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         /** @var LocalisationCache */
  79         private $localisationCache;
  80
  81         public static $mLangObjCache = [];
  82
  83         /**
  84          * Return a fallback chain for messages in getFallbacksFor
  85          * @since 1.32
  86          */
  87         const MESSAGES_FALLBACKS = 0;
  88
  89         /**
  90          * Return a strict fallback chain in getFallbacksFor
  91          * @since 1.32
  92          */
  93         const STRICT_FALLBACKS = 1;
  94
  95         public static $mWeekdayMsgs = [
  96                 'sunday', 'monday', 'tuesday', 'wednesday', 'thursday',
  97                 'friday', 'saturday'
  98         ];
  99
 100         public static $mWeekdayAbbrevMsgs = [
 101                 'sun', 'mon', 'tue', 'wed', 'thu', 'fri', 'sat'
 102         ];
 103
 104         public static $mMonthMsgs = [
 105                 'january', 'february', 'march', 'april', 'may_long', 'june',
 106                 'july', 'august', 'september', 'october', 'november',
 107                 'december'
 108         ];
 109         public static $mMonthGenMsgs = [
 110                 'january-gen', 'february-gen', 'march-gen', 'april-gen', 'may-gen', 'june-gen',
 111                 'july-gen', 'august-gen', 'september-gen', 'october-gen', 'november-gen',
 112                 'december-gen'
 113         ];
 114         public static $mMonthAbbrevMsgs = [
 115                 'jan', 'feb', 'mar', 'apr', 'may', 'jun', 'jul', 'aug',
 116                 'sep', 'oct', 'nov', 'dec'
 117         ];
 118
 119         public static $mIranianCalendarMonthMsgs = [
 120                 'iranian-calendar-m1', 'iranian-calendar-m2', 'iranian-calendar-m3',
 121                 'iranian-calendar-m4', 'iranian-calendar-m5', 'iranian-calendar-m6',
 122                 'iranian-calendar-m7', 'iranian-calendar-m8', 'iranian-calendar-m9',
 123                 'iranian-calendar-m10', 'iranian-calendar-m11', 'iranian-calendar-m12'
 124         ];
 125
 126         public static $mHebrewCalendarMonthMsgs = [
 127                 'hebrew-calendar-m1', 'hebrew-calendar-m2', 'hebrew-calendar-m3',
 128                 'hebrew-calendar-m4', 'hebrew-calendar-m5', 'hebrew-calendar-m6',
 129                 'hebrew-calendar-m7', 'hebrew-calendar-m8', 'hebrew-calendar-m9',
 130                 'hebrew-calendar-m10', 'hebrew-calendar-m11', 'hebrew-calendar-m12',
 131                 'hebrew-calendar-m6a', 'hebrew-calendar-m6b'
 132         ];
 133
 134         public static $mHebrewCalendarMonthGenMsgs = [
 135                 'hebrew-calendar-m1-gen', 'hebrew-calendar-m2-gen', 'hebrew-calendar-m3-gen',
 136                 'hebrew-calendar-m4-gen', 'hebrew-calendar-m5-gen', 'hebrew-calendar-m6-gen',
 137                 'hebrew-calendar-m7-gen', 'hebrew-calendar-m8-gen', 'hebrew-calendar-m9-gen',
 138                 'hebrew-calendar-m10-gen', 'hebrew-calendar-m11-gen', 'hebrew-calendar-m12-gen',
 139                 'hebrew-calendar-m6a-gen', 'hebrew-calendar-m6b-gen'
 140         ];
 141
 142         public static $mHijriCalendarMonthMsgs = [
 143                 'hijri-calendar-m1', 'hijri-calendar-m2', 'hijri-calendar-m3',
 144                 'hijri-calendar-m4', 'hijri-calendar-m5', 'hijri-calendar-m6',
 145                 'hijri-calendar-m7', 'hijri-calendar-m8', 'hijri-calendar-m9',
 146                 'hijri-calendar-m10', 'hijri-calendar-m11', 'hijri-calendar-m12'
 147         ];
 148
 149         /**
 150          * @since 1.20
 151          * @var array
 152          */
 153         public static $durationIntervals = [
 154                 'millennia' => 31556952000,
 155                 'centuries' => 3155695200,
 156                 'decades' => 315569520,
 157                 'years' => 31556952, // 86400 * ( 365 + ( 24 * 3 + 25 ) / 400 )
 158                 'weeks' => 604800,
 159                 'days' => 86400,
 160                 'hours' => 3600,
 161                 'minutes' => 60,
 162                 'seconds' => 1,
 163         ];
 164
 165         /**
 166          * Cache for language fallbacks.
 167          * @see Language::getFallbacksIncludingSiteLanguage
 168          * @since 1.21
 169          * @var array
 170          */
 171         private static $fallbackLanguageCache = [];
 172
 173         /**
 174          * Cache for grammar rules data
 175          * @var MapCacheLRU|null
 176          */
 177         private static $grammarTransformations;
 178
 179         /**
 180          * Cache for language names
 181          * @var HashBagOStuff|null
 182          */
 183         private static $languageNameCache;
 184
 185         /**
 186          * Unicode directional formatting characters, for embedBidi()
 187          */
 188         private static $lre = "\u{202A}"; // U+202A LEFT-TO-RIGHT EMBEDDING
 189         private static $rle = "\u{202B}"; // U+202B RIGHT-TO-LEFT EMBEDDING
 190         private static $pdf = "\u{202C}"; // U+202C POP DIRECTIONAL FORMATTING
 191
 192         /**
 193          * Directionality test regex for embedBidi(). Matches the first strong directionality codepoint:
 194          * - in group 1 if it is LTR
 195          * - in group 2 if it is RTL
 196          * Does not match if there is no strong directionality codepoint.
 197          *
 198          * The form is '/(?:([strong ltr codepoint])|([strong rtl codepoint]))/u' .
 199          *
 200          * Generated by UnicodeJS (see tools/strongDir) from the UCD; see
 201          * https://phabricator.wikimedia.org/diffusion/GUJS/ .
 202          */
 203         // @codeCoverageIgnoreStart
 204         // phpcs:ignore Generic.Files.LineLength
 205         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';
 206         // @codeCoverageIgnoreEnd
 207
 208         /**
 209          * Get a cached or new language object for a given language code
 210          * @param string $code
 211          * @throws MWException
 212          * @return Language
 213          */
 214         static function factory( $code ) {
 215                 global $wgDummyLanguageCodes, $wgLangObjCacheSize;
 216
 217                 if ( isset( $wgDummyLanguageCodes[$code] ) ) {
 218                         $code = $wgDummyLanguageCodes[$code];
 219                 }
 220
 221                 // get the language object to process
 222                 $langObj = self::$mLangObjCache[$code] ?? self::newFromCode( $code );
 223
 224                 // merge the language object in to get it up front in the cache
 225                 self::$mLangObjCache = array_merge( [ $code => $langObj ], self::$mLangObjCache );
 226                 // get rid of the oldest ones in case we have an overflow
 227                 self::$mLangObjCache = array_slice( self::$mLangObjCache, 0, $wgLangObjCacheSize, true );
 228
 229                 return $langObj;
 230         }
 231
 232         /**
 233          * Create a language object for a given language code
 234          * @param string $code
 235          * @param bool $fallback Whether we're going through language fallback chain
 236          * @throws MWException
 237          * @return Language
 238          */
 239         protected static function newFromCode( $code, $fallback = false ) {
 240                 if ( !self::isValidCode( $code ) ) {
 241                         throw new MWException( "Invalid language code \"$code\"" );
 242                 }
 243
 244                 if ( !self::isValidBuiltInCode( $code ) ) {
 245                         // It's not possible to customise this code with class files, so
 246                         // just return a Language object. This is to support uselang= hacks.
 247                         $lang = new Language;
 248                         $lang->mCode = $code;
 249                         return $lang;
 250                 }
 251
 252                 // Check if there is a language class for the code
 253                 $class = self::classFromCode( $code, $fallback );
 254                 // LanguageCode does not inherit Language
 255                 if ( class_exists( $class ) && is_a( $class, 'Language', true ) ) {
 256                         $lang = new $class;
 257                         return $lang;
 258                 }
 259
 260                 // Keep trying the fallback list until we find an existing class
 261                 $fallbacks = self::getFallbacksFor( $code );
 262                 foreach ( $fallbacks as $fallbackCode ) {
 263                         if ( !self::isValidBuiltInCode( $fallbackCode ) ) {
 264                                 throw new MWException( "Invalid fallback '$fallbackCode' in fallback sequence for '$code'" );
 265                         }
 266
 267                         $class = self::classFromCode( $fallbackCode );
 268                         if ( class_exists( $class ) ) {
 269                                 $lang = new $class;
 270                                 $lang->mCode = $code;
 271                                 return $lang;
 272                         }
 273                 }
 274
 275                 throw new MWException( "Invalid fallback sequence for language '$code'" );
 276         }
 277
 278         /**
 279          * Intended for tests that may change configuration in a way that invalidates caches.
 280          *
 281          * @since 1.32
 282          */
 283         public static function clearCaches() {
 284                 if ( !defined( 'MW_PHPUNIT_TEST' ) && !defined( 'MEDIAWIKI_INSTALL' ) ) {
 285                         throw new MWException( __METHOD__ . ' must not be used outside tests/installer' );
 286                 }
 287                 if ( !defined( 'MEDIAWIKI_INSTALL' ) ) {
 288                         MediaWikiServices::getInstance()->resetServiceForTesting( 'LocalisationCache' );
 289                 }
 290                 self::$mLangObjCache = [];
 291                 self::$fallbackLanguageCache = [];
 292                 self::$grammarTransformations = null;
 293                 self::$languageNameCache = null;
 294         }
 295
 296         /**
 297          * Checks whether any localisation is available for that language tag
 298          * in MediaWiki (MessagesXx.php exists).
 299          *
 300          * @param string $code Language tag (in lower case)
 301          * @return bool Whether language is supported
 302          * @since 1.21
 303          */
 304         public static function isSupportedLanguage( $code ) {
 305                 if ( !self::isValidBuiltInCode( $code ) ) {
 306                         return false;
 307                 }
 308
 309                 if ( $code === 'qqq' ) {
 310                         return false;
 311                 }
 312
 313                 return is_readable( self::getMessagesFileName( $code ) ) ||
 314                         is_readable( self::getJsonMessagesFileName( $code ) );
 315         }
 316
 317         /**
 318          * Returns true if a language code string is a well-formed language tag
 319          * according to RFC 5646.
 320          * This function only checks well-formedness; it doesn't check that
 321          * language, script or variant codes actually exist in the repositories.
 322          *
 323          * Based on regexes by Mark Davis of the Unicode Consortium:
 324          * https://www.unicode.org/repos/cldr/trunk/tools/java/org/unicode/cldr/util/data/langtagRegex.txt
 325          *
 326          * @param string $code
 327          * @param bool $lenient Whether to allow '_' as separator. The default is only '-'.
 328          *
 329          * @return bool
 330          * @since 1.21
 331          */
 332         public static function isWellFormedLanguageTag( $code, $lenient = false ) {
 333                 $alpha = '[a-z]';
 334                 $digit = '[0-9]';
 335                 $alphanum = '[a-z0-9]';
 336                 $x = 'x'; # private use singleton
 337                 $singleton = '[a-wy-z]'; # other singleton
 338                 $s = $lenient ? '[-_]' : '-';
 339
 340                 $language = "$alpha{2,8}|$alpha{2,3}$s$alpha{3}";
 341                 $script = "$alpha{4}"; # ISO 15924
 342                 $region = "(?:$alpha{2}|$digit{3})"; # ISO 3166-1 alpha-2 or UN M.49
 343                 $variant = "(?:$alphanum{5,8}|$digit$alphanum{3})";
 344                 $extension = "$singleton(?:$s$alphanum{2,8})+";
 345                 $privateUse = "$x(?:$s$alphanum{1,8})+";
 346
 347                 # Define certain grandfathered codes, since otherwise the regex is pretty useless.
 348                 # Since these are limited, this is safe even later changes to the registry --
 349                 # the only oddity is that it might change the type of the tag, and thus
 350                 # the results from the capturing groups.
 351                 # https://www.iana.org/assignments/language-subtag-registry
 352
 353                 $grandfathered = "en{$s}GB{$s}oed"
 354                         . "|i{$s}(?:ami|bnn|default|enochian|hak|klingon|lux|mingo|navajo|pwn|tao|tay|tsu)"
 355                         . "|no{$s}(?:bok|nyn)"
 356                         . "|sgn{$s}(?:BE{$s}(?:fr|nl)|CH{$s}de)"
 357                         . "|zh{$s}min{$s}nan";
 358
 359                 $variantList = "$variant(?:$s$variant)*";
 360                 $extensionList = "$extension(?:$s$extension)*";
 361
 362                 $langtag = "(?:($language)"
 363                         . "(?:$s$script)?"
 364                         . "(?:$s$region)?"
 365                         . "(?:$s$variantList)?"
 366                         . "(?:$s$extensionList)?"
 367                         . "(?:$s$privateUse)?)";
 368
 369                 # The final breakdown, with capturing groups for each of these components
 370                 # The variants, extensions, grandfathered, and private-use may have interior '-'
 371
 372                 $root = "^(?:$langtag|$privateUse|$grandfathered)$";
 373
 374                 return (bool)preg_match( "/$root/", strtolower( $code ) );
 375         }
 376
 377         /**
 378          * Returns true if a language code string is of a valid form, whether or
 379          * not it exists. This includes codes which are used solely for
 380          * customisation via the MediaWiki namespace.
 381          *
 382          * @param string $code
 383          *
 384          * @return bool
 385          */
 386         public static function isValidCode( $code ) {
 387                 static $cache = [];
 388                 Assert::parameterType( 'string', $code, '$code' );
 389                 if ( !isset( $cache[$code] ) ) {
 390                         // People think language codes are html safe, so enforce it.
 391                         // Ideally we should only allow a-zA-Z0-9-
 392                         // but, .+ and other chars are often used for {{int:}} hacks
 393                         // see bugs T39564, T39587, T38938
 394                         $cache[$code] =
 395                                 // Protect against path traversal
 396                                 strcspn( $code, ":/\\\000&<>'\"" ) === strlen( $code )
 397                                 && !preg_match( MediaWikiTitleCodec::getTitleInvalidRegex(), $code );
 398                 }
 399                 return $cache[$code];
 400         }
 401
 402         /**
 403          * Returns true if a language code is of a valid form for the purposes of
 404          * internal customisation of MediaWiki, via Messages*.php or *.json.
 405          *
 406          * @param string $code
 407          *
 408          * @since 1.18
 409          * @return bool
 410          */
 411         public static function isValidBuiltInCode( $code ) {
 412                 Assert::parameterType( 'string', $code, '$code' );
 413
 414                 return (bool)preg_match( '/^[a-z0-9-]{2,}$/', $code );
 415         }
 416
 417         /**
 418          * Returns true if a language code is an IETF tag known to MediaWiki.
 419          *
 420          * @param string $tag
 421          *
 422          * @since 1.21
 423          * @return bool
 424          */
 425         public static function isKnownLanguageTag( $tag ) {
 426                 // Quick escape for invalid input to avoid exceptions down the line
 427                 // when code tries to process tags which are not valid at all.
 428                 if ( !self::isValidBuiltInCode( $tag ) ) {
 429                         return false;
 430                 }
 431
 432                 if ( isset( MediaWiki\Languages\Data\Names::$names[$tag] )
 433                         || self::fetchLanguageName( $tag, $tag ) !== ''
 434                 ) {
 435                         return true;
 436                 }
 437
 438                 return false;
 439         }
 440
 441         /**
 442          * Get the LocalisationCache instance
 443          *
 444          * @deprecated since 1.34, use MediaWikiServices
 445          * @return LocalisationCache
 446          */
 447         public static function getLocalisationCache() {
 448                 return MediaWikiServices::getInstance()->getLocalisationCache();
 449         }
 450
 451         function __construct() {
 452                 $this->mConverter = new FakeConverter( $this );
 453                 // Set the code to the name of the descendant
 454                 if ( static::class === 'Language' ) {
 455                         $this->mCode = 'en';
 456                 } else {
 457                         $this->mCode = str_replace( '_', '-', strtolower( substr( static::class, 8 ) ) );
 458                 }
 459                 $this->localisationCache = MediaWikiServices::getInstance()->getLocalisationCache();
 460         }
 461
 462         /**
 463          * Reduce memory usage
 464          */
 465         function __destruct() {
 466                 foreach ( $this as $name => $value ) {
 467                         unset( $this->$name );
 468                 }
 469         }
 470
 471         /**
 472          * Hook which will be called if this is the content language.
 473          * Descendants can use this to register hook functions or modify globals
 474          */
 475         function initContLang() {
 476         }
 477
 478         /**
 479          * @return array
 480          * @since 1.19
 481          */
 482         public function getFallbackLanguages() {
 483                 return self::getFallbacksFor( $this->mCode );
 484         }
 485
 486         /**
 487          * Exports $wgBookstoreListEn
 488          * @return array
 489          */
 490         public function getBookstoreList() {
 491                 return $this->localisationCache->getItem( $this->mCode, 'bookstoreList' );
 492         }
 493
 494         /**
 495          * Returns an array of localised namespaces indexed by their numbers. If the namespace is not
 496          * available in localised form, it will be included in English.
 497          *
 498          * @return string[] List of localized namespace names, indexed by numeric namespace ID.
 499          */
 500         public function getNamespaces() {
 501                 if ( is_null( $this->namespaceNames ) ) {
 502                         global $wgMetaNamespace, $wgMetaNamespaceTalk, $wgExtraNamespaces;
 503
 504                         $validNamespaces = MediaWikiServices::getInstance()->getNamespaceInfo()->
 505                                 getCanonicalNamespaces();
 506
 507                         $this->namespaceNames = $wgExtraNamespaces +
 508                                 $this->localisationCache->getItem( $this->mCode, 'namespaceNames' );
 509                         $this->namespaceNames += $validNamespaces;
 510
 511                         $this->namespaceNames[NS_PROJECT] = $wgMetaNamespace;
 512                         if ( $wgMetaNamespaceTalk ) {
 513                                 $this->namespaceNames[NS_PROJECT_TALK] = $wgMetaNamespaceTalk;
 514                         } else {
 515                                 $talk = $this->namespaceNames[NS_PROJECT_TALK];
 516                                 $this->namespaceNames[NS_PROJECT_TALK] =
 517                                         $this->fixVariableInNamespace( $talk );
 518                         }
 519
 520                         # Sometimes a language will be localised but not actually exist on this wiki.
 521                         foreach ( $this->namespaceNames as $key => $text ) {
 522                                 if ( !isset( $validNamespaces[$key] ) ) {
 523                                         unset( $this->namespaceNames[$key] );
 524                                 }
 525                         }
 526
 527                         # The above mixing may leave namespaces out of canonical order.
 528                         # Re-order by namespace ID number...
 529                         ksort( $this->namespaceNames );
 530
 531                         Hooks::run( 'LanguageGetNamespaces', [ &$this->namespaceNames ] );
 532                 }
 533
 534                 return $this->namespaceNames;
 535         }
 536
 537         /**
 538          * Arbitrarily set all of the namespace names at once. Mainly used for testing
 539          * @param array $namespaces Array of namespaces (id => name)
 540          */
 541         public function setNamespaces( array $namespaces ) {
 542                 $this->namespaceNames = $namespaces;
 543                 $this->mNamespaceIds = null;
 544         }
 545
 546         /**
 547          * Resets all of the namespace caches. Mainly used for testing
 548          */
 549         public function resetNamespaces() {
 550                 $this->namespaceNames = null;
 551                 $this->mNamespaceIds = null;
 552                 $this->namespaceAliases = null;
 553         }
 554
 555         /**
 556          * A convenience function that returns getNamespaces() with spaces instead of underscores
 557          * in values. Useful for producing output to be displayed e.g. in `<select>` forms.
 558          *
 559          * @return array
 560          */
 561         public function getFormattedNamespaces() {
 562                 $ns = $this->getNamespaces();
 563                 foreach ( $ns as $k => $v ) {
 564                         $ns[$k] = strtr( $v, '_', ' ' );
 565                 }
 566                 return $ns;
 567         }
 568
 569         /**
 570          * Get a namespace value by key
 571          *
 572          * <code>
 573          * $mw_ns = $lang->getNsText( NS_MEDIAWIKI );
 574          * echo $mw_ns; // prints 'MediaWiki'
 575          * </code>
 576          *
 577          * @param int $index The array key of the namespace to return
 578          * @return string|bool String if the namespace value exists, otherwise false
 579          */
 580         public function getNsText( $index ) {
 581                 $ns = $this->getNamespaces();
 582                 return $ns[$index] ?? false;
 583         }
 584
 585         /**
 586          * A convenience function that returns the same thing as
 587          * getNsText() except with '_' changed to ' ', useful for
 588          * producing output.
 589          *
 590          * <code>
 591          * $mw_ns = $lang->getFormattedNsText( NS_MEDIAWIKI_TALK );
 592          * echo $mw_ns; // prints 'MediaWiki talk'
 593          * </code>
 594          *
 595          * @param int $index The array key of the namespace to return
 596          * @return string Namespace name without underscores (empty string if namespace does not exist)
 597          */
 598         public function getFormattedNsText( $index ) {
 599                 $ns = $this->getNsText( $index );
 600                 return strtr( $ns, '_', ' ' );
 601         }
 602
 603         /**
 604          * Returns gender-dependent namespace alias if available.
 605          * See https://www.mediawiki.org/wiki/Manual:$wgExtraGenderNamespaces
 606          * @param int $index Namespace index
 607          * @param string $gender Gender key (male, female... )
 608          * @return string
 609          * @since 1.18
 610          */
 611         public function getGenderNsText( $index, $gender ) {
 612                 global $wgExtraGenderNamespaces;
 613
 614                 $ns = $wgExtraGenderNamespaces +
 615                         (array)$this->localisationCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 616
 617                 return $ns[$index][$gender] ?? $this->getNsText( $index );
 618         }
 619
 620         /**
 621          * Whether this language uses gender-dependent namespace aliases.
 622          * See https://www.mediawiki.org/wiki/Manual:$wgExtraGenderNamespaces
 623          * @return bool
 624          * @since 1.18
 625          */
 626         public function needsGenderDistinction() {
 627                 global $wgExtraGenderNamespaces, $wgExtraNamespaces;
 628                 if ( count( $wgExtraGenderNamespaces ) > 0 ) {
 629                         // $wgExtraGenderNamespaces overrides everything
 630                         return true;
 631                 } elseif ( isset( $wgExtraNamespaces[NS_USER] ) && isset( $wgExtraNamespaces[NS_USER_TALK] ) ) {
 632                         /// @todo There may be other gender namespace than NS_USER & NS_USER_TALK in the future
 633                         // $wgExtraNamespaces overrides any gender aliases specified in i18n files
 634                         return false;
 635                 } else {
 636                         // Check what is in i18n files
 637                         $aliases = $this->localisationCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 638                         return count( $aliases ) > 0;
 639                 }
 640         }
 641
 642         /**
 643          * Get a namespace key by value, case insensitive.
 644          * Only matches namespace names for the current language, not the
 645          * canonical ones defined in Namespace.php.
