languages/Language.php

   1 <?php
   2 /**
   3  * Internationalisation code.
   4  * See https://www.mediawiki.org/wiki/Special:MyLanguage/Localisation for more information.
   5  *
   6  * This program is free software; you can redistribute it and/or modify
   7  * it under the terms of the GNU General Public License as published by
   8  * the Free Software Foundation; either version 2 of the License, or
   9  * (at your option) any later version.
  10  *
  11  * This program is distributed in the hope that it will be useful,
  12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  14  * GNU General Public License for more details.
  15  *
  16  * You should have received a copy of the GNU General Public License along
  17  * with this program; if not, write to the Free Software Foundation, Inc.,
  18  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  19  * http://www.gnu.org/copyleft/gpl.html
  20  *
  21  * @file
  22  * @ingroup Language
  23  */
  24
  25 /**
  26  * @defgroup Language Language
  27  */
  28
  29 use CLDRPluralRuleParser\Evaluator;
  30 use MediaWiki\MediaWikiServices;
  31 use Wikimedia\Assert\Assert;
  32
  33 /**
  34  * Internationalisation code
  35  * @ingroup Language
  36  */
  37 class Language {
  38         /**
  39          * Return autonyms in fetchLanguageName(s).
  40          * @since 1.32
  41          */
  42         const AS_AUTONYMS = null;
  43
  44         /**
  45          * Return all known languages in fetchLanguageName(s).
  46          * @since 1.32
  47          */
  48         const ALL = 'all';
  49
  50         /**
  51          * Return in fetchLanguageName(s) only the languages for which we have at
  52          * least some localisation.
  53          * @since 1.32
  54          */
  55         const SUPPORTED = 'mwfile';
  56
  57         /**
  58          * @var LanguageConverter
  59          */
  60         public $mConverter;
  61
  62         public $mVariants, $mCode, $mLoaded = false;
  63         public $mMagicExtensions = [];
  64         private $mHtmlCode = null, $mParentLanguage = false;
  65
  66         public $dateFormatStrings = [];
  67         public $mExtendedSpecialPageAliases;
  68
  69         /** @var array|null */
  70         protected $namespaceNames;
  71         protected $mNamespaceIds, $namespaceAliases;
  72
  73         /**
  74          * ReplacementArray object caches
  75          */
  76         public $transformData = [];
  77
  78         /**
  79          * @var LocalisationCache
  80          */
  81         public static $dataCache;
  82
  83         public static $mLangObjCache = [];
  84
  85         /**
  86          * Return a fallback chain for messages in getFallbacksFor
  87          * @since 1.32
  88          */
  89         const MESSAGES_FALLBACKS = 0;
  90
  91         /**
  92          * Return a strict fallback chain in getFallbacksFor
  93          * @since 1.32
  94          */
  95         const STRICT_FALLBACKS = 1;
  96
  97         public static $mWeekdayMsgs = [
  98                 'sunday', 'monday', 'tuesday', 'wednesday', 'thursday',
  99                 'friday', 'saturday'
 100         ];
 101
 102         public static $mWeekdayAbbrevMsgs = [
 103                 'sun', 'mon', 'tue', 'wed', 'thu', 'fri', 'sat'
 104         ];
 105
 106         public static $mMonthMsgs = [
 107                 'january', 'february', 'march', 'april', 'may_long', 'june',
 108                 'july', 'august', 'september', 'october', 'november',
 109                 'december'
 110         ];
 111         public static $mMonthGenMsgs = [
 112                 'january-gen', 'february-gen', 'march-gen', 'april-gen', 'may-gen', 'june-gen',
 113                 'july-gen', 'august-gen', 'september-gen', 'october-gen', 'november-gen',
 114                 'december-gen'
 115         ];
 116         public static $mMonthAbbrevMsgs = [
 117                 'jan', 'feb', 'mar', 'apr', 'may', 'jun', 'jul', 'aug',
 118                 'sep', 'oct', 'nov', 'dec'
 119         ];
 120
 121         public static $mIranianCalendarMonthMsgs = [
 122                 'iranian-calendar-m1', 'iranian-calendar-m2', 'iranian-calendar-m3',
 123                 'iranian-calendar-m4', 'iranian-calendar-m5', 'iranian-calendar-m6',
 124                 'iranian-calendar-m7', 'iranian-calendar-m8', 'iranian-calendar-m9',
 125                 'iranian-calendar-m10', 'iranian-calendar-m11', 'iranian-calendar-m12'
 126         ];
 127
 128         public static $mHebrewCalendarMonthMsgs = [
 129                 'hebrew-calendar-m1', 'hebrew-calendar-m2', 'hebrew-calendar-m3',
 130                 'hebrew-calendar-m4', 'hebrew-calendar-m5', 'hebrew-calendar-m6',
 131                 'hebrew-calendar-m7', 'hebrew-calendar-m8', 'hebrew-calendar-m9',
 132                 'hebrew-calendar-m10', 'hebrew-calendar-m11', 'hebrew-calendar-m12',
 133                 'hebrew-calendar-m6a', 'hebrew-calendar-m6b'
 134         ];
 135
 136         public static $mHebrewCalendarMonthGenMsgs = [
 137                 'hebrew-calendar-m1-gen', 'hebrew-calendar-m2-gen', 'hebrew-calendar-m3-gen',
 138                 'hebrew-calendar-m4-gen', 'hebrew-calendar-m5-gen', 'hebrew-calendar-m6-gen',
 139                 'hebrew-calendar-m7-gen', 'hebrew-calendar-m8-gen', 'hebrew-calendar-m9-gen',
 140                 'hebrew-calendar-m10-gen', 'hebrew-calendar-m11-gen', 'hebrew-calendar-m12-gen',
 141                 'hebrew-calendar-m6a-gen', 'hebrew-calendar-m6b-gen'
 142         ];
 143
 144         public static $mHijriCalendarMonthMsgs = [
 145                 'hijri-calendar-m1', 'hijri-calendar-m2', 'hijri-calendar-m3',
 146                 'hijri-calendar-m4', 'hijri-calendar-m5', 'hijri-calendar-m6',
 147                 'hijri-calendar-m7', 'hijri-calendar-m8', 'hijri-calendar-m9',
 148                 'hijri-calendar-m10', 'hijri-calendar-m11', 'hijri-calendar-m12'
 149         ];
 150
 151         /**
 152          * @since 1.20
 153          * @var array
 154          */
 155         public static $durationIntervals = [
 156                 'millennia' => 31556952000,
 157                 'centuries' => 3155695200,
 158                 'decades' => 315569520,
 159                 'years' => 31556952, // 86400 * ( 365 + ( 24 * 3 + 25 ) / 400 )
 160                 'weeks' => 604800,
 161                 'days' => 86400,
 162                 'hours' => 3600,
 163                 'minutes' => 60,
 164                 'seconds' => 1,
 165         ];
 166
 167         /**
 168          * Cache for language fallbacks.
 169          * @see Language::getFallbacksIncludingSiteLanguage
 170          * @since 1.21
 171          * @var array
 172          */
 173         private static $fallbackLanguageCache = [];
 174
 175         /**
 176          * Cache for grammar rules data
 177          * @var MapCacheLRU|null
 178          */
 179         private static $grammarTransformations;
 180
 181         /**
 182          * Cache for language names
 183          * @var HashBagOStuff|null
 184          */
 185         private static $languageNameCache;
 186
 187         /**
 188          * Unicode directional formatting characters, for embedBidi()
 189          */
 190         private static $lre = "\u{202A}"; // U+202A LEFT-TO-RIGHT EMBEDDING
 191         private static $rle = "\u{202B}"; // U+202B RIGHT-TO-LEFT EMBEDDING
 192         private static $pdf = "\u{202C}"; // U+202C POP DIRECTIONAL FORMATTING
 193
 194         /**
 195          * Directionality test regex for embedBidi(). Matches the first strong directionality codepoint:
 196          * - in group 1 if it is LTR
 197          * - in group 2 if it is RTL
 198          * Does not match if there is no strong directionality codepoint.
 199          *
 200          * The form is '/(?:([strong ltr codepoint])|([strong rtl codepoint]))/u' .
 201          *
 202          * Generated by UnicodeJS (see tools/strongDir) from the UCD; see
 203          * https://phabricator.wikimedia.org/diffusion/GUJS/ .
 204          */
 205         // @codeCoverageIgnoreStart
 206         // phpcs:ignore Generic.Files.LineLength
 207         private static $strongDirRegex = '/(?:([\x{41}-\x{5a}\x{61}-\x{7a}\x{aa}\x{b5}\x{ba}\x{c0}-\x{d6}\x{d8}-\x{f6}\x{f8}-\x{2b8}\x{2bb}-\x{2c1}\x{2d0}\x{2d1}\x{2e0}-\x{2e4}\x{2ee}\x{370}-\x{373}\x{376}\x{377}\x{37a}-\x{37d}\x{37f}\x{386}\x{388}-\x{38a}\x{38c}\x{38e}-\x{3a1}\x{3a3}-\x{3f5}\x{3f7}-\x{482}\x{48a}-\x{52f}\x{531}-\x{556}\x{559}-\x{55f}\x{561}-\x{587}\x{589}\x{903}-\x{939}\x{93b}\x{93d}-\x{940}\x{949}-\x{94c}\x{94e}-\x{950}\x{958}-\x{961}\x{964}-\x{980}\x{982}\x{983}\x{985}-\x{98c}\x{98f}\x{990}\x{993}-\x{9a8}\x{9aa}-\x{9b0}\x{9b2}\x{9b6}-\x{9b9}\x{9bd}-\x{9c0}\x{9c7}\x{9c8}\x{9cb}\x{9cc}\x{9ce}\x{9d7}\x{9dc}\x{9dd}\x{9df}-\x{9e1}\x{9e6}-\x{9f1}\x{9f4}-\x{9fa}\x{a03}\x{a05}-\x{a0a}\x{a0f}\x{a10}\x{a13}-\x{a28}\x{a2a}-\x{a30}\x{a32}\x{a33}\x{a35}\x{a36}\x{a38}\x{a39}\x{a3e}-\x{a40}\x{a59}-\x{a5c}\x{a5e}\x{a66}-\x{a6f}\x{a72}-\x{a74}\x{a83}\x{a85}-\x{a8d}\x{a8f}-\x{a91}\x{a93}-\x{aa8}\x{aaa}-\x{ab0}\x{ab2}\x{ab3}\x{ab5}-\x{ab9}\x{abd}-\x{ac0}\x{ac9}\x{acb}\x{acc}\x{ad0}\x{ae0}\x{ae1}\x{ae6}-\x{af0}\x{af9}\x{b02}\x{b03}\x{b05}-\x{b0c}\x{b0f}\x{b10}\x{b13}-\x{b28}\x{b2a}-\x{b30}\x{b32}\x{b33}\x{b35}-\x{b39}\x{b3d}\x{b3e}\x{b40}\x{b47}\x{b48}\x{b4b}\x{b4c}\x{b57}\x{b5c}\x{b5d}\x{b5f}-\x{b61}\x{b66}-\x{b77}\x{b83}\x{b85}-\x{b8a}\x{b8e}-\x{b90}\x{b92}-\x{b95}\x{b99}\x{b9a}\x{b9c}\x{b9e}\x{b9f}\x{ba3}\x{ba4}\x{ba8}-\x{baa}\x{bae}-\x{bb9}\x{bbe}\x{bbf}\x{bc1}\x{bc2}\x{bc6}-\x{bc8}\x{bca}-\x{bcc}\x{bd0}\x{bd7}\x{be6}-\x{bf2}\x{c01}-\x{c03}\x{c05}-\x{c0c}\x{c0e}-\x{c10}\x{c12}-\x{c28}\x{c2a}-\x{c39}\x{c3d}\x{c41}-\x{c44}\x{c58}-\x{c5a}\x{c60}\x{c61}\x{c66}-\x{c6f}\x{c7f}\x{c82}\x{c83}\x{c85}-\x{c8c}\x{c8e}-\x{c90}\x{c92}-\x{ca8}\x{caa}-\x{cb3}\x{cb5}-\x{cb9}\x{cbd}-\x{cc4}\x{cc6}-\x{cc8}\x{cca}\x{ccb}\x{cd5}\x{cd6}\x{cde}\x{ce0}\x{ce1}\x{ce6}-\x{cef}\x{cf1}\x{cf2}\x{d02}\x{d03}\x{d05}-\x{d0c}\x{d0e}-\x{d10}\x{d12}-\x{d3a}\x{d3d}-\x{d40}\x{d46}-\x{d48}\x{d4a}-\x{d4c}\x{d4e}\x{d57}\x{d5f}-\x{d61}\x{d66}-\x{d75}\x{d79}-\x{d7f}\x{d82}\x{d83}\x{d85}-\x{d96}\x{d9a}-\x{db1}\x{db3}-\x{dbb}\x{dbd}\x{dc0}-\x{dc6}\x{dcf}-\x{dd1}\x{dd8}-\x{ddf}\x{de6}-\x{def}\x{df2}-\x{df4}\x{e01}-\x{e30}\x{e32}\x{e33}\x{e40}-\x{e46}\x{e4f}-\x{e5b}\x{e81}\x{e82}\x{e84}\x{e87}\x{e88}\x{e8a}\x{e8d}\x{e94}-\x{e97}\x{e99}-\x{e9f}\x{ea1}-\x{ea3}\x{ea5}\x{ea7}\x{eaa}\x{eab}\x{ead}-\x{eb0}\x{eb2}\x{eb3}\x{ebd}\x{ec0}-\x{ec4}\x{ec6}\x{ed0}-\x{ed9}\x{edc}-\x{edf}\x{f00}-\x{f17}\x{f1a}-\x{f34}\x{f36}\x{f38}\x{f3e}-\x{f47}\x{f49}-\x{f6c}\x{f7f}\x{f85}\x{f88}-\x{f8c}\x{fbe}-\x{fc5}\x{fc7}-\x{fcc}\x{fce}-\x{fda}\x{1000}-\x{102c}\x{1031}\x{1038}\x{103b}\x{103c}\x{103f}-\x{1057}\x{105a}-\x{105d}\x{1061}-\x{1070}\x{1075}-\x{1081}\x{1083}\x{1084}\x{1087}-\x{108c}\x{108e}-\x{109c}\x{109e}-\x{10c5}\x{10c7}\x{10cd}\x{10d0}-\x{1248}\x{124a}-\x{124d}\x{1250}-\x{1256}\x{1258}\x{125a}-\x{125d}\x{1260}-\x{1288}\x{128a}-\x{128d}\x{1290}-\x{12b0}\x{12b2}-\x{12b5}\x{12b8}-\x{12be}\x{12c0}\x{12c2}-\x{12c5}\x{12c8}-\x{12d6}\x{12d8}-\x{1310}\x{1312}-\x{1315}\x{1318}-\x{135a}\x{1360}-\x{137c}\x{1380}-\x{138f}\x{13a0}-\x{13f5}\x{13f8}-\x{13fd}\x{1401}-\x{167f}\x{1681}-\x{169a}\x{16a0}-\x{16f8}\x{1700}-\x{170c}\x{170e}-\x{1711}\x{1720}-\x{1731}\x{1735}\x{1736}\x{1740}-\x{1751}\x{1760}-\x{176c}\x{176e}-\x{1770}\x{1780}-\x{17b3}\x{17b6}\x{17be}-\x{17c5}\x{17c7}\x{17c8}\x{17d4}-\x{17da}\x{17dc}\x{17e0}-\x{17e9}\x{1810}-\x{1819}\x{1820}-\x{1877}\x{1880}-\x{18a8}\x{18aa}\x{18b0}-\x{18f5}\x{1900}-\x{191e}\x{1923}-\x{1926}\x{1929}-\x{192b}\x{1930}\x{1931}\x{1933}-\x{1938}\x{1946}-\x{196d}\x{1970}-\x{1974}\x{1980}-\x{19ab}\x{19b0}-\x{19c9}\x{19d0}-\x{19da}\x{1a00}-\x{1a16}\x{1a19}\x{1a1a}\x{1a1e}-\x{1a55}\x{1a57}\x{1a61}\x{1a63}\x{1a64}\x{1a6d}-\x{1a72}\x{1a80}-\x{1a89}\x{1a90}-\x{1a99}\x{1aa0}-\x{1aad}\x{1b04}-\x{1b33}\x{1b35}\x{1b3b}\x{1b3d}-\x{1b41}\x{1b43}-\x{1b4b}\x{1b50}-\x{1b6a}\x{1b74}-\x{1b7c}\x{1b82}-\x{1ba1}\x{1ba6}\x{1ba7}\x{1baa}\x{1bae}-\x{1be5}\x{1be7}\x{1bea}-\x{1bec}\x{1bee}\x{1bf2}\x{1bf3}\x{1bfc}-\x{1c2b}\x{1c34}\x{1c35}\x{1c3b}-\x{1c49}\x{1c4d}-\x{1c7f}\x{1cc0}-\x{1cc7}\x{1cd3}\x{1ce1}\x{1ce9}-\x{1cec}\x{1cee}-\x{1cf3}\x{1cf5}\x{1cf6}\x{1d00}-\x{1dbf}\x{1e00}-\x{1f15}\x{1f18}-\x{1f1d}\x{1f20}-\x{1f45}\x{1f48}-\x{1f4d}\x{1f50}-\x{1f57}\x{1f59}\x{1f5b}\x{1f5d}\x{1f5f}-\x{1f7d}\x{1f80}-\x{1fb4}\x{1fb6}-\x{1fbc}\x{1fbe}\x{1fc2}-\x{1fc4}\x{1fc6}-\x{1fcc}\x{1fd0}-\x{1fd3}\x{1fd6}-\x{1fdb}\x{1fe0}-\x{1fec}\x{1ff2}-\x{1ff4}\x{1ff6}-\x{1ffc}\x{200e}\x{2071}\x{207f}\x{2090}-\x{209c}\x{2102}\x{2107}\x{210a}-\x{2113}\x{2115}\x{2119}-\x{211d}\x{2124}\x{2126}\x{2128}\x{212a}-\x{212d}\x{212f}-\x{2139}\x{213c}-\x{213f}\x{2145}-\x{2149}\x{214e}\x{214f}\x{2160}-\x{2188}\x{2336}-\x{237a}\x{2395}\x{249c}-\x{24e9}\x{26ac}\x{2800}-\x{28ff}\x{2c00}-\x{2c2e}\x{2c30}-\x{2c5e}\x{2c60}-\x{2ce4}\x{2ceb}-\x{2cee}\x{2cf2}\x{2cf3}\x{2d00}-\x{2d25}\x{2d27}\x{2d2d}\x{2d30}-\x{2d67}\x{2d6f}\x{2d70}\x{2d80}-\x{2d96}\x{2da0}-\x{2da6}\x{2da8}-\x{2dae}\x{2db0}-\x{2db6}\x{2db8}-\x{2dbe}\x{2dc0}-\x{2dc6}\x{2dc8}-\x{2dce}\x{2dd0}-\x{2dd6}\x{2dd8}-\x{2dde}\x{3005}-\x{3007}\x{3021}-\x{3029}\x{302e}\x{302f}\x{3031}-\x{3035}\x{3038}-\x{303c}\x{3041}-\x{3096}\x{309d}-\x{309f}\x{30a1}-\x{30fa}\x{30fc}-\x{30ff}\x{3105}-\x{312d}\x{3131}-\x{318e}\x{3190}-\x{31ba}\x{31f0}-\x{321c}\x{3220}-\x{324f}\x{3260}-\x{327b}\x{327f}-\x{32b0}\x{32c0}-\x{32cb}\x{32d0}-\x{32fe}\x{3300}-\x{3376}\x{337b}-\x{33dd}\x{33e0}-\x{33fe}\x{3400}-\x{4db5}\x{4e00}-\x{9fd5}\x{a000}-\x{a48c}\x{a4d0}-\x{a60c}\x{a610}-\x{a62b}\x{a640}-\x{a66e}\x{a680}-\x{a69d}\x{a6a0}-\x{a6ef}\x{a6f2}-\x{a6f7}\x{a722}-\x{a787}\x{a789}-\x{a7ad}\x{a7b0}-\x{a7b7}\x{a7f7}-\x{a801}\x{a803}-\x{a805}\x{a807}-\x{a80a}\x{a80c}-\x{a824}\x{a827}\x{a830}-\x{a837}\x{a840}-\x{a873}\x{a880}-\x{a8c3}\x{a8ce}-\x{a8d9}\x{a8f2}-\x{a8fd}\x{a900}-\x{a925}\x{a92e}-\x{a946}\x{a952}\x{a953}\x{a95f}-\x{a97c}\x{a983}-\x{a9b2}\x{a9b4}\x{a9b5}\x{a9ba}\x{a9bb}\x{a9bd}-\x{a9cd}\x{a9cf}-\x{a9d9}\x{a9de}-\x{a9e4}\x{a9e6}-\x{a9fe}\x{aa00}-\x{aa28}\x{aa2f}\x{aa30}\x{aa33}\x{aa34}\x{aa40}-\x{aa42}\x{aa44}-\x{aa4b}\x{aa4d}\x{aa50}-\x{aa59}\x{aa5c}-\x{aa7b}\x{aa7d}-\x{aaaf}\x{aab1}\x{aab5}\x{aab6}\x{aab9}-\x{aabd}\x{aac0}\x{aac2}\x{aadb}-\x{aaeb}\x{aaee}-\x{aaf5}\x{ab01}-\x{ab06}\x{ab09}-\x{ab0e}\x{ab11}-\x{ab16}\x{ab20}-\x{ab26}\x{ab28}-\x{ab2e}\x{ab30}-\x{ab65}\x{ab70}-\x{abe4}\x{abe6}\x{abe7}\x{abe9}-\x{abec}\x{abf0}-\x{abf9}\x{ac00}-\x{d7a3}\x{d7b0}-\x{d7c6}\x{d7cb}-\x{d7fb}\x{e000}-\x{fa6d}\x{fa70}-\x{fad9}\x{fb00}-\x{fb06}\x{fb13}-\x{fb17}\x{ff21}-\x{ff3a}\x{ff41}-\x{ff5a}\x{ff66}-\x{ffbe}\x{ffc2}-\x{ffc7}\x{ffca}-\x{ffcf}\x{ffd2}-\x{ffd7}\x{ffda}-\x{ffdc}\x{10000}-\x{1000b}\x{1000d}-\x{10026}\x{10028}-\x{1003a}\x{1003c}\x{1003d}\x{1003f}-\x{1004d}\x{10050}-\x{1005d}\x{10080}-\x{100fa}\x{10100}\x{10102}\x{10107}-\x{10133}\x{10137}-\x{1013f}\x{101d0}-\x{101fc}\x{10280}-\x{1029c}\x{102a0}-\x{102d0}\x{10300}-\x{10323}\x{10330}-\x{1034a}\x{10350}-\x{10375}\x{10380}-\x{1039d}\x{1039f}-\x{103c3}\x{103c8}-\x{103d5}\x{10400}-\x{1049d}\x{104a0}-\x{104a9}\x{10500}-\x{10527}\x{10530}-\x{10563}\x{1056f}\x{10600}-\x{10736}\x{10740}-\x{10755}\x{10760}-\x{10767}\x{11000}\x{11002}-\x{11037}\x{11047}-\x{1104d}\x{11066}-\x{1106f}\x{11082}-\x{110b2}\x{110b7}\x{110b8}\x{110bb}-\x{110c1}\x{110d0}-\x{110e8}\x{110f0}-\x{110f9}\x{11103}-\x{11126}\x{1112c}\x{11136}-\x{11143}\x{11150}-\x{11172}\x{11174}-\x{11176}\x{11182}-\x{111b5}\x{111bf}-\x{111c9}\x{111cd}\x{111d0}-\x{111df}\x{111e1}-\x{111f4}\x{11200}-\x{11211}\x{11213}-\x{1122e}\x{11232}\x{11233}\x{11235}\x{11238}-\x{1123d}\x{11280}-\x{11286}\x{11288}\x{1128a}-\x{1128d}\x{1128f}-\x{1129d}\x{1129f}-\x{112a9}\x{112b0}-\x{112de}\x{112e0}-\x{112e2}\x{112f0}-\x{112f9}\x{11302}\x{11303}\x{11305}-\x{1130c}\x{1130f}\x{11310}\x{11313}-\x{11328}\x{1132a}-\x{11330}\x{11332}\x{11333}\x{11335}-\x{11339}\x{1133d}-\x{1133f}\x{11341}-\x{11344}\x{11347}\x{11348}\x{1134b}-\x{1134d}\x{11350}\x{11357}\x{1135d}-\x{11363}\x{11480}-\x{114b2}\x{114b9}\x{114bb}-\x{114be}\x{114c1}\x{114c4}-\x{114c7}\x{114d0}-\x{114d9}\x{11580}-\x{115b1}\x{115b8}-\x{115bb}\x{115be}\x{115c1}-\x{115db}\x{11600}-\x{11632}\x{1163b}\x{1163c}\x{1163e}\x{11641}-\x{11644}\x{11650}-\x{11659}\x{11680}-\x{116aa}\x{116ac}\x{116ae}\x{116af}\x{116b6}\x{116c0}-\x{116c9}\x{11700}-\x{11719}\x{11720}\x{11721}\x{11726}\x{11730}-\x{1173f}\x{118a0}-\x{118f2}\x{118ff}\x{11ac0}-\x{11af8}\x{12000}-\x{12399}\x{12400}-\x{1246e}\x{12470}-\x{12474}\x{12480}-\x{12543}\x{13000}-\x{1342e}\x{14400}-\x{14646}\x{16800}-\x{16a38}\x{16a40}-\x{16a5e}\x{16a60}-\x{16a69}\x{16a6e}\x{16a6f}\x{16ad0}-\x{16aed}\x{16af5}\x{16b00}-\x{16b2f}\x{16b37}-\x{16b45}\x{16b50}-\x{16b59}\x{16b5b}-\x{16b61}\x{16b63}-\x{16b77}\x{16b7d}-\x{16b8f}\x{16f00}-\x{16f44}\x{16f50}-\x{16f7e}\x{16f93}-\x{16f9f}\x{1b000}\x{1b001}\x{1bc00}-\x{1bc6a}\x{1bc70}-\x{1bc7c}\x{1bc80}-\x{1bc88}\x{1bc90}-\x{1bc99}\x{1bc9c}\x{1bc9f}\x{1d000}-\x{1d0f5}\x{1d100}-\x{1d126}\x{1d129}-\x{1d166}\x{1d16a}-\x{1d172}\x{1d183}\x{1d184}\x{1d18c}-\x{1d1a9}\x{1d1ae}-\x{1d1e8}\x{1d360}-\x{1d371}\x{1d400}-\x{1d454}\x{1d456}-\x{1d49c}\x{1d49e}\x{1d49f}\x{1d4a2}\x{1d4a5}\x{1d4a6}\x{1d4a9}-\x{1d4ac}\x{1d4ae}-\x{1d4b9}\x{1d4bb}\x{1d4bd}-\x{1d4c3}\x{1d4c5}-\x{1d505}\x{1d507}-\x{1d50a}\x{1d50d}-\x{1d514}\x{1d516}-\x{1d51c}\x{1d51e}-\x{1d539}\x{1d53b}-\x{1d53e}\x{1d540}-\x{1d544}\x{1d546}\x{1d54a}-\x{1d550}\x{1d552}-\x{1d6a5}\x{1d6a8}-\x{1d6da}\x{1d6dc}-\x{1d714}\x{1d716}-\x{1d74e}\x{1d750}-\x{1d788}\x{1d78a}-\x{1d7c2}\x{1d7c4}-\x{1d7cb}\x{1d800}-\x{1d9ff}\x{1da37}-\x{1da3a}\x{1da6d}-\x{1da74}\x{1da76}-\x{1da83}\x{1da85}-\x{1da8b}\x{1f110}-\x{1f12e}\x{1f130}-\x{1f169}\x{1f170}-\x{1f19a}\x{1f1e6}-\x{1f202}\x{1f210}-\x{1f23a}\x{1f240}-\x{1f248}\x{1f250}\x{1f251}\x{20000}-\x{2a6d6}\x{2a700}-\x{2b734}\x{2b740}-\x{2b81d}\x{2b820}-\x{2cea1}\x{2f800}-\x{2fa1d}\x{f0000}-\x{ffffd}\x{100000}-\x{10fffd}])|([\x{590}\x{5be}\x{5c0}\x{5c3}\x{5c6}\x{5c8}-\x{5ff}\x{7c0}-\x{7ea}\x{7f4}\x{7f5}\x{7fa}-\x{815}\x{81a}\x{824}\x{828}\x{82e}-\x{858}\x{85c}-\x{89f}\x{200f}\x{fb1d}\x{fb1f}-\x{fb28}\x{fb2a}-\x{fb4f}\x{10800}-\x{1091e}\x{10920}-\x{10a00}\x{10a04}\x{10a07}-\x{10a0b}\x{10a10}-\x{10a37}\x{10a3b}-\x{10a3e}\x{10a40}-\x{10ae4}\x{10ae7}-\x{10b38}\x{10b40}-\x{10e5f}\x{10e7f}-\x{10fff}\x{1e800}-\x{1e8cf}\x{1e8d7}-\x{1edff}\x{1ef00}-\x{1efff}\x{608}\x{60b}\x{60d}\x{61b}-\x{64a}\x{66d}-\x{66f}\x{671}-\x{6d5}\x{6e5}\x{6e6}\x{6ee}\x{6ef}\x{6fa}-\x{710}\x{712}-\x{72f}\x{74b}-\x{7a5}\x{7b1}-\x{7bf}\x{8a0}-\x{8e2}\x{fb50}-\x{fd3d}\x{fd40}-\x{fdcf}\x{fdf0}-\x{fdfc}\x{fdfe}\x{fdff}\x{fe70}-\x{fefe}\x{1ee00}-\x{1eeef}\x{1eef2}-\x{1eeff}]))/u';
 208         // @codeCoverageIgnoreEnd
 209
 210         /**
 211          * Get a cached or new language object for a given language code
 212          * @param string $code
 213          * @throws MWException
 214          * @return Language
 215          */
 216         static function factory( $code ) {
 217                 global $wgDummyLanguageCodes, $wgLangObjCacheSize;
 218
 219                 if ( isset( $wgDummyLanguageCodes[$code] ) ) {
 220                         $code = $wgDummyLanguageCodes[$code];
 221                 }
 222
 223                 // get the language object to process
 224                 $langObj = self::$mLangObjCache[$code] ?? self::newFromCode( $code );
 225
 226                 // merge the language object in to get it up front in the cache
 227                 self::$mLangObjCache = array_merge( [ $code => $langObj ], self::$mLangObjCache );
 228                 // get rid of the oldest ones in case we have an overflow
 229                 self::$mLangObjCache = array_slice( self::$mLangObjCache, 0, $wgLangObjCacheSize, true );
 230
 231                 return $langObj;
 232         }
 233
 234         /**
 235          * Create a language object for a given language code
 236          * @param string $code
 237          * @param bool $fallback Whether we're going through language fallback chain
 238          * @throws MWException
 239          * @return Language
 240          */
 241         protected static function newFromCode( $code, $fallback = false ) {
 242                 if ( !self::isValidCode( $code ) ) {
 243                         throw new MWException( "Invalid language code \"$code\"" );
 244                 }
 245
 246                 if ( !self::isValidBuiltInCode( $code ) ) {
 247                         // It's not possible to customise this code with class files, so
 248                         // just return a Language object. This is to support uselang= hacks.
