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 Wikimedia\Assert\Assert;
  31
  32 /**
  33  * Internationalisation code
  34  * @ingroup Language
  35  */
  36 class Language {
  37         /**
  38          * Return autonyms in fetchLanguageName(s).
  39          * @since 1.32
  40          */
  41         const AS_AUTONYMS = null;
  42
  43         /**
  44          * Return all known languages in fetchLanguageName(s).
  45          * @since 1.32
  46          */
  47         const ALL = 'all';
  48
  49         /**
  50          * Return in fetchLanguageName(s) only the languages for which we have at
  51          * least some localisation.
  52          * @since 1.32
  53          */
  54         const SUPPORTED = 'mwfile';
  55
  56         /**
  57          * @var LanguageConverter
  58          */
  59         public $mConverter;
  60
  61         public $mVariants, $mCode, $mLoaded = false;
  62         public $mMagicExtensions = [];
  63         private $mHtmlCode = null, $mParentLanguage = false;
  64
  65         public $dateFormatStrings = [];
  66         public $mExtendedSpecialPageAliases;
  67
  68         /** @var array|null */
  69         protected $namespaceNames;
  70         protected $mNamespaceIds, $namespaceAliases;
  71
  72         /**
  73          * ReplacementArray object caches
  74          */
  75         public $transformData = [];
  76
  77         /**
  78          * @var LocalisationCache
  79          */
  80         public static $dataCache;
  81
  82         public static $mLangObjCache = [];
  83
  84         /**
  85          * Return a fallback chain for messages in getFallbacksFor
  86          * @since 1.32
  87          */
  88         const MESSAGES_FALLBACKS = 0;
  89
  90         /**
  91          * Return a strict fallback chain in getFallbacksFor
  92          * @since 1.32
  93          */
  94         const STRICT_FALLBACKS = 1;
  95
  96         public static $mWeekdayMsgs = [
  97                 'sunday', 'monday', 'tuesday', 'wednesday', 'thursday',
  98                 'friday', 'saturday'
  99         ];
 100
 101         public static $mWeekdayAbbrevMsgs = [
 102                 'sun', 'mon', 'tue', 'wed', 'thu', 'fri', 'sat'
 103         ];
 104
 105         public static $mMonthMsgs = [
 106                 'january', 'february', 'march', 'april', 'may_long', 'june',
 107                 'july', 'august', 'september', 'october', 'november',
 108                 'december'
 109         ];
 110         public static $mMonthGenMsgs = [
 111                 'january-gen', 'february-gen', 'march-gen', 'april-gen', 'may-gen', 'june-gen',
 112                 'july-gen', 'august-gen', 'september-gen', 'october-gen', 'november-gen',
 113                 'december-gen'
 114         ];
 115         public static $mMonthAbbrevMsgs = [
 116                 'jan', 'feb', 'mar', 'apr', 'may', 'jun', 'jul', 'aug',
 117                 'sep', 'oct', 'nov', 'dec'
 118         ];
 119
 120         public static $mIranianCalendarMonthMsgs = [
 121                 'iranian-calendar-m1', 'iranian-calendar-m2', 'iranian-calendar-m3',
 122                 'iranian-calendar-m4', 'iranian-calendar-m5', 'iranian-calendar-m6',
 123                 'iranian-calendar-m7', 'iranian-calendar-m8', 'iranian-calendar-m9',
 124                 'iranian-calendar-m10', 'iranian-calendar-m11', 'iranian-calendar-m12'
 125         ];
 126
 127         public static $mHebrewCalendarMonthMsgs = [
 128                 'hebrew-calendar-m1', 'hebrew-calendar-m2', 'hebrew-calendar-m3',
 129                 'hebrew-calendar-m4', 'hebrew-calendar-m5', 'hebrew-calendar-m6',
 130                 'hebrew-calendar-m7', 'hebrew-calendar-m8', 'hebrew-calendar-m9',
 131                 'hebrew-calendar-m10', 'hebrew-calendar-m11', 'hebrew-calendar-m12',
 132                 'hebrew-calendar-m6a', 'hebrew-calendar-m6b'
 133         ];
 134
 135         public static $mHebrewCalendarMonthGenMsgs = [
 136                 'hebrew-calendar-m1-gen', 'hebrew-calendar-m2-gen', 'hebrew-calendar-m3-gen',
 137                 'hebrew-calendar-m4-gen', 'hebrew-calendar-m5-gen', 'hebrew-calendar-m6-gen',
 138                 'hebrew-calendar-m7-gen', 'hebrew-calendar-m8-gen', 'hebrew-calendar-m9-gen',
 139                 'hebrew-calendar-m10-gen', 'hebrew-calendar-m11-gen', 'hebrew-calendar-m12-gen',
 140                 'hebrew-calendar-m6a-gen', 'hebrew-calendar-m6b-gen'
 141         ];
 142
 143         public static $mHijriCalendarMonthMsgs = [
 144                 'hijri-calendar-m1', 'hijri-calendar-m2', 'hijri-calendar-m3',
 145                 'hijri-calendar-m4', 'hijri-calendar-m5', 'hijri-calendar-m6',
 146                 'hijri-calendar-m7', 'hijri-calendar-m8', 'hijri-calendar-m9',
 147                 'hijri-calendar-m10', 'hijri-calendar-m11', 'hijri-calendar-m12'
 148         ];
 149
 150         /**
 151          * @since 1.20
 152          * @var array
 153          */
 154         public static $durationIntervals = [
 155                 'millennia' => 31556952000,
 156                 'centuries' => 3155695200,
 157                 'decades' => 315569520,
 158                 'years' => 31556952, // 86400 * ( 365 + ( 24 * 3 + 25 ) / 400 )
 159                 'weeks' => 604800,
 160                 'days' => 86400,
 161                 'hours' => 3600,
 162                 'minutes' => 60,
 163                 'seconds' => 1,
 164         ];
 165
 166         /**
 167          * Cache for language fallbacks.
 168          * @see Language::getFallbacksIncludingSiteLanguage
 169          * @since 1.21
 170          * @var array
 171          */
 172         private static $fallbackLanguageCache = [];
 173
 174         /**
 175          * Cache for grammar rules data
 176          * @var MapCacheLRU|null
 177          */
 178         private static $grammarTransformations;
 179
 180         /**
 181          * Cache for language names
 182          * @var HashBagOStuff|null
 183          */
 184         private static $languageNameCache;
 185
 186         /**
 187          * Unicode directional formatting characters, for embedBidi()
 188          */
 189         private static $lre = "\u{202A}"; // U+202A LEFT-TO-RIGHT EMBEDDING
 190         private static $rle = "\u{202B}"; // U+202B RIGHT-TO-LEFT EMBEDDING
 191         private static $pdf = "\u{202C}"; // U+202C POP DIRECTIONAL FORMATTING
 192
 193         /**
 194          * Directionality test regex for embedBidi(). Matches the first strong directionality codepoint:
 195          * - in group 1 if it is LTR
 196          * - in group 2 if it is RTL
 197          * Does not match if there is no strong directionality codepoint.
 198          *
 199          * The form is '/(?:([strong ltr codepoint])|([strong rtl codepoint]))/u' .
 200          *
 201          * Generated by UnicodeJS (see tools/strongDir) from the UCD; see
 202          * https://phabricator.wikimedia.org/diffusion/GUJS/ .
 203          */
 204         // @codeCoverageIgnoreStart
 205         // phpcs:ignore Generic.Files.LineLength
 206         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';
 207         // @codeCoverageIgnoreEnd
 208
 209         /**
 210          * Get a cached or new language object for a given language code
 211          * @param string $code
 212          * @throws MWException
 213          * @return Language
 214          */
 215         static function factory( $code ) {
 216                 global $wgDummyLanguageCodes, $wgLangObjCacheSize;
 217
 218                 if ( isset( $wgDummyLanguageCodes[$code] ) ) {
 219                         $code = $wgDummyLanguageCodes[$code];
 220                 }
 221
 222                 // get the language object to process
 223                 $langObj = self::$mLangObjCache[$code] ?? self::newFromCode( $code );
 224
 225                 // merge the language object in to get it up front in the cache
 226                 self::$mLangObjCache = array_merge( [ $code => $langObj ], self::$mLangObjCache );
 227                 // get rid of the oldest ones in case we have an overflow
 228                 self::$mLangObjCache = array_slice( self::$mLangObjCache, 0, $wgLangObjCacheSize, true );
 229
 230                 return $langObj;
 231         }
 232
 233         /**
 234          * Create a language object for a given language code
 235          * @param string $code
 236          * @param bool $fallback Whether we're going through language fallback chain
 237          * @throws MWException
 238          * @return Language
 239          */
 240         protected static function newFromCode( $code, $fallback = false ) {
 241                 if ( !self::isValidCode( $code ) ) {
 242                         throw new MWException( "Invalid language code \"$code\"" );
 243                 }
 244
 245                 if ( !self::isValidBuiltInCode( $code ) ) {
 246                         // It's not possible to customise this code with class files, so
 247                         // just return a Language object. This is to support uselang= hacks.
 248                         $lang = new Language;
 249                         $lang->mCode = $code;
 250                         return $lang;
 251                 }
 252
 253                 // Check if there is a language class for the code
 254                 $class = self::classFromCode( $code, $fallback );
 255                 // LanguageCode does not inherit Language
 256                 if ( class_exists( $class ) && is_a( $class, 'Language', true ) ) {
 257                         $lang = new $class;
 258                         return $lang;
 259                 }
 260
 261                 // Keep trying the fallback list until we find an existing class
 262                 $fallbacks = self::getFallbacksFor( $code );
 263                 foreach ( $fallbacks as $fallbackCode ) {
 264                         if ( !self::isValidBuiltInCode( $fallbackCode ) ) {
 265                                 throw new MWException( "Invalid fallback '$fallbackCode' in fallback sequence for '$code'" );
 266                         }
 267
 268                         $class = self::classFromCode( $fallbackCode );
 269                         if ( class_exists( $class ) ) {
 270                                 $lang = new $class;
 271                                 $lang->mCode = $code;
 272                                 return $lang;
 273                         }
 274                 }
 275
 276                 throw new MWException( "Invalid fallback sequence for language '$code'" );
 277         }
 278
 279         /**
 280          * Intended for tests that may change configuration in a way that invalidates caches.
 281          *
 282          * @since 1.32
 283          */
 284         public static function clearCaches() {
 285                 if ( !defined( 'MW_PHPUNIT_TEST' ) ) {
 286                         throw new MWException( __METHOD__ . ' must not be used outside tests' );
 287                 }
 288                 self::$dataCache = null;
 289                 // Reinitialize $dataCache, since it's expected to always be available
 290                 self::getLocalisationCache();
 291                 self::$mLangObjCache = [];
 292                 self::$fallbackLanguageCache = [];
 293                 self::$grammarTransformations = null;
 294                 self::$languageNameCache = null;
 295         }
 296
 297         /**
 298          * Checks whether any localisation is available for that language tag
 299          * in MediaWiki (MessagesXx.php exists).
 300          *
 301          * @param string $code Language tag (in lower case)
 302          * @return bool Whether language is supported
 303          * @since 1.21
 304          */
 305         public static function isSupportedLanguage( $code ) {
 306                 if ( !self::isValidBuiltInCode( $code ) ) {
 307                         return false;
 308                 }
 309
 310                 if ( $code === 'qqq' ) {
 311                         return false;
 312                 }
 313
 314                 return is_readable( self::getMessagesFileName( $code ) ) ||
 315                         is_readable( self::getJsonMessagesFileName( $code ) );
 316         }
 317
 318         /**
 319          * Returns true if a language code string is a well-formed language tag
 320          * according to RFC 5646.
 321          * This function only checks well-formedness; it doesn't check that
 322          * language, script or variant codes actually exist in the repositories.
 323          *
 324          * Based on regexes by Mark Davis of the Unicode Consortium:
 325          * https://www.unicode.org/repos/cldr/trunk/tools/java/org/unicode/cldr/util/data/langtagRegex.txt
 326          *
 327          * @param string $code
 328          * @param bool $lenient Whether to allow '_' as separator. The default is only '-'.