 646          *
 647          * @param string $text
 648          * @return int|bool An integer if $text is a valid value otherwise false
 649          */
 650         function getLocalNsIndex( $text ) {
 651                 $lctext = $this->lc( $text );
 652                 $ids = $this->getNamespaceIds();
 653                 return $ids[$lctext] ?? false;
 654         }
 655
 656         /**
 657          * @return array
 658          */
 659         public function getNamespaceAliases() {
 660                 if ( is_null( $this->namespaceAliases ) ) {
 661                         $aliases = $this->localisationCache->getItem( $this->mCode, 'namespaceAliases' );
 662                         if ( !$aliases ) {
 663                                 $aliases = [];
 664                         } else {
 665                                 foreach ( $aliases as $name => $index ) {
 666                                         if ( $index === NS_PROJECT_TALK ) {
 667                                                 unset( $aliases[$name] );
 668                                                 $name = $this->fixVariableInNamespace( $name );
 669                                                 $aliases[$name] = $index;
 670                                         }
 671                                 }
 672                         }
 673
 674                         global $wgExtraGenderNamespaces;
 675                         $genders = $wgExtraGenderNamespaces + (array)$this->localisationCache
 676                                 ->getItem( $this->mCode, 'namespaceGenderAliases' );
 677                         foreach ( $genders as $index => $forms ) {
 678                                 foreach ( $forms as $alias ) {
 679                                         $aliases[$alias] = $index;
 680                                 }
 681                         }
 682
 683                         # Also add converted namespace names as aliases, to avoid confusion.
 684                         $convertedNames = [];
 685                         foreach ( $this->getVariants() as $variant ) {
 686                                 if ( $variant === $this->mCode ) {
 687                                         continue;
 688                                 }
 689                                 foreach ( $this->getNamespaces() as $ns => $_ ) {
 690                                         $convertedNames[$this->getConverter()->convertNamespace( $ns, $variant )] = $ns;
 691                                 }
 692                         }
 693
 694                         $this->namespaceAliases = $aliases + $convertedNames;
 695
 696                         # Filter out aliases to namespaces that don't exist, e.g. from extensions
 697                         # that aren't loaded here but are included in the l10n cache.
 698                         # (array_intersect preserves keys from its first argument)
 699                         $this->namespaceAliases = array_intersect(
 700                                 $this->namespaceAliases,
 701                                 array_keys( $this->getNamespaces() )
 702                         );
 703                 }
 704
 705                 return $this->namespaceAliases;
 706         }
 707
 708         /**
 709          * @return array
 710          */
 711         public function getNamespaceIds() {
 712                 if ( is_null( $this->mNamespaceIds ) ) {
 713                         global $wgNamespaceAliases;
 714                         # Put namespace names and aliases into a hashtable.
 715                         # If this is too slow, then we should arrange it so that it is done
 716                         # before caching. The catch is that at pre-cache time, the above
 717                         # class-specific fixup hasn't been done.
 718                         $this->mNamespaceIds = [];
 719                         foreach ( $this->getNamespaces() as $index => $name ) {
 720                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 721                         }
 722                         foreach ( $this->getNamespaceAliases() as $name => $index ) {
 723                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 724                         }
 725                         if ( $wgNamespaceAliases ) {
 726                                 foreach ( $wgNamespaceAliases as $name => $index ) {
 727                                         $this->mNamespaceIds[$this->lc( $name )] = $index;
 728                                 }
 729                         }
 730                 }
 731                 return $this->mNamespaceIds;
 732         }
 733
 734         /**
 735          * Get a namespace key by value, case insensitive.  Canonical namespace
 736          * names override custom ones defined for the current language.
 737          *
 738          * @param string $text
 739          * @return int|bool An integer if $text is a valid value otherwise false
 740          */
 741         public function getNsIndex( $text ) {
 742                 $lctext = $this->lc( $text );
 743                 $ns = MediaWikiServices::getInstance()->getNamespaceInfo()->
 744                         getCanonicalIndex( $lctext );
 745                 if ( $ns !== null ) {
 746                         return $ns;
 747                 }
 748                 $ids = $this->getNamespaceIds();
 749                 return $ids[$lctext] ?? false;
 750         }
 751
 752         /**
 753          * short names for language variants used for language conversion links.
 754          *
 755          * @param string $code
 756          * @param bool $usemsg Use the "variantname-xyz" message if it exists
 757          * @return string
 758          */
 759         public function getVariantname( $code, $usemsg = true ) {
 760                 $msg = "variantname-$code";
 761                 if ( $usemsg && wfMessage( $msg )->exists() ) {
 762                         return $this->getMessageFromDB( $msg );
 763                 }
 764                 $name = self::fetchLanguageName( $code );
 765                 if ( $name ) {
 766                         return $name; # if it's defined as a language name, show that
 767                 } else {
 768                         # otherwise, output the language code
 769                         return $code;
 770                 }
 771         }
 772
 773         /**
 774          * @return string[]|bool List of date format preference keys, or false if disabled.
 775          */
 776         public function getDatePreferences() {
 777                 return $this->localisationCache->getItem( $this->mCode, 'datePreferences' );
 778         }
 779
 780         /**
 781          * @return array
 782          */
 783         function getDateFormats() {
 784                 return $this->localisationCache->getItem( $this->mCode, 'dateFormats' );
 785         }
 786
 787         /**
 788          * @return array|string
 789          */
 790         public function getDefaultDateFormat() {
 791                 $df = $this->localisationCache->getItem( $this->mCode, 'defaultDateFormat' );
 792                 if ( $df === 'dmy or mdy' ) {
 793                         global $wgAmericanDates;
 794                         return $wgAmericanDates ? 'mdy' : 'dmy';
 795                 } else {
 796                         return $df;
 797                 }
 798         }
 799
 800         /**
 801          * @return array
 802          */
 803         public function getDatePreferenceMigrationMap() {
 804                 return $this->localisationCache->getItem( $this->mCode, 'datePreferenceMigrationMap' );
 805         }
 806
 807         /**
 808          * @return array
 809          */
 810         public function getExtraUserToggles() {
 811                 wfDeprecated( __METHOD__, '1.34' );
 812                 return [];
 813         }
 814
 815         /**
 816          * @param string $tog
 817          * @return string
 818          */
 819         function getUserToggle( $tog ) {
 820                 return $this->getMessageFromDB( "tog-$tog" );
 821         }
 822
 823         /**
 824          * Get an array of language names, indexed by code.
 825          * @param null|string $inLanguage Code of language in which to return the names
 826          *              Use self::AS_AUTONYMS for autonyms (native names)
 827          * @param string $include One of:
 828          *              self::ALL all available languages
 829          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 830          *              self::SUPPORTED only if the language is in 'mw' *and* has a message file
 831          * @return array Language code => language name (sorted by key)
 832          * @since 1.20
 833          */
 834         public static function fetchLanguageNames( $inLanguage = self::AS_AUTONYMS, $include = 'mw' ) {
 835                 $cacheKey = $inLanguage === self::AS_AUTONYMS ? 'null' : $inLanguage;
 836                 $cacheKey .= ":$include";
 837                 if ( self::$languageNameCache === null ) {
 838                         self::$languageNameCache = new HashBagOStuff( [ 'maxKeys' => 20 ] );
 839                 }
 840
 841                 $ret = self::$languageNameCache->get( $cacheKey );
 842                 if ( !$ret ) {
 843                         $ret = self::fetchLanguageNamesUncached( $inLanguage, $include );
 844                         self::$languageNameCache->set( $cacheKey, $ret );
 845                 }
 846                 return $ret;
 847         }
 848
 849         /**
 850          * Uncached helper for fetchLanguageNames
 851          * @param null|string $inLanguage Code of language in which to return the names
 852          *              Use self::AS_AUTONYMS for autonyms (native names)
 853          * @param string $include One of:
 854          *              self::ALL all available languages
 855          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 856          *              self::SUPPORTED only if the language is in 'mw' *and* has a message file
 857          * @return array Language code => language name (sorted by key)
 858          */
 859         private static function fetchLanguageNamesUncached(
 860                 $inLanguage = self::AS_AUTONYMS,
 861                 $include = 'mw'
 862         ) {
 863                 global $wgExtraLanguageNames, $wgUsePigLatinVariant;
 864
 865                 // If passed an invalid language code to use, fallback to en
 866                 if ( $inLanguage !== self::AS_AUTONYMS && !self::isValidCode( $inLanguage ) ) {
 867                         $inLanguage = 'en';
 868                 }
 869
 870                 $names = [];
 871
 872                 if ( $inLanguage ) {
 873                         # TODO: also include when $inLanguage is null, when this code is more efficient
 874                         Hooks::run( 'LanguageGetTranslatedLanguageNames', [ &$names, $inLanguage ] );
 875                 }
 876
 877                 $mwNames = $wgExtraLanguageNames + MediaWiki\Languages\Data\Names::$names;
 878                 if ( $wgUsePigLatinVariant ) {
 879                         // Pig Latin (for variant development)
 880                         $mwNames['en-x-piglatin'] = 'Igpay Atinlay';
 881                 }
 882
 883                 foreach ( $mwNames as $mwCode => $mwName ) {
 884                         # - Prefer own MediaWiki native name when not using the hook
 885                         # - For other names just add if not added through the hook
 886                         if ( $mwCode === $inLanguage || !isset( $names[$mwCode] ) ) {
 887                                 $names[$mwCode] = $mwName;
 888                         }
 889                 }
 890
 891                 if ( $include === self::ALL ) {
 892                         ksort( $names );
 893                         return $names;
 894                 }
 895
 896                 $returnMw = [];
 897                 $coreCodes = array_keys( $mwNames );
 898                 foreach ( $coreCodes as $coreCode ) {
 899                         $returnMw[$coreCode] = $names[$coreCode];
 900                 }
 901
 902                 if ( $include === self::SUPPORTED ) {
 903                         $namesMwFile = [];
 904                         # We do this using a foreach over the codes instead of a directory
 905                         # loop so that messages files in extensions will work correctly.
 906                         foreach ( $returnMw as $code => $value ) {
 907                                 if ( is_readable( self::getMessagesFileName( $code ) )
 908                                         || is_readable( self::getJsonMessagesFileName( $code ) )
 909                                 ) {
 910                                         $namesMwFile[$code] = $names[$code];
 911                                 }
 912                         }
 913
 914                         ksort( $namesMwFile );
 915                         return $namesMwFile;
 916                 }
 917
 918                 ksort( $returnMw );
 919                 # 'mw' option; default if it's not one of the other two options (all/mwfile)
 920                 return $returnMw;
 921         }
 922
 923         /**
 924          * @param string $code The code of the language for which to get the name
 925          * @param null|string $inLanguage Code of language in which to return the name
 926          *   (SELF::AS_AUTONYMS for autonyms)
 927          * @param string $include See fetchLanguageNames()
 928          * @return string Language name or empty
 929          * @since 1.20
 930          */
 931         public static function fetchLanguageName(
 932                 $code,
 933                 $inLanguage = self::AS_AUTONYMS,
 934                 $include = self::ALL
 935         ) {
 936                 $code = strtolower( $code );
 937                 $array = self::fetchLanguageNames( $inLanguage, $include );
 938                 return !array_key_exists( $code, $array ) ? '' : $array[$code];
 939         }
 940
 941         /**
 942          * Get a message from the MediaWiki namespace.
 943          *
 944          * @param string $msg Message name
 945          * @return string
 946          */
 947         public function getMessageFromDB( $msg ) {
 948                 return $this->msg( $msg )->text();
 949         }
 950
 951         /**
 952          * Get message object in this language. Only for use inside this class.
 953          *
 954          * @param string $msg Message name
 955          * @return Message
 956          */
 957         protected function msg( $msg ) {
 958                 return wfMessage( $msg )->inLanguage( $this );
 959         }
 960
 961         /**
 962          * @param string $key
 963          * @return string
 964          */
 965         public function getMonthName( $key ) {
 966                 return $this->getMessageFromDB( self::$mMonthMsgs[$key - 1] );
 967         }
 968
 969         /**
 970          * @return array
 971          */
 972         public function getMonthNamesArray() {
 973                 $monthNames = [ '' ];
 974                 for ( $i = 1; $i < 13; $i++ ) {
 975                         $monthNames[] = $this->getMonthName( $i );
 976                 }
 977                 return $monthNames;
 978         }
 979
 980         /**
 981          * @param string $key
 982          * @return string
 983          */
 984         public function getMonthNameGen( $key ) {
 985                 return $this->getMessageFromDB( self::$mMonthGenMsgs[$key - 1] );
 986         }
 987
 988         /**
 989          * @param string $key
 990          * @return string
 991          */
 992         public function getMonthAbbreviation( $key ) {
 993                 return $this->getMessageFromDB( self::$mMonthAbbrevMsgs[$key - 1] );
 994         }
 995
 996         /**
 997          * @return array
 998          */
 999         public function getMonthAbbreviationsArray() {
1000                 $monthNames = [ '' ];
1001                 for ( $i = 1; $i < 13; $i++ ) {
1002                         $monthNames[] = $this->getMonthAbbreviation( $i );
1003                 }
1004                 return $monthNames;
1005         }
1006
1007         /**
1008          * @param string $key
1009          * @return string
1010          */
1011         public function getWeekdayName( $key ) {
1012                 return $this->getMessageFromDB( self::$mWeekdayMsgs[$key - 1] );
1013         }
1014
1015         /**
1016          * @param string $key
1017          * @return string
1018          */
1019         function getWeekdayAbbreviation( $key ) {
1020                 return $this->getMessageFromDB( self::$mWeekdayAbbrevMsgs[$key - 1] );
1021         }
1022
1023         /**
1024          * @param string $key
1025          * @return string
1026          */
1027         function getIranianCalendarMonthName( $key ) {
1028                 return $this->getMessageFromDB( self::$mIranianCalendarMonthMsgs[$key - 1] );
1029         }
1030
1031         /**
1032          * @param string $key
1033          * @return string
1034          */
1035         function getHebrewCalendarMonthName( $key ) {
1036                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthMsgs[$key - 1] );
1037         }
1038
1039         /**
1040          * @param string $key
1041          * @return string
1042          */
1043         function getHebrewCalendarMonthNameGen( $key ) {
1044                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthGenMsgs[$key - 1] );
1045         }
1046
1047         /**
1048          * @param string $key
1049          * @return string
1050          */
1051         function getHijriCalendarMonthName( $key ) {
1052                 return $this->getMessageFromDB( self::$mHijriCalendarMonthMsgs[$key - 1] );
1053         }
1054
1055         /**
1056          * Pass through result from $dateTimeObj->format()
1057          * @param DateTime|bool|null &$dateTimeObj
1058          * @param string $ts
1059          * @param DateTimeZone|bool|null $zone
1060          * @param string $code
1061          * @return string
1062          */
1063         private static function dateTimeObjFormat( &$dateTimeObj, $ts, $zone, $code ) {
1064                 if ( !$dateTimeObj ) {
1065                         $dateTimeObj = DateTime::createFromFormat(
1066                                 'YmdHis', $ts, $zone ?: new DateTimeZone( 'UTC' )
1067                         );
1068                 }
1069                 return $dateTimeObj->format( $code );
1070         }
1071
1072         /**
1073          * This is a workalike of PHP's date() function, but with better
1074          * internationalisation, a reduced set of format characters, and a better
1075          * escaping format.
1076          *
1077          * Supported format characters are dDjlNwzWFmMntLoYyaAgGhHiscrUeIOPTZ. See
1078          * the PHP manual for definitions. There are a number of extensions, which
1079          * start with "x":
1080          *
1081          *    xn   Do not translate digits of the next numeric format character
1082          *    xN   Toggle raw digit (xn) flag, stays set until explicitly unset
1083          *    xr   Use roman numerals for the next numeric format character
1084          *    xh   Use hebrew numerals for the next numeric format character
1085          *    xx   Literal x
1086          *    xg   Genitive month name
1087          *
1088          *    xij  j (day number) in Iranian calendar
1089          *    xiF  F (month name) in Iranian calendar
1090          *    xin  n (month number) in Iranian calendar
1091          *    xiy  y (two digit year) in Iranian calendar
1092          *    xiY  Y (full year) in Iranian calendar
1093          *    xit  t (days in month) in Iranian calendar
1094          *    xiz  z (day of the year) in Iranian calendar
1095          *
1096          *    xjj  j (day number) in Hebrew calendar
1097          *    xjF  F (month name) in Hebrew calendar
1098          *    xjt  t (days in month) in Hebrew calendar
1099          *    xjx  xg (genitive month name) in Hebrew calendar
1100          *    xjn  n (month number) in Hebrew calendar
1101          *    xjY  Y (full year) in Hebrew calendar
1102          *
1103          *    xmj  j (day number) in Hijri calendar
1104          *    xmF  F (month name) in Hijri calendar
1105          *    xmn  n (month number) in Hijri calendar
1106          *    xmY  Y (full year) in Hijri calendar
1107          *
1108          *    xkY  Y (full year) in Thai solar calendar. Months and days are
1109          *                       identical to the Gregorian calendar
1110          *    xoY  Y (full year) in Minguo calendar or Juche year.
1111          *                       Months and days are identical to the
1112          *                       Gregorian calendar
1113          *    xtY  Y (full year) in Japanese nengo. Months and days are
1114          *                       identical to the Gregorian calendar
1115          *
1116          * Characters enclosed in double quotes will be considered literal (with
1117          * the quotes themselves removed). Unmatched quotes will be considered
1118          * literal quotes. Example:
1119          *
1120          * "The month is" F       => The month is January
1121          * i's"                   => 20'11"
1122          *
1123          * Backslash escaping is also supported.
1124          *
1125          * Input timestamp is assumed to be pre-normalized to the desired local
1126          * time zone, if any. Note that the format characters crUeIOPTZ will assume
1127          * $ts is UTC if $zone is not given.
1128          *
1129          * @param string $format
1130          * @param string $ts 14-character timestamp
1131          *      YYYYMMDDHHMMSS
1132          *      01234567890123
1133          * @param DateTimeZone|null $zone Timezone of $ts
1134          * @param int &$ttl The amount of time (in seconds) the output may be cached for.
1135          * Only makes sense if $ts is the current time.
1136          * @todo handling of "o" format character for Iranian, Hebrew, Hijri & Thai?