 249                         $lang = new Language;
 250                         $lang->mCode = $code;
 251                         return $lang;
 252                 }
 253
 254                 // Check if there is a language class for the code
 255                 $class = self::classFromCode( $code, $fallback );
 256                 // LanguageCode does not inherit Language
 257                 if ( class_exists( $class ) && is_a( $class, 'Language', true ) ) {
 258                         $lang = new $class;
 259                         return $lang;
 260                 }
 261
 262                 // Keep trying the fallback list until we find an existing class
 263                 $fallbacks = self::getFallbacksFor( $code );
 264                 foreach ( $fallbacks as $fallbackCode ) {
 265                         if ( !self::isValidBuiltInCode( $fallbackCode ) ) {
 266                                 throw new MWException( "Invalid fallback '$fallbackCode' in fallback sequence for '$code'" );
 267                         }
 268
 269                         $class = self::classFromCode( $fallbackCode );
 270                         if ( class_exists( $class ) ) {
 271                                 $lang = new $class;
 272                                 $lang->mCode = $code;
 273                                 return $lang;
 274                         }
 275                 }
 276
 277                 throw new MWException( "Invalid fallback sequence for language '$code'" );
 278         }
 279
 280         /**
 281          * Intended for tests that may change configuration in a way that invalidates caches.
 282          *
 283          * @since 1.32
 284          */
 285         public static function clearCaches() {
 286                 if ( !defined( 'MW_PHPUNIT_TEST' ) ) {
 287                         throw new MWException( __METHOD__ . ' must not be used outside tests' );
 288                 }
 289                 self::$dataCache = null;
 290                 // Reinitialize $dataCache, since it's expected to always be available
 291                 self::getLocalisationCache();
 292                 self::$mLangObjCache = [];
 293                 self::$fallbackLanguageCache = [];
 294                 self::$grammarTransformations = null;
 295                 self::$languageNameCache = null;
 296         }
 297
 298         /**
 299          * Checks whether any localisation is available for that language tag
 300          * in MediaWiki (MessagesXx.php exists).
 301          *
 302          * @param string $code Language tag (in lower case)
 303          * @return bool Whether language is supported
 304          * @since 1.21
 305          */
 306         public static function isSupportedLanguage( $code ) {
 307                 if ( !self::isValidBuiltInCode( $code ) ) {
 308                         return false;
 309                 }
 310
 311                 if ( $code === 'qqq' ) {
 312                         return false;
 313                 }
 314
 315                 return is_readable( self::getMessagesFileName( $code ) ) ||
 316                         is_readable( self::getJsonMessagesFileName( $code ) );
 317         }
 318
 319         /**
 320          * Returns true if a language code string is a well-formed language tag
 321          * according to RFC 5646.
 322          * This function only checks well-formedness; it doesn't check that
 323          * language, script or variant codes actually exist in the repositories.
 324          *
 325          * Based on regexes by Mark Davis of the Unicode Consortium:
 326          * https://www.unicode.org/repos/cldr/trunk/tools/java/org/unicode/cldr/util/data/langtagRegex.txt
 327          *
 328          * @param string $code
 329          * @param bool $lenient Whether to allow '_' as separator. The default is only '-'.
 330          *
 331          * @return bool
 332          * @since 1.21
 333          */
 334         public static function isWellFormedLanguageTag( $code, $lenient = false ) {
 335                 $alpha = '[a-z]';
 336                 $digit = '[0-9]';
 337                 $alphanum = '[a-z0-9]';
 338                 $x = 'x'; # private use singleton
 339                 $singleton = '[a-wy-z]'; # other singleton
 340                 $s = $lenient ? '[-_]' : '-';
 341
 342                 $language = "$alpha{2,8}|$alpha{2,3}$s$alpha{3}";
 343                 $script = "$alpha{4}"; # ISO 15924
 344                 $region = "(?:$alpha{2}|$digit{3})"; # ISO 3166-1 alpha-2 or UN M.49
 345                 $variant = "(?:$alphanum{5,8}|$digit$alphanum{3})";
 346                 $extension = "$singleton(?:$s$alphanum{2,8})+";
 347                 $privateUse = "$x(?:$s$alphanum{1,8})+";
 348
 349                 # Define certain grandfathered codes, since otherwise the regex is pretty useless.
 350                 # Since these are limited, this is safe even later changes to the registry --
 351                 # the only oddity is that it might change the type of the tag, and thus
 352                 # the results from the capturing groups.
 353                 # https://www.iana.org/assignments/language-subtag-registry
 354
 355                 $grandfathered = "en{$s}GB{$s}oed"
 356                         . "|i{$s}(?:ami|bnn|default|enochian|hak|klingon|lux|mingo|navajo|pwn|tao|tay|tsu)"
 357                         . "|no{$s}(?:bok|nyn)"
 358                         . "|sgn{$s}(?:BE{$s}(?:fr|nl)|CH{$s}de)"
 359                         . "|zh{$s}min{$s}nan";
 360
 361                 $variantList = "$variant(?:$s$variant)*";
 362                 $extensionList = "$extension(?:$s$extension)*";
 363
 364                 $langtag = "(?:($language)"
 365                         . "(?:$s$script)?"
 366                         . "(?:$s$region)?"
 367                         . "(?:$s$variantList)?"
 368                         . "(?:$s$extensionList)?"
 369                         . "(?:$s$privateUse)?)";
 370
 371                 # The final breakdown, with capturing groups for each of these components
 372                 # The variants, extensions, grandfathered, and private-use may have interior '-'
 373
 374                 $root = "^(?:$langtag|$privateUse|$grandfathered)$";
 375
 376                 return (bool)preg_match( "/$root/", strtolower( $code ) );
 377         }
 378
 379         /**
 380          * Returns true if a language code string is of a valid form, whether or
 381          * not it exists. This includes codes which are used solely for
 382          * customisation via the MediaWiki namespace.
 383          *
 384          * @param string $code
 385          *
 386          * @return bool
 387          */
 388         public static function isValidCode( $code ) {
 389                 static $cache = [];
 390                 Assert::parameterType( 'string', $code, '$code' );
 391                 if ( !isset( $cache[$code] ) ) {
 392                         // People think language codes are html safe, so enforce it.
 393                         // Ideally we should only allow a-zA-Z0-9-
 394                         // but, .+ and other chars are often used for {{int:}} hacks
 395                         // see bugs T39564, T39587, T38938
 396                         $cache[$code] =
 397                                 // Protect against path traversal
 398                                 strcspn( $code, ":/\\\000&<>'\"" ) === strlen( $code )
 399                                 && !preg_match( MediaWikiTitleCodec::getTitleInvalidRegex(), $code );
 400                 }
 401                 return $cache[$code];
 402         }
 403
 404         /**
 405          * Returns true if a language code is of a valid form for the purposes of
 406          * internal customisation of MediaWiki, via Messages*.php or *.json.
 407          *
 408          * @param string $code
 409          *
 410          * @since 1.18
 411          * @return bool
 412          */
 413         public static function isValidBuiltInCode( $code ) {
 414                 Assert::parameterType( 'string', $code, '$code' );
 415
 416                 return (bool)preg_match( '/^[a-z0-9-]{2,}$/', $code );
 417         }
 418
 419         /**
 420          * Returns true if a language code is an IETF tag known to MediaWiki.
 421          *
 422          * @param string $tag
 423          *
 424          * @since 1.21
 425          * @return bool
 426          */
 427         public static function isKnownLanguageTag( $tag ) {
 428                 // Quick escape for invalid input to avoid exceptions down the line
 429                 // when code tries to process tags which are not valid at all.
 430                 if ( !self::isValidBuiltInCode( $tag ) ) {
 431                         return false;
 432                 }
 433
 434                 if ( isset( MediaWiki\Languages\Data\Names::$names[$tag] )
 435                         || self::fetchLanguageName( $tag, $tag ) !== ''
 436                 ) {
 437                         return true;
 438                 }
 439
 440                 return false;
 441         }
 442
 443         /**
 444          * Get the LocalisationCache instance
 445          *
 446          * @return LocalisationCache
 447          */
 448         public static function getLocalisationCache() {
 449                 if ( is_null( self::$dataCache ) ) {
 450                         global $wgLocalisationCacheConf;
 451                         $class = $wgLocalisationCacheConf['class'];
 452                         self::$dataCache = new $class( $wgLocalisationCacheConf );
 453                 }
 454                 return self::$dataCache;
 455         }
 456
 457         function __construct() {
 458                 $this->mConverter = new FakeConverter( $this );
 459                 // Set the code to the name of the descendant
 460                 if ( static::class === 'Language' ) {
 461                         $this->mCode = 'en';
 462                 } else {
 463                         $this->mCode = str_replace( '_', '-', strtolower( substr( static::class, 8 ) ) );
 464                 }
 465                 self::getLocalisationCache();
 466         }
 467
 468         /**
 469          * Reduce memory usage
 470          */
 471         function __destruct() {
 472                 foreach ( $this as $name => $value ) {
 473                         unset( $this->$name );
 474                 }
 475         }
 476
 477         /**
 478          * Hook which will be called if this is the content language.
 479          * Descendants can use this to register hook functions or modify globals
 480          */
 481         function initContLang() {
 482         }
 483
 484         /**
 485          * @return array
 486          * @since 1.19
 487          */
 488         public function getFallbackLanguages() {
 489                 return self::getFallbacksFor( $this->mCode );
 490         }
 491
 492         /**
 493          * Exports $wgBookstoreListEn
 494          * @return array
 495          */
 496         public function getBookstoreList() {
 497                 return self::$dataCache->getItem( $this->mCode, 'bookstoreList' );
 498         }
 499
 500         /**
 501          * Returns an array of localised namespaces indexed by their numbers. If the namespace is not
 502          * available in localised form, it will be included in English.
 503          *
 504          * @return string[] List of localized namespace names, indexed by numeric namespace ID.
 505          */
 506         public function getNamespaces() {
 507                 if ( is_null( $this->namespaceNames ) ) {
 508                         global $wgMetaNamespace, $wgMetaNamespaceTalk, $wgExtraNamespaces;
 509
 510                         $validNamespaces = MediaWikiServices::getInstance()->getNamespaceInfo()->
 511                                 getCanonicalNamespaces();
 512
 513                         $this->namespaceNames = $wgExtraNamespaces +
 514                                 self::$dataCache->getItem( $this->mCode, 'namespaceNames' );
 515                         $this->namespaceNames += $validNamespaces;
 516
 517                         $this->namespaceNames[NS_PROJECT] = $wgMetaNamespace;
 518                         if ( $wgMetaNamespaceTalk ) {
 519                                 $this->namespaceNames[NS_PROJECT_TALK] = $wgMetaNamespaceTalk;
 520                         } else {
 521                                 $talk = $this->namespaceNames[NS_PROJECT_TALK];
 522                                 $this->namespaceNames[NS_PROJECT_TALK] =
 523                                         $this->fixVariableInNamespace( $talk );
 524                         }
 525
 526                         # Sometimes a language will be localised but not actually exist on this wiki.
 527                         foreach ( $this->namespaceNames as $key => $text ) {
 528                                 if ( !isset( $validNamespaces[$key] ) ) {
 529                                         unset( $this->namespaceNames[$key] );
 530                                 }
 531                         }
 532
 533                         # The above mixing may leave namespaces out of canonical order.
 534                         # Re-order by namespace ID number...
 535                         ksort( $this->namespaceNames );
 536
 537                         Hooks::run( 'LanguageGetNamespaces', [ &$this->namespaceNames ] );
 538                 }
 539
 540                 return $this->namespaceNames;
 541         }
 542
 543         /**
 544          * Arbitrarily set all of the namespace names at once. Mainly used for testing
 545          * @param array $namespaces Array of namespaces (id => name)
 546          */
 547         public function setNamespaces( array $namespaces ) {
 548                 $this->namespaceNames = $namespaces;
 549                 $this->mNamespaceIds = null;
 550         }
 551
 552         /**
 553          * Resets all of the namespace caches. Mainly used for testing
 554          */
 555         public function resetNamespaces() {
 556                 $this->namespaceNames = null;
 557                 $this->mNamespaceIds = null;
 558                 $this->namespaceAliases = null;
 559         }
 560
 561         /**
 562          * A convenience function that returns getNamespaces() with spaces instead of underscores
 563          * in values. Useful for producing output to be displayed e.g. in `<select>` forms.
 564          *
 565          * @return array
 566          */
 567         public function getFormattedNamespaces() {
 568                 $ns = $this->getNamespaces();
 569                 foreach ( $ns as $k => $v ) {
 570                         $ns[$k] = strtr( $v, '_', ' ' );
 571                 }
 572                 return $ns;
 573         }
 574
 575         /**
 576          * Get a namespace value by key
 577          *
 578          * <code>
 579          * $mw_ns = $lang->getNsText( NS_MEDIAWIKI );
 580          * echo $mw_ns; // prints 'MediaWiki'
 581          * </code>
 582          *
 583          * @param int $index The array key of the namespace to return
 584          * @return string|bool String if the namespace value exists, otherwise false
 585          */
 586         public function getNsText( $index ) {
 587                 $ns = $this->getNamespaces();
 588                 return $ns[$index] ?? false;
 589         }
 590
 591         /**
 592          * A convenience function that returns the same thing as
 593          * getNsText() except with '_' changed to ' ', useful for
 594          * producing output.
 595          *
 596          * <code>
 597          * $mw_ns = $lang->getFormattedNsText( NS_MEDIAWIKI_TALK );
 598          * echo $mw_ns; // prints 'MediaWiki talk'
 599          * </code>
 600          *
 601          * @param int $index The array key of the namespace to return
 602          * @return string Namespace name without underscores (empty string if namespace does not exist)
 603          */
 604         public function getFormattedNsText( $index ) {
 605                 $ns = $this->getNsText( $index );
 606                 return strtr( $ns, '_', ' ' );
 607         }
 608
 609         /**
 610          * Returns gender-dependent namespace alias if available.
 611          * See https://www.mediawiki.org/wiki/Manual:$wgExtraGenderNamespaces
 612          * @param int $index Namespace index
 613          * @param string $gender Gender key (male, female... )
 614          * @return string
 615          * @since 1.18
 616          */
 617         public function getGenderNsText( $index, $gender ) {
 618                 global $wgExtraGenderNamespaces;
 619
 620                 $ns = $wgExtraGenderNamespaces +
 621                         (array)self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 622
 623                 return $ns[$index][$gender] ?? $this->getNsText( $index );
 624         }
 625
 626         /**
 627          * Whether this language uses gender-dependent namespace aliases.
 628          * See https://www.mediawiki.org/wiki/Manual:$wgExtraGenderNamespaces
 629          * @return bool
 630          * @since 1.18
 631          */
 632         public function needsGenderDistinction() {
 633                 global $wgExtraGenderNamespaces, $wgExtraNamespaces;
 634                 if ( count( $wgExtraGenderNamespaces ) > 0 ) {
 635                         // $wgExtraGenderNamespaces overrides everything
 636                         return true;
 637                 } elseif ( isset( $wgExtraNamespaces[NS_USER] ) && isset( $wgExtraNamespaces[NS_USER_TALK] ) ) {
 638                         /// @todo There may be other gender namespace than NS_USER & NS_USER_TALK in the future
 639                         // $wgExtraNamespaces overrides any gender aliases specified in i18n files
 640                         return false;
 641                 } else {
 642                         // Check what is in i18n files
 643                         $aliases = self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 644                         return count( $aliases ) > 0;
 645                 }
 646         }
 647
 648         /**
 649          * Get a namespace key by value, case insensitive.
 650          * Only matches namespace names for the current language, not the
 651          * canonical ones defined in Namespace.php.
 652          *
 653          * @param string $text
 654          * @return int|bool An integer if $text is a valid value otherwise false
 655          */
 656         function getLocalNsIndex( $text ) {
 657                 $lctext = $this->lc( $text );
 658                 $ids = $this->getNamespaceIds();
 659                 return $ids[$lctext] ?? false;
 660         }
 661
 662         /**
 663          * @return array
 664          */
 665         public function getNamespaceAliases() {
 666                 if ( is_null( $this->namespaceAliases ) ) {
 667                         $aliases = self::$dataCache->getItem( $this->mCode, 'namespaceAliases' );
 668                         if ( !$aliases ) {
 669                                 $aliases = [];
 670                         } else {
 671                                 foreach ( $aliases as $name => $index ) {
 672                                         if ( $index === NS_PROJECT_TALK ) {
 673                                                 unset( $aliases[$name] );
 674                                                 $name = $this->fixVariableInNamespace( $name );
 675                                                 $aliases[$name] = $index;
 676                                         }
 677                                 }
 678                         }
 679
 680                         global $wgExtraGenderNamespaces;
 681                         $genders = $wgExtraGenderNamespaces +
 682                                 (array)self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 683                         foreach ( $genders as $index => $forms ) {
 684                                 foreach ( $forms as $alias ) {
 685                                         $aliases[$alias] = $index;
 686                                 }
 687                         }
 688
 689                         # Also add converted namespace names as aliases, to avoid confusion.
 690                         $convertedNames = [];
 691                         foreach ( $this->getVariants() as $variant ) {
 692                                 if ( $variant === $this->mCode ) {
 693                                         continue;
 694                                 }
 695                                 foreach ( $this->getNamespaces() as $ns => $_ ) {
 696                                         $convertedNames[$this->getConverter()->convertNamespace( $ns, $variant )] = $ns;
 697                                 }
 698                         }
 699
 700                         $this->namespaceAliases = $aliases + $convertedNames;
 701
 702                         # Filter out aliases to namespaces that don't exist, e.g. from extensions
 703                         # that aren't loaded here but are included in the l10n cache.
 704                         # (array_intersect preserves keys from its first argument)
 705                         $this->namespaceAliases = array_intersect(
 706                                 $this->namespaceAliases,
 707                                 array_keys( $this->getNamespaces() )
 708                         );
 709                 }
 710
 711                 return $this->namespaceAliases;
 712         }
 713
 714         /**
 715          * @return array
 716          */
 717         public function getNamespaceIds() {
 718                 if ( is_null( $this->mNamespaceIds ) ) {
 719                         global $wgNamespaceAliases;
 720                         # Put namespace names and aliases into a hashtable.
 721                         # If this is too slow, then we should arrange it so that it is done
 722                         # before caching. The catch is that at pre-cache time, the above
 723                         # class-specific fixup hasn't been done.
 724                         $this->mNamespaceIds = [];
 725                         foreach ( $this->getNamespaces() as $index => $name ) {
 726                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 727                         }
 728                         foreach ( $this->getNamespaceAliases() as $name => $index ) {
 729                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 730                         }
 731                         if ( $wgNamespaceAliases ) {
 732                                 foreach ( $wgNamespaceAliases as $name => $index ) {
 733                                         $this->mNamespaceIds[$this->lc( $name )] = $index;
 734                                 }
 735                         }
 736                 }
 737                 return $this->mNamespaceIds;
 738         }
 739
 740         /**
 741          * Get a namespace key by value, case insensitive.  Canonical namespace
 742          * names override custom ones defined for the current language.
 743          *
 744          * @param string $text
 745          * @return int|bool An integer if $text is a valid value otherwise false
 746          */
 747         public function getNsIndex( $text ) {
 748                 $lctext = $this->lc( $text );
 749                 $ns = MediaWikiServices::getInstance()->getNamespaceInfo()->
 750                         getCanonicalIndex( $lctext );
 751                 if ( $ns !== null ) {
 752                         return $ns;
 753                 }
 754                 $ids = $this->getNamespaceIds();
 755                 return $ids[$lctext] ?? false;
 756         }
 757
 758         /**
 759          * short names for language variants used for language conversion links.
 760          *
 761          * @param string $code
 762          * @param bool $usemsg Use the "variantname-xyz" message if it exists
 763          * @return string
 764          */
 765         public function getVariantname( $code, $usemsg = true ) {
 766                 $msg = "variantname-$code";
 767                 if ( $usemsg && wfMessage( $msg )->exists() ) {
 768                         return $this->getMessageFromDB( $msg );
 769                 }
 770                 $name = self::fetchLanguageName( $code );
 771                 if ( $name ) {
 772                         return $name; # if it's defined as a language name, show that
 773                 } else {
 774                         # otherwise, output the language code
 775                         return $code;
 776                 }
 777         }
 778
 779         /**
 780          * @return string[]|bool List of date format preference keys, or false if disabled.
 781          */
 782         public function getDatePreferences() {
 783                 return self::$dataCache->getItem( $this->mCode, 'datePreferences' );
 784         }
 785
 786         /**
 787          * @return array
 788          */
 789         function getDateFormats() {
 790                 return self::$dataCache->getItem( $this->mCode, 'dateFormats' );
 791         }
 792
 793         /**
 794          * @return array|string
 795          */
 796         public function getDefaultDateFormat() {
 797                 $df = self::$dataCache->getItem( $this->mCode, 'defaultDateFormat' );
 798                 if ( $df === 'dmy or mdy' ) {
 799                         global $wgAmericanDates;
 800                         return $wgAmericanDates ? 'mdy' : 'dmy';
 801                 } else {
 802                         return $df;
 803                 }
 804         }
 805
 806         /**
 807          * @return array
 808          */
 809         public function getDatePreferenceMigrationMap() {
 810                 return self::$dataCache->getItem( $this->mCode, 'datePreferenceMigrationMap' );
 811         }
 812
 813         /**
 814          * @return array
 815          */
 816         public function getExtraUserToggles() {
 817                 wfDeprecated( __METHOD__, '1.34' );
 818                 return [];
 819         }
 820
 821         /**
 822          * @param string $tog
 823          * @return string
 824          */
 825         function getUserToggle( $tog ) {
 826                 return $this->getMessageFromDB( "tog-$tog" );
 827         }
 828
 829         /**
 830          * Get an array of language names, indexed by code.