 329          *
 330          * @return bool
 331          * @since 1.21
 332          */
 333         public static function isWellFormedLanguageTag( $code, $lenient = false ) {
 334                 $alpha = '[a-z]';
 335                 $digit = '[0-9]';
 336                 $alphanum = '[a-z0-9]';
 337                 $x = 'x'; # private use singleton
 338                 $singleton = '[a-wy-z]'; # other singleton
 339                 $s = $lenient ? '[-_]' : '-';
 340
 341                 $language = "$alpha{2,8}|$alpha{2,3}$s$alpha{3}";
 342                 $script = "$alpha{4}"; # ISO 15924
 343                 $region = "(?:$alpha{2}|$digit{3})"; # ISO 3166-1 alpha-2 or UN M.49
 344                 $variant = "(?:$alphanum{5,8}|$digit$alphanum{3})";
 345                 $extension = "$singleton(?:$s$alphanum{2,8})+";
 346                 $privateUse = "$x(?:$s$alphanum{1,8})+";
 347
 348                 # Define certain grandfathered codes, since otherwise the regex is pretty useless.
 349                 # Since these are limited, this is safe even later changes to the registry --
 350                 # the only oddity is that it might change the type of the tag, and thus
 351                 # the results from the capturing groups.
 352                 # https://www.iana.org/assignments/language-subtag-registry
 353
 354                 $grandfathered = "en{$s}GB{$s}oed"
 355                         . "|i{$s}(?:ami|bnn|default|enochian|hak|klingon|lux|mingo|navajo|pwn|tao|tay|tsu)"
 356                         . "|no{$s}(?:bok|nyn)"
 357                         . "|sgn{$s}(?:BE{$s}(?:fr|nl)|CH{$s}de)"
 358                         . "|zh{$s}min{$s}nan";
 359
 360                 $variantList = "$variant(?:$s$variant)*";
 361                 $extensionList = "$extension(?:$s$extension)*";
 362
 363                 $langtag = "(?:($language)"
 364                         . "(?:$s$script)?"
 365                         . "(?:$s$region)?"
 366                         . "(?:$s$variantList)?"
 367                         . "(?:$s$extensionList)?"
 368                         . "(?:$s$privateUse)?)";
 369
 370                 # The final breakdown, with capturing groups for each of these components
 371                 # The variants, extensions, grandfathered, and private-use may have interior '-'
 372
 373                 $root = "^(?:$langtag|$privateUse|$grandfathered)$";
 374
 375                 return (bool)preg_match( "/$root/", strtolower( $code ) );
 376         }
 377
 378         /**
 379          * Returns true if a language code string is of a valid form, whether or
 380          * not it exists. This includes codes which are used solely for
 381          * customisation via the MediaWiki namespace.
 382          *
 383          * @param string $code
 384          *
 385          * @return bool
 386          */
 387         public static function isValidCode( $code ) {
 388                 static $cache = [];
 389                 Assert::parameterType( 'string', $code, '$code' );
 390                 if ( !isset( $cache[$code] ) ) {
 391                         // People think language codes are html safe, so enforce it.
 392                         // Ideally we should only allow a-zA-Z0-9-
 393                         // but, .+ and other chars are often used for {{int:}} hacks
 394                         // see bugs T39564, T39587, T38938
 395                         $cache[$code] =
 396                                 // Protect against path traversal
 397                                 strcspn( $code, ":/\\\000&<>'\"" ) === strlen( $code )
 398                                 && !preg_match( MediaWikiTitleCodec::getTitleInvalidRegex(), $code );
 399                 }
 400                 return $cache[$code];
 401         }
 402
 403         /**
 404          * Returns true if a language code is of a valid form for the purposes of
 405          * internal customisation of MediaWiki, via Messages*.php or *.json.
 406          *
 407          * @param string $code
 408          *
 409          * @since 1.18
 410          * @return bool
 411          */
 412         public static function isValidBuiltInCode( $code ) {
 413                 Assert::parameterType( 'string', $code, '$code' );
 414
 415                 return (bool)preg_match( '/^[a-z0-9-]{2,}$/', $code );
 416         }
 417
 418         /**
 419          * Returns true if a language code is an IETF tag known to MediaWiki.
 420          *
 421          * @param string $tag
 422          *
 423          * @since 1.21
 424          * @return bool
 425          */
 426         public static function isKnownLanguageTag( $tag ) {
 427                 // Quick escape for invalid input to avoid exceptions down the line
 428                 // when code tries to process tags which are not valid at all.
 429                 if ( !self::isValidBuiltInCode( $tag ) ) {
 430                         return false;
 431                 }
 432
 433                 if ( isset( MediaWiki\Languages\Data\Names::$names[$tag] )
 434                         || self::fetchLanguageName( $tag, $tag ) !== ''
 435                 ) {
 436                         return true;
 437                 }
 438
 439                 return false;
 440         }
 441
 442         /**
 443          * Get the LocalisationCache instance
 444          *
 445          * @return LocalisationCache
 446          */
 447         public static function getLocalisationCache() {
 448                 if ( is_null( self::$dataCache ) ) {
 449                         global $wgLocalisationCacheConf;
 450                         $class = $wgLocalisationCacheConf['class'];
 451                         self::$dataCache = new $class( $wgLocalisationCacheConf );
 452                 }
 453                 return self::$dataCache;
 454         }
 455
 456         function __construct() {
 457                 $this->mConverter = new FakeConverter( $this );
 458                 // Set the code to the name of the descendant
 459                 if ( static::class === 'Language' ) {
 460                         $this->mCode = 'en';
 461                 } else {
 462                         $this->mCode = str_replace( '_', '-', strtolower( substr( static::class, 8 ) ) );
 463                 }
 464                 self::getLocalisationCache();
 465         }
 466
 467         /**
 468          * Reduce memory usage
 469          */
 470         function __destruct() {
 471                 foreach ( $this as $name => $value ) {
 472                         unset( $this->$name );
 473                 }
 474         }
 475
 476         /**
 477          * Hook which will be called if this is the content language.
 478          * Descendants can use this to register hook functions or modify globals
 479          */
 480         function initContLang() {
 481         }
 482
 483         /**
 484          * @return array
 485          * @since 1.19
 486          */
 487         public function getFallbackLanguages() {
 488                 return self::getFallbacksFor( $this->mCode );
 489         }
 490
 491         /**
 492          * Exports $wgBookstoreListEn
 493          * @return array
 494          */
 495         public function getBookstoreList() {
 496                 return self::$dataCache->getItem( $this->mCode, 'bookstoreList' );
 497         }
 498
 499         /**
 500          * Returns an array of localised namespaces indexed by their numbers. If the namespace is not
 501          * available in localised form, it will be included in English.
 502          *
 503          * @return string[] List of localized namespace names, indexed by numeric namespace ID.
 504          */
 505         public function getNamespaces() {
 506                 if ( is_null( $this->namespaceNames ) ) {
 507                         global $wgMetaNamespace, $wgMetaNamespaceTalk, $wgExtraNamespaces;
 508
 509                         $validNamespaces = MWNamespace::getCanonicalNamespaces();
 510
 511                         $this->namespaceNames = $wgExtraNamespaces +
 512                                 self::$dataCache->getItem( $this->mCode, 'namespaceNames' );
 513                         $this->namespaceNames += $validNamespaces;
 514
 515                         $this->namespaceNames[NS_PROJECT] = $wgMetaNamespace;
 516                         if ( $wgMetaNamespaceTalk ) {
 517                                 $this->namespaceNames[NS_PROJECT_TALK] = $wgMetaNamespaceTalk;
 518                         } else {
 519                                 $talk = $this->namespaceNames[NS_PROJECT_TALK];
 520                                 $this->namespaceNames[NS_PROJECT_TALK] =
 521                                         $this->fixVariableInNamespace( $talk );
 522                         }
 523
 524                         # Sometimes a language will be localised but not actually exist on this wiki.
 525                         foreach ( $this->namespaceNames as $key => $text ) {
 526                                 if ( !isset( $validNamespaces[$key] ) ) {
 527                                         unset( $this->namespaceNames[$key] );
 528                                 }
 529                         }
 530
 531                         # The above mixing may leave namespaces out of canonical order.
 532                         # Re-order by namespace ID number...
 533                         ksort( $this->namespaceNames );
 534
 535                         Hooks::run( 'LanguageGetNamespaces', [ &$this->namespaceNames ] );
 536                 }
 537
 538                 return $this->namespaceNames;
 539         }
 540
 541         /**
 542          * Arbitrarily set all of the namespace names at once. Mainly used for testing
 543          * @param array $namespaces Array of namespaces (id => name)
 544          */
 545         public function setNamespaces( array $namespaces ) {
 546                 $this->namespaceNames = $namespaces;
 547                 $this->mNamespaceIds = null;
 548         }
 549
 550         /**
 551          * Resets all of the namespace caches. Mainly used for testing
 552          */
 553         public function resetNamespaces() {
 554                 $this->namespaceNames = null;
 555                 $this->mNamespaceIds = null;
 556                 $this->namespaceAliases = null;
 557         }
 558
 559         /**
 560          * A convenience function that returns getNamespaces() with spaces instead of underscores
 561          * in values. Useful for producing output to be displayed e.g. in `<select>` forms.
 562          *
 563          * @return array
 564          */
 565         public function getFormattedNamespaces() {
 566                 $ns = $this->getNamespaces();
 567                 foreach ( $ns as $k => $v ) {
 568                         $ns[$k] = strtr( $v, '_', ' ' );
 569                 }
 570                 return $ns;
 571         }
 572
 573         /**
 574          * Get a namespace value by key
 575          *
 576          * <code>
 577          * $mw_ns = $lang->getNsText( NS_MEDIAWIKI );
 578          * echo $mw_ns; // prints 'MediaWiki'
 579          * </code>
 580          *
 581          * @param int $index The array key of the namespace to return
 582          * @return string|bool String if the namespace value exists, otherwise false
 583          */
 584         public function getNsText( $index ) {
 585                 $ns = $this->getNamespaces();
 586                 return $ns[$index] ?? false;
 587         }
 588
 589         /**
 590          * A convenience function that returns the same thing as
 591          * getNsText() except with '_' changed to ' ', useful for
 592          * producing output.
 593          *
 594          * <code>
 595          * $mw_ns = $lang->getFormattedNsText( NS_MEDIAWIKI_TALK );
 596          * echo $mw_ns; // prints 'MediaWiki talk'
 597          * </code>
 598          *
 599          * @param int $index The array key of the namespace to return
 600          * @return string Namespace name without underscores (empty string if namespace does not exist)
 601          */
 602         public function getFormattedNsText( $index ) {
 603                 $ns = $this->getNsText( $index );
 604                 return strtr( $ns, '_', ' ' );
 605         }
 606
 607         /**
 608          * Returns gender-dependent namespace alias if available.
 609          * See https://www.mediawiki.org/wiki/Manual:$wgExtraGenderNamespaces
 610          * @param int $index Namespace index
 611          * @param string $gender Gender key (male, female... )
 612          * @return string
 613          * @since 1.18
 614          */
 615         public function getGenderNsText( $index, $gender ) {
 616                 global $wgExtraGenderNamespaces;
 617
 618                 $ns = $wgExtraGenderNamespaces +
 619                         (array)self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 620
 621                 return $ns[$index][$gender] ?? $this->getNsText( $index );
 622         }
 623
 624         /**
 625          * Whether this language uses gender-dependent namespace aliases.
 626          * See https://www.mediawiki.org/wiki/Manual:$wgExtraGenderNamespaces
 627          * @return bool
 628          * @since 1.18
 629          */
 630         public function needsGenderDistinction() {
 631                 global $wgExtraGenderNamespaces, $wgExtraNamespaces;
 632                 if ( count( $wgExtraGenderNamespaces ) > 0 ) {
 633                         // $wgExtraGenderNamespaces overrides everything
 634                         return true;
 635                 } elseif ( isset( $wgExtraNamespaces[NS_USER] ) && isset( $wgExtraNamespaces[NS_USER_TALK] ) ) {
 636                         /// @todo There may be other gender namespace than NS_USER & NS_USER_TALK in the future
 637                         // $wgExtraNamespaces overrides any gender aliases specified in i18n files
 638                         return false;
 639                 } else {
 640                         // Check what is in i18n files
 641                         $aliases = self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 642                         return count( $aliases ) > 0;
 643                 }
 644         }
 645
 646         /**
 647          * Get a namespace key by value, case insensitive.