1137          *
1138          * @throws MWException
1139          * @return string
1140          */
1141         public function sprintfDate( $format, $ts, DateTimeZone $zone = null, &$ttl = 'unused' ) {
1142                 $s = '';
1143                 $raw = false;
1144                 $roman = false;
1145                 $hebrewNum = false;
1146                 $dateTimeObj = false;
1147                 $rawToggle = false;
1148                 $iranian = false;
1149                 $hebrew = false;
1150                 $hijri = false;
1151                 $thai = false;
1152                 $minguo = false;
1153                 $tenno = false;
1154
1155                 $usedSecond = false;
1156                 $usedMinute = false;
1157                 $usedHour = false;
1158                 $usedAMPM = false;
1159                 $usedDay = false;
1160                 $usedWeek = false;
1161                 $usedMonth = false;
1162                 $usedYear = false;
1163                 $usedISOYear = false;
1164                 $usedIsLeapYear = false;
1165
1166                 $usedHebrewMonth = false;
1167                 $usedIranianMonth = false;
1168                 $usedHijriMonth = false;
1169                 $usedHebrewYear = false;
1170                 $usedIranianYear = false;
1171                 $usedHijriYear = false;
1172                 $usedTennoYear = false;
1173
1174                 if ( strlen( $ts ) !== 14 ) {
1175                         throw new MWException( __METHOD__ . ": The timestamp $ts should have 14 characters" );
1176                 }
1177
1178                 if ( !ctype_digit( $ts ) ) {
1179                         throw new MWException( __METHOD__ . ": The timestamp $ts should be a number" );
1180                 }
1181
1182                 $formatLength = strlen( $format );
1183                 for ( $p = 0; $p < $formatLength; $p++ ) {
1184                         $num = false;
1185                         $code = $format[$p];
1186                         if ( $code == 'x' && $p < $formatLength - 1 ) {
1187                                 $code .= $format[++$p];
1188                         }
1189
1190                         if ( ( $code === 'xi'
1191                                         || $code === 'xj'
1192                                         || $code === 'xk'
1193                                         || $code === 'xm'
1194                                         || $code === 'xo'
1195                                         || $code === 'xt' )
1196                                 && $p < $formatLength - 1 ) {
1197                                 $code .= $format[++$p];
1198                         }
1199
1200                         switch ( $code ) {
1201                                 case 'xx':
1202                                         $s .= 'x';
1203                                         break;
1204                                 case 'xn':
1205                                         $raw = true;
1206                                         break;
1207                                 case 'xN':
1208                                         $rawToggle = !$rawToggle;
1209                                         break;
1210                                 case 'xr':
1211                                         $roman = true;
1212                                         break;
1213                                 case 'xh':
1214                                         $hebrewNum = true;
1215                                         break;
1216                                 case 'xg':
1217                                         $usedMonth = true;
1218                                         $s .= $this->getMonthNameGen( substr( $ts, 4, 2 ) );
1219                                         break;
1220                                 case 'xjx':
1221                                         $usedHebrewMonth = true;
1222                                         if ( !$hebrew ) {
1223                                                 $hebrew = self::tsToHebrew( $ts );
1224                                         }
1225                                         $s .= $this->getHebrewCalendarMonthNameGen( $hebrew[1] );
1226                                         break;
1227                                 case 'd':
1228                                         $usedDay = true;
1229                                         $num = substr( $ts, 6, 2 );
1230                                         break;
1231                                 case 'D':
1232                                         $usedDay = true;
1233                                         $s .= $this->getWeekdayAbbreviation(
1234                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1235                                         );
1236                                         break;
1237                                 case 'j':
1238                                         $usedDay = true;
1239                                         $num = intval( substr( $ts, 6, 2 ) );
1240                                         break;
1241                                 case 'xij':
1242                                         $usedDay = true;
1243                                         if ( !$iranian ) {
1244                                                 $iranian = self::tsToIranian( $ts );
1245                                         }
1246                                         $num = $iranian[2];
1247                                         break;
1248                                 case 'xmj':
1249                                         $usedDay = true;
1250                                         if ( !$hijri ) {
1251                                                 $hijri = self::tsToHijri( $ts );
1252                                         }
1253                                         $num = $hijri[2];
1254                                         break;
1255                                 case 'xjj':
1256                                         $usedDay = true;
1257                                         if ( !$hebrew ) {
1258                                                 $hebrew = self::tsToHebrew( $ts );
1259                                         }
1260                                         $num = $hebrew[2];
1261                                         break;
1262                                 case 'l':
1263                                         $usedDay = true;
1264                                         $s .= $this->getWeekdayName(
1265                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1266                                         );
1267                                         break;
1268                                 case 'F':
1269                                         $usedMonth = true;
1270                                         $s .= $this->getMonthName( substr( $ts, 4, 2 ) );
1271                                         break;
1272                                 case 'xiF':
1273                                         $usedIranianMonth = true;
1274                                         if ( !$iranian ) {
1275                                                 $iranian = self::tsToIranian( $ts );
1276                                         }
1277                                         $s .= $this->getIranianCalendarMonthName( $iranian[1] );
1278                                         break;
1279                                 case 'xmF':
1280                                         $usedHijriMonth = true;
1281                                         if ( !$hijri ) {
1282                                                 $hijri = self::tsToHijri( $ts );
1283                                         }
1284                                         $s .= $this->getHijriCalendarMonthName( $hijri[1] );
1285                                         break;
1286                                 case 'xjF':
1287                                         $usedHebrewMonth = true;
1288                                         if ( !$hebrew ) {
1289                                                 $hebrew = self::tsToHebrew( $ts );
1290                                         }
1291                                         $s .= $this->getHebrewCalendarMonthName( $hebrew[1] );
1292                                         break;
1293                                 case 'm':
1294                                         $usedMonth = true;
1295                                         $num = substr( $ts, 4, 2 );
1296                                         break;
1297                                 case 'M':
1298                                         $usedMonth = true;
1299                                         $s .= $this->getMonthAbbreviation( substr( $ts, 4, 2 ) );
1300                                         break;
1301                                 case 'n':
1302                                         $usedMonth = true;
1303                                         $num = intval( substr( $ts, 4, 2 ) );
1304                                         break;
1305                                 case 'xin':
1306                                         $usedIranianMonth = true;
1307                                         if ( !$iranian ) {
1308                                                 $iranian = self::tsToIranian( $ts );
1309                                         }
1310                                         $num = $iranian[1];
1311                                         break;
1312                                 case 'xmn':
1313                                         $usedHijriMonth = true;
1314                                         if ( !$hijri ) {
1315                                                 $hijri = self::tsToHijri( $ts );
1316                                         }
1317                                         $num = $hijri[1];
1318                                         break;
1319                                 case 'xjn':
1320                                         $usedHebrewMonth = true;
1321                                         if ( !$hebrew ) {
1322                                                 $hebrew = self::tsToHebrew( $ts );
1323                                         }
1324                                         $num = $hebrew[1];
1325                                         break;
1326                                 case 'xjt':
1327                                         $usedHebrewMonth = true;
1328                                         if ( !$hebrew ) {
1329                                                 $hebrew = self::tsToHebrew( $ts );
1330                                         }
1331                                         $num = $hebrew[3];
1332                                         break;
1333                                 case 'Y':
1334                                         $usedYear = true;
1335                                         $num = substr( $ts, 0, 4 );
1336                                         break;
1337                                 case 'xiY':
1338                                         $usedIranianYear = true;
1339                                         if ( !$iranian ) {
1340                                                 $iranian = self::tsToIranian( $ts );
1341                                         }
1342                                         $num = $iranian[0];
1343                                         break;
1344                                 case 'xmY':
1345                                         $usedHijriYear = true;
1346                                         if ( !$hijri ) {
1347                                                 $hijri = self::tsToHijri( $ts );
1348                                         }
1349                                         $num = $hijri[0];
1350                                         break;
1351                                 case 'xjY':
1352                                         $usedHebrewYear = true;
1353                                         if ( !$hebrew ) {
1354                                                 $hebrew = self::tsToHebrew( $ts );
1355                                         }
1356                                         $num = $hebrew[0];
1357                                         break;
1358                                 case 'xkY':
1359                                         $usedYear = true;
1360                                         if ( !$thai ) {
1361                                                 $thai = self::tsToYear( $ts, 'thai' );
1362                                         }
1363                                         $num = $thai[0];
1364                                         break;
1365                                 case 'xoY':
1366                                         $usedYear = true;
1367                                         if ( !$minguo ) {
1368                                                 $minguo = self::tsToYear( $ts, 'minguo' );
1369                                         }
1370                                         $num = $minguo[0];
1371                                         break;
1372                                 case 'xtY':
1373                                         $usedTennoYear = true;
1374                                         if ( !$tenno ) {
1375                                                 $tenno = self::tsToYear( $ts, 'tenno' );
1376                                         }
1377                                         $num = $tenno[0];
1378                                         break;
1379                                 case 'y':
1380                                         $usedYear = true;
1381                                         $num = substr( $ts, 2, 2 );
1382                                         break;
1383                                 case 'xiy':
1384                                         $usedIranianYear = true;
1385                                         if ( !$iranian ) {
1386                                                 $iranian = self::tsToIranian( $ts );
1387                                         }
1388                                         $num = substr( $iranian[0], -2 );
1389                                         break;
1390                                 case 'xit':
1391                                         $usedIranianYear = true;
1392                                         if ( !$iranian ) {
1393                                                 $iranian = self::tsToIranian( $ts );
1394                                         }
1395                                         $num = self::$IRANIAN_DAYS[$iranian[1] - 1];
1396                                         break;
1397                                 case 'xiz':
1398                                         $usedIranianYear = true;
1399                                         if ( !$iranian ) {
1400                                                 $iranian = self::tsToIranian( $ts );
1401                                         }
1402                                         $num = $iranian[3];
1403                                         break;
1404                                 case 'a':
1405                                         $usedAMPM = true;
1406                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'am' : 'pm';
1407                                         break;
1408                                 case 'A':
1409                                         $usedAMPM = true;
1410                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'AM' : 'PM';
1411                                         break;
1412                                 case 'g':
1413                                         $usedHour = true;
1414                                         $h = substr( $ts, 8, 2 );
1415                                         $num = $h % 12 ?: 12;
1416                                         break;
1417                                 case 'G':
1418                                         $usedHour = true;
1419                                         $num = intval( substr( $ts, 8, 2 ) );
1420                                         break;
1421                                 case 'h':
1422                                         $usedHour = true;
1423                                         $h = substr( $ts, 8, 2 );
1424                                         $num = sprintf( '%02d', $h % 12 ?: 12 );
1425                                         break;
1426                                 case 'H':
1427                                         $usedHour = true;
1428                                         $num = substr( $ts, 8, 2 );
1429                                         break;
1430                                 case 'i':
1431                                         $usedMinute = true;
1432                                         $num = substr( $ts, 10, 2 );
1433                                         break;
1434                                 case 's':
1435                                         $usedSecond = true;
1436                                         $num = substr( $ts, 12, 2 );
1437                                         break;
1438                                 case 'c':
1439                                 case 'r':
1440                                         $usedSecond = true;
1441                                         // fall through
1442                                 case 'e':
1443                                 case 'O':
1444                                 case 'P':
1445                                 case 'T':
1446                                         $s .= self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1447                                         break;
1448                                 case 'w':
1449                                 case 'N':
1450                                 case 'z':
1451                                         $usedDay = true;
1452                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1453                                         break;
1454                                 case 'W':
1455                                         $usedWeek = true;
1456                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1457                                         break;
1458                                 case 't':
1459                                         $usedMonth = true;
1460                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1461                                         break;
1462                                 case 'L':
1463                                         $usedIsLeapYear = true;
1464                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1465                                         break;
1466                                 case 'o':
1467                                         $usedISOYear = true;
1468                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1469                                         break;
1470                                 case 'U':
1471                                         $usedSecond = true;
1472                                         // fall through
1473                                 case 'I':
1474                                 case 'Z':
1475                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1476                                         break;
1477                                 case '\\':
1478                                         # Backslash escaping
1479                                         if ( $p < $formatLength - 1 ) {
1480                                                 $s .= $format[++$p];
1481                                         } else {
1482                                                 $s .= '\\';
1483                                         }
1484                                         break;
1485                                 case '"':
1486                                         # Quoted literal
1487                                         if ( $p < $formatLength - 1 ) {
1488                                                 $endQuote = strpos( $format, '"', $p + 1 );
1489                                                 if ( $endQuote === false ) {
1490                                                         # No terminating quote, assume literal "
1491                                                         $s .= '"';
1492                                                 } else {
1493                                                         $s .= substr( $format, $p + 1, $endQuote - $p - 1 );
1494                                                         $p = $endQuote;
1495                                                 }
1496                                         } else {
1497                                                 # Quote at end of string, assume literal "
1498                                                 $s .= '"';
1499                                         }
1500                                         break;
1501                                 default:
1502                                         $s .= $format[$p];
1503                         }
1504                         if ( $num !== false ) {
1505                                 if ( $rawToggle || $raw ) {
1506                                         $s .= $num;
1507                                         $raw = false;
1508                                 } elseif ( $roman ) {
1509                                         $s .= self::romanNumeral( $num );
1510                                         $roman = false;
1511                                 } elseif ( $hebrewNum ) {
1512                                         $s .= self::hebrewNumeral( $num );
1513                                         $hebrewNum = false;
1514                                 } else {
1515                                         $s .= $this->formatNum( $num, true );
1516                                 }
1517                         }
1518                 }
1519
1520                 if ( $ttl === 'unused' ) {
1521                         // No need to calculate the TTL, the caller wont use it anyway.
1522                 } elseif ( $usedSecond ) {
1523                         $ttl = 1;
1524                 } elseif ( $usedMinute ) {
1525                         $ttl = 60 - substr( $ts, 12, 2 );
1526                 } elseif ( $usedHour ) {
1527                         $ttl = 3600 - substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1528                 } elseif ( $usedAMPM ) {
1529                         $ttl = 43200 - ( substr( $ts, 8, 2 ) % 12 ) * 3600 -
1530                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1531                 } elseif (
1532                         $usedDay ||
1533                         $usedHebrewMonth ||
1534                         $usedIranianMonth ||
1535                         $usedHijriMonth ||
1536                         $usedHebrewYear ||
1537                         $usedIranianYear ||
1538                         $usedHijriYear ||
1539                         $usedTennoYear
1540                 ) {
1541                         // @todo Someone who understands the non-Gregorian calendars
1542                         // should write proper logic for them so that they don't need purged every day.
1543                         $ttl = 86400 - substr( $ts, 8, 2 ) * 3600 -
1544                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1545                 } else {
1546                         $possibleTtls = [];
1547                         $timeRemainingInDay = 86400 - substr( $ts, 8, 2 ) * 3600 -
1548                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1549                         if ( $usedWeek ) {
1550                                 $possibleTtls[] =
1551                                         ( 7 - self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400 +
1552                                         $timeRemainingInDay;
1553                         } elseif ( $usedISOYear ) {
1554                                 // December 28th falls on the last ISO week of the year, every year.
1555                                 // The last ISO week of a year can be 52 or 53.
1556                                 $lastWeekOfISOYear = DateTime::createFromFormat(
1557                                         'Ymd',
1558                                         substr( $ts, 0, 4 ) . '1228',
1559                                         $zone ?: new DateTimeZone( 'UTC' )
1560                                 )->format( 'W' );
1561                                 $currentISOWeek = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'W' );
1562                                 $weeksRemaining = $lastWeekOfISOYear - $currentISOWeek;
1563                                 $timeRemainingInWeek =
1564                                         ( 7 - self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400
1565                                         + $timeRemainingInDay;
1566                                 $possibleTtls[] = $weeksRemaining * 604800 + $timeRemainingInWeek;
1567                         }
1568
1569                         if ( $usedMonth ) {
1570                                 $possibleTtls[] =
1571                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 't' ) -
1572                                                 substr( $ts, 6, 2 ) ) * 86400
1573                                         + $timeRemainingInDay;
1574                         } elseif ( $usedYear ) {
1575                                 $possibleTtls[] =
1576                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1577                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1578                                         + $timeRemainingInDay;
1579                         } elseif ( $usedIsLeapYear ) {
1580                                 $year = substr( $ts, 0, 4 );
1581                                 $timeRemainingInYear =
1582                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1583                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1584                                         + $timeRemainingInDay;
1585                                 $mod = $year % 4;
1586                                 if ( $mod || ( !( $year % 100 ) && $year % 400 ) ) {
1587                                         // this isn't a leap year. see when the next one starts
1588                                         $nextCandidate = $year - $mod + 4;
1589                                         if ( $nextCandidate % 100 || !( $nextCandidate % 400 ) ) {
1590                                                 $possibleTtls[] = ( $nextCandidate - $year - 1 ) * 365 * 86400 +
1591                                                         $timeRemainingInYear;
1592                                         } else {
1593                                                 $possibleTtls[] = ( $nextCandidate - $year + 3 ) * 365 * 86400 +
1594                                                         $timeRemainingInYear;
1595                                         }
1596                                 } else {
1597                                         // this is a leap year, so the next year isn't
1598                                         $possibleTtls[] = $timeRemainingInYear;
1599                                 }
1600                         }
1601
1602                         if ( $possibleTtls ) {
1603                                 $ttl = min( $possibleTtls );
1604                         }
1605                 }
1606
1607                 return $s;
1608         }
1609
1610         private static $GREG_DAYS = [ 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 ];
1611         private static $IRANIAN_DAYS = [ 31, 31, 31, 31, 31, 31, 30, 30, 30, 30, 30, 29 ];
1612
1613         /**
1614          * Algorithm by Roozbeh Pournader and Mohammad Toossi to convert
1615          * Gregorian dates to Iranian dates. Originally written in C, it
1616          * is released under the terms of GNU Lesser General Public
1617          * License. Conversion to PHP was performed by Niklas Laxström.
1618          *
1619          * Link: http://www.farsiweb.info/jalali/jalali.c
1620          *
1621          * @param string $ts
1622          *
1623          * @return int[]
1624          */
1625         private static function tsToIranian( $ts ) {
1626                 $gy = substr( $ts, 0, 4 ) - 1600;
1627                 $gm = substr( $ts, 4, 2 ) - 1;
1628                 $gd = substr( $ts, 6, 2 ) - 1;
1629
1630                 # Days passed from the beginning (including leap years)
1631                 $gDayNo = 365 * $gy
1632                         + floor( ( $gy + 3 ) / 4 )
1633                         - floor( ( $gy + 99 ) / 100 )
1634                         + floor( ( $gy + 399 ) / 400 );
1635
1636                 // Add days of the past months of this year
1637                 for ( $i = 0; $i < $gm; $i++ ) {
1638                         $gDayNo += self::$GREG_DAYS[$i];
1639                 }
1640
1641                 // Leap years
1642                 if ( $gm > 1 && ( ( $gy % 4 === 0 && $gy % 100 !== 0 || ( $gy % 400 == 0 ) ) ) ) {
1643                         $gDayNo++;
1644                 }
1645
1646                 // Days passed in current month
1647                 $gDayNo += (int)$gd;
1648
1649                 $jDayNo = $gDayNo - 79;
1650
1651                 $jNp = floor( $jDayNo / 12053 );
1652                 $jDayNo %= 12053;
1653
1654                 $jy = 979 + 33 * $jNp + 4 * floor( $jDayNo / 1461 );
1655                 $jDayNo %= 1461;
1656
1657                 if ( $jDayNo >= 366 ) {
1658                         $jy += floor( ( $jDayNo - 1 ) / 365 );
1659                         $jDayNo = floor( ( $jDayNo - 1 ) % 365 );
1660                 }
1661
1662                 $jz = $jDayNo;
1663
1664                 for ( $i = 0; $i < 11 && $jDayNo >= self::$IRANIAN_DAYS[$i]; $i++ ) {
1665                         $jDayNo -= self::$IRANIAN_DAYS[$i];
1666                 }
1667
1668                 $jm = $i + 1;
1669                 $jd = $jDayNo + 1;
1670
1671                 return [ $jy, $jm, $jd, $jz ];
1672         }
1673
1674         /**
1675          * Converting Gregorian dates to Hijri dates.
1676          *
1677          * Based on a PHP-Nuke block by Sharjeel which is released under GNU/GPL license
1678          *
1679          * @see https://phpnuke.org/modules.php?name=News&file=article&sid=8234&mode=thread&order=0&thold=0
1680          *
1681          * @param string $ts
1682          *
1683          * @return int[]
1684          */
1685         private static function tsToHijri( $ts ) {
1686                 $year = substr( $ts, 0, 4 );
1687                 $month = substr( $ts, 4, 2 );
1688                 $day = substr( $ts, 6, 2 );
1689
1690                 $zyr = $year;
1691                 $zd = $day;
1692                 $zm = $month;
1693                 $zy = $zyr;
1694
1695                 if (
1696                         ( $zy > 1582 ) || ( ( $zy == 1582 ) && ( $zm > 10 ) ) ||
1697                         ( ( $zy == 1582 ) && ( $zm == 10 ) && ( $zd > 14 ) )
1698                 ) {
1699                         $zjd = (int)( ( 1461 * ( $zy + 4800 + (int)( ( $zm - 14 ) / 12 ) ) ) / 4 ) +
1700                                         (int)( ( 367 * ( $zm - 2 - 12 * ( (int)( ( $zm - 14 ) / 12 ) ) ) ) / 12 ) -
1701                                         (int)( ( 3 * (int)( ( ( $zy + 4900 + (int)( ( $zm - 14 ) / 12 ) ) / 100 ) ) ) / 4 ) +
1702                                         $zd - 32075;
1703                 } else {
1704                         $zjd = 367 * $zy - (int)( ( 7 * ( $zy + 5001 + (int)( ( $zm - 9 ) / 7 ) ) ) / 4 ) +
1705                                                                 (int)( ( 275 * $zm ) / 9 ) + $zd + 1729777;
1706                 }
1707
1708                 $zl = $zjd - 1948440 + 10632;
1709                 $zn = (int)( ( $zl - 1 ) / 10631 );
1710                 $zl = $zl - 10631 * $zn + 354;
1711                 $zj = ( (int)( ( 10985 - $zl ) / 5316 ) ) * ( (int)( ( 50 * $zl ) / 17719 ) ) +
1712                         ( (int)( $zl / 5670 ) ) * ( (int)( ( 43 * $zl ) / 15238 ) );
1713                 $zl = $zl - ( (int)( ( 30 - $zj ) / 15 ) ) * ( (int)( ( 17719 * $zj ) / 50 ) ) -
1714                         ( (int)( $zj / 16 ) ) * ( (int)( ( 15238 * $zj ) / 43 ) ) + 29;
1715                 $zm = (int)( ( 24 * $zl ) / 709 );
1716                 $zd = $zl - (int)( ( 709 * $zm ) / 24 );
1717                 $zy = 30 * $zn + $zj - 30;
1718
1719                 return [ $zy, $zm, $zd ];
1720         }
1721
1722         /**
1723          * Converting Gregorian dates to Hebrew dates.
1724          *
1725          * Based on a JavaScript code by Abu Mami and Yisrael Hersch
1726          * (abu-mami@kaluach.net, http://www.kaluach.net), who permitted
1727          * to translate the relevant functions into PHP and release them under
1728          * GNU GPL.
1729          *
1730          * The months are counted from Tishrei = 1. In a leap year, Adar I is 13
1731          * and Adar II is 14. In a non-leap year, Adar is 6.
1732          *
1733          * @param string $ts
1734          *
1735          * @return int[]
1736          */
1737         private static function tsToHebrew( $ts ) {
1738                 # Parse date
1739                 $year = substr( $ts, 0, 4 );
1740                 $month = substr( $ts, 4, 2 );
1741                 $day = substr( $ts, 6, 2 );
1742
1743                 # Calculate Hebrew year
1744                 $hebrewYear = $year + 3760;
1745
1746                 # Month number when September = 1, August = 12
1747                 $month += 4;
1748                 if ( $month > 12 ) {
1749                         # Next year
1750                         $month -= 12;
1751                         $year++;
1752                         $hebrewYear++;
1753                 }
1754
1755                 # Calculate day of year from 1 September
1756                 $dayOfYear = $day;
1757                 for ( $i = 1; $i < $month; $i++ ) {
1758                         if ( $i == 6 ) {
1759                                 # February
1760                                 $dayOfYear += 28;
1761                                 # Check if the year is leap
1762                                 if ( $year % 400 == 0 || ( $year % 4 == 0 && $year % 100 > 0 ) ) {
1763                                         $dayOfYear++;
1764                                 }
1765                         } elseif ( $i == 8 || $i == 10 || $i == 1 || $i == 3 ) {
1766                                 $dayOfYear += 30;
1767                         } else {
1768                                 $dayOfYear += 31;
1769                         }
1770                 }
1771
1772                 # Calculate the start of the Hebrew year
1773                 $start = self::hebrewYearStart( $hebrewYear );
1774
1775                 # Calculate next year's start
1776                 if ( $dayOfYear <= $start ) {
1777                         # Day is before the start of the year - it is the previous year
1778                         # Next year's start
1779                         $nextStart = $start;
1780                         # Previous year
1781                         $year--;
1782                         $hebrewYear--;
1783                         # Add days since previous year's 1 September
1784                         $dayOfYear += 365;
1785                         if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1786                                 # Leap year
1787                                 $dayOfYear++;
1788                         }
1789                         # Start of the new (previous) year
1790                         $start = self::hebrewYearStart( $hebrewYear );
1791                 } else {
1792                         # Next year's start
1793                         $nextStart = self::hebrewYearStart( $hebrewYear + 1 );
1794                 }
1795
1796                 # Calculate Hebrew day of year
1797                 $hebrewDayOfYear = $dayOfYear - $start;
1798
1799                 # Difference between year's days
1800                 $diff = $nextStart - $start;
1801                 # Add 12 (or 13 for leap years) days to ignore the difference between
1802                 # Hebrew and Gregorian year (353 at least vs. 365/6) - now the
1803                 # difference is only about the year type
1804                 if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1805                         $diff += 13;
1806                 } else {
1807                         $diff += 12;
1808                 }
1809
1810                 # Check the year pattern, and is leap year
1811                 # 0 means an incomplete year, 1 means a regular year, 2 means a complete year
1812                 # This is mod 30, to work on both leap years (which add 30 days of Adar I)
1813                 # and non-leap years
1814                 $yearPattern = $diff % 30;
1815                 # Check if leap year
1816                 $isLeap = $diff >= 30;
1817
1818                 # Calculate day in the month from number of day in the Hebrew year
1819                 # Don't check Adar - if the day is not in Adar, we will stop before;
1820                 # if it is in Adar, we will use it to check if it is Adar I or Adar II
1821                 $hebrewDay = $hebrewDayOfYear;
1822                 $hebrewMonth = 1;
1823                 $days = 0;
1824                 while ( $hebrewMonth <= 12 ) {
1825                         # Calculate days in this month
1826                         if ( $isLeap && $hebrewMonth == 6 ) {
1827                                 # Leap year - has Adar I, with 30 days, and Adar II, with 29 days
1828                                 $days = 30;
1829                                 if ( $hebrewDay <= $days ) {
1830                                         # Day in Adar I
1831                                         $hebrewMonth = 13;
1832                                 } else {
1833                                         # Subtract the days of Adar I
1834                                         $hebrewDay -= $days;
1835                                         # Try Adar II
1836                                         $days = 29;
1837                                         if ( $hebrewDay <= $days ) {
1838                                                 # Day in Adar II
1839                                                 $hebrewMonth = 14;
1840                                         }
1841                                 }
1842                         } elseif ( $hebrewMonth == 2 && $yearPattern == 2 ) {
1843                                 # Cheshvan in a complete year (otherwise as the rule below)
1844                                 $days = 30;
1845                         } elseif ( $hebrewMonth == 3 && $yearPattern == 0 ) {
1846                                 # Kislev in an incomplete year (otherwise as the rule below)
1847                                 $days = 29;
1848                         } else {
1849                                 # Odd months have 30 days, even have 29
1850                                 $days = 30 - ( $hebrewMonth - 1 ) % 2;
1851                         }
1852                         if ( $hebrewDay <= $days ) {
1853                                 # In the current month
1854                                 break;
1855                         } else {
1856                                 # Subtract the days of the current month
1857                                 $hebrewDay -= $days;
1858                                 # Try in the next month
1859                                 $hebrewMonth++;
1860                         }
1861                 }
1862
1863                 return [ $hebrewYear, $hebrewMonth, $hebrewDay, $days ];
1864         }
1865
1866         /**
1867          * This calculates the Hebrew year start, as days since 1 September.
1868          * Based on Carl Friedrich Gauss algorithm for finding Easter date.
1869          * Used for Hebrew date.
