .= $this->getMonthAbbreviation( substr( $ts, 4, 2 ) );
1307                                         break;
1308                                 case 'n':
1309                                         $usedMonth = true;
1310                                         $num = intval( substr( $ts, 4, 2 ) );
1311                                         break;
1312                                 case 'xin':
1313                                         $usedIranianMonth = true;
1314                                         if ( !$iranian ) {
1315                                                 $iranian = self::tsToIranian( $ts );
1316                                         }
1317                                         $num = $iranian[1];
1318                                         break;
1319                                 case 'xmn':
1320                                         $usedHijriMonth = true;
1321                                         if ( !$hijri ) {
1322                                                 $hijri = self::tsToHijri( $ts );
1323                                         }
1324                                         $num = $hijri[1];
1325                                         break;
1326                                 case 'xjn':
1327                                         $usedHebrewMonth = true;
1328                                         if ( !$hebrew ) {
1329                                                 $hebrew = self::tsToHebrew( $ts );
1330                                         }
1331                                         $num = $hebrew[1];
1332                                         break;
1333                                 case 'xjt':
1334                                         $usedHebrewMonth = true;
1335                                         if ( !$hebrew ) {
1336                                                 $hebrew = self::tsToHebrew( $ts );
1337                                         }
1338                                         $num = $hebrew[3];
1339                                         break;
1340                                 case 'Y':
1341                                         $usedYear = true;
1342                                         $num = substr( $ts, 0, 4 );
1343                                         break;
1344                                 case 'xiY':
1345                                         $usedIranianYear = true;
1346                                         if ( !$iranian ) {
1347                                                 $iranian = self::tsToIranian( $ts );
1348                                         }
1349                                         $num = $iranian[0];
1350                                         break;
1351                                 case 'xmY':
1352                                         $usedHijriYear = true;
1353                                         if ( !$hijri ) {
1354                                                 $hijri = self::tsToHijri( $ts );
1355                                         }
1356                                         $num = $hijri[0];
1357                                         break;
1358                                 case 'xjY':
1359                                         $usedHebrewYear = true;
1360                                         if ( !$hebrew ) {
1361                                                 $hebrew = self::tsToHebrew( $ts );
1362                                         }
1363                                         $num = $hebrew[0];
1364                                         break;
1365                                 case 'xkY':
1366                                         $usedYear = true;
1367                                         if ( !$thai ) {
1368                                                 $thai = self::tsToYear( $ts, 'thai' );
1369                                         }
1370                                         $num = $thai[0];
1371                                         break;
1372                                 case 'xoY':
1373                                         $usedYear = true;
1374                                         if ( !$minguo ) {
1375                                                 $minguo = self::tsToYear( $ts, 'minguo' );
1376                                         }
1377                                         $num = $minguo[0];
1378                                         break;
1379                                 case 'xtY':
1380                                         $usedTennoYear = true;
1381                                         if ( !$tenno ) {
1382                                                 $tenno = self::tsToYear( $ts, 'tenno' );
1383                                         }
1384                                         $num = $tenno[0];
1385                                         break;
1386                                 case 'y':
1387                                         $usedYear = true;
1388                                         $num = substr( $ts, 2, 2 );
1389                                         break;
1390                                 case 'xiy':
1391                                         $usedIranianYear = true;
1392                                         if ( !$iranian ) {
1393                                                 $iranian = self::tsToIranian( $ts );
1394                                         }
1395                                         $num = substr( $iranian[0], -2 );
1396                                         break;
1397                                 case 'xit':
1398                                         $usedIranianYear = true;
1399                                         if ( !$iranian ) {
1400                                                 $iranian = self::tsToIranian( $ts );
1401                                         }
1402                                         $num = self::$IRANIAN_DAYS[$iranian[1] - 1];
1403                                         break;
1404                                 case 'xiz':
1405                                         $usedIranianYear = true;
1406                                         if ( !$iranian ) {
1407                                                 $iranian = self::tsToIranian( $ts );
1408                                         }
1409                                         $num = $iranian[3];
1410                                         break;
1411                                 case 'a':
1412                                         $usedAMPM = true;
1413                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'am' : 'pm';
1414                                         break;
1415                                 case 'A':
1416                                         $usedAMPM = true;
1417                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'AM' : 'PM';
1418                                         break;
1419                                 case 'g':
1420                                         $usedHour = true;
1421                                         $h = substr( $ts, 8, 2 );
1422                                         $num = $h % 12 ?: 12;
1423                                         break;
1424                                 case 'G':
1425                                         $usedHour = true;
1426                                         $num = intval( substr( $ts, 8, 2 ) );
1427                                         break;
1428                                 case 'h':
1429                                         $usedHour = true;
1430                                         $h = substr( $ts, 8, 2 );
1431                                         $num = sprintf( '%02d', $h % 12 ?: 12 );
1432                                         break;
1433                                 case 'H':
1434                                         $usedHour = true;
1435                                         $num = substr( $ts, 8, 2 );
1436                                         break;
1437                                 case 'i':
1438                                         $usedMinute = true;
1439                                         $num = substr( $ts, 10, 2 );
1440                                         break;
1441                                 case 's':
1442                                         $usedSecond = true;
1443                                         $num = substr( $ts, 12, 2 );
1444                                         break;
1445                                 case 'c':
1446                                 case 'r':
1447                                         $usedSecond = true;
1448                                         // fall through
1449                                 case 'e':
1450                                 case 'O':
1451                                 case 'P':
1452                                 case 'T':
1453                                         $s .= self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1454                                         break;
1455                                 case 'w':
1456                                 case 'N':
1457                                 case 'z':
1458                                         $usedDay = true;
1459                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1460                                         break;
1461                                 case 'W':
1462                                         $usedWeek = true;
1463                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1464                                         break;
1465                                 case 't':
1466                                         $usedMonth = true;
1467                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1468                                         break;
1469                                 case 'L':
1470                                         $usedIsLeapYear = true;
1471                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1472                                         break;
1473                                 case 'o':
1474                                         $usedISOYear = true;
1475                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1476                                         break;
1477                                 case 'U':
1478                                         $usedSecond = true;
1479                                         // fall through
1480                                 case 'I':
1481                                 case 'Z':
1482                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1483                                         break;
1484                                 case '\\':
1485                                         # Backslash escaping
1486                                         if ( $p < $formatLength - 1 ) {
1487                                                 $s .= $format[++$p];
1488                                         } else {
1489                                                 $s .= '\\';
1490                                         }
1491                                         break;
1492                                 case '"':
1493                                         # Quoted literal
1494                                         if ( $p < $formatLength - 1 ) {
1495                                                 $endQuote = strpos( $format, '"', $p + 1 );
1496                                                 if ( $endQuote === false ) {
1497                                                         # No terminating quote, assume literal "
1498                                                         $s .= '"';
1499                                                 } else {
1500                                                         $s .= substr( $format, $p + 1, $endQuote - $p - 1 );
1501                                                         $p = $endQuote;
1502                                                 }
1503                                         } else {
1504                                                 # Quote at end of string, assume literal "
1505                                                 $s .= '"';
1506                                         }
1507                                         break;
1508                                 default:
1509                                         $s .= $format[$p];
1510                         }
1511                         if ( $num !== false ) {
1512                                 if ( $rawToggle || $raw ) {
1513                                         $s .= $num;
1514                                         $raw = false;
1515                                 } elseif ( $roman ) {
1516                                         $s .= self::romanNumeral( $num );
1517                                         $roman = false;
1518                                 } elseif ( $hebrewNum ) {
1519                                         $s .= self::hebrewNumeral( $num );
1520                                         $hebrewNum = false;
1521                                 } else {
1522                                         $s .= $this->formatNum( $num, true );
1523                                 }
1524                         }
1525                 }
1526
1527                 if ( $ttl === 'unused' ) {
1528                         // No need to calculate the TTL, the caller wont use it anyway.
1529                 } elseif ( $usedSecond ) {
1530                         $ttl = 1;
1531                 } elseif ( $usedMinute ) {
1532                         $ttl = 60 - substr( $ts, 12, 2 );
1533                 } elseif ( $usedHour ) {
1534                         $ttl = 3600 - substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1535                 } elseif ( $usedAMPM ) {
1536                         $ttl = 43200 - ( substr( $ts, 8, 2 ) % 12 ) * 3600 -
1537                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1538                 } elseif (
1539                         $usedDay ||
1540                         $usedHebrewMonth ||
1541                         $usedIranianMonth ||
1542                         $usedHijriMonth ||
1543                         $usedHebrewYear ||
1544                         $usedIranianYear ||
1545                         $usedHijriYear ||
1546                         $usedTennoYear
1547                 ) {
1548                         // @todo Someone who understands the non-Gregorian calendars
1549                         // should write proper logic for them so that they don't need purged every day.
1550                         $ttl = 86400 - substr( $ts, 8, 2 ) * 3600 -
1551                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1552                 } else {
1553                         $possibleTtls = [];
1554                         $timeRemainingInDay = 86400 - substr( $ts, 8, 2 ) * 3600 -
1555                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1556                         if ( $usedWeek ) {
1557                                 $possibleTtls[] =
1558                                         ( 7 - self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400 +
1559                                         $timeRemainingInDay;
1560                         } elseif ( $usedISOYear ) {
1561                                 // December 28th falls on the last ISO week of the year, every year.
1562                                 // The last ISO week of a year can be 52 or 53.
1563                                 $lastWeekOfISOYear = DateTime::createFromFormat(
1564                                         'Ymd',
1565                                         substr( $ts, 0, 4 ) . '1228',
1566                                         $zone ?: new DateTimeZone( 'UTC' )
1567                                 )->format( 'W' );
1568                                 $currentISOWeek = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'W' );
1569                                 $weeksRemaining = $lastWeekOfISOYear - $currentISOWeek;
1570                                 $timeRemainingInWeek =
1571                                         ( 7 - self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400
1572                                         + $timeRemainingInDay;
1573                                 $possibleTtls[] = $weeksRemaining * 604800 + $timeRemainingInWeek;
1574                         }
1575
1576                         if ( $usedMonth ) {
1577                                 $possibleTtls[] =
1578                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 't' ) -
1579                                                 substr( $ts, 6, 2 ) ) * 86400
1580                                         + $timeRemainingInDay;
1581                         } elseif ( $usedYear ) {
1582                                 $possibleTtls[] =
1583                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1584                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1585                                         + $timeRemainingInDay;
1586                         } elseif ( $usedIsLeapYear ) {
1587                                 $year = substr( $ts, 0, 4 );
1588                                 $timeRemainingInYear =
1589                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1590                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1591                                         + $timeRemainingInDay;
1592                                 $mod = $year % 4;
1593                                 if ( $mod || ( !( $year % 100 ) && $year % 400 ) ) {
1594                                         // this isn't a leap year. see when the next one starts
1595                                         $nextCandidate = $year - $mod + 4;
1596                                         if ( $nextCandidate % 100 || !( $nextCandidate % 400 ) ) {
1597                                                 $possibleTtls[] = ( $nextCandidate - $year - 1 ) * 365 * 86400 +
1598                                                         $timeRemainingInYear;
1599                                         } else {
1600                                                 $possibleTtls[] = ( $nextCandidate - $year + 3 ) * 365 * 86400 +
1601                                                         $timeRemainingInYear;
1602                                         }
1603                                 } else {
1604                                         // this is a leap year, so the next year isn't
1605                                         $possibleTtls[] = $timeRemainingInYear;
1606                                 }
1607                         }
1608
1609                         if ( $possibleTtls ) {
1610                                 $ttl = min( $possibleTtls );
1611                         }
1612                 }
1613
1614                 return $s;
1615         }
1616
1617         private static $GREG_DAYS = [ 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 ];
1618         private static $IRANIAN_DAYS = [ 31, 31, 31, 31, 31, 31, 30, 30, 30, 30, 30, 29 ];
1619
1620         /**
1621          * Algorithm by Roozbeh Pournader and Mohammad Toossi to convert
1622          * Gregorian dates to Iranian dates. Originally written in C, it
1623          * is released under the terms of GNU Lesser General Public
1624          * License. Conversion to PHP was performed by Niklas Laxström.
1625          *
1626          * Link: http://www.farsiweb.info/jalali/jalali.c
1627          *
1628          * @param string $ts
1629          *
1630          * @return int[]
1631          */
1632         private static function tsToIranian( $ts ) {
1633                 $gy = substr( $ts, 0, 4 ) - 1600;
1634                 $gm = substr( $ts, 4, 2 ) - 1;
1635                 $gd = substr( $ts, 6, 2 ) - 1;
1636
1637                 # Days passed from the beginning (including leap years)
1638                 $gDayNo = 365 * $gy
1639                         + floor( ( $gy + 3 ) / 4 )
1640                         - floor( ( $gy + 99 ) / 100 )
1641                         + floor( ( $gy + 399 ) / 400 );
1642
1643                 // Add days of the past months of this year
1644                 for ( $i = 0; $i < $gm; $i++ ) {
1645                         $gDayNo += self::$GREG_DAYS[$i];
1646                 }
1647
1648                 // Leap years
1649                 if ( $gm > 1 && ( ( $gy % 4 === 0 && $gy % 100 !== 0 || ( $gy % 400 == 0 ) ) ) ) {
1650                         $gDayNo++;
1651                 }
1652
1653                 // Days passed in current month
1654                 $gDayNo += (int)$gd;
1655
1656                 $jDayNo = $gDayNo - 79;
1657
1658                 $jNp = floor( $jDayNo / 12053 );
1659                 $jDayNo %= 12053;
1660
1661                 $jy = 979 + 33 * $jNp + 4 * floor( $jDayNo / 1461 );
1662                 $jDayNo %= 1461;
1663
1664                 if ( $jDayNo >= 366 ) {
1665                         $jy += floor( ( $jDayNo - 1 ) / 365 );
1666                         $jDayNo = floor( ( $jDayNo - 1 ) % 365 );
1667                 }
1668
1669                 $jz = $jDayNo;
1670
1671                 for ( $i = 0; $i < 11 && $jDayNo >= self::$IRANIAN_DAYS[$i]; $i++ ) {
1672                         $jDayNo -= self::$IRANIAN_DAYS[$i];
1673                 }
1674
1675                 $jm = $i + 1;
1676                 $jd = $jDayNo + 1;
1677
1678                 return [ $jy, $jm, $jd, $jz ];
1679         }
1680
1681         /**
1682          * Converting Gregorian dates to Hijri dates.
1683          *
1684          * Based on a PHP-Nuke block by Sharjeel which is released under GNU/GPL license
1685          *
1686          * @see https://phpnuke.org/modules.php?name=News&file=article&sid=8234&mode=thread&order=0&thold=0
1687          *
1688          * @param string $ts
1689          *
1690          * @return int[]
1691          */
1692         private static function tsToHijri( $ts ) {
1693                 $year = substr( $ts, 0, 4 );
1694                 $month = substr( $ts, 4, 2 );
1695                 $day = substr( $ts, 6, 2 );
1696
1697                 $zyr = $year;
1698                 $zd = $day;
1699                 $zm = $month;
1700                 $zy = $zyr;
1701
1702                 if (
1703                         ( $zy > 1582 ) || ( ( $zy == 1582 ) && ( $zm > 10 ) ) ||
1704                         ( ( $zy == 1582 ) && ( $zm == 10 ) && ( $zd > 14 ) )
1705                 ) {
1706                         $zjd = (int)( ( 1461 * ( $zy + 4800 + (int)( ( $zm - 14 ) / 12 ) ) ) / 4 ) +
1707                                         (int)( ( 367 * ( $zm - 2 - 12 * ( (int)( ( $zm - 14 ) / 12 ) ) ) ) / 12 ) -
1708                                         (int)( ( 3 * (int)( ( ( $zy + 4900 + (int)( ( $zm - 14 ) / 12 ) ) / 100 ) ) ) / 4 ) +
1709                                         $zd - 32075;
1710                 } else {
1711                         $zjd = 367 * $zy - (int)( ( 7 * ( $zy + 5001 + (int)( ( $zm - 9 ) / 7 ) ) ) / 4 ) +
1712                                                                 (int)( ( 275 * $zm ) / 9 ) + $zd + 1729777;
1713                 }
1714
1715                 $zl = $zjd - 1948440 + 10632;
1716                 $zn = (int)( ( $zl - 1 ) / 10631 );
1717                 $zl = $zl - 10631 * $zn + 354;
1718                 $zj = ( (int)( ( 10985 - $zl ) / 5316 ) ) * ( (int)( ( 50 * $zl ) / 17719 ) ) +
1719                         ( (int)( $zl / 5670 ) ) * ( (int)( ( 43 * $zl ) / 15238 ) );
1720                 $zl = $zl - ( (int)( ( 30 - $zj ) / 15 ) ) * ( (int)( ( 17719 * $zj ) / 50 ) ) -
1721                         ( (int)( $zj / 16 ) ) * ( (int)( ( 15238 * $zj ) / 43 ) ) + 29;
1722                 $zm = (int)( ( 24 * $zl ) / 709 );
1723                 $zd = $zl - (int)( ( 709 * $zm ) / 24 );
1724                 $zy = 30 * $zn + $zj - 30;
1725
1726                 return [ $zy, $zm, $zd ];
1727         }
1728
1729         /**
1730          * Converting Gregorian dates to Hebrew dates.
1731          *
1732          * Based on a JavaScript code by Abu Mami and Yisrael Hersch
1733          * (abu-mami@kaluach.net, http://www.kaluach.net), who permitted
1734          * to translate the relevant functions into PHP and release them under
1735          * GNU GPL.
1736          *
1737          * The months are counted from Tishrei = 1. In a leap year, Adar I is 13
1738          * and Adar II is 14. In a non-leap year, Adar is 6.