 648          * Only matches namespace names for the current language, not the
 649          * canonical ones defined in Namespace.php.
 650          *
 651          * @param string $text
 652          * @return int|bool An integer if $text is a valid value otherwise false
 653          */
 654         function getLocalNsIndex( $text ) {
 655                 $lctext = $this->lc( $text );
 656                 $ids = $this->getNamespaceIds();
 657                 return $ids[$lctext] ?? false;
 658         }
 659
 660         /**
 661          * @return array
 662          */
 663         public function getNamespaceAliases() {
 664                 if ( is_null( $this->namespaceAliases ) ) {
 665                         $aliases = self::$dataCache->getItem( $this->mCode, 'namespaceAliases' );
 666                         if ( !$aliases ) {
 667                                 $aliases = [];
 668                         } else {
 669                                 foreach ( $aliases as $name => $index ) {
 670                                         if ( $index === NS_PROJECT_TALK ) {
 671                                                 unset( $aliases[$name] );
 672                                                 $name = $this->fixVariableInNamespace( $name );
 673                                                 $aliases[$name] = $index;
 674                                         }
 675                                 }
 676                         }
 677
 678                         global $wgExtraGenderNamespaces;
 679                         $genders = $wgExtraGenderNamespaces +
 680                                 (array)self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 681                         foreach ( $genders as $index => $forms ) {
 682                                 foreach ( $forms as $alias ) {
 683                                         $aliases[$alias] = $index;
 684                                 }
 685                         }
 686
 687                         # Also add converted namespace names as aliases, to avoid confusion.
 688                         $convertedNames = [];
 689                         foreach ( $this->getVariants() as $variant ) {
 690                                 if ( $variant === $this->mCode ) {
 691                                         continue;
 692                                 }
 693                                 foreach ( $this->getNamespaces() as $ns => $_ ) {
 694                                         $convertedNames[$this->getConverter()->convertNamespace( $ns, $variant )] = $ns;
 695                                 }
 696                         }
 697
 698                         $this->namespaceAliases = $aliases + $convertedNames;
 699
 700                         # Filter out aliases to namespaces that don't exist, e.g. from extensions
 701                         # that aren't loaded here but are included in the l10n cache.
 702                         # (array_intersect preserves keys from its first argument)
 703                         $this->namespaceAliases = array_intersect(
 704                                 $this->namespaceAliases,
 705                                 array_keys( $this->getNamespaces() )
 706                         );
 707                 }
 708
 709                 return $this->namespaceAliases;
 710         }
 711
 712         /**
 713          * @return array
 714          */
 715         public function getNamespaceIds() {
 716                 if ( is_null( $this->mNamespaceIds ) ) {
 717                         global $wgNamespaceAliases;
 718                         # Put namespace names and aliases into a hashtable.
 719                         # If this is too slow, then we should arrange it so that it is done
 720                         # before caching. The catch is that at pre-cache time, the above
 721                         # class-specific fixup hasn't been done.
 722                         $this->mNamespaceIds = [];
 723                         foreach ( $this->getNamespaces() as $index => $name ) {
 724                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 725                         }
 726                         foreach ( $this->getNamespaceAliases() as $name => $index ) {
 727                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 728                         }
 729                         if ( $wgNamespaceAliases ) {
 730                                 foreach ( $wgNamespaceAliases as $name => $index ) {
 731                                         $this->mNamespaceIds[$this->lc( $name )] = $index;
 732                                 }
 733                         }
 734                 }
 735                 return $this->mNamespaceIds;
 736         }
 737
 738         /**
 739          * Get a namespace key by value, case insensitive.  Canonical namespace
 740          * names override custom ones defined for the current language.
 741          *
 742          * @param string $text
 743          * @return int|bool An integer if $text is a valid value otherwise false
 744          */
 745         public function getNsIndex( $text ) {
 746                 $lctext = $this->lc( $text );
 747                 $ns = MWNamespace::getCanonicalIndex( $lctext );
 748                 if ( $ns !== null ) {
 749                         return $ns;
 750                 }
 751                 $ids = $this->getNamespaceIds();
 752                 return $ids[$lctext] ?? false;
 753         }
 754
 755         /**
 756          * short names for language variants used for language conversion links.
 757          *
 758          * @param string $code
 759          * @param bool $usemsg Use the "variantname-xyz" message if it exists
 760          * @return string
 761          */
 762         public function getVariantname( $code, $usemsg = true ) {
 763                 $msg = "variantname-$code";
 764                 if ( $usemsg && wfMessage( $msg )->exists() ) {
 765                         return $this->getMessageFromDB( $msg );
 766                 }
 767                 $name = self::fetchLanguageName( $code );
 768                 if ( $name ) {
 769                         return $name; # if it's defined as a language name, show that
 770                 } else {
 771                         # otherwise, output the language code
 772                         return $code;
 773                 }
 774         }
 775
 776         /**
 777          * @return string[]|bool List of date format preference keys, or false if disabled.
 778          */
 779         public function getDatePreferences() {
 780                 return self::$dataCache->getItem( $this->mCode, 'datePreferences' );
 781         }
 782
 783         /**
 784          * @return array
 785          */
 786         function getDateFormats() {
 787                 return self::$dataCache->getItem( $this->mCode, 'dateFormats' );
 788         }
 789
 790         /**
 791          * @return array|string
 792          */
 793         public function getDefaultDateFormat() {
 794                 $df = self::$dataCache->getItem( $this->mCode, 'defaultDateFormat' );
 795                 if ( $df === 'dmy or mdy' ) {
 796                         global $wgAmericanDates;
 797                         return $wgAmericanDates ? 'mdy' : 'dmy';
 798                 } else {
 799                         return $df;
 800                 }
 801         }
 802
 803         /**
 804          * @return array
 805          */
 806         public function getDatePreferenceMigrationMap() {
 807                 return self::$dataCache->getItem( $this->mCode, 'datePreferenceMigrationMap' );
 808         }
 809
 810         /**
 811          * @return array
 812          */
 813         public function getExtraUserToggles() {
 814                 return (array)self::$dataCache->getItem( $this->mCode, 'extraUserToggles' );
 815         }
 816
 817         /**
 818          * @param string $tog
 819          * @return string
 820          */
 821         function getUserToggle( $tog ) {
 822                 return $this->getMessageFromDB( "tog-$tog" );
 823         }
 824
 825         /**
 826          * Get an array of language names, indexed by code.
 827          * @param null|string $inLanguage Code of language in which to return the names
 828          *              Use self::AS_AUTONYMS for autonyms (native names)
 829          * @param string $include One of:
 830          *              self::ALL all available languages
 831          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 832          *              self::SUPPORTED only if the language is in 'mw' *and* has a message file
 833          * @return array Language code => language name (sorted by key)
 834          * @since 1.20
 835          */
 836         public static function fetchLanguageNames( $inLanguage = self::AS_AUTONYMS, $include = 'mw' ) {
 837                 $cacheKey = $inLanguage === self::AS_AUTONYMS ? 'null' : $inLanguage;
 838                 $cacheKey .= ":$include";
 839                 if ( self::$languageNameCache === null ) {
 840                         self::$languageNameCache = new HashBagOStuff( [ 'maxKeys' => 20 ] );
 841                 }
 842
 843                 $ret = self::$languageNameCache->get( $cacheKey );
 844                 if ( !$ret ) {
 845                         $ret = self::fetchLanguageNamesUncached( $inLanguage, $include );
 846                         self::$languageNameCache->set( $cacheKey, $ret );
 847                 }
 848                 return $ret;
 849         }
 850
 851         /**
 852          * Uncached helper for fetchLanguageNames
 853          * @param null|string $inLanguage Code of language in which to return the names
 854          *              Use self::AS_AUTONYMS for autonyms (native names)
 855          * @param string $include One of:
 856          *              self::ALL all available languages
 857          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 858          *              self::SUPPORTED only if the language is in 'mw' *and* has a message file
 859          * @return array Language code => language name (sorted by key)
 860          */
 861         private static function fetchLanguageNamesUncached(
 862                 $inLanguage = self::AS_AUTONYMS,
 863                 $include = 'mw'
 864         ) {
 865                 global $wgExtraLanguageNames, $wgUsePigLatinVariant;
 866
 867                 // If passed an invalid language code to use, fallback to en
 868                 if ( $inLanguage !== self::AS_AUTONYMS && !self::isValidCode( $inLanguage ) ) {
 869                         $inLanguage = 'en';
 870                 }
 871
 872                 $names = [];
 873
 874                 if ( $inLanguage ) {
 875                         # TODO: also include when $inLanguage is null, when this code is more efficient
 876                         Hooks::run( 'LanguageGetTranslatedLanguageNames', [ &$names, $inLanguage ] );
 877                 }
 878
 879                 $mwNames = $wgExtraLanguageNames + MediaWiki\Languages\Data\Names::$names;
 880                 if ( $wgUsePigLatinVariant ) {
 881                         // Pig Latin (for variant development)
 882                         $mwNames['en-x-piglatin'] = 'Igpay Atinlay';
 883                 }
 884
 885                 foreach ( $mwNames as $mwCode => $mwName ) {
 886                         # - Prefer own MediaWiki native name when not using the hook
 887                         # - For other names just add if not added through the hook
 888                         if ( $mwCode === $inLanguage || !isset( $names[$mwCode] ) ) {
 889                                 $names[$mwCode] = $mwName;
 890                         }
 891                 }
 892
 893                 if ( $include === self::ALL ) {
 894                         ksort( $names );
 895                         return $names;
 896                 }
 897
 898                 $returnMw = [];
 899                 $coreCodes = array_keys( $mwNames );
 900                 foreach ( $coreCodes as $coreCode ) {
 901                         $returnMw[$coreCode] = $names[$coreCode];
 902                 }
 903
 904                 if ( $include === self::SUPPORTED ) {
 905                         $namesMwFile = [];
 906                         # We do this using a foreach over the codes instead of a directory
 907                         # loop so that messages files in extensions will work correctly.
 908                         foreach ( $returnMw as $code => $value ) {
 909                                 if ( is_readable( self::getMessagesFileName( $code ) )
 910                                         || is_readable( self::getJsonMessagesFileName( $code ) )
 911                                 ) {
 912                                         $namesMwFile[$code] = $names[$code];
 913                                 }
 914                         }
 915
 916                         ksort( $namesMwFile );
 917                         return $namesMwFile;
 918                 }
 919
 920                 ksort( $returnMw );
 921                 # 'mw' option; default if it's not one of the other two options (all/mwfile)
 922                 return $returnMw;
 923         }
 924
 925         /**
 926          * @param string $code The code of the language for which to get the name
 927          * @param null|string $inLanguage Code of language in which to return the name
 928          *   (SELF::AS_AUTONYMS for autonyms)
 929          * @param string $include See fetchLanguageNames()
 930          * @return string Language name or empty
 931          * @since 1.20
 932          */
 933         public static function fetchLanguageName(
 934                 $code,
 935                 $inLanguage = self::AS_AUTONYMS,
 936                 $include = self::ALL
 937         ) {
 938                 $code = strtolower( $code );
 939                 $array = self::fetchLanguageNames( $inLanguage, $include );
 940                 return !array_key_exists( $code, $array ) ? '' : $array[$code];
 941         }
 942
 943         /**
 944          * Get a message from the MediaWiki namespace.
 945          *
 946          * @param string $msg Message name
 947          * @return string
 948          */
 949         public function getMessageFromDB( $msg ) {
 950                 return $this->msg( $msg )->text();
 951         }
 952
 953         /**
 954          * Get message object in this language. Only for use inside this class.