1870          *
1871          * @param int $year
1872          *
1873          * @return string
1874          */
1875         private static function hebrewYearStart( $year ) {
1876                 $a = intval( ( 12 * ( $year - 1 ) + 17 ) % 19 );
1877                 $b = intval( ( $year - 1 ) % 4 );
1878                 $m = 32.044093161144 + 1.5542417966212 * $a + $b / 4.0 - 0.0031777940220923 * ( $year - 1 );
1879                 if ( $m < 0 ) {
1880                         $m--;
1881                 }
1882                 $Mar = intval( $m );
1883                 if ( $m < 0 ) {
1884                         $m++;
1885                 }
1886                 $m -= $Mar;
1887
1888                 $c = intval( ( $Mar + 3 * ( $year - 1 ) + 5 * $b + 5 ) % 7 );
1889                 if ( $c == 0 && $a > 11 && $m >= 0.89772376543210 ) {
1890                         $Mar++;
1891                 } elseif ( $c == 1 && $a > 6 && $m >= 0.63287037037037 ) {
1892                         $Mar += 2;
1893                 } elseif ( $c == 2 || $c == 4 || $c == 6 ) {
1894                         $Mar++;
1895                 }
1896
1897                 $Mar += intval( ( $year - 3761 ) / 100 ) - intval( ( $year - 3761 ) / 400 ) - 24;
1898                 return $Mar;
1899         }
1900
1901         /**
1902          * Algorithm to convert Gregorian dates to Thai solar dates,
1903          * Minguo dates or Minguo dates.
1904          *
1905          * Link: https://en.wikipedia.org/wiki/Thai_solar_calendar
1906          *       https://en.wikipedia.org/wiki/Minguo_calendar
1907          *       https://en.wikipedia.org/wiki/Japanese_era_name
1908          *
1909          * @param string $ts 14-character timestamp
1910          * @param string $cName Calender name
1911          * @return array Converted year, month, day
1912          */
1913         private static function tsToYear( $ts, $cName ) {
1914                 $gy = substr( $ts, 0, 4 );
1915                 $gm = substr( $ts, 4, 2 );
1916                 $gd = substr( $ts, 6, 2 );
1917
1918                 if ( !strcmp( $cName, 'thai' ) ) {
1919                         # Thai solar dates
1920                         # Add 543 years to the Gregorian calendar
1921                         # Months and days are identical
1922                         $gy_offset = $gy + 543;
1923                         # fix for dates between 1912 and 1941
1924                         # https://en.wikipedia.org/?oldid=836596673#New_year
1925                         if ( $gy >= 1912 && $gy <= 1940 ) {
1926                                 if ( $gm <= 3 ) {
1927                                         $gy_offset--;
1928                                 }
1929                                 $gm = ( $gm - 3 ) % 12;
1930                         }
1931                 } elseif ( ( !strcmp( $cName, 'minguo' ) ) || !strcmp( $cName, 'juche' ) ) {
1932                         # Minguo dates
1933                         # Deduct 1911 years from the Gregorian calendar
1934                         # Months and days are identical
1935                         $gy_offset = $gy - 1911;
1936                 } elseif ( !strcmp( $cName, 'tenno' ) ) {
1937                         # Nengō dates up to Meiji period
1938                         # Deduct years from the Gregorian calendar
1939                         # depending on the nengo periods
1940                         # Months and days are identical
1941                         if ( ( $gy < 1912 )
1942                                 || ( ( $gy == 1912 ) && ( $gm < 7 ) )
1943                                 || ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd < 31 ) )
1944                         ) {
1945                                 # Meiji period
1946                                 $gy_gannen = $gy - 1868 + 1;
1947                                 $gy_offset = $gy_gannen;
1948                                 if ( $gy_gannen == 1 ) {
1949                                         $gy_offset = '元';
1950                                 }
1951                                 $gy_offset = '明治' . $gy_offset;
1952                         } elseif (
1953                                 ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd == 31 ) ) ||
1954                                 ( ( $gy == 1912 ) && ( $gm >= 8 ) ) ||
1955                                 ( ( $gy > 1912 ) && ( $gy < 1926 ) ) ||
1956                                 ( ( $gy == 1926 ) && ( $gm < 12 ) ) ||
1957                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd < 26 ) )
1958                         ) {
1959                                 # Taishō period
1960                                 $gy_gannen = $gy - 1912 + 1;
1961                                 $gy_offset = $gy_gannen;
1962                                 if ( $gy_gannen == 1 ) {
1963                                         $gy_offset = '元';
1964                                 }
1965                                 $gy_offset = '大正' . $gy_offset;
1966                         } elseif (
1967                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd >= 26 ) ) ||
1968                                 ( ( $gy > 1926 ) && ( $gy < 1989 ) ) ||
1969                                 ( ( $gy == 1989 ) && ( $gm == 1 ) && ( $gd < 8 ) )
1970                         ) {
1971                                 # Shōwa period
1972                                 $gy_gannen = $gy - 1926 + 1;
1973                                 $gy_offset = $gy_gannen;
1974                                 if ( $gy_gannen == 1 ) {
1975                                         $gy_offset = '元';
1976                                 }
1977                                 $gy_offset = '昭和' . $gy_offset;
1978                         } elseif (
1979                                 ( ( $gy == 1989 ) && ( $gm == 1 ) && ( $gd >= 8 ) ) ||
1980                                 ( ( $gy > 1989 ) && ( $gy < 2019 ) ) ||
1981                                 ( ( $gy == 2019 ) && ( $gm < 5 ) )
1982                         ) {
1983                                 # Heisei period
1984                                 $gy_gannen = $gy - 1989 + 1;
1985                                 $gy_offset = $gy_gannen;
1986                                 if ( $gy_gannen == 1 ) {
1987                                         $gy_offset = '元';
1988                                 }
1989                                 $gy_offset = '平成' . $gy_offset;
1990                         } else {
1991                                 # Reiwa period
1992                                 $gy_gannen = $gy - 2019 + 1;
1993                                 $gy_offset = $gy_gannen;
1994                                 if ( $gy_gannen == 1 ) {
1995                                         $gy_offset = '元';
1996                                 }
1997                                 $gy_offset = '令和' . $gy_offset;
1998                         }
1999                 } else {
2000                         $gy_offset = $gy;
2001                 }
2002
2003                 return [ $gy_offset, $gm, $gd ];
2004         }
2005
2006         /**
2007          * Gets directionality of the first strongly directional codepoint, for embedBidi()
2008          *
2009          * This is the rule the BIDI algorithm uses to determine the directionality of
2010          * paragraphs ( https://www.unicode.org/reports/tr9/#The_Paragraph_Level ) and
2011          * FSI isolates ( https://www.unicode.org/reports/tr9/#Explicit_Directional_Isolates ).
2012          *
2013          * TODO: Does not handle BIDI control characters inside the text.
2014          * TODO: Does not handle unallocated characters.
2015          *
2016          * @param string $text Text to test
2017          * @return null|string Directionality ('ltr' or 'rtl') or null
2018          */
2019         private static function strongDirFromContent( $text = '' ) {
2020                 if ( !preg_match( self::$strongDirRegex, $text, $matches ) ) {
2021                         return null;
2022                 }
2023                 if ( $matches[1] === '' ) {
2024                         return 'rtl';
2025                 }
2026                 return 'ltr';
2027         }
2028
2029         /**
2030          * Roman number formatting up to 10000
2031          *
2032          * @param int $num
2033          *
2034          * @return string
2035          */
2036         static function romanNumeral( $num ) {
2037                 static $table = [
2038                         [ '', 'I', 'II', 'III', 'IV', 'V', 'VI', 'VII', 'VIII', 'IX', 'X' ],
2039                         [ '', 'X', 'XX', 'XXX', 'XL', 'L', 'LX', 'LXX', 'LXXX', 'XC', 'C' ],
2040                         [ '', 'C', 'CC', 'CCC', 'CD', 'D', 'DC', 'DCC', 'DCCC', 'CM', 'M' ],
2041                         [ '', 'M', 'MM', 'MMM', 'MMMM', 'MMMMM', 'MMMMMM', 'MMMMMMM',
2042                                 'MMMMMMMM', 'MMMMMMMMM', 'MMMMMMMMMM' ]
2043                 ];
2044
2045                 $num = intval( $num );
2046                 if ( $num > 10000 || $num <= 0 ) {
2047                         return $num;
2048                 }
2049
2050                 $s = '';
2051                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
2052                         if ( $num >= $pow10 ) {
2053                                 $s .= $table[$i][(int)floor( $num / $pow10 )];
2054                         }
2055                         $num = $num % $pow10;
2056                 }
2057                 return $s;
2058         }
2059
2060         /**
2061          * Hebrew Gematria number formatting up to 9999
2062          *
2063          * @param int $num
2064          *
2065          * @return string
2066          */
2067         static function hebrewNumeral( $num ) {
2068                 static $table = [
2069                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ],
2070                         [ '', 'י', 'כ', 'ל', 'מ', 'נ', 'ס', 'ע', 'פ', 'צ', 'ק' ],
2071                         [ '',
2072                                 [ 'ק' ],
2073                                 [ 'ר' ],
2074                                 [ 'ש' ],
2075                                 [ 'ת' ],
2076                                 [ 'ת', 'ק' ],
2077                                 [ 'ת', 'ר' ],
2078                                 [ 'ת', 'ש' ],
2079                                 [ 'ת', 'ת' ],
2080                                 [ 'ת', 'ת', 'ק' ],
2081                                 [ 'ת', 'ת', 'ר' ],
2082                         ],
2083                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ]
2084                 ];
2085
2086                 $num = intval( $num );
2087                 if ( $num > 9999 || $num <= 0 ) {
2088                         return $num;
2089                 }
2090
2091                 // Round thousands have special notations
2092                 if ( $num === 1000 ) {
2093                         return "א' אלף";
2094                 } elseif ( $num % 1000 === 0 ) {
2095                         return $table[0][$num / 1000] . "' אלפים";
2096                 }
2097
2098                 $letters = [];
2099
2100                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
2101                         if ( $num >= $pow10 ) {
2102                                 if ( $num === 15 || $num === 16 ) {
2103                                         $letters[] = $table[0][9];
2104                                         $letters[] = $table[0][$num - 9];
2105                                         $num = 0;
2106                                 } else {
2107                                         $letters = array_merge(
2108                                                 $letters,
2109                                                 (array)$table[$i][intval( $num / $pow10 )]
2110                                         );
2111
2112                                         if ( $pow10 === 1000 ) {
2113                                                 $letters[] = "'";
2114                                         }
2115                                 }
2116                         }
2117
2118                         $num = $num % $pow10;
2119                 }
2120
2121                 $preTransformLength = count( $letters );
2122                 if ( $preTransformLength === 1 ) {
2123                         // Add geresh (single quote) to one-letter numbers
2124                         $letters[] = "'";
2125                 } else {
2126                         $lastIndex = $preTransformLength - 1;
2127                         $letters[$lastIndex] = str_replace(
2128                                 [ 'כ', 'מ', 'נ', 'פ', 'צ' ],
2129                                 [ 'ך', 'ם', 'ן', 'ף', 'ץ' ],
2130                                 $letters[$lastIndex]
2131                         );
2132
2133                         // Add gershayim (double quote) to multiple-letter numbers,
2134                         // but exclude numbers with only one letter after the thousands
2135                         // (1001-1009, 1020, 1030, 2001-2009, etc.)
2136                         if ( $letters[1] === "'" && $preTransformLength === 3 ) {
2137                                 $letters[] = "'";
2138                         } else {
2139                                 array_splice( $letters, -1, 0, '"' );
2140                         }
2141                 }
2142
2143                 return implode( $letters );
2144         }
2145
2146         /**
2147          * Used by date() and time() to adjust the time output.
2148          *
2149          * @param string $ts The time in date('YmdHis') format
2150          * @param mixed $tz Adjust the time by this amount (default false, mean we
2151          *   get user timecorrection setting)
2152          * @return int
2153          */
2154         public function userAdjust( $ts, $tz = false ) {
2155                 global $wgUser, $wgLocalTZoffset;
2156
2157                 if ( $tz === false ) {
2158                         $tz = $wgUser->getOption( 'timecorrection' );
2159                 }
2160
2161                 $data = explode( '|', $tz, 3 );
2162
2163                 if ( $data[0] == 'ZoneInfo' ) {
2164                         try {
2165                                 $userTZ = new DateTimeZone( $data[2] );
2166                                 $date = new DateTime( $ts, new DateTimeZone( 'UTC' ) );
2167                                 $date->setTimezone( $userTZ );
2168                                 return $date->format( 'YmdHis' );
2169                         } catch ( Exception $e ) {
2170                                 // Unrecognized timezone, default to 'Offset' with the stored offset.
2171                                 $data[0] = 'Offset';
2172                         }
2173                 }
2174
2175                 if ( $data[0] == 'System' || $tz == '' ) {
2176                         # Global offset in minutes.
2177                         $minDiff = $wgLocalTZoffset;
2178                 } elseif ( $data[0] == 'Offset' ) {
2179                         $minDiff = intval( $data[1] );
2180                 } else {
2181                         $data = explode( ':', $tz );
2182                         if ( count( $data ) == 2 ) {
2183                                 $data[0] = intval( $data[0] );
2184                                 $data[1] = intval( $data[1] );
2185                                 $minDiff = abs( $data[0] ) * 60 + $data[1];
2186                                 if ( $data[0] < 0 ) {
2187                                         $minDiff = -$minDiff;
2188                                 }
2189                         } else {
2190                                 $minDiff = intval( $data[0] ) * 60;
2191                         }
2192                 }
2193
2194                 # No difference ? Return time unchanged
2195                 if ( $minDiff == 0 ) {
2196                         return $ts;
2197                 }
2198
2199                 Wikimedia\suppressWarnings(); // E_STRICT system time bitching
2200                 # Generate an adjusted date; take advantage of the fact that mktime
2201                 # will normalize out-of-range values so we don't have to split $minDiff
2202                 # into hours and minutes.
2203                 $t = mktime( (
2204                         (int)substr( $ts, 8, 2 ) ), # Hours
2205                         (int)substr( $ts, 10, 2 ) + $minDiff, # Minutes
2206                         (int)substr( $ts, 12, 2 ), # Seconds
2207                         (int)substr( $ts, 4, 2 ), # Month
2208                         (int)substr( $ts, 6, 2 ), # Day
2209                         (int)substr( $ts, 0, 4 ) ); # Year
2210
2211                 $date = date( 'YmdHis', $t );
2212                 Wikimedia\restoreWarnings();
2213
2214                 return $date;
2215         }
2216
2217         /**
2218          * This is meant to be used by time(), date(), and timeanddate() to get
2219          * the date preference they're supposed to use, it should be used in
2220          * all children.
2221          *
2222          *     function timeanddate([...], $format = true) {
2223          *       $datePreference = $this->dateFormat($format);
2224          *       [...]
2225          *     }
2226          *
2227          * @param int|string|bool $usePrefs If true, the user's preference is used
2228          *   if false, the site/language default is used
2229          *   if int/string, assumed to be a format.
2230          * @return string
2231          */
2232         function dateFormat( $usePrefs = true ) {
2233                 global $wgUser;
2234
2235                 if ( is_bool( $usePrefs ) ) {
2236                         if ( $usePrefs ) {
2237                                 $datePreference = $wgUser->getDatePreference();
2238                         } else {
2239                                 $datePreference = (string)User::getDefaultOption( 'date' );
2240                         }
2241                 } else {
2242                         $datePreference = (string)$usePrefs;
2243                 }
2244
2245                 // return int
2246                 if ( $datePreference == '' ) {
2247                         return 'default';
2248                 }
2249
2250                 return $datePreference;
2251         }
2252
2253         /**
2254          * Get a format string for a given type and preference
2255          * @param string $type May be 'date', 'time', 'both', or 'pretty'.
2256          * @param string $pref The format name as it appears in Messages*.php under
2257          *  $datePreferences.
2258          *
2259          * @since 1.22 New type 'pretty' that provides a more readable timestamp format
2260          *
2261          * @return string
2262          */
2263         function getDateFormatString( $type, $pref ) {
2264                 $wasDefault = false;
2265                 if ( $pref == 'default' ) {
2266                         $wasDefault = true;
2267                         $pref = $this->getDefaultDateFormat();
2268                 }
2269
2270                 if ( !isset( $this->dateFormatStrings[$type][$pref] ) ) {
2271                         $df =
2272                                 $this->localisationCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2273
2274                         if ( $type === 'pretty' && $df === null ) {
2275                                 $df = $this->getDateFormatString( 'date', $pref );
2276                         }
2277
2278                         if ( !$wasDefault && $df === null ) {
2279                                 $pref = $this->getDefaultDateFormat();
2280                                 $df = $this->getLocalisationCache()
2281                                         ->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2282                         }
2283
2284                         $this->dateFormatStrings[$type][$pref] = $df;
2285                 }
2286                 return $this->dateFormatStrings[$type][$pref];
2287         }
2288
2289         /**
2290          * @param string $ts The time format which needs to be turned into a
2291          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2292          * @param bool $adj Whether to adjust the time output according to the
2293          *   user configured offset ($timecorrection)
2294          * @param mixed $format True to use user's date format preference
2295          * @param string|bool $timecorrection The time offset as returned by
2296          *   validateTimeZone() in Special:Preferences
2297          * @return string
2298          */
2299         public function date( $ts, $adj = false, $format = true, $timecorrection = false ) {
2300                 $ts = wfTimestamp( TS_MW, $ts );
2301                 if ( $adj ) {
2302                         $ts = $this->userAdjust( $ts, $timecorrection );
2303                 }
2304                 $df = $this->getDateFormatString( 'date', $this->dateFormat( $format ) );
2305                 return $this->sprintfDate( $df, $ts );
2306         }
2307
2308         /**
2309          * @param string $ts The time format which needs to be turned into a
2310          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2311          * @param bool $adj Whether to adjust the time output according to the
2312          *   user configured offset ($timecorrection)
2313          * @param mixed $format True to use user's date format preference
2314          * @param string|bool $timecorrection The time offset as returned by
2315          *   validateTimeZone() in Special:Preferences
2316          * @return string
2317          */
2318         public function time( $ts, $adj = false, $format = true, $timecorrection = false ) {
2319                 $ts = wfTimestamp( TS_MW, $ts );
2320                 if ( $adj ) {
2321                         $ts = $this->userAdjust( $ts, $timecorrection );
2322                 }
2323                 $df = $this->getDateFormatString( 'time', $this->dateFormat( $format ) );
2324                 return $this->sprintfDate( $df, $ts );
2325         }
2326
2327         /**
2328          * @param string $ts The time format which needs to be turned into a
2329          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2330          * @param bool $adj Whether to adjust the time output according to the
2331          *   user configured offset ($timecorrection)
2332          * @param mixed $format What format to return, if it's false output the
2333          *   default one (default true)
2334          * @param string|bool $timecorrection The time offset as returned by
2335          *   validateTimeZone() in Special:Preferences
2336          * @return string
2337          */
2338         public function timeanddate( $ts, $adj = false, $format = true, $timecorrection = false ) {
2339                 $ts = wfTimestamp( TS_MW, $ts );
2340                 if ( $adj ) {
2341                         $ts = $this->userAdjust( $ts, $timecorrection );
2342                 }
2343                 $df = $this->getDateFormatString( 'both', $this->dateFormat( $format ) );
2344                 return $this->sprintfDate( $df, $ts );
2345         }
2346
2347         /**
2348          * Takes a number of seconds and turns it into a text using values such as hours and minutes.
2349          *
2350          * @since 1.20
2351          *
2352          * @param int $seconds The amount of seconds.
2353          * @param array $chosenIntervals The intervals to enable.
2354          *
2355          * @return string
2356          */
2357         public function formatDuration( $seconds, array $chosenIntervals = [] ) {
2358                 $intervals = $this->getDurationIntervals( $seconds, $chosenIntervals );
2359
2360                 $segments = [];
2361
2362                 foreach ( $intervals as $intervalName => $intervalValue ) {
2363                         // Messages: duration-seconds, duration-minutes, duration-hours, duration-days, duration-weeks,
2364                         // duration-years, duration-decades, duration-centuries, duration-millennia
2365                         $message = wfMessage( 'duration-' . $intervalName )->numParams( $intervalValue );
2366                         $segments[] = $message->inLanguage( $this )->escaped();
2367                 }
2368
2369                 return $this->listToText( $segments );
2370         }
2371
2372         /**
2373          * Takes a number of seconds and returns an array with a set of corresponding intervals.
2374          * For example 65 will be turned into [ minutes => 1, seconds => 5 ].
2375          *
2376          * @since 1.20
2377          *
2378          * @param int $seconds The amount of seconds.
2379          * @param array $chosenIntervals The intervals to enable.
2380          *
2381          * @return array
2382          */
2383         public function getDurationIntervals( $seconds, array $chosenIntervals = [] ) {
2384                 if ( empty( $chosenIntervals ) ) {
2385                         $chosenIntervals = [
2386                                 'millennia',
2387                                 'centuries',
2388                                 'decades',
2389                                 'years',
2390                                 'days',
2391                                 'hours',
2392                                 'minutes',
2393                                 'seconds'
2394                         ];
2395                 }
2396
2397                 $intervals = array_intersect_key( self::$durationIntervals, array_flip( $chosenIntervals ) );
2398                 $sortedNames = array_keys( $intervals );
2399                 $smallestInterval = array_pop( $sortedNames );
2400
2401                 $segments = [];
2402
2403                 foreach ( $intervals as $name => $length ) {
2404                         $value = floor( $seconds / $length );
2405
2406                         if ( $value > 0 || ( $name == $smallestInterval && empty( $segments ) ) ) {
2407                                 $seconds -= $value * $length;
2408                                 $segments[$name] = $value;
2409                         }
2410                 }
2411
2412                 return $segments;
2413         }
2414
2415         /**
2416          * Internal helper function for userDate(), userTime() and userTimeAndDate()
2417          *
2418          * @param string $type Can be 'date', 'time' or 'both'
2419          * @param string $ts The time format which needs to be turned into a
2420          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2421          * @param User $user User object used to get preferences for timezone and format
2422          * @param array $options Array, can contain the following keys:
2423          *   - 'timecorrection': time correction, can have the following values:
2424          *     - true: use user's preference
2425          *     - false: don't use time correction
2426          *     - int: value of time correction in minutes
2427          *   - 'format': format to use, can have the following values:
2428          *     - true: use user's preference
2429          *     - false: use default preference
2430          *     - string: format to use
2431          * @since 1.19
2432          * @return string
2433          */
2434         private function internalUserTimeAndDate( $type, $ts, User $user, array $options ) {
2435                 $ts = wfTimestamp( TS_MW, $ts );
2436                 $options += [ 'timecorrection' => true, 'format' => true ];
2437                 if ( $options['timecorrection'] !== false ) {
2438                         if ( $options['timecorrection'] === true ) {
2439                                 $offset = $user->getOption( 'timecorrection' );
2440                         } else {
2441                                 $offset = $options['timecorrection'];
2442                         }
2443                         $ts = $this->userAdjust( $ts, $offset );
2444                 }
2445                 if ( $options['format'] === true ) {
2446                         $format = $user->getDatePreference();
2447                 } else {
2448                         $format = $options['format'];
2449                 }
2450                 $df = $this->getDateFormatString( $type, $this->dateFormat( $format ) );
2451                 return $this->sprintfDate( $df, $ts );
2452         }
2453
2454         /**
2455          * Get the formatted date for the given timestamp and formatted for
2456          * the given user.
2457          *
2458          * @param mixed $ts Mixed: the time format which needs to be turned into a
2459          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2460          * @param User $user User object used to get preferences for timezone and format
2461          * @param array $options Array, can contain the following keys:
2462          *   - 'timecorrection': time correction, can have the following values:
2463          *     - true: use user's preference
2464          *     - false: don't use time correction
2465          *     - int: value of time correction in minutes
2466          *   - 'format': format to use, can have the following values:
2467          *     - true: use user's preference
2468          *     - false: use default preference
2469          *     - string: format to use
2470          * @since 1.19
2471          * @return string
2472          */
2473         public function userDate( $ts, User $user, array $options = [] ) {
2474                 return $this->internalUserTimeAndDate( 'date', $ts, $user, $options );
2475         }
2476
2477         /**
2478          * Get the formatted time for the given timestamp and formatted for
2479          * the given user.