1739          *
1740          * @param string $ts
1741          *
1742          * @return int[]
1743          */
1744         private static function tsToHebrew( $ts ) {
1745                 # Parse date
1746                 $year = substr( $ts, 0, 4 );
1747                 $month = substr( $ts, 4, 2 );
1748                 $day = substr( $ts, 6, 2 );
1749
1750                 # Calculate Hebrew year
1751                 $hebrewYear = $year + 3760;
1752
1753                 # Month number when September = 1, August = 12
1754                 $month += 4;
1755                 if ( $month > 12 ) {
1756                         # Next year
1757                         $month -= 12;
1758                         $year++;
1759                         $hebrewYear++;
1760                 }
1761
1762                 # Calculate day of year from 1 September
1763                 $dayOfYear = $day;
1764                 for ( $i = 1; $i < $month; $i++ ) {
1765                         if ( $i == 6 ) {
1766                                 # February
1767                                 $dayOfYear += 28;
1768                                 # Check if the year is leap
1769                                 if ( $year % 400 == 0 || ( $year % 4 == 0 && $year % 100 > 0 ) ) {
1770                                         $dayOfYear++;
1771                                 }
1772                         } elseif ( $i == 8 || $i == 10 || $i == 1 || $i == 3 ) {
1773                                 $dayOfYear += 30;
1774                         } else {
1775                                 $dayOfYear += 31;
1776                         }
1777                 }
1778
1779                 # Calculate the start of the Hebrew year
1780                 $start = self::hebrewYearStart( $hebrewYear );
1781
1782                 # Calculate next year's start
1783                 if ( $dayOfYear <= $start ) {
1784                         # Day is before the start of the year - it is the previous year
1785                         # Next year's start
1786                         $nextStart = $start;
1787                         # Previous year
1788                         $year--;
1789                         $hebrewYear--;
1790                         # Add days since previous year's 1 September
1791                         $dayOfYear += 365;
1792                         if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1793                                 # Leap year
1794                                 $dayOfYear++;
1795                         }
1796                         # Start of the new (previous) year
1797                         $start = self::hebrewYearStart( $hebrewYear );
1798                 } else {
1799                         # Next year's start
1800                         $nextStart = self::hebrewYearStart( $hebrewYear + 1 );
1801                 }
1802
1803                 # Calculate Hebrew day of year
1804                 $hebrewDayOfYear = $dayOfYear - $start;
1805
1806                 # Difference between year's days
1807                 $diff = $nextStart - $start;
1808                 # Add 12 (or 13 for leap years) days to ignore the difference between
1809                 # Hebrew and Gregorian year (353 at least vs. 365/6) - now the
1810                 # difference is only about the year type
1811                 if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1812                         $diff += 13;
1813                 } else {
1814                         $diff += 12;
1815                 }
1816
1817                 # Check the year pattern, and is leap year
1818                 # 0 means an incomplete year, 1 means a regular year, 2 means a complete year
1819                 # This is mod 30, to work on both leap years (which add 30 days of Adar I)
1820                 # and non-leap years
1821                 $yearPattern = $diff % 30;
1822                 # Check if leap year
1823                 $isLeap = $diff >= 30;
1824
1825                 # Calculate day in the month from number of day in the Hebrew year
1826                 # Don't check Adar - if the day is not in Adar, we will stop before;
1827                 # if it is in Adar, we will use it to check if it is Adar I or Adar II
1828                 $hebrewDay = $hebrewDayOfYear;
1829                 $hebrewMonth = 1;
1830                 $days = 0;
1831                 while ( $hebrewMonth <= 12 ) {
1832                         # Calculate days in this month
1833                         if ( $isLeap && $hebrewMonth == 6 ) {
1834                                 # Leap year - has Adar I, with 30 days, and Adar II, with 29 days
1835                                 $days = 30;
1836                                 if ( $hebrewDay <= $days ) {
1837                                         # Day in Adar I
1838                                         $hebrewMonth = 13;
1839                                 } else {
1840                                         # Subtract the days of Adar I
1841                                         $hebrewDay -= $days;
1842                                         # Try Adar II
1843                                         $days = 29;
1844                                         if ( $hebrewDay <= $days ) {
1845                                                 # Day in Adar II
1846                                                 $hebrewMonth = 14;
1847                                         }
1848                                 }
1849                         } elseif ( $hebrewMonth == 2 && $yearPattern == 2 ) {
1850                                 # Cheshvan in a complete year (otherwise as the rule below)
1851                                 $days = 30;
1852                         } elseif ( $hebrewMonth == 3 && $yearPattern == 0 ) {
1853                                 # Kislev in an incomplete year (otherwise as the rule below)
1854                                 $days = 29;
1855                         } else {
1856                                 # Odd months have 30 days, even have 29
1857                                 $days = 30 - ( $hebrewMonth - 1 ) % 2;
1858                         }
1859                         if ( $hebrewDay <= $days ) {
1860                                 # In the current month
1861                                 break;
1862                         } else {
1863                                 # Subtract the days of the current month
1864                                 $hebrewDay -= $days;
1865                                 # Try in the next month
1866                                 $hebrewMonth++;
1867                         }
1868                 }
1869
1870                 return [ $hebrewYear, $hebrewMonth, $hebrewDay, $days ];
1871         }
1872
1873         /**
1874          * This calculates the Hebrew year start, as days since 1 September.
1875          * Based on Carl Friedrich Gauss algorithm for finding Easter date.
1876          * Used for Hebrew date.
1877          *
1878          * @param int $year
1879          *
1880          * @return string
1881          */
1882         private static function hebrewYearStart( $year ) {
1883                 $a = intval( ( 12 * ( $year - 1 ) + 17 ) % 19 );
1884                 $b = intval( ( $year - 1 ) % 4 );
1885                 $m = 32.044093161144 + 1.5542417966212 * $a + $b / 4.0 - 0.0031777940220923 * ( $year - 1 );
1886                 if ( $m < 0 ) {
1887                         $m--;
1888                 }
1889                 $Mar = intval( $m );
1890                 if ( $m < 0 ) {
1891                         $m++;
1892                 }
1893                 $m -= $Mar;
1894
1895                 $c = intval( ( $Mar + 3 * ( $year - 1 ) + 5 * $b + 5 ) % 7 );
1896                 if ( $c == 0 && $a > 11 && $m >= 0.89772376543210 ) {
1897                         $Mar++;
1898                 } elseif ( $c == 1 && $a > 6 && $m >= 0.63287037037037 ) {
1899                         $Mar += 2;
1900                 } elseif ( $c == 2 || $c == 4 || $c == 6 ) {
1901                         $Mar++;
1902                 }
1903
1904                 $Mar += intval( ( $year - 3761 ) / 100 ) - intval( ( $year - 3761 ) / 400 ) - 24;
1905                 return $Mar;
1906         }
1907
1908         /**
1909          * Algorithm to convert Gregorian dates to Thai solar dates,
1910          * Minguo dates or Minguo dates.
1911          *
1912          * Link: https://en.wikipedia.org/wiki/Thai_solar_calendar
1913          *       https://en.wikipedia.org/wiki/Minguo_calendar
1914          *       https://en.wikipedia.org/wiki/Japanese_era_name
1915          *
1916          * @param string $ts 14-character timestamp
1917          * @param string $cName Calender name
1918          * @return array Converted year, month, day
1919          */
1920         private static function tsToYear( $ts, $cName ) {
1921                 $gy = substr( $ts, 0, 4 );
1922                 $gm = substr( $ts, 4, 2 );
1923                 $gd = substr( $ts, 6, 2 );
1924
1925                 if ( !strcmp( $cName, 'thai' ) ) {
1926                         # Thai solar dates
1927                         # Add 543 years to the Gregorian calendar
1928                         # Months and days are identical
1929                         $gy_offset = $gy + 543;
1930                         # fix for dates between 1912 and 1941
1931                         # https://en.wikipedia.org/?oldid=836596673#New_year
1932                         if ( $gy >= 1912 && $gy <= 1940 ) {
1933                                 if ( $gm <= 3 ) {
1934                                         $gy_offset--;
1935                                 }
1936                                 $gm = ( $gm - 3 ) % 12;
1937                         }
1938                 } elseif ( ( !strcmp( $cName, 'minguo' ) ) || !strcmp( $cName, 'juche' ) ) {
1939                         # Minguo dates
1940                         # Deduct 1911 years from the Gregorian calendar
1941                         # Months and days are identical
1942                         $gy_offset = $gy - 1911;
1943                 } elseif ( !strcmp( $cName, 'tenno' ) ) {
1944                         # Nengō dates up to Meiji period
1945                         # Deduct years from the Gregorian calendar
1946                         # depending on the nengo periods
1947                         # Months and days are identical
1948                         if ( ( $gy < 1912 )
1949                                 || ( ( $gy == 1912 ) && ( $gm < 7 ) )
1950                                 || ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd < 31 ) )
1951                         ) {
1952                                 # Meiji period
1953                                 $gy_gannen = $gy - 1868 + 1;
1954                                 $gy_offset = $gy_gannen;
1955                                 if ( $gy_gannen == 1 ) {
1956                                         $gy_offset = '元';
1957                                 }
1958                                 $gy_offset = '明治' . $gy_offset;
1959                         } elseif (
1960                                 ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd == 31 ) ) ||
1961                                 ( ( $gy == 1912 ) && ( $gm >= 8 ) ) ||
1962                                 ( ( $gy > 1912 ) && ( $gy < 1926 ) ) ||
1963                                 ( ( $gy == 1926 ) && ( $gm < 12 ) ) ||
1964                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd < 26 ) )
1965                         ) {
1966                                 # Taishō period
1967                                 $gy_gannen = $gy - 1912 + 1;
1968                                 $gy_offset = $gy_gannen;
1969                                 if ( $gy_gannen == 1 ) {
1970                                         $gy_offset = '元';
1971                                 }
1972                                 $gy_offset = '大正' . $gy_offset;
1973                         } elseif (
1974                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd >= 26 ) ) ||
1975                                 ( ( $gy > 1926 ) && ( $gy < 1989 ) ) ||
1976                                 ( ( $gy == 1989 ) && ( $gm == 1 ) && ( $gd < 8 ) )
1977                         ) {
1978                                 # Shōwa period
1979                                 $gy_gannen = $gy - 1926 + 1;
1980                                 $gy_offset = $gy_gannen;
1981                                 if ( $gy_gannen == 1 ) {
1982                                         $gy_offset = '元';
1983                                 }
1984                                 $gy_offset = '昭和' . $gy_offset;
1985                         } elseif (
1986                                 ( ( $gy == 1989 ) && ( $gm == 1 ) && ( $gd >= 8 ) ) ||
1987                                 ( ( $gy > 1989 ) && ( $gy < 2019 ) ) ||
1988                                 ( ( $gy == 2019 ) && ( $gm < 5 ) )
1989                         ) {
1990                                 # Heisei period
1991                                 $gy_gannen = $gy - 1989 + 1;
1992                                 $gy_offset = $gy_gannen;
1993                                 if ( $gy_gannen == 1 ) {
1994                                         $gy_offset = '元';
1995                                 }
1996                                 $gy_offset = '平成' . $gy_offset;
1997                         } else {
1998                                 # Reiwa period
1999                                 $gy_gannen = $gy - 2019 + 1;
2000                                 $gy_offset = $gy_gannen;
2001                                 if ( $gy_gannen == 1 ) {
2002                                         $gy_offset = '元';
2003                                 }
2004                                 $gy_offset = '令和' . $gy_offset;
2005                         }
2006                 } else {
2007                         $gy_offset = $gy;
2008                 }
2009
2010                 return [ $gy_offset, $gm, $gd ];
2011         }
2012
2013         /**
2014          * Gets directionality of the first strongly directional codepoint, for embedBidi()
2015          *
2016          * This is the rule the BIDI algorithm uses to determine the directionality of
2017          * paragraphs ( https://www.unicode.org/reports/tr9/#The_Paragraph_Level ) and
2018          * FSI isolates ( https://www.unicode.org/reports/tr9/#Explicit_Directional_Isolates ).
2019          *
2020          * TODO: Does not handle BIDI control characters inside the text.
2021          * TODO: Does not handle unallocated characters.
2022          *
2023          * @param string $text Text to test
2024          * @return null|string Directionality ('ltr' or 'rtl') or null
2025          */
2026         private static function strongDirFromContent( $text = '' ) {
2027                 if ( !preg_match( self::$strongDirRegex, $text, $matches ) ) {
2028                         return null;
2029                 }
2030                 if ( $matches[1] === '' ) {
2031                         return 'rtl';
2032                 }
2033                 return 'ltr';
2034         }
2035
2036         /**
2037          * Roman number formatting up to 10000
2038          *
2039          * @param int $num
2040          *
2041          * @return string
2042          */
2043         static function romanNumeral( $num ) {
2044                 static $table = [
2045                         [ '', 'I', 'II', 'III', 'IV', 'V', 'VI', 'VII', 'VIII', 'IX', 'X' ],
2046                         [ '', 'X', 'XX', 'XXX', 'XL', 'L', 'LX', 'LXX', 'LXXX', 'XC', 'C' ],
2047                         [ '', 'C', 'CC', 'CCC', 'CD', 'D', 'DC', 'DCC', 'DCCC', 'CM', 'M' ],
2048                         [ '', 'M', 'MM', 'MMM', 'MMMM', 'MMMMM', 'MMMMMM', 'MMMMMMM',
2049                                 'MMMMMMMM', 'MMMMMMMMM', 'MMMMMMMMMM' ]
2050                 ];
2051
2052                 $num = intval( $num );
2053                 if ( $num > 10000 || $num <= 0 ) {
2054                         return $num;
2055                 }
2056
2057                 $s = '';
2058                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
2059                         if ( $num >= $pow10 ) {
2060                                 $s .= $table[$i][(int)floor( $num / $pow10 )];
2061                         }
2062                         $num = $num % $pow10;
2063                 }
2064                 return $s;
2065         }
2066
2067         /**
2068          * Hebrew Gematria number formatting up to 9999
2069          *
2070          * @param int $num
2071          *
2072          * @return string
2073          */
2074         static function hebrewNumeral( $num ) {
2075                 static $table = [
2076                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ],
2077                         [ '', 'י', 'כ', 'ל', 'מ', 'נ', 'ס', 'ע', 'פ', 'צ', 'ק' ],
2078                         [ '',
2079                                 [ 'ק' ],
2080                                 [ 'ר' ],
2081                                 [ 'ש' ],
2082                                 [ 'ת' ],
2083                                 [ 'ת', 'ק' ],
2084                                 [ 'ת', 'ר' ],
2085                                 [ 'ת', 'ש' ],
2086                                 [ 'ת', 'ת' ],
2087                                 [ 'ת', 'ת', 'ק' ],
2088                                 [ 'ת', 'ת', 'ר' ],
2089                         ],
2090                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ]
2091                 ];
2092
2093                 $num = intval( $num );
2094                 if ( $num > 9999 || $num <= 0 ) {
2095                         return $num;
2096                 }
2097
2098                 // Round thousands have special notations
2099                 if ( $num === 1000 ) {
2100                         return "א' אלף";
2101                 } elseif ( $num % 1000 === 0 ) {
2102                         return $table[0][$num / 1000] . "' אלפים";
2103                 }
2104
2105                 $letters = [];
2106
2107                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
2108                         if ( $num >= $pow10 ) {
2109                                 if ( $num === 15 || $num === 16 ) {
2110                                         $letters[] = $table[0][9];
2111                                         $letters[] = $table[0][$num - 9];
2112                                         $num = 0;
2113                                 } else {
2114                                         $letters = array_merge(
2115                                                 $letters,
2116                                                 (array)$table[$i][intval( $num / $pow10 )]
2117                                         );
2118
2119                                         if ( $pow10 === 1000 ) {
2120                                                 $letters[] = "'";
2121                                         }
2122                                 }
2123                         }
2124
2125                         $num = $num % $pow10;
2126                 }
2127
2128                 $preTransformLength = count( $letters );
2129                 if ( $preTransformLength === 1 ) {
2130                         // Add geresh (single quote) to one-letter numbers
2131                         $letters[] = "'";
2132                 } else {
2133                         $lastIndex = $preTransformLength - 1;
2134                         $letters[$lastIndex] = str_replace(
2135                                 [ 'כ', 'מ', 'נ', 'פ', 'צ' ],
2136                                 [ 'ך', 'ם', 'ן', 'ף', 'ץ' ],
2137                                 $letters[$lastIndex]
2138                         );
2139
2140                         // Add gershayim (double quote) to multiple-letter numbers,
2141                         // but exclude numbers with only one letter after the thousands
2142                         // (1001-1009, 1020, 1030, 2001-2009, etc.)
2143                         if ( $letters[1] === "'" && $preTransformLength === 3 ) {
2144                                 $letters[] = "'";
2145                         } else {
2146                                 array_splice( $letters, -1, 0, '"' );
2147                         }
2148                 }
2149
2150                 return implode( $letters );
2151         }
2152
2153         /**
2154          * Used by date() and time() to adjust the time output.
2155          *
2156          * @param string $ts The time in date('YmdHis') format
2157          * @param mixed $tz Adjust the time by this amount (default false, mean we
2158          *   get user timecorrection setting)
2159          * @return int
2160          */
2161         public function userAdjust( $ts, $tz = false ) {
2162                 global $wgUser, $wgLocalTZoffset;
2163
2164                 if ( $tz === false ) {
2165                         $tz = $wgUser->getOption( 'timecorrection' );
2166                 }
2167
2168                 $data = explode( '|', $tz, 3 );
2169
2170                 if ( $data[0] == 'ZoneInfo' ) {
2171                         try {
2172                                 $userTZ = new DateTimeZone( $data[2] );
2173                                 $date = new DateTime( $ts, new DateTimeZone( 'UTC' ) );
2174                                 $date->setTimezone( $userTZ );
2175                                 return $date->format( 'YmdHis' );
2176                         } catch ( Exception $e ) {
2177                                 // Unrecognized timezone, default to 'Offset' with the stored offset.