 831          * @param null|string $inLanguage Code of language in which to return the names
 832          *              Use self::AS_AUTONYMS for autonyms (native names)
 833          * @param string $include One of:
 834          *              self::ALL all available languages
 835          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 836          *              self::SUPPORTED only if the language is in 'mw' *and* has a message file
 837          * @return array Language code => language name (sorted by key)
 838          * @since 1.20
 839          */
 840         public static function fetchLanguageNames( $inLanguage = self::AS_AUTONYMS, $include = 'mw' ) {
 841                 $cacheKey = $inLanguage === self::AS_AUTONYMS ? 'null' : $inLanguage;
 842                 $cacheKey .= ":$include";
 843                 if ( self::$languageNameCache === null ) {
 844                         self::$languageNameCache = new HashBagOStuff( [ 'maxKeys' => 20 ] );
 845                 }
 846
 847                 $ret = self::$languageNameCache->get( $cacheKey );
 848                 if ( !$ret ) {
 849                         $ret = self::fetchLanguageNamesUncached( $inLanguage, $include );
 850                         self::$languageNameCache->set( $cacheKey, $ret );
 851                 }
 852                 return $ret;
 853         }
 854
 855         /**
 856          * Uncached helper for fetchLanguageNames
 857          * @param null|string $inLanguage Code of language in which to return the names
 858          *              Use self::AS_AUTONYMS for autonyms (native names)
 859          * @param string $include One of:
 860          *              self::ALL all available languages
 861          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 862          *              self::SUPPORTED only if the language is in 'mw' *and* has a message file
 863          * @return array Language code => language name (sorted by key)
 864          */
 865         private static function fetchLanguageNamesUncached(
 866                 $inLanguage = self::AS_AUTONYMS,
 867                 $include = 'mw'
 868         ) {
 869                 global $wgExtraLanguageNames, $wgUsePigLatinVariant;
 870
 871                 // If passed an invalid language code to use, fallback to en
 872                 if ( $inLanguage !== self::AS_AUTONYMS && !self::isValidCode( $inLanguage ) ) {
 873                         $inLanguage = 'en';
 874                 }
 875
 876                 $names = [];
 877
 878                 if ( $inLanguage ) {
 879                         # TODO: also include when $inLanguage is null, when this code is more efficient
 880                         Hooks::run( 'LanguageGetTranslatedLanguageNames', [ &$names, $inLanguage ] );
 881                 }
 882
 883                 $mwNames = $wgExtraLanguageNames + MediaWiki\Languages\Data\Names::$names;
 884                 if ( $wgUsePigLatinVariant ) {
 885                         // Pig Latin (for variant development)
 886                         $mwNames['en-x-piglatin'] = 'Igpay Atinlay';
 887                 }
 888
 889                 foreach ( $mwNames as $mwCode => $mwName ) {
 890                         # - Prefer own MediaWiki native name when not using the hook
 891                         # - For other names just add if not added through the hook
 892                         if ( $mwCode === $inLanguage || !isset( $names[$mwCode] ) ) {
 893                                 $names[$mwCode] = $mwName;
 894                         }
 895                 }
 896
 897                 if ( $include === self::ALL ) {
 898                         ksort( $names );
 899                         return $names;
 900                 }
 901
 902                 $returnMw = [];
 903                 $coreCodes = array_keys( $mwNames );
 904                 foreach ( $coreCodes as $coreCode ) {
 905                         $returnMw[$coreCode] = $names[$coreCode];
 906                 }
 907
 908                 if ( $include === self::SUPPORTED ) {
 909                         $namesMwFile = [];
 910                         # We do this using a foreach over the codes instead of a directory
 911                         # loop so that messages files in extensions will work correctly.
 912                         foreach ( $returnMw as $code => $value ) {
 913                                 if ( is_readable( self::getMessagesFileName( $code ) )
 914                                         || is_readable( self::getJsonMessagesFileName( $code ) )
 915                                 ) {
 916                                         $namesMwFile[$code] = $names[$code];
 917                                 }
 918                         }
 919
 920                         ksort( $namesMwFile );
 921                         return $namesMwFile;
 922                 }
 923
 924                 ksort( $returnMw );
 925                 # 'mw' option; default if it's not one of the other two options (all/mwfile)
 926                 return $returnMw;
 927         }
 928
 929         /**
 930          * @param string $code The code of the language for which to get the name
 931          * @param null|string $inLanguage Code of language in which to return the name
 932          *   (SELF::AS_AUTONYMS for autonyms)
 933          * @param string $include See fetchLanguageNames()
 934          * @return string Language name or empty
 935          * @since 1.20
 936          */
 937         public static function fetchLanguageName(
 938                 $code,
 939                 $inLanguage = self::AS_AUTONYMS,
 940                 $include = self::ALL
 941         ) {
 942                 $code = strtolower( $code );
 943                 $array = self::fetchLanguageNames( $inLanguage, $include );
 944                 return !array_key_exists( $code, $array ) ? '' : $array[$code];
 945         }
 946
 947         /**
 948          * Get a message from the MediaWiki namespace.
 949          *
 950          * @param string $msg Message name
 951          * @return string
 952          */
 953         public function getMessageFromDB( $msg ) {
 954                 return $this->msg( $msg )->text();
 955         }
 956
 957         /**
 958          * Get message object in this language. Only for use inside this class.
 959          *
 960          * @param string $msg Message name
 961          * @return Message
 962          */
 963         protected function msg( $msg ) {
 964                 return wfMessage( $msg )->inLanguage( $this );
 965         }
 966
 967         /**
 968          * @param string $key
 969          * @return string
 970          */
 971         public function getMonthName( $key ) {
 972                 return $this->getMessageFromDB( self::$mMonthMsgs[$key - 1] );
 973         }
 974
 975         /**
 976          * @return array
 977          */
 978         public function getMonthNamesArray() {
 979                 $monthNames = [ '' ];
 980                 for ( $i = 1; $i < 13; $i++ ) {
 981                         $monthNames[] = $this->getMonthName( $i );
 982                 }
 983                 return $monthNames;
 984         }
 985
 986         /**
 987          * @param string $key
 988          * @return string
 989          */
 990         public function getMonthNameGen( $key ) {
 991                 return $this->getMessageFromDB( self::$mMonthGenMsgs[$key - 1] );
 992         }
 993
 994         /**
 995          * @param string $key
 996          * @return string
 997          */
 998         public function getMonthAbbreviation( $key ) {
 999                 return $this->getMessageFromDB( self::$mMonthAbbrevMsgs[$key - 1] );
1000         }
1001
1002         /**
1003          * @return array
1004          */
1005         public function getMonthAbbreviationsArray() {
1006                 $monthNames = [ '' ];
1007                 for ( $i = 1; $i < 13; $i++ ) {
1008                         $monthNames[] = $this->getMonthAbbreviation( $i );
1009                 }
1010                 return $monthNames;
1011         }
1012
1013         /**
1014          * @param string $key
1015          * @return string
1016          */
1017         public function getWeekdayName( $key ) {
1018                 return $this->getMessageFromDB( self::$mWeekdayMsgs[$key - 1] );
1019         }
1020
1021         /**
1022          * @param string $key
1023          * @return string
1024          */
1025         function getWeekdayAbbreviation( $key ) {
1026                 return $this->getMessageFromDB( self::$mWeekdayAbbrevMsgs[$key - 1] );
1027         }
1028
1029         /**
1030          * @param string $key
1031          * @return string
1032          */
1033         function getIranianCalendarMonthName( $key ) {
1034                 return $this->getMessageFromDB( self::$mIranianCalendarMonthMsgs[$key - 1] );
1035         }
1036
1037         /**
1038          * @param string $key
1039          * @return string
1040          */
1041         function getHebrewCalendarMonthName( $key ) {
1042                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthMsgs[$key - 1] );
1043         }
1044
1045         /**
1046          * @param string $key
1047          * @return string
1048          */
1049         function getHebrewCalendarMonthNameGen( $key ) {
1050                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthGenMsgs[$key - 1] );
1051         }
1052
1053         /**
1054          * @param string $key
1055          * @return string
1056          */
1057         function getHijriCalendarMonthName( $key ) {
1058                 return $this->getMessageFromDB( self::$mHijriCalendarMonthMsgs[$key - 1] );
1059         }
1060
1061         /**
1062          * Pass through result from $dateTimeObj->format()
1063          * @param DateTime|bool|null &$dateTimeObj
1064          * @param string $ts
1065          * @param DateTimeZone|bool|null $zone
1066          * @param string $code
1067          * @return string
1068          */
1069         private static function dateTimeObjFormat( &$dateTimeObj, $ts, $zone, $code ) {
1070                 if ( !$dateTimeObj ) {
1071                         $dateTimeObj = DateTime::createFromFormat(
1072                                 'YmdHis', $ts, $zone ?: new DateTimeZone( 'UTC' )
1073                         );
1074                 }
1075                 return $dateTimeObj->format( $code );
1076         }
1077
1078         /**
1079          * This is a workalike of PHP's date() function, but with better
1080          * internationalisation, a reduced set of format characters, and a better
1081          * escaping format.
1082          *
1083          * Supported format characters are dDjlNwzWFmMntLoYyaAgGhHiscrUeIOPTZ. See
1084          * the PHP manual for definitions. There are a number of extensions, which
1085          * start with "x":
1086          *
1087          *    xn   Do not translate digits of the next numeric format character
1088          *    xN   Toggle raw digit (xn) flag, stays set until explicitly unset
1089          *    xr   Use roman numerals for the next numeric format character
1090          *    xh   Use hebrew numerals for the next numeric format character
1091          *    xx   Literal x
1092          *    xg   Genitive month name
1093          *
1094          *    xij  j (day number) in Iranian calendar
1095          *    xiF  F (month name) in Iranian calendar
1096          *    xin  n (month number) in Iranian calendar
1097          *    xiy  y (two digit year) in Iranian calendar
1098          *    xiY  Y (full year) in Iranian calendar
1099          *    xit  t (days in month) in Iranian calendar
1100          *    xiz  z (day of the year) in Iranian calendar
1101          *
1102          *    xjj  j (day number) in Hebrew calendar
1103          *    xjF  F (month name) in Hebrew calendar
1104          *    xjt  t (days in month) in Hebrew calendar
1105          *    xjx  xg (genitive month name) in Hebrew calendar
1106          *    xjn  n (month number) in Hebrew calendar
1107          *    xjY  Y (full year) in Hebrew calendar
1108          *
1109          *    xmj  j (day number) in Hijri calendar
1110          *    xmF  F (month name) in Hijri calendar
1111          *    xmn  n (month number) in Hijri calendar
1112          *    xmY  Y (full year) in Hijri calendar
1113          *
1114          *    xkY  Y (full year) in Thai solar calendar. Months and days are
1115          *                       identical to the Gregorian calendar
1116          *    xoY  Y (full year) in Minguo calendar or Juche year.
1117          *                       Months and days are identical to the
1118          *                       Gregorian calendar
1119          *    xtY  Y (full year) in Japanese nengo. Months and days are
1120          *                       identical to the Gregorian calendar
1121          *
1122          * Characters enclosed in double quotes will be considered literal (with
1123          * the quotes themselves removed). Unmatched quotes will be considered
1124          * literal quotes. Example:
1125          *
1126          * "The month is" F       => The month is January
1127          * i's"                   => 20'11"
1128          *
1129          * Backslash escaping is also supported.
1130          *
1131          * Input timestamp is assumed to be pre-normalized to the desired local
1132          * time zone, if any. Note that the format characters crUeIOPTZ will assume
1133          * $ts is UTC if $zone is not given.
1134          *
1135          * @param string $format
1136          * @param string $ts 14-character timestamp
1137          *      YYYYMMDDHHMMSS
1138          *      01234567890123
1139          * @param DateTimeZone|null $zone Timezone of $ts
1140          * @param int &$ttl The amount of time (in seconds) the output may be cached for.
1141          * Only makes sense if $ts is the current time.
1142          * @todo handling of "o" format character for Iranian, Hebrew, Hijri & Thai?
1143          *
1144          * @throws MWException
1145          * @return string
1146          */
1147         public function sprintfDate( $format, $ts, DateTimeZone $zone = null, &$ttl = 'unused' ) {
1148                 $s = '';
1149                 $raw = false;
1150                 $roman = false;
1151                 $hebrewNum = false;
1152                 $dateTimeObj = false;
1153                 $rawToggle = false;
1154                 $iranian = false;
1155                 $hebrew = false;
1156                 $hijri = false;
1157                 $thai = false;
1158                 $minguo = false;
1159                 $tenno = false;
1160
1161                 $usedSecond = false;
1162                 $usedMinute = false;
1163                 $usedHour = false;
1164                 $usedAMPM = false;
1165                 $usedDay = false;
1166                 $usedWeek = false;
1167                 $usedMonth = false;
1168                 $usedYear = false;
1169                 $usedISOYear = false;
1170                 $usedIsLeapYear = false;
1171
1172                 $usedHebrewMonth = false;
1173                 $usedIranianMonth = false;
1174                 $usedHijriMonth = false;
1175                 $usedHebrewYear = false;
1176                 $usedIranianYear = false;
1177                 $usedHijriYear = false;
1178                 $usedTennoYear = false;
1179
1180                 if ( strlen( $ts ) !== 14 ) {
1181                         throw new MWException( __METHOD__ . ": The timestamp $ts should have 14 characters" );
1182                 }
1183
1184                 if ( !ctype_digit( $ts ) ) {
1185                         throw new MWException( __METHOD__ . ": The timestamp $ts should be a number" );
1186                 }
1187
1188                 $formatLength = strlen( $format );
1189                 for ( $p = 0; $p < $formatLength; $p++ ) {
1190                         $num = false;
1191                         $code = $format[$p];
1192                         if ( $code == 'x' && $p < $formatLength - 1 ) {
1193                                 $code .= $format[++$p];
1194                         }
1195
1196                         if ( ( $code === 'xi'
1197                                         || $code === 'xj'
1198                                         || $code === 'xk'
1199                                         || $code === 'xm'
1200                                         || $code === 'xo'
1201                                         || $code === 'xt' )
1202                                 && $p < $formatLength - 1 ) {
1203                                 $code .= $format[++$p];
1204                         }
1205
1206                         switch ( $code ) {
1207                                 case 'xx':
1208                                         $s .= 'x';
1209                                         break;
1210                                 case 'xn':
1211                                         $raw = true;
1212                                         break;
1213                                 case 'xN':
1214                                         $rawToggle = !$rawToggle;
1215                                         break;
1216                                 case 'xr':
1217                                         $roman = true;
1218                                         break;
1219                                 case 'xh':
1220                                         $hebrewNum = true;
1221                                         break;
1222                                 case 'xg':
1223                                         $usedMonth = true;
1224                                         $s .= $this->getMonthNameGen( substr( $ts, 4, 2 ) );
1225                                         break;
1226                                 case 'xjx':
1227                                         $usedHebrewMonth = true;
1228                                         if ( !$hebrew ) {
1229                                                 $hebrew = self::tsToHebrew( $ts );
1230                                         }
1231                                         $s .= $this->getHebrewCalendarMonthNameGen( $hebrew[1] );
1232                                         break;
1233                                 case 'd':
1234                                         $usedDay = true;
1235                                         $num = substr( $ts, 6, 2 );
1236                                         break;
1237                                 case 'D':
1238                                         $usedDay = true;
1239                                         $s .= $this->getWeekdayAbbreviation(
1240                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1241                                         );
1242                                         break;
1243                                 case 'j':
1244                                         $usedDay = true;
1245                                         $num = intval( substr( $ts, 6, 2 ) );
1246                                         break;
1247                                 case 'xij':
1248                                         $usedDay = true;
1249                                         if ( !$iranian ) {
1250                                                 $iranian = self::tsToIranian( $ts );
1251                                         }
1252                                         $num = $iranian[2];
1253                                         break;
1254                                 case 'xmj':
1255                                         $usedDay = true;
1256                                         if ( !$hijri ) {
1257                                                 $hijri = self::tsToHijri( $ts );
1258                                         }
1259                                         $num = $hijri[2];
1260                                         break;
1261                                 case 'xjj':
1262                                         $usedDay = true;
1263                                         if ( !$hebrew ) {
1264                                                 $hebrew = self::tsToHebrew( $ts );
1265                                         }
1266                                         $num = $hebrew[2];
1267                                         break;
1268                                 case 'l':
1269                                         $usedDay = true;
1270                                         $s .= $this->getWeekdayName(
1271                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1272                                         );
1273                                         break;
1274                                 case 'F':
1275                                         $usedMonth = true;
1276                                         $s .= $this->getMonthName( substr( $ts, 4, 2 ) );
1277                                         break;
1278                                 case 'xiF':
1279                                         $usedIranianMonth = true;
1280                                         if ( !$iranian ) {
1281                                                 $iranian = self::tsToIranian( $ts );
1282                                         }
1283                                         $s .= $this->getIranianCalendarMonthName( $iranian[1] );
1284                                         break;
1285                                 case 'xmF':
1286                                         $usedHijriMonth = true;
1287                                         if ( !$hijri ) {
1288                                                 $hijri = self::tsToHijri( $ts );
1289                                         }
1290                                         $s .= $this->getHijriCalendarMonthName( $hijri[1] );
1291                                         break;
1292                                 case 'xjF':
1293                                         $usedHebrewMonth = true;
1294                                         if ( !$hebrew ) {
1295                                                 $hebrew = self::tsToHebrew( $ts );
1296                                         }
1297                                         $s .= $this->getHebrewCalendarMonthName( $hebrew[1] );
1298                                         break;
1299                                 case 'm':
1300                                         $usedMonth = true;
1301                                         $num = substr( $ts, 4, 2 );
1302                                         break;
1303                                 case 'M':
1304                                         $usedMonth = true;
1305                                         $s .= $this->getMonthAbbreviation( substr( $ts, 4, 2 ) );
1306                                         break;
1307                                 case 'n':
1308                                         $usedMonth = true;
1309                                         $num = intval( substr( $ts, 4, 2 ) );
1310                                         break;
1311                                 case 'xin':
1312                                         $usedIranianMonth = true;
1313                                         if ( !$iranian ) {
1314                                                 $iranian = self::tsToIranian( $ts );
1315                                         }
1316                                         $num = $iranian[1];
1317                                         break;
1318                                 case 'xmn':
1319                                         $usedHijriMonth = true;
1320                                         if ( !$hijri ) {
1321                                                 $hijri = self::tsToHijri( $ts );
1322                                         }
1323                                         $num = $hijri[1];
1324                                         break;
1325                                 case 'xjn':
1326                                         $usedHebrewMonth = true;
1327                                         if ( !$hebrew ) {
1328                                                 $hebrew = self::tsToHebrew( $ts );
1329                                         }
1330                                         $num = $hebrew[1];
1331                                         break;
1332                                 case 'xjt':
1333                                         $usedHebrewMonth = true;
1334                                         if ( !$hebrew ) {
1335                                                 $hebrew = self::tsToHebrew( $ts );
1336                                         }
1337                                         $num = $hebrew[3];
1338                                         break;
1339                                 case 'Y':
1340                                         $usedYear = true;
1341                                         $num = substr( $ts, 0, 4 );
1342                                         break;
1343                                 case 'xiY':
1344                                         $usedIranianYear = true;
1345                                         if ( !$iranian ) {
1346                                                 $iranian = self::tsToIranian( $ts );
1347                                         }
1348                                         $num = $iranian[0];
1349                                         break;
1350                                 case 'xmY':
1351                                         $usedHijriYear = true;
1352                                         if ( !$hijri ) {
1353                                                 $hijri = self::tsToHijri( $ts );
1354                                         }
1355                                         $num = $hijri[0];
1356                                         break;
1357                                 case 'xjY':
1358                                         $usedHebrewYear = true;
1359                                         if ( !$hebrew ) {
1360                                                 $hebrew = self::tsToHebrew( $ts );
1361                                         }
1362                                         $num = $hebrew[0];
1363                                         break;
1364                                 case 'xkY':
1365                                         $usedYear = true;
1366                                         if ( !$thai ) {
1367                                                 $thai = self::tsToYear( $ts, 'thai' );
1368                                         }
1369                                         $num = $thai[0];
1370                                         break;
1371                                 case 'xoY':
1372                                         $usedYear = true;
1373                                         if ( !$minguo ) {
1374                                                 $minguo = self::tsToYear( $ts, 'minguo' );
1375                                         }
1376                                         $num = $minguo[0];
1377                                         break;
1378                                 case 'xtY':
1379                                         $usedTennoYear = true;
1380                                         if ( !$tenno ) {
1381                                                 $tenno = self::tsToYear( $ts, 'tenno' );
1382                                         }
1383                                         $num = $tenno[0];
1384                                         break;
1385                                 case 'y':
1386                                         $usedYear = true;
1387                                         $num = substr( $ts, 2, 2 );
1388                                         break;
1389                                 case 'xiy':
1390                                         $usedIranianYear = true;
1391                                         if ( !$iranian ) {
1392                                                 $iranian = self::tsToIranian( $ts );
1393                                         }
1394                                         $num = substr( $iranian[0], -2 );
1395                                         break;
1396                                 case 'xit':
1397                                         $usedIranianYear = true;
1398                                         if ( !$iranian ) {
1399                                                 $iranian = self::tsToIranian( $ts );
1400                                         }
1401                                         $num = self::$IRANIAN_DAYS[$iranian[1] - 1];
1402                                         break;
1403                                 case 'xiz':
1404                                         $usedIranianYear = true;
1405                                         if ( !$iranian ) {
1406                                                 $iranian = self::tsToIranian( $ts );
1407                                         }
1408                                         $num = $iranian[3];
1409                                         break;
1410                                 case 'a':
1411                                         $usedAMPM = true;
1412                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'am' : 'pm';
1413                                         break;
1414                                 case 'A':
1415                                         $usedAMPM = true;
1416                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'AM' : 'PM';
1417                                         break;
1418                                 case 'g':
1419                                         $usedHour = true;
1420                                         $h = substr( $ts, 8, 2 );
1421                                         $num = $h % 12 ?: 12;
1422                                         break;
1423                                 case 'G':
1424                                         $usedHour = true;
1425                                         $num = intval( substr( $ts, 8, 2 ) );
1426                                         break;
1427                                 case 'h':
1428                                         $usedHour = true;
1429                                         $h = substr( $ts, 8, 2 );
1430                                         $num = sprintf( '%02d', $h % 12 ?: 12 );
1431                                         break;
1432                                 case 'H':
1433                                         $usedHour = true;
1434                                         $num = substr( $ts, 8, 2 );
1435                                         break;
1436                                 case 'i':
1437                                         $usedMinute = true;
1438                                         $num = substr( $ts, 10, 2 );
1439                                         break;
1440                                 case 's':
1441                                         $usedSecond = true;
1442                                         $num = substr( $ts, 12, 2 );
1443                                         break;
1444                                 case 'c':
1445                                 case 'r':
1446                                         $usedSecond = true;
1447                                         // fall through
1448                                 case 'e':
1449                                 case 'O':
1450                                 case 'P':
1451                                 case 'T':
1452                                         $s .= self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1453                                         break;
1454                                 case 'w':
1455                                 case 'N':
1456                                 case 'z':
1457                                         $usedDay = true;
1458                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1459                                         break;
1460                                 case 'W':
1461                                         $usedWeek = true;
1462                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1463                                         break;
1464                                 case 't':
1465                                         $usedMonth = true;
1466                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1467                                         break;
1468                                 case 'L':
1469                                         $usedIsLeapYear = true;
1470                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1471                                         break;
1472                                 case 'o':
1473                                         $usedISOYear = true;
1474                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1475                                         break;
1476                                 case 'U':
1477                                         $usedSecond = true;
1478                                         // fall through
1479                                 case 'I':
1480                                 case 'Z':
1481                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1482                                         break;
1483                                 case '\\':
1484                                         # Backslash escaping
1485                                         if ( $p < $formatLength - 1 ) {
1486                                                 $s .= $format[++$p];
1487                                         } else {
1488                                                 $s .= '\\';
1489                                         }
1490                                         break;
1491                                 case '"':
1492                                         # Quoted literal
1493                                         if ( $p < $formatLength - 1 ) {
1494                                                 $endQuote = strpos( $format, '"', $p + 1 );
1495                                                 if ( $endQuote === false ) {
1496                                                         # No terminating quote, assume literal "
1497                                                         $s .= '"';
1498                                                 } else {
1499                                                         $s .= substr( $format, $p + 1, $endQuote - $p - 1 );
1500                                                         $p = $endQuote;
1501                                                 }
1502                                         } else {
1503                                                 # Quote at end of string, assume literal "
1504                                                 $s .= '"';
1505                                         }
1506                                         break;
1507                                 default:
1508                                         $s .= $format[$p];
1509                         }
1510                         if ( $num !== false ) {
1511                                 if ( $rawToggle || $raw ) {
1512                                         $s .= $num;
1513                                         $raw = false;
1514                                 } elseif ( $roman ) {
1515                                         $s .= self::romanNumeral( $num );
1516                                         $roman = false;
1517                                 } elseif ( $hebrewNum ) {
1518                                         $s .= self::hebrewNumeral( $num );
1519                                         $hebrewNum = false;
1520                                 } else {
1521                                         $s .= $this->formatNum( $num, true );
1522                                 }
1523                         }
1524                 }
1525
1526                 if ( $ttl === 'unused' ) {
1527                         // No need to calculate the TTL, the caller wont use it anyway.
1528                 } elseif ( $usedSecond ) {
1529                         $ttl = 1;
1530                 } elseif ( $usedMinute ) {
1531                         $ttl = 60 - substr( $ts, 12, 2 );
1532                 } elseif ( $usedHour ) {
1533                         $ttl = 3600 - substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1534                 } elseif ( $usedAMPM ) {
1535                         $ttl = 43200 - ( substr( $ts, 8, 2 ) % 12 ) * 3600 -
1536                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1537                 } elseif (
1538                         $usedDay ||
1539                         $usedHebrewMonth ||
1540                         $usedIranianMonth ||
1541                         $usedHijriMonth ||
1542                         $usedHebrewYear ||
1543                         $usedIranianYear ||
1544                         $usedHijriYear ||
1545                         $usedTennoYear
1546                 ) {
1547                         // @todo Someone who understands the non-Gregorian calendars
1548                         // should write proper logic for them so that they don't need purged every day.