2480          *
2481          * @param mixed $ts The time format which needs to be turned into a
2482          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2483          * @param User $user User object used to get preferences for timezone and format
2484          * @param array $options Array, can contain the following keys:
2485          *   - 'timecorrection': time correction, can have the following values:
2486          *     - true: use user's preference
2487          *     - false: don't use time correction
2488          *     - int: value of time correction in minutes
2489          *   - 'format': format to use, can have the following values:
2490          *     - true: use user's preference
2491          *     - false: use default preference
2492          *     - string: format to use
2493          * @since 1.19
2494          * @return string
2495          */
2496         public function userTime( $ts, User $user, array $options = [] ) {
2497                 return $this->internalUserTimeAndDate( 'time', $ts, $user, $options );
2498         }
2499
2500         /**
2501          * Get the formatted date and time for the given timestamp and formatted for
2502          * the given user.
2503          *
2504          * @param mixed $ts The time format which needs to be turned into a
2505          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2506          * @param User $user User object used to get preferences for timezone and format
2507          * @param array $options Array, can contain the following keys:
2508          *   - 'timecorrection': time correction, can have the following values:
2509          *     - true: use user's preference
2510          *     - false: don't use time correction
2511          *     - int: value of time correction in minutes
2512          *   - 'format': format to use, can have the following values:
2513          *     - true: use user's preference
2514          *     - false: use default preference
2515          *     - string: format to use
2516          * @since 1.19
2517          * @return string
2518          */
2519         public function userTimeAndDate( $ts, User $user, array $options = [] ) {
2520                 return $this->internalUserTimeAndDate( 'both', $ts, $user, $options );
2521         }
2522
2523         /**
2524          * Get the timestamp in a human-friendly relative format, e.g., "3 days ago".
2525          *
2526          * Determine the difference between the timestamp and the current time, and
2527          * generate a readable timestamp by returning "<N> <units> ago", where the
2528          * largest possible unit is used.
2529          *
2530          * @since 1.26 (Prior to 1.26 method existed but was not meant to be used directly)
2531          *
2532          * @param MWTimestamp $time
2533          * @param MWTimestamp|null $relativeTo The base timestamp to compare to (defaults to now)
2534          * @param User|null $user User the timestamp is being generated for
2535          *  (or null to use main context's user)
2536          * @return string Formatted timestamp
2537          */
2538         public function getHumanTimestamp(
2539                 MWTimestamp $time, MWTimestamp $relativeTo = null, User $user = null
2540         ) {
2541                 if ( $relativeTo === null ) {
2542                         $relativeTo = new MWTimestamp();
2543                 }
2544                 if ( $user === null ) {
2545                         $user = RequestContext::getMain()->getUser();
2546                 }
2547
2548                 // Adjust for the user's timezone.
2549                 $offsetThis = $time->offsetForUser( $user );
2550                 $offsetRel = $relativeTo->offsetForUser( $user );
2551
2552                 $ts = '';
2553                 if ( Hooks::run( 'GetHumanTimestamp', [ &$ts, $time, $relativeTo, $user, $this ] ) ) {
2554                         $ts = $this->getHumanTimestampInternal( $time, $relativeTo, $user );
2555                 }
2556
2557                 // Reset the timezone on the objects.
2558                 $time->timestamp->sub( $offsetThis );
2559                 $relativeTo->timestamp->sub( $offsetRel );
2560
2561                 return $ts;
2562         }
2563
2564         /**
2565          * Convert an MWTimestamp into a pretty human-readable timestamp using
2566          * the given user preferences and relative base time.
2567          *
2568          * @see Language::getHumanTimestamp
2569          * @param MWTimestamp $ts Timestamp to prettify
2570          * @param MWTimestamp $relativeTo Base timestamp
2571          * @param User $user User preferences to use
2572          * @return string Human timestamp
2573          * @since 1.26
2574          */
2575         private function getHumanTimestampInternal(
2576                 MWTimestamp $ts, MWTimestamp $relativeTo, User $user
2577         ) {
2578                 $diff = $ts->diff( $relativeTo );
2579                 $diffDay = (bool)( (int)$ts->timestamp->format( 'w' ) -
2580                         (int)$relativeTo->timestamp->format( 'w' ) );
2581                 $days = $diff->days ?: (int)$diffDay;
2582                 if ( $diff->invert || $days > 5
2583                         && $ts->timestamp->format( 'Y' ) !== $relativeTo->timestamp->format( 'Y' )
2584                 ) {
2585                         // Timestamps are in different years: use full timestamp
2586                         // Also do full timestamp for future dates
2587                         /**
2588                          * @todo FIXME: Add better handling of future timestamps.
2589                          */
2590                         $format = $this->getDateFormatString( 'both', $user->getDatePreference() ?: 'default' );
2591                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2592                 } elseif ( $days > 5 ) {
2593                         // Timestamps are in same year,  but more than 5 days ago: show day and month only.
2594                         $format = $this->getDateFormatString( 'pretty', $user->getDatePreference() ?: 'default' );
2595                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2596                 } elseif ( $days > 1 ) {
2597                         // Timestamp within the past week: show the day of the week and time
2598                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2599                         $weekday = self::$mWeekdayMsgs[$ts->timestamp->format( 'w' )];
2600                         // Messages:
2601                         // sunday-at, monday-at, tuesday-at, wednesday-at, thursday-at, friday-at, saturday-at
2602                         $ts = wfMessage( "$weekday-at" )
2603                                 ->inLanguage( $this )
2604                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2605                                 ->text();
2606                 } elseif ( $days == 1 ) {
2607                         // Timestamp was yesterday: say 'yesterday' and the time.
2608                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2609                         $ts = wfMessage( 'yesterday-at' )
2610                                 ->inLanguage( $this )
2611                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2612                                 ->text();
2613                 } elseif ( $diff->h > 1 || $diff->h == 1 && $diff->i > 30 ) {
2614                         // Timestamp was today, but more than 90 minutes ago: say 'today' and the time.
2615                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2616                         $ts = wfMessage( 'today-at' )
2617                                 ->inLanguage( $this )
2618                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2619                                 ->text();
2620
2621                 // From here on in, the timestamp was soon enough ago so that we can simply say
2622                 // XX units ago, e.g., "2 hours ago" or "5 minutes ago"
2623                 } elseif ( $diff->h == 1 ) {
2624                         // Less than 90 minutes, but more than an hour ago.
2625                         $ts = wfMessage( 'hours-ago' )->inLanguage( $this )->numParams( 1 )->text();
2626                 } elseif ( $diff->i >= 1 ) {
2627                         // A few minutes ago.
2628                         $ts = wfMessage( 'minutes-ago' )->inLanguage( $this )->numParams( $diff->i )->text();
2629                 } elseif ( $diff->s >= 30 ) {
2630                         // Less than a minute, but more than 30 sec ago.
2631                         $ts = wfMessage( 'seconds-ago' )->inLanguage( $this )->numParams( $diff->s )->text();
2632                 } else {
2633                         // Less than 30 seconds ago.
2634                         $ts = wfMessage( 'just-now' )->text();
2635                 }
2636
2637                 return $ts;
2638         }
2639
2640         /**
2641          * @param string $key
2642          * @return string|null
2643          */
2644         public function getMessage( $key ) {
2645                 return $this->localisationCache->getSubitem( $this->mCode, 'messages', $key );
2646         }
2647
2648         /**
2649          * @return array
2650          */
2651         function getAllMessages() {
2652                 return $this->localisationCache->getItem( $this->mCode, 'messages' );
2653         }
2654
2655         /**
2656          * @param string $in
2657          * @param string $out
2658          * @param string $string
2659          * @return string
2660          */
2661         public function iconv( $in, $out, $string ) {
2662                 # Even with //IGNORE iconv can whine about illegal characters in
2663                 # *input* string. We just ignore those too.
2664                 # REF: https://bugs.php.net/bug.php?id=37166
2665                 # REF: https://phabricator.wikimedia.org/T18885
2666                 Wikimedia\suppressWarnings();
2667                 $text = iconv( $in, $out . '//IGNORE', $string );
2668                 Wikimedia\restoreWarnings();
2669                 return $text;
2670         }
2671
2672         // callback functions for ucwords(), ucwordbreaks()
2673
2674         /**
2675          * @param array $matches
2676          * @return mixed|string
2677          */
2678         function ucwordbreaksCallbackAscii( $matches ) {
2679                 return $this->ucfirst( $matches[1] );
2680         }
2681
2682         /**
2683          * @param array $matches
2684          * @return string
2685          */
2686         function ucwordbreaksCallbackMB( $matches ) {
2687                 return mb_strtoupper( $matches[0] );
2688         }
2689
2690         /**
2691          * @param array $matches
2692          * @return string
2693          */
2694         function ucwordsCallbackMB( $matches ) {
2695                 return mb_strtoupper( $matches[0] );
2696         }
2697
2698         /**
2699          * Make a string's first character uppercase
2700          *
2701          * @param string $str
2702          *
2703          * @return string
2704          */
2705         public function ucfirst( $str ) {
2706                 $o = ord( $str );
2707                 if ( $o < 96 ) { // if already uppercase...
2708                         return $str;
2709                 } elseif ( $o < 128 ) {
2710                         return ucfirst( $str ); // use PHP's ucfirst()
2711                 } else {
2712                         // fall back to more complex logic in case of multibyte strings
2713                         return $this->uc( $str, true );
2714                 }
2715         }
2716
2717         /**
2718          * Convert a string to uppercase
2719          *
2720          * @param string $str
2721          * @param bool $first
2722          *
2723          * @return string
2724          */
2725         public function uc( $str, $first = false ) {
2726                 if ( $first ) {
2727                         if ( $this->isMultibyte( $str ) ) {
2728                                 return $this->mbUpperChar( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2729                         } else {
2730                                 return ucfirst( $str );
2731                         }
2732                 } else {
2733                         return $this->isMultibyte( $str ) ? mb_strtoupper( $str ) : strtoupper( $str );
2734                 }
2735         }
2736
2737         /**
2738          * Convert character to uppercase, allowing overrides of the default mb_upper
2739          * behaviour, which is buggy in many ways. Having a conversion table can be
2740          * useful during transitions between PHP versions where unicode changes happen.
2741          * This can make some resources unreachable on-wiki, see discussion at T219279.
2742          * Providing such a conversion table can allow to manage the transition period.
2743          *
2744          * @since 1.34
2745          *
2746          * @param string $char
2747          *
2748          * @return string
2749          */
2750         protected function mbUpperChar( $char ) {
2751                 global $wgOverrideUcfirstCharacters;
2752                 if ( array_key_exists( $char, $wgOverrideUcfirstCharacters ) ) {
2753                         return $wgOverrideUcfirstCharacters[$char];
2754                 } else {
2755                         return mb_strtoupper( $char );
2756                 }
2757         }
2758
2759         /**
2760          * @param string $str
2761          * @return mixed|string
2762          */
2763         function lcfirst( $str ) {
2764                 $o = ord( $str );
2765                 if ( !$o ) {
2766                         return strval( $str );
2767                 } elseif ( $o >= 128 ) {
2768                         return $this->lc( $str, true );
2769                 } elseif ( $o > 96 ) {
2770                         return $str;
2771                 } else {
2772                         $str[0] = strtolower( $str[0] );
2773                         return $str;
2774                 }
2775         }
2776
2777         /**
2778          * @param string $str
2779          * @param bool $first
2780          * @return mixed|string
2781          */
2782         function lc( $str, $first = false ) {
2783                 if ( $first ) {
2784                         if ( $this->isMultibyte( $str ) ) {
2785                                 return mb_strtolower( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2786                         } else {
2787                                 return strtolower( substr( $str, 0, 1 ) ) . substr( $str, 1 );
2788                         }
2789                 } else {
2790                         return $this->isMultibyte( $str ) ? mb_strtolower( $str ) : strtolower( $str );
2791                 }
2792         }
2793
2794         /**
2795          * @param string $str
2796          * @return bool
2797          */
2798         function isMultibyte( $str ) {
2799                 return strlen( $str ) !== mb_strlen( $str );
2800         }
2801
2802         /**
2803          * @param string $str
2804          * @return mixed|string
2805          */
2806         function ucwords( $str ) {
2807                 if ( $this->isMultibyte( $str ) ) {
2808                         $str = $this->lc( $str );
2809
2810                         // regexp to find first letter in each word (i.e. after each space)
2811                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)| ([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2812
2813                         // function to use to capitalize a single char
2814                         return preg_replace_callback(
2815                                 $replaceRegexp,
2816                                 [ $this, 'ucwordsCallbackMB' ],
2817                                 $str
2818                         );
2819                 } else {
2820                         return ucwords( strtolower( $str ) );
2821                 }
2822         }
2823
2824         /**
2825          * capitalize words at word breaks
2826          *
2827          * @param string $str
2828          * @return mixed
2829          */
2830         function ucwordbreaks( $str ) {
2831                 if ( $this->isMultibyte( $str ) ) {
2832                         $str = $this->lc( $str );
2833
2834                         // since \b doesn't work for UTF-8, we explicitely define word break chars
2835                         $breaks = "[ \-\(\)\}\{\.,\?!]";
2836
2837                         // find first letter after word break
2838                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)|" .
2839                                 "$breaks([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2840
2841                         return preg_replace_callback(
2842                                 $replaceRegexp,
2843                                 [ $this, 'ucwordbreaksCallbackMB' ],
2844                                 $str
2845                         );
2846                 } else {
2847                         return preg_replace_callback(
2848                                 '/\b([\w\x80-\xff]+)\b/',
2849                                 [ $this, 'ucwordbreaksCallbackAscii' ],
2850                                 $str
2851                         );
2852                 }
2853         }
2854
2855         /**
2856          * Return a case-folded representation of $s
2857          *
2858          * This is a representation such that caseFold($s1)==caseFold($s2) if $s1
2859          * and $s2 are the same except for the case of their characters. It is not
2860          * necessary for the value returned to make sense when displayed.
2861          *
2862          * Do *not* perform any other normalisation in this function. If a caller
2863          * uses this function when it should be using a more general normalisation
2864          * function, then fix the caller.
2865          *
2866          * @param string $s
2867          *
2868          * @return string
2869          */
2870         function caseFold( $s ) {
2871                 return $this->uc( $s );
2872         }
2873
2874         /**
2875          * TODO: $s is not always a string per T218883
2876          * @param string $s
2877          * @return string
2878          */
2879         function checkTitleEncoding( $s ) {
2880                 if ( is_array( $s ) ) {
2881                         throw new MWException( 'Given array to checkTitleEncoding.' );
2882                 }
2883                 if ( StringUtils::isUtf8( $s ) ) {
2884                         return $s;
2885                 }
2886
2887                 return $this->iconv( $this->fallback8bitEncoding(), 'utf-8', $s );
2888         }
2889
2890         /**
2891          * @return string
2892          */
2893         function fallback8bitEncoding() {
2894                 return $this->localisationCache->getItem( $this->mCode, 'fallback8bitEncoding' );
2895         }
2896
2897         /**
2898          * Most writing systems use whitespace to break up words.
2899          * Some languages such as Chinese don't conventionally do this,
2900          * which requires special handling when breaking up words for
2901          * searching etc.
2902          *
2903          * @return bool
2904          */
2905         function hasWordBreaks() {
2906                 return true;
2907         }
2908
2909         /**
2910          * Some languages such as Chinese require word segmentation,
2911          * Specify such segmentation when overridden in derived class.
2912          *
2913          * @param string $string
2914          * @return string
2915          */
2916         function segmentByWord( $string ) {
2917                 return $string;
2918         }
2919
2920         /**
2921          * Some languages have special punctuation need to be normalized.
2922          * Make such changes here.
2923          *
2924          * @param string $string
2925          * @return string
2926          */
2927         function normalizeForSearch( $string ) {
2928                 return self::convertDoubleWidth( $string );
2929         }
2930
2931         /**
2932          * convert double-width roman characters to single-width.
2933          * range: ff00-ff5f ~= 0020-007f
2934          *
2935          * @param string $string
2936          *
2937          * @return string
2938          */
2939         protected static function convertDoubleWidth( $string ) {
2940                 static $full = null;
2941                 static $half = null;
2942
2943                 if ( $full === null ) {
2944                         $fullWidth = "０１２３４５６７８９ＡＢＣＤＥＦＧＨＩＪＫＬＭＮＯＰＱＲＳＴＵＶＷＸＹＺａｂｃｄｅｆｇｈｉｊｋｌｍｎｏｐｑｒｓｔｕｖｗｘｙｚ";
2945                         $halfWidth = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz";
2946                         $full = str_split( $fullWidth, 3 );
2947                         $half = str_split( $halfWidth );
2948                 }
2949
2950                 $string = str_replace( $full, $half, $string );
2951                 return $string;
2952         }
2953
2954         /**
2955          * @param string $string
2956          * @param string $pattern
2957          * @return string
2958          */
2959         protected static function insertSpace( $string, $pattern ) {
2960                 $string = preg_replace( $pattern, " $1 ", $string );
2961                 $string = preg_replace( '/ +/', ' ', $string );
2962                 return $string;
2963         }
2964
2965         /**
2966          * @param string[] $termsArray
2967          * @return string[]
2968          */
2969         function convertForSearchResult( $termsArray ) {
2970                 # some languages, e.g. Chinese, need to do a conversion
2971                 # in order for search results to be displayed correctly
2972                 return $termsArray;
2973         }
2974
2975         /**
2976          * Get the first character of a string.
2977          *
2978          * @param string $s
2979          * @return string
2980          */
2981         function firstChar( $s ) {
2982                 $matches = [];
2983                 preg_match(
2984                         '/^([\x00-\x7f]|[\xc0-\xdf][\x80-\xbf]|' .
2985                                 '[\xe0-\xef][\x80-\xbf]{2}|[\xf0-\xf7][\x80-\xbf]{3})/',
2986                         $s,
2987                         $matches
2988                 );
2989
2990                 if ( isset( $matches[1] ) ) {
2991                         if ( strlen( $matches[1] ) != 3 ) {
2992                                 return $matches[1];
2993                         }
2994
2995                         // Break down Hangul syllables to grab the first jamo
2996                         $code = UtfNormal\Utils::utf8ToCodepoint( $matches[1] );
2997                         if ( $code < 0xac00 || 0xd7a4 <= $code ) {
2998                                 return $matches[1];
2999                         } elseif ( $code < 0xb098 ) {
3000                                 return "\u{3131}";
3001                         } elseif ( $code < 0xb2e4 ) {
3002                                 return "\u{3134}";
3003                         } elseif ( $code < 0xb77c ) {
3004                                 return "\u{3137}";
3005                         } elseif ( $code < 0xb9c8 ) {
3006                                 return "\u{3139}";
3007                         } elseif ( $code < 0xbc14 ) {
3008                                 return "\u{3141}";
3009                         } elseif ( $code < 0xc0ac ) {
3010                                 return "\u{3142}";
3011                         } elseif ( $code < 0xc544 ) {
3012                                 return "\u{3145}";
3013                         } elseif ( $code < 0xc790 ) {
3014                                 return "\u{3147}";
3015                         } elseif ( $code < 0xcc28 ) {
3016                                 return "\u{3148}";
3017                         } elseif ( $code < 0xce74 ) {
3018                                 return "\u{314A}";
3019                         } elseif ( $code < 0xd0c0 ) {
3020                                 return "\u{314B}";
3021                         } elseif ( $code < 0xd30c ) {
3022                                 return "\u{314C}";
3023                         } elseif ( $code < 0xd558 ) {
3024                                 return "\u{314D}";
3025                         } else {
3026                                 return "\u{314E}";
3027                         }
3028                 } else {
3029                         return '';
3030                 }
3031         }
3032
3033         /**
3034          * Convert a UTF-8 string to normal form C. In Malayalam and Arabic, this
3035          * also cleans up certain backwards-compatible sequences, converting them
3036          * to the modern Unicode equivalent.
3037          *
3038          * This is language-specific for performance reasons only.
3039          *
3040          * @param string $s
3041          *
3042          * @return string
3043          */
3044         public function normalize( $s ) {
3045                 global $wgAllUnicodeFixes;
3046                 $s = UtfNormal\Validator::cleanUp( $s );
3047                 if ( $wgAllUnicodeFixes ) {
3048                         $s = $this->transformUsingPairFile( 'normalize-ar.php', $s );
3049                         $s = $this->transformUsingPairFile( 'normalize-ml.php', $s );
3050                 }
3051
3052                 return $s;
3053         }
3054
3055         /**
3056          * Transform a string using serialized data stored in the given file (which
3057          * must be in the serialized subdirectory of $IP). The file contains pairs
3058          * mapping source characters to destination characters.
3059          *
3060          * The data is cached in process memory. This will go faster if you have the
3061          * FastStringSearch extension.
3062          *
3063          * @param string $file
3064          * @param string $string
3065          *
3066          * @throws MWException
3067          * @return string
3068          */
3069         protected function transformUsingPairFile( $file, $string ) {
3070                 if ( !isset( $this->transformData[$file] ) ) {
3071                         global $IP;
3072                         $data = require "$IP/languages/data/{$file}";
3073                         $this->transformData[$file] = new ReplacementArray( $data );
3074                 }
3075                 return $this->transformData[$file]->replace( $string );
3076         }
3077
3078         /**
3079          * For right-to-left language support
3080          *
3081          * @return bool
3082          */
3083         function isRTL() {
3084                 return $this->localisationCache->getItem( $this->mCode, 'rtl' );
3085         }
3086
3087         /**
3088          * Return the correct HTML 'dir' attribute value for this language.
3089          * @return string
3090          */
3091         function getDir() {
3092                 return $this->isRTL() ? 'rtl' : 'ltr';
3093         }
3094
3095         /**
3096          * Return 'left' or 'right' as appropriate alignment for line-start
3097          * for this language's text direction.
3098          *
3099          * Should be equivalent to CSS3 'start' text-align value....
3100          *
3101          * @return string
3102          */
3103         function alignStart() {
3104                 return $this->isRTL() ? 'right' : 'left';
3105         }
3106
3107         /**
3108          * Return 'right' or 'left' as appropriate alignment for line-end
3109          * for this language's text direction.
3110          *
3111          * Should be equivalent to CSS3 'end' text-align value....
3112          *
3113          * @return string
3114          */
3115         function alignEnd() {
3116                 return $this->isRTL() ? 'left' : 'right';
3117         }
3118
3119         /**
3120          * A hidden direction mark (LRM or RLM), depending on the language direction.
3121          * Unlike getDirMark(), this function returns the character as an HTML entity.
3122          * This function should be used when the output is guaranteed to be HTML,
3123          * because it makes the output HTML source code more readable. When
3124          * the output is plain text or can be escaped, getDirMark() should be used.
3125          *
3126          * @param bool $opposite Get the direction mark opposite to your language
3127          * @return string
3128          * @since 1.20
3129          */
3130         function getDirMarkEntity( $opposite = false ) {
3131                 if ( $opposite ) {
3132                         return $this->isRTL() ? '&lrm;' : '&rlm;';
3133                 }
3134                 return $this->isRTL() ? '&rlm;' : '&lrm;';
3135         }
3136
3137         /**
3138          * A hidden direction mark (LRM or RLM), depending on the language direction.
3139          * This function produces them as invisible Unicode characters and
3140          * the output may be hard to read and debug, so it should only be used
3141          * when the output is plain text or can be escaped. When the output is
3142          * HTML, use getDirMarkEntity() instead.