 955          *
 956          * @param string $msg Message name
 957          * @return Message
 958          */
 959         protected function msg( $msg ) {
 960                 return wfMessage( $msg )->inLanguage( $this );
 961         }
 962
 963         /**
 964          * @param string $key
 965          * @return string
 966          */
 967         public function getMonthName( $key ) {
 968                 return $this->getMessageFromDB( self::$mMonthMsgs[$key - 1] );
 969         }
 970
 971         /**
 972          * @return array
 973          */
 974         public function getMonthNamesArray() {
 975                 $monthNames = [ '' ];
 976                 for ( $i = 1; $i < 13; $i++ ) {
 977                         $monthNames[] = $this->getMonthName( $i );
 978                 }
 979                 return $monthNames;
 980         }
 981
 982         /**
 983          * @param string $key
 984          * @return string
 985          */
 986         public function getMonthNameGen( $key ) {
 987                 return $this->getMessageFromDB( self::$mMonthGenMsgs[$key - 1] );
 988         }
 989
 990         /**
 991          * @param string $key
 992          * @return string
 993          */
 994         public function getMonthAbbreviation( $key ) {
 995                 return $this->getMessageFromDB( self::$mMonthAbbrevMsgs[$key - 1] );
 996         }
 997
 998         /**
 999          * @return array
1000          */
1001         public function getMonthAbbreviationsArray() {
1002                 $monthNames = [ '' ];
1003                 for ( $i = 1; $i < 13; $i++ ) {
1004                         $monthNames[] = $this->getMonthAbbreviation( $i );
1005                 }
1006                 return $monthNames;
1007         }
1008
1009         /**
1010          * @param string $key
1011          * @return string
1012          */
1013         public function getWeekdayName( $key ) {
1014                 return $this->getMessageFromDB( self::$mWeekdayMsgs[$key - 1] );
1015         }
1016
1017         /**
1018          * @param string $key
1019          * @return string
1020          */
1021         function getWeekdayAbbreviation( $key ) {
1022                 return $this->getMessageFromDB( self::$mWeekdayAbbrevMsgs[$key - 1] );
1023         }
1024
1025         /**
1026          * @param string $key
1027          * @return string
1028          */
1029         function getIranianCalendarMonthName( $key ) {
1030                 return $this->getMessageFromDB( self::$mIranianCalendarMonthMsgs[$key - 1] );
1031         }
1032
1033         /**
1034          * @param string $key
1035          * @return string
1036          */
1037         function getHebrewCalendarMonthName( $key ) {
1038                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthMsgs[$key - 1] );
1039         }
1040
1041         /**
1042          * @param string $key
1043          * @return string
1044          */
1045         function getHebrewCalendarMonthNameGen( $key ) {
1046                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthGenMsgs[$key - 1] );
1047         }
1048
1049         /**
1050          * @param string $key
1051          * @return string
1052          */
1053         function getHijriCalendarMonthName( $key ) {
1054                 return $this->getMessageFromDB( self::$mHijriCalendarMonthMsgs[$key - 1] );
1055         }
1056
1057         /**
1058          * Pass through result from $dateTimeObj->format()
1059          * @param DateTime|bool|null &$dateTimeObj
1060          * @param string $ts
1061          * @param DateTimeZone|bool|null $zone
1062          * @param string $code
1063          * @return string
1064          */
1065         private static function dateTimeObjFormat( &$dateTimeObj, $ts, $zone, $code ) {
1066                 if ( !$dateTimeObj ) {
1067                         $dateTimeObj = DateTime::createFromFormat(
1068                                 'YmdHis', $ts, $zone ?: new DateTimeZone( 'UTC' )
1069                         );
1070                 }
1071                 return $dateTimeObj->format( $code );
1072         }
1073
1074         /**
1075          * This is a workalike of PHP's date() function, but with better
1076          * internationalisation, a reduced set of format characters, and a better
1077          * escaping format.
1078          *
1079          * Supported format characters are dDjlNwzWFmMntLoYyaAgGhHiscrUeIOPTZ. See
1080          * the PHP manual for definitions. There are a number of extensions, which
1081          * start with "x":
1082          *
1083          *    xn   Do not translate digits of the next numeric format character
1084          *    xN   Toggle raw digit (xn) flag, stays set until explicitly unset
1085          *    xr   Use roman numerals for the next numeric format character
1086          *    xh   Use hebrew numerals for the next numeric format character
1087          *    xx   Literal x
1088          *    xg   Genitive month name
1089          *
1090          *    xij  j (day number) in Iranian calendar
1091          *    xiF  F (month name) in Iranian calendar
1092          *    xin  n (month number) in Iranian calendar
1093          *    xiy  y (two digit year) in Iranian calendar
1094          *    xiY  Y (full year) in Iranian calendar
1095          *    xit  t (days in month) in Iranian calendar
1096          *    xiz  z (day of the year) in Iranian calendar
1097          *
1098          *    xjj  j (day number) in Hebrew calendar
1099          *    xjF  F (month name) in Hebrew calendar
1100          *    xjt  t (days in month) in Hebrew calendar
1101          *    xjx  xg (genitive month name) in Hebrew calendar
1102          *    xjn  n (month number) in Hebrew calendar
1103          *    xjY  Y (full year) in Hebrew calendar
1104          *
1105          *    xmj  j (day number) in Hijri calendar
1106          *    xmF  F (month name) in Hijri calendar
1107          *    xmn  n (month number) in Hijri calendar
1108          *    xmY  Y (full year) in Hijri calendar
1109          *
1110          *    xkY  Y (full year) in Thai solar calendar. Months and days are
1111          *                       identical to the Gregorian calendar
1112          *    xoY  Y (full year) in Minguo calendar or Juche year.
1113          *                       Months and days are identical to the
1114          *                       Gregorian calendar
1115          *    xtY  Y (full year) in Japanese nengo. Months and days are
1116          *                       identical to the Gregorian calendar
1117          *
1118          * Characters enclosed in double quotes will be considered literal (with
1119          * the quotes themselves removed). Unmatched quotes will be considered
1120          * literal quotes. Example:
1121          *
1122          * "The month is" F       => The month is January
1123          * i's"                   => 20'11"
1124          *
1125          * Backslash escaping is also supported.
1126          *
1127          * Input timestamp is assumed to be pre-normalized to the desired local
1128          * time zone, if any. Note that the format characters crUeIOPTZ will assume
1129          * $ts is UTC if $zone is not given.
1130          *
1131          * @param string $format
1132          * @param string $ts 14-character timestamp
1133          *      YYYYMMDDHHMMSS
1134          *      01234567890123
1135          * @param DateTimeZone|null $zone Timezone of $ts
1136          * @param int &$ttl The amount of time (in seconds) the output may be cached for.
1137          * Only makes sense if $ts is the current time.
1138          * @todo handling of "o" format character for Iranian, Hebrew, Hijri & Thai?
1139          *
1140          * @throws MWException
1141          * @return string
1142          */
1143         public function sprintfDate( $format, $ts, DateTimeZone $zone = null, &$ttl = 'unused' ) {
1144                 $s = '';
1145                 $raw = false;
1146                 $roman = false;
1147                 $hebrewNum = false;
1148                 $dateTimeObj = false;
1149                 $rawToggle = false;
1150                 $iranian = false;
1151                 $hebrew = false;
1152                 $hijri = false;
1153                 $thai = false;
1154                 $minguo = false;
1155                 $tenno = false;
1156
1157                 $usedSecond = false;
1158                 $usedMinute = false;
1159                 $usedHour = false;
1160                 $usedAMPM = false;
1161                 $usedDay = false;
1162                 $usedWeek = false;
1163                 $usedMonth = false;
1164                 $usedYear = false;
1165                 $usedISOYear = false;
1166                 $usedIsLeapYear = false;
1167
1168                 $usedHebrewMonth = false;
1169                 $usedIranianMonth = false;
1170                 $usedHijriMonth = false;
1171                 $usedHebrewYear = false;
1172                 $usedIranianYear = false;
1173                 $usedHijriYear = false;
1174                 $usedTennoYear = false;
1175
1176                 if ( strlen( $ts ) !== 14 ) {
1177                         throw new MWException( __METHOD__ . ": The timestamp $ts should have 14 characters" );
1178                 }
1179
1180                 if ( !ctype_digit( $ts ) ) {
1181                         throw new MWException( __METHOD__ . ": The timestamp $ts should be a number" );
1182                 }
1183
1184                 $formatLength = strlen( $format );
1185                 for ( $p = 0; $p < $formatLength; $p++ ) {
1186                         $num = false;
1187                         $code = $format[$p];
1188                         if ( $code == 'x' && $p < $formatLength - 1 ) {
1189                                 $code .= $format[++$p];
1190                         }
1191
1192                         if ( ( $code === 'xi'
1193                                         || $code === 'xj'
1194                                         || $code === 'xk'
1195                                         || $code === 'xm'
1196                                         || $code === 'xo'
1197                                         || $code === 'xt' )
1198                                 && $p < $formatLength - 1 ) {
1199                                 $code .= $format[++$p];
1200                         }
1201
1202                         switch ( $code ) {
1203                                 case 'xx':
1204                                         $s .= 'x';
1205                                         break;
1206                                 case 'xn':
1207                                         $raw = true;
1208                                         break;
1209                                 case 'xN':
1210                                         $rawToggle = !$rawToggle;
1211                                         break;
1212                                 case 'xr':
1213                                         $roman = true;
1214                                         break;
1215                                 case 'xh':
1216                                         $hebrewNum = true;
1217                                         break;
1218                                 case 'xg':
1219                                         $usedMonth = true;
1220                                         $s .= $this->getMonthNameGen( substr( $ts, 4, 2 ) );
1221                                         break;
1222                                 case 'xjx':
1223                                         $usedHebrewMonth = true;
1224                                         if ( !$hebrew ) {
1225                                                 $hebrew = self::tsToHebrew( $ts );
1226                                         }
1227                                         $s .= $this->getHebrewCalendarMonthNameGen( $hebrew[1] );
1228                                         break;
1229                                 case 'd':
1230                                         $usedDay = true;
1231                                         $num = substr( $ts, 6, 2 );
1232                                         break;
1233                                 case 'D':
1234                                         $usedDay = true;
1235                                         $s .= $this->getWeekdayAbbreviation(
1236                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1237                                         );
1238                                         break;
1239                                 case 'j':
1240                                         $usedDay = true;
1241                                         $num = intval( substr( $ts, 6, 2 ) );
1242                                         break;
1243                                 case 'xij':
1244                                         $usedDay = true;
1245                                         if ( !$iranian ) {
1246                                                 $iranian = self::tsToIranian( $ts );
1247                                         }
1248                                         $num = $iranian[2];
1249                                         break;
1250                                 case 'xmj':
1251                                         $usedDay = true;
1252                                         if ( !$hijri ) {
1253                                                 $hijri = self::tsToHijri( $ts );
1254                                         }
1255                                         $num = $hijri[2];
1256                                         break;
1257                                 case 'xjj':
1258                                         $usedDay = true;
1259                                         if ( !$hebrew ) {
1260                                                 $hebrew = self::tsToHebrew( $ts );
1261                                         }
1262                                         $num = $hebrew[2];
1263                                         break;
1264                                 case 'l':
1265                                         $usedDay = true;
1266                                         $s .= $this->getWeekdayName(
1267                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1268                                         );
1269                                         break;
1270                                 case 'F':
1271                                         $usedMonth = true;
1272                                         $s .= $this->getMonthName( substr( $ts, 4, 2 ) );
1273                                         break;
1274                                 case 'xiF':
1275                                         $usedIranianMonth = true;
1276                                         if ( !$iranian ) {
1277                                                 $iranian = self::tsToIranian( $ts );
1278                                         }
1279                                         $s .= $this->getIranianCalendarMonthName( $iranian[1] );
1280                                         break;
1281                                 case 'xmF':
1282                                         $usedHijriMonth = true;
1283                                         if ( !$hijri ) {
1284                                                 $hijri = self::tsToHijri( $ts );
1285                                         }
1286                                         $s .= $this->getHijriCalendarMonthName( $hijri[1] );
1287                                         break;
1288                                 case 'xjF':
1289                                         $usedHebrewMonth = true;
1290                                         if ( !$hebrew ) {
1291                                                 $hebrew = self::tsToHebrew( $ts );
1292                                         }
1293                                         $s .= $this->getHebrewCalendarMonthName( $hebrew[1] );
1294                                         break;
1295                                 case 'm':
1296                                         $usedMonth = true;
1297                                         $num = substr( $ts, 4, 2 );
1298                                         break;
1299                                 case 'M':
1300                                         $usedMonth = true;
1301                                         $s .= $this->getMonthAbbreviation( substr( $ts, 4, 2 ) );
1302                                         break;
1303                                 case 'n':
1304                                         $usedMonth = true;
1305                                         $num = intval( substr( $ts, 4, 2 ) );
1306                                         break;
1307                                 case 'xin':
1308                                         $usedIranianMonth = true;
1309                                         if ( !$iranian ) {
1310                                                 $iranian = self::tsToIranian( $ts );
1311                                         }
1312                                         $num = $iranian[1];
1313                                         break;
1314                                 case 'xmn':
1315                                         $usedHijriMonth = true;
1316                                         if ( !$hijri ) {
1317                                                 $hijri = self::tsToHijri( $ts );
1318                                         }
1319                                         $num = $hijri[1];
1320                                         break;
1321                                 case 'xjn':
1322                                         $usedHebrewMonth = true;
1323                                         if ( !$hebrew ) {
1324                                                 $hebrew = self::tsToHebrew( $ts );
1325                                         }