1549                         $ttl = 86400 - substr( $ts, 8, 2 ) * 3600 -
1550                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1551                 } else {
1552                         $possibleTtls = [];
1553                         $timeRemainingInDay = 86400 - substr( $ts, 8, 2 ) * 3600 -
1554                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1555                         if ( $usedWeek ) {
1556                                 $possibleTtls[] =
1557                                         ( 7 - self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400 +
1558                                         $timeRemainingInDay;
1559                         } elseif ( $usedISOYear ) {
1560                                 // December 28th falls on the last ISO week of the year, every year.
1561                                 // The last ISO week of a year can be 52 or 53.
1562                                 $lastWeekOfISOYear = DateTime::createFromFormat(
1563                                         'Ymd',
1564                                         substr( $ts, 0, 4 ) . '1228',
1565                                         $zone ?: new DateTimeZone( 'UTC' )
1566                                 )->format( 'W' );
1567                                 $currentISOWeek = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'W' );
1568                                 $weeksRemaining = $lastWeekOfISOYear - $currentISOWeek;
1569                                 $timeRemainingInWeek =
1570                                         ( 7 - self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400
1571                                         + $timeRemainingInDay;
1572                                 $possibleTtls[] = $weeksRemaining * 604800 + $timeRemainingInWeek;
1573                         }
1574
1575                         if ( $usedMonth ) {
1576                                 $possibleTtls[] =
1577                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 't' ) -
1578                                                 substr( $ts, 6, 2 ) ) * 86400
1579                                         + $timeRemainingInDay;
1580                         } elseif ( $usedYear ) {
1581                                 $possibleTtls[] =
1582                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1583                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1584                                         + $timeRemainingInDay;
1585                         } elseif ( $usedIsLeapYear ) {
1586                                 $year = substr( $ts, 0, 4 );
1587                                 $timeRemainingInYear =
1588                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1589                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1590                                         + $timeRemainingInDay;
1591                                 $mod = $year % 4;
1592                                 if ( $mod || ( !( $year % 100 ) && $year % 400 ) ) {
1593                                         // this isn't a leap year. see when the next one starts
1594                                         $nextCandidate = $year - $mod + 4;
1595                                         if ( $nextCandidate % 100 || !( $nextCandidate % 400 ) ) {
1596                                                 $possibleTtls[] = ( $nextCandidate - $year - 1 ) * 365 * 86400 +
1597                                                         $timeRemainingInYear;
1598                                         } else {
1599                                                 $possibleTtls[] = ( $nextCandidate - $year + 3 ) * 365 * 86400 +
1600                                                         $timeRemainingInYear;
1601                                         }
1602                                 } else {
1603                                         // this is a leap year, so the next year isn't
1604                                         $possibleTtls[] = $timeRemainingInYear;
1605                                 }
1606                         }
1607
1608                         if ( $possibleTtls ) {
1609                                 $ttl = min( $possibleTtls );
1610                         }
1611                 }
1612
1613                 return $s;
1614         }
1615
1616         private static $GREG_DAYS = [ 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 ];
1617         private static $IRANIAN_DAYS = [ 31, 31, 31, 31, 31, 31, 30, 30, 30, 30, 30, 29 ];
1618
1619         /**
1620          * Algorithm by Roozbeh Pournader and Mohammad Toossi to convert
1621          * Gregorian dates to Iranian dates. Originally written in C, it
1622          * is released under the terms of GNU Lesser General Public
1623          * License. Conversion to PHP was performed by Niklas Laxström.
1624          *
1625          * Link: http://www.farsiweb.info/jalali/jalali.c
1626          *
1627          * @param string $ts
1628          *
1629          * @return int[]
1630          */
1631         private static function tsToIranian( $ts ) {
1632                 $gy = substr( $ts, 0, 4 ) - 1600;
1633                 $gm = substr( $ts, 4, 2 ) - 1;
1634                 $gd = substr( $ts, 6, 2 ) - 1;
1635
1636                 # Days passed from the beginning (including leap years)
1637                 $gDayNo = 365 * $gy
1638                         + floor( ( $gy + 3 ) / 4 )
1639                         - floor( ( $gy + 99 ) / 100 )
1640                         + floor( ( $gy + 399 ) / 400 );
1641
1642                 // Add days of the past months of this year
1643                 for ( $i = 0; $i < $gm; $i++ ) {
1644                         $gDayNo += self::$GREG_DAYS[$i];
1645                 }
1646
1647                 // Leap years
1648                 if ( $gm > 1 && ( ( $gy % 4 === 0 && $gy % 100 !== 0 || ( $gy % 400 == 0 ) ) ) ) {
1649                         $gDayNo++;
1650                 }
1651
1652                 // Days passed in current month
1653                 $gDayNo += (int)$gd;
1654
1655                 $jDayNo = $gDayNo - 79;
1656
1657                 $jNp = floor( $jDayNo / 12053 );
1658                 $jDayNo %= 12053;
1659
1660                 $jy = 979 + 33 * $jNp + 4 * floor( $jDayNo / 1461 );
1661                 $jDayNo %= 1461;
1662
1663                 if ( $jDayNo >= 366 ) {
1664                         $jy += floor( ( $jDayNo - 1 ) / 365 );
1665                         $jDayNo = floor( ( $jDayNo - 1 ) % 365 );
1666                 }
1667
1668                 $jz = $jDayNo;
1669
1670                 for ( $i = 0; $i < 11 && $jDayNo >= self::$IRANIAN_DAYS[$i]; $i++ ) {
1671                         $jDayNo -= self::$IRANIAN_DAYS[$i];
1672                 }
1673
1674                 $jm = $i + 1;
1675                 $jd = $jDayNo + 1;
1676
1677                 return [ $jy, $jm, $jd, $jz ];
1678         }
1679
1680         /**
1681          * Converting Gregorian dates to Hijri dates.
1682          *
1683          * Based on a PHP-Nuke block by Sharjeel which is released under GNU/GPL license
1684          *
1685          * @see https://phpnuke.org/modules.php?name=News&file=article&sid=8234&mode=thread&order=0&thold=0
1686          *
1687          * @param string $ts
1688          *
1689          * @return int[]
1690          */
1691         private static function tsToHijri( $ts ) {
1692                 $year = substr( $ts, 0, 4 );
1693                 $month = substr( $ts, 4, 2 );
1694                 $day = substr( $ts, 6, 2 );
1695
1696                 $zyr = $year;
1697                 $zd = $day;
1698                 $zm = $month;
1699                 $zy = $zyr;
1700
1701                 if (
1702                         ( $zy > 1582 ) || ( ( $zy == 1582 ) && ( $zm > 10 ) ) ||
1703                         ( ( $zy == 1582 ) && ( $zm == 10 ) && ( $zd > 14 ) )
1704                 ) {
1705                         $zjd = (int)( ( 1461 * ( $zy + 4800 + (int)( ( $zm - 14 ) / 12 ) ) ) / 4 ) +
1706                                         (int)( ( 367 * ( $zm - 2 - 12 * ( (int)( ( $zm - 14 ) / 12 ) ) ) ) / 12 ) -
1707                                         (int)( ( 3 * (int)( ( ( $zy + 4900 + (int)( ( $zm - 14 ) / 12 ) ) / 100 ) ) ) / 4 ) +
1708                                         $zd - 32075;
1709                 } else {
1710                         $zjd = 367 * $zy - (int)( ( 7 * ( $zy + 5001 + (int)( ( $zm - 9 ) / 7 ) ) ) / 4 ) +
1711                                                                 (int)( ( 275 * $zm ) / 9 ) + $zd + 1729777;
1712                 }
1713
1714                 $zl = $zjd - 1948440 + 10632;
1715                 $zn = (int)( ( $zl - 1 ) / 10631 );
1716                 $zl = $zl - 10631 * $zn + 354;
1717                 $zj = ( (int)( ( 10985 - $zl ) / 5316 ) ) * ( (int)( ( 50 * $zl ) / 17719 ) ) +
1718                         ( (int)( $zl / 5670 ) ) * ( (int)( ( 43 * $zl ) / 15238 ) );
1719                 $zl = $zl - ( (int)( ( 30 - $zj ) / 15 ) ) * ( (int)( ( 17719 * $zj ) / 50 ) ) -
1720                         ( (int)( $zj / 16 ) ) * ( (int)( ( 15238 * $zj ) / 43 ) ) + 29;
1721                 $zm = (int)( ( 24 * $zl ) / 709 );
1722                 $zd = $zl - (int)( ( 709 * $zm ) / 24 );
1723                 $zy = 30 * $zn + $zj - 30;
1724
1725                 return [ $zy, $zm, $zd ];
1726         }
1727
1728         /**
1729          * Converting Gregorian dates to Hebrew dates.
1730          *
1731          * Based on a JavaScript code by Abu Mami and Yisrael Hersch
1732          * (abu-mami@kaluach.net, http://www.kaluach.net), who permitted
1733          * to translate the relevant functions into PHP and release them under
1734          * GNU GPL.
1735          *
1736          * The months are counted from Tishrei = 1. In a leap year, Adar I is 13
1737          * and Adar II is 14. In a non-leap year, Adar is 6.
1738          *
1739          * @param string $ts
1740          *
1741          * @return int[]
1742          */
1743         private static function tsToHebrew( $ts ) {
1744                 # Parse date
1745                 $year = substr( $ts, 0, 4 );
1746                 $month = substr( $ts, 4, 2 );
1747                 $day = substr( $ts, 6, 2 );
1748
1749                 # Calculate Hebrew year
1750                 $hebrewYear = $year + 3760;
1751
1752                 # Month number when September = 1, August = 12
1753                 $month += 4;
1754                 if ( $month > 12 ) {
1755                         # Next year
1756                         $month -= 12;
1757                         $year++;
1758                         $hebrewYear++;
1759                 }
1760
1761                 # Calculate day of year from 1 September
1762                 $dayOfYear = $day;
1763                 for ( $i = 1; $i < $month; $i++ ) {
1764                         if ( $i == 6 ) {
1765                                 # February
1766                                 $dayOfYear += 28;
1767                                 # Check if the year is leap
1768                                 if ( $year % 400 == 0 || ( $year % 4 == 0 && $year % 100 > 0 ) ) {
1769                                         $dayOfYear++;
1770                                 }
1771                         } elseif ( $i == 8 || $i == 10 || $i == 1 || $i == 3 ) {
1772                                 $dayOfYear += 30;
1773                         } else {
1774                                 $dayOfYear += 31;
1775                         }
1776                 }
1777
1778                 # Calculate the start of the Hebrew year
1779                 $start = self::hebrewYearStart( $hebrewYear );
1780
1781                 # Calculate next year's start
1782                 if ( $dayOfYear <= $start ) {
1783                         # Day is before the start of the year - it is the previous year
1784                         # Next year's start
1785                         $nextStart = $start;
1786                         # Previous year
1787                         $year--;
1788                         $hebrewYear--;
1789                         # Add days since previous year's 1 September
1790                         $dayOfYear += 365;
1791                         if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1792                                 # Leap year
1793                                 $dayOfYear++;
1794                         }
1795                         # Start of the new (previous) year
1796                         $start = self::hebrewYearStart( $hebrewYear );
1797                 } else {
1798                         # Next year's start
1799                         $nextStart = self::hebrewYearStart( $hebrewYear + 1 );
1800                 }
1801
1802                 # Calculate Hebrew day of year
1803                 $hebrewDayOfYear = $dayOfYear - $start;
1804
1805                 # Difference between year's days
1806                 $diff = $nextStart - $start;
1807                 # Add 12 (or 13 for leap years) days to ignore the difference between
1808                 # Hebrew and Gregorian year (353 at least vs. 365/6) - now the
1809                 # difference is only about the year type
1810                 if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1811                         $diff += 13;
1812                 } else {
1813                         $diff += 12;
1814                 }
1815
1816                 # Check the year pattern, and is leap year
1817                 # 0 means an incomplete year, 1 means a regular year, 2 means a complete year
1818                 # This is mod 30, to work on both leap years (which add 30 days of Adar I)
1819                 # and non-leap years
1820                 $yearPattern = $diff % 30;
1821                 # Check if leap year
1822                 $isLeap = $diff >= 30;
1823
1824                 # Calculate day in the month from number of day in the Hebrew year
1825                 # Don't check Adar - if the day is not in Adar, we will stop before;
1826                 # if it is in Adar, we will use it to check if it is Adar I or Adar II
1827                 $hebrewDay = $hebrewDayOfYear;
1828                 $hebrewMonth = 1;
1829                 $days = 0;
1830                 while ( $hebrewMonth <= 12 ) {
1831                         # Calculate days in this month
1832                         if ( $isLeap && $hebrewMonth == 6 ) {
1833                                 # Leap year - has Adar I, with 30 days, and Adar II, with 29 days
1834                                 $days = 30;
1835                                 if ( $hebrewDay <= $days ) {
1836                                         # Day in Adar I
1837                                         $hebrewMonth = 13;
1838                                 } else {
1839                                         # Subtract the days of Adar I
1840                                         $hebrewDay -= $days;
1841                                         # Try Adar II
1842                                         $days = 29;
1843                                         if ( $hebrewDay <= $days ) {
1844                                                 # Day in Adar II
1845                                                 $hebrewMonth = 14;
1846                                         }
1847                                 }
1848                         } elseif ( $hebrewMonth == 2 && $yearPattern == 2 ) {
1849                                 # Cheshvan in a complete year (otherwise as the rule below)
1850                                 $days = 30;
1851                         } elseif ( $hebrewMonth == 3 && $yearPattern == 0 ) {
1852                                 # Kislev in an incomplete year (otherwise as the rule below)
1853                                 $days = 29;
1854                         } else {
1855                                 # Odd months have 30 days, even have 29
1856                                 $days = 30 - ( $hebrewMonth - 1 ) % 2;
1857                         }
1858                         if ( $hebrewDay <= $days ) {
1859                                 # In the current month
1860                                 break;
1861                         } else {
1862                                 # Subtract the days of the current month
1863                                 $hebrewDay -= $days;
1864                                 # Try in the next month
1865                                 $hebrewMonth++;
1866                         }
1867                 }
1868
1869                 return [ $hebrewYear, $hebrewMonth, $hebrewDay, $days ];
1870         }
1871
1872         /**
1873          * This calculates the Hebrew year start, as days since 1 September.
1874          * Based on Carl Friedrich Gauss algorithm for finding Easter date.
1875          * Used for Hebrew date.
1876          *
1877          * @param int $year
1878          *
1879          * @return string
1880          */
1881         private static function hebrewYearStart( $year ) {
1882                 $a = intval( ( 12 * ( $year - 1 ) + 17 ) % 19 );
1883                 $b = intval( ( $year - 1 ) % 4 );
1884                 $m = 32.044093161144 + 1.5542417966212 * $a + $b / 4.0 - 0.0031777940220923 * ( $year - 1 );
1885                 if ( $m < 0 ) {
1886                         $m--;
1887                 }
1888                 $Mar = intval( $m );
1889                 if ( $m < 0 ) {
1890                         $m++;
1891                 }
1892                 $m -= $Mar;
1893
1894                 $c = intval( ( $Mar + 3 * ( $year - 1 ) + 5 * $b + 5 ) % 7 );
1895                 if ( $c == 0 && $a > 11 && $m >= 0.89772376543210 ) {
1896                         $Mar++;
1897                 } elseif ( $c == 1 && $a > 6 && $m >= 0.63287037037037 ) {
1898                         $Mar += 2;
1899                 } elseif ( $c == 2 || $c == 4 || $c == 6 ) {
1900                         $Mar++;
1901                 }
1902
1903                 $Mar += intval( ( $year - 3761 ) / 100 ) - intval( ( $year - 3761 ) / 400 ) - 24;
1904                 return $Mar;
1905         }
1906
1907         /**
1908          * Algorithm to convert Gregorian dates to Thai solar dates,
1909          * Minguo dates or Minguo dates.
1910          *
1911          * Link: https://en.wikipedia.org/wiki/Thai_solar_calendar
1912          *       https://en.wikipedia.org/wiki/Minguo_calendar
1913          *       https://en.wikipedia.org/wiki/Japanese_era_name
1914          *
1915          * @param string $ts 14-character timestamp
1916          * @param string $cName Calender name
1917          * @return array Converted year, month, day
1918          */
1919         private static function tsToYear( $ts, $cName ) {
1920                 $gy = substr( $ts, 0, 4 );
1921                 $gm = substr( $ts, 4, 2 );
1922                 $gd = substr( $ts, 6, 2 );
1923
1924                 if ( !strcmp( $cName, 'thai' ) ) {
1925                         # Thai solar dates
1926                         # Add 543 years to the Gregorian calendar
1927                         # Months and days are identical
1928                         $gy_offset = $gy + 543;
1929                         # fix for dates between 1912 and 1941
1930                         # https://en.wikipedia.org/?oldid=836596673#New_year
1931                         if ( $gy >= 1912 && $gy <= 1940 ) {
1932                                 if ( $gm <= 3 ) {
1933                                         $gy_offset--;
1934                                 }
1935                                 $gm = ( $gm - 3 ) % 12;
1936                         }
1937                 } elseif ( ( !strcmp( $cName, 'minguo' ) ) || !strcmp( $cName, 'juche' ) ) {
1938                         # Minguo dates
1939                         # Deduct 1911 years from the Gregorian calendar
1940                         # Months and days are identical
1941                         $gy_offset = $gy - 1911;
1942                 } elseif ( !strcmp( $cName, 'tenno' ) ) {
1943                         # Nengō dates up to Meiji period
1944                         # Deduct years from the Gregorian calendar
1945                         # depending on the nengo periods
1946                         # Months and days are identical
1947                         if ( ( $gy < 1912 )
1948                                 || ( ( $gy == 1912 ) && ( $gm < 7 ) )
1949                                 || ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd < 31 ) )
1950                         ) {
1951                                 # Meiji period
1952                                 $gy_gannen = $gy - 1868 + 1;
1953                                 $gy_offset = $gy_gannen;
1954                                 if ( $gy_gannen == 1 ) {
1955                                         $gy_offset = '元';
1956                                 }
1957                                 $gy_offset = '明治' . $gy_offset;
1958                         } elseif (
1959                                 ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd == 31 ) ) ||
1960                                 ( ( $gy == 1912 ) && ( $gm >= 8 ) ) ||
1961                                 ( ( $gy > 1912 ) && ( $gy < 1926 ) ) ||
1962                                 ( ( $gy == 1926 ) && ( $gm < 12 ) ) ||
1963                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd < 26 ) )
1964                         ) {
1965                                 # Taishō period
1966                                 $gy_gannen = $gy - 1912 + 1;
1967                                 $gy_offset = $gy_gannen;
1968                                 if ( $gy_gannen == 1 ) {
1969                                         $gy_offset = '元';
1970                                 }
1971                                 $gy_offset = '大正' . $gy_offset;
1972                         } elseif (
1973                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd >= 26 ) ) ||
1974                                 ( ( $gy > 1926 ) && ( $gy < 1989 ) ) ||
1975                                 ( ( $gy == 1989 ) && ( $gm == 1 ) && ( $gd < 8 ) )
1976                         ) {
1977                                 # Shōwa period
1978                                 $gy_gannen = $gy - 1926 + 1;
1979                                 $gy_offset = $gy_gannen;
1980                                 if ( $gy_gannen == 1 ) {
1981                                         $gy_offset = '元';
1982                                 }
1983                                 $gy_offset = '昭和' . $gy_offset;
1984                         } elseif (
1985                                 ( ( $gy == 1989 ) && ( $gm == 1 ) && ( $gd >= 8 ) ) ||
1986                                 ( ( $gy > 1989 ) && ( $gy < 2019 ) ) ||
1987                                 ( ( $gy == 2019 ) && ( $gm < 5 ) )
1988                         ) {
1989                                 # Heisei period
1990                                 $gy_gannen = $gy - 1989 + 1;
1991                                 $gy_offset = $gy_gannen;
1992                                 if ( $gy_gannen == 1 ) {
1993                                         $gy_offset = '元';
1994                                 }
1995                                 $gy_offset = '平成' . $gy_offset;
1996                         } else {
1997                                 # Reiwa period
1998                                 $gy_gannen = $gy - 2019 + 1;
1999                                 $gy_offset = $gy_gannen;
2000                                 if ( $gy_gannen == 1 ) {
2001                                         $gy_offset = '元';
2002                                 }
2003                                 $gy_offset = '令和' . $gy_offset;
2004                         }
2005                 } else {
2006                         $gy_offset = $gy;
2007                 }
2008
2009                 return [ $gy_offset, $gm, $gd ];
2010         }
2011
2012         /**
2013          * Gets directionality of the first strongly directional codepoint, for embedBidi()
2014          *
2015          * This is the rule the BIDI algorithm uses to determine the directionality of
2016          * paragraphs ( https://www.unicode.org/reports/tr9/#The_Paragraph_Level ) and
2017          * FSI isolates ( https://www.unicode.org/reports/tr9/#Explicit_Directional_Isolates ).
2018          *
2019          * TODO: Does not handle BIDI control characters inside the text.
2020          * TODO: Does not handle unallocated characters.
2021          *
2022          * @param string $text Text to test
2023          * @return null|string Directionality ('ltr' or 'rtl') or null
2024          */
2025         private static function strongDirFromContent( $text = '' ) {
2026                 if ( !preg_match( self::$strongDirRegex, $text, $matches ) ) {
2027                         return null;
2028                 }
2029                 if ( $matches[1] === '' ) {
2030                         return 'rtl';
2031                 }
2032                 return 'ltr';
2033         }
2034
2035         /**
2036          * Roman number formatting up to 10000
2037          *
2038          * @param int $num
2039          *
2040          * @return string
2041          */
2042         static function romanNumeral( $num ) {
2043                 static $table = [
2044                         [ '', 'I', 'II', 'III', 'IV', 'V', 'VI', 'VII', 'VIII', 'IX', 'X' ],
2045                         [ '', 'X', 'XX', 'XXX', 'XL', 'L', 'LX', 'LXX', 'LXXX', 'XC', 'C' ],
2046                         [ '', 'C', 'CC', 'CCC', 'CD', 'D', 'DC', 'DCC', 'DCCC', 'CM', 'M' ],
2047                         [ '', 'M', 'MM', 'MMM', 'MMMM', 'MMMMM', 'MMMMMM', 'MMMMMMM',
2048                                 'MMMMMMMM', 'MMMMMMMMM', 'MMMMMMMMMM' ]
2049                 ];
2050
2051                 $num = intval( $num );
2052                 if ( $num > 10000 || $num <= 0 ) {
2053                         return $num;
2054                 }
2055
2056                 $s = '';
2057                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
2058                         if ( $num >= $pow10 ) {
2059                                 $s .= $table[$i][(int)floor( $num / $pow10 )];
2060                         }
2061                         $num = $num % $pow10;
2062                 }
2063                 return $s;
2064         }
2065
2066         /**
2067          * Hebrew Gematria number formatting up to 9999
2068          *
2069          * @param int $num
2070          *
2071          * @return string
2072          */
2073         static function hebrewNumeral( $num ) {
2074                 static $table = [
2075                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ],
2076                         [ '', 'י', 'כ', 'ל', 'מ', 'נ', 'ס', 'ע', 'פ', 'צ', 'ק' ],
2077                         [ '',
2078                                 [ 'ק' ],
2079                                 [ 'ר' ],
2080                                 [ 'ש' ],
2081                                 [ 'ת' ],
2082                                 [ 'ת', 'ק' ],
2083                                 [ 'ת', 'ר' ],
2084                                 [ 'ת', 'ש' ],
2085                                 [ 'ת', 'ת' ],
2086                                 [ 'ת', 'ת', 'ק' ],
2087                                 [ 'ת', 'ת', 'ר' ],
2088                         ],
2089                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ]
2090                 ];
2091
2092                 $num = intval( $num );
2093                 if ( $num > 9999 || $num <= 0 ) {
2094                         return $num;
2095                 }
2096
2097                 // Round thousands have special notations
2098                 if ( $num === 1000 ) {
2099                         return "א' אלף";
2100                 } elseif ( $num % 1000 === 0 ) {
2101                         return $table[0][$num / 1000] . "' אלפים";
2102                 }
2103
2104                 $letters = [];
2105
2106                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
2107                         if ( $num >= $pow10 ) {
2108                                 if ( $num === 15 || $num === 16 ) {
2109                                         $letters[] = $table[0][9];
2110                                         $letters[] = $table[0][$num - 9];
2111                                         $num = 0;
2112                                 } else {
2113                                         $letters = array_merge(
2114                                                 $letters,
2115                                                 (array)$table[$i][intval( $num / $pow10 )]
2116                                         );
2117
2118                                         if ( $pow10 === 1000 ) {
2119                                                 $letters[] = "'";
2120                                         }
2121                                 }
2122                         }
2123
2124                         $num = $num % $pow10;
2125                 }
2126
2127                 $preTransformLength = count( $letters );
2128                 if ( $preTransformLength === 1 ) {
2129                         // Add geresh (single quote) to one-letter numbers
2130                         $letters[] = "'";
2131                 } else {
2132                         $lastIndex = $preTransformLength - 1;
2133                         $letters[$lastIndex] = str_replace(
2134                                 [ 'כ', 'מ', 'נ', 'פ', 'צ' ],
2135                                 [ 'ך', 'ם', 'ן', 'ף', 'ץ' ],
2136                                 $letters[$lastIndex]
2137                         );
2138
2139                         // Add gershayim (double quote) to multiple-letter numbers,
2140                         // but exclude numbers with only one letter after the thousands
2141                         // (1001-1009, 1020, 1030, 2001-2009, etc.)
2142                         if ( $letters[1] === "'" && $preTransformLength === 3 ) {
2143                                 $letters[] = "'";
2144                         } else {
2145                                 array_splice( $letters, -1, 0, '"' );
2146                         }
2147                 }
2148
2149                 return implode( $letters );
2150         }
2151
2152         /**
2153          * Used by date() and time() to adjust the time output.