3143          *
3144          * @param bool $opposite Get the direction mark opposite to your language
3145          * @return string
3146          */
3147         function getDirMark( $opposite = false ) {
3148                 $lrm = "\u{200E}"; # LEFT-TO-RIGHT MARK, commonly abbreviated LRM
3149                 $rlm = "\u{200F}"; # RIGHT-TO-LEFT MARK, commonly abbreviated RLM
3150                 if ( $opposite ) {
3151                         return $this->isRTL() ? $lrm : $rlm;
3152                 }
3153                 return $this->isRTL() ? $rlm : $lrm;
3154         }
3155
3156         /**
3157          * @return array
3158          */
3159         function capitalizeAllNouns() {
3160                 return $this->localisationCache->getItem( $this->mCode, 'capitalizeAllNouns' );
3161         }
3162
3163         /**
3164          * An arrow, depending on the language direction.
3165          *
3166          * @param string $direction The direction of the arrow: forwards (default),
3167          *   backwards, left, right, up, down.
3168          * @return string
3169          */
3170         function getArrow( $direction = 'forwards' ) {
3171                 switch ( $direction ) {
3172                         case 'forwards':
3173                                 return $this->isRTL() ? '←' : '→';
3174                         case 'backwards':
3175                                 return $this->isRTL() ? '→' : '←';
3176                         case 'left':
3177                                 return '←';
3178                         case 'right':
3179                                 return '→';
3180                         case 'up':
3181                                 return '↑';
3182                         case 'down':
3183                                 return '↓';
3184                 }
3185         }
3186
3187         /**
3188          * To allow "foo[[bar]]" to extend the link over the whole word "foobar"
3189          *
3190          * @return bool
3191          */
3192         function linkPrefixExtension() {
3193                 return $this->localisationCache->getItem( $this->mCode, 'linkPrefixExtension' );
3194         }
3195
3196         /**
3197          * Get all magic words from cache.
3198          * @return array
3199          */
3200         function getMagicWords() {
3201                 return $this->localisationCache->getItem( $this->mCode, 'magicWords' );
3202         }
3203
3204         /**
3205          * Fill a MagicWord object with data from here
3206          *
3207          * @param MagicWord $mw
3208          */
3209         function getMagic( $mw ) {
3210                 $rawEntry = $this->mMagicExtensions[$mw->mId] ??
3211                         $this->localisationCache->getSubitem( $this->mCode, 'magicWords', $mw->mId );
3212
3213                 if ( !is_array( $rawEntry ) ) {
3214                         wfWarn( "\"$rawEntry\" is not a valid magic word for \"$mw->mId\"" );
3215                 } else {
3216                         $mw->mCaseSensitive = $rawEntry[0];
3217                         $mw->mSynonyms = array_slice( $rawEntry, 1 );
3218                 }
3219         }
3220
3221         /**
3222          * Add magic words to the extension array
3223          *
3224          * @param array $newWords
3225          */
3226         function addMagicWordsByLang( $newWords ) {
3227                 $fallbackChain = $this->getFallbackLanguages();
3228                 $fallbackChain = array_reverse( $fallbackChain );
3229                 foreach ( $fallbackChain as $code ) {
3230                         if ( isset( $newWords[$code] ) ) {
3231                                 $this->mMagicExtensions = $newWords[$code] + $this->mMagicExtensions;
3232                         }
3233                 }
3234         }
3235
3236         /**
3237          * Get special page names, as an associative array
3238          *   canonical name => array of valid names, including aliases
3239          * @return array
3240          */
3241         function getSpecialPageAliases() {
3242                 // Cache aliases because it may be slow to load them
3243                 if ( is_null( $this->mExtendedSpecialPageAliases ) ) {
3244                         // Initialise array
3245                         $this->mExtendedSpecialPageAliases =
3246                                 $this->localisationCache->getItem( $this->mCode, 'specialPageAliases' );
3247                 }
3248
3249                 return $this->mExtendedSpecialPageAliases;
3250         }
3251
3252         /**
3253          * Italic is unsuitable for some languages
3254          *
3255          * @param string $text The text to be emphasized.
3256          * @return string
3257          */
3258         function emphasize( $text ) {
3259                 return "<em>$text</em>";
3260         }
3261
3262         /**
3263          * Normally we output all numbers in plain en_US style, that is
3264          * 293,291.235 for twohundredninetythreethousand-twohundredninetyone
3265          * point twohundredthirtyfive. However this is not suitable for all
3266          * languages, some such as Bengali (bn) want ২,৯৩,২৯১.২৩৫ and others such as
3267          * Icelandic just want to use commas instead of dots, and dots instead
3268          * of commas like "293.291,235".
3269          *
3270          * An example of this function being called:
3271          * <code>
3272          * wfMessage( 'message' )->numParams( $num )->text()
3273          * </code>
3274          *
3275          * See $separatorTransformTable on MessageIs.php for
3276          * the , => . and . => , implementation.
3277          *
3278          * @todo check if it's viable to use localeconv() for the decimal separator thing.
3279          * @param int|float $number The string to be formatted, should be an integer
3280          *   or a floating point number.
3281          * @param bool $nocommafy Set to true for special numbers like dates
3282          * @return string
3283          */
3284         public function formatNum( $number, $nocommafy = false ) {
3285                 global $wgTranslateNumerals;
3286                 if ( !$nocommafy ) {
3287                         $number = $this->commafy( $number );
3288                         $s = $this->separatorTransformTable();
3289                         if ( $s ) {
3290                                 $number = strtr( $number, $s );
3291                         }
3292                 }
3293
3294                 if ( $wgTranslateNumerals ) {
3295                         $s = $this->digitTransformTable();
3296                         if ( $s ) {
3297                                 $number = strtr( $number, $s );
3298                         }
3299                 }
3300
3301                 return (string)$number;
3302         }
3303
3304         /**
3305          * Front-end for non-commafied formatNum
3306          *
3307          * @param int|float $number The string to be formatted, should be an integer
3308          *        or a floating point number.
3309          * @since 1.21
3310          * @return string
3311          */
3312         public function formatNumNoSeparators( $number ) {
3313                 return $this->formatNum( $number, true );
3314         }
3315
3316         /**
3317          * @param string $number
3318          * @return string
3319          */
3320         public function parseFormattedNumber( $number ) {
3321                 $s = $this->digitTransformTable();
3322                 if ( $s ) {
3323                         // eliminate empty array values such as ''. (T66347)
3324                         $s = array_filter( $s );
3325                         $number = strtr( $number, array_flip( $s ) );
3326                 }
3327
3328                 $s = $this->separatorTransformTable();
3329                 if ( $s ) {
3330                         // eliminate empty array values such as ''. (T66347)
3331                         $s = array_filter( $s );
3332                         $number = strtr( $number, array_flip( $s ) );
3333                 }
3334
3335                 $number = strtr( $number, [ ',' => '' ] );
3336                 return $number;
3337         }
3338
3339         /**
3340          * Adds commas to a given number
3341          * @since 1.19
3342          * @param mixed $number
3343          * @return string
3344          */
3345         function commafy( $number ) {
3346                 $digitGroupingPattern = $this->digitGroupingPattern();
3347                 $minimumGroupingDigits = $this->minimumGroupingDigits();
3348                 if ( $number === null ) {
3349                         return '';
3350                 }
3351
3352                 if ( !$digitGroupingPattern || $digitGroupingPattern === "###,###,###" ) {
3353                         // Default grouping is at thousands, use the same for ###,###,### pattern too.
3354                         // In some languages it's conventional not to insert a thousands separator
3355                         // in numbers that are four digits long (1000-9999).
3356                         if ( $minimumGroupingDigits ) {
3357                                 // Number of '#' characters after last comma in the grouping pattern.
3358                                 // The pattern is hardcoded here, but this would vary for different patterns.
3359                                 $primaryGroupingSize = 3;
3360                                 // Maximum length of a number to suppress digit grouping for.
3361                                 $maximumLength = $minimumGroupingDigits + $primaryGroupingSize - 1;
3362                                 if ( preg_match( '/^\-?\d{1,' . $maximumLength . '}(\.\d+)?$/', $number ) ) {
3363                                         return $number;
3364                                 }
3365                         }
3366                         return strrev( (string)preg_replace( '/(\d{3})(?=\d)(?!\d*\.)/', '$1,', strrev( $number ) ) );
3367                 } else {
3368                         // Ref: http://cldr.unicode.org/translation/number-patterns
3369                         $sign = "";
3370                         if ( intval( $number ) < 0 ) {
3371                                 // For negative numbers apply the algorithm like positive number and add sign.
3372                                 $sign = "-";
3373                                 $number = substr( $number, 1 );
3374                         }
3375                         $integerPart = [];
3376                         $decimalPart = [];
3377                         $numMatches = preg_match_all( "/(#+)/", $digitGroupingPattern, $matches );
3378                         preg_match( "/\d+/", $number, $integerPart );
3379                         preg_match( "/\.\d*/", $number, $decimalPart );
3380                         $groupedNumber = ( count( $decimalPart ) > 0 ) ? $decimalPart[0] : "";
3381                         if ( $groupedNumber === $number ) {
3382                                 // the string does not have any number part. Eg: .12345
3383                                 return $sign . $groupedNumber;
3384                         }
3385                         $start = $end = ( $integerPart ) ? strlen( $integerPart[0] ) : 0;
3386                         while ( $start > 0 ) {
3387                                 $match = $matches[0][$numMatches - 1];
3388                                 $matchLen = strlen( $match );
3389                                 $start = $end - $matchLen;
3390                                 if ( $start < 0 ) {
3391                                         $start = 0;
3392                                 }
3393                                 $groupedNumber = substr( $number, $start, $end - $start ) . $groupedNumber;
3394                                 $end = $start;
3395                                 if ( $numMatches > 1 ) {
3396                                         // use the last pattern for the rest of the number
3397                                         $numMatches--;
3398                                 }
3399                                 if ( $start > 0 ) {
3400                                         $groupedNumber = "," . $groupedNumber;
3401                                 }
3402                         }
3403                         return $sign . $groupedNumber;
3404                 }
3405         }
3406
3407         /**
3408          * @return string
3409          */
3410         function digitGroupingPattern() {
3411                 return $this->localisationCache->getItem( $this->mCode, 'digitGroupingPattern' );
3412         }
3413
3414         /**
3415          * @return array
3416          */
3417         function digitTransformTable() {
3418                 return $this->localisationCache->getItem( $this->mCode, 'digitTransformTable' );
3419         }
3420
3421         /**
3422          * @return array
3423          */
3424         function separatorTransformTable() {
3425                 return $this->localisationCache->getItem( $this->mCode, 'separatorTransformTable' );
3426         }
3427
3428         /**
3429          * @return int|null
3430          */
3431         function minimumGroupingDigits() {
3432                 return $this->localisationCache->getItem( $this->mCode, 'minimumGroupingDigits' );
3433         }
3434
3435         /**
3436          * Take a list of strings and build a locale-friendly comma-separated
3437          * list, using the local comma-separator message.
3438          * The last two strings are chained with an "and".
3439          *
3440          * @param string[] $list
3441          * @return string
3442          */
3443         public function listToText( array $list ) {
3444                 $itemCount = count( $list );
3445                 if ( $itemCount < 1 ) {
3446                         return '';
3447                 }
3448                 $text = array_pop( $list );
3449                 if ( $itemCount > 1 ) {
3450                         $and = $this->msg( 'and' )->escaped();
3451                         $space = $this->msg( 'word-separator' )->escaped();
3452                         $comma = '';
3453                         if ( $itemCount > 2 ) {
3454                                 $comma = $this->msg( 'comma-separator' )->escaped();
3455                         }
3456                         $text = implode( $comma, $list ) . $and . $space . $text;
3457                 }
3458                 return $text;
3459         }
3460
3461         /**
3462          * Take a list of strings and build a locale-friendly comma-separated
3463          * list, using the local comma-separator message.
3464          * @param string[] $list Array of strings to put in a comma list
3465          * @return string
3466          */
3467         function commaList( array $list ) {
3468                 return implode(
3469                         wfMessage( 'comma-separator' )->inLanguage( $this )->escaped(),
3470                         $list
3471                 );
3472         }
3473
3474         /**
3475          * Take a list of strings and build a locale-friendly semicolon-separated
3476          * list, using the local semicolon-separator message.
3477          * @param string[] $list Array of strings to put in a semicolon list
3478          * @return string
3479          */
3480         function semicolonList( array $list ) {
3481                 return implode(
3482                         wfMessage( 'semicolon-separator' )->inLanguage( $this )->escaped(),
3483                         $list
3484                 );
3485         }
3486
3487         /**
3488          * Same as commaList, but separate it with the pipe instead.
3489          * @param string[] $list Array of strings to put in a pipe list
3490          * @return string
3491          */
3492         function pipeList( array $list ) {
3493                 return implode(
3494                         wfMessage( 'pipe-separator' )->inLanguage( $this )->escaped(),
3495                         $list
3496                 );
3497         }
3498
3499         /**
3500          * Truncate a string to a specified length in bytes, appending an optional
3501          * string (e.g. for ellipsis)
3502          *
3503          * If $length is negative, the string will be truncated from the beginning
3504          *
3505          * @since 1.31
3506          *
3507          * @param string $string String to truncate
3508          * @param int $length Maximum length in bytes
3509          * @param string $ellipsis String to append to the end of truncated text
3510          * @param bool $adjustLength Subtract length of ellipsis from $length
3511          *
3512          * @return string
3513          */
3514         function truncateForDatabase( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3515                 return $this->truncateInternal(
3516                         $string, $length, $ellipsis, $adjustLength, 'strlen', 'substr'
3517                 );
3518         }
3519
3520         /**
3521          * Truncate a string to a specified number of characters, appending an optional
3522          * string (e.g. for ellipsis).
3523          *
3524          * This provides multibyte version of truncateForDatabase() method of this class,
3525          * suitable for truncation based on number of characters, instead of number of bytes.
3526          *
3527          * If $length is negative, the string will be truncated from the beginning.
3528          *
3529          * @since 1.31
3530          *
3531          * @param string $string String to truncate
3532          * @param int $length Maximum number of characters
3533          * @param string $ellipsis String to append to the end of truncated text
3534          * @param bool $adjustLength Subtract length of ellipsis from $length
3535          *
3536          * @return string
3537          */
3538         function truncateForVisual( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3539                 // Passing encoding to mb_strlen and mb_substr is optional.
3540                 // Encoding defaults to mb_internal_encoding(), which is set to UTF-8 in Setup.php, so
3541                 // explicit specification of encoding is skipped.
3542                 // Note: Both multibyte methods are callables invoked in truncateInternal.
3543                 return $this->truncateInternal(
3544                         $string, $length, $ellipsis, $adjustLength, 'mb_strlen', 'mb_substr'
3545                 );
3546         }
3547
3548         /**
3549          * Internal method used for truncation. This method abstracts text truncation into
3550          * one common method, allowing users to provide length measurement function and
3551          * function for finding substring.
3552          *
3553          * For usages, see truncateForDatabase and truncateForVisual.
3554          *
3555          * @param string $string String to truncate
3556          * @param int $length Maximum length of final text
3557          * @param string $ellipsis String to append to the end of truncated text
3558          * @param bool $adjustLength Subtract length of ellipsis from $length
3559          * @param callable $measureLength Callable function used for determining the length of text
3560          * @param callable $getSubstring Callable function used for getting the substrings
3561          *
3562          * @return string
3563          */
3564         private function truncateInternal(
3565                 $string, $length, $ellipsis, $adjustLength, callable $measureLength, callable $getSubstring
3566         ) {
3567                 # Check if there is no need to truncate
3568                 if ( $measureLength( $string ) <= abs( $length ) ) {
3569                         return $string; // no need to truncate
3570                 }
3571
3572                 # Use the localized ellipsis character
3573                 if ( $ellipsis == '...' ) {
3574                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3575                 }
3576                 if ( $length == 0 ) {
3577                         return $ellipsis; // convention
3578                 }
3579
3580                 $stringOriginal = $string;
3581                 # If ellipsis length is >= $length then we can't apply $adjustLength
3582                 if ( $adjustLength && $measureLength( $ellipsis ) >= abs( $length ) ) {
3583                         $string = $ellipsis; // this can be slightly unexpected
3584                 # Otherwise, truncate and add ellipsis...
3585                 } else {
3586                         $ellipsisLength = $adjustLength ? $measureLength( $ellipsis ) : 0;
3587                         if ( $length > 0 ) {
3588                                 $length -= $ellipsisLength;
3589                                 $string = $getSubstring( $string, 0, $length ); // xyz...
3590                                 $string = $this->removeBadCharLast( $string );
3591                                 $string = rtrim( $string ) . $ellipsis;
3592                         } else {
3593                                 $length += $ellipsisLength;
3594                                 $string = $getSubstring( $string, $length ); // ...xyz
3595                                 $string = $this->removeBadCharFirst( $string );
3596                                 $string = $ellipsis . ltrim( $string );
3597                         }
3598                 }
3599
3600                 # Do not truncate if the ellipsis makes the string longer/equal (T24181).
3601                 # This check is *not* redundant if $adjustLength, due to the single case where
3602                 # LEN($ellipsis) > ABS($limit arg); $stringOriginal could be shorter than $string.
3603                 if ( $measureLength( $string ) < $measureLength( $stringOriginal ) ) {
3604                         return $string;
3605                 } else {
3606                         return $stringOriginal;
3607                 }
3608         }
3609
3610         /**
3611          * Remove bytes that represent an incomplete Unicode character
3612          * at the end of string (e.g. bytes of the char are missing)
3613          *
3614          * @param string $string
3615          * @return string
3616          */
3617         protected function removeBadCharLast( $string ) {
3618                 if ( $string != '' ) {
3619                         $char = ord( $string[strlen( $string ) - 1] );
3620                         $m = [];
3621                         if ( $char >= 0xc0 ) {
3622                                 # We got the first byte only of a multibyte char; remove it.
3623                                 $string = substr( $string, 0, -1 );
3624                         } elseif ( $char >= 0x80 &&
3625                                 // Use the /s modifier (PCRE_DOTALL) so (.*) also matches newlines
3626                                 preg_match( '/^(.*)(?:[\xe0-\xef][\x80-\xbf]|' .
3627                                         '[\xf0-\xf7][\x80-\xbf]{1,2})$/s', $string, $m )
3628                         ) {
3629                                 # We chopped in the middle of a character; remove it
3630                                 $string = $m[1];
3631                         }
3632                 }
3633                 return $string;
3634         }
3635
3636         /**
3637          * Remove bytes that represent an incomplete Unicode character
3638          * at the start of string (e.g. bytes of the char are missing)
3639          *
3640          * @param string $string
3641          * @return string
3642          */
3643         protected function removeBadCharFirst( $string ) {
3644                 if ( $string != '' ) {
3645                         $char = ord( $string[0] );
3646                         if ( $char >= 0x80 && $char < 0xc0 ) {
3647                                 # We chopped in the middle of a character; remove the whole thing
3648                                 $string = preg_replace( '/^[\x80-\xbf]+/', '', $string );
3649                         }
3650                 }
3651                 return $string;
3652         }
3653
3654         /**
3655          * Truncate a string of valid HTML to a specified length in bytes,
3656          * appending an optional string (e.g. for ellipses), and return valid HTML
3657          *
3658          * This is only intended for styled/linked text, such as HTML with
3659          * tags like <span> and <a>, were the tags are self-contained (valid HTML).
3660          * Also, this will not detect things like "display:none" CSS.
3661          *
3662          * Note: since 1.18 you do not need to leave extra room in $length for ellipses.
3663          *
3664          * @param string $text HTML string to truncate
3665          * @param int $length (zero/positive) Maximum length (including ellipses)
3666          * @param string $ellipsis String to append to the truncated text
3667          * @return string
3668          */
3669         function truncateHtml( $text, $length, $ellipsis = '...' ) {
3670                 # Use the localized ellipsis character
3671                 if ( $ellipsis == '...' ) {
3672                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3673                 }
3674                 # Check if there is clearly no need to truncate
3675                 if ( $length <= 0 ) {
3676                         return $ellipsis; // no text shown, nothing to format (convention)
3677                 } elseif ( strlen( $text ) <= $length ) {
3678                         return $text; // string short enough even *with* HTML (short-circuit)
3679                 }
3680
3681                 $dispLen = 0; // innerHTML legth so far
3682                 $testingEllipsis = false; // checking if ellipses will make string longer/equal?
3683                 $tagType = 0; // 0-open, 1-close
3684                 $bracketState = 0; // 1-tag start, 2-tag name, 0-neither
3685                 $entityState = 0; // 0-not entity, 1-entity
3686                 $tag = $ret = ''; // accumulated tag name, accumulated result string
3687                 $openTags = []; // open tag stack
3688                 $maybeState = null; // possible truncation state
3689
3690                 $textLen = strlen( $text );
3691                 $neLength = max( 0, $length - strlen( $ellipsis ) ); // non-ellipsis len if truncated
3692                 for ( $pos = 0; true; ++$pos ) {
3693                         # Consider truncation once the display length has reached the maximim.
3694                         # We check if $dispLen > 0 to grab tags for the $neLength = 0 case.
3695                         # Check that we're not in the middle of a bracket/entity...
3696                         if ( $dispLen && $dispLen >= $neLength && $bracketState == 0 && !$entityState ) {
3697                                 if ( !$testingEllipsis ) {
3698                                         $testingEllipsis = true;
3699                                         # Save where we are; we will truncate here unless there turn out to
3700                                         # be so few remaining characters that truncation is not necessary.
3701                                         if ( !$maybeState ) { // already saved? ($neLength = 0 case)
3702                                                 $maybeState = [ $ret, $openTags ]; // save state
3703                                         }
3704                                 } elseif ( $dispLen > $length && $dispLen > strlen( $ellipsis ) ) {
3705                                         # String in fact does need truncation, the truncation point was OK.
3706                                         // @phan-suppress-next-line PhanTypeInvalidExpressionArrayDestructuring
3707                                         list( $ret, $openTags ) = $maybeState; // reload state
3708                                         $ret = $this->removeBadCharLast( $ret ); // multi-byte char fix
3709                                         $ret .= $ellipsis; // add ellipsis
3710                                         break;
3711                                 }
3712                         }
3713                         if ( $pos >= $textLen ) {
3714                                 break; // extra iteration just for above checks
3715                         }
3716
3717                         # Read the next char...
3718                         $ch = $text[$pos];
3719                         $lastCh = $pos ? $text[$pos - 1] : '';
3720                         $ret .= $ch; // add to result string
3721                         if ( $ch == '<' ) {
3722                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags ); // for bad HTML
3723                                 $entityState = 0; // for bad HTML
3724                                 $bracketState = 1; // tag started (checking for backslash)
3725                         } elseif ( $ch == '>' ) {
3726                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags );
3727                                 $entityState = 0; // for bad HTML
3728                                 $bracketState = 0; // out of brackets
3729                         } elseif ( $bracketState == 1 ) {
3730                                 if ( $ch == '/' ) {
3731                                         $tagType = 1; // close tag (e.g. "</span>")
3732                                 } else {
3733                                         $tagType = 0; // open tag (e.g. "<span>")
3734                                         $tag .= $ch;
3735                                 }
3736                                 $bracketState = 2; // building tag name
3737                         } elseif ( $bracketState == 2 ) {
3738                                 if ( $ch != ' ' ) {
3739                                         $tag .= $ch;
3740                                 } else {
3741                                         // Name found (e.g. "<a href=..."), add on tag attributes...
3742                                         $pos += $this->truncate_skip( $ret, $text, "<>", $pos + 1 );
3743                                 }
3744                         } elseif ( $bracketState == 0 ) {
3745                                 if ( $entityState ) {
3746                                         if ( $ch == ';' ) {
3747                                                 $entityState = 0;
3748                                                 $dispLen++; // entity is one displayed char
3749                                         }
3750                                 } else {
3751                                         if ( $neLength == 0 && !$maybeState ) {
3752                                                 // Save state without $ch. We want to *hit* the first
3753                                                 // display char (to get tags) but not *use* it if truncating.