1326                                         $num = $hebrew[1];
1327                                         break;
1328                                 case 'xjt':
1329                                         $usedHebrewMonth = true;
1330                                         if ( !$hebrew ) {
1331                                                 $hebrew = self::tsToHebrew( $ts );
1332                                         }
1333                                         $num = $hebrew[3];
1334                                         break;
1335                                 case 'Y':
1336                                         $usedYear = true;
1337                                         $num = substr( $ts, 0, 4 );
1338                                         break;
1339                                 case 'xiY':
1340                                         $usedIranianYear = true;
1341                                         if ( !$iranian ) {
1342                                                 $iranian = self::tsToIranian( $ts );
1343                                         }
1344                                         $num = $iranian[0];
1345                                         break;
1346                                 case 'xmY':
1347                                         $usedHijriYear = true;
1348                                         if ( !$hijri ) {
1349                                                 $hijri = self::tsToHijri( $ts );
1350                                         }
1351                                         $num = $hijri[0];
1352                                         break;
1353                                 case 'xjY':
1354                                         $usedHebrewYear = true;
1355                                         if ( !$hebrew ) {
1356                                                 $hebrew = self::tsToHebrew( $ts );
1357                                         }
1358                                         $num = $hebrew[0];
1359                                         break;
1360                                 case 'xkY':
1361                                         $usedYear = true;
1362                                         if ( !$thai ) {
1363                                                 $thai = self::tsToYear( $ts, 'thai' );
1364                                         }
1365                                         $num = $thai[0];
1366                                         break;
1367                                 case 'xoY':
1368                                         $usedYear = true;
1369                                         if ( !$minguo ) {
1370                                                 $minguo = self::tsToYear( $ts, 'minguo' );
1371                                         }
1372                                         $num = $minguo[0];
1373                                         break;
1374                                 case 'xtY':
1375                                         $usedTennoYear = true;
1376                                         if ( !$tenno ) {
1377                                                 $tenno = self::tsToYear( $ts, 'tenno' );
1378                                         }
1379                                         $num = $tenno[0];
1380                                         break;
1381                                 case 'y':
1382                                         $usedYear = true;
1383                                         $num = substr( $ts, 2, 2 );
1384                                         break;
1385                                 case 'xiy':
1386                                         $usedIranianYear = true;
1387                                         if ( !$iranian ) {
1388                                                 $iranian = self::tsToIranian( $ts );
1389                                         }
1390                                         $num = substr( $iranian[0], -2 );
1391                                         break;
1392                                 case 'xit':
1393                                         $usedIranianYear = true;
1394                                         if ( !$iranian ) {
1395                                                 $iranian = self::tsToIranian( $ts );
1396                                         }
1397                                         $num = self::$IRANIAN_DAYS[$iranian[1] - 1];
1398                                         break;
1399                                 case 'xiz':
1400                                         $usedIranianYear = true;
1401                                         if ( !$iranian ) {
1402                                                 $iranian = self::tsToIranian( $ts );
1403                                         }
1404                                         $num = $iranian[3];
1405                                         break;
1406                                 case 'a':
1407                                         $usedAMPM = true;
1408                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'am' : 'pm';
1409                                         break;
1410                                 case 'A':
1411                                         $usedAMPM = true;
1412                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'AM' : 'PM';
1413                                         break;
1414                                 case 'g':
1415                                         $usedHour = true;
1416                                         $h = substr( $ts, 8, 2 );
1417                                         $num = $h % 12 ?: 12;
1418                                         break;
1419                                 case 'G':
1420                                         $usedHour = true;
1421                                         $num = intval( substr( $ts, 8, 2 ) );
1422                                         break;
1423                                 case 'h':
1424                                         $usedHour = true;
1425                                         $h = substr( $ts, 8, 2 );
1426                                         $num = sprintf( '%02d', $h % 12 ?: 12 );
1427                                         break;
1428                                 case 'H':
1429                                         $usedHour = true;
1430                                         $num = substr( $ts, 8, 2 );
1431                                         break;
1432                                 case 'i':
1433                                         $usedMinute = true;
1434                                         $num = substr( $ts, 10, 2 );
1435                                         break;
1436                                 case 's':
1437                                         $usedSecond = true;
1438                                         $num = substr( $ts, 12, 2 );
1439                                         break;
1440                                 case 'c':
1441                                 case 'r':
1442                                         $usedSecond = true;
1443                                         // fall through
1444                                 case 'e':
1445                                 case 'O':
1446                                 case 'P':
1447                                 case 'T':
1448                                         $s .= self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1449                                         break;
1450                                 case 'w':
1451                                 case 'N':
1452                                 case 'z':
1453                                         $usedDay = true;
1454                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1455                                         break;
1456                                 case 'W':
1457                                         $usedWeek = true;
1458                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1459                                         break;
1460                                 case 't':
1461                                         $usedMonth = true;
1462                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1463                                         break;
1464                                 case 'L':
1465                                         $usedIsLeapYear = true;
1466                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1467                                         break;
1468                                 case 'o':
1469                                         $usedISOYear = true;
1470                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1471                                         break;
1472                                 case 'U':
1473                                         $usedSecond = true;
1474                                         // fall through
1475                                 case 'I':
1476                                 case 'Z':
1477                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1478                                         break;
1479                                 case '\\':
1480                                         # Backslash escaping
1481                                         if ( $p < $formatLength - 1 ) {
1482                                                 $s .= $format[++$p];
1483                                         } else {
1484                                                 $s .= '\\';
1485                                         }
1486                                         break;
1487                                 case '"':
1488                                         # Quoted literal
1489                                         if ( $p < $formatLength - 1 ) {
1490                                                 $endQuote = strpos( $format, '"', $p + 1 );
1491                                                 if ( $endQuote === false ) {
1492                                                         # No terminating quote, assume literal "
1493                                                         $s .= '"';
1494                                                 } else {
1495                                                         $s .= substr( $format, $p + 1, $endQuote - $p - 1 );
1496                                                         $p = $endQuote;
1497                                                 }
1498                                         } else {
1499                                                 # Quote at end of string, assume literal "
1500                                                 $s .= '"';
1501                                         }
1502                                         break;
1503                                 default:
1504                                         $s .= $format[$p];
1505                         }
1506                         if ( $num !== false ) {
1507                                 if ( $rawToggle || $raw ) {
1508                                         $s .= $num;
1509                                         $raw = false;
1510                                 } elseif ( $roman ) {
1511                                         $s .= self::romanNumeral( $num );
1512                                         $roman = false;
1513                                 } elseif ( $hebrewNum ) {
1514                                         $s .= self::hebrewNumeral( $num );
1515                                         $hebrewNum = false;
1516                                 } else {
1517                                         $s .= $this->formatNum( $num, true );
1518                                 }
1519                         }
1520                 }
1521
1522                 if ( $ttl === 'unused' ) {
1523                         // No need to calculate the TTL, the caller wont use it anyway.
1524                 } elseif ( $usedSecond ) {
1525                         $ttl = 1;
1526                 } elseif ( $usedMinute ) {
1527                         $ttl = 60 - substr( $ts, 12, 2 );
1528                 } elseif ( $usedHour ) {
1529                         $ttl = 3600 - substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1530                 } elseif ( $usedAMPM ) {
1531                         $ttl = 43200 - ( substr( $ts, 8, 2 ) % 12 ) * 3600 -
1532                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1533                 } elseif (
1534                         $usedDay ||
1535                         $usedHebrewMonth ||
1536                         $usedIranianMonth ||
1537                         $usedHijriMonth ||
1538                         $usedHebrewYear ||
1539                         $usedIranianYear ||
1540                         $usedHijriYear ||
1541                         $usedTennoYear
1542                 ) {
1543                         // @todo Someone who understands the non-Gregorian calendars
1544                         // should write proper logic for them so that they don't need purged every day.
1545                         $ttl = 86400 - substr( $ts, 8, 2 ) * 3600 -
1546                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1547                 } else {
1548                         $possibleTtls = [];
1549                         $timeRemainingInDay = 86400 - substr( $ts, 8, 2 ) * 3600 -
1550                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1551                         if ( $usedWeek ) {
1552                                 $possibleTtls[] =
1553                                         ( 7 - self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400 +
1554                                         $timeRemainingInDay;
1555                         } elseif ( $usedISOYear ) {
1556                                 // December 28th falls on the last ISO week of the year, every year.
1557                                 // The last ISO week of a year can be 52 or 53.
1558                                 $lastWeekOfISOYear = DateTime::createFromFormat(
1559                                         'Ymd',
1560                                         substr( $ts, 0, 4 ) . '1228',
1561                                         $zone ?: new DateTimeZone( 'UTC' )
1562                                 )->format( 'W' );
1563                                 $currentISOWeek = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'W' );
1564                                 $weeksRemaining = $lastWeekOfISOYear - $currentISOWeek;
1565                                 $timeRemainingInWeek =
1566                                         ( 7 - self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400
1567                                         + $timeRemainingInDay;
1568                                 $possibleTtls[] = $weeksRemaining * 604800 + $timeRemainingInWeek;
1569                         }
1570
1571                         if ( $usedMonth ) {
1572                                 $possibleTtls[] =
1573                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 't' ) -
1574                                                 substr( $ts, 6, 2 ) ) * 86400
1575                                         + $timeRemainingInDay;
1576                         } elseif ( $usedYear ) {
1577                                 $possibleTtls[] =
1578                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1579                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1580                                         + $timeRemainingInDay;
1581                         } elseif ( $usedIsLeapYear ) {
1582                                 $year = substr( $ts, 0, 4 );
1583                                 $timeRemainingInYear =
1584                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1585                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1586                                         + $timeRemainingInDay;
1587                                 $mod = $year % 4;
1588                                 if ( $mod || ( !( $year % 100 ) && $year % 400 ) ) {
1589                                         // this isn't a leap year. see when the next one starts
1590                                         $nextCandidate = $year - $mod + 4;
1591                                         if ( $nextCandidate % 100 || !( $nextCandidate % 400 ) ) {
1592                                                 $possibleTtls[] = ( $nextCandidate - $year - 1 ) * 365 * 86400 +
1593                                                         $timeRemainingInYear;
1594                                         } else {
1595                                                 $possibleTtls[] = ( $nextCandidate - $year + 3 ) * 365 * 86400 +
1596                                                         $timeRemainingInYear;
1597                                         }
1598                                 } else {
1599                                         // this is a leap year, so the next year isn't
1600                                         $possibleTtls[] = $timeRemainingInYear;
1601                                 }
1602                         }
1603
1604                         if ( $possibleTtls ) {
1605                                 $ttl = min( $possibleTtls );
1606                         }
1607                 }
1608
1609                 return $s;
1610         }
1611
1612         private static $GREG_DAYS = [ 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 ];
1613         private static $IRANIAN_DAYS = [ 31, 31, 31, 31, 31, 31, 30, 30, 30, 30, 30, 29 ];
1614
1615         /**
1616          * Algorithm by Roozbeh Pournader and Mohammad Toossi to convert
1617          * Gregorian dates to Iranian dates. Originally written in C, it
1618          * is released under the terms of GNU Lesser General Public
1619          * License. Conversion to PHP was performed by Niklas Laxström.