2178                                 $data[0] = 'Offset';
2179                         }
2180                 }
2181
2182                 if ( $data[0] == 'System' || $tz == '' ) {
2183                         # Global offset in minutes.
2184                         $minDiff = $wgLocalTZoffset;
2185                 } elseif ( $data[0] == 'Offset' ) {
2186                         $minDiff = intval( $data[1] );
2187                 } else {
2188                         $data = explode( ':', $tz );
2189                         if ( count( $data ) == 2 ) {
2190                                 $data[0] = intval( $data[0] );
2191                                 $data[1] = intval( $data[1] );
2192                                 $minDiff = abs( $data[0] ) * 60 + $data[1];
2193                                 if ( $data[0] < 0 ) {
2194                                         $minDiff = -$minDiff;
2195                                 }
2196                         } else {
2197                                 $minDiff = intval( $data[0] ) * 60;
2198                         }
2199                 }
2200
2201                 # No difference ? Return time unchanged
2202                 if ( $minDiff == 0 ) {
2203                         return $ts;
2204                 }
2205
2206                 Wikimedia\suppressWarnings(); // E_STRICT system time bitching
2207                 # Generate an adjusted date; take advantage of the fact that mktime
2208                 # will normalize out-of-range values so we don't have to split $minDiff
2209                 # into hours and minutes.
2210                 $t = mktime( (
2211                         (int)substr( $ts, 8, 2 ) ), # Hours
2212                         (int)substr( $ts, 10, 2 ) + $minDiff, # Minutes
2213                         (int)substr( $ts, 12, 2 ), # Seconds
2214                         (int)substr( $ts, 4, 2 ), # Month
2215                         (int)substr( $ts, 6, 2 ), # Day
2216                         (int)substr( $ts, 0, 4 ) ); # Year
2217
2218                 $date = date( 'YmdHis', $t );
2219                 Wikimedia\restoreWarnings();
2220
2221                 return $date;
2222         }
2223
2224         /**
2225          * This is meant to be used by time(), date(), and timeanddate() to get
2226          * the date preference they're supposed to use, it should be used in
2227          * all children.
2228          *
2229          *     function timeanddate([...], $format = true) {
2230          *       $datePreference = $this->dateFormat($format);
2231          *       [...]
2232          *     }
2233          *
2234          * @param int|string|bool $usePrefs If true, the user's preference is used
2235          *   if false, the site/language default is used
2236          *   if int/string, assumed to be a format.
2237          * @return string
2238          */
2239         function dateFormat( $usePrefs = true ) {
2240                 global $wgUser;
2241
2242                 if ( is_bool( $usePrefs ) ) {
2243                         if ( $usePrefs ) {
2244                                 $datePreference = $wgUser->getDatePreference();
2245                         } else {
2246                                 $datePreference = (string)User::getDefaultOption( 'date' );
2247                         }
2248                 } else {
2249                         $datePreference = (string)$usePrefs;
2250                 }
2251
2252                 // return int
2253                 if ( $datePreference == '' ) {
2254                         return 'default';
2255                 }
2256
2257                 return $datePreference;
2258         }
2259
2260         /**
2261          * Get a format string for a given type and preference
2262          * @param string $type May be 'date', 'time', 'both', or 'pretty'.
2263          * @param string $pref The format name as it appears in Messages*.php under
2264          *  $datePreferences.
2265          *
2266          * @since 1.22 New type 'pretty' that provides a more readable timestamp format
2267          *
2268          * @return string
2269          */
2270         function getDateFormatString( $type, $pref ) {
2271                 $wasDefault = false;
2272                 if ( $pref == 'default' ) {
2273                         $wasDefault = true;
2274                         $pref = $this->getDefaultDateFormat();
2275                 }
2276
2277                 if ( !isset( $this->dateFormatStrings[$type][$pref] ) ) {
2278                         $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2279
2280                         if ( $type === 'pretty' && $df === null ) {
2281                                 $df = $this->getDateFormatString( 'date', $pref );
2282                         }
2283
2284                         if ( !$wasDefault && $df === null ) {
2285                                 $pref = $this->getDefaultDateFormat();
2286                                 $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2287                         }
2288
2289                         $this->dateFormatStrings[$type][$pref] = $df;
2290                 }
2291                 return $this->dateFormatStrings[$type][$pref];
2292         }
2293
2294         /**
2295          * @param string $ts The time format which needs to be turned into a
2296          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2297          * @param bool $adj Whether to adjust the time output according to the
2298          *   user configured offset ($timecorrection)
2299          * @param mixed $format True to use user's date format preference
2300          * @param string|bool $timecorrection The time offset as returned by
2301          *   validateTimeZone() in Special:Preferences
2302          * @return string
2303          */
2304         public function date( $ts, $adj = false, $format = true, $timecorrection = false ) {
2305                 $ts = wfTimestamp( TS_MW, $ts );
2306                 if ( $adj ) {
2307                         $ts = $this->userAdjust( $ts, $timecorrection );
2308                 }
2309                 $df = $this->getDateFormatString( 'date', $this->dateFormat( $format ) );
2310                 return $this->sprintfDate( $df, $ts );
2311         }
2312
2313         /**
2314          * @param string $ts The time format which needs to be turned into a
2315          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2316          * @param bool $adj Whether to adjust the time output according to the
2317          *   user configured offset ($timecorrection)
2318          * @param mixed $format True to use user's date format preference
2319          * @param string|bool $timecorrection The time offset as returned by
2320          *   validateTimeZone() in Special:Preferences
2321          * @return string
2322          */
2323         public function time( $ts, $adj = false, $format = true, $timecorrection = false ) {
2324                 $ts = wfTimestamp( TS_MW, $ts );
2325                 if ( $adj ) {
2326                         $ts = $this->userAdjust( $ts, $timecorrection );
2327                 }
2328                 $df = $this->getDateFormatString( 'time', $this->dateFormat( $format ) );
2329                 return $this->sprintfDate( $df, $ts );
2330         }
2331
2332         /**
2333          * @param string $ts The time format which needs to be turned into a
2334          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2335          * @param bool $adj Whether to adjust the time output according to the
2336          *   user configured offset ($timecorrection)
2337          * @param mixed $format What format to return, if it's false output the
2338          *   default one (default true)
2339          * @param string|bool $timecorrection The time offset as returned by
2340          *   validateTimeZone() in Special:Preferences
2341          * @return string
2342          */
2343         public function timeanddate( $ts, $adj = false, $format = true, $timecorrection = false ) {
2344                 $ts = wfTimestamp( TS_MW, $ts );
2345                 if ( $adj ) {
2346                         $ts = $this->userAdjust( $ts, $timecorrection );
2347                 }
2348                 $df = $this->getDateFormatString( 'both', $this->dateFormat( $format ) );
2349                 return $this->sprintfDate( $df, $ts );
2350         }
2351
2352         /**
2353          * Takes a number of seconds and turns it into a text using values such as hours and minutes.
2354          *
2355          * @since 1.20
2356          *
2357          * @param int $seconds The amount of seconds.
2358          * @param array $chosenIntervals The intervals to enable.
2359          *
2360          * @return string
2361          */
2362         public function formatDuration( $seconds, array $chosenIntervals = [] ) {
2363                 $intervals = $this->getDurationIntervals( $seconds, $chosenIntervals );
2364
2365                 $segments = [];
2366
2367                 foreach ( $intervals as $intervalName => $intervalValue ) {
2368                         // Messages: duration-seconds, duration-minutes, duration-hours, duration-days, duration-weeks,
2369                         // duration-years, duration-decades, duration-centuries, duration-millennia
2370                         $message = wfMessage( 'duration-' . $intervalName )->numParams( $intervalValue );
2371                         $segments[] = $message->inLanguage( $this )->escaped();
2372                 }
2373
2374                 return $this->listToText( $segments );
2375         }
2376
2377         /**
2378          * Takes a number of seconds and returns an array with a set of corresponding intervals.
2379          * For example 65 will be turned into [ minutes => 1, seconds => 5 ].
2380          *
2381          * @since 1.20
2382          *
2383          * @param int $seconds The amount of seconds.
2384          * @param array $chosenIntervals The intervals to enable.
2385          *
2386          * @return array
2387          */
2388         public function getDurationIntervals( $seconds, array $chosenIntervals = [] ) {
2389                 if ( empty( $chosenIntervals ) ) {
2390                         $chosenIntervals = [
2391                                 'millennia',
2392                                 'centuries',
2393                                 'decades',
2394                                 'years',
2395                                 'days',
2396                                 'hours',
2397                                 'minutes',
2398                                 'seconds'
2399                         ];
2400                 }
2401
2402                 $intervals = array_intersect_key( self::$durationIntervals, array_flip( $chosenIntervals ) );
2403                 $sortedNames = array_keys( $intervals );
2404                 $smallestInterval = array_pop( $sortedNames );
2405
2406                 $segments = [];
2407
2408                 foreach ( $intervals as $name => $length ) {
2409                         $value = floor( $seconds / $length );
2410
2411                         if ( $value > 0 || ( $name == $smallestInterval && empty( $segments ) ) ) {
2412                                 $seconds -= $value * $length;
2413                                 $segments[$name] = $value;
2414                         }
2415                 }
2416
2417                 return $segments;
2418         }
2419
2420         /**
2421          * Internal helper function for userDate(), userTime() and userTimeAndDate()
2422          *
2423          * @param string $type Can be 'date', 'time' or 'both'
2424          * @param string $ts The time format which needs to be turned into a
2425          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2426          * @param User $user User object used to get preferences for timezone and format
2427          * @param array $options Array, can contain the following keys:
2428          *   - 'timecorrection': time correction, can have the following values:
2429          *     - true: use user's preference
2430          *     - false: don't use time correction
2431          *     - int: value of time correction in minutes
2432          *   - 'format': format to use, can have the following values:
2433          *     - true: use user's preference
2434          *     - false: use default preference
2435          *     - string: format to use
2436          * @since 1.19
2437          * @return string
2438          */
2439         private function internalUserTimeAndDate( $type, $ts, User $user, array $options ) {
2440                 $ts = wfTimestamp( TS_MW, $ts );
2441                 $options += [ 'timecorrection' => true, 'format' => true ];
2442                 if ( $options['timecorrection'] !== false ) {
2443                         if ( $options['timecorrection'] === true ) {
2444                                 $offset = $user->getOption( 'timecorrection' );
2445                         } else {
2446                                 $offset = $options['timecorrection'];
2447                         }
2448                         $ts = $this->userAdjust( $ts, $offset );
2449                 }
2450                 if ( $options['format'] === true ) {
2451                         $format = $user->getDatePreference();
2452                 } else {
2453                         $format = $options['format'];
2454                 }
2455                 $df = $this->getDateFormatString( $type, $this->dateFormat( $format ) );
2456                 return $this->sprintfDate( $df, $ts );
2457         }
2458
2459         /**
2460          * Get the formatted date for the given timestamp and formatted for
2461          * the given user.
2462          *
2463          * @param mixed $ts Mixed: the time format which needs to be turned into a
2464          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2465          * @param User $user User object used to get preferences for timezone and format
2466          * @param array $options Array, can contain the following keys:
2467          *   - 'timecorrection': time correction, can have the following values:
2468          *     - true: use user's preference
2469          *     - false: don't use time correction
2470          *     - int: value of time correction in minutes
2471          *   - 'format': format to use, can have the following values:
2472          *     - true: use user's preference
2473          *     - false: use default preference
2474          *     - string: format to use
2475          * @since 1.19
2476          * @return string
2477          */
2478         public function userDate( $ts, User $user, array $options = [] ) {
2479                 return $this->internalUserTimeAndDate( 'date', $ts, $user, $options );
2480         }
2481
2482         /**
2483          * Get the formatted time for the given timestamp and formatted for
2484          * the given user.
2485          *
2486          * @param mixed $ts The time format which needs to be turned into a
2487          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2488          * @param User $user User object used to get preferences for timezone and format
2489          * @param array $options Array, can contain the following keys:
2490          *   - 'timecorrection': time correction, can have the following values:
2491          *     - true: use user's preference
2492          *     - false: don't use time correction
2493          *     - int: value of time correction in minutes
2494          *   - 'format': format to use, can have the following values:
2495          *     - true: use user's preference
2496          *     - false: use default preference
2497          *     - string: format to use
2498          * @since 1.19
2499          * @return string
2500          */
2501         public function userTime( $ts, User $user, array $options = [] ) {
2502                 return $this->internalUserTimeAndDate( 'time', $ts, $user, $options );
2503         }
2504
2505         /**
2506          * Get the formatted date and time for the given timestamp and formatted for
2507          * the given user.
2508          *
2509          * @param mixed $ts The time format which needs to be turned into a
2510          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2511          * @param User $user User object used to get preferences for timezone and format
2512          * @param array $options Array, can contain the following keys:
2513          *   - 'timecorrection': time correction, can have the following values:
2514          *     - true: use user's preference
2515          *     - false: don't use time correction
2516          *     - int: value of time correction in minutes
2517          *   - 'format': format to use, can have the following values:
2518          *     - true: use user's preference
2519          *     - false: use default preference
2520          *     - string: format to use
2521          * @since 1.19
2522          * @return string
2523          */
2524         public function userTimeAndDate( $ts, User $user, array $options = [] ) {
2525                 return $this->internalUserTimeAndDate( 'both', $ts, $user, $options );
2526         }
2527
2528         /**
2529          * Get the timestamp in a human-friendly relative format, e.g., "3 days ago".
2530          *
2531          * Determine the difference between the timestamp and the current time, and
2532          * generate a readable timestamp by returning "<N> <units> ago", where the
2533          * largest possible unit is used.
2534          *
2535          * @since 1.26 (Prior to 1.26 method existed but was not meant to be used directly)
2536          *
2537          * @param MWTimestamp $time
2538          * @param MWTimestamp|null $relativeTo The base timestamp to compare to (defaults to now)
2539          * @param User|null $user User the timestamp is being generated for
2540          *  (or null to use main context's user)
2541          * @return string Formatted timestamp
2542          */
2543         public function getHumanTimestamp(
2544                 MWTimestamp $time, MWTimestamp $relativeTo = null, User $user = null
2545         ) {
2546                 if ( $relativeTo === null ) {
2547                         $relativeTo = new MWTimestamp();
2548                 }
2549                 if ( $user === null ) {
2550                         $user = RequestContext::getMain()->getUser();
2551                 }
2552
2553                 // Adjust for the user's timezone.
2554                 $offsetThis = $time->offsetForUser( $user );
2555                 $offsetRel = $relativeTo->offsetForUser( $user );
2556
2557                 $ts = '';
2558                 if ( Hooks::run( 'GetHumanTimestamp', [ &$ts, $time, $relativeTo, $user, $this ] ) ) {
2559                         $ts = $this->getHumanTimestampInternal( $time, $relativeTo, $user );
2560                 }
2561
2562                 // Reset the timezone on the objects.
2563                 $time->timestamp->sub( $offsetThis );
2564                 $relativeTo->timestamp->sub( $offsetRel );
2565
2566                 return $ts;
2567         }
2568
2569         /**
2570          * Convert an MWTimestamp into a pretty human-readable timestamp using
2571          * the given user preferences and relative base time.
2572          *
2573          * @see Language::getHumanTimestamp
2574          * @param MWTimestamp $ts Timestamp to prettify
2575          * @param MWTimestamp $relativeTo Base timestamp
2576          * @param User $user User preferences to use
2577          * @return string Human timestamp
2578          * @since 1.26
2579          */
2580         private function getHumanTimestampInternal(
2581                 MWTimestamp $ts, MWTimestamp $relativeTo, User $user
2582         ) {
2583                 $diff = $ts->diff( $relativeTo );
2584                 $diffDay = (bool)( (int)$ts->timestamp->format( 'w' ) -
2585                         (int)$relativeTo->timestamp->format( 'w' ) );
2586                 $days = $diff->days ?: (int)$diffDay;
2587                 if ( $diff->invert || $days > 5
2588                         && $ts->timestamp->format( 'Y' ) !== $relativeTo->timestamp->format( 'Y' )
2589                 ) {
2590                         // Timestamps are in different years: use full timestamp
2591                         // Also do full timestamp for future dates
2592                         /**
2593                          * @todo FIXME: Add better handling of future timestamps.
2594                          */
2595                         $format = $this->getDateFormatString( 'both', $user->getDatePreference() ?: 'default' );
2596                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2597                 } elseif ( $days > 5 ) {
2598                         // Timestamps are in same year,  but more than 5 days ago: show day and month only.
2599                         $format = $this->getDateFormatString( 'pretty', $user->getDatePreference() ?: 'default' );
2600                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2601                 } elseif ( $days > 1 ) {
2602                         // Timestamp within the past week: show the day of the week and time
2603                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2604                         $weekday = self::$mWeekdayMsgs[$ts->timestamp->format( 'w' )];
2605                         // Messages:
2606                         // sunday-at, monday-at, tuesday-at, wednesday-at, thursday-at, friday-at, saturday-at
2607                         $ts = wfMessage( "$weekday-at" )
2608                                 ->inLanguage( $this )
2609                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2610                                 ->text();
2611                 } elseif ( $days == 1 ) {
2612                         // Timestamp was yesterday: say 'yesterday' and the time.