2154          *
2155          * @param string $ts The time in date('YmdHis') format
2156          * @param mixed $tz Adjust the time by this amount (default false, mean we
2157          *   get user timecorrection setting)
2158          * @return int
2159          */
2160         public function userAdjust( $ts, $tz = false ) {
2161                 global $wgUser, $wgLocalTZoffset;
2162
2163                 if ( $tz === false ) {
2164                         $tz = $wgUser->getOption( 'timecorrection' );
2165                 }
2166
2167                 $data = explode( '|', $tz, 3 );
2168
2169                 if ( $data[0] == 'ZoneInfo' ) {
2170                         try {
2171                                 $userTZ = new DateTimeZone( $data[2] );
2172                                 $date = new DateTime( $ts, new DateTimeZone( 'UTC' ) );
2173                                 $date->setTimezone( $userTZ );
2174                                 return $date->format( 'YmdHis' );
2175                         } catch ( Exception $e ) {
2176                                 // Unrecognized timezone, default to 'Offset' with the stored offset.
2177                                 $data[0] = 'Offset';
2178                         }
2179                 }
2180
2181                 if ( $data[0] == 'System' || $tz == '' ) {
2182                         # Global offset in minutes.
2183                         $minDiff = $wgLocalTZoffset;
2184                 } elseif ( $data[0] == 'Offset' ) {
2185                         $minDiff = intval( $data[1] );
2186                 } else {
2187                         $data = explode( ':', $tz );
2188                         if ( count( $data ) == 2 ) {
2189                                 $data[0] = intval( $data[0] );
2190                                 $data[1] = intval( $data[1] );
2191                                 $minDiff = abs( $data[0] ) * 60 + $data[1];
2192                                 if ( $data[0] < 0 ) {
2193                                         $minDiff = -$minDiff;
2194                                 }
2195                         } else {
2196                                 $minDiff = intval( $data[0] ) * 60;
2197                         }
2198                 }
2199
2200                 # No difference ? Return time unchanged
2201                 if ( $minDiff == 0 ) {
2202                         return $ts;
2203                 }
2204
2205                 Wikimedia\suppressWarnings(); // E_STRICT system time bitching
2206                 # Generate an adjusted date; take advantage of the fact that mktime
2207                 # will normalize out-of-range values so we don't have to split $minDiff
2208                 # into hours and minutes.
2209                 $t = mktime( (
2210                         (int)substr( $ts, 8, 2 ) ), # Hours
2211                         (int)substr( $ts, 10, 2 ) + $minDiff, # Minutes
2212                         (int)substr( $ts, 12, 2 ), # Seconds
2213                         (int)substr( $ts, 4, 2 ), # Month
2214                         (int)substr( $ts, 6, 2 ), # Day
2215                         (int)substr( $ts, 0, 4 ) ); # Year
2216
2217                 $date = date( 'YmdHis', $t );
2218                 Wikimedia\restoreWarnings();
2219
2220                 return $date;
2221         }
2222
2223         /**
2224          * This is meant to be used by time(), date(), and timeanddate() to get
2225          * the date preference they're supposed to use, it should be used in
2226          * all children.
2227          *
2228          *     function timeanddate([...], $format = true) {
2229          *       $datePreference = $this->dateFormat($format);
2230          *       [...]
2231          *     }
2232          *
2233          * @param int|string|bool $usePrefs If true, the user's preference is used
2234          *   if false, the site/language default is used
2235          *   if int/string, assumed to be a format.
2236          * @return string
2237          */
2238         function dateFormat( $usePrefs = true ) {
2239                 global $wgUser;
2240
2241                 if ( is_bool( $usePrefs ) ) {
2242                         if ( $usePrefs ) {
2243                                 $datePreference = $wgUser->getDatePreference();
2244                         } else {
2245                                 $datePreference = (string)User::getDefaultOption( 'date' );
2246                         }
2247                 } else {
2248                         $datePreference = (string)$usePrefs;
2249                 }
2250
2251                 // return int
2252                 if ( $datePreference == '' ) {
2253                         return 'default';
2254                 }
2255
2256                 return $datePreference;
2257         }
2258
2259         /**
2260          * Get a format string for a given type and preference
2261          * @param string $type May be 'date', 'time', 'both', or 'pretty'.
2262          * @param string $pref The format name as it appears in Messages*.php under
2263          *  $datePreferences.
2264          *
2265          * @since 1.22 New type 'pretty' that provides a more readable timestamp format
2266          *
2267          * @return string
2268          */
2269         function getDateFormatString( $type, $pref ) {
2270                 $wasDefault = false;
2271                 if ( $pref == 'default' ) {
2272                         $wasDefault = true;
2273                         $pref = $this->getDefaultDateFormat();
2274                 }
2275
2276                 if ( !isset( $this->dateFormatStrings[$type][$pref] ) ) {
2277                         $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2278
2279                         if ( $type === 'pretty' && $df === null ) {
2280                                 $df = $this->getDateFormatString( 'date', $pref );
2281                         }
2282
2283                         if ( !$wasDefault && $df === null ) {
2284                                 $pref = $this->getDefaultDateFormat();
2285                                 $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2286                         }
2287
2288                         $this->dateFormatStrings[$type][$pref] = $df;
2289                 }
2290                 return $this->dateFormatStrings[$type][$pref];
2291         }
2292
2293         /**
2294          * @param string $ts The time format which needs to be turned into a
2295          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2296          * @param bool $adj Whether to adjust the time output according to the
2297          *   user configured offset ($timecorrection)
2298          * @param mixed $format True to use user's date format preference
2299          * @param string|bool $timecorrection The time offset as returned by
2300          *   validateTimeZone() in Special:Preferences
2301          * @return string
2302          */
2303         public function date( $ts, $adj = false, $format = true, $timecorrection = false ) {
2304                 $ts = wfTimestamp( TS_MW, $ts );
2305                 if ( $adj ) {
2306                         $ts = $this->userAdjust( $ts, $timecorrection );
2307                 }
2308                 $df = $this->getDateFormatString( 'date', $this->dateFormat( $format ) );
2309                 return $this->sprintfDate( $df, $ts );
2310         }
2311
2312         /**
2313          * @param string $ts The time format which needs to be turned into a
2314          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2315          * @param bool $adj Whether to adjust the time output according to the
2316          *   user configured offset ($timecorrection)
2317          * @param mixed $format True to use user's date format preference
2318          * @param string|bool $timecorrection The time offset as returned by
2319          *   validateTimeZone() in Special:Preferences
2320          * @return string
2321          */
2322         public function time( $ts, $adj = false, $format = true, $timecorrection = false ) {
2323                 $ts = wfTimestamp( TS_MW, $ts );
2324                 if ( $adj ) {
2325                         $ts = $this->userAdjust( $ts, $timecorrection );
2326                 }
2327                 $df = $this->getDateFormatString( 'time', $this->dateFormat( $format ) );
2328                 return $this->sprintfDate( $df, $ts );
2329         }
2330
2331         /**
2332          * @param string $ts The time format which needs to be turned into a
2333          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2334          * @param bool $adj Whether to adjust the time output according to the
2335          *   user configured offset ($timecorrection)
2336          * @param mixed $format What format to return, if it's false output the
2337          *   default one (default true)
2338          * @param string|bool $timecorrection The time offset as returned by
2339          *   validateTimeZone() in Special:Preferences
2340          * @return string
2341          */
2342         public function timeanddate( $ts, $adj = false, $format = true, $timecorrection = false ) {
2343                 $ts = wfTimestamp( TS_MW, $ts );
2344                 if ( $adj ) {
2345                         $ts = $this->userAdjust( $ts, $timecorrection );
2346                 }
2347                 $df = $this->getDateFormatString( 'both', $this->dateFormat( $format ) );
2348                 return $this->sprintfDate( $df, $ts );
2349         }
2350
2351         /**
2352          * Takes a number of seconds and turns it into a text using values such as hours and minutes.
2353          *
2354          * @since 1.20
2355          *
2356          * @param int $seconds The amount of seconds.
2357          * @param array $chosenIntervals The intervals to enable.
2358          *
2359          * @return string
2360          */
2361         public function formatDuration( $seconds, array $chosenIntervals = [] ) {
2362                 $intervals = $this->getDurationIntervals( $seconds, $chosenIntervals );
2363
2364                 $segments = [];
2365
2366                 foreach ( $intervals as $intervalName => $intervalValue ) {
2367                         // Messages: duration-seconds, duration-minutes, duration-hours, duration-days, duration-weeks,
2368                         // duration-years, duration-decades, duration-centuries, duration-millennia
2369                         $message = wfMessage( 'duration-' . $intervalName )->numParams( $intervalValue );
2370                         $segments[] = $message->inLanguage( $this )->escaped();
2371                 }
2372
2373                 return $this->listToText( $segments );
2374         }
2375
2376         /**
2377          * Takes a number of seconds and returns an array with a set of corresponding intervals.
2378          * For example 65 will be turned into [ minutes => 1, seconds => 5 ].
2379          *
2380          * @since 1.20
2381          *
2382          * @param int $seconds The amount of seconds.
2383          * @param array $chosenIntervals The intervals to enable.
2384          *
2385          * @return array
2386          */
2387         public function getDurationIntervals( $seconds, array $chosenIntervals = [] ) {
2388                 if ( empty( $chosenIntervals ) ) {
2389                         $chosenIntervals = [
2390                                 'millennia',
2391                                 'centuries',
2392                                 'decades',
2393                                 'years',
2394                                 'days',
2395                                 'hours',
2396                                 'minutes',
2397                                 'seconds'
2398                         ];
2399                 }
2400
2401                 $intervals = array_intersect_key( self::$durationIntervals, array_flip( $chosenIntervals ) );
2402                 $sortedNames = array_keys( $intervals );
2403                 $smallestInterval = array_pop( $sortedNames );
2404
2405                 $segments = [];
2406
2407                 foreach ( $intervals as $name => $length ) {
2408                         $value = floor( $seconds / $length );
2409
2410                         if ( $value > 0 || ( $name == $smallestInterval && empty( $segments ) ) ) {
2411                                 $seconds -= $value * $length;
2412                                 $segments[$name] = $value;
2413                         }
2414                 }
2415
2416                 return $segments;
2417         }
2418
2419         /**
2420          * Internal helper function for userDate(), userTime() and userTimeAndDate()
2421          *
2422          * @param string $type Can be 'date', 'time' or 'both'
2423          * @param string $ts The time format which needs to be turned into a
2424          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2425          * @param User $user User object used to get preferences for timezone and format
2426          * @param array $options Array, can contain the following keys:
2427          *   - 'timecorrection': time correction, can have the following values:
2428          *     - true: use user's preference
2429          *     - false: don't use time correction
2430          *     - int: value of time correction in minutes
2431          *   - 'format': format to use, can have the following values:
2432          *     - true: use user's preference
2433          *     - false: use default preference
2434          *     - string: format to use
2435          * @since 1.19
2436          * @return string
2437          */
2438         private function internalUserTimeAndDate( $type, $ts, User $user, array $options ) {
2439                 $ts = wfTimestamp( TS_MW, $ts );
2440                 $options += [ 'timecorrection' => true, 'format' => true ];
2441                 if ( $options['timecorrection'] !== false ) {
2442                         if ( $options['timecorrection'] === true ) {
2443                                 $offset = $user->getOption( 'timecorrection' );
2444                         } else {
2445                                 $offset = $options['timecorrection'];
2446                         }
2447                         $ts = $this->userAdjust( $ts, $offset );
2448                 }
2449                 if ( $options['format'] === true ) {
2450                         $format = $user->getDatePreference();
2451                 } else {
2452                         $format = $options['format'];
2453                 }
2454                 $df = $this->getDateFormatString( $type, $this->dateFormat( $format ) );
2455                 return $this->sprintfDate( $df, $ts );
2456         }
2457
2458         /**
2459          * Get the formatted date for the given timestamp and formatted for
2460          * the given user.
2461          *
2462          * @param mixed $ts Mixed: the time format which needs to be turned into a
2463          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2464          * @param User $user User object used to get preferences for timezone and format
2465          * @param array $options Array, can contain the following keys:
2466          *   - 'timecorrection': time correction, can have the following values:
2467          *     - true: use user's preference
2468          *     - false: don't use time correction
2469          *     - int: value of time correction in minutes
2470          *   - 'format': format to use, can have the following values:
2471          *     - true: use user's preference
2472          *     - false: use default preference
2473          *     - string: format to use
2474          * @since 1.19
2475          * @return string
2476          */
2477         public function userDate( $ts, User $user, array $options = [] ) {
2478                 return $this->internalUserTimeAndDate( 'date', $ts, $user, $options );
2479         }
2480
2481         /**
2482          * Get the formatted time for the given timestamp and formatted for
2483          * the given user.
2484          *
2485          * @param mixed $ts The time format which needs to be turned into a
2486          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2487          * @param User $user User object used to get preferences for timezone and format
2488          * @param array $options Array, can contain the following keys:
2489          *   - 'timecorrection': time correction, can have the following values:
2490          *     - true: use user's preference
2491          *     - false: don't use time correction
2492          *     - int: value of time correction in minutes
2493          *   - 'format': format to use, can have the following values:
2494          *     - true: use user's preference
2495          *     - false: use default preference
2496          *     - string: format to use
2497          * @since 1.19
2498          * @return string
2499          */
2500         public function userTime( $ts, User $user, array $options = [] ) {
2501                 return $this->internalUserTimeAndDate( 'time', $ts, $user, $options );
2502         }
2503
2504         /**
2505          * Get the formatted date and time for the given timestamp and formatted for
2506          * the given user.
2507          *
2508          * @param mixed $ts The time format which needs to be turned into a
2509          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2510          * @param User $user User object used to get preferences for timezone and format
2511          * @param array $options Array, can contain the following keys:
2512          *   - 'timecorrection': time correction, can have the following values:
2513          *     - true: use user's preference
2514          *     - false: don't use time correction
2515          *     - int: value of time correction in minutes
2516          *   - 'format': format to use, can have the following values:
2517          *     - true: use user's preference
2518          *     - false: use default preference
2519          *     - string: format to use
2520          * @since 1.19
2521          * @return string
2522          */
2523         public function userTimeAndDate( $ts, User $user, array $options = [] ) {
2524                 return $this->internalUserTimeAndDate( 'both', $ts, $user, $options );
2525         }
2526
2527         /**
2528          * Get the timestamp in a human-friendly relative format, e.g., "3 days ago".
2529          *
2530          * Determine the difference between the timestamp and the current time, and
2531          * generate a readable timestamp by returning "<N> <units> ago", where the
2532          * largest possible unit is used.
2533          *
2534          * @since 1.26 (Prior to 1.26 method existed but was not meant to be used directly)
2535          *
2536          * @param MWTimestamp $time
2537          * @param MWTimestamp|null $relativeTo The base timestamp to compare to (defaults to now)
2538          * @param User|null $user User the timestamp is being generated for
2539          *  (or null to use main context's user)
2540          * @return string Formatted timestamp
2541          */
2542         public function getHumanTimestamp(
2543                 MWTimestamp $time, MWTimestamp $relativeTo = null, User $user = null
2544         ) {
2545                 if ( $relativeTo === null ) {
2546                         $relativeTo = new MWTimestamp();
2547                 }
2548                 if ( $user === null ) {
2549                         $user = RequestContext::getMain()->getUser();
2550                 }
2551
2552                 // Adjust for the user's timezone.
2553                 $offsetThis = $time->offsetForUser( $user );
2554                 $offsetRel = $relativeTo->offsetForUser( $user );
2555
2556                 $ts = '';
2557                 if ( Hooks::run( 'GetHumanTimestamp', [ &$ts, $time, $relativeTo, $user, $this ] ) ) {
2558                         $ts = $this->getHumanTimestampInternal( $time, $relativeTo, $user );
2559                 }
2560
2561                 // Reset the timezone on the objects.
2562                 $time->timestamp->sub( $offsetThis );
2563                 $relativeTo->timestamp->sub( $offsetRel );
2564
2565                 return $ts;
2566         }
2567
2568         /**
2569          * Convert an MWTimestamp into a pretty human-readable timestamp using
2570          * the given user preferences and relative base time.
2571          *
2572          * @see Language::getHumanTimestamp
2573          * @param MWTimestamp $ts Timestamp to prettify
2574          * @param MWTimestamp $relativeTo Base timestamp
2575          * @param User $user User preferences to use
2576          * @return string Human timestamp
2577          * @since 1.26
2578          */
2579         private function getHumanTimestampInternal(
2580                 MWTimestamp $ts, MWTimestamp $relativeTo, User $user
2581         ) {
2582                 $diff = $ts->diff( $relativeTo );
2583                 $diffDay = (bool)( (int)$ts->timestamp->format( 'w' ) -
2584                         (int)$relativeTo->timestamp->format( 'w' ) );
2585                 $days = $diff->days ?: (int)$diffDay;
2586                 if ( $diff->invert || $days > 5
2587                         && $ts->timestamp->format( 'Y' ) !== $relativeTo->timestamp->format( 'Y' )
2588                 ) {
2589                         // Timestamps are in different years: use full timestamp
2590                         // Also do full timestamp for future dates
2591                         /**
2592                          * @todo FIXME: Add better handling of future timestamps.
2593                          */
2594                         $format = $this->getDateFormatString( 'both', $user->getDatePreference() ?: 'default' );
2595                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2596                 } elseif ( $days > 5 ) {
2597                         // Timestamps are in same year,  but more than 5 days ago: show day and month only.
2598                         $format = $this->getDateFormatString( 'pretty', $user->getDatePreference() ?: 'default' );
2599                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2600                 } elseif ( $days > 1 ) {
2601                         // Timestamp within the past week: show the day of the week and time
2602                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2603                         $weekday = self::$mWeekdayMsgs[$ts->timestamp->format( 'w' )];
2604                         // Messages:
2605                         // sunday-at, monday-at, tuesday-at, wednesday-at, thursday-at, friday-at, saturday-at
2606                         $ts = wfMessage( "$weekday-at" )
2607                                 ->inLanguage( $this )
2608                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2609                                 ->text();
2610                 } elseif ( $days == 1 ) {
2611                         // Timestamp was yesterday: say 'yesterday' and the time.
2612                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2613                         $ts = wfMessage( 'yesterday-at' )
2614                                 ->inLanguage( $this )
2615                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2616                                 ->text();
2617                 } elseif ( $diff->h > 1 || $diff->h == 1 && $diff->i > 30 ) {
2618                         // Timestamp was today, but more than 90 minutes ago: say 'today' and the time.
2619                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2620                         $ts = wfMessage( 'today-at' )
2621                                 ->inLanguage( $this )
2622                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2623                                 ->text();
2624
2625                 // From here on in, the timestamp was soon enough ago so that we can simply say
2626                 // XX units ago, e.g., "2 hours ago" or "5 minutes ago"
2627                 } elseif ( $diff->h == 1 ) {
2628                         // Less than 90 minutes, but more than an hour ago.
2629                         $ts = wfMessage( 'hours-ago' )->inLanguage( $this )->numParams( 1 )->text();
2630                 } elseif ( $diff->i >= 1 ) {
2631                         // A few minutes ago.
2632                         $ts = wfMessage( 'minutes-ago' )->inLanguage( $this )->numParams( $diff->i )->text();
2633                 } elseif ( $diff->s >= 30 ) {
2634                         // Less than a minute, but more than 30 sec ago.
2635                         $ts = wfMessage( 'seconds-ago' )->inLanguage( $this )->numParams( $diff->s )->text();
2636                 } else {
2637                         // Less than 30 seconds ago.
2638                         $ts = wfMessage( 'just-now' )->text();
2639                 }
2640
2641                 return $ts;
2642         }
2643
2644         /**
2645          * @param string $key
2646          * @return string|null
2647          */
2648         public function getMessage( $key ) {
2649                 return self::$dataCache->getSubitem( $this->mCode, 'messages', $key );
2650         }
2651
2652         /**
2653          * @return array
2654          */
2655         function getAllMessages() {
2656                 return self::$dataCache->getItem( $this->mCode, 'messages' );
2657         }
2658
2659         /**
2660          * @param string $in
2661          * @param string $out
2662          * @param string $string
2663          * @return string
2664          */
2665         public function iconv( $in, $out, $string ) {
2666                 # Even with //IGNORE iconv can whine about illegal characters in
2667                 # *input* string. We just ignore those too.
2668                 # REF: https://bugs.php.net/bug.php?id=37166
2669                 # REF: https://phabricator.wikimedia.org/T18885
2670                 Wikimedia\suppressWarnings();
2671                 $text = iconv( $in, $out . '//IGNORE', $string );
2672                 Wikimedia\restoreWarnings();
2673                 return $text;
2674         }
2675
2676         // callback functions for ucwords(), ucwordbreaks()
2677
2678         /**
2679          * @param array $matches
2680          * @return mixed|string
2681          */
2682         function ucwordbreaksCallbackAscii( $matches ) {
2683                 return $this->ucfirst( $matches[1] );
2684         }
2685
2686         /**
2687          * @param array $matches
2688          * @return string
2689          */
2690         function ucwordbreaksCallbackMB( $matches ) {
2691                 return mb_strtoupper( $matches[0] );
2692         }
2693
2694         /**
2695          * @param array $matches
2696          * @return string
2697          */
2698         function ucwordsCallbackMB( $matches ) {
2699                 return mb_strtoupper( $matches[0] );
2700         }
2701
2702         /**
2703          * Make a string's first character uppercase
2704          *
2705          * @param string $str
2706          *
2707          * @return string
2708          */
2709         public function ucfirst( $str ) {
2710                 $o = ord( $str );
2711                 if ( $o < 96 ) { // if already uppercase...
2712                         return $str;
2713                 } elseif ( $o < 128 ) {
2714                         return ucfirst( $str ); // use PHP's ucfirst()
2715                 } else {
2716                         // fall back to more complex logic in case of multibyte strings
2717                         return $this->uc( $str, true );
2718                 }
2719         }
2720
2721         /**
2722          * Convert a string to uppercase
2723          *
2724          * @param string $str
2725          * @param bool $first
2726          *
2727          * @return string
2728          */
2729         public function uc( $str, $first = false ) {
2730                 if ( $first ) {
2731                         if ( $this->isMultibyte( $str ) ) {
2732                                 return $this->mbUpperChar( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2733                         } else {
2734                                 return ucfirst( $str );
2735                         }
2736                 } else {
2737                         return $this->isMultibyte( $str ) ? mb_strtoupper( $str ) : strtoupper( $str );
2738                 }
2739         }
2740
2741         /**
2742          * Convert character to uppercase, allowing overrides of the default mb_upper
2743          * behaviour, which is buggy in many ways. Having a conversion table can be
2744          * useful during transitions between PHP versions where unicode changes happen.
2745          * This can make some resources unreachable on-wiki, see discussion at T219279.
2746          * Providing such a conversion table can allow to manage the transition period.
2747          *
2748          * @since 1.34
2749          *
2750          * @param string $char
2751          *
2752          * @return string
2753          */
2754         protected function mbUpperChar( $char ) {
2755                 global $wgOverrideUcfirstCharacters;
2756                 if ( array_key_exists( $char, $wgOverrideUcfirstCharacters ) ) {
2757                         return $wgOverrideUcfirstCharacters[$char];
2758                 } else {
2759                         return mb_strtoupper( $char );
2760                 }
2761         }
2762
2763         /**
2764          * @param string $str
2765          * @return mixed|string
2766          */
2767         function lcfirst( $str ) {
2768                 $o = ord( $str );
2769                 if ( !$o ) {
2770                         return strval( $str );
2771                 } elseif ( $o >= 128 ) {
2772                         return $this->lc( $str, true );
2773                 } elseif ( $o > 96 ) {
2774                         return $str;
2775                 } else {
2776                         $str[0] = strtolower( $str[0] );
2777                         return $str;
2778                 }
2779         }
2780
2781         /**
2782          * @param string $str
2783          * @param bool $first
2784          * @return mixed|string
2785          */
2786         function lc( $str, $first = false ) {
2787                 if ( $first ) {
2788                         if ( $this->isMultibyte( $str ) ) {
2789                                 return mb_strtolower( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2790                         } else {
2791                                 return strtolower( substr( $str, 0, 1 ) ) . substr( $str, 1 );
2792                         }
2793                 } else {
2794                         return $this->isMultibyte( $str ) ? mb_strtolower( $str ) : strtolower( $str );
2795                 }
2796         }
2797
2798         /**
2799          * @param string $str
2800          * @return bool
2801          */
2802         function isMultibyte( $str ) {
2803                 return strlen( $str ) !== mb_strlen( $str );
2804         }
2805
2806         /**
2807          * @param string $str
2808          * @return mixed|string
2809          */
2810         function ucwords( $str ) {
2811                 if ( $this->isMultibyte( $str ) ) {
2812                         $str = $this->lc( $str );
2813
2814                         // regexp to find first letter in each word (i.e. after each space)
2815                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)| ([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2816
2817                         // function to use to capitalize a single char
2818                         return preg_replace_callback(
2819                                 $replaceRegexp,
2820                                 [ $this, 'ucwordsCallbackMB' ],
2821                                 $str
2822                         );
2823                 } else {
2824                         return ucwords( strtolower( $str ) );
2825                 }
2826         }
2827
2828         /**
2829          * capitalize words at word breaks
2830          *
2831          * @param string $str
2832          * @return mixed
2833          */
2834         function ucwordbreaks( $str ) {
2835                 if ( $this->isMultibyte( $str ) ) {
2836                         $str = $this->lc( $str );
2837
2838                         // since \b doesn't work for UTF-8, we explicitely define word break chars
2839                         $breaks = "[ \-\(\)\}\{\.,\?!]";
2840
2841                         // find first letter after word break
2842                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)|" .
2843                                 "$breaks([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2844
2845                         return preg_replace_callback(
2846                                 $replaceRegexp,
2847                                 [ $this, 'ucwordbreaksCallbackMB' ],
2848                                 $str
2849                         );
2850                 } else {
2851                         return preg_replace_callback(
2852                                 '/\b([\w\x80-\xff]+)\b/',
2853                                 [ $this, 'ucwordbreaksCallbackAscii' ],
2854                                 $str
2855                         );
2856                 }
2857         }
2858
2859         /**
2860          * Return a case-folded representation of $s
2861          *
2862          * This is a representation such that caseFold($s1)==caseFold($s2) if $s1
2863          * and $s2 are the same except for the case of their characters. It is not
2864          * necessary for the value returned to make sense when displayed.
2865          *
2866          * Do *not* perform any other normalisation in this function. If a caller
2867          * uses this function when it should be using a more general normalisation
2868          * function, then fix the caller.