1620          *
1621          * Link: http://www.farsiweb.info/jalali/jalali.c
1622          *
1623          * @param string $ts
1624          *
1625          * @return int[]
1626          */
1627         private static function tsToIranian( $ts ) {
1628                 $gy = substr( $ts, 0, 4 ) - 1600;
1629                 $gm = substr( $ts, 4, 2 ) - 1;
1630                 $gd = substr( $ts, 6, 2 ) - 1;
1631
1632                 # Days passed from the beginning (including leap years)
1633                 $gDayNo = 365 * $gy
1634                         + floor( ( $gy + 3 ) / 4 )
1635                         - floor( ( $gy + 99 ) / 100 )
1636                         + floor( ( $gy + 399 ) / 400 );
1637
1638                 // Add days of the past months of this year
1639                 for ( $i = 0; $i < $gm; $i++ ) {
1640                         $gDayNo += self::$GREG_DAYS[$i];
1641                 }
1642
1643                 // Leap years
1644                 if ( $gm > 1 && ( ( $gy % 4 === 0 && $gy % 100 !== 0 || ( $gy % 400 == 0 ) ) ) ) {
1645                         $gDayNo++;
1646                 }
1647
1648                 // Days passed in current month
1649                 $gDayNo += (int)$gd;
1650
1651                 $jDayNo = $gDayNo - 79;
1652
1653                 $jNp = floor( $jDayNo / 12053 );
1654                 $jDayNo %= 12053;
1655
1656                 $jy = 979 + 33 * $jNp + 4 * floor( $jDayNo / 1461 );
1657                 $jDayNo %= 1461;
1658
1659                 if ( $jDayNo >= 366 ) {
1660                         $jy += floor( ( $jDayNo - 1 ) / 365 );
1661                         $jDayNo = floor( ( $jDayNo - 1 ) % 365 );
1662                 }
1663
1664                 $jz = $jDayNo;
1665
1666                 for ( $i = 0; $i < 11 && $jDayNo >= self::$IRANIAN_DAYS[$i]; $i++ ) {
1667                         $jDayNo -= self::$IRANIAN_DAYS[$i];
1668                 }
1669
1670                 $jm = $i + 1;
1671                 $jd = $jDayNo + 1;
1672
1673                 return [ $jy, $jm, $jd, $jz ];
1674         }
1675
1676         /**
1677          * Converting Gregorian dates to Hijri dates.
1678          *
1679          * Based on a PHP-Nuke block by Sharjeel which is released under GNU/GPL license
1680          *
1681          * @see https://phpnuke.org/modules.php?name=News&file=article&sid=8234&mode=thread&order=0&thold=0
1682          *
1683          * @param string $ts
1684          *
1685          * @return int[]
1686          */
1687         private static function tsToHijri( $ts ) {
1688                 $year = substr( $ts, 0, 4 );
1689                 $month = substr( $ts, 4, 2 );
1690                 $day = substr( $ts, 6, 2 );
1691
1692                 $zyr = $year;
1693                 $zd = $day;
1694                 $zm = $month;
1695                 $zy = $zyr;
1696
1697                 if (
1698                         ( $zy > 1582 ) || ( ( $zy == 1582 ) && ( $zm > 10 ) ) ||
1699                         ( ( $zy == 1582 ) && ( $zm == 10 ) && ( $zd > 14 ) )
1700                 ) {
1701                         $zjd = (int)( ( 1461 * ( $zy + 4800 + (int)( ( $zm - 14 ) / 12 ) ) ) / 4 ) +
1702                                         (int)( ( 367 * ( $zm - 2 - 12 * ( (int)( ( $zm - 14 ) / 12 ) ) ) ) / 12 ) -
1703                                         (int)( ( 3 * (int)( ( ( $zy + 4900 + (int)( ( $zm - 14 ) / 12 ) ) / 100 ) ) ) / 4 ) +
1704                                         $zd - 32075;
1705                 } else {
1706                         $zjd = 367 * $zy - (int)( ( 7 * ( $zy + 5001 + (int)( ( $zm - 9 ) / 7 ) ) ) / 4 ) +
1707                                                                 (int)( ( 275 * $zm ) / 9 ) + $zd + 1729777;
1708                 }
1709
1710                 $zl = $zjd - 1948440 + 10632;
1711                 $zn = (int)( ( $zl - 1 ) / 10631 );
1712                 $zl = $zl - 10631 * $zn + 354;
1713                 $zj = ( (int)( ( 10985 - $zl ) / 5316 ) ) * ( (int)( ( 50 * $zl ) / 17719 ) ) +
1714                         ( (int)( $zl / 5670 ) ) * ( (int)( ( 43 * $zl ) / 15238 ) );
1715                 $zl = $zl - ( (int)( ( 30 - $zj ) / 15 ) ) * ( (int)( ( 17719 * $zj ) / 50 ) ) -
1716                         ( (int)( $zj / 16 ) ) * ( (int)( ( 15238 * $zj ) / 43 ) ) + 29;
1717                 $zm = (int)( ( 24 * $zl ) / 709 );
1718                 $zd = $zl - (int)( ( 709 * $zm ) / 24 );
1719                 $zy = 30 * $zn + $zj - 30;
1720
1721                 return [ $zy, $zm, $zd ];
1722         }
1723
1724         /**
1725          * Converting Gregorian dates to Hebrew dates.
1726          *
1727          * Based on a JavaScript code by Abu Mami and Yisrael Hersch
1728          * (abu-mami@kaluach.net, http://www.kaluach.net), who permitted
1729          * to translate the relevant functions into PHP and release them under
1730          * GNU GPL.
1731          *
1732          * The months are counted from Tishrei = 1. In a leap year, Adar I is 13
1733          * and Adar II is 14. In a non-leap year, Adar is 6.
1734          *
1735          * @param string $ts
1736          *
1737          * @return int[]
1738          */
1739         private static function tsToHebrew( $ts ) {
1740                 # Parse date
1741                 $year = substr( $ts, 0, 4 );
1742                 $month = substr( $ts, 4, 2 );
1743                 $day = substr( $ts, 6, 2 );
1744
1745                 # Calculate Hebrew year
1746                 $hebrewYear = $year + 3760;
1747
1748                 # Month number when September = 1, August = 12
1749                 $month += 4;
1750                 if ( $month > 12 ) {
1751                         # Next year
1752                         $month -= 12;
1753                         $year++;
1754                         $hebrewYear++;
1755                 }
1756
1757                 # Calculate day of year from 1 September
1758                 $dayOfYear = $day;
1759                 for ( $i = 1; $i < $month; $i++ ) {
1760                         if ( $i == 6 ) {
1761                                 # February
1762                                 $dayOfYear += 28;
1763                                 # Check if the year is leap
1764                                 if ( $year % 400 == 0 || ( $year % 4 == 0 && $year % 100 > 0 ) ) {
1765                                         $dayOfYear++;
1766                                 }
1767                         } elseif ( $i == 8 || $i == 10 || $i == 1 || $i == 3 ) {
1768                                 $dayOfYear += 30;
1769                         } else {
1770                                 $dayOfYear += 31;
1771                         }
1772                 }
1773
1774                 # Calculate the start of the Hebrew year
1775                 $start = self::hebrewYearStart( $hebrewYear );
1776
1777                 # Calculate next year's start
1778                 if ( $dayOfYear <= $start ) {
1779                         # Day is before the start of the year - it is the previous year
1780                         # Next year's start
1781                         $nextStart = $start;
1782                         # Previous year
1783                         $year--;
1784                         $hebrewYear--;
1785                         # Add days since previous year's 1 September
1786                         $dayOfYear += 365;
1787                         if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1788                                 # Leap year
1789                                 $dayOfYear++;
1790                         }
1791                         # Start of the new (previous) year
1792                         $start = self::hebrewYearStart( $hebrewYear );
1793                 } else {
1794                         # Next year's start
1795                         $nextStart = self::hebrewYearStart( $hebrewYear + 1 );
1796                 }
1797
1798                 # Calculate Hebrew day of year
1799                 $hebrewDayOfYear = $dayOfYear - $start;
1800
1801                 # Difference between year's days
1802                 $diff = $nextStart - $start;
1803                 # Add 12 (or 13 for leap years) days to ignore the difference between
1804                 # Hebrew and Gregorian year (353 at least vs. 365/6) - now the
1805                 # difference is only about the year type
1806                 if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1807                         $diff += 13;
1808                 } else {
1809                         $diff += 12;
1810                 }
1811
1812                 # Check the year pattern, and is leap year
1813                 # 0 means an incomplete year, 1 means a regular year, 2 means a complete year
1814                 # This is mod 30, to work on both leap years (which add 30 days of Adar I)
1815                 # and non-leap years
1816                 $yearPattern = $diff % 30;
1817                 # Check if leap year
1818                 $isLeap = $diff >= 30;
1819
1820                 # Calculate day in the month from number of day in the Hebrew year
1821                 # Don't check Adar - if the day is not in Adar, we will stop before;
1822                 # if it is in Adar, we will use it to check if it is Adar I or Adar II
1823                 $hebrewDay = $hebrewDayOfYear;
1824                 $hebrewMonth = 1;
1825                 $days = 0;
1826                 while ( $hebrewMonth <= 12 ) {
1827                         # Calculate days in this month
1828                         if ( $isLeap && $hebrewMonth == 6 ) {
1829                                 # Leap year - has Adar I, with 30 days, and Adar II, with 29 days
1830                                 $days = 30;
1831                                 if ( $hebrewDay <= $days ) {
1832                                         # Day in Adar I
1833                                         $hebrewMonth = 13;
1834                                 } else {
1835                                         # Subtract the days of Adar I
1836                                         $hebrewDay -= $days;
1837                                         # Try Adar II
1838                                         $days = 29;
1839                                         if ( $hebrewDay <= $days ) {
1840                                                 # Day in Adar II
1841                                                 $hebrewMonth = 14;
1842                                         }
1843                                 }
1844                         } elseif ( $hebrewMonth == 2 && $yearPattern == 2 ) {
1845                                 # Cheshvan in a complete year (otherwise as the rule below)
1846                                 $days = 30;
1847                         } elseif ( $hebrewMonth == 3 && $yearPattern == 0 ) {
1848                                 # Kislev in an incomplete year (otherwise as the rule below)
1849                                 $days = 29;
1850                         } else {
1851                                 # Odd months have 30 days, even have 29
1852                                 $days = 30 - ( $hebrewMonth - 1 ) % 2;
1853                         }
1854                         if ( $hebrewDay <= $days ) {
1855                                 # In the current month
1856                                 break;
1857                         } else {
1858                                 # Subtract the days of the current month
1859                                 $hebrewDay -= $days;
1860                                 # Try in the next month
1861                                 $hebrewMonth++;
1862                         }
1863                 }
1864
1865                 return [ $hebrewYear, $hebrewMonth, $hebrewDay, $days ];
1866         }
1867
1868         /**
1869          * This calculates the Hebrew year start, as days since 1 September.
1870          * Based on Carl Friedrich Gauss algorithm for finding Easter date.
1871          * Used for Hebrew date.
1872          *
1873          * @param int $year
1874          *
1875          * @return string
1876          */
1877         private static function hebrewYearStart( $year ) {
1878                 $a = intval( ( 12 * ( $year - 1 ) + 17 ) % 19 );
1879                 $b = intval( ( $year - 1 ) % 4 );
1880                 $m = 32.044093161144 + 1.5542417966212 * $a + $b / 4.0 - 0.0031777940220923 * ( $year - 1 );
1881                 if ( $m < 0 ) {
1882                         $m--;
1883                 }
1884                 $Mar = intval( $m );
1885                 if ( $m < 0 ) {
1886                         $m++;
1887                 }
1888                 $m -= $Mar;
1889
1890                 $c = intval( ( $Mar + 3 * ( $year - 1 ) + 5 * $b + 5 ) % 7 );
1891                 if ( $c == 0 && $a > 11 && $m >= 0.89772376543210 ) {
1892                         $Mar++;
1893                 } elseif ( $c == 1 && $a > 6 && $m >= 0.63287037037037 ) {
1894                         $Mar += 2;
1895                 } elseif ( $c == 2 || $c == 4 || $c == 6 ) {
1896                         $Mar++;
1897                 }
1898
1899                 $Mar += intval( ( $year - 3761 ) / 100 ) - intval( ( $year - 3761 ) / 400 ) - 24;
1900                 return $Mar;
1901         }
1902
1903         /**
1904          * Algorithm to convert Gregorian dates to Thai solar dates,
1905          * Minguo dates or Minguo dates.