2613                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2614                         $ts = wfMessage( 'yesterday-at' )
2615                                 ->inLanguage( $this )
2616                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2617                                 ->text();
2618                 } elseif ( $diff->h > 1 || $diff->h == 1 && $diff->i > 30 ) {
2619                         // Timestamp was today, but more than 90 minutes ago: say 'today' and the time.
2620                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2621                         $ts = wfMessage( 'today-at' )
2622                                 ->inLanguage( $this )
2623                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2624                                 ->text();
2625
2626                 // From here on in, the timestamp was soon enough ago so that we can simply say
2627                 // XX units ago, e.g., "2 hours ago" or "5 minutes ago"
2628                 } elseif ( $diff->h == 1 ) {
2629                         // Less than 90 minutes, but more than an hour ago.
2630                         $ts = wfMessage( 'hours-ago' )->inLanguage( $this )->numParams( 1 )->text();
2631                 } elseif ( $diff->i >= 1 ) {
2632                         // A few minutes ago.
2633                         $ts = wfMessage( 'minutes-ago' )->inLanguage( $this )->numParams( $diff->i )->text();
2634                 } elseif ( $diff->s >= 30 ) {
2635                         // Less than a minute, but more than 30 sec ago.
2636                         $ts = wfMessage( 'seconds-ago' )->inLanguage( $this )->numParams( $diff->s )->text();
2637                 } else {
2638                         // Less than 30 seconds ago.
2639                         $ts = wfMessage( 'just-now' )->text();
2640                 }
2641
2642                 return $ts;
2643         }
2644
2645         /**
2646          * @param string $key
2647          * @return string|null
2648          */
2649         public function getMessage( $key ) {
2650                 return self::$dataCache->getSubitem( $this->mCode, 'messages', $key );
2651         }
2652
2653         /**
2654          * @return array
2655          */
2656         function getAllMessages() {
2657                 return self::$dataCache->getItem( $this->mCode, 'messages' );
2658         }
2659
2660         /**
2661          * @param string $in
2662          * @param string $out
2663          * @param string $string
2664          * @return string
2665          */
2666         public function iconv( $in, $out, $string ) {
2667                 # Even with //IGNORE iconv can whine about illegal characters in
2668                 # *input* string. We just ignore those too.
2669                 # REF: https://bugs.php.net/bug.php?id=37166
2670                 # REF: https://phabricator.wikimedia.org/T18885
2671                 Wikimedia\suppressWarnings();
2672                 $text = iconv( $in, $out . '//IGNORE', $string );
2673                 Wikimedia\restoreWarnings();
2674                 return $text;
2675         }
2676
2677         // callback functions for ucwords(), ucwordbreaks()
2678
2679         /**
2680          * @param array $matches
2681          * @return mixed|string
2682          */
2683         function ucwordbreaksCallbackAscii( $matches ) {
2684                 return $this->ucfirst( $matches[1] );
2685         }
2686
2687         /**
2688          * @param array $matches
2689          * @return string
2690          */
2691         function ucwordbreaksCallbackMB( $matches ) {
2692                 return mb_strtoupper( $matches[0] );
2693         }
2694
2695         /**
2696          * @param array $matches
2697          * @return string
2698          */
2699         function ucwordsCallbackMB( $matches ) {
2700                 return mb_strtoupper( $matches[0] );
2701         }
2702
2703         /**
2704          * Make a string's first character uppercase
2705          *
2706          * @param string $str
2707          *
2708          * @return string
2709          */
2710         public function ucfirst( $str ) {
2711                 $o = ord( $str );
2712                 if ( $o < 96 ) { // if already uppercase...
2713                         return $str;
2714                 } elseif ( $o < 128 ) {
2715                         return ucfirst( $str ); // use PHP's ucfirst()
2716                 } else {
2717                         // fall back to more complex logic in case of multibyte strings
2718                         return $this->uc( $str, true );
2719                 }
2720         }
2721
2722         /**
2723          * Convert a string to uppercase
2724          *
2725          * @param string $str
2726          * @param bool $first
2727          *
2728          * @return string
2729          */
2730         public function uc( $str, $first = false ) {
2731                 if ( $first ) {
2732                         if ( $this->isMultibyte( $str ) ) {
2733                                 return $this->mbUpperChar( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2734                         } else {
2735                                 return ucfirst( $str );
2736                         }
2737                 } else {
2738                         return $this->isMultibyte( $str ) ? mb_strtoupper( $str ) : strtoupper( $str );
2739                 }
2740         }
2741
2742         /**
2743          * Convert character to uppercase, allowing overrides of the default mb_upper
2744          * behaviour, which is buggy in many ways. Having a conversion table can be
2745          * useful during transitions between PHP versions where unicode changes happen.
2746          * This can make some resources unreachable on-wiki, see discussion at T219279.
2747          * Providing such a conversion table can allow to manage the transition period.
2748          *
2749          * @since 1.34
2750          *
2751          * @param string $char
2752          *
2753          * @return string
2754          */
2755         protected function mbUpperChar( $char ) {
2756                 global $wgOverrideUcfirstCharacters;
2757                 if ( array_key_exists( $char, $wgOverrideUcfirstCharacters ) ) {
2758                         return $wgOverrideUcfirstCharacters[$char];
2759                 } else {
2760                         return mb_strtoupper( $char );
2761                 }
2762         }
2763
2764         /**
2765          * @param string $str
2766          * @return mixed|string
2767          */
2768         function lcfirst( $str ) {
2769                 $o = ord( $str );
2770                 if ( !$o ) {
2771                         return strval( $str );
2772                 } elseif ( $o >= 128 ) {
2773                         return $this->lc( $str, true );
2774                 } elseif ( $o > 96 ) {
2775                         return $str;
2776                 } else {
2777                         $str[0] = strtolower( $str[0] );
2778                         return $str;
2779                 }
2780         }
2781
2782         /**
2783          * @param string $str
2784          * @param bool $first
2785          * @return mixed|string
2786          */
2787         function lc( $str, $first = false ) {
2788                 if ( $first ) {
2789                         if ( $this->isMultibyte( $str ) ) {
2790                                 return mb_strtolower( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2791                         } else {
2792                                 return strtolower( substr( $str, 0, 1 ) ) . substr( $str, 1 );
2793                         }
2794                 } else {
2795                         return $this->isMultibyte( $str ) ? mb_strtolower( $str ) : strtolower( $str );
2796                 }
2797         }
2798
2799         /**
2800          * @param string $str
2801          * @return bool
2802          */
2803         function isMultibyte( $str ) {
2804                 return strlen( $str ) !== mb_strlen( $str );
2805         }
2806
2807         /**
2808          * @param string $str
2809          * @return mixed|string
2810          */
2811         function ucwords( $str ) {
2812                 if ( $this->isMultibyte( $str ) ) {
2813                         $str = $this->lc( $str );
2814
2815                         // regexp to find first letter in each word (i.e. after each space)
2816                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)| ([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2817
2818                         // function to use to capitalize a single char
2819                         return preg_replace_callback(
2820                                 $replaceRegexp,
2821                                 [ $this, 'ucwordsCallbackMB' ],
2822                                 $str
2823                         );
2824                 } else {
2825                         return ucwords( strtolower( $str ) );
2826                 }
2827         }
2828
2829         /**
2830          * capitalize words at word breaks
2831          *
2832          * @param string $str
2833          * @return mixed
2834          */
2835         function ucwordbreaks( $str ) {
2836                 if ( $this->isMultibyte( $str ) ) {
2837                         $str = $this->lc( $str );
2838
2839                         // since \b doesn't work for UTF-8, we explicitely define word break chars
2840                         $breaks = "[ \-\(\)\}\{\.,\?!]";
2841
2842                         // find first letter after word break
2843                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)|" .
2844                                 "$breaks([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2845
2846                         return preg_replace_callback(
2847                                 $replaceRegexp,
2848                                 [ $this, 'ucwordbreaksCallbackMB' ],
2849                                 $str
2850                         );
2851                 } else {
2852                         return preg_replace_callback(
2853                                 '/\b([\w\x80-\xff]+)\b/',
2854                                 [ $this, 'ucwordbreaksCallbackAscii' ],
2855                                 $str
2856                         );
2857                 }
2858         }
2859
2860         /**
2861          * Return a case-folded representation of $s
2862          *
2863          * This is a representation such that caseFold($s1)==caseFold($s2) if $s1
2864          * and $s2 are the same except for the case of their characters. It is not
2865          * necessary for the value returned to make sense when displayed.
2866          *
2867          * Do *not* perform any other normalisation in this function. If a caller
2868          * uses this function when it should be using a more general normalisation
2869          * function, then fix the caller.
2870          *
2871          * @param string $s
2872          *
2873          * @return string
2874          */
2875         function caseFold( $s ) {
2876                 return $this->uc( $s );
2877         }
2878
2879         /**
2880          * TODO: $s is not always a string per T218883
2881          * @param string $s
2882          * @return string
2883          */
2884         function checkTitleEncoding( $s ) {
2885                 if ( is_array( $s ) ) {
2886                         throw new MWException( 'Given array to checkTitleEncoding.' );
2887                 }
2888                 if ( StringUtils::isUtf8( $s ) ) {
2889                         return $s;
2890                 }
2891
2892                 return $this->iconv( $this->fallback8bitEncoding(), 'utf-8', $s );
2893         }
2894
2895         /**
2896          * @return string
2897          */
2898         function fallback8bitEncoding() {
2899                 return self::$dataCache->getItem( $this->mCode, 'fallback8bitEncoding' );
2900         }
2901
2902         /**
2903          * Most writing systems use whitespace to break up words.
2904          * Some languages such as Chinese don't conventionally do this,
2905          * which requires special handling when breaking up words for
2906          * searching etc.
2907          *
2908          * @return bool
2909          */
2910         function hasWordBreaks() {
2911                 return true;
2912         }
2913
2914         /**
2915          * Some languages such as Chinese require word segmentation,
2916          * Specify such segmentation when overridden in derived class.
2917          *
2918          * @param string $string
2919          * @return string
2920          */
2921         function segmentByWord( $string ) {
2922                 return $string;
2923         }
2924
2925         /**
2926          * Some languages have special punctuation need to be normalized.
2927          * Make such changes here.
2928          *
2929          * @param string $string
2930          * @return string
2931          */
2932         function normalizeForSearch( $string ) {
2933                 return self::convertDoubleWidth( $string );
2934         }
2935
2936         /**
2937          * convert double-width roman characters to single-width.
2938          * range: ff00-ff5f ~= 0020-007f
2939          *
2940          * @param string $string
2941          *
2942          * @return string
2943          */
2944         protected static function convertDoubleWidth( $string ) {
2945                 static $full = null;
2946                 static $half = null;
2947
2948                 if ( $full === null ) {
2949                         $fullWidth = "０１２３４５６７８９ＡＢＣＤＥＦＧＨＩＪＫＬＭＮＯＰＱＲＳＴＵＶＷＸＹＺａｂｃｄｅｆｇｈｉｊｋｌｍｎｏｐｑｒｓｔｕｖｗｘｙｚ";
2950                         $halfWidth = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz";
2951                         $full = str_split( $fullWidth, 3 );
2952                         $half = str_split( $halfWidth );
2953                 }
2954
2955                 $string = str_replace( $full, $half, $string );
2956                 return $string;
2957         }
2958
2959         /**
2960          * @param string $string
2961          * @param string $pattern
2962          * @return string
2963          */
2964         protected static function insertSpace( $string, $pattern ) {
2965                 $string = preg_replace( $pattern, " $1 ", $string );
2966                 $string = preg_replace( '/ +/', ' ', $string );
2967                 return $string;
2968         }
2969
2970         /**
2971          * @param string[] $termsArray
2972          * @return string[]
2973          */
2974         function convertForSearchResult( $termsArray ) {
2975                 # some languages, e.g. Chinese, need to do a conversion
2976                 # in order for search results to be displayed correctly
2977                 return $termsArray;
2978         }
2979
2980         /**
2981          * Get the first character of a string.
2982          *
2983          * @param string $s
2984          * @return string
2985          */
2986         function firstChar( $s ) {
2987                 $matches = [];
2988                 preg_match(
2989                         '/^([\x00-\x7f]|[\xc0-\xdf][\x80-\xbf]|' .
2990                                 '[\xe0-\xef][\x80-\xbf]{2}|[\xf0-\xf7][\x80-\xbf]{3})/',
2991                         $s,
2992                         $matches
2993                 );
2994
2995                 if ( isset( $matches[1] ) ) {
2996                         if ( strlen( $matches[1] ) != 3 ) {
2997                                 return $matches[1];
2998                         }
2999
3000                         // Break down Hangul syllables to grab the first jamo
3001                         $code = UtfNormal\Utils::utf8ToCodepoint( $matches[1] );
3002                         if ( $code < 0xac00 || 0xd7a4 <= $code ) {
3003                                 return $matches[1];
3004                         } elseif ( $code < 0xb098 ) {
3005                                 return "\u{3131}";
3006                         } elseif ( $code < 0xb2e4 ) {
3007                                 return "\u{3134}";
3008                         } elseif ( $code < 0xb77c ) {
3009                                 return "\u{3137}";
3010                         } elseif ( $code < 0xb9c8 ) {
3011                                 return "\u{3139}";
3012                         } elseif ( $code < 0xbc14 ) {
3013                                 return "\u{3141}";
3014                         } elseif ( $code < 0xc0ac ) {
3015                                 return "\u{3142}";
3016                         } elseif ( $code < 0xc544 ) {
3017                                 return "\u{3145}";
3018                         } elseif ( $code < 0xc790 ) {
3019                                 return "\u{3147}";
3020                         } elseif ( $code < 0xcc28 ) {
3021                                 return "\u{3148}";
3022                         } elseif ( $code < 0xce74 ) {
3023                                 return "\u{314A}";
3024                         } elseif ( $code < 0xd0c0 ) {
3025                                 return "\u{314B}";
3026                         } elseif ( $code < 0xd30c ) {
3027                                 return "\u{314C}";
3028                         } elseif ( $code < 0xd558 ) {
3029                                 return "\u{314D}";
3030                         } else {
3031                                 return "\u{314E}";
3032                         }
3033                 } else {
3034                         return '';
3035                 }
3036         }
3037
3038         /**
3039          * Convert a UTF-8 string to normal form C. In Malayalam and Arabic, this
3040          * also cleans up certain backwards-compatible sequences, converting them
3041          * to the modern Unicode equivalent.
3042          *
3043          * This is language-specific for performance reasons only.
3044          *
3045          * @param string $s
3046          *
3047          * @return string
3048          */
3049         public function normalize( $s ) {
3050                 global $wgAllUnicodeFixes;
3051                 $s = UtfNormal\Validator::cleanUp( $s );
3052                 if ( $wgAllUnicodeFixes ) {
3053                         $s = $this->transformUsingPairFile( 'normalize-ar.php', $s );
3054                         $s = $this->transformUsingPairFile( 'normalize-ml.php', $s );
3055                 }
3056
3057                 return $s;
3058         }
3059
3060         /**
3061          * Transform a string using serialized data stored in the given file (which
3062          * must be in the serialized subdirectory of $IP). The file contains pairs
3063          * mapping source characters to destination characters.
3064          *
3065          * The data is cached in process memory. This will go faster if you have the
3066          * FastStringSearch extension.
3067          *
3068          * @param string $file
3069          * @param string $string
3070          *
3071          * @throws MWException
3072          * @return string
3073          */
3074         protected function transformUsingPairFile( $file, $string ) {
3075                 if ( !isset( $this->transformData[$file] ) ) {
3076                         global $IP;
3077                         $data = require "$IP/languages/data/{$file}";
3078                         $this->transformData[$file] = new ReplacementArray( $data );
3079                 }
3080                 return $this->transformData[$file]->replace( $string );
3081         }
3082
3083         /**
3084          * For right-to-left language support
3085          *
3086          * @return bool
3087          */
3088         function isRTL() {
3089                 return self::$dataCache->getItem( $this->mCode, 'rtl' );
3090         }
3091
3092         /**
3093          * Return the correct HTML 'dir' attribute value for this language.
3094          * @return string
3095          */
3096         function getDir() {
3097                 return $this->isRTL() ? 'rtl' : 'ltr';
3098         }
3099
3100         /**
3101          * Return 'left' or 'right' as appropriate alignment for line-start
3102          * for this language's text direction.
3103          *
3104          * Should be equivalent to CSS3 'start' text-align value....
3105          *
3106          * @return string
3107          */
3108         function alignStart() {
3109                 return $this->isRTL() ? 'right' : 'left';
3110         }
3111
3112         /**
3113          * Return 'right' or 'left' as appropriate alignment for line-end
3114          * for this language's text direction.
3115          *
3116          * Should be equivalent to CSS3 'end' text-align value....
3117          *
3118          * @return string
3119          */
3120         function alignEnd() {
3121                 return $this->isRTL() ? 'left' : 'right';
3122         }
3123
3124         /**
3125          * A hidden direction mark (LRM or RLM), depending on the language direction.
3126          * Unlike getDirMark(), this function returns the character as an HTML entity.
3127          * This function should be used when the output is guaranteed to be HTML,
3128          * because it makes the output HTML source code more readable. When
3129          * the output is plain text or can be escaped, getDirMark() should be used.
3130          *
3131          * @param bool $opposite Get the direction mark opposite to your language
3132          * @return string
3133          * @since 1.20
3134          */
3135         function getDirMarkEntity( $opposite = false ) {
3136                 if ( $opposite ) {
3137                         return $this->isRTL() ? '&lrm;' : '&rlm;';
3138                 }
3139                 return $this->isRTL() ? '&rlm;' : '&lrm;';
3140         }
3141
3142         /**
3143          * A hidden direction mark (LRM or RLM), depending on the language direction.