2869          *
2870          * @param string $s
2871          *
2872          * @return string
2873          */
2874         function caseFold( $s ) {
2875                 return $this->uc( $s );
2876         }
2877
2878         /**
2879          * TODO: $s is not always a string per T218883
2880          * @param string $s
2881          * @return string
2882          */
2883         function checkTitleEncoding( $s ) {
2884                 if ( is_array( $s ) ) {
2885                         throw new MWException( 'Given array to checkTitleEncoding.' );
2886                 }
2887                 if ( StringUtils::isUtf8( $s ) ) {
2888                         return $s;
2889                 }
2890
2891                 return $this->iconv( $this->fallback8bitEncoding(), 'utf-8', $s );
2892         }
2893
2894         /**
2895          * @return string
2896          */
2897         function fallback8bitEncoding() {
2898                 return self::$dataCache->getItem( $this->mCode, 'fallback8bitEncoding' );
2899         }
2900
2901         /**
2902          * Most writing systems use whitespace to break up words.
2903          * Some languages such as Chinese don't conventionally do this,
2904          * which requires special handling when breaking up words for
2905          * searching etc.
2906          *
2907          * @return bool
2908          */
2909         function hasWordBreaks() {
2910                 return true;
2911         }
2912
2913         /**
2914          * Some languages such as Chinese require word segmentation,
2915          * Specify such segmentation when overridden in derived class.
2916          *
2917          * @param string $string
2918          * @return string
2919          */
2920         function segmentByWord( $string ) {
2921                 return $string;
2922         }
2923
2924         /**
2925          * Some languages have special punctuation need to be normalized.
2926          * Make such changes here.
2927          *
2928          * @param string $string
2929          * @return string
2930          */
2931         function normalizeForSearch( $string ) {
2932                 return self::convertDoubleWidth( $string );
2933         }
2934
2935         /**
2936          * convert double-width roman characters to single-width.
2937          * range: ff00-ff5f ~= 0020-007f
2938          *
2939          * @param string $string
2940          *
2941          * @return string
2942          */
2943         protected static function convertDoubleWidth( $string ) {
2944                 static $full = null;
2945                 static $half = null;
2946
2947                 if ( $full === null ) {
2948                         $fullWidth = "０１２３４５６７８９ＡＢＣＤＥＦＧＨＩＪＫＬＭＮＯＰＱＲＳＴＵＶＷＸＹＺａｂｃｄｅｆｇｈｉｊｋｌｍｎｏｐｑｒｓｔｕｖｗｘｙｚ";
2949                         $halfWidth = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz";
2950                         $full = str_split( $fullWidth, 3 );
2951                         $half = str_split( $halfWidth );
2952                 }
2953
2954                 $string = str_replace( $full, $half, $string );
2955                 return $string;
2956         }
2957
2958         /**
2959          * @param string $string
2960          * @param string $pattern
2961          * @return string
2962          */
2963         protected static function insertSpace( $string, $pattern ) {
2964                 $string = preg_replace( $pattern, " $1 ", $string );
2965                 $string = preg_replace( '/ +/', ' ', $string );
2966                 return $string;
2967         }
2968
2969         /**
2970          * @param string[] $termsArray
2971          * @return string[]
2972          */
2973         function convertForSearchResult( $termsArray ) {
2974                 # some languages, e.g. Chinese, need to do a conversion
2975                 # in order for search results to be displayed correctly
2976                 return $termsArray;
2977         }
2978
2979         /**
2980          * Get the first character of a string.
2981          *
2982          * @param string $s
2983          * @return string
2984          */
2985         function firstChar( $s ) {
2986                 $matches = [];
2987                 preg_match(
2988                         '/^([\x00-\x7f]|[\xc0-\xdf][\x80-\xbf]|' .
2989                                 '[\xe0-\xef][\x80-\xbf]{2}|[\xf0-\xf7][\x80-\xbf]{3})/',
2990                         $s,
2991                         $matches
2992                 );
2993
2994                 if ( isset( $matches[1] ) ) {
2995                         if ( strlen( $matches[1] ) != 3 ) {
2996                                 return $matches[1];
2997                         }
2998
2999                         // Break down Hangul syllables to grab the first jamo
3000                         $code = UtfNormal\Utils::utf8ToCodepoint( $matches[1] );
3001                         if ( $code < 0xac00 || 0xd7a4 <= $code ) {
3002                                 return $matches[1];
3003                         } elseif ( $code < 0xb098 ) {
3004                                 return "\u{3131}";
3005                         } elseif ( $code < 0xb2e4 ) {
3006                                 return "\u{3134}";
3007                         } elseif ( $code < 0xb77c ) {
3008                                 return "\u{3137}";
3009                         } elseif ( $code < 0xb9c8 ) {
3010                                 return "\u{3139}";
3011                         } elseif ( $code < 0xbc14 ) {
3012                                 return "\u{3141}";
3013                         } elseif ( $code < 0xc0ac ) {
3014                                 return "\u{3142}";
3015                         } elseif ( $code < 0xc544 ) {
3016                                 return "\u{3145}";
3017                         } elseif ( $code < 0xc790 ) {
3018                                 return "\u{3147}";
3019                         } elseif ( $code < 0xcc28 ) {
3020                                 return "\u{3148}";
3021                         } elseif ( $code < 0xce74 ) {
3022                                 return "\u{314A}";
3023                         } elseif ( $code < 0xd0c0 ) {
3024                                 return "\u{314B}";
3025                         } elseif ( $code < 0xd30c ) {
3026                                 return "\u{314C}";
3027                         } elseif ( $code < 0xd558 ) {
3028                                 return "\u{314D}";
3029                         } else {
3030                                 return "\u{314E}";
3031                         }
3032                 } else {
3033                         return '';
3034                 }
3035         }
3036
3037         /**
3038          * Convert a UTF-8 string to normal form C. In Malayalam and Arabic, this
3039          * also cleans up certain backwards-compatible sequences, converting them
3040          * to the modern Unicode equivalent.
3041          *
3042          * This is language-specific for performance reasons only.
3043          *
3044          * @param string $s
3045          *
3046          * @return string
3047          */
3048         public function normalize( $s ) {
3049                 global $wgAllUnicodeFixes;
3050                 $s = UtfNormal\Validator::cleanUp( $s );
3051                 if ( $wgAllUnicodeFixes ) {
3052                         $s = $this->transformUsingPairFile( 'normalize-ar.php', $s );
3053                         $s = $this->transformUsingPairFile( 'normalize-ml.php', $s );
3054                 }
3055
3056                 return $s;
3057         }
3058
3059         /**
3060          * Transform a string using serialized data stored in the given file (which
3061          * must be in the serialized subdirectory of $IP). The file contains pairs
3062          * mapping source characters to destination characters.
3063          *
3064          * The data is cached in process memory. This will go faster if you have the
3065          * FastStringSearch extension.
3066          *
3067          * @param string $file
3068          * @param string $string
3069          *
3070          * @throws MWException
3071          * @return string
3072          */
3073         protected function transformUsingPairFile( $file, $string ) {
3074                 if ( !isset( $this->transformData[$file] ) ) {
3075                         global $IP;
3076                         $data = require "$IP/languages/data/{$file}";
3077                         $this->transformData[$file] = new ReplacementArray( $data );
3078                 }
3079                 return $this->transformData[$file]->replace( $string );
3080         }
3081
3082         /**
3083          * For right-to-left language support
3084          *
3085          * @return bool
3086          */
3087         function isRTL() {
3088                 return self::$dataCache->getItem( $this->mCode, 'rtl' );
3089         }
3090
3091         /**
3092          * Return the correct HTML 'dir' attribute value for this language.
3093          * @return string
3094          */
3095         function getDir() {
3096                 return $this->isRTL() ? 'rtl' : 'ltr';
3097         }
3098
3099         /**
3100          * Return 'left' or 'right' as appropriate alignment for line-start
3101          * for this language's text direction.
3102          *
3103          * Should be equivalent to CSS3 'start' text-align value....
3104          *
3105          * @return string
3106          */
3107         function alignStart() {
3108                 return $this->isRTL() ? 'right' : 'left';
3109         }
3110
3111         /**
3112          * Return 'right' or 'left' as appropriate alignment for line-end
3113          * for this language's text direction.
3114          *
3115          * Should be equivalent to CSS3 'end' text-align value....
3116          *
3117          * @return string
3118          */
3119         function alignEnd() {
3120                 return $this->isRTL() ? 'left' : 'right';
3121         }
3122
3123         /**
3124          * A hidden direction mark (LRM or RLM), depending on the language direction.
3125          * Unlike getDirMark(), this function returns the character as an HTML entity.
3126          * This function should be used when the output is guaranteed to be HTML,
3127          * because it makes the output HTML source code more readable. When
3128          * the output is plain text or can be escaped, getDirMark() should be used.
3129          *
3130          * @param bool $opposite Get the direction mark opposite to your language
3131          * @return string
3132          * @since 1.20
3133          */
3134         function getDirMarkEntity( $opposite = false ) {
3135                 if ( $opposite ) {
3136                         return $this->isRTL() ? '&lrm;' : '&rlm;';
3137                 }
3138                 return $this->isRTL() ? '&rlm;' : '&lrm;';
3139         }
3140
3141         /**
3142          * A hidden direction mark (LRM or RLM), depending on the language direction.
3143          * This function produces them as invisible Unicode characters and
3144          * the output may be hard to read and debug, so it should only be used
3145          * when the output is plain text or can be escaped. When the output is
3146          * HTML, use getDirMarkEntity() instead.
3147          *
3148          * @param bool $opposite Get the direction mark opposite to your language
3149          * @return string
3150          */
3151         function getDirMark( $opposite = false ) {
3152                 $lrm = "\u{200E}"; # LEFT-TO-RIGHT MARK, commonly abbreviated LRM
3153                 $rlm = "\u{200F}"; # RIGHT-TO-LEFT MARK, commonly abbreviated RLM
3154                 if ( $opposite ) {
3155                         return $this->isRTL() ? $lrm : $rlm;
3156                 }
3157                 return $this->isRTL() ? $rlm : $lrm;
3158         }
3159
3160         /**
3161          * @return array
3162          */
3163         function capitalizeAllNouns() {
3164                 return self::$dataCache->getItem( $this->mCode, 'capitalizeAllNouns' );
3165         }
3166
3167         /**
3168          * An arrow, depending on the language direction.
3169          *
3170          * @param string $direction The direction of the arrow: forwards (default),
3171          *   backwards, left, right, up, down.
3172          * @return string
3173          */
3174         function getArrow( $direction = 'forwards' ) {
3175                 switch ( $direction ) {
3176                         case 'forwards':
3177                                 return $this->isRTL() ? '←' : '→';
3178                         case 'backwards':
3179                                 return $this->isRTL() ? '→' : '←';
3180                         case 'left':
3181                                 return '←';
3182                         case 'right':
3183                                 return '→';
3184                         case 'up':
3185                                 return '↑';
3186                         case 'down':
3187                                 return '↓';
3188                 }
3189         }
3190
3191         /**
3192          * To allow "foo[[bar]]" to extend the link over the whole word "foobar"
3193          *
3194          * @return bool
3195          */
3196         function linkPrefixExtension() {
3197                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixExtension' );
3198         }
3199
3200         /**
3201          * Get all magic words from cache.
3202          * @return array
3203          */
3204         function getMagicWords() {
3205                 return self::$dataCache->getItem( $this->mCode, 'magicWords' );
3206         }
3207
3208         /**
3209          * Fill a MagicWord object with data from here
3210          *
3211          * @param MagicWord $mw
3212          */
3213         function getMagic( $mw ) {
3214                 $rawEntry = $this->mMagicExtensions[$mw->mId] ??
3215                         self::$dataCache->getSubitem( $this->mCode, 'magicWords', $mw->mId );
3216
3217                 if ( !is_array( $rawEntry ) ) {
3218                         wfWarn( "\"$rawEntry\" is not a valid magic word for \"$mw->mId\"" );
3219                 } else {
3220                         $mw->mCaseSensitive = $rawEntry[0];
3221                         $mw->mSynonyms = array_slice( $rawEntry, 1 );
3222                 }
3223         }
3224
3225         /**
3226          * Add magic words to the extension array
3227          *
3228          * @param array $newWords
3229          */
3230         function addMagicWordsByLang( $newWords ) {
3231                 $fallbackChain = $this->getFallbackLanguages();
3232                 $fallbackChain = array_reverse( $fallbackChain );
3233                 foreach ( $fallbackChain as $code ) {
3234                         if ( isset( $newWords[$code] ) ) {
3235                                 $this->mMagicExtensions = $newWords[$code] + $this->mMagicExtensions;
3236                         }
3237                 }
3238         }
3239
3240         /**
3241          * Get special page names, as an associative array
3242          *   canonical name => array of valid names, including aliases
3243          * @return array
3244          */
3245         function getSpecialPageAliases() {
3246                 // Cache aliases because it may be slow to load them
3247                 if ( is_null( $this->mExtendedSpecialPageAliases ) ) {
3248                         // Initialise array
3249                         $this->mExtendedSpecialPageAliases =
3250                                 self::$dataCache->getItem( $this->mCode, 'specialPageAliases' );
3251                 }
3252
3253                 return $this->mExtendedSpecialPageAliases;
3254         }
3255
3256         /**
3257          * Italic is unsuitable for some languages
3258          *
3259          * @param string $text The text to be emphasized.
3260          * @return string
3261          */
3262         function emphasize( $text ) {
3263                 return "<em>$text</em>";
3264         }
3265
3266         /**
3267          * Normally we output all numbers in plain en_US style, that is
3268          * 293,291.235 for twohundredninetythreethousand-twohundredninetyone
3269          * point twohundredthirtyfive. However this is not suitable for all
3270          * languages, some such as Bengali (bn) want ২,৯৩,২৯১.২৩৫ and others such as
3271          * Icelandic just want to use commas instead of dots, and dots instead
3272          * of commas like "293.291,235".
3273          *
3274          * An example of this function being called:
3275          * <code>
3276          * wfMessage( 'message' )->numParams( $num )->text()
3277          * </code>
3278          *
3279          * See $separatorTransformTable on MessageIs.php for
3280          * the , => . and . => , implementation.
3281          *
3282          * @todo check if it's viable to use localeconv() for the decimal separator thing.
3283          * @param int|float $number The string to be formatted, should be an integer
3284          *   or a floating point number.
3285          * @param bool $nocommafy Set to true for special numbers like dates
3286          * @return string
3287          */
3288         public function formatNum( $number, $nocommafy = false ) {
3289                 global $wgTranslateNumerals;
3290                 if ( !$nocommafy ) {
3291                         $number = $this->commafy( $number );
3292                         $s = $this->separatorTransformTable();
3293                         if ( $s ) {
3294                                 $number = strtr( $number, $s );
3295                         }
3296                 }
3297
3298                 if ( $wgTranslateNumerals ) {
3299                         $s = $this->digitTransformTable();
3300                         if ( $s ) {
3301                                 $number = strtr( $number, $s );
3302                         }
3303                 }
3304
3305                 return (string)$number;
3306         }
3307
3308         /**
3309          * Front-end for non-commafied formatNum
3310          *
3311          * @param int|float $number The string to be formatted, should be an integer
3312          *        or a floating point number.
3313          * @since 1.21
3314          * @return string
3315          */
3316         public function formatNumNoSeparators( $number ) {
3317                 return $this->formatNum( $number, true );
3318         }
3319
3320         /**
3321          * @param string $number
3322          * @return string
3323          */
3324         public function parseFormattedNumber( $number ) {
3325                 $s = $this->digitTransformTable();
3326                 if ( $s ) {
3327                         // eliminate empty array values such as ''. (T66347)
3328                         $s = array_filter( $s );
3329                         $number = strtr( $number, array_flip( $s ) );
3330                 }
3331
3332                 $s = $this->separatorTransformTable();
3333                 if ( $s ) {
3334                         // eliminate empty array values such as ''. (T66347)
3335                         $s = array_filter( $s );
3336                         $number = strtr( $number, array_flip( $s ) );
3337                 }
3338
3339                 $number = strtr( $number, [ ',' => '' ] );
3340                 return $number;
3341         }
3342
3343         /**
3344          * Adds commas to a given number
3345          * @since 1.19
3346          * @param mixed $number
3347          * @return string
3348          */
3349         function commafy( $number ) {
3350                 $digitGroupingPattern = $this->digitGroupingPattern();
3351                 $minimumGroupingDigits = $this->minimumGroupingDigits();
3352                 if ( $number === null ) {
3353                         return '';
3354                 }
3355
3356                 if ( !$digitGroupingPattern || $digitGroupingPattern === "###,###,###" ) {
3357                         // Default grouping is at thousands, use the same for ###,###,### pattern too.
3358                         // In some languages it's conventional not to insert a thousands separator
3359                         // in numbers that are four digits long (1000-9999).
3360                         if ( $minimumGroupingDigits ) {
3361                                 // Number of '#' characters after last comma in the grouping pattern.
3362                                 // The pattern is hardcoded here, but this would vary for different patterns.
3363                                 $primaryGroupingSize = 3;
3364                                 // Maximum length of a number to suppress digit grouping for.
3365                                 $maximumLength = $minimumGroupingDigits + $primaryGroupingSize - 1;
3366                                 if ( preg_match( '/^\-?\d{1,' . $maximumLength . '}(\.\d+)?$/', $number ) ) {
3367                                         return $number;
3368                                 }
3369                         }
3370                         return strrev( (string)preg_replace( '/(\d{3})(?=\d)(?!\d*\.)/', '$1,', strrev( $number ) ) );
3371                 } else {
3372                         // Ref: http://cldr.unicode.org/translation/number-patterns
3373                         $sign = "";
3374                         if ( intval( $number ) < 0 ) {
3375                                 // For negative numbers apply the algorithm like positive number and add sign.
3376                                 $sign = "-";
3377                                 $number = substr( $number, 1 );
3378                         }
3379                         $integerPart = [];
3380                         $decimalPart = [];
3381                         $numMatches = preg_match_all( "/(#+)/", $digitGroupingPattern, $matches );
3382                         preg_match( "/\d+/", $number, $integerPart );
3383                         preg_match( "/\.\d*/", $number, $decimalPart );
3384                         $groupedNumber = ( count( $decimalPart ) > 0 ) ? $decimalPart[0] : "";
3385                         if ( $groupedNumber === $number ) {
3386                                 // the string does not have any number part. Eg: .12345
3387                                 return $sign . $groupedNumber;
3388                         }
3389                         $start = $end = ( $integerPart ) ? strlen( $integerPart[0] ) : 0;
3390                         while ( $start > 0 ) {
3391                                 $match = $matches[0][$numMatches - 1];
3392                                 $matchLen = strlen( $match );
3393                                 $start = $end - $matchLen;
3394                                 if ( $start < 0 ) {
3395                                         $start = 0;
3396                                 }
3397                                 $groupedNumber = substr( $number, $start, $end - $start ) . $groupedNumber;
3398                                 $end = $start;
3399                                 if ( $numMatches > 1 ) {
3400                                         // use the last pattern for the rest of the number
3401                                         $numMatches--;
3402                                 }
3403                                 if ( $start > 0 ) {
3404                                         $groupedNumber = "," . $groupedNumber;
3405                                 }
3406                         }
3407                         return $sign . $groupedNumber;
3408                 }
3409         }
3410
3411         /**
3412          * @return string
3413          */
3414         function digitGroupingPattern() {
3415                 return self::$dataCache->getItem( $this->mCode, 'digitGroupingPattern' );
3416         }
3417
3418         /**
3419          * @return array
3420          */
3421         function digitTransformTable() {
3422                 return self::$dataCache->getItem( $this->mCode, 'digitTransformTable' );
3423         }
3424
3425         /**
3426          * @return array
3427          */
3428         function separatorTransformTable() {
3429                 return self::$dataCache->getItem( $this->mCode, 'separatorTransformTable' );
3430         }
3431
3432         /**
3433          * @return int|null
3434          */
3435         function minimumGroupingDigits() {
3436                 return self::$dataCache->getItem( $this->mCode, 'minimumGroupingDigits' );
3437         }
3438
3439         /**
3440          * Take a list of strings and build a locale-friendly comma-separated
3441          * list, using the local comma-separator message.
3442          * The last two strings are chained with an "and".
3443          *
3444          * @param string[] $list
3445          * @return string
3446          */
3447         public function listToText( array $list ) {
3448                 $itemCount = count( $list );
3449                 if ( $itemCount < 1 ) {
3450                         return '';
3451                 }
3452                 $text = array_pop( $list );
3453                 if ( $itemCount > 1 ) {
3454                         $and = $this->msg( 'and' )->escaped();
3455                         $space = $this->msg( 'word-separator' )->escaped();
3456                         $comma = '';
3457                         if ( $itemCount > 2 ) {
3458                                 $comma = $this->msg( 'comma-separator' )->escaped();
3459                         }
3460                         $text = implode( $comma, $list ) . $and . $space . $text;
3461                 }
3462                 return $text;
3463         }
3464
3465         /**
3466          * Take a list of strings and build a locale-friendly comma-separated
3467          * list, using the local comma-separator message.
3468          * @param string[] $list Array of strings to put in a comma list
3469          * @return string
3470          */
3471         function commaList( array $list ) {
3472                 return implode(
3473                         wfMessage( 'comma-separator' )->inLanguage( $this )->escaped(),
3474                         $list
3475                 );
3476         }
3477
3478         /**
3479          * Take a list of strings and build a locale-friendly semicolon-separated
3480          * list, using the local semicolon-separator message.
3481          * @param string[] $list Array of strings to put in a semicolon list
3482          * @return string
3483          */
3484         function semicolonList( array $list ) {
3485                 return implode(
3486                         wfMessage( 'semicolon-separator' )->inLanguage( $this )->escaped(),
3487                         $list
3488                 );
3489         }
3490
3491         /**
3492          * Same as commaList, but separate it with the pipe instead.
3493          * @param string[] $list Array of strings to put in a pipe list
3494          * @return string
3495          */
3496         function pipeList( array $list ) {
3497                 return implode(
3498                         wfMessage( 'pipe-separator' )->inLanguage( $this )->escaped(),
3499                         $list
3500                 );
3501         }
3502
3503         /**
3504          * Truncate a string to a specified length in bytes, appending an optional
3505          * string (e.g. for ellipsis)
3506          *
3507          * If $length is negative, the string will be truncated from the beginning
3508          *
3509          * @since 1.31
3510          *
3511          * @param string $string String to truncate
3512          * @param int $length Maximum length in bytes
3513          * @param string $ellipsis String to append to the end of truncated text
3514          * @param bool $adjustLength Subtract length of ellipsis from $length
3515          *
3516          * @return string
3517          */
3518         function truncateForDatabase( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3519                 return $this->truncateInternal(
3520                         $string, $length, $ellipsis, $adjustLength, 'strlen', 'substr'
3521                 );
3522         }
3523
3524         /**
3525          * Truncate a string to a specified number of characters, appending an optional
3526          * string (e.g. for ellipsis).
3527          *
3528          * This provides multibyte version of truncateForDatabase() method of this class,
3529          * suitable for truncation based on number of characters, instead of number of bytes.
3530          *
3531          * If $length is negative, the string will be truncated from the beginning.
3532          *
3533          * @since 1.31
3534          *
3535          * @param string $string String to truncate
3536          * @param int $length Maximum number of characters
3537          * @param string $ellipsis String to append to the end of truncated text
3538          * @param bool $adjustLength Subtract length of ellipsis from $length
3539          *
3540          * @return string
3541          */
3542         function truncateForVisual( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3543                 // Passing encoding to mb_strlen and mb_substr is optional.
3544                 // Encoding defaults to mb_internal_encoding(), which is set to UTF-8 in Setup.php, so
3545                 // explicit specification of encoding is skipped.
3546                 // Note: Both multibyte methods are callables invoked in truncateInternal.
3547                 return $this->truncateInternal(
3548                         $string, $length, $ellipsis, $adjustLength, 'mb_strlen', 'mb_substr'
3549                 );
3550         }
3551
3552         /**
3553          * Internal method used for truncation. This method abstracts text truncation into
3554          * one common method, allowing users to provide length measurement function and
3555          * function for finding substring.
3556          *
3557          * For usages, see truncateForDatabase and truncateForVisual.
3558          *
3559          * @param string $string String to truncate
3560          * @param int $length Maximum length of final text
3561          * @param string $ellipsis String to append to the end of truncated text
3562          * @param bool $adjustLength Subtract length of ellipsis from $length
3563          * @param callable $measureLength Callable function used for determining the length of text
3564          * @param callable $getSubstring Callable function used for getting the substrings
3565          *
3566          * @return string
3567          */
3568         private function truncateInternal(
3569                 $string, $length, $ellipsis, $adjustLength, callable $measureLength, callable $getSubstring
3570         ) {
3571                 # Check if there is no need to truncate
3572                 if ( $measureLength( $string ) <= abs( $length ) ) {
3573                         return $string; // no need to truncate
3574                 }
3575
3576                 # Use the localized ellipsis character
3577                 if ( $ellipsis == '...' ) {
3578                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3579                 }
3580                 if ( $length == 0 ) {
3581                         return $ellipsis; // convention
3582                 }
3583
3584                 $stringOriginal = $string;
3585                 # If ellipsis length is >= $length then we can't apply $adjustLength
3586                 if ( $adjustLength && $measureLength( $ellipsis ) >= abs( $length ) ) {
3587                         $string = $ellipsis; // this can be slightly unexpected
3588                 # Otherwise, truncate and add ellipsis...
3589                 } else {
3590                         $ellipsisLength = $adjustLength ? $measureLength( $ellipsis ) : 0;
3591                         if ( $length > 0 ) {
3592                                 $length -= $ellipsisLength;
3593                                 $string = $getSubstring( $string, 0, $length ); // xyz...
3594                                 $string = $this->removeBadCharLast( $string );
3595                                 $string = rtrim( $string ) . $ellipsis;
3596                         } else {
3597                                 $length += $ellipsisLength;
3598                                 $string = $getSubstring( $string, $length ); // ...xyz
3599                                 $string = $this->removeBadCharFirst( $string );
3600                                 $string = $ellipsis . ltrim( $string );
3601                         }
3602                 }
3603
3604                 # Do not truncate if the ellipsis makes the string longer/equal (T24181).
3605                 # This check is *not* redundant if $adjustLength, due to the single case where
3606                 # LEN($ellipsis) > ABS($limit arg); $stringOriginal could be shorter than $string.