1906          *
1907          * Link: https://en.wikipedia.org/wiki/Thai_solar_calendar
1908          *       https://en.wikipedia.org/wiki/Minguo_calendar
1909          *       https://en.wikipedia.org/wiki/Japanese_era_name
1910          *
1911          * @param string $ts 14-character timestamp
1912          * @param string $cName Calender name
1913          * @return array Converted year, month, day
1914          */
1915         private static function tsToYear( $ts, $cName ) {
1916                 $gy = substr( $ts, 0, 4 );
1917                 $gm = substr( $ts, 4, 2 );
1918                 $gd = substr( $ts, 6, 2 );
1919
1920                 if ( !strcmp( $cName, 'thai' ) ) {
1921                         # Thai solar dates
1922                         # Add 543 years to the Gregorian calendar
1923                         # Months and days are identical
1924                         $gy_offset = $gy + 543;
1925                         # fix for dates between 1912 and 1941
1926                         # https://en.wikipedia.org/?oldid=836596673#New_year
1927                         if ( $gy >= 1912 && $gy <= 1940 ) {
1928                                 if ( $gm <= 3 ) {
1929                                         $gy_offset--;
1930                                 }
1931                                 $gm = ( $gm - 3 ) % 12;
1932                         }
1933                 } elseif ( ( !strcmp( $cName, 'minguo' ) ) || !strcmp( $cName, 'juche' ) ) {
1934                         # Minguo dates
1935                         # Deduct 1911 years from the Gregorian calendar
1936                         # Months and days are identical
1937                         $gy_offset = $gy - 1911;
1938                 } elseif ( !strcmp( $cName, 'tenno' ) ) {
1939                         # Nengō dates up to Meiji period
1940                         # Deduct years from the Gregorian calendar
1941                         # depending on the nengo periods
1942                         # Months and days are identical
1943                         if ( ( $gy < 1912 )
1944                                 || ( ( $gy == 1912 ) && ( $gm < 7 ) )
1945                                 || ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd < 31 ) )
1946                         ) {
1947                                 # Meiji period
1948                                 $gy_gannen = $gy - 1868 + 1;
1949                                 $gy_offset = $gy_gannen;
1950                                 if ( $gy_gannen == 1 ) {
1951                                         $gy_offset = '元';
1952                                 }
1953                                 $gy_offset = '明治' . $gy_offset;
1954                         } elseif (
1955                                 ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd == 31 ) ) ||
1956                                 ( ( $gy == 1912 ) && ( $gm >= 8 ) ) ||
1957                                 ( ( $gy > 1912 ) && ( $gy < 1926 ) ) ||
1958                                 ( ( $gy == 1926 ) && ( $gm < 12 ) ) ||
1959                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd < 26 ) )
1960                         ) {
1961                                 # Taishō period
1962                                 $gy_gannen = $gy - 1912 + 1;
1963                                 $gy_offset = $gy_gannen;
1964                                 if ( $gy_gannen == 1 ) {
1965                                         $gy_offset = '元';
1966                                 }
1967                                 $gy_offset = '大正' . $gy_offset;
1968                         } elseif (
1969                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd >= 26 ) ) ||
1970                                 ( ( $gy > 1926 ) && ( $gy < 1989 ) ) ||
1971                                 ( ( $gy == 1989 ) && ( $gm == 1 ) && ( $gd < 8 ) )
1972                         ) {
1973                                 # Shōwa period
1974                                 $gy_gannen = $gy - 1926 + 1;
1975                                 $gy_offset = $gy_gannen;
1976                                 if ( $gy_gannen == 1 ) {
1977                                         $gy_offset = '元';
1978                                 }
1979                                 $gy_offset = '昭和' . $gy_offset;
1980                         } elseif (
1981                                 ( ( $gy == 1989 ) && ( $gm == 1 ) && ( $gd >= 8 ) ) ||
1982                                 ( ( $gy > 1989 ) && ( $gy < 2019 ) ) ||
1983                                 ( ( $gy == 2019 ) && ( $gm < 5 ) )
1984                         ) {
1985                                 # Heisei period
1986                                 $gy_gannen = $gy - 1989 + 1;
1987                                 $gy_offset = $gy_gannen;
1988                                 if ( $gy_gannen == 1 ) {
1989                                         $gy_offset = '元';
1990                                 }
1991                                 $gy_offset = '平成' . $gy_offset;
1992                         } else {
1993                                 # Reiwa period
1994                                 $gy_gannen = $gy - 2019 + 1;
1995                                 $gy_offset = $gy_gannen;
1996                                 if ( $gy_gannen == 1 ) {
1997                                         $gy_offset = '元';
1998                                 }
1999                                 $gy_offset = '令和' . $gy_offset;
2000                         }
2001                 } else {
2002                         $gy_offset = $gy;
2003                 }
2004
2005                 return [ $gy_offset, $gm, $gd ];
2006         }
2007
2008         /**
2009          * Gets directionality of the first strongly directional codepoint, for embedBidi()
2010          *
2011          * This is the rule the BIDI algorithm uses to determine the directionality of
2012          * paragraphs ( https://www.unicode.org/reports/tr9/#The_Paragraph_Level ) and
2013          * FSI isolates ( https://www.unicode.org/reports/tr9/#Explicit_Directional_Isolates ).
2014          *
2015          * TODO: Does not handle BIDI control characters inside the text.
2016          * TODO: Does not handle unallocated characters.
2017          *
2018          * @param string $text Text to test
2019          * @return null|string Directionality ('ltr' or 'rtl') or null
2020          */
2021         private static function strongDirFromContent( $text = '' ) {
2022                 if ( !preg_match( self::$strongDirRegex, $text, $matches ) ) {
2023                         return null;
2024                 }
2025                 if ( $matches[1] === '' ) {
2026                         return 'rtl';
2027                 }
2028                 return 'ltr';
2029         }
2030
2031         /**
2032          * Roman number formatting up to 10000
2033          *
2034          * @param int $num
2035          *
2036          * @return string
2037          */
2038         static function romanNumeral( $num ) {
2039                 static $table = [
2040                         [ '', 'I', 'II', 'III', 'IV', 'V', 'VI', 'VII', 'VIII', 'IX', 'X' ],
2041                         [ '', 'X', 'XX', 'XXX', 'XL', 'L', 'LX', 'LXX', 'LXXX', 'XC', 'C' ],
2042                         [ '', 'C', 'CC', 'CCC', 'CD', 'D', 'DC', 'DCC', 'DCCC', 'CM', 'M' ],
2043                         [ '', 'M', 'MM', 'MMM', 'MMMM', 'MMMMM', 'MMMMMM', 'MMMMMMM',
2044                                 'MMMMMMMM', 'MMMMMMMMM', 'MMMMMMMMMM' ]
2045                 ];
2046
2047                 $num = intval( $num );
2048                 if ( $num > 10000 || $num <= 0 ) {
2049                         return $num;
2050                 }
2051
2052                 $s = '';
2053                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
2054                         if ( $num >= $pow10 ) {
2055                                 $s .= $table[$i][(int)floor( $num / $pow10 )];
2056                         }
2057                         $num = $num % $pow10;
2058                 }
2059                 return $s;
2060         }
2061
2062         /**
2063          * Hebrew Gematria number formatting up to 9999
2064          *
2065          * @param int $num
2066          *
2067          * @return string
2068          */
2069         static function hebrewNumeral( $num ) {
2070                 static $table = [
2071                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ],
2072                         [ '', 'י', 'כ', 'ל', 'מ', 'נ', 'ס', 'ע', 'פ', 'צ', 'ק' ],
2073                         [ '',
2074                                 [ 'ק' ],
2075                                 [ 'ר' ],
2076                                 [ 'ש' ],
2077                                 [ 'ת' ],
2078                                 [ 'ת', 'ק' ],
2079                                 [ 'ת', 'ר' ],
2080                                 [ 'ת', 'ש' ],
2081                                 [ 'ת', 'ת' ],
2082                                 [ 'ת', 'ת', 'ק' ],
2083                                 [ 'ת', 'ת', 'ר' ],
2084                         ],
2085                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ]
2086                 ];
2087
2088                 $num = intval( $num );
2089                 if ( $num > 9999 || $num <= 0 ) {
2090                         return $num;
2091                 }
2092
2093                 // Round thousands have special notations
2094                 if ( $num === 1000 ) {
2095                         return "א' אלף";
2096                 } elseif ( $num % 1000 === 0 ) {
2097                         return $table[0][$num / 1000] . "' אלפים";
2098                 }
2099
2100                 $letters = [];
2101
2102                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
2103                         if ( $num >= $pow10 ) {
2104                                 if ( $num === 15 || $num === 16 ) {
2105                                         $letters[] = $table[0][9];
2106                                         $letters[] = $table[0][$num - 9];
2107                                         $num = 0;
2108                                 } else {
2109                                         $letters = array_merge(
2110                                                 $letters,
2111                                                 (array)$table[$i][intval( $num / $pow10 )]
2112                                         );
2113
2114                                         if ( $pow10 === 1000 ) {
2115                                                 $letters[] = "'";
2116                                         }
2117                                 }
2118                         }
2119
2120                         $num = $num % $pow10;
2121                 }
2122
2123                 $preTransformLength = count( $letters );
2124                 if ( $preTransformLength === 1 ) {
2125                         // Add geresh (single quote) to one-letter numbers
2126                         $letters[] = "'";
2127                 } else {
2128                         $lastIndex = $preTransformLength - 1;
2129                         $letters[$lastIndex] = str_replace(
2130                                 [ 'כ', 'מ', 'נ', 'פ', 'צ' ],
2131                                 [ 'ך', 'ם', 'ן', 'ף', 'ץ' ],
2132                                 $letters[$lastIndex]
2133                         );
2134
2135                         // Add gershayim (double quote) to multiple-letter numbers,
2136                         // but exclude numbers with only one letter after the thousands
2137                         // (1001-1009, 1020, 1030, 2001-2009, etc.)
2138                         if ( $letters[1] === "'" && $preTransformLength === 3 ) {
2139                                 $letters[] = "'";
2140                         } else {
2141                                 array_splice( $letters, -1, 0, '"' );
2142                         }
2143                 }
2144
2145                 return implode( $letters );
2146         }
2147
2148         /**
2149          * Used by date() and time() to adjust the time output.
2150          *
2151          * @param string $ts The time in date('YmdHis') format
2152          * @param mixed $tz Adjust the time by this amount (default false, mean we
2153          *   get user timecorrection setting)
2154          * @return int
2155          */
2156         public function userAdjust( $ts, $tz = false ) {
2157                 global $wgUser, $wgLocalTZoffset;
2158
2159                 if ( $tz === false ) {
2160                         $tz = $wgUser->getOption( 'timecorrection' );
2161                 }
2162
2163                 $data = explode( '|', $tz, 3 );
2164
2165                 if ( $data[0] == 'ZoneInfo' ) {
2166                         try {
2167                                 $userTZ = new DateTimeZone( $data[2] );
2168                                 $date = new DateTime( $ts, new DateTimeZone( 'UTC' ) );
2169                                 $date->setTimezone( $userTZ );
2170                                 return $date->format( 'YmdHis' );
2171                         } catch ( Exception $e ) {
2172                                 // Unrecognized timezone, default to 'Offset' with the stored offset.
2173                                 $data[0] = 'Offset';
2174                         }
2175                 }
2176
2177                 if ( $data[0] == 'System' || $tz == '' ) {
2178                         # Global offset in minutes.