3144          * This function produces them as invisible Unicode characters and
3145          * the output may be hard to read and debug, so it should only be used
3146          * when the output is plain text or can be escaped. When the output is
3147          * HTML, use getDirMarkEntity() instead.
3148          *
3149          * @param bool $opposite Get the direction mark opposite to your language
3150          * @return string
3151          */
3152         function getDirMark( $opposite = false ) {
3153                 $lrm = "\u{200E}"; # LEFT-TO-RIGHT MARK, commonly abbreviated LRM
3154                 $rlm = "\u{200F}"; # RIGHT-TO-LEFT MARK, commonly abbreviated RLM
3155                 if ( $opposite ) {
3156                         return $this->isRTL() ? $lrm : $rlm;
3157                 }
3158                 return $this->isRTL() ? $rlm : $lrm;
3159         }
3160
3161         /**
3162          * @return array
3163          */
3164         function capitalizeAllNouns() {
3165                 return self::$dataCache->getItem( $this->mCode, 'capitalizeAllNouns' );
3166         }
3167
3168         /**
3169          * An arrow, depending on the language direction.
3170          *
3171          * @param string $direction The direction of the arrow: forwards (default),
3172          *   backwards, left, right, up, down.
3173          * @return string
3174          */
3175         function getArrow( $direction = 'forwards' ) {
3176                 switch ( $direction ) {
3177                         case 'forwards':
3178                                 return $this->isRTL() ? '←' : '→';
3179                         case 'backwards':
3180                                 return $this->isRTL() ? '→' : '←';
3181                         case 'left':
3182                                 return '←';
3183                         case 'right':
3184                                 return '→';
3185                         case 'up':
3186                                 return '↑';
3187                         case 'down':
3188                                 return '↓';
3189                 }
3190         }
3191
3192         /**
3193          * To allow "foo[[bar]]" to extend the link over the whole word "foobar"
3194          *
3195          * @return bool
3196          */
3197         function linkPrefixExtension() {
3198                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixExtension' );
3199         }
3200
3201         /**
3202          * Get all magic words from cache.
3203          * @return array
3204          */
3205         function getMagicWords() {
3206                 return self::$dataCache->getItem( $this->mCode, 'magicWords' );
3207         }
3208
3209         /**
3210          * Fill a MagicWord object with data from here
3211          *
3212          * @param MagicWord $mw
3213          */
3214         function getMagic( $mw ) {
3215                 $rawEntry = $this->mMagicExtensions[$mw->mId] ??
3216                         self::$dataCache->getSubitem( $this->mCode, 'magicWords', $mw->mId );
3217
3218                 if ( !is_array( $rawEntry ) ) {
3219                         wfWarn( "\"$rawEntry\" is not a valid magic word for \"$mw->mId\"" );
3220                 } else {
3221                         $mw->mCaseSensitive = $rawEntry[0];
3222                         $mw->mSynonyms = array_slice( $rawEntry, 1 );
3223                 }
3224         }
3225
3226         /**
3227          * Add magic words to the extension array
3228          *
3229          * @param array $newWords
3230          */
3231         function addMagicWordsByLang( $newWords ) {
3232                 $fallbackChain = $this->getFallbackLanguages();
3233                 $fallbackChain = array_reverse( $fallbackChain );
3234                 foreach ( $fallbackChain as $code ) {
3235                         if ( isset( $newWords[$code] ) ) {
3236                                 $this->mMagicExtensions = $newWords[$code] + $this->mMagicExtensions;
3237                         }
3238                 }
3239         }
3240
3241         /**
3242          * Get special page names, as an associative array
3243          *   canonical name => array of valid names, including aliases
3244          * @return array
3245          */
3246         function getSpecialPageAliases() {
3247                 // Cache aliases because it may be slow to load them
3248                 if ( is_null( $this->mExtendedSpecialPageAliases ) ) {
3249                         // Initialise array
3250                         $this->mExtendedSpecialPageAliases =
3251                                 self::$dataCache->getItem( $this->mCode, 'specialPageAliases' );
3252                 }
3253
3254                 return $this->mExtendedSpecialPageAliases;
3255         }
3256
3257         /**
3258          * Italic is unsuitable for some languages
3259          *
3260          * @param string $text The text to be emphasized.
3261          * @return string
3262          */
3263         function emphasize( $text ) {
3264                 return "<em>$text</em>";
3265         }
3266
3267         /**
3268          * Normally we output all numbers in plain en_US style, that is
3269          * 293,291.235 for twohundredninetythreethousand-twohundredninetyone
3270          * point twohundredthirtyfive. However this is not suitable for all
3271          * languages, some such as Bengali (bn) want ২,৯৩,২৯১.২৩৫ and others such as
3272          * Icelandic just want to use commas instead of dots, and dots instead
3273          * of commas like "293.291,235".
3274          *
3275          * An example of this function being called:
3276          * <code>
3277          * wfMessage( 'message' )->numParams( $num )->text()
3278          * </code>
3279          *
3280          * See $separatorTransformTable on MessageIs.php for
3281          * the , => . and . => , implementation.
3282          *
3283          * @todo check if it's viable to use localeconv() for the decimal separator thing.
3284          * @param int|float $number The string to be formatted, should be an integer
3285          *   or a floating point number.
3286          * @param bool $nocommafy Set to true for special numbers like dates
3287          * @return string
3288          */
3289         public function formatNum( $number, $nocommafy = false ) {
3290                 global $wgTranslateNumerals;
3291                 if ( !$nocommafy ) {
3292                         $number = $this->commafy( $number );
3293                         $s = $this->separatorTransformTable();
3294                         if ( $s ) {
3295                                 $number = strtr( $number, $s );
3296                         }
3297                 }
3298
3299                 if ( $wgTranslateNumerals ) {
3300                         $s = $this->digitTransformTable();
3301                         if ( $s ) {
3302                                 $number = strtr( $number, $s );
3303                         }
3304                 }
3305
3306                 return (string)$number;
3307         }
3308
3309         /**
3310          * Front-end for non-commafied formatNum
3311          *
3312          * @param int|float $number The string to be formatted, should be an integer
3313          *        or a floating point number.
3314          * @since 1.21
3315          * @return string
3316          */
3317         public function formatNumNoSeparators( $number ) {
3318                 return $this->formatNum( $number, true );
3319         }
3320
3321         /**
3322          * @param string $number
3323          * @return string
3324          */
3325         public function parseFormattedNumber( $number ) {
3326                 $s = $this->digitTransformTable();
3327                 if ( $s ) {
3328                         // eliminate empty array values such as ''. (T66347)
3329                         $s = array_filter( $s );
3330                         $number = strtr( $number, array_flip( $s ) );
3331                 }
3332
3333                 $s = $this->separatorTransformTable();
3334                 if ( $s ) {
3335                         // eliminate empty array values such as ''. (T66347)
3336                         $s = array_filter( $s );
3337                         $number = strtr( $number, array_flip( $s ) );
3338                 }
3339
3340                 $number = strtr( $number, [ ',' => '' ] );
3341                 return $number;
3342         }
3343
3344         /**
3345          * Adds commas to a given number
3346          * @since 1.19
3347          * @param mixed $number
3348          * @return string
3349          */
3350         function commafy( $number ) {
3351                 $digitGroupingPattern = $this->digitGroupingPattern();
3352                 $minimumGroupingDigits = $this->minimumGroupingDigits();
3353                 if ( $number === null ) {
3354                         return '';
3355                 }
3356
3357                 if ( !$digitGroupingPattern || $digitGroupingPattern === "###,###,###" ) {
3358                         // Default grouping is at thousands, use the same for ###,###,### pattern too.
3359                         // In some languages it's conventional not to insert a thousands separator
3360                         // in numbers that are four digits long (1000-9999).
3361                         if ( $minimumGroupingDigits ) {
3362                                 // Number of '#' characters after last comma in the grouping pattern.
3363                                 // The pattern is hardcoded here, but this would vary for different patterns.
3364                                 $primaryGroupingSize = 3;
3365                                 // Maximum length of a number to suppress digit grouping for.
3366                                 $maximumLength = $minimumGroupingDigits + $primaryGroupingSize - 1;
3367                                 if ( preg_match( '/^\-?\d{1,' . $maximumLength . '}(\.\d+)?$/', $number ) ) {
3368                                         return $number;
3369                                 }
3370                         }
3371                         return strrev( (string)preg_replace( '/(\d{3})(?=\d)(?!\d*\.)/', '$1,', strrev( $number ) ) );
3372                 } else {
3373                         // Ref: http://cldr.unicode.org/translation/number-patterns
3374                         $sign = "";
3375                         if ( intval( $number ) < 0 ) {
3376                                 // For negative numbers apply the algorithm like positive number and add sign.
3377                                 $sign = "-";
3378                                 $number = substr( $number, 1 );
3379                         }
3380                         $integerPart = [];
3381                         $decimalPart = [];
3382                         $numMatches = preg_match_all( "/(#+)/", $digitGroupingPattern, $matches );
3383                         preg_match( "/\d+/", $number, $integerPart );
3384                         preg_match( "/\.\d*/", $number, $decimalPart );
3385                         $groupedNumber = ( count( $decimalPart ) > 0 ) ? $decimalPart[0] : "";
3386                         if ( $groupedNumber === $number ) {
3387                                 // the string does not have any number part. Eg: .12345
3388                                 return $sign . $groupedNumber;
3389                         }
3390                         $start = $end = ( $integerPart ) ? strlen( $integerPart[0] ) : 0;
3391                         while ( $start > 0 ) {
3392                                 $match = $matches[0][$numMatches - 1];
3393                                 $matchLen = strlen( $match );
3394                                 $start = $end - $matchLen;
3395                                 if ( $start < 0 ) {
3396                                         $start = 0;
3397                                 }
3398                                 $groupedNumber = substr( $number, $start, $end - $start ) . $groupedNumber;
3399                                 $end = $start;
3400                                 if ( $numMatches > 1 ) {
3401                                         // use the last pattern for the rest of the number
3402                                         $numMatches--;
3403                                 }
3404                                 if ( $start > 0 ) {
3405                                         $groupedNumber = "," . $groupedNumber;
3406                                 }
3407                         }
3408                         return $sign . $groupedNumber;
3409                 }
3410         }
3411
3412         /**
3413          * @return string
3414          */
3415         function digitGroupingPattern() {
3416                 return self::$dataCache->getItem( $this->mCode, 'digitGroupingPattern' );
3417         }
3418
3419         /**
3420          * @return array
3421          */
3422         function digitTransformTable() {
3423                 return self::$dataCache->getItem( $this->mCode, 'digitTransformTable' );
3424         }
3425
3426         /**
3427          * @return array
3428          */
3429         function separatorTransformTable() {
3430                 return self::$dataCache->getItem( $this->mCode, 'separatorTransformTable' );
3431         }
3432
3433         /**
3434          * @return int|null
3435          */
3436         function minimumGroupingDigits() {
3437                 return self::$dataCache->getItem( $this->mCode, 'minimumGroupingDigits' );
3438         }
3439
3440         /**
3441          * Take a list of strings and build a locale-friendly comma-separated
3442          * list, using the local comma-separator message.
3443          * The last two strings are chained with an "and".
3444          *
3445          * @param string[] $list
3446          * @return string
3447          */
3448         public function listToText( array $list ) {
3449                 $itemCount = count( $list );
3450                 if ( $itemCount < 1 ) {
3451                         return '';
3452                 }
3453                 $text = array_pop( $list );
3454                 if ( $itemCount > 1 ) {
3455                         $and = $this->msg( 'and' )->escaped();
3456                         $space = $this->msg( 'word-separator' )->escaped();
3457                         $comma = '';
3458                         if ( $itemCount > 2 ) {
3459                                 $comma = $this->msg( 'comma-separator' )->escaped();
3460                         }
3461                         $text = implode( $comma, $list ) . $and . $space . $text;
3462                 }
3463                 return $text;
3464         }
3465
3466         /**
3467          * Take a list of strings and build a locale-friendly comma-separated
3468          * list, using the local comma-separator message.
3469          * @param string[] $list Array of strings to put in a comma list
3470          * @return string
3471          */
3472         function commaList( array $list ) {
3473                 return implode(
3474                         wfMessage( 'comma-separator' )->inLanguage( $this )->escaped(),
3475                         $list
3476                 );
3477         }
3478
3479         /**
3480          * Take a list of strings and build a locale-friendly semicolon-separated
3481          * list, using the local semicolon-separator message.
3482          * @param string[] $list Array of strings to put in a semicolon list
3483          * @return string
3484          */
3485         function semicolonList( array $list ) {
3486                 return implode(
3487                         wfMessage( 'semicolon-separator' )->inLanguage( $this )->escaped(),
3488                         $list
3489                 );
3490         }
3491
3492         /**
3493          * Same as commaList, but separate it with the pipe instead.
3494          * @param string[] $list Array of strings to put in a pipe list
3495          * @return string
3496          */
3497         function pipeList( array $list ) {
3498                 return implode(
3499                         wfMessage( 'pipe-separator' )->inLanguage( $this )->escaped(),
3500                         $list
3501                 );
3502         }
3503
3504         /**
3505          * Truncate a string to a specified length in bytes, appending an optional
3506          * string (e.g. for ellipsis)
3507          *
3508          * If $length is negative, the string will be truncated from the beginning
3509          *
3510          * @since 1.31
3511          *
3512          * @param string $string String to truncate
3513          * @param int $length Maximum length in bytes
3514          * @param string $ellipsis String to append to the end of truncated text
3515          * @param bool $adjustLength Subtract length of ellipsis from $length
3516          *
3517          * @return string
3518          */
3519         function truncateForDatabase( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3520                 return $this->truncateInternal(
3521                         $string, $length, $ellipsis, $adjustLength, 'strlen', 'substr'
3522                 );
3523         }
3524
3525         /**
3526          * Truncate a string to a specified number of characters, appending an optional
3527          * string (e.g. for ellipsis).
3528          *
3529          * This provides multibyte version of truncateForDatabase() method of this class,
3530          * suitable for truncation based on number of characters, instead of number of bytes.
3531          *
3532          * If $length is negative, the string will be truncated from the beginning.
3533          *
3534          * @since 1.31
3535          *
3536          * @param string $string String to truncate
3537          * @param int $length Maximum number of characters
3538          * @param string $ellipsis String to append to the end of truncated text
3539          * @param bool $adjustLength Subtract length of ellipsis from $length
3540          *
3541          * @return string
3542          */
3543         function truncateForVisual( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3544                 // Passing encoding to mb_strlen and mb_substr is optional.
3545                 // Encoding defaults to mb_internal_encoding(), which is set to UTF-8 in Setup.php, so
3546                 // explicit specification of encoding is skipped.
3547                 // Note: Both multibyte methods are callables invoked in truncateInternal.
3548                 return $this->truncateInternal(
3549                         $string, $length, $ellipsis, $adjustLength, 'mb_strlen', 'mb_substr'
3550                 );
3551         }
3552
3553         /**
3554          * Internal method used for truncation. This method abstracts text truncation into
3555          * one common method, allowing users to provide length measurement function and
3556          * function for finding substring.
3557          *
3558          * For usages, see truncateForDatabase and truncateForVisual.
3559          *
3560          * @param string $string String to truncate
3561          * @param int $length Maximum length of final text
3562          * @param string $ellipsis String to append to the end of truncated text
3563          * @param bool $adjustLength Subtract length of ellipsis from $length
3564          * @param callable $measureLength Callable function used for determining the length of text
3565          * @param callable $getSubstring Callable function used for getting the substrings
3566          *
3567          * @return string
3568          */
3569         private function truncateInternal(
3570                 $string, $length, $ellipsis, $adjustLength, callable $measureLength, callable $getSubstring
3571         ) {
3572                 # Check if there is no need to truncate
3573                 if ( $measureLength( $string ) <= abs( $length ) ) {
3574                         return $string; // no need to truncate
3575                 }
3576
3577                 # Use the localized ellipsis character
3578                 if ( $ellipsis == '...' ) {
3579                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3580                 }
3581                 if ( $length == 0 ) {
3582                         return $ellipsis; // convention
3583                 }
3584
3585                 $stringOriginal = $string;
3586                 # If ellipsis length is >= $length then we can't apply $adjustLength
3587                 if ( $adjustLength && $measureLength( $ellipsis ) >= abs( $length ) ) {
3588                         $string = $ellipsis; // this can be slightly unexpected
3589                 # Otherwise, truncate and add ellipsis...
3590                 } else {
3591                         $ellipsisLength = $adjustLength ? $measureLength( $ellipsis ) : 0;
3592                         if ( $length > 0 ) {
3593                                 $length -= $ellipsisLength;
3594                                 $string = $getSubstring( $string, 0, $length ); // xyz...
3595                                 $string = $this->removeBadCharLast( $string );
3596                                 $string = rtrim( $string ) . $ellipsis;
3597                         } else {
3598                                 $length += $ellipsisLength;
3599                                 $string = $getSubstring( $string, $length ); // ...xyz
3600                                 $string = $this->removeBadCharFirst( $string );
3601                                 $string = $ellipsis . ltrim( $string );
3602                         }
3603                 }
3604
3605                 # Do not truncate if the ellipsis makes the string longer/equal (T24181).
3606                 # This check is *not* redundant if $adjustLength, due to the single case where
3607                 # LEN($ellipsis) > ABS($limit arg); $stringOriginal could be shorter than $string.
3608                 if ( $measureLength( $string ) < $measureLength( $stringOriginal ) ) {
3609                         return $string;
3610                 } else {
3611                         return $stringOriginal;
3612                 }
3613         }
3614
3615         /**
3616          * Remove bytes that represent an incomplete Unicode character
3617          * at the end of string (e.g. bytes of the char are missing)
3618          *
3619          * @param string $string
3620          * @return string
3621          */
3622         protected function removeBadCharLast( $string ) {
3623                 if ( $string != '' ) {
3624                         $char = ord( $string[strlen( $string ) - 1] );
3625                         $m = [];
3626                         if ( $char >= 0xc0 ) {
3627                                 # We got the first byte only of a multibyte char; remove it.