3607                 if ( $measureLength( $string ) < $measureLength( $stringOriginal ) ) {
3608                         return $string;
3609                 } else {
3610                         return $stringOriginal;
3611                 }
3612         }
3613
3614         /**
3615          * Remove bytes that represent an incomplete Unicode character
3616          * at the end of string (e.g. bytes of the char are missing)
3617          *
3618          * @param string $string
3619          * @return string
3620          */
3621         protected function removeBadCharLast( $string ) {
3622                 if ( $string != '' ) {
3623                         $char = ord( $string[strlen( $string ) - 1] );
3624                         $m = [];
3625                         if ( $char >= 0xc0 ) {
3626                                 # We got the first byte only of a multibyte char; remove it.
3627                                 $string = substr( $string, 0, -1 );
3628                         } elseif ( $char >= 0x80 &&
3629                                 // Use the /s modifier (PCRE_DOTALL) so (.*) also matches newlines
3630                                 preg_match( '/^(.*)(?:[\xe0-\xef][\x80-\xbf]|' .
3631                                         '[\xf0-\xf7][\x80-\xbf]{1,2})$/s', $string, $m )
3632                         ) {
3633                                 # We chopped in the middle of a character; remove it
3634                                 $string = $m[1];
3635                         }
3636                 }
3637                 return $string;
3638         }
3639
3640         /**
3641          * Remove bytes that represent an incomplete Unicode character
3642          * at the start of string (e.g. bytes of the char are missing)
3643          *
3644          * @param string $string
3645          * @return string
3646          */
3647         protected function removeBadCharFirst( $string ) {
3648                 if ( $string != '' ) {
3649                         $char = ord( $string[0] );
3650                         if ( $char >= 0x80 && $char < 0xc0 ) {
3651                                 # We chopped in the middle of a character; remove the whole thing
3652                                 $string = preg_replace( '/^[\x80-\xbf]+/', '', $string );
3653                         }
3654                 }
3655                 return $string;
3656         }
3657
3658         /**
3659          * Truncate a string of valid HTML to a specified length in bytes,
3660          * appending an optional string (e.g. for ellipses), and return valid HTML
3661          *
3662          * This is only intended for styled/linked text, such as HTML with
3663          * tags like <span> and <a>, were the tags are self-contained (valid HTML).
3664          * Also, this will not detect things like "display:none" CSS.
3665          *
3666          * Note: since 1.18 you do not need to leave extra room in $length for ellipses.
3667          *
3668          * @param string $text HTML string to truncate
3669          * @param int $length (zero/positive) Maximum length (including ellipses)
3670          * @param string $ellipsis String to append to the truncated text
3671          * @return string
3672          */
3673         function truncateHtml( $text, $length, $ellipsis = '...' ) {
3674                 # Use the localized ellipsis character
3675                 if ( $ellipsis == '...' ) {
3676                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3677                 }
3678                 # Check if there is clearly no need to truncate
3679                 if ( $length <= 0 ) {
3680                         return $ellipsis; // no text shown, nothing to format (convention)
3681                 } elseif ( strlen( $text ) <= $length ) {
3682                         return $text; // string short enough even *with* HTML (short-circuit)
3683                 }
3684
3685                 $dispLen = 0; // innerHTML legth so far
3686                 $testingEllipsis = false; // checking if ellipses will make string longer/equal?
3687                 $tagType = 0; // 0-open, 1-close
3688                 $bracketState = 0; // 1-tag start, 2-tag name, 0-neither
3689                 $entityState = 0; // 0-not entity, 1-entity
3690                 $tag = $ret = ''; // accumulated tag name, accumulated result string
3691                 $openTags = []; // open tag stack
3692                 $maybeState = null; // possible truncation state
3693
3694                 $textLen = strlen( $text );
3695                 $neLength = max( 0, $length - strlen( $ellipsis ) ); // non-ellipsis len if truncated
3696                 for ( $pos = 0; true; ++$pos ) {
3697                         # Consider truncation once the display length has reached the maximim.
3698                         # We check if $dispLen > 0 to grab tags for the $neLength = 0 case.
3699                         # Check that we're not in the middle of a bracket/entity...
3700                         if ( $dispLen && $dispLen >= $neLength && $bracketState == 0 && !$entityState ) {
3701                                 if ( !$testingEllipsis ) {
3702                                         $testingEllipsis = true;
3703                                         # Save where we are; we will truncate here unless there turn out to
3704                                         # be so few remaining characters that truncation is not necessary.
3705                                         if ( !$maybeState ) { // already saved? ($neLength = 0 case)
3706                                                 $maybeState = [ $ret, $openTags ]; // save state
3707                                         }
3708                                 } elseif ( $dispLen > $length && $dispLen > strlen( $ellipsis ) ) {
3709                                         # String in fact does need truncation, the truncation point was OK.
3710                                         // @phan-suppress-next-line PhanTypeInvalidExpressionArrayDestructuring
3711                                         list( $ret, $openTags ) = $maybeState; // reload state
3712                                         $ret = $this->removeBadCharLast( $ret ); // multi-byte char fix
3713                                         $ret .= $ellipsis; // add ellipsis
3714                                         break;
3715                                 }
3716                         }
3717                         if ( $pos >= $textLen ) {
3718                                 break; // extra iteration just for above checks
3719                         }
3720
3721                         # Read the next char...
3722                         $ch = $text[$pos];
3723                         $lastCh = $pos ? $text[$pos - 1] : '';
3724                         $ret .= $ch; // add to result string
3725                         if ( $ch == '<' ) {
3726                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags ); // for bad HTML
3727                                 $entityState = 0; // for bad HTML
3728                                 $bracketState = 1; // tag started (checking for backslash)
3729                         } elseif ( $ch == '>' ) {
3730                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags );
3731                                 $entityState = 0; // for bad HTML
3732                                 $bracketState = 0; // out of brackets
3733                         } elseif ( $bracketState == 1 ) {
3734                                 if ( $ch == '/' ) {
3735                                         $tagType = 1; // close tag (e.g. "</span>")
3736                                 } else {
3737                                         $tagType = 0; // open tag (e.g. "<span>")
3738                                         $tag .= $ch;
3739                                 }
3740                                 $bracketState = 2; // building tag name
3741                         } elseif ( $bracketState == 2 ) {
3742                                 if ( $ch != ' ' ) {
3743                                         $tag .= $ch;
3744                                 } else {
3745                                         // Name found (e.g. "<a href=..."), add on tag attributes...
3746                                         $pos += $this->truncate_skip( $ret, $text, "<>", $pos + 1 );
3747                                 }
3748                         } elseif ( $bracketState == 0 ) {
3749                                 if ( $entityState ) {
3750                                         if ( $ch == ';' ) {
3751                                                 $entityState = 0;
3752                                                 $dispLen++; // entity is one displayed char
3753                                         }
3754                                 } else {
3755                                         if ( $neLength == 0 && !$maybeState ) {
3756                                                 // Save state without $ch. We want to *hit* the first
3757                                                 // display char (to get tags) but not *use* it if truncating.
3758                                                 $maybeState = [ substr( $ret, 0, -1 ), $openTags ];
3759                                         }
3760                                         if ( $ch == '&' ) {
3761                                                 $entityState = 1; // entity found, (e.g. "&#160;")
3762                                         } else {
3763                                                 $dispLen++; // this char is displayed
3764                                                 // Add the next $max display text chars after this in one swoop...
3765                                                 $max = ( $testingEllipsis ? $length : $neLength ) - $dispLen;
3766                                                 $skipped = $this->truncate_skip( $ret, $text, "<>&", $pos + 1, $max );
3767                                                 $dispLen += $skipped;
3768                                                 $pos += $skipped;
3769                                         }
3770                                 }
3771                         }
3772                 }
3773                 // Close the last tag if left unclosed by bad HTML
3774                 $this->truncate_endBracket( $tag, $text[$textLen - 1], $tagType, $openTags );
3775                 while ( count( $openTags ) > 0 ) {
3776                         $ret .= '</' . array_pop( $openTags ) . '>'; // close open tags
3777                 }
3778                 return $ret;
3779         }
3780
3781         /**
3782          * truncateHtml() helper function
3783          * like strcspn() but adds the skipped chars to $ret
3784          *
3785          * @param string $ret
3786          * @param string $text
3787          * @param string $search
3788          * @param int $start
3789          * @param null|int $len
3790          * @return int
3791          */
3792         private function truncate_skip( &$ret, $text, $search, $start, $len = null ) {
3793                 if ( $len === null ) {
3794                         $len = -1; // -1 means "no limit" for strcspn
3795                 } elseif ( $len < 0 ) {
3796                         $len = 0; // sanity
3797                 }
3798                 $skipCount = 0;
3799                 if ( $start < strlen( $text ) ) {
3800                         $skipCount = strcspn( $text, $search, $start, $len );
3801                         $ret .= substr( $text, $start, $skipCount );
3802                 }
3803                 return $skipCount;
3804         }
3805
3806         /**
3807          * truncateHtml() helper function
3808          * (a) push or pop $tag from $openTags as needed
3809          * (b) clear $tag value
3810          * @param string &$tag Current HTML tag name we are looking at
3811          * @param int $tagType (0-open tag, 1-close tag)
3812          * @param string $lastCh Character before the '>' that ended this tag
3813          * @param array &$openTags Open tag stack (not accounting for $tag)
3814          */
3815         private function truncate_endBracket( &$tag, $tagType, $lastCh, &$openTags ) {
3816                 $tag = ltrim( $tag );
3817                 if ( $tag != '' ) {
3818                         if ( $tagType == 0 && $lastCh != '/' ) {
3819                                 $openTags[] = $tag; // tag opened (didn't close itself)
3820                         } elseif ( $tagType == 1 ) {
3821                                 if ( $openTags && $tag == $openTags[count( $openTags ) - 1] ) {
3822                                         array_pop( $openTags ); // tag closed
3823                                 }
3824                         }
3825                         $tag = '';
3826                 }
3827         }
3828
3829         /**
3830          * Grammatical transformations, needed for inflected languages
3831          * Invoked by putting {{grammar:case|word}} in a message
3832          *
3833          * @param string $word
3834          * @param string $case
3835          * @return string
3836          */
3837         function convertGrammar( $word, $case ) {
3838                 global $wgGrammarForms;
3839                 if ( isset( $wgGrammarForms[$this->getCode()][$case][$word] ) ) {
3840                         return $wgGrammarForms[$this->getCode()][$case][$word];
3841                 }
3842
3843                 $grammarTransformations = $this->getGrammarTransformations();
3844
3845                 if ( isset( $grammarTransformations[$case] ) ) {
3846                         $forms = $grammarTransformations[$case];
3847
3848                         // Some names of grammar rules are aliases for other rules.
3849                         // In such cases the value is a string rather than object,
3850                         // so load the actual rules.
3851                         if ( is_string( $forms ) ) {
3852                                 $forms = $grammarTransformations[$forms];
3853                         }
3854
3855                         foreach ( array_values( $forms ) as $rule ) {
3856                                 $form = $rule[0];
3857
3858                                 if ( $form === '@metadata' ) {
3859                                         continue;
3860                                 }
3861
3862                                 $replacement = $rule[1];
3863
3864                                 $regex = '/' . addcslashes( $form, '/' ) . '/u';
3865                                 $patternMatches = preg_match( $regex, $word );
3866
3867                                 if ( $patternMatches === false ) {
3868                                         wfLogWarning(
3869                                                 'An error occurred while processing grammar. ' .
3870                                                 "Word: '$word'. Regex: /$form/."
3871                                         );
3872                                 } elseif ( $patternMatches === 1 ) {
3873                                         $word = preg_replace( $regex, $replacement, $word );
3874
3875                                         break;
3876                                 }
3877                         }
3878                 }
3879
3880                 return $word;
3881         }
3882
3883         /**
3884          * Get the grammar forms for the content language
3885          * @return array Array of grammar forms
3886          * @since 1.20
3887          */
3888         function getGrammarForms() {
3889                 global $wgGrammarForms;
3890                 if ( isset( $wgGrammarForms[$this->getCode()] )
3891                         && is_array( $wgGrammarForms[$this->getCode()] )
3892                 ) {
3893                         return $wgGrammarForms[$this->getCode()];
3894                 }
3895
3896                 return [];
3897         }
3898
3899         /**
3900          * Get the grammar transformations data for the language.
3901          * Used like grammar forms, with {{GRAMMAR}} and cases,
3902          * but uses pairs of regexes and replacements instead of code.
3903          *
3904          * @return array[] Array of grammar transformations.
3905          * @throws MWException
3906          * @since 1.28
3907          */
3908         public function getGrammarTransformations() {
3909                 $languageCode = $this->getCode();
3910
3911                 if ( self::$grammarTransformations === null ) {
3912                         self::$grammarTransformations = new MapCacheLRU( 10 );
3913                 }
3914
3915                 if ( self::$grammarTransformations->has( $languageCode ) ) {
3916                         return self::$grammarTransformations->get( $languageCode );
3917                 }
3918
3919                 $data = [];
3920
3921                 $grammarDataFile = __DIR__ . "/data/grammarTransformations/$languageCode.json";
3922                 if ( is_readable( $grammarDataFile ) ) {
3923                         $data = FormatJson::decode(
3924                                 file_get_contents( $grammarDataFile ),
3925                                 true
3926                         );
3927
3928                         if ( $data === null ) {
3929                                 throw new MWException( "Invalid grammar data for \"$languageCode\"." );
3930                         }
3931
3932                         self::$grammarTransformations->set( $languageCode, $data );
3933                 }
3934
3935                 return $data;
3936         }
3937
3938         /**
3939          * Provides an alternative text depending on specified gender.
3940          * Usage {{gender:username|masculine|feminine|unknown}}.
3941          * username is optional, in which case the gender of current user is used,
3942          * but only in (some) interface messages; otherwise default gender is used.
3943          *
3944          * If no forms are given, an empty string is returned. If only one form is
3945          * given, it will be returned unconditionally. These details are implied by
3946          * the caller and cannot be overridden in subclasses.
3947          *
3948          * If three forms are given, the default is to use the third (unknown) form.
3949          * If fewer than three forms are given, the default is to use the first (masculine) form.
3950          * These details can be overridden in subclasses.
3951          *
3952          * @param string $gender
3953          * @param array $forms
3954          *
3955          * @return string
3956          */
3957         function gender( $gender, $forms ) {
3958                 if ( !count( $forms ) ) {
3959                         return '';
3960                 }
3961                 $forms = $this->preConvertPlural( $forms, 2 );
3962                 if ( $gender === 'male' ) {
3963                         return $forms[0];
3964                 }
3965                 if ( $gender === 'female' ) {
3966                         return $forms[1];
3967                 }
3968                 return $forms[2] ?? $forms[0];
3969         }
3970
3971         /**
3972          * Plural form transformations, needed for some languages.
3973          * For example, there are 3 form of plural in Russian and Polish,
3974          * depending on "count mod 10". See [[w:Plural]]
3975          * For English it is pretty simple.
3976          *
3977          * Invoked by putting {{plural:count|wordform1|wordform2}}
3978          * or {{plural:count|wordform1|wordform2|wordform3}}
3979          *
3980          * Example: {{plural:{{NUMBEROFARTICLES}}|article|articles}}
3981          *
3982          * @param int $count Non-localized number
3983          * @param array $forms Different plural forms
3984          * @return string Correct form of plural for $count in this language
3985          */
3986         function convertPlural( $count, $forms ) {
3987                 // Handle explicit n=pluralform cases
3988                 $forms = $this->handleExplicitPluralForms( $count, $forms );
3989                 if ( is_string( $forms ) ) {
3990                         return $forms;
3991                 }
3992                 if ( !count( $forms ) ) {
3993                         return '';
3994                 }
3995
3996                 $pluralForm = $this->getPluralRuleIndexNumber( $count );
3997                 $pluralForm = min( $pluralForm, count( $forms ) - 1 );
3998                 return $forms[$pluralForm];
3999         }
4000
4001         /**
4002          * Handles explicit plural forms for Language::convertPlural()
4003          *
4004          * In {{PLURAL:$1|0=nothing|one|many}}, 0=nothing will be returned if $1 equals zero.
4005          * If an explicitly defined plural form matches the $count, then
4006          * string value returned, otherwise array returned for further consideration
4007          * by CLDR rules or overridden convertPlural().
4008          *
4009          * @since 1.23
4010          *
4011          * @param int $count Non-localized number
4012          * @param array $forms Different plural forms
4013          *
4014          * @return array|string
4015          */
4016         protected function handleExplicitPluralForms( $count, array $forms ) {
4017                 foreach ( $forms as $index => $form ) {
4018                         if ( preg_match( '/\d+=/i', $form ) ) {
4019                                 $pos = strpos( $form, '=' );
4020                                 if ( substr( $form, 0, $pos ) === (string)$count ) {
4021                                         return substr( $form, $pos + 1 );
4022                                 }
4023                                 unset( $forms[$index] );
4024                         }
4025                 }
4026                 return array_values( $forms );
4027         }
4028
4029         /**
4030          * Checks that convertPlural was given an array and pads it to requested
4031          * amount of forms by copying the last one.
4032          *
4033          * @param array $forms
4034          * @param int $count How many forms should there be at least
4035          * @return array Padded array of forms
4036          */
4037         protected function preConvertPlural( /* Array */ $forms, $count ) {
4038                 return array_pad( $forms, $count, end( $forms ) );
4039         }
4040
4041         /**
4042          * Wraps argument with unicode control characters for directionality safety
4043          *
4044          * This solves the problem where directionality-neutral characters at the edge of
4045          * the argument string get interpreted with the wrong directionality from the
4046          * enclosing context, giving renderings that look corrupted like "(Ben_(WMF".
4047          *
4048          * The wrapping is LRE...PDF or RLE...PDF, depending on the detected
4049          * directionality of the argument string, using the BIDI algorithm's own "First
4050          * strong directional codepoint" rule. Essentially, this works round the fact that
4051          * there is no embedding equivalent of U+2068 FSI (isolation with heuristic
4052          * direction inference). The latter is cleaner but still not widely supported.
4053          *
4054          * @param string $text Text to wrap
4055          * @return string Text, wrapped in LRE...PDF or RLE...PDF or nothing
4056          */
4057         public function embedBidi( $text = '' ) {
4058                 $dir = self::strongDirFromContent( $text );
4059                 if ( $dir === 'ltr' ) {
4060                         // Wrap in LEFT-TO-RIGHT EMBEDDING ... POP DIRECTIONAL FORMATTING
4061                         return self::$lre . $text . self::$pdf;
4062                 }
4063                 if ( $dir === 'rtl' ) {
4064                         // Wrap in RIGHT-TO-LEFT EMBEDDING ... POP DIRECTIONAL FORMATTING
4065                         return self::$rle . $text . self::$pdf;
4066                 }
4067                 // No strong directionality: do not wrap
4068                 return $text;
4069         }
4070
4071         /**
4072          * @todo Maybe translate block durations.  Note that this function is somewhat misnamed: it
4073          * deals with translating the *duration* ("1 week", "4 days", etc), not the expiry time
4074          * (which is an absolute timestamp). Please note: do NOT add this blindly, as it is used
4075          * on old expiry lengths recorded in log entries. You'd need to provide the start date to
4076          * match up with it.
4077          *
4078          * @param string $str The validated block duration in English
4079          * @param User|null $user User object to use timezone from or null for $wgUser
4080          * @param int $now Current timestamp, for formatting relative block durations
4081          * @return string Somehow translated block duration
4082          * @see LanguageFi.php for example implementation
4083          */
4084         function translateBlockExpiry( $str, User $user = null, $now = 0 ) {
4085                 $duration = SpecialBlock::getSuggestedDurations( $this );
4086                 foreach ( $duration as $show => $value ) {
4087                         if ( strcmp( $str, $value ) == 0 ) {
4088                                 return htmlspecialchars( trim( $show ) );
4089                         }
4090                 }
4091
4092                 if ( wfIsInfinity( $str ) ) {
4093                         foreach ( $duration as $show => $value ) {
4094                                 if ( wfIsInfinity( $value ) ) {
4095                                         return htmlspecialchars( trim( $show ) );
4096                                 }
4097                         }
4098                 }
4099
4100                 // If all else fails, return a standard duration or timestamp description.
4101                 $time = strtotime( $str, $now );
4102                 if ( $time === false ) { // Unknown format. Return it as-is in case.
4103                         return $str;
4104                 } elseif ( $time !== strtotime( $str, $now + 1 ) ) { // It's a relative timestamp.
4105                         // The result differs based on current time, so the difference
4106                         // is a fixed duration length.
4107                         return $this->formatDuration( $time - $now );
4108                 } else { // It's an absolute timestamp.
4109                         if ( $time === 0 ) {
4110                                 // wfTimestamp() handles 0 as current time instead of epoch.
4111                                 $time = '19700101000000';
4112                         }
4113                         if ( $user ) {
4114                                 return $this->userTimeAndDate( $time, $user );
4115                         }
4116                         return $this->timeanddate( $time );
4117                 }
4118         }
4119
4120         /**
4121          * languages like Chinese need to be segmented in order for the diff
4122          * to be of any use
4123          *
4124          * @param string $text
4125          * @return string
4126          */
4127         public function segmentForDiff( $text ) {
4128                 return $text;
4129         }
4130
4131         /**
4132          * and unsegment to show the result
4133          *
4134          * @param string $text
4135          * @return string
4136          */
4137         public function unsegmentForDiff( $text ) {
4138                 return $text;
4139         }
4140
4141         /**
4142          * Return the LanguageConverter used in the Language
4143          *
4144          * @since 1.19
4145          * @return LanguageConverter
4146          */
4147         public function getConverter() {
4148                 return $this->mConverter;
4149         }
4150
4151         /**
4152          * convert text to a variant
4153          *
4154          * @param string $text text to convert
4155          * @param string|bool $variant variant to convert to, or false to use the user's preferred
4156          *      variant (if logged in), or the project default variant
4157          * @return string the converted string
4158          */
4159         public function autoConvert( $text, $variant = false ) {
4160                 return $this->mConverter->autoConvert( $text, $variant );
4161         }
4162
4163         /**
4164          * convert text to all supported variants
4165          *
4166          * @param string $text
4167          * @return array
4168          */
4169         public function autoConvertToAllVariants( $text ) {
4170                 return $this->mConverter->autoConvertToAllVariants( $text );
4171         }
4172
4173         /**
4174          * convert text to different variants of a language.
4175          *
4176          * @warning Glossary state is maintained between calls. This means
4177          *  if you pass unescaped text to this method it can cause an XSS
4178          *  in later calls to this method, even if the later calls have properly
4179          *  escaped the input. Never feed this method user controlled text that
4180          *  is not properly escaped!
4181          * @param string $text Content that has been already escaped for use in HTML
4182          * @return string HTML
4183          */
4184         public function convert( $text ) {
4185                 return $this->mConverter->convert( $text );
4186         }
4187
4188         /**
4189          * Convert a Title object to a string in the preferred variant
4190          *
4191          * @param Title $title
4192          * @return string
4193          */
4194         public function convertTitle( $title ) {
4195                 return $this->mConverter->convertTitle( $title );
4196         }
4197
4198         /**
4199          * Convert a namespace index to a string in the preferred variant
4200          *
4201          * @param int $ns namespace index (https://www.mediawiki.org/wiki/Manual:Namespace)
4202          * @param string|null $variant variant to convert to, or null to use the user's preferred
4203          *      variant (if logged in), or the project default variant
4204          * @return string a string representation of the namespace
4205          */
4206         public function convertNamespace( $ns, $variant = null ) {
4207                 return $this->mConverter->convertNamespace( $ns, $variant );
4208         }
4209
4210         /**
4211          * Check if this is a language with variants
4212          *
4213          * @return bool
4214          */
4215         public function hasVariants() {
4216                 return count( $this->getVariants() ) > 1;
4217         }
4218
4219         /**
4220          * Strict check if the language has the specific variant.
4221          *
4222          * Compare to LanguageConverter::validateVariant() which does a more
4223          * lenient check and attempts to coerce the given code to a valid one.
4224          *
4225          * @since 1.19
4226          * @param string $variant
4227          * @return bool
4228          */
4229         public function hasVariant( $variant ) {
4230                 return $variant && ( $variant === $this->mConverter->validateVariant( $variant ) );
4231         }
4232
4233         /**
4234          * Perform output conversion on a string, and encode for safe HTML output.
4235          * @param string $text Text to be converted
4236          * @return string
4237          * @todo this should get integrated somewhere sane
4238          */
4239         public function convertHtml( $text ) {
4240                 return htmlspecialchars( $this->convert( $text ) );
4241         }
4242
4243         /**
4244          * @param string $key
4245          * @return string
4246          */
4247         public function convertCategoryKey( $key ) {
4248                 return $this->mConverter->convertCategoryKey( $key );
4249         }
4250
4251         /**
4252          * Get the list of variants supported by this language
4253          * see sample implementation in LanguageZh.php
4254          *
4255          * @return string[] An array of language codes
4256          */
4257         public function getVariants() {
4258                 return $this->mConverter->getVariants();
4259         }
4260
4261         /**
4262          * @return string
4263          */
4264         public function getPreferredVariant() {
4265                 return $this->mConverter->getPreferredVariant();
4266         }
4267
4268         /**
4269          * @return string
4270          */
4271         public function getDefaultVariant() {
4272                 return $this->mConverter->getDefaultVariant();
4273         }
4274
4275         /**
4276          * @return string
4277          */
4278         public function getURLVariant() {
4279                 return $this->mConverter->getURLVariant();
4280         }
4281
4282         /**
4283          * If a language supports multiple variants, it is
4284          * possible that non-existing link in one variant
4285          * actually exists in another variant. this function
4286          * tries to find it. See e.g. LanguageZh.php
4287          * The input parameters may be modified upon return
4288          *
4289          * @param string &$link The name of the link
4290          * @param Title &$nt The title object of the link
4291          * @param bool $ignoreOtherCond To disable other conditions when
4292          *   we need to transclude a template or update a category's link
4293          */
4294         public function findVariantLink( &$link, &$nt, $ignoreOtherCond = false ) {
4295                 $this->mConverter->findVariantLink( $link, $nt, $ignoreOtherCond );
4296         }
4297
4298         /**
4299          * returns language specific options used by User::getPageRenderHash()
4300          * for example, the preferred language variant
4301          *
4302          * @return string
4303          */
4304         function getExtraHashOptions() {
4305                 return $this->mConverter->getExtraHashOptions();
4306         }
4307
4308         /**
4309          * For languages that support multiple variants, the title of an
4310          * article may be displayed differently in different variants. this
4311          * function returns the apporiate title defined in the body of the article.