3754                                                 $maybeState = [ substr( $ret, 0, -1 ), $openTags ];
3755                                         }
3756                                         if ( $ch == '&' ) {
3757                                                 $entityState = 1; // entity found, (e.g. "&#160;")
3758                                         } else {
3759                                                 $dispLen++; // this char is displayed
3760                                                 // Add the next $max display text chars after this in one swoop...
3761                                                 $max = ( $testingEllipsis ? $length : $neLength ) - $dispLen;
3762                                                 $skipped = $this->truncate_skip( $ret, $text, "<>&", $pos + 1, $max );
3763                                                 $dispLen += $skipped;
3764                                                 $pos += $skipped;
3765                                         }
3766                                 }
3767                         }
3768                 }
3769                 // Close the last tag if left unclosed by bad HTML
3770                 $this->truncate_endBracket( $tag, $text[$textLen - 1], $tagType, $openTags );
3771                 while ( count( $openTags ) > 0 ) {
3772                         $ret .= '</' . array_pop( $openTags ) . '>'; // close open tags
3773                 }
3774                 return $ret;
3775         }
3776
3777         /**
3778          * truncateHtml() helper function
3779          * like strcspn() but adds the skipped chars to $ret
3780          *
3781          * @param string $ret
3782          * @param string $text
3783          * @param string $search
3784          * @param int $start
3785          * @param null|int $len
3786          * @return int
3787          */
3788         private function truncate_skip( &$ret, $text, $search, $start, $len = null ) {
3789                 if ( $len === null ) {
3790                         $len = -1; // -1 means "no limit" for strcspn
3791                 } elseif ( $len < 0 ) {
3792                         $len = 0; // sanity
3793                 }
3794                 $skipCount = 0;
3795                 if ( $start < strlen( $text ) ) {
3796                         $skipCount = strcspn( $text, $search, $start, $len );
3797                         $ret .= substr( $text, $start, $skipCount );
3798                 }
3799                 return $skipCount;
3800         }
3801
3802         /**
3803          * truncateHtml() helper function
3804          * (a) push or pop $tag from $openTags as needed
3805          * (b) clear $tag value
3806          * @param string &$tag Current HTML tag name we are looking at
3807          * @param int $tagType (0-open tag, 1-close tag)
3808          * @param string $lastCh Character before the '>' that ended this tag
3809          * @param array &$openTags Open tag stack (not accounting for $tag)
3810          */
3811         private function truncate_endBracket( &$tag, $tagType, $lastCh, &$openTags ) {
3812                 $tag = ltrim( $tag );
3813                 if ( $tag != '' ) {
3814                         if ( $tagType == 0 && $lastCh != '/' ) {
3815                                 $openTags[] = $tag; // tag opened (didn't close itself)
3816                         } elseif ( $tagType == 1 ) {
3817                                 if ( $openTags && $tag == $openTags[count( $openTags ) - 1] ) {
3818                                         array_pop( $openTags ); // tag closed
3819                                 }
3820                         }
3821                         $tag = '';
3822                 }
3823         }
3824
3825         /**
3826          * Grammatical transformations, needed for inflected languages
3827          * Invoked by putting {{grammar:case|word}} in a message
3828          *
3829          * @param string $word
3830          * @param string $case
3831          * @return string
3832          */
3833         function convertGrammar( $word, $case ) {
3834                 global $wgGrammarForms;
3835                 if ( isset( $wgGrammarForms[$this->getCode()][$case][$word] ) ) {
3836                         return $wgGrammarForms[$this->getCode()][$case][$word];
3837                 }
3838
3839                 $grammarTransformations = $this->getGrammarTransformations();
3840
3841                 if ( isset( $grammarTransformations[$case] ) ) {
3842                         $forms = $grammarTransformations[$case];
3843
3844                         // Some names of grammar rules are aliases for other rules.
3845                         // In such cases the value is a string rather than object,
3846                         // so load the actual rules.
3847                         if ( is_string( $forms ) ) {
3848                                 $forms = $grammarTransformations[$forms];
3849                         }
3850
3851                         foreach ( array_values( $forms ) as $rule ) {
3852                                 $form = $rule[0];
3853
3854                                 if ( $form === '@metadata' ) {
3855                                         continue;
3856                                 }
3857
3858                                 $replacement = $rule[1];
3859
3860                                 $regex = '/' . addcslashes( $form, '/' ) . '/u';
3861                                 $patternMatches = preg_match( $regex, $word );
3862
3863                                 if ( $patternMatches === false ) {
3864                                         wfLogWarning(
3865                                                 'An error occurred while processing grammar. ' .
3866                                                 "Word: '$word'. Regex: /$form/."
3867                                         );
3868                                 } elseif ( $patternMatches === 1 ) {
3869                                         $word = preg_replace( $regex, $replacement, $word );
3870
3871                                         break;
3872                                 }
3873                         }
3874                 }
3875
3876                 return $word;
3877         }
3878
3879         /**
3880          * Get the grammar forms for the content language
3881          * @return array Array of grammar forms
3882          * @since 1.20
3883          */
3884         function getGrammarForms() {
3885                 global $wgGrammarForms;
3886                 if ( isset( $wgGrammarForms[$this->getCode()] )
3887                         && is_array( $wgGrammarForms[$this->getCode()] )
3888                 ) {
3889                         return $wgGrammarForms[$this->getCode()];
3890                 }
3891
3892                 return [];
3893         }
3894
3895         /**
3896          * Get the grammar transformations data for the language.
3897          * Used like grammar forms, with {{GRAMMAR}} and cases,
3898          * but uses pairs of regexes and replacements instead of code.
3899          *
3900          * @return array[] Array of grammar transformations.
3901          * @throws MWException
3902          * @since 1.28
3903          */
3904         public function getGrammarTransformations() {
3905                 $languageCode = $this->getCode();
3906
3907                 if ( self::$grammarTransformations === null ) {
3908                         self::$grammarTransformations = new MapCacheLRU( 10 );
3909                 }
3910
3911                 if ( self::$grammarTransformations->has( $languageCode ) ) {
3912                         return self::$grammarTransformations->get( $languageCode );
3913                 }
3914
3915                 $data = [];
3916
3917                 $grammarDataFile = __DIR__ . "/data/grammarTransformations/$languageCode.json";
3918                 if ( is_readable( $grammarDataFile ) ) {
3919                         $data = FormatJson::decode(
3920                                 file_get_contents( $grammarDataFile ),
3921                                 true
3922                         );
3923
3924                         if ( $data === null ) {
3925                                 throw new MWException( "Invalid grammar data for \"$languageCode\"." );
3926                         }
3927
3928                         self::$grammarTransformations->set( $languageCode, $data );
3929                 }
3930
3931                 return $data;
3932         }
3933
3934         /**
3935          * Provides an alternative text depending on specified gender.
3936          * Usage {{gender:username|masculine|feminine|unknown}}.
3937          * username is optional, in which case the gender of current user is used,
3938          * but only in (some) interface messages; otherwise default gender is used.
3939          *
3940          * If no forms are given, an empty string is returned. If only one form is
3941          * given, it will be returned unconditionally. These details are implied by
3942          * the caller and cannot be overridden in subclasses.
3943          *
3944          * If three forms are given, the default is to use the third (unknown) form.
3945          * If fewer than three forms are given, the default is to use the first (masculine) form.
3946          * These details can be overridden in subclasses.
3947          *
3948          * @param string $gender
3949          * @param array $forms
3950          *
3951          * @return string
3952          */
3953         function gender( $gender, $forms ) {
3954                 if ( !count( $forms ) ) {
3955                         return '';
3956                 }
3957                 $forms = $this->preConvertPlural( $forms, 2 );
3958                 if ( $gender === 'male' ) {
3959                         return $forms[0];
3960                 }
3961                 if ( $gender === 'female' ) {
3962                         return $forms[1];
3963                 }
3964                 return $forms[2] ?? $forms[0];
3965         }
3966
3967         /**
3968          * Plural form transformations, needed for some languages.
3969          * For example, there are 3 form of plural in Russian and Polish,
3970          * depending on "count mod 10". See [[w:Plural]]
3971          * For English it is pretty simple.
3972          *
3973          * Invoked by putting {{plural:count|wordform1|wordform2}}
3974          * or {{plural:count|wordform1|wordform2|wordform3}}
3975          *
3976          * Example: {{plural:{{NUMBEROFARTICLES}}|article|articles}}
3977          *
3978          * @param int $count Non-localized number
3979          * @param array $forms Different plural forms
3980          * @return string Correct form of plural for $count in this language
3981          */
3982         function convertPlural( $count, $forms ) {
3983                 // Handle explicit n=pluralform cases
3984                 $forms = $this->handleExplicitPluralForms( $count, $forms );
3985                 if ( is_string( $forms ) ) {
3986                         return $forms;
3987                 }
3988                 if ( !count( $forms ) ) {
3989                         return '';
3990                 }
3991
3992                 $pluralForm = $this->getPluralRuleIndexNumber( $count );
3993                 $pluralForm = min( $pluralForm, count( $forms ) - 1 );
3994                 return $forms[$pluralForm];
3995         }
3996
3997         /**
3998          * Handles explicit plural forms for Language::convertPlural()
3999          *
4000          * In {{PLURAL:$1|0=nothing|one|many}}, 0=nothing will be returned if $1 equals zero.
4001          * If an explicitly defined plural form matches the $count, then
4002          * string value returned, otherwise array returned for further consideration
4003          * by CLDR rules or overridden convertPlural().
4004          *
4005          * @since 1.23
4006          *
4007          * @param int $count Non-localized number
4008          * @param array $forms Different plural forms
4009          *
4010          * @return array|string
4011          */
4012         protected function handleExplicitPluralForms( $count, array $forms ) {
4013                 foreach ( $forms as $index => $form ) {
4014                         if ( preg_match( '/\d+=/i', $form ) ) {
4015                                 $pos = strpos( $form, '=' );
4016                                 if ( substr( $form, 0, $pos ) === (string)$count ) {
4017                                         return substr( $form, $pos + 1 );
4018                                 }
4019                                 unset( $forms[$index] );
4020                         }
4021                 }
4022                 return array_values( $forms );
4023         }
4024
4025         /**
4026          * Checks that convertPlural was given an array and pads it to requested
4027          * amount of forms by copying the last one.
4028          *
4029          * @param array $forms
4030          * @param int $count How many forms should there be at least
4031          * @return array Padded array of forms
4032          */
4033         protected function preConvertPlural( /* Array */ $forms, $count ) {
4034                 return array_pad( $forms, $count, end( $forms ) );
4035         }
4036
4037         /**
4038          * Wraps argument with unicode control characters for directionality safety
4039          *
4040          * This solves the problem where directionality-neutral characters at the edge of
4041          * the argument string get interpreted with the wrong directionality from the
4042          * enclosing context, giving renderings that look corrupted like "(Ben_(WMF".
4043          *
4044          * The wrapping is LRE...PDF or RLE...PDF, depending on the detected
4045          * directionality of the argument string, using the BIDI algorithm's own "First
4046          * strong directional codepoint" rule. Essentially, this works round the fact that
4047          * there is no embedding equivalent of U+2068 FSI (isolation with heuristic
4048          * direction inference). The latter is cleaner but still not widely supported.
4049          *
4050          * @param string $text Text to wrap
4051          * @return string Text, wrapped in LRE...PDF or RLE...PDF or nothing
4052          */
4053         public function embedBidi( $text = '' ) {
4054                 $dir = self::strongDirFromContent( $text );
4055                 if ( $dir === 'ltr' ) {
4056                         // Wrap in LEFT-TO-RIGHT EMBEDDING ... POP DIRECTIONAL FORMATTING
4057                         return self::$lre . $text . self::$pdf;
4058                 }
4059                 if ( $dir === 'rtl' ) {
4060                         // Wrap in RIGHT-TO-LEFT EMBEDDING ... POP DIRECTIONAL FORMATTING
4061                         return self::$rle . $text . self::$pdf;
4062                 }
4063                 // No strong directionality: do not wrap
4064                 return $text;
4065         }
4066
4067         /**
4068          * @todo Maybe translate block durations.  Note that this function is somewhat misnamed: it
4069          * deals with translating the *duration* ("1 week", "4 days", etc), not the expiry time
4070          * (which is an absolute timestamp). Please note: do NOT add this blindly, as it is used
4071          * on old expiry lengths recorded in log entries. You'd need to provide the start date to
4072          * match up with it.
4073          *
4074          * @param string $str The validated block duration in English
4075          * @param User|null $user User object to use timezone from or null for $wgUser
4076          * @param int $now Current timestamp, for formatting relative block durations
4077          * @return string Somehow translated block duration
4078          * @see LanguageFi.php for example implementation
4079          */
4080         function translateBlockExpiry( $str, User $user = null, $now = 0 ) {
4081                 $duration = SpecialBlock::getSuggestedDurations( $this );
4082                 foreach ( $duration as $show => $value ) {
4083                         if ( strcmp( $str, $value ) == 0 ) {
4084                                 return htmlspecialchars( trim( $show ) );
4085                         }
4086                 }
4087
4088                 if ( wfIsInfinity( $str ) ) {
4089                         foreach ( $duration as $show => $value ) {
4090                                 if ( wfIsInfinity( $value ) ) {
4091                                         return htmlspecialchars( trim( $show ) );
4092                                 }
4093                         }
4094                 }
4095
4096                 // If all else fails, return a standard duration or timestamp description.
4097                 $time = strtotime( $str, $now );
4098                 if ( $time === false ) { // Unknown format. Return it as-is in case.
4099                         return $str;
4100                 } elseif ( $time !== strtotime( $str, $now + 1 ) ) { // It's a relative timestamp.
4101                         // The result differs based on current time, so the difference
4102                         // is a fixed duration length.
4103                         return $this->formatDuration( $time - $now );
4104                 } else { // It's an absolute timestamp.
4105                         if ( $time === 0 ) {
4106                                 // wfTimestamp() handles 0 as current time instead of epoch.
4107                                 $time = '19700101000000';
4108                         }
4109                         if ( $user ) {
4110                                 return $this->userTimeAndDate( $time, $user );
4111                         }
4112                         return $this->timeanddate( $time );
4113                 }
4114         }
4115
4116         /**
4117          * languages like Chinese need to be segmented in order for the diff
4118          * to be of any use
4119          *
4120          * @param string $text
4121          * @return string
4122          */
4123         public function segmentForDiff( $text ) {
4124                 return $text;
4125         }
4126
4127         /**
4128          * and unsegment to show the result
4129          *
4130          * @param string $text
4131          * @return string
4132          */
4133         public function unsegmentForDiff( $text ) {
4134                 return $text;
4135         }
4136
4137         /**
4138          * Return the LanguageConverter used in the Language
4139          *
4140          * @since 1.19
4141          * @return LanguageConverter
4142          */
4143         public function getConverter() {
4144                 return $this->mConverter;
4145         }
4146
4147         /**
4148          * convert text to a variant
4149          *
4150          * @param string $text text to convert
4151          * @param string|bool $variant variant to convert to, or false to use the user's preferred
4152          *      variant (if logged in), or the project default variant
4153          * @return string the converted string
4154          */
4155         public function autoConvert( $text, $variant = false ) {
4156                 return $this->mConverter->autoConvert( $text, $variant );
4157         }
4158
4159         /**
4160          * convert text to all supported variants
4161          *
4162          * @param string $text
4163          * @return array
4164          */
4165         public function autoConvertToAllVariants( $text ) {
4166                 return $this->mConverter->autoConvertToAllVariants( $text );
4167         }
4168
4169         /**
4170          * convert text to different variants of a language.
4171          *
4172          * @warning Glossary state is maintained between calls. This means
4173          *  if you pass unescaped text to this method it can cause an XSS
4174          *  in later calls to this method, even if the later calls have properly
4175          *  escaped the input. Never feed this method user controlled text that
4176          *  is not properly escaped!
4177          * @param string $text Content that has been already escaped for use in HTML
4178          * @return string HTML
4179          */
4180         public function convert( $text ) {
4181                 return $this->mConverter->convert( $text );
4182         }
4183
4184         /**
4185          * Convert a Title object to a string in the preferred variant
4186          *
4187          * @param Title $title
4188          * @return string
4189          */
4190         public function convertTitle( $title ) {
4191                 return $this->mConverter->convertTitle( $title );
4192         }
4193
4194         /**
4195          * Convert a namespace index to a string in the preferred variant
4196          *
4197          * @param int $ns namespace index (https://www.mediawiki.org/wiki/Manual:Namespace)
4198          * @param string|null $variant variant to convert to, or null to use the user's preferred
4199          *      variant (if logged in), or the project default variant
4200          * @return string a string representation of the namespace
4201          */
4202         public function convertNamespace( $ns, $variant = null ) {
4203                 return $this->mConverter->convertNamespace( $ns, $variant );
4204         }
4205
4206         /**
4207          * Check if this is a language with variants
4208          *
4209          * @return bool
4210          */
4211         public function hasVariants() {
4212                 return count( $this->getVariants() ) > 1;
4213         }
4214
4215         /**
4216          * Strict check if the language has the specific variant.
4217          *
4218          * Compare to LanguageConverter::validateVariant() which does a more
4219          * lenient check and attempts to coerce the given code to a valid one.
4220          *
4221          * @since 1.19
4222          * @param string $variant
4223          * @return bool
4224          */
4225         public function hasVariant( $variant ) {
4226                 return $variant && ( $variant === $this->mConverter->validateVariant( $variant ) );
4227         }
4228
4229         /**
4230          * Perform output conversion on a string, and encode for safe HTML output.
4231          * @param string $text Text to be converted
4232          * @return string
4233          * @todo this should get integrated somewhere sane
4234          */
4235         public function convertHtml( $text ) {
4236                 return htmlspecialchars( $this->convert( $text ) );
4237         }
4238
4239         /**
4240          * @param string $key
4241          * @return string
4242          */
4243         public function convertCategoryKey( $key ) {
4244                 return $this->mConverter->convertCategoryKey( $key );
4245         }
4246
4247         /**
4248          * Get the list of variants supported by this language
4249          * see sample implementation in LanguageZh.php
4250          *
4251          * @return string[] An array of language codes
4252          */
4253         public function getVariants() {
4254                 return $this->mConverter->getVariants();
4255         }
4256
4257         /**
4258          * @return string
4259          */
4260         public function getPreferredVariant() {
4261                 return $this->mConverter->getPreferredVariant();
4262         }
4263
4264         /**
4265          * @return string
4266          */
4267         public function getDefaultVariant() {
4268                 return $this->mConverter->getDefaultVariant();
4269         }
4270
4271         /**
4272          * @return string
4273          */
4274         public function getURLVariant() {
4275                 return $this->mConverter->getURLVariant();
4276         }
4277
4278         /**
4279          * If a language supports multiple variants, it is
4280          * possible that non-existing link in one variant
4281          * actually exists in another variant. this function
4282          * tries to find it. See e.g. LanguageZh.php
4283          * The input parameters may be modified upon return
4284          *
4285          * @param string &$link The name of the link
4286          * @param Title &$nt The title object of the link
4287          * @param bool $ignoreOtherCond To disable other conditions when
4288          *   we need to transclude a template or update a category's link
4289          */
4290         public function findVariantLink( &$link, &$nt, $ignoreOtherCond = false ) {
4291                 $this->mConverter->findVariantLink( $link, $nt, $ignoreOtherCond );
4292         }
4293
4294         /**
4295          * returns language specific options used by User::getPageRenderHash()
4296          * for example, the preferred language variant
4297          *
4298          * @return string
4299          */
4300         function getExtraHashOptions() {
4301                 return $this->mConverter->getExtraHashOptions();
4302         }
4303
4304         /**
4305          * For languages that support multiple variants, the title of an
4306          * article may be displayed differently in different variants. this
4307          * function returns the apporiate title defined in the body of the article.
4308          *
4309          * @return string
4310          */
4311         public function getParsedTitle() {
4312                 return $this->mConverter->getParsedTitle();
4313         }
4314
4315         /**
4316          * Refresh the cache of conversion tables when
4317          * MediaWiki:Conversiontable* is updated.
4318          *
4319          * @param Title $title The Title of the page being updated
4320          */
4321         public function updateConversionTable( Title $title ) {
4322                 $this->mConverter->updateConversionTable( $title );
4323         }
4324
4325         /**
4326          * A regular expression to match legal word-trailing characters
4327          * which should be merged onto a link of the form [[foo]]bar.
4328          *
4329          * @return string
4330          */
4331         public function linkTrail() {
4332                 return $this->localisationCache->getItem( $this->mCode, 'linkTrail' );
4333         }
4334
4335         /**
4336          * A regular expression character set to match legal word-prefixing
4337          * characters which should be merged onto a link of the form foo[[bar]].
4338          *
4339          * @return string
4340          */
4341         public function linkPrefixCharset() {
4342                 return $this->localisationCache->getItem( $this->mCode, 'linkPrefixCharset' );
4343         }
4344
4345         /**
4346          * Get the "parent" language which has a converter to convert a "compatible" language
4347          * (in another variant) to this language (eg. zh for zh-cn, but not en for en-gb).
4348          *
4349          * @return Language|null
4350          * @since 1.22
4351          */
4352         public function getParentLanguage() {
4353                 if ( $this->mParentLanguage !== false ) {
4354                         return $this->mParentLanguage;
4355                 }
4356
4357                 $code = explode( '-', $this->getCode() )[0];
4358                 if ( !in_array( $code, LanguageConverter::$languagesWithVariants ) ) {
4359                         $this->mParentLanguage = null;
4360                         return null;
4361                 }
4362                 $lang = self::factory( $code );
4363                 if ( !$lang->hasVariant( $this->getCode() ) ) {
4364                         $this->mParentLanguage = null;
4365                         return null;
4366                 }
4367
4368                 $this->mParentLanguage = $lang;
4369                 return $lang;
4370         }
4371
4372         /**
4373          * Compare with an other language object
4374          *
4375          * @since 1.28
4376          * @param Language $lang
4377          * @return bool
4378          */
4379         public function equals( Language $lang ) {
4380                 return $lang === $this || $lang->getCode() === $this->mCode;
4381         }
4382
4383         /**
4384          * Get the internal language code for this language object
4385          *
4386          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4387          * htmlspecialchars() or similar
4388          *
4389          * @return string
4390          */
4391         public function getCode() {
4392                 return $this->mCode;
4393         }
4394
4395         /**
4396          * Get the code in BCP 47 format which we can use
4397          * inside of html lang="" tags.
4398          *
4399          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4400          * htmlspecialchars() or similar.
4401          *
4402          * @since 1.19
4403          * @return string
4404          */
4405         public function getHtmlCode() {
4406                 if ( is_null( $this->mHtmlCode ) ) {
4407                         $this->mHtmlCode = LanguageCode::bcp47( $this->getCode() );
4408                 }
4409                 return $this->mHtmlCode;
4410         }
4411
4412         /**
4413          * Get the language code from a file name. Inverse of getFileName()
4414          * @param string $filename $prefix . $languageCode . $suffix
4415          * @param string $prefix Prefix before the language code
4416          * @param string $suffix Suffix after the language code
4417          * @return string Language code, or false if $prefix or $suffix isn't found
4418          */
4419         public static function getCodeFromFileName( $filename, $prefix = 'Language', $suffix = '.php' ) {
4420                 $m = null;
4421                 preg_match( '/' . preg_quote( $prefix, '/' ) . '([A-Z][a-z_]+)' .