3628                                 $string = substr( $string, 0, -1 );
3629                         } elseif ( $char >= 0x80 &&
3630                                 // Use the /s modifier (PCRE_DOTALL) so (.*) also matches newlines
3631                                 preg_match( '/^(.*)(?:[\xe0-\xef][\x80-\xbf]|' .
3632                                         '[\xf0-\xf7][\x80-\xbf]{1,2})$/s', $string, $m )
3633                         ) {
3634                                 # We chopped in the middle of a character; remove it
3635                                 $string = $m[1];
3636                         }
3637                 }
3638                 return $string;
3639         }
3640
3641         /**
3642          * Remove bytes that represent an incomplete Unicode character
3643          * at the start of string (e.g. bytes of the char are missing)
3644          *
3645          * @param string $string
3646          * @return string
3647          */
3648         protected function removeBadCharFirst( $string ) {
3649                 if ( $string != '' ) {
3650                         $char = ord( $string[0] );
3651                         if ( $char >= 0x80 && $char < 0xc0 ) {
3652                                 # We chopped in the middle of a character; remove the whole thing
3653                                 $string = preg_replace( '/^[\x80-\xbf]+/', '', $string );
3654                         }
3655                 }
3656                 return $string;
3657         }
3658
3659         /**
3660          * Truncate a string of valid HTML to a specified length in bytes,
3661          * appending an optional string (e.g. for ellipses), and return valid HTML
3662          *
3663          * This is only intended for styled/linked text, such as HTML with
3664          * tags like <span> and <a>, were the tags are self-contained (valid HTML).
3665          * Also, this will not detect things like "display:none" CSS.
3666          *
3667          * Note: since 1.18 you do not need to leave extra room in $length for ellipses.
3668          *
3669          * @param string $text HTML string to truncate
3670          * @param int $length (zero/positive) Maximum length (including ellipses)
3671          * @param string $ellipsis String to append to the truncated text
3672          * @return string
3673          */
3674         function truncateHtml( $text, $length, $ellipsis = '...' ) {
3675                 # Use the localized ellipsis character
3676                 if ( $ellipsis == '...' ) {
3677                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3678                 }
3679                 # Check if there is clearly no need to truncate
3680                 if ( $length <= 0 ) {
3681                         return $ellipsis; // no text shown, nothing to format (convention)
3682                 } elseif ( strlen( $text ) <= $length ) {
3683                         return $text; // string short enough even *with* HTML (short-circuit)
3684                 }
3685
3686                 $dispLen = 0; // innerHTML legth so far
3687                 $testingEllipsis = false; // checking if ellipses will make string longer/equal?
3688                 $tagType = 0; // 0-open, 1-close
3689                 $bracketState = 0; // 1-tag start, 2-tag name, 0-neither
3690                 $entityState = 0; // 0-not entity, 1-entity
3691                 $tag = $ret = ''; // accumulated tag name, accumulated result string
3692                 $openTags = []; // open tag stack
3693                 $maybeState = null; // possible truncation state
3694
3695                 $textLen = strlen( $text );
3696                 $neLength = max( 0, $length - strlen( $ellipsis ) ); // non-ellipsis len if truncated
3697                 for ( $pos = 0; true; ++$pos ) {
3698                         # Consider truncation once the display length has reached the maximim.
3699                         # We check if $dispLen > 0 to grab tags for the $neLength = 0 case.
3700                         # Check that we're not in the middle of a bracket/entity...
3701                         if ( $dispLen && $dispLen >= $neLength && $bracketState == 0 && !$entityState ) {
3702                                 if ( !$testingEllipsis ) {
3703                                         $testingEllipsis = true;
3704                                         # Save where we are; we will truncate here unless there turn out to
3705                                         # be so few remaining characters that truncation is not necessary.
3706                                         if ( !$maybeState ) { // already saved? ($neLength = 0 case)
3707                                                 $maybeState = [ $ret, $openTags ]; // save state
3708                                         }
3709                                 } elseif ( $dispLen > $length && $dispLen > strlen( $ellipsis ) ) {
3710                                         # String in fact does need truncation, the truncation point was OK.
3711                                         // @phan-suppress-next-line PhanTypeInvalidExpressionArrayDestructuring
3712                                         list( $ret, $openTags ) = $maybeState; // reload state
3713                                         $ret = $this->removeBadCharLast( $ret ); // multi-byte char fix
3714                                         $ret .= $ellipsis; // add ellipsis
3715                                         break;
3716                                 }
3717                         }
3718                         if ( $pos >= $textLen ) {
3719                                 break; // extra iteration just for above checks
3720                         }
3721
3722                         # Read the next char...
3723                         $ch = $text[$pos];
3724                         $lastCh = $pos ? $text[$pos - 1] : '';
3725                         $ret .= $ch; // add to result string
3726                         if ( $ch == '<' ) {
3727                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags ); // for bad HTML
3728                                 $entityState = 0; // for bad HTML
3729                                 $bracketState = 1; // tag started (checking for backslash)
3730                         } elseif ( $ch == '>' ) {
3731                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags );
3732                                 $entityState = 0; // for bad HTML
3733                                 $bracketState = 0; // out of brackets
3734                         } elseif ( $bracketState == 1 ) {
3735                                 if ( $ch == '/' ) {
3736                                         $tagType = 1; // close tag (e.g. "</span>")
3737                                 } else {
3738                                         $tagType = 0; // open tag (e.g. "<span>")
3739                                         $tag .= $ch;
3740                                 }
3741                                 $bracketState = 2; // building tag name
3742                         } elseif ( $bracketState == 2 ) {
3743                                 if ( $ch != ' ' ) {
3744                                         $tag .= $ch;
3745                                 } else {
3746                                         // Name found (e.g. "<a href=..."), add on tag attributes...
3747                                         $pos += $this->truncate_skip( $ret, $text, "<>", $pos + 1 );
3748                                 }
3749                         } elseif ( $bracketState == 0 ) {
3750                                 if ( $entityState ) {
3751                                         if ( $ch == ';' ) {
3752                                                 $entityState = 0;
3753                                                 $dispLen++; // entity is one displayed char
3754                                         }
3755                                 } else {
3756                                         if ( $neLength == 0 && !$maybeState ) {
3757                                                 // Save state without $ch. We want to *hit* the first
3758                                                 // display char (to get tags) but not *use* it if truncating.
3759                                                 $maybeState = [ substr( $ret, 0, -1 ), $openTags ];
3760                                         }
3761                                         if ( $ch == '&' ) {
3762                                                 $entityState = 1; // entity found, (e.g. "&#160;")
3763                                         } else {
3764                                                 $dispLen++; // this char is displayed
3765                                                 // Add the next $max display text chars after this in one swoop...
3766                                                 $max = ( $testingEllipsis ? $length : $neLength ) - $dispLen;
3767                                                 $skipped = $this->truncate_skip( $ret, $text, "<>&", $pos + 1, $max );
3768                                                 $dispLen += $skipped;
3769                                                 $pos += $skipped;
3770                                         }
3771                                 }
3772                         }
3773                 }
3774                 // Close the last tag if left unclosed by bad HTML
3775                 $this->truncate_endBracket( $tag, $text[$textLen - 1], $tagType, $openTags );
3776                 while ( count( $openTags ) > 0 ) {
3777                         $ret .= '</' . array_pop( $openTags ) . '>'; // close open tags
3778                 }
3779                 return $ret;
3780         }
3781
3782         /**
3783          * truncateHtml() helper function
3784          * like strcspn() but adds the skipped chars to $ret
3785          *
3786          * @param string $ret
3787          * @param string $text
3788          * @param string $search
3789          * @param int $start
3790          * @param null|int $len
3791          * @return int
3792          */
3793         private function truncate_skip( &$ret, $text, $search, $start, $len = null ) {
3794                 if ( $len === null ) {
3795                         $len = -1; // -1 means "no limit" for strcspn
3796                 } elseif ( $len < 0 ) {
3797                         $len = 0; // sanity
3798                 }
3799                 $skipCount = 0;
3800                 if ( $start < strlen( $text ) ) {
3801                         $skipCount = strcspn( $text, $search, $start, $len );
3802                         $ret .= substr( $text, $start, $skipCount );
3803                 }
3804                 return $skipCount;
3805         }
3806
3807         /**
3808          * truncateHtml() helper function
3809          * (a) push or pop $tag from $openTags as needed
3810          * (b) clear $tag value
3811          * @param string &$tag Current HTML tag name we are looking at
3812          * @param int $tagType (0-open tag, 1-close tag)
3813          * @param string $lastCh Character before the '>' that ended this tag
3814          * @param array &$openTags Open tag stack (not accounting for $tag)
3815          */
3816         private function truncate_endBracket( &$tag, $tagType, $lastCh, &$openTags ) {
3817                 $tag = ltrim( $tag );
3818                 if ( $tag != '' ) {
3819                         if ( $tagType == 0 && $lastCh != '/' ) {
3820                                 $openTags[] = $tag; // tag opened (didn't close itself)
3821                         } elseif ( $tagType == 1 ) {
3822                                 if ( $openTags && $tag == $openTags[count( $openTags ) - 1] ) {
3823                                         array_pop( $openTags ); // tag closed
3824                                 }
3825                         }
3826                         $tag = '';
3827                 }
3828         }
3829
3830         /**
3831          * Grammatical transformations, needed for inflected languages
3832          * Invoked by putting {{grammar:case|word}} in a message
3833          *
3834          * @param string $word
3835          * @param string $case
3836          * @return string
3837          */
3838         function convertGrammar( $word, $case ) {
3839                 global $wgGrammarForms;
3840                 if ( isset( $wgGrammarForms[$this->getCode()][$case][$word] ) ) {
3841                         return $wgGrammarForms[$this->getCode()][$case][$word];
3842                 }
3843
3844                 $grammarTransformations = $this->getGrammarTransformations();
3845
3846                 if ( isset( $grammarTransformations[$case] ) ) {
3847                         $forms = $grammarTransformations[$case];
3848
3849                         // Some names of grammar rules are aliases for other rules.
3850                         // In such cases the value is a string rather than object,
3851                         // so load the actual rules.
3852                         if ( is_string( $forms ) ) {
3853                                 $forms = $grammarTransformations[$forms];
3854                         }
3855
3856                         foreach ( array_values( $forms ) as $rule ) {
3857                                 $form = $rule[0];
3858
3859                                 if ( $form === '@metadata' ) {
3860                                         continue;
3861                                 }
3862
3863                                 $replacement = $rule[1];
3864
3865                                 $regex = '/' . addcslashes( $form, '/' ) . '/u';
3866                                 $patternMatches = preg_match( $regex, $word );
3867
3868                                 if ( $patternMatches === false ) {
3869                                         wfLogWarning(
3870                                                 'An error occurred while processing grammar. ' .
3871                                                 "Word: '$word'. Regex: /$form/."
3872                                         );
3873                                 } elseif ( $patternMatches === 1 ) {
3874                                         $word = preg_replace( $regex, $replacement, $word );
3875
3876                                         break;
3877                                 }
3878                         }
3879                 }
3880
3881                 return $word;
3882         }
3883
3884         /**
3885          * Get the grammar forms for the content language
3886          * @return array Array of grammar forms
3887          * @since 1.20
3888          */
3889         function getGrammarForms() {
3890                 global $wgGrammarForms;
3891                 if ( isset( $wgGrammarForms[$this->getCode()] )
3892                         && is_array( $wgGrammarForms[$this->getCode()] )
3893                 ) {
3894                         return $wgGrammarForms[$this->getCode()];
3895                 }
3896
3897                 return [];
3898         }
3899
3900         /**
3901          * Get the grammar transformations data for the language.
3902          * Used like grammar forms, with {{GRAMMAR}} and cases,
3903          * but uses pairs of regexes and replacements instead of code.
3904          *
3905          * @return array[] Array of grammar transformations.
3906          * @throws MWException
3907          * @since 1.28
3908          */
3909         public function getGrammarTransformations() {
3910                 $languageCode = $this->getCode();
3911
3912                 if ( self::$grammarTransformations === null ) {
3913                         self::$grammarTransformations = new MapCacheLRU( 10 );
3914                 }
3915
3916                 if ( self::$grammarTransformations->has( $languageCode ) ) {
3917                         return self::$grammarTransformations->get( $languageCode );
3918                 }
3919
3920                 $data = [];
3921
3922                 $grammarDataFile = __DIR__ . "/data/grammarTransformations/$languageCode.json";
3923                 if ( is_readable( $grammarDataFile ) ) {
3924                         $data = FormatJson::decode(
3925                                 file_get_contents( $grammarDataFile ),
3926                                 true
3927                         );
3928
3929                         if ( $data === null ) {
3930                                 throw new MWException( "Invalid grammar data for \"$languageCode\"." );
3931                         }
3932
3933                         self::$grammarTransformations->set( $languageCode, $data );
3934                 }
3935
3936                 return $data;
3937         }
3938
3939         /**
3940          * Provides an alternative text depending on specified gender.
3941          * Usage {{gender:username|masculine|feminine|unknown}}.
3942          * username is optional, in which case the gender of current user is used,
3943          * but only in (some) interface messages; otherwise default gender is used.
3944          *
3945          * If no forms are given, an empty string is returned. If only one form is
3946          * given, it will be returned unconditionally. These details are implied by
3947          * the caller and cannot be overridden in subclasses.
3948          *
3949          * If three forms are given, the default is to use the third (unknown) form.
3950          * If fewer than three forms are given, the default is to use the first (masculine) form.
3951          * These details can be overridden in subclasses.
3952          *
3953          * @param string $gender
3954          * @param array $forms
3955          *
3956          * @return string
3957          */
3958         function gender( $gender, $forms ) {
3959                 if ( !count( $forms ) ) {
3960                         return '';
3961                 }
3962                 $forms = $this->preConvertPlural( $forms, 2 );
3963                 if ( $gender === 'male' ) {
3964                         return $forms[0];
3965                 }
3966                 if ( $gender === 'female' ) {
3967                         return $forms[1];
3968                 }
3969                 return $forms[2] ?? $forms[0];
3970         }
3971
3972         /**
3973          * Plural form transformations, needed for some languages.
3974          * For example, there are 3 form of plural in Russian and Polish,
3975          * depending on "count mod 10". See [[w:Plural]]
3976          * For English it is pretty simple.
3977          *
3978          * Invoked by putting {{plural:count|wordform1|wordform2}}
3979          * or {{plural:count|wordform1|wordform2|wordform3}}
3980          *
3981          * Example: {{plural:{{NUMBEROFARTICLES}}|article|articles}}
3982          *
3983          * @param int $count Non-localized number
3984          * @param array $forms Different plural forms
3985          * @return string Correct form of plural for $count in this language
3986          */
3987         function convertPlural( $count, $forms ) {
3988                 // Handle explicit n=pluralform cases
3989                 $forms = $this->handleExplicitPluralForms( $count, $forms );
3990                 if ( is_string( $forms ) ) {
3991                         return $forms;
3992                 }
3993                 if ( !count( $forms ) ) {
3994                         return '';
3995                 }
3996
3997                 $pluralForm = $this->getPluralRuleIndexNumber( $count );
3998                 $pluralForm = min( $pluralForm, count( $forms ) - 1 );
3999                 return $forms[$pluralForm];
4000         }
4001
4002         /**
4003          * Handles explicit plural forms for Language::convertPlural()
4004          *
4005          * In {{PLURAL:$1|0=nothing|one|many}}, 0=nothing will be returned if $1 equals zero.
4006          * If an explicitly defined plural form matches the $count, then
4007          * string value returned, otherwise array returned for further consideration
4008          * by CLDR rules or overridden convertPlural().
4009          *
4010          * @since 1.23
4011          *
4012          * @param int $count Non-localized number
4013          * @param array $forms Different plural forms
4014          *
4015          * @return array|string
4016          */
4017         protected function handleExplicitPluralForms( $count, array $forms ) {
4018                 foreach ( $forms as $index => $form ) {
4019                         if ( preg_match( '/\d+=/i', $form ) ) {
4020                                 $pos = strpos( $form, '=' );
4021                                 if ( substr( $form, 0, $pos ) === (string)$count ) {
4022                                         return substr( $form, $pos + 1 );
4023                                 }
4024                                 unset( $forms[$index] );
4025                         }
4026                 }
4027                 return array_values( $forms );
4028         }
4029
4030         /**
4031          * Checks that convertPlural was given an array and pads it to requested
4032          * amount of forms by copying the last one.
4033          *
4034          * @param array $forms
4035          * @param int $count How many forms should there be at least
4036          * @return array Padded array of forms
4037          */
4038         protected function preConvertPlural( /* Array */ $forms, $count ) {
4039                 return array_pad( $forms, $count, end( $forms ) );
4040         }
4041
4042         /**
4043          * Wraps argument with unicode control characters for directionality safety
4044          *
4045          * This solves the problem where directionality-neutral characters at the edge of
4046          * the argument string get interpreted with the wrong directionality from the
4047          * enclosing context, giving renderings that look corrupted like "(Ben_(WMF".
4048          *
4049          * The wrapping is LRE...PDF or RLE...PDF, depending on the detected
4050          * directionality of the argument string, using the BIDI algorithm's own "First
4051          * strong directional codepoint" rule. Essentially, this works round the fact that
4052          * there is no embedding equivalent of U+2068 FSI (isolation with heuristic
4053          * direction inference). The latter is cleaner but still not widely supported.