4312          *
4313          * @return string
4314          */
4315         public function getParsedTitle() {
4316                 return $this->mConverter->getParsedTitle();
4317         }
4318
4319         /**
4320          * Refresh the cache of conversion tables when
4321          * MediaWiki:Conversiontable* is updated.
4322          *
4323          * @param Title $title The Title of the page being updated
4324          */
4325         public function updateConversionTable( Title $title ) {
4326                 $this->mConverter->updateConversionTable( $title );
4327         }
4328
4329         /**
4330          * A regular expression to match legal word-trailing characters
4331          * which should be merged onto a link of the form [[foo]]bar.
4332          *
4333          * @return string
4334          */
4335         public function linkTrail() {
4336                 return self::$dataCache->getItem( $this->mCode, 'linkTrail' );
4337         }
4338
4339         /**
4340          * A regular expression character set to match legal word-prefixing
4341          * characters which should be merged onto a link of the form foo[[bar]].
4342          *
4343          * @return string
4344          */
4345         public function linkPrefixCharset() {
4346                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixCharset' );
4347         }
4348
4349         /**
4350          * Get the "parent" language which has a converter to convert a "compatible" language
4351          * (in another variant) to this language (eg. zh for zh-cn, but not en for en-gb).
4352          *
4353          * @return Language|null
4354          * @since 1.22
4355          */
4356         public function getParentLanguage() {
4357                 if ( $this->mParentLanguage !== false ) {
4358                         return $this->mParentLanguage;
4359                 }
4360
4361                 $code = explode( '-', $this->getCode() )[0];
4362                 if ( !in_array( $code, LanguageConverter::$languagesWithVariants ) ) {
4363                         $this->mParentLanguage = null;
4364                         return null;
4365                 }
4366                 $lang = self::factory( $code );
4367                 if ( !$lang->hasVariant( $this->getCode() ) ) {
4368                         $this->mParentLanguage = null;
4369                         return null;
4370                 }
4371
4372                 $this->mParentLanguage = $lang;
4373                 return $lang;
4374         }
4375
4376         /**
4377          * Compare with an other language object
4378          *
4379          * @since 1.28
4380          * @param Language $lang
4381          * @return bool
4382          */
4383         public function equals( Language $lang ) {
4384                 return $lang === $this || $lang->getCode() === $this->mCode;
4385         }
4386
4387         /**
4388          * Get the internal language code for this language object
4389          *
4390          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4391          * htmlspecialchars() or similar
4392          *
4393          * @return string
4394          */
4395         public function getCode() {
4396                 return $this->mCode;
4397         }
4398
4399         /**
4400          * Get the code in BCP 47 format which we can use
4401          * inside of html lang="" tags.
4402          *
4403          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4404          * htmlspecialchars() or similar.
4405          *
4406          * @since 1.19
4407          * @return string
4408          */
4409         public function getHtmlCode() {
4410                 if ( is_null( $this->mHtmlCode ) ) {
4411                         $this->mHtmlCode = LanguageCode::bcp47( $this->getCode() );
4412                 }
4413                 return $this->mHtmlCode;
4414         }
4415
4416         /**
4417          * Get the language code from a file name. Inverse of getFileName()
4418          * @param string $filename $prefix . $languageCode . $suffix
4419          * @param string $prefix Prefix before the language code
4420          * @param string $suffix Suffix after the language code
4421          * @return string Language code, or false if $prefix or $suffix isn't found
4422          */
4423         public static function getCodeFromFileName( $filename, $prefix = 'Language', $suffix = '.php' ) {
4424                 $m = null;
4425                 preg_match( '/' . preg_quote( $prefix, '/' ) . '([A-Z][a-z_]+)' .
4426                         preg_quote( $suffix, '/' ) . '/', $filename, $m );
4427                 if ( !count( $m ) ) {
4428                         return false;
4429                 }
4430                 return str_replace( '_', '-', strtolower( $m[1] ) );
4431         }
4432
4433         /**
4434          * @param string $code
4435          * @param bool $fallback Whether we're going through language fallback chain
4436          * @return string Name of the language class
4437          */
4438         public static function classFromCode( $code, $fallback = true ) {
4439                 if ( $fallback && $code == 'en' ) {
4440                         return 'Language';
4441                 } else {
4442                         return 'Language' . str_replace( '-', '_', ucfirst( $code ) );
4443                 }
4444         }
4445
4446         /**
4447          * Get the name of a file for a certain language code
4448          * @param string $prefix Prepend this to the filename
4449          * @param string $code Language code
4450          * @param string $suffix Append this to the filename
4451          * @throws MWException
4452          * @return string $prefix . $mangledCode . $suffix
4453          */
4454         public static function getFileName( $prefix, $code, $suffix = '.php' ) {
4455                 if ( !self::isValidBuiltInCode( $code ) ) {
4456                         throw new MWException( "Invalid language code \"$code\"" );
4457                 }
4458
4459                 return $prefix . str_replace( '-', '_', ucfirst( $code ) ) . $suffix;
4460         }
4461
4462         /**
4463          * @param string $code
4464          * @return string
4465          */
4466         public static function getMessagesFileName( $code ) {
4467                 global $IP;
4468                 $file = self::getFileName( "$IP/languages/messages/Messages", $code, '.php' );
4469                 Hooks::run( 'Language::getMessagesFileName', [ $code, &$file ] );
4470                 return $file;
4471         }
4472
4473         /**
4474          * @param string $code
4475          * @return string
4476          * @throws MWException
4477          * @since 1.23
4478          */
4479         public static function getJsonMessagesFileName( $code ) {
4480                 global $IP;
4481
4482                 if ( !self::isValidBuiltInCode( $code ) ) {
4483                         throw new MWException( "Invalid language code \"$code\"" );
4484                 }
4485
4486                 return "$IP/languages/i18n/$code.json";
4487         }
4488
4489         /**
4490          * Get the first fallback for a given language.
4491          *
4492          * @param string $code
4493          *
4494          * @return bool|string
4495          */
4496         public static function getFallbackFor( $code ) {
4497                 $fallbacks = self::getFallbacksFor( $code );
4498                 if ( $fallbacks ) {
4499                         return $fallbacks[0];
4500                 }
4501                 return false;
4502         }
4503
4504         /**
4505          * Get the ordered list of fallback languages.
4506          *
4507          * @since 1.19
4508          * @param string $code Language code
4509          * @param int $mode Fallback mode, either MESSAGES_FALLBACKS (which always falls back to 'en'),
4510          * or STRICT_FALLBACKS (whic honly falls back to 'en' when explicitly defined)
4511          * @throws MWException
4512          * @return array List of language codes
4513          */
4514         public static function getFallbacksFor( $code, $mode = self::MESSAGES_FALLBACKS ) {
4515                 if ( $code === 'en' || !self::isValidBuiltInCode( $code ) ) {
4516                         return [];
4517                 }
4518                 switch ( $mode ) {
4519                         case self::MESSAGES_FALLBACKS:
4520                                 // For unknown languages, fallbackSequence returns an empty array,
4521                                 // hardcode fallback to 'en' in that case as English messages are
4522                                 // always defined.
4523                                 return self::getLocalisationCache()->getItem( $code, 'fallbackSequence' ) ?: [ 'en' ];
4524                         case self::STRICT_FALLBACKS:
4525                                 // Use this mode when you don't want to fallback to English unless
4526                                 // explicitly defined, for example when you have language-variant icons
4527                                 // and an international language-independent fallback.
4528                                 return self::getLocalisationCache()->getItem( $code, 'originalFallbackSequence' );
4529                         default:
4530                                 throw new MWException( "Invalid fallback mode \"$mode\"" );
4531                 }
4532         }
4533
4534         /**
4535          * Get the ordered list of fallback languages, ending with the fallback
4536          * language chain for the site language.
4537          *
4538          * @since 1.22
4539          * @param string $code Language code
4540          * @return array [ fallbacks, site fallbacks ]
4541          */
4542         public static function getFallbacksIncludingSiteLanguage( $code ) {
4543                 global $wgLanguageCode;
4544
4545                 // Usually, we will only store a tiny number of fallback chains, so we
4546                 // keep them in static memory.
4547                 $cacheKey = "{$code}-{$wgLanguageCode}";
4548
4549                 if ( !array_key_exists( $cacheKey, self::$fallbackLanguageCache ) ) {
4550                         $fallbacks = self::getFallbacksFor( $code );
4551
4552                         // Append the site's fallback chain, including the site language itself
4553                         $siteFallbacks = self::getFallbacksFor( $wgLanguageCode );
4554                         array_unshift( $siteFallbacks, $wgLanguageCode );
4555
4556                         // Eliminate any languages already included in the chain
4557                         $siteFallbacks = array_diff( $siteFallbacks, $fallbacks );
4558
4559                         self::$fallbackLanguageCache[$cacheKey] = [ $fallbacks, $siteFallbacks ];
4560                 }
4561                 return self::$fallbackLanguageCache[$cacheKey];
4562         }
4563
4564         /**
4565          * Get all messages for a given language
4566          * WARNING: this may take a long time. If you just need all message *keys*
4567          * but need the *contents* of only a few messages, consider using getMessageKeysFor().
4568          *
4569          * @param string $code
4570          *
4571          * @return array
4572          */
4573         public static function getMessagesFor( $code ) {
4574                 return self::getLocalisationCache()->getItem( $code, 'messages' );
4575         }
4576
4577         /**
4578          * Get a message for a given language
4579          *
4580          * @param string $key
4581          * @param string $code
4582          *
4583          * @return string
4584          */
4585         public static function getMessageFor( $key, $code ) {
4586                 return self::getLocalisationCache()->getSubitem( $code, 'messages', $key );
4587         }
4588
4589         /**
4590          * Get all message keys for a given language. This is a faster alternative to
4591          * array_keys( Language::getMessagesFor( $code ) )
4592          *
4593          * @since 1.19
4594          * @param string $code Language code
4595          * @return array Array of message keys (strings)
4596          */
4597         public static function getMessageKeysFor( $code ) {
4598                 return self::getLocalisationCache()->getSubitemList( $code, 'messages' );
4599         }
4600
4601         /**
4602          * @param string $talk
4603          * @return string
4604          */
4605         function fixVariableInNamespace( $talk ) {
4606                 if ( strpos( $talk, '$1' ) === false ) {
4607                         return $talk;
4608                 }
4609
4610                 global $wgMetaNamespace;
4611                 $talk = str_replace( '$1', $wgMetaNamespace, $talk );
4612
4613                 # Allow grammar transformations
4614                 # Allowing full message-style parsing would make simple requests
4615                 # such as action=raw much more expensive than they need to be.
4616                 # This will hopefully cover most cases.
4617                 $talk = preg_replace_callback( '/{{grammar:(.*?)\|(.*?)}}/i',
4618                         [ $this, 'replaceGrammarInNamespace' ], $talk );
4619                 return str_replace( ' ', '_', $talk );
4620         }
4621
4622         /**
4623          * @param string $m
4624          * @return string
4625          */
4626         function replaceGrammarInNamespace( $m ) {
4627                 return $this->convertGrammar( trim( $m[2] ), trim( $m[1] ) );
4628         }
4629
4630         /**
4631          * Decode an expiry (block, protection, etc) which has come from the DB
4632          *
4633          * @param string $expiry Database expiry String
4634          * @param bool|int $format True to process using language functions, or TS_ constant
4635          *     to return the expiry in a given timestamp
4636          * @param string $infinity If $format is not true, use this string for infinite expiry
4637          * @return string
4638          * @since 1.18
4639          */
4640         public function formatExpiry( $expiry, $format = true, $infinity = 'infinity' ) {
4641                 static $dbInfinity;
4642                 if ( $dbInfinity === null ) {
4643                         $dbInfinity = wfGetDB( DB_REPLICA )->getInfinity();
4644                 }
4645
4646                 if ( $expiry == '' || $expiry === 'infinity' || $expiry == $dbInfinity ) {
4647                         return $format === true
4648                                 ? $this->getMessageFromDB( 'infiniteblock' )
4649                                 : $infinity;
4650                 } else {
4651                         return $format === true
4652                                 ? $this->timeanddate( $expiry, /* User preference timezone */ true )
4653                                 : wfTimestamp( $format, $expiry );
4654                 }
4655         }
4656
4657         /**
4658          * Formats a time given in seconds into a string representation of that time.
4659          *
4660          * @param int|float $seconds
4661          * @param array $format An optional argument that formats the returned string in different ways:
4662          *   If $format['avoid'] === 'avoidhours': don't show hours, just show days
4663          *   If $format['avoid'] === 'avoidseconds': don't show seconds if $seconds >= 1 hour,
4664          *   If $format['avoid'] === 'avoidminutes': don't show seconds/minutes if $seconds > 48 hours,
4665          *   If $format['noabbrevs'] is true: use 'seconds' and friends instead of 'seconds-abbrev'
4666          *     and friends.
4667          * @note For backwards compatibility, $format may also be one of the strings 'avoidseconds'
4668          *     or 'avoidminutes'.
4669          * @return string
4670          */
4671         function formatTimePeriod( $seconds, $format = [] ) {
4672                 if ( !is_array( $format ) ) {
4673                         $format = [ 'avoid' => $format ]; // For backwards compatibility
4674                 }
4675                 if ( !isset( $format['avoid'] ) ) {
4676                         $format['avoid'] = false;
4677                 }
4678                 if ( !isset( $format['noabbrevs'] ) ) {
4679                         $format['noabbrevs'] = false;
4680                 }
4681                 $secondsMsg = wfMessage(
4682                         $format['noabbrevs'] ? 'seconds' : 'seconds-abbrev' )->inLanguage( $this );
4683                 $minutesMsg = wfMessage(
4684                         $format['noabbrevs'] ? 'minutes' : 'minutes-abbrev' )->inLanguage( $this );
4685                 $hoursMsg = wfMessage(
4686                         $format['noabbrevs'] ? 'hours' : 'hours-abbrev' )->inLanguage( $this );
4687                 $daysMsg = wfMessage(
4688                         $format['noabbrevs'] ? 'days' : 'days-abbrev' )->inLanguage( $this );
4689
4690                 if ( round( $seconds * 10 ) < 100 ) {
4691                         $s = $this->formatNum( sprintf( "%.1f", round( $seconds * 10 ) / 10 ) );
4692                         $s = $secondsMsg->params( $s )->text();
4693                 } elseif ( round( $seconds ) < 60 ) {
4694                         $s = $this->formatNum( round( $seconds ) );
4695                         $s = $secondsMsg->params( $s )->text();
4696                 } elseif ( round( $seconds ) < 3600 ) {
4697                         $minutes = floor( $seconds / 60 );
4698                         $secondsPart = round( fmod( $seconds, 60 ) );
4699                         if ( $secondsPart == 60 ) {
4700                                 $secondsPart = 0;
4701                                 $minutes++;
4702                         }
4703                         $s = $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4704                         $s .= ' ';
4705                         $s .= $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4706                 } elseif ( round( $seconds ) <= 2 * 86400 ) {
4707                         $hours = floor( $seconds / 3600 );
4708                         $minutes = floor( ( $seconds - $hours * 3600 ) / 60 );
4709                         $secondsPart = round( $seconds - $hours * 3600 - $minutes * 60 );
4710                         if ( $secondsPart == 60 ) {
4711                                 $secondsPart = 0;
4712                                 $minutes++;
4713                         }
4714                         if ( $minutes == 60 ) {
4715                                 $minutes = 0;
4716                                 $hours++;
4717                         }
4718                         $s = $hoursMsg->params( $this->formatNum( $hours ) )->text();
4719                         $s .= ' ';
4720                         $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4721                         if ( !in_array( $format['avoid'], [ 'avoidseconds', 'avoidminutes', 'avoidhours' ] ) ) {
4722                                 $s .= ' ' . $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4723                         }
4724                 } else {
4725                         $days = floor( $seconds / 86400 );
4726                         if ( $format['avoid'] === 'avoidhours' ) {
4727                                 $hours = round( ( $seconds - $days * 86400 ) / 3600 );
4728                                 if ( $hours == 24 ) {
4729                                         $hours = 0;
4730                                         $days++;
4731                                 }
4732                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4733                         } elseif ( $format['avoid'] === 'avoidminutes' ) {
4734                                 $hours = round( ( $seconds - $days * 86400 ) / 3600 );
4735                                 if ( $hours == 24 ) {
4736                                         $hours = 0;
4737                                         $days++;
4738                                 }
4739                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4740                                 $s .= ' ';
4741                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4742                         } elseif ( $format['avoid'] === 'avoidseconds' ) {
4743                                 $hours = floor( ( $seconds - $days * 86400 ) / 3600 );
4744                                 $minutes = round( ( $seconds - $days * 86400 - $hours * 3600 ) / 60 );
4745                                 if ( $minutes == 60 ) {
4746                                         $minutes = 0;
4747                                         $hours++;
4748                                 }
4749                                 if ( $hours == 24 ) {
4750                                         $hours = 0;
4751                                         $days++;
4752                                 }
4753                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4754                                 $s .= ' ';
4755                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4756                                 $s .= ' ';
4757                                 $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4758                         } else {
4759                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4760                                 $s .= ' ';
4761                                 $s .= $this->formatTimePeriod( $seconds - $days * 86400, $format );
4762                         }
4763                 }
4764                 return $s;
4765         }
4766
4767         /**
4768          * Format a bitrate for output, using an appropriate
4769          * unit (bps, kbps, Mbps, Gbps, Tbps, Pbps, Ebps, Zbps or Ybps) according to
4770          *   the magnitude in question.
4771          *
4772          * This use base 1000. For base 1024 use formatSize(), for another base
4773          * see formatComputingNumbers().
4774          *
4775          * @param int $bps
4776          * @return string
4777          */
4778         function formatBitrate( $bps ) {
4779                 return $this->formatComputingNumbers( $bps, 1000, "bitrate-$1bits" );
4780         }
4781
4782         /**
4783          * @param int $size Size of the unit
4784          * @param int $boundary Size boundary (1000, or 1024 in most cases)
4785          * @param string $messageKey Message key to be uesd
4786          * @return string
4787          */
4788         function formatComputingNumbers( $size, $boundary, $messageKey ) {
4789                 if ( $size <= 0 ) {
4790                         return str_replace( '$1', $this->formatNum( $size ),
4791                                 $this->getMessageFromDB( str_replace( '$1', '', $messageKey ) )
4792                         );
4793                 }
4794                 $sizes = [ '', 'kilo', 'mega', 'giga', 'tera', 'peta', 'exa', 'zeta', 'yotta' ];
4795                 $index = 0;
4796
4797                 $maxIndex = count( $sizes ) - 1;
4798                 while ( $size >= $boundary && $index < $maxIndex ) {
4799                         $index++;
4800                         $size /= $boundary;
4801                 }
4802
4803                 // For small sizes no decimal places necessary
4804                 $round = 0;
4805                 if ( $index > 1 ) {
4806                         // For MB and bigger two decimal places are smarter
4807                         $round = 2;
4808                 }
4809                 $msg = str_replace( '$1', $sizes[$index], $messageKey );
4810
4811                 $size = round( $size, $round );
4812                 $text = $this->getMessageFromDB( $msg );
4813                 return str_replace( '$1', $this->formatNum( $size ), $text );
4814         }
4815
4816         /**
4817          * Format a size in bytes for output, using an appropriate
4818          * unit (B, KB, MB, GB, TB, PB, EB, ZB or YB) according to the magnitude in question
4819          *
4820          * This method use base 1024. For base 1000 use formatBitrate(), for
4821          * another base see formatComputingNumbers()
4822          *
4823          * @param int $size Size to format
4824          * @return string Plain text (not HTML)
4825          */
4826         function formatSize( $size ) {
4827                 return $this->formatComputingNumbers( $size, 1024, "size-$1bytes" );
4828         }
4829
4830         /**
4831          * Make a list item, used by various special pages
4832          *
4833          * @param string $page Page link
4834          * @param string $details HTML safe text between brackets
4835          * @param bool $oppositedm Add the direction mark opposite to your
4836          *   language, to display text properly
4837          * @return string HTML escaped
4838          */
4839         function specialList( $page, $details, $oppositedm = true ) {
4840                 if ( !$details ) {
4841                         return $page;
4842                 }
4843
4844                 $dirmark = ( $oppositedm ? $this->getDirMark( true ) : '' ) . $this->getDirMark();
4845                 return $page .
4846                         $dirmark .
4847                         $this->msg( 'word-separator' )->escaped() .
4848                         $this->msg( 'parentheses' )->rawParams( $details )->escaped();
4849         }
4850
4851         /**
4852          * Generate (prev x| next x) (20|50|100...) type links for paging
4853          *
4854          * @param Title $title Title object to link
4855          * @param int $offset
4856          * @param int $limit
4857          * @param array $query Optional URL query parameter string
4858          * @param bool $atend Optional param for specified if this is the last page
4859          * @return string
4860          * @deprecated since 1.34, use PrevNextNavigationRenderer::buildPrevNextNavigation()
4861          *  instead.
4862          */
4863         public function viewPrevNext( Title $title, $offset, $limit,
4864                 array $query = [], $atend = false
4865         ) {
4866                 wfDeprecated( __METHOD__, '1.34' );
4867                 // @todo FIXME: Why on earth this needs one message for the text and another one for tooltip?
4868
4869                 # Make 'previous' link
4870                 $prev = wfMessage( 'prevn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4871                 if ( $offset > 0 ) {
4872                         $plink = $this->numLink( $title, max( $offset - $limit, 0 ), $limit,
4873                                 $query, $prev, 'prevn-title', 'mw-prevlink' );
4874                 } else {
4875                         $plink = htmlspecialchars( $prev );
4876                 }
4877
4878                 # Make 'next' link
4879                 $next = wfMessage( 'nextn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4880                 if ( $atend ) {
4881                         $nlink = htmlspecialchars( $next );
4882                 } else {
4883                         $nlink = $this->numLink( $title, $offset + $limit, $limit,
4884                                 $query, $next, 'nextn-title', 'mw-nextlink' );
4885                 }
4886
4887                 # Make links to set number of items per page
4888                 $numLinks = [];
4889                 foreach ( [ 20, 50, 100, 250, 500 ] as $num ) {
4890                         $numLinks[] = $this->numLink( $title, $offset, $num,
4891                                 $query, $this->formatNum( $num ), 'shown-title', 'mw-numlink' );
4892                 }
4893
4894                 return wfMessage( 'viewprevnext' )->inLanguage( $this )->title( $title
4895                         )->rawParams( $plink, $nlink, $this->pipeList( $numLinks ) )->escaped();
4896         }
4897
4898         /**
4899          * Helper function for viewPrevNext() that generates links
4900          *
4901          * @param Title $title Title object to link
4902          * @param int $offset
4903          * @param int $limit
4904          * @param array $query Extra query parameters
4905          * @param string $link Text to use for the link; will be escaped
4906          * @param string $tooltipMsg Name of the message to use as tooltip
4907          * @param string $class Value of the "class" attribute of the link
4908          * @return string HTML fragment
4909          */
4910         private function numLink( Title $title, $offset, $limit, array $query, $link,
4911                 $tooltipMsg, $class
4912         ) {
4913                 $query = [ 'limit' => $limit, 'offset' => $offset ] + $query;
4914                 $tooltip = wfMessage( $tooltipMsg )->inLanguage( $this )->title( $title )
4915                         ->numParams( $limit )->text();
4916
4917                 return Html::element( 'a', [ 'href' => $title->getLocalURL( $query ),
4918                         'title' => $tooltip, 'class' => $class ], $link );
4919         }
4920
4921         /**
4922          * Get the conversion rule title, if any.
4923          *
4924          * @return string
4925          */
4926         public function getConvRuleTitle() {
4927                 return $this->mConverter->getConvRuleTitle();
4928         }
4929
4930         /**
4931          * Get the compiled plural rules for the language
4932          * @since 1.20
4933          * @return array Associative array with plural form, and plural rule as key-value pairs
4934          */
4935         public function getCompiledPluralRules() {
4936                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'compiledPluralRules' );
4937                 $fallbacks = self::getFallbacksFor( $this->mCode );
4938                 if ( !$pluralRules ) {
4939                         foreach ( $fallbacks as $fallbackCode ) {
4940                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'compiledPluralRules' );
4941                                 if ( $pluralRules ) {
4942                                         break;
4943                                 }
4944                         }
4945                 }
4946                 return $pluralRules;
4947         }
4948
4949         /**
4950          * Get the plural rules for the language
4951          * @since 1.20
4952          * @return array Associative array with plural form number and plural rule as key-value pairs
4953          */
4954         public function getPluralRules() {
4955                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRules' );
4956                 $fallbacks = self::getFallbacksFor( $this->mCode );
4957                 if ( !$pluralRules ) {
4958                         foreach ( $fallbacks as $fallbackCode ) {
4959                                 $pluralRules = self::$dataCache->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                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRuleTypes' );
4975                 $fallbacks = self::getFallbacksFor( $this->mCode );
4976                 if ( !$pluralRuleTypes ) {
4977                         foreach ( $fallbacks as $fallbackCode ) {
4978                                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRuleTypes' );
4979                                 if ( $pluralRuleTypes ) {
4980                                         break;
4981                                 }
4982                         }
4983                 }
4984                 return $pluralRuleTypes;
4985         }
4986
4987         /**
4988          * Find the index number of the plural rule appropriate for the given number
4989          * @param int $number
4990          * @return int The index number of the plural rule
4991          */
4992         public function getPluralRuleIndexNumber( $number ) {
4993                 $pluralRules = $this->getCompiledPluralRules();
4994                 $form = Evaluator::evaluateCompiled( $number, $pluralRules );
4995                 return $form;
4996         }
4997
4998         /**
4999          * Find the plural rule type appropriate for the given number
5000          * For example, if the language is set to Arabic, getPluralType(5) should
5001          * return 'few'.
5002          * @since 1.22
5003          * @param int $number
5004          * @return string The name of the plural rule type, e.g. one, two, few, many
5005          */
5006         public function getPluralRuleType( $number ) {
5007                 $index = $this->getPluralRuleIndexNumber( $number );
5008                 $pluralRuleTypes = $this->getPluralRuleTypes();
5009                 return $pluralRuleTypes[$index] ?? 'other';
5010         }
5011 }