4422                         preg_quote( $suffix, '/' ) . '/', $filename, $m );
4423                 if ( !count( $m ) ) {
4424                         return false;
4425                 }
4426                 return str_replace( '_', '-', strtolower( $m[1] ) );
4427         }
4428
4429         /**
4430          * @param string $code
4431          * @param bool $fallback Whether we're going through language fallback chain
4432          * @return string Name of the language class
4433          */
4434         public static function classFromCode( $code, $fallback = true ) {
4435                 if ( $fallback && $code == 'en' ) {
4436                         return 'Language';
4437                 } else {
4438                         return 'Language' . str_replace( '-', '_', ucfirst( $code ) );
4439                 }
4440         }
4441
4442         /**
4443          * Get the name of a file for a certain language code
4444          * @param string $prefix Prepend this to the filename
4445          * @param string $code Language code
4446          * @param string $suffix Append this to the filename
4447          * @throws MWException
4448          * @return string $prefix . $mangledCode . $suffix
4449          */
4450         public static function getFileName( $prefix, $code, $suffix = '.php' ) {
4451                 if ( !self::isValidBuiltInCode( $code ) ) {
4452                         throw new MWException( "Invalid language code \"$code\"" );
4453                 }
4454
4455                 return $prefix . str_replace( '-', '_', ucfirst( $code ) ) . $suffix;
4456         }
4457
4458         /**
4459          * @param string $code
4460          * @return string
4461          */
4462         public static function getMessagesFileName( $code ) {
4463                 global $IP;
4464                 $file = self::getFileName( "$IP/languages/messages/Messages", $code, '.php' );
4465                 Hooks::run( 'Language::getMessagesFileName', [ $code, &$file ] );
4466                 return $file;
4467         }
4468
4469         /**
4470          * @param string $code
4471          * @return string
4472          * @throws MWException
4473          * @since 1.23
4474          */
4475         public static function getJsonMessagesFileName( $code ) {
4476                 global $IP;
4477
4478                 if ( !self::isValidBuiltInCode( $code ) ) {
4479                         throw new MWException( "Invalid language code \"$code\"" );
4480                 }
4481
4482                 return "$IP/languages/i18n/$code.json";
4483         }
4484
4485         /**
4486          * Get the first fallback for a given language.
4487          *
4488          * @param string $code
4489          *
4490          * @return bool|string
4491          */
4492         public static function getFallbackFor( $code ) {
4493                 $fallbacks = self::getFallbacksFor( $code );
4494                 if ( $fallbacks ) {
4495                         return $fallbacks[0];
4496                 }
4497                 return false;
4498         }
4499
4500         /**
4501          * Get the ordered list of fallback languages.
4502          *
4503          * @since 1.19
4504          * @param string $code Language code
4505          * @param int $mode Fallback mode, either MESSAGES_FALLBACKS (which always falls back to 'en'),
4506          * or STRICT_FALLBACKS (whic honly falls back to 'en' when explicitly defined)
4507          * @throws MWException
4508          * @return array List of language codes
4509          */
4510         public static function getFallbacksFor( $code, $mode = self::MESSAGES_FALLBACKS ) {
4511                 if ( $code === 'en' || !self::isValidBuiltInCode( $code ) ) {
4512                         return [];
4513                 }
4514                 switch ( $mode ) {
4515                         case self::MESSAGES_FALLBACKS:
4516                                 // For unknown languages, fallbackSequence returns an empty array,
4517                                 // hardcode fallback to 'en' in that case as English messages are
4518                                 // always defined.
4519                                 return self::getLocalisationCache()->getItem( $code, 'fallbackSequence' ) ?: [ 'en' ];
4520                         case self::STRICT_FALLBACKS:
4521                                 // Use this mode when you don't want to fallback to English unless
4522                                 // explicitly defined, for example when you have language-variant icons
4523                                 // and an international language-independent fallback.
4524                                 return self::getLocalisationCache()->getItem( $code, 'originalFallbackSequence' );
4525                         default:
4526                                 throw new MWException( "Invalid fallback mode \"$mode\"" );
4527                 }
4528         }
4529
4530         /**
4531          * Get the ordered list of fallback languages, ending with the fallback
4532          * language chain for the site language.
4533          *
4534          * @since 1.22
4535          * @param string $code Language code
4536          * @return array [ fallbacks, site fallbacks ]
4537          */
4538         public static function getFallbacksIncludingSiteLanguage( $code ) {
4539                 global $wgLanguageCode;
4540
4541                 // Usually, we will only store a tiny number of fallback chains, so we
4542                 // keep them in static memory.
4543                 $cacheKey = "{$code}-{$wgLanguageCode}";
4544
4545                 if ( !array_key_exists( $cacheKey, self::$fallbackLanguageCache ) ) {
4546                         $fallbacks = self::getFallbacksFor( $code );
4547
4548                         // Append the site's fallback chain, including the site language itself
4549                         $siteFallbacks = self::getFallbacksFor( $wgLanguageCode );
4550                         array_unshift( $siteFallbacks, $wgLanguageCode );
4551
4552                         // Eliminate any languages already included in the chain
4553                         $siteFallbacks = array_diff( $siteFallbacks, $fallbacks );
4554
4555                         self::$fallbackLanguageCache[$cacheKey] = [ $fallbacks, $siteFallbacks ];
4556                 }
4557                 return self::$fallbackLanguageCache[$cacheKey];
4558         }
4559
4560         /**
4561          * Get all messages for a given language
4562          * WARNING: this may take a long time. If you just need all message *keys*
4563          * but need the *contents* of only a few messages, consider using getMessageKeysFor().
4564          *
4565          * @param string $code
4566          *
4567          * @return array
4568          */
4569         public static function getMessagesFor( $code ) {
4570                 return self::getLocalisationCache()->getItem( $code, 'messages' );
4571         }
4572
4573         /**
4574          * Get a message for a given language
4575          *
4576          * @param string $key
4577          * @param string $code
4578          *
4579          * @return string
4580          */
4581         public static function getMessageFor( $key, $code ) {
4582                 return self::getLocalisationCache()->getSubitem( $code, 'messages', $key );
4583         }
4584
4585         /**
4586          * Get all message keys for a given language. This is a faster alternative to
4587          * array_keys( Language::getMessagesFor( $code ) )
4588          *
4589          * @since 1.19
4590          * @param string $code Language code
4591          * @return array Array of message keys (strings)
4592          */
4593         public static function getMessageKeysFor( $code ) {
4594                 return self::getLocalisationCache()->getSubitemList( $code, 'messages' );
4595         }
4596
4597         /**
4598          * @param string $talk
4599          * @return string
4600          */
4601         function fixVariableInNamespace( $talk ) {
4602                 if ( strpos( $talk, '$1' ) === false ) {
4603                         return $talk;
4604                 }
4605
4606                 global $wgMetaNamespace;
4607                 $talk = str_replace( '$1', $wgMetaNamespace, $talk );
4608
4609                 # Allow grammar transformations
4610                 # Allowing full message-style parsing would make simple requests
4611                 # such as action=raw much more expensive than they need to be.
4612                 # This will hopefully cover most cases.
4613                 $talk = preg_replace_callback( '/{{grammar:(.*?)\|(.*?)}}/i',
4614                         [ $this, 'replaceGrammarInNamespace' ], $talk );
4615                 return str_replace( ' ', '_', $talk );
4616         }
4617
4618         /**
4619          * @param string $m
4620          * @return string
4621          */
4622         function replaceGrammarInNamespace( $m ) {
4623                 return $this->convertGrammar( trim( $m[2] ), trim( $m[1] ) );
4624         }
4625
4626         /**
4627          * Decode an expiry (block, protection, etc) which has come from the DB
4628          *
4629          * @param string $expiry Database expiry String
4630          * @param bool|int $format True to process using language functions, or TS_ constant
4631          *     to return the expiry in a given timestamp
4632          * @param string $infinity If $format is not true, use this string for infinite expiry
4633          * @return string
4634          * @since 1.18
4635          */
4636         public function formatExpiry( $expiry, $format = true, $infinity = 'infinity' ) {
4637                 static $dbInfinity;
4638                 if ( $dbInfinity === null ) {
4639                         $dbInfinity = wfGetDB( DB_REPLICA )->getInfinity();
4640                 }
4641
4642                 if ( $expiry == '' || $expiry === 'infinity' || $expiry == $dbInfinity ) {
4643                         return $format === true
4644                                 ? $this->getMessageFromDB( 'infiniteblock' )
4645                                 : $infinity;
4646                 } else {
4647                         return $format === true
4648                                 ? $this->timeanddate( $expiry, /* User preference timezone */ true )
4649                                 : wfTimestamp( $format, $expiry );
4650                 }
4651         }
4652
4653         /**
4654          * Formats a time given in seconds into a string representation of that time.
4655          *
4656          * @param int|float $seconds
4657          * @param array $format An optional argument that formats the returned string in different ways:
4658          *   If $format['avoid'] === 'avoidhours': don't show hours, just show days
4659          *   If $format['avoid'] === 'avoidseconds': don't show seconds if $seconds >= 1 hour,
4660          *   If $format['avoid'] === 'avoidminutes': don't show seconds/minutes if $seconds > 48 hours,
4661          *   If $format['noabbrevs'] is true: use 'seconds' and friends instead of 'seconds-abbrev'
4662          *     and friends.
4663          * @note For backwards compatibility, $format may also be one of the strings 'avoidseconds'
4664          *     or 'avoidminutes'.
4665          * @return string
4666          */
4667         function formatTimePeriod( $seconds, $format = [] ) {
4668                 if ( !is_array( $format ) ) {
4669                         $format = [ 'avoid' => $format ]; // For backwards compatibility
4670                 }
4671                 if ( !isset( $format['avoid'] ) ) {
4672                         $format['avoid'] = false;
4673                 }
4674                 if ( !isset( $format['noabbrevs'] ) ) {
4675                         $format['noabbrevs'] = false;
4676                 }
4677                 $secondsMsg = wfMessage(
4678                         $format['noabbrevs'] ? 'seconds' : 'seconds-abbrev' )->inLanguage( $this );
4679                 $minutesMsg = wfMessage(
4680                         $format['noabbrevs'] ? 'minutes' : 'minutes-abbrev' )->inLanguage( $this );
4681                 $hoursMsg = wfMessage(
4682                         $format['noabbrevs'] ? 'hours' : 'hours-abbrev' )->inLanguage( $this );
4683                 $daysMsg = wfMessage(
4684                         $format['noabbrevs'] ? 'days' : 'days-abbrev' )->inLanguage( $this );
4685
4686                 if ( round( $seconds * 10 ) < 100 ) {
4687                         $s = $this->formatNum( sprintf( "%.1f", round( $seconds * 10 ) / 10 ) );
4688                         $s = $secondsMsg->params( $s )->text();
4689                 } elseif ( round( $seconds ) < 60 ) {
4690                         $s = $this->formatNum( round( $seconds ) );
4691                         $s = $secondsMsg->params( $s )->text();
4692                 } elseif ( round( $seconds ) < 3600 ) {
4693                         $minutes = floor( $seconds / 60 );
4694                         $secondsPart = round( fmod( $seconds, 60 ) );
4695                         if ( $secondsPart == 60 ) {
4696                                 $secondsPart = 0;
4697                                 $minutes++;
4698                         }
4699                         $s = $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4700                         $s .= ' ';
4701                         $s .= $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4702                 } elseif ( round( $seconds ) <= 2 * 86400 ) {
4703                         $hours = floor( $seconds / 3600 );
4704                         $minutes = floor( ( $seconds - $hours * 3600 ) / 60 );
4705                         $secondsPart = round( $seconds - $hours * 3600 - $minutes * 60 );
4706                         if ( $secondsPart == 60 ) {
4707                                 $secondsPart = 0;
4708                                 $minutes++;
4709                         }
4710                         if ( $minutes == 60 ) {
4711                                 $minutes = 0;
4712                                 $hours++;
4713                         }
4714                         $s = $hoursMsg->params( $this->formatNum( $hours ) )->text();
4715                         $s .= ' ';
4716                         $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4717                         if ( !in_array( $format['avoid'], [ 'avoidseconds', 'avoidminutes', 'avoidhours' ] ) ) {
4718                                 $s .= ' ' . $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4719                         }
4720                 } else {
4721                         $days = floor( $seconds / 86400 );
4722                         if ( $format['avoid'] === 'avoidhours' ) {
4723                                 $hours = round( ( $seconds - $days * 86400 ) / 3600 );
4724                                 if ( $hours == 24 ) {
4725                                         $hours = 0;
4726                                         $days++;
4727                                 }
4728                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4729                         } elseif ( $format['avoid'] === 'avoidminutes' ) {
4730                                 $hours = round( ( $seconds - $days * 86400 ) / 3600 );
4731                                 if ( $hours == 24 ) {
4732                                         $hours = 0;
4733                                         $days++;
4734                                 }
4735                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4736                                 $s .= ' ';
4737                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4738                         } elseif ( $format['avoid'] === 'avoidseconds' ) {
4739                                 $hours = floor( ( $seconds - $days * 86400 ) / 3600 );
4740                                 $minutes = round( ( $seconds - $days * 86400 - $hours * 3600 ) / 60 );
4741                                 if ( $minutes == 60 ) {
4742                                         $minutes = 0;
4743                                         $hours++;
4744                                 }
4745                                 if ( $hours == 24 ) {
4746                                         $hours = 0;
4747                                         $days++;
4748                                 }
4749                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4750                                 $s .= ' ';
4751                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4752                                 $s .= ' ';
4753                                 $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4754                         } else {
4755                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4756                                 $s .= ' ';
4757                                 $s .= $this->formatTimePeriod( $seconds - $days * 86400, $format );
4758                         }
4759                 }
4760                 return $s;
4761         }
4762
4763         /**
4764          * Format a bitrate for output, using an appropriate
4765          * unit (bps, kbps, Mbps, Gbps, Tbps, Pbps, Ebps, Zbps or Ybps) according to
4766          *   the magnitude in question.
4767          *
4768          * This use base 1000. For base 1024 use formatSize(), for another base
4769          * see formatComputingNumbers().
4770          *
4771          * @param int $bps
4772          * @return string
4773          */
4774         function formatBitrate( $bps ) {
4775                 return $this->formatComputingNumbers( $bps, 1000, "bitrate-$1bits" );
4776         }
4777
4778         /**
4779          * @param int $size Size of the unit
4780          * @param int $boundary Size boundary (1000, or 1024 in most cases)
4781          * @param string $messageKey Message key to be uesd
4782          * @return string
4783          */
4784         function formatComputingNumbers( $size, $boundary, $messageKey ) {
4785                 if ( $size <= 0 ) {
4786                         return str_replace( '$1', $this->formatNum( $size ),
4787                                 $this->getMessageFromDB( str_replace( '$1', '', $messageKey ) )
4788                         );
4789                 }
4790                 $sizes = [ '', 'kilo', 'mega', 'giga', 'tera', 'peta', 'exa', 'zeta', 'yotta' ];
4791                 $index = 0;
4792
4793                 $maxIndex = count( $sizes ) - 1;
4794                 while ( $size >= $boundary && $index < $maxIndex ) {
4795                         $index++;
4796                         $size /= $boundary;
4797                 }
4798
4799                 // For small sizes no decimal places necessary
4800                 $round = 0;
4801                 if ( $index > 1 ) {
4802                         // For MB and bigger two decimal places are smarter
4803                         $round = 2;
4804                 }
4805                 $msg = str_replace( '$1', $sizes[$index], $messageKey );
4806
4807                 $size = round( $size, $round );
4808                 $text = $this->getMessageFromDB( $msg );
4809                 return str_replace( '$1', $this->formatNum( $size ), $text );
4810         }
4811
4812         /**
4813          * Format a size in bytes for output, using an appropriate
4814          * unit (B, KB, MB, GB, TB, PB, EB, ZB or YB) according to the magnitude in question
4815          *
4816          * This method use base 1024. For base 1000 use formatBitrate(), for
4817          * another base see formatComputingNumbers()
4818          *
4819          * @param int $size Size to format
4820          * @return string Plain text (not HTML)
4821          */
4822         function formatSize( $size ) {
4823                 return $this->formatComputingNumbers( $size, 1024, "size-$1bytes" );
4824         }
4825
4826         /**
4827          * Make a list item, used by various special pages
4828          *
4829          * @param string $page Page link
4830          * @param string $details HTML safe text between brackets
4831          * @param bool $oppositedm Add the direction mark opposite to your
4832          *   language, to display text properly
4833          * @return string HTML escaped
4834          */
4835         function specialList( $page, $details, $oppositedm = true ) {
4836                 if ( !$details ) {
4837                         return $page;
4838                 }
4839
4840                 $dirmark = ( $oppositedm ? $this->getDirMark( true ) : '' ) . $this->getDirMark();
4841                 return $page .
4842                         $dirmark .
4843                         $this->msg( 'word-separator' )->escaped() .
4844                         $this->msg( 'parentheses' )->rawParams( $details )->escaped();
4845         }
4846
4847         /**
4848          * Generate (prev x| next x) (20|50|100...) type links for paging
4849          *
4850          * @param Title $title Title object to link
4851          * @param int $offset
4852          * @param int $limit
4853          * @param array $query Optional URL query parameter string
4854          * @param bool $atend Optional param for specified if this is the last page
4855          * @return string
4856          * @deprecated since 1.34, use PrevNextNavigationRenderer::buildPrevNextNavigation()
4857          *  instead.
4858          */
4859         public function viewPrevNext( Title $title, $offset, $limit,
4860                 array $query = [], $atend = false
4861         ) {
4862                 wfDeprecated( __METHOD__, '1.34' );
4863                 // @todo FIXME: Why on earth this needs one message for the text and another one for tooltip?
4864
4865                 # Make 'previous' link
4866                 $prev = wfMessage( 'prevn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4867                 if ( $offset > 0 ) {
4868                         $plink = $this->numLink( $title, max( $offset - $limit, 0 ), $limit,
4869                                 $query, $prev, 'prevn-title', 'mw-prevlink' );
4870                 } else {
4871                         $plink = htmlspecialchars( $prev );
4872                 }
4873
4874                 # Make 'next' link
4875                 $next = wfMessage( 'nextn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4876                 if ( $atend ) {
4877                         $nlink = htmlspecialchars( $next );
4878                 } else {
4879                         $nlink = $this->numLink( $title, $offset + $limit, $limit,
4880                                 $query, $next, 'nextn-title', 'mw-nextlink' );
4881                 }
4882
4883                 # Make links to set number of items per page
4884                 $numLinks = [];
4885                 foreach ( [ 20, 50, 100, 250, 500 ] as $num ) {
4886                         $numLinks[] = $this->numLink( $title, $offset, $num,
4887                                 $query, $this->formatNum( $num ), 'shown-title', 'mw-numlink' );
4888                 }
4889
4890                 return wfMessage( 'viewprevnext' )->inLanguage( $this )->title( $title
4891                         )->rawParams( $plink, $nlink, $this->pipeList( $numLinks ) )->escaped();
4892         }
4893
4894         /**
4895          * Helper function for viewPrevNext() that generates links
4896          *
4897          * @param Title $title Title object to link
4898          * @param int $offset
4899          * @param int $limit
4900          * @param array $query Extra query parameters
4901          * @param string $link Text to use for the link; will be escaped
4902          * @param string $tooltipMsg Name of the message to use as tooltip
4903          * @param string $class Value of the "class" attribute of the link
4904          * @return string HTML fragment
4905          */
4906         private function numLink( Title $title, $offset, $limit, array $query, $link,
4907                 $tooltipMsg, $class
4908         ) {
4909                 $query = [ 'limit' => $limit, 'offset' => $offset ] + $query;
4910                 $tooltip = wfMessage( $tooltipMsg )->inLanguage( $this )->title( $title )
4911                         ->numParams( $limit )->text();
4912
4913                 return Html::element( 'a', [ 'href' => $title->getLocalURL( $query ),
4914                         'title' => $tooltip, 'class' => $class ], $link );
4915         }
4916
4917         /**
4918          * Get the conversion rule title, if any.
4919          *
4920          * @return string
4921          */
4922         public function getConvRuleTitle() {
4923                 return $this->mConverter->getConvRuleTitle();
4924         }
4925
4926         /**
4927          * Get the compiled plural rules for the language
4928          * @since 1.20
4929          * @return array Associative array with plural form, and plural rule as key-value pairs
4930          */
4931         public function getCompiledPluralRules() {
4932                 $pluralRules =
4933                         $this->localisationCache->getItem( strtolower( $this->mCode ), 'compiledPluralRules' );
4934                 $fallbacks = self::getFallbacksFor( $this->mCode );
4935                 if ( !$pluralRules ) {
4936                         foreach ( $fallbacks as $fallbackCode ) {
4937                                 $pluralRules = $this->localisationCache
4938                                         ->getItem( strtolower( $fallbackCode ), 'compiledPluralRules' );
4939                                 if ( $pluralRules ) {
4940                                         break;
4941                                 }
4942                         }
4943                 }
4944                 return $pluralRules;
4945         }
4946
4947         /**
4948          * Get the plural rules for the language
4949          * @since 1.20
4950          * @return array Associative array with plural form number and plural rule as key-value pairs
4951          */
4952         public function getPluralRules() {
4953                 $pluralRules =
4954                         $this->localisationCache->getItem( strtolower( $this->mCode ), 'pluralRules' );
4955                 $fallbacks = self::getFallbacksFor( $this->mCode );
4956                 if ( !$pluralRules ) {
4957                         foreach ( $fallbacks as $fallbackCode ) {
4958                                 $pluralRules = $this->localisationCache
4959                                         ->getItem( strtolower( $fallbackCode ), 'pluralRules' );
4960                                 if ( $pluralRules ) {
4961                                         break;
4962                                 }
4963                         }
4964                 }
4965                 return $pluralRules;
4966         }
4967
4968         /**
4969          * Get the plural rule types for the language
4970          * @since 1.22
4971          * @return array Associative array with plural form number and plural rule type as key-value pairs
4972          */
4973         public function getPluralRuleTypes() {
4974                 $pluralRuleTypes =
4975                         $this->localisationCache->getItem( strtolower( $this->mCode ), 'pluralRuleTypes' );
4976                 $fallbacks = self::getFallbacksFor( $this->mCode );
4977                 if ( !$pluralRuleTypes ) {
4978                         foreach ( $fallbacks as $fallbackCode ) {
4979                                 $pluralRuleTypes = $this->localisationCache
4980                                         ->getItem( strtolower( $fallbackCode ), 'pluralRuleTypes' );
4981                                 if ( $pluralRuleTypes ) {
4982                                         break;
4983                                 }
4984                         }
4985                 }
4986                 return $pluralRuleTypes;
4987         }
4988
4989         /**
4990          * Find the index number of the plural rule appropriate for the given number
4991          * @param int $number
4992          * @return int The index number of the plural rule
4993          */
4994         public function getPluralRuleIndexNumber( $number ) {
4995                 $pluralRules = $this->getCompiledPluralRules();
4996                 $form = Evaluator::evaluateCompiled( $number, $pluralRules );
4997                 return $form;
4998         }
4999
5000         /**
5001          * Find the plural rule type appropriate for the given number
5002          * For example, if the language is set to Arabic, getPluralType(5) should
5003          * return 'few'.
5004          * @since 1.22
5005          * @param int $number
5006          * @return string The name of the plural rule type, e.g. one, two, few, many
5007          */
5008         public function getPluralRuleType( $number ) {
5009                 $index = $this->getPluralRuleIndexNumber( $number );
5010                 $pluralRuleTypes = $this->getPluralRuleTypes();
5011                 return $pluralRuleTypes[$index] ?? 'other';
5012         }
5013 }