4054          *
4055          * @param string $text Text to wrap
4056          * @return string Text, wrapped in LRE...PDF or RLE...PDF or nothing
4057          */
4058         public function embedBidi( $text = '' ) {
4059                 $dir = self::strongDirFromContent( $text );
4060                 if ( $dir === 'ltr' ) {
4061                         // Wrap in LEFT-TO-RIGHT EMBEDDING ... POP DIRECTIONAL FORMATTING
4062                         return self::$lre . $text . self::$pdf;
4063                 }
4064                 if ( $dir === 'rtl' ) {
4065                         // Wrap in RIGHT-TO-LEFT EMBEDDING ... POP DIRECTIONAL FORMATTING
4066                         return self::$rle . $text . self::$pdf;
4067                 }
4068                 // No strong directionality: do not wrap
4069                 return $text;
4070         }
4071
4072         /**
4073          * @todo Maybe translate block durations.  Note that this function is somewhat misnamed: it
4074          * deals with translating the *duration* ("1 week", "4 days", etc), not the expiry time
4075          * (which is an absolute timestamp). Please note: do NOT add this blindly, as it is used
4076          * on old expiry lengths recorded in log entries. You'd need to provide the start date to
4077          * match up with it.
4078          *
4079          * @param string $str The validated block duration in English
4080          * @param User|null $user User object to use timezone from or null for $wgUser
4081          * @param int $now Current timestamp, for formatting relative block durations
4082          * @return string Somehow translated block duration
4083          * @see LanguageFi.php for example implementation
4084          */
4085         function translateBlockExpiry( $str, User $user = null, $now = 0 ) {
4086                 $duration = SpecialBlock::getSuggestedDurations( $this );
4087                 foreach ( $duration as $show => $value ) {
4088                         if ( strcmp( $str, $value ) == 0 ) {
4089                                 return htmlspecialchars( trim( $show ) );
4090                         }
4091                 }
4092
4093                 if ( wfIsInfinity( $str ) ) {
4094                         foreach ( $duration as $show => $value ) {
4095                                 if ( wfIsInfinity( $value ) ) {
4096                                         return htmlspecialchars( trim( $show ) );
4097                                 }
4098                         }
4099                 }
4100
4101                 // If all else fails, return a standard duration or timestamp description.
4102                 $time = strtotime( $str, $now );
4103                 if ( $time === false ) { // Unknown format. Return it as-is in case.
4104                         return $str;
4105                 } elseif ( $time !== strtotime( $str, $now + 1 ) ) { // It's a relative timestamp.
4106                         // The result differs based on current time, so the difference
4107                         // is a fixed duration length.
4108                         return $this->formatDuration( $time - $now );
4109                 } else { // It's an absolute timestamp.
4110                         if ( $time === 0 ) {
4111                                 // wfTimestamp() handles 0 as current time instead of epoch.
4112                                 $time = '19700101000000';
4113                         }
4114                         if ( $user ) {
4115                                 return $this->userTimeAndDate( $time, $user );
4116                         }
4117                         return $this->timeanddate( $time );
4118                 }
4119         }
4120
4121         /**
4122          * languages like Chinese need to be segmented in order for the diff
4123          * to be of any use
4124          *
4125          * @param string $text
4126          * @return string
4127          */
4128         public function segmentForDiff( $text ) {
4129                 return $text;
4130         }
4131
4132         /**
4133          * and unsegment to show the result
4134          *
4135          * @param string $text
4136          * @return string
4137          */
4138         public function unsegmentForDiff( $text ) {
4139                 return $text;
4140         }
4141
4142         /**
4143          * Return the LanguageConverter used in the Language
4144          *
4145          * @since 1.19
4146          * @return LanguageConverter
4147          */
4148         public function getConverter() {
4149                 return $this->mConverter;
4150         }
4151
4152         /**
4153          * convert text to a variant
4154          *
4155          * @param string $text text to convert
4156          * @param string|bool $variant variant to convert to, or false to use the user's preferred
4157          *      variant (if logged in), or the project default variant
4158          * @return string the converted string
4159          */
4160         public function autoConvert( $text, $variant = false ) {
4161                 return $this->mConverter->autoConvert( $text, $variant );
4162         }
4163
4164         /**
4165          * convert text to all supported variants
4166          *
4167          * @param string $text
4168          * @return array
4169          */
4170         public function autoConvertToAllVariants( $text ) {
4171                 return $this->mConverter->autoConvertToAllVariants( $text );
4172         }
4173
4174         /**
4175          * convert text to different variants of a language.
4176          *
4177          * @warning Glossary state is maintained between calls. This means
4178          *  if you pass unescaped text to this method it can cause an XSS
4179          *  in later calls to this method, even if the later calls have properly
4180          *  escaped the input. Never feed this method user controlled text that
4181          *  is not properly escaped!
4182          * @param string $text Content that has been already escaped for use in HTML
4183          * @return string HTML
4184          */
4185         public function convert( $text ) {
4186                 return $this->mConverter->convert( $text );
4187         }
4188
4189         /**
4190          * Convert a Title object to a string in the preferred variant
4191          *
4192          * @param Title $title
4193          * @return string
4194          */
4195         public function convertTitle( $title ) {
4196                 return $this->mConverter->convertTitle( $title );
4197         }
4198
4199         /**
4200          * Convert a namespace index to a string in the preferred variant
4201          *
4202          * @param int $ns namespace index (https://www.mediawiki.org/wiki/Manual:Namespace)
4203          * @param string|null $variant variant to convert to, or null to use the user's preferred
4204          *      variant (if logged in), or the project default variant
4205          * @return string a string representation of the namespace
4206          */
4207         public function convertNamespace( $ns, $variant = null ) {
4208                 return $this->mConverter->convertNamespace( $ns, $variant );
4209         }
4210
4211         /**
4212          * Check if this is a language with variants
4213          *
4214          * @return bool
4215          */
4216         public function hasVariants() {
4217                 return count( $this->getVariants() ) > 1;
4218         }
4219
4220         /**
4221          * Strict check if the language has the specific variant.
4222          *
4223          * Compare to LanguageConverter::validateVariant() which does a more
4224          * lenient check and attempts to coerce the given code to a valid one.
4225          *
4226          * @since 1.19
4227          * @param string $variant
4228          * @return bool
4229          */
4230         public function hasVariant( $variant ) {
4231                 return $variant && ( $variant === $this->mConverter->validateVariant( $variant ) );
4232         }
4233
4234         /**
4235          * Perform output conversion on a string, and encode for safe HTML output.
4236          * @param string $text Text to be converted
4237          * @return string
4238          * @todo this should get integrated somewhere sane
4239          */
4240         public function convertHtml( $text ) {
4241                 return htmlspecialchars( $this->convert( $text ) );
4242         }
4243
4244         /**
4245          * @param string $key
4246          * @return string
4247          */
4248         public function convertCategoryKey( $key ) {
4249                 return $this->mConverter->convertCategoryKey( $key );
4250         }
4251
4252         /**
4253          * Get the list of variants supported by this language
4254          * see sample implementation in LanguageZh.php
4255          *
4256          * @return string[] An array of language codes
4257          */
4258         public function getVariants() {
4259                 return $this->mConverter->getVariants();
4260         }
4261
4262         /**
4263          * @return string
4264          */
4265         public function getPreferredVariant() {
4266                 return $this->mConverter->getPreferredVariant();
4267         }
4268
4269         /**
4270          * @return string
4271          */
4272         public function getDefaultVariant() {
4273                 return $this->mConverter->getDefaultVariant();
4274         }
4275
4276         /**
4277          * @return string
4278          */
4279         public function getURLVariant() {
4280                 return $this->mConverter->getURLVariant();
4281         }
4282
4283         /**
4284          * If a language supports multiple variants, it is
4285          * possible that non-existing link in one variant
4286          * actually exists in another variant. this function
4287          * tries to find it. See e.g. LanguageZh.php
4288          * The input parameters may be modified upon return
4289          *
4290          * @param string &$link The name of the link
4291          * @param Title &$nt The title object of the link
4292          * @param bool $ignoreOtherCond To disable other conditions when
4293          *   we need to transclude a template or update a category's link
4294          */
4295         public function findVariantLink( &$link, &$nt, $ignoreOtherCond = false ) {
4296                 $this->mConverter->findVariantLink( $link, $nt, $ignoreOtherCond );
4297         }
4298
4299         /**
4300          * returns language specific options used by User::getPageRenderHash()
4301          * for example, the preferred language variant
4302          *
4303          * @return string
4304          */
4305         function getExtraHashOptions() {
4306                 return $this->mConverter->getExtraHashOptions();
4307         }
4308
4309         /**
4310          * For languages that support multiple variants, the title of an
4311          * article may be displayed differently in different variants. this
4312          * function returns the apporiate title defined in the body of the article.
4313          *
4314          * @return string
4315          */
4316         public function getParsedTitle() {
4317                 return $this->mConverter->getParsedTitle();
4318         }
4319
4320         /**
4321          * Refresh the cache of conversion tables when
4322          * MediaWiki:Conversiontable* is updated.
4323          *
4324          * @param Title $title The Title of the page being updated
4325          */
4326         public function updateConversionTable( Title $title ) {
4327                 $this->mConverter->updateConversionTable( $title );
4328         }
4329
4330         /**
4331          * A regular expression to match legal word-trailing characters
4332          * which should be merged onto a link of the form [[foo]]bar.
4333          *
4334          * @return string
4335          */
4336         public function linkTrail() {
4337                 return self::$dataCache->getItem( $this->mCode, 'linkTrail' );
4338         }
4339
4340         /**
4341          * A regular expression character set to match legal word-prefixing
4342          * characters which should be merged onto a link of the form foo[[bar]].
4343          *
4344          * @return string
4345          */
4346         public function linkPrefixCharset() {
4347                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixCharset' );
4348         }
4349
4350         /**
4351          * Get the "parent" language which has a converter to convert a "compatible" language
4352          * (in another variant) to this language (eg. zh for zh-cn, but not en for en-gb).
4353          *
4354          * @return Language|null
4355          * @since 1.22
4356          */
4357         public function getParentLanguage() {
4358                 if ( $this->mParentLanguage !== false ) {
4359                         return $this->mParentLanguage;
4360                 }
4361
4362                 $code = explode( '-', $this->getCode() )[0];
4363                 if ( !in_array( $code, LanguageConverter::$languagesWithVariants ) ) {
4364                         $this->mParentLanguage = null;
4365                         return null;
4366                 }
4367                 $lang = self::factory( $code );
4368                 if ( !$lang->hasVariant( $this->getCode() ) ) {
4369                         $this->mParentLanguage = null;
4370                         return null;
4371                 }
4372
4373                 $this->mParentLanguage = $lang;
4374                 return $lang;
4375         }
4376
4377         /**
4378          * Compare with an other language object
4379          *
4380          * @since 1.28
4381          * @param Language $lang
4382          * @return bool
4383          */
4384         public function equals( Language $lang ) {
4385                 return $lang === $this || $lang->getCode() === $this->mCode;
4386         }
4387
4388         /**
4389          * Get the internal language code for this language object
4390          *
4391          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4392          * htmlspecialchars() or similar
4393          *
4394          * @return string
4395          */
4396         public function getCode() {
4397                 return $this->mCode;
4398         }
4399
4400         /**
4401          * Get the code in BCP 47 format which we can use
4402          * inside of html lang="" tags.
4403          *
4404          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4405          * htmlspecialchars() or similar.
4406          *
4407          * @since 1.19
4408          * @return string
4409          */
4410         public function getHtmlCode() {
4411                 if ( is_null( $this->mHtmlCode ) ) {
4412                         $this->mHtmlCode = LanguageCode::bcp47( $this->getCode() );
4413                 }
4414                 return $this->mHtmlCode;
4415         }
4416
4417         /**
4418          * Get the language code from a file name. Inverse of getFileName()
4419          * @param string $filename $prefix . $languageCode . $suffix
4420          * @param string $prefix Prefix before the language code
4421          * @param string $suffix Suffix after the language code
4422          * @return string Language code, or false if $prefix or $suffix isn't found
4423          */
4424         public static function getCodeFromFileName( $filename, $prefix = 'Language', $suffix = '.php' ) {
4425                 $m = null;
4426                 preg_match( '/' . preg_quote( $prefix, '/' ) . '([A-Z][a-z_]+)' .
4427                         preg_quote( $suffix, '/' ) . '/', $filename, $m );
4428                 if ( !count( $m ) ) {
4429                         return false;
4430                 }
4431                 return str_replace( '_', '-', strtolower( $m[1] ) );
4432         }
4433
4434         /**
4435          * @param string $code
4436          * @param bool $fallback Whether we're going through language fallback chain
4437          * @return string Name of the language class
4438          */
4439         public static function classFromCode( $code, $fallback = true ) {
4440                 if ( $fallback && $code == 'en' ) {
4441                         return 'Language';
4442                 } else {
4443                         return 'Language' . str_replace( '-', '_', ucfirst( $code ) );
4444                 }
4445         }
4446
4447         /**
4448          * Get the name of a file for a certain language code
4449          * @param string $prefix Prepend this to the filename
4450          * @param string $code Language code
4451          * @param string $suffix Append this to the filename
4452          * @throws MWException
4453          * @return string $prefix . $mangledCode . $suffix
4454          */
4455         public static function getFileName( $prefix, $code, $suffix = '.php' ) {
4456                 if ( !self::isValidBuiltInCode( $code ) ) {
4457                         throw new MWException( "Invalid language code \"$code\"" );
4458                 }
4459
4460                 return $prefix . str_replace( '-', '_', ucfirst( $code ) ) . $suffix;
4461         }
4462
4463         /**
4464          * @param string $code
4465          * @return string
4466          */
4467         public static function getMessagesFileName( $code ) {
4468                 global $IP;
4469                 $file = self::getFileName( "$IP/languages/messages/Messages", $code, '.php' );
4470                 Hooks::run( 'Language::getMessagesFileName', [ $code, &$file ] );
4471                 return $file;
4472         }
4473
4474         /**
4475          * @param string $code
4476          * @return string
4477          * @throws MWException
4478          * @since 1.23
4479          */
4480         public static function getJsonMessagesFileName( $code ) {
4481                 global $IP;
4482
4483                 if ( !self::isValidBuiltInCode( $code ) ) {
4484                         throw new MWException( "Invalid language code \"$code\"" );
4485                 }
4486
4487                 return "$IP/languages/i18n/$code.json";
4488         }
4489
4490         /**
4491          * Get the first fallback for a given language.
4492          *
4493          * @param string $code
4494          *
4495          * @return bool|string
4496          */
4497         public static function getFallbackFor( $code ) {
4498                 $fallbacks = self::getFallbacksFor( $code );
4499                 if ( $fallbacks ) {
4500                         return $fallbacks[0];
4501                 }
4502                 return false;
4503         }
4504
4505         /**
4506          * Get the ordered list of fallback languages.
4507          *
4508          * @since 1.19
4509          * @param string $code Language code
4510          * @param int $mode Fallback mode, either MESSAGES_FALLBACKS (which always falls back to 'en'),
4511          * or STRICT_FALLBACKS (whic honly falls back to 'en' when explicitly defined)
4512          * @throws MWException
4513          * @return array List of language codes
4514          */
4515         public static function getFallbacksFor( $code, $mode = self::MESSAGES_FALLBACKS ) {
4516                 if ( $code === 'en' || !self::isValidBuiltInCode( $code ) ) {
4517                         return [];
4518                 }
4519                 switch ( $mode ) {
4520                         case self::MESSAGES_FALLBACKS:
4521                                 // For unknown languages, fallbackSequence returns an empty array,
4522                                 // hardcode fallback to 'en' in that case as English messages are
4523                                 // always defined.
4524                                 return self::getLocalisationCache()->getItem( $code, 'fallbackSequence' ) ?: [ 'en' ];
4525                         case self::STRICT_FALLBACKS:
4526                                 // Use this mode when you don't want to fallback to English unless
4527                                 // explicitly defined, for example when you have language-variant icons
4528                                 // and an international language-independent fallback.
4529                                 return self::getLocalisationCache()->getItem( $code, 'originalFallbackSequence' );
4530                         default:
4531                                 throw new MWException( "Invalid fallback mode \"$mode\"" );
4532                 }
4533         }
4534
4535         /**
4536          * Get the ordered list of fallback languages, ending with the fallback
4537          * language chain for the site language.
4538          *
4539          * @since 1.22
4540          * @param string $code Language code
4541          * @return array [ fallbacks, site fallbacks ]
4542          */
4543         public static function getFallbacksIncludingSiteLanguage( $code ) {
4544                 global $wgLanguageCode;
4545
4546                 // Usually, we will only store a tiny number of fallback chains, so we
4547                 // keep them in static memory.
4548                 $cacheKey = "{$code}-{$wgLanguageCode}";
4549
4550                 if ( !array_key_exists( $cacheKey, self::$fallbackLanguageCache ) ) {
4551                         $fallbacks = self::getFallbacksFor( $code );
4552
4553                         // Append the site's fallback chain, including the site language itself
4554                         $siteFallbacks = self::getFallbacksFor( $wgLanguageCode );
4555                         array_unshift( $siteFallbacks, $wgLanguageCode );
4556
4557                         // Eliminate any languages already included in the chain
4558                         $siteFallbacks = array_diff( $siteFallbacks, $fallbacks );
4559
4560                         self::$fallbackLanguageCache[$cacheKey] = [ $fallbacks, $siteFallbacks ];
4561                 }
4562                 return self::$fallbackLanguageCache[$cacheKey];
4563         }
4564
4565         /**
4566          * Get all messages for a given language
4567          * WARNING: this may take a long time. If you just need all message *keys*
4568          * but need the *contents* of only a few messages, consider using getMessageKeysFor().
4569          *
4570          * @param string $code
4571          *
4572          * @return array
4573          */
4574         public static function getMessagesFor( $code ) {
4575                 return self::getLocalisationCache()->getItem( $code, 'messages' );
4576         }
4577
4578         /**
4579          * Get a message for a given language
4580          *
4581          * @param string $key
4582          * @param string $code
4583          *
4584          * @return string
4585          */
4586         public static function getMessageFor( $key, $code ) {
4587                 return self::getLocalisationCache()->getSubitem( $code, 'messages', $key );
4588         }
4589
4590         /**
4591          * Get all message keys for a given language. This is a faster alternative to
4592          * array_keys( Language::getMessagesFor( $code ) )
4593          *
4594          * @since 1.19
4595          * @param string $code Language code
4596          * @return array Array of message keys (strings)
4597          */
4598         public static function getMessageKeysFor( $code ) {
4599                 return self::getLocalisationCache()->getSubitemList( $code, 'messages' );
4600         }
4601
4602         /**
4603          * @param string $talk
4604          * @return string
4605          */
4606         function fixVariableInNamespace( $talk ) {
4607                 if ( strpos( $talk, '$1' ) === false ) {
4608                         return $talk;
4609                 }
4610
4611                 global $wgMetaNamespace;
4612                 $talk = str_replace( '$1', $wgMetaNamespace, $talk );
4613
4614                 # Allow grammar transformations
4615                 # Allowing full message-style parsing would make simple requests
4616                 # such as action=raw much more expensive than they need to be.
4617                 # This will hopefully cover most cases.
4618                 $talk = preg_replace_callback( '/{{grammar:(.*?)\|(.*?)}}/i',
4619                         [ $this, 'replaceGrammarInNamespace' ], $talk );
4620                 return str_replace( ' ', '_', $talk );
4621         }
4622
4623         /**
4624          * @param string $m
4625          * @return string
4626          */
4627         function replaceGrammarInNamespace( $m ) {
4628                 return $this->convertGrammar( trim( $m[2] ), trim( $m[1] ) );
4629         }
4630
4631         /**
4632          * Decode an expiry (block, protection, etc) which has come from the DB
4633          *
4634          * @param string $expiry Database expiry String
4635          * @param bool|int $format True to process using language functions, or TS_ constant
4636          *     to return the expiry in a given timestamp
4637          * @param string $infinity If $format is not true, use this string for infinite expiry
4638          * @return string
4639          * @since 1.18
4640          */
4641         public function formatExpiry( $expiry, $format = true, $infinity = 'infinity' ) {
4642                 static $dbInfinity;
4643                 if ( $dbInfinity === null ) {
4644                         $dbInfinity = wfGetDB( DB_REPLICA )->getInfinity();
4645                 }
4646
4647                 if ( $expiry == '' || $expiry === 'infinity' || $expiry == $dbInfinity ) {
4648                         return $format === true
4649                                 ? $this->getMessageFromDB( 'infiniteblock' )
4650                                 : $infinity;
4651                 } else {
4652                         return $format === true
4653                                 ? $this->timeanddate( $expiry, /* User preference timezone */ true )
4654                                 : wfTimestamp( $format, $expiry );
4655                 }
4656         }
4657
4658         /**
4659          * Formats a time given in seconds into a string representation of that time.
4660          *
4661          * @param int|float $seconds
4662          * @param array $format An optional argument that formats the returned string in different ways:
4663          *   If $format['avoid'] === 'avoidhours': don't show hours, just show days
4664          *   If $format['avoid'] === 'avoidseconds': don't show seconds if $seconds >= 1 hour,
4665          *   If $format['avoid'] === 'avoidminutes': don't show seconds/minutes if $seconds > 48 hours,
4666          *   If $format['noabbrevs'] is true: use 'seconds' and friends instead of 'seconds-abbrev'
4667          *     and friends.
4668          * @note For backwards compatibility, $format may also be one of the strings 'avoidseconds'
4669          *     or 'avoidminutes'.
4670          * @return string
4671          */
4672         function formatTimePeriod( $seconds, $format = [] ) {
4673                 if ( !is_array( $format ) ) {
4674                         $format = [ 'avoid' => $format ]; // For backwards compatibility
4675                 }
4676                 if ( !isset( $format['avoid'] ) ) {
4677                         $format['avoid'] = false;
4678                 }
4679                 if ( !isset( $format['noabbrevs'] ) ) {
4680                         $format['noabbrevs'] = false;
4681                 }
4682                 $secondsMsg = wfMessage(
4683                         $format['noabbrevs'] ? 'seconds' : 'seconds-abbrev' )->inLanguage( $this );
4684                 $minutesMsg = wfMessage(
4685                         $format['noabbrevs'] ? 'minutes' : 'minutes-abbrev' )->inLanguage( $this );
4686                 $hoursMsg = wfMessage(
4687                         $format['noabbrevs'] ? 'hours' : 'hours-abbrev' )->inLanguage( $this );
4688                 $daysMsg = wfMessage(
4689                         $format['noabbrevs'] ? 'days' : 'days-abbrev' )->inLanguage( $this );
4690
4691                 if ( round( $seconds * 10 ) < 100 ) {
4692                         $s = $this->formatNum( sprintf( "%.1f", round( $seconds * 10 ) / 10 ) );
4693                         $s = $secondsMsg->params( $s )->text();
4694                 } elseif ( round( $seconds ) < 60 ) {
4695                         $s = $this->formatNum( round( $seconds ) );
4696                         $s = $secondsMsg->params( $s )->text();
4697                 } elseif ( round( $seconds ) < 3600 ) {
4698                         $minutes = floor( $seconds / 60 );
4699                         $secondsPart = round( fmod( $seconds, 60 ) );
4700                         if ( $secondsPart == 60 ) {
4701                                 $secondsPart = 0;
4702                                 $minutes++;
4703                         }
4704                         $s = $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4705                         $s .= ' ';
4706                         $s .= $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4707                 } elseif ( round( $seconds ) <= 2 * 86400 ) {
4708                         $hours = floor( $seconds / 3600 );
4709                         $minutes = floor( ( $seconds - $hours * 3600 ) / 60 );
4710                         $secondsPart = round( $seconds - $hours * 3600 - $minutes * 60 );
4711                         if ( $secondsPart == 60 ) {
4712                                 $secondsPart = 0;
4713                                 $minutes++;
4714                         }
4715                         if ( $minutes == 60 ) {
4716                                 $minutes = 0;
4717                                 $hours++;
4718                         }
4719                         $s = $hoursMsg->params( $this->formatNum( $hours ) )->text();
4720                         $s .= ' ';
4721                         $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4722                         if ( !in_array( $format['avoid'], [ 'avoidseconds', 'avoidminutes', 'avoidhours' ] ) ) {
4723                                 $s .= ' ' . $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4724                         }
4725                 } else {
4726                         $days = floor( $seconds / 86400 );
4727                         if ( $format['avoid'] === 'avoidhours' ) {
4728                                 $hours = round( ( $seconds - $days * 86400 ) / 3600 );
4729                                 if ( $hours == 24 ) {
4730                                         $hours = 0;
4731                                         $days++;
4732                                 }
4733                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4734                         } elseif ( $format['avoid'] === 'avoidminutes' ) {
4735                                 $hours = round( ( $seconds - $days * 86400 ) / 3600 );
4736                                 if ( $hours == 24 ) {
4737                                         $hours = 0;
4738                                         $days++;
4739                                 }
4740                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4741                                 $s .= ' ';
4742                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4743                         } elseif ( $format['avoid'] === 'avoidseconds' ) {
4744                                 $hours = floor( ( $seconds - $days * 86400 ) / 3600 );
4745                                 $minutes = round( ( $seconds - $days * 86400 - $hours * 3600 ) / 60 );
4746                                 if ( $minutes == 60 ) {
4747                                         $minutes = 0;
4748                                         $hours++;
4749                                 }
4750                                 if ( $hours == 24 ) {
4751                                         $hours = 0;
4752                                         $days++;
4753                                 }
4754                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4755                                 $s .= ' ';
4756                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4757                                 $s .= ' ';
4758                                 $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4759                         } else {
4760                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4761                                 $s .= ' ';
4762                                 $s .= $this->formatTimePeriod( $seconds - $days * 86400, $format );
4763                         }
4764                 }
4765                 return $s;
4766         }
4767
4768         /**
4769          * Format a bitrate for output, using an appropriate
4770          * unit (bps, kbps, Mbps, Gbps, Tbps, Pbps, Ebps, Zbps or Ybps) according to
4771          *   the magnitude in question.
4772          *
4773          * This use base 1000. For base 1024 use formatSize(), for another base
4774          * see formatComputingNumbers().
4775          *
4776          * @param int $bps
4777          * @return string
4778          */
4779         function formatBitrate( $bps ) {
4780                 return $this->formatComputingNumbers( $bps, 1000, "bitrate-$1bits" );
4781         }
4782
4783         /**
4784          * @param int $size Size of the unit
4785          * @param int $boundary Size boundary (1000, or 1024 in most cases)
4786          * @param string $messageKey Message key to be uesd
4787          * @return string
4788          */
4789         function formatComputingNumbers( $size, $boundary, $messageKey ) {
4790                 if ( $size <= 0 ) {
4791                         return str_replace( '$1', $this->formatNum( $size ),
4792                                 $this->getMessageFromDB( str_replace( '$1', '', $messageKey ) )
4793                         );
4794                 }
4795                 $sizes = [ '', 'kilo', 'mega', 'giga', 'tera', 'peta', 'exa', 'zeta', 'yotta' ];
4796                 $index = 0;
4797
4798                 $maxIndex = count( $sizes ) - 1;
4799                 while ( $size >= $boundary && $index < $maxIndex ) {
4800                         $index++;
4801                         $size /= $boundary;
4802                 }
4803
4804                 // For small sizes no decimal places necessary
4805                 $round = 0;
4806                 if ( $index > 1 ) {
4807                         // For MB and bigger two decimal places are smarter
4808                         $round = 2;
4809                 }
4810                 $msg = str_replace( '$1', $sizes[$index], $messageKey );
4811
4812                 $size = round( $size, $round );
4813                 $text = $this->getMessageFromDB( $msg );
4814                 return str_replace( '$1', $this->formatNum( $size ), $text );
4815         }
4816
4817         /**
4818          * Format a size in bytes for output, using an appropriate
4819          * unit (B, KB, MB, GB, TB, PB, EB, ZB or YB) according to the magnitude in question
4820          *
4821          * This method use base 1024. For base 1000 use formatBitrate(), for
4822          * another base see formatComputingNumbers()
4823          *
4824          * @param int $size Size to format
4825          * @return string Plain text (not HTML)
4826          */
4827         function formatSize( $size ) {
4828                 return $this->formatComputingNumbers( $size, 1024, "size-$1bytes" );
4829         }
4830
4831         /**
4832          * Make a list item, used by various special pages
4833          *
4834          * @param string $page Page link
4835          * @param string $details HTML safe text between brackets
4836          * @param bool $oppositedm Add the direction mark opposite to your
4837          *   language, to display text properly
4838          * @return string HTML escaped
4839          */
4840         function specialList( $page, $details, $oppositedm = true ) {
4841                 if ( !$details ) {
4842                         return $page;
4843                 }
4844
4845                 $dirmark = ( $oppositedm ? $this->getDirMark( true ) : '' ) . $this->getDirMark();
4846                 return $page .
4847                         $dirmark .
4848                         $this->msg( 'word-separator' )->escaped() .
4849                         $this->msg( 'parentheses' )->rawParams( $details )->escaped();
4850         }
4851
4852         /**
4853          * Generate (prev x| next x) (20|50|100...) type links for paging
4854          *
4855          * @param Title $title Title object to link
4856          * @param int $offset
4857          * @param int $limit
4858          * @param array $query Optional URL query parameter string
4859          * @param bool $atend Optional param for specified if this is the last page
4860          * @return string
4861          * @deprecated since 1.34, use PrevNextNavigationRenderer::buildPrevNextNavigation()
4862          *  instead.
4863          */
4864         public function viewPrevNext( Title $title, $offset, $limit,
4865                 array $query = [], $atend = false
4866         ) {
4867                 wfDeprecated( __METHOD__, '1.34' );
4868                 // @todo FIXME: Why on earth this needs one message for the text and another one for tooltip?
4869
4870                 # Make 'previous' link
4871                 $prev = wfMessage( 'prevn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4872                 if ( $offset > 0 ) {
4873                         $plink = $this->numLink( $title, max( $offset - $limit, 0 ), $limit,
4874                                 $query, $prev, 'prevn-title', 'mw-prevlink' );
4875                 } else {
4876                         $plink = htmlspecialchars( $prev );
4877                 }
4878
4879                 # Make 'next' link
4880                 $next = wfMessage( 'nextn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4881                 if ( $atend ) {
4882                         $nlink = htmlspecialchars( $next );
4883                 } else {
4884                         $nlink = $this->numLink( $title, $offset + $limit, $limit,
4885                                 $query, $next, 'nextn-title', 'mw-nextlink' );
4886                 }
4887
4888                 # Make links to set number of items per page
4889                 $numLinks = [];
4890                 foreach ( [ 20, 50, 100, 250, 500 ] as $num ) {
4891                         $numLinks[] = $this->numLink( $title, $offset, $num,
4892                                 $query, $this->formatNum( $num ), 'shown-title', 'mw-numlink' );
4893                 }
4894
4895                 return wfMessage( 'viewprevnext' )->inLanguage( $this )->title( $title
4896                         )->rawParams( $plink, $nlink, $this->pipeList( $numLinks ) )->escaped();
4897         }
4898
4899         /**
4900          * Helper function for viewPrevNext() that generates links
4901          *
4902          * @param Title $title Title object to link
4903          * @param int $offset
4904          * @param int $limit
4905          * @param array $query Extra query parameters
4906          * @param string $link Text to use for the link; will be escaped
4907          * @param string $tooltipMsg Name of the message to use as tooltip
4908          * @param string $class Value of the "class" attribute of the link
4909          * @return string HTML fragment
4910          */
4911         private function numLink( Title $title, $offset, $limit, array $query, $link,
4912                 $tooltipMsg, $class
4913         ) {
4914                 $query = [ 'limit' => $limit, 'offset' => $offset ] + $query;
4915                 $tooltip = wfMessage( $tooltipMsg )->inLanguage( $this )->title( $title )
4916                         ->numParams( $limit )->text();
4917
4918                 return Html::element( 'a', [ 'href' => $title->getLocalURL( $query ),
4919                         'title' => $tooltip, 'class' => $class ], $link );
4920         }
4921
4922         /**
4923          * Get the conversion rule title, if any.
4924          *
4925          * @return string
4926          */
4927         public function getConvRuleTitle() {
4928                 return $this->mConverter->getConvRuleTitle();
4929         }
4930
4931         /**
4932          * Get the compiled plural rules for the language
4933          * @since 1.20
4934          * @return array Associative array with plural form, and plural rule as key-value pairs
4935          */
4936         public function getCompiledPluralRules() {
4937                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'compiledPluralRules' );
4938                 $fallbacks = self::getFallbacksFor( $this->mCode );
4939                 if ( !$pluralRules ) {
4940                         foreach ( $fallbacks as $fallbackCode ) {
4941                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'compiledPluralRules' );
4942                                 if ( $pluralRules ) {
4943                                         break;
4944                                 }
4945                         }
4946                 }
4947                 return $pluralRules;
4948         }
4949
4950         /**
4951          * Get the plural rules for the language
4952          * @since 1.20
4953          * @return array Associative array with plural form number and plural rule as key-value pairs
4954          */
4955         public function getPluralRules() {
4956                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRules' );
4957                 $fallbacks = self::getFallbacksFor( $this->mCode );
4958                 if ( !$pluralRules ) {
4959                         foreach ( $fallbacks as $fallbackCode ) {
4960                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRules' );
4961                                 if ( $pluralRules ) {
4962                                         break;
4963                                 }
4964                         }
4965                 }
4966                 return $pluralRules;
4967         }
4968
4969         /**
4970          * Get the plural rule types for the language
4971          * @since 1.22
4972          * @return array Associative array with plural form number and plural rule type as key-value pairs
4973          */
4974         public function getPluralRuleTypes() {
4975                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRuleTypes' );
4976                 $fallbacks = self::getFallbacksFor( $this->mCode );
4977                 if ( !$pluralRuleTypes ) {
4978                         foreach ( $fallbacks as $fallbackCode ) {
4979                                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRuleTypes' );
4980                                 if ( $pluralRuleTypes ) {
4981                                         break;
4982                                 }
4983                         }
4984                 }
4985                 return $pluralRuleTypes;
4986         }
4987
4988         /**
4989          * Find the index number of the plural rule appropriate for the given number
4990          * @param int $number
4991          * @return int The index number of the plural rule
4992          */
4993         public function getPluralRuleIndexNumber( $number ) {
4994                 $pluralRules = $this->getCompiledPluralRules();
4995                 $form = Evaluator::evaluateCompiled( $number, $pluralRules );
4996                 return $form;
4997         }
4998
4999         /**
5000          * Find the plural rule type appropriate for the given number
5001          * For example, if the language is set to Arabic, getPluralType(5) should
5002          * return 'few'.
5003          * @since 1.22
5004          * @param int $number
5005          * @return string The name of the plural rule type, e.g. one, two, few, many
5006          */
5007         public function getPluralRuleType( $number ) {
5008                 $index = $this->getPluralRuleIndexNumber( $number );
5009                 $pluralRuleTypes = $this->getPluralRuleTypes();
5010                 return $pluralRuleTypes[$index] ?? 'other';
5011         }
5012 }