2179                         $minDiff = $wgLocalTZoffset;
2180                 } elseif ( $data[0] == 'Offset' ) {
2181                         $minDiff = intval( $data[1] );
2182                 } else {
2183                         $data = explode( ':', $tz );
2184                         if ( count( $data ) == 2 ) {
2185                                 $data[0] = intval( $data[0] );
2186                                 $data[1] = intval( $data[1] );
2187                                 $minDiff = abs( $data[0] ) * 60 + $data[1];
2188                                 if ( $data[0] < 0 ) {
2189                                         $minDiff = -$minDiff;
2190                                 }
2191                         } else {
2192                                 $minDiff = intval( $data[0] ) * 60;
2193                         }
2194                 }
2195
2196                 # No difference ? Return time unchanged
2197                 if ( $minDiff == 0 ) {
2198                         return $ts;
2199                 }
2200
2201                 Wikimedia\suppressWarnings(); // E_STRICT system time bitching
2202                 # Generate an adjusted date; take advantage of the fact that mktime
2203                 # will normalize out-of-range values so we don't have to split $minDiff
2204                 # into hours and minutes.
2205                 $t = mktime( (
2206                         (int)substr( $ts, 8, 2 ) ), # Hours
2207                         (int)substr( $ts, 10, 2 ) + $minDiff, # Minutes
2208                         (int)substr( $ts, 12, 2 ), # Seconds
2209                         (int)substr( $ts, 4, 2 ), # Month
2210                         (int)substr( $ts, 6, 2 ), # Day
2211                         (int)substr( $ts, 0, 4 ) ); # Year
2212
2213                 $date = date( 'YmdHis', $t );
2214                 Wikimedia\restoreWarnings();
2215
2216                 return $date;
2217         }
2218
2219         /**
2220          * This is meant to be used by time(), date(), and timeanddate() to get
2221          * the date preference they're supposed to use, it should be used in
2222          * all children.
2223          *
2224          *     function timeanddate([...], $format = true) {
2225          *       $datePreference = $this->dateFormat($format);
2226          *       [...]
2227          *     }
2228          *
2229          * @param int|string|bool $usePrefs If true, the user's preference is used
2230          *   if false, the site/language default is used
2231          *   if int/string, assumed to be a format.
2232          * @return string
2233          */
2234         function dateFormat( $usePrefs = true ) {
2235                 global $wgUser;
2236
2237                 if ( is_bool( $usePrefs ) ) {
2238                         if ( $usePrefs ) {
2239                                 $datePreference = $wgUser->getDatePreference();
2240                         } else {
2241                                 $datePreference = (string)User::getDefaultOption( 'date' );
2242                         }
2243                 } else {
2244                         $datePreference = (string)$usePrefs;
2245                 }
2246
2247                 // return int
2248                 if ( $datePreference == '' ) {
2249                         return 'default';
2250                 }
2251
2252                 return $datePreference;
2253         }
2254
2255         /**
2256          * Get a format string for a given type and preference
2257          * @param string $type May be 'date', 'time', 'both', or 'pretty'.
2258          * @param string $pref The format name as it appears in Messages*.php under
2259          *  $datePreferences.
2260          *
2261          * @since 1.22 New type 'pretty' that provides a more readable timestamp format
2262          *
2263          * @return string
2264          */
2265         function getDateFormatString( $type, $pref ) {
2266                 $wasDefault = false;
2267                 if ( $pref == 'default' ) {
2268                         $wasDefault = true;
2269                         $pref = $this->getDefaultDateFormat();
2270                 }
2271
2272                 if ( !isset( $this->dateFormatStrings[$type][$pref] ) ) {
2273                         $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2274
2275                         if ( $type === 'pretty' && $df === null ) {
2276                                 $df = $this->getDateFormatString( 'date', $pref );
2277                         }
2278
2279                         if ( !$wasDefault && $df === null ) {
2280                                 $pref = $this->getDefaultDateFormat();
2281                                 $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2282                         }
2283
2284                         $this->dateFormatStrings[$type][$pref] = $df;
2285                 }
2286                 return $this->dateFormatStrings[$type][$pref];
2287         }
2288
2289         /**
2290          * @param string $ts The time format which needs to be turned into a
2291          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2292          * @param bool $adj Whether to adjust the time output according to the
2293          *   user configured offset ($timecorrection)
2294          * @param mixed $format True to use user's date format preference
2295          * @param string|bool $timecorrection The time offset as returned by
2296          *   validateTimeZone() in Special:Preferences
2297          * @return string
2298          */
2299         public function date( $ts, $adj = false, $format = true, $timecorrection = false ) {
2300                 $ts = wfTimestamp( TS_MW, $ts );
2301                 if ( $adj ) {
2302                         $ts = $this->userAdjust( $ts, $timecorrection );
2303                 }
2304                 $df = $this->getDateFormatString( 'date', $this->dateFormat( $format ) );
2305                 return $this->sprintfDate( $df, $ts );
2306         }
2307
2308         /**
2309          * @param string $ts The time format which needs to be turned into a
2310          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2311          * @param bool $adj Whether to adjust the time output according to the
2312          *   user configured offset ($timecorrection)
2313          * @param mixed $format True to use user's date format preference
2314          * @param string|bool $timecorrection The time offset as returned by
2315          *   validateTimeZone() in Special:Preferences
2316          * @return string
2317          */
2318         public function time( $ts, $adj = false, $format = true, $timecorrection = false ) {
2319                 $ts = wfTimestamp( TS_MW, $ts );
2320                 if ( $adj ) {
2321                         $ts = $this->userAdjust( $ts, $timecorrection );
2322                 }
2323                 $df = $this->getDateFormatString( 'time', $this->dateFormat( $format ) );
2324                 return $this->sprintfDate( $df, $ts );
2325         }
2326
2327         /**
2328          * @param string $ts The time format which needs to be turned into a
2329          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2330          * @param bool $adj Whether to adjust the time output according to the
2331          *   user configured offset ($timecorrection)
2332          * @param mixed $format What format to return, if it's false output the
2333          *   default one (default true)
2334          * @param string|bool $timecorrection The time offset as returned by
2335          *   validateTimeZone() in Special:Preferences
2336          * @return string
2337          */
2338         public function timeanddate( $ts, $adj = false, $format = true, $timecorrection = false ) {
2339                 $ts = wfTimestamp( TS_MW, $ts );
2340                 if ( $adj ) {
2341                         $ts = $this->userAdjust( $ts, $timecorrection );
2342                 }
2343                 $df = $this->getDateFormatString( 'both', $this->dateFormat( $format ) );
2344                 return $this->sprintfDate( $df, $ts );
2345         }
2346
2347         /**
2348          * Takes a number of seconds and turns it into a text using values such as hours and minutes.
2349          *
2350          * @since 1.20
2351          *
2352          * @param int $seconds The amount of seconds.
2353          * @param array $chosenIntervals The intervals to enable.
2354          *
2355          * @return string
2356          */
2357         public function formatDuration( $seconds, array $chosenIntervals = [] ) {
2358                 $intervals = $this->getDurationIntervals( $seconds, $chosenIntervals );
2359
2360                 $segments = [];
2361
2362                 foreach ( $intervals as $intervalName => $intervalValue ) {
2363                         // Messages: duration-seconds, duration-minutes, duration-hours, duration-days, duration-weeks,
2364                         // duration-years, duration-decades, duration-centuries, duration-millennia
2365                         $message = wfMessage( 'duration-' . $intervalName )->numParams( $intervalValue );
2366                         $segments[] = $message->inLanguage( $this )->escaped();
2367                 }
2368
2369                 return $this->listToText( $segments );
2370         }
2371
2372         /**
2373          * Takes a number of seconds and returns an array with a set of corresponding intervals.
2374          * For example 65 will be turned into [ minutes => 1, seconds => 5 ].
2375          *
2376          * @since 1.20
2377          *
2378          * @param int $seconds The amount of seconds.
2379          * @param array $chosenIntervals The intervals to enable.
2380          *
2381          * @return array
2382          */
2383         public function getDurationIntervals( $seconds, array $chosenIntervals = [] ) {
2384                 if ( empty( $chosenIntervals ) ) {
2385                         $chosenIntervals = [
2386                                 'millennia',
2387                                 'centuries',
2388                                 'decades',
2389                                 'years',
2390                                 'days',
2391                                 'hours',
2392                                 'minutes',
2393                                 'seconds'
2394                         ];
2395                 }
2396
2397                 $intervals = array_intersect_key( self::$durationIntervals, array_flip( $chosenIntervals ) );
2398                 $sortedNames = array_keys( $intervals );
2399                 $smallestInterval = array_pop( $sortedNames );
2400
2401                 $segments = [];
2402
2403                 foreach ( $intervals as $name => $length ) {
2404                         $value = floor( $seconds / $length );
2405
2406                         if ( $value > 0 || ( $name == $smallestInterval && empty( $segments ) ) ) {
2407                                 $seconds -= $value * $length;
2408                                 $segments[$name] = $value;
2409                         }
2410                 }
2411
2412                 return $segments;
2413         }
2414
2415         /**
2416          * Internal helper function for userDate(), userTime() and userTimeAndDate()
2417          *
2418          * @param string $type Can be 'date', 'time' or 'both'
2419          * @param string $ts The time format which needs to be turned into a
2420          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2421          * @param User $user User object used to get preferences for timezone and format
2422          * @param array $options Array, can contain the following keys:
2423          *   - 'timecorrection': time correction, can have the following values:
2424          *     - true: use user's preference
2425          *     - false: don't use time correction
2426          *     - int: value of time correction in minutes
2427          *   - 'format': format to use, can have the following values:
2428          *     - true: use user's preference
2429          *     - false: use default preference
2430          *     - string: format to use
2431          * @since 1.19
2432          * @return string
2433          */
2434         private function internalUserTimeAndDate( $type, $ts, User $user, array $options ) {
2435                 $ts = wfTimestamp( TS_MW, $ts );
2436                 $options += [ 'timecorrection' => true, 'format' => true ];
2437                 if ( $options['timecorrection'] !== false ) {
2438                         if ( $options['timecorrection'] === true ) {
2439                                 $offset = $user->getOption( 'timecorrection' );
2440                         } else {
2441                                 $offset = $options['timecorrection'];
2442                         }
2443                         $ts = $this->userAdjust( $ts, $offset );
2444                 }
2445                 if ( $options['format'] === true ) {
2446                         $format = $user->getDatePreference();
2447                 } else {
2448                         $format = $options['format'];
2449                 }
2450                 $df = $this->getDateFormatString( $type, $this->dateFormat( $format ) );
2451                 return $this->sprintfDate( $df, $ts );
2452         }
2453
2454         /**
2455          * Get the formatted date for the given timestamp and formatted for
2456          * the given user.
2457          *
2458          * @param mixed $ts Mixed: the time format which needs to be turned into a
2459          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2460          * @param User $user User object used to get preferences for timezone and format
2461          * @param array $options Array, can contain the following keys:
2462          *   - 'timecorrection': time correction, can have the following values:
2463          *     - true: use user's preference
2464          *     - false: don't use time correction
2465          *     - int: value of time correction in minutes
2466          *   - 'format': format to use, can have the following values:
2467          *     - true: use user's preference
2468          *     - false: use default preference
2469          *     - string: format to use
2470          * @since 1.19
2471          * @return string
2472          */
2473         public function userDate( $ts, User $user, array $options = [] ) {
2474                 return $this->internalUserTimeAndDate( 'date', $ts, $user, $options );
2475         }
2476
2477         /**
2478          * Get the formatted time for the given timestamp and formatted for
2479          * the given user.
2480          *
2481          * @param mixed $ts The time format which needs to be turned into a
2482          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2483          * @param User $user User object used to get preferences for timezone and format
2484          * @param array $options Array, can contain the following keys:
2485          *   - 'timecorrection': time correction, can have the following values:
2486          *     - true: use user's preference
2487          *     - false: don't use time correction
2488          *     - int: value of time correction in minutes
2489          *   - 'format': format to use, can have the following values:
2490          *     - true: use user's preference
2491          *     - false: use default preference
2492          *     - string: format to use
2493          * @since 1.19
2494          * @return string
2495          */
2496         public function userTime( $ts, User $user, array $options = [] ) {
2497                 return $this->internalUserTimeAndDate( 'time', $ts, $user, $options );
2498         }
2499
2500         /**
2501          * Get the formatted date and time for the given timestamp and formatted for
2502          * the given user.
2503          *
2504          * @param mixed $ts The time format which needs to be turned into a
2505          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2506          * @param User $user User object used to get preferences for timezone and format
2507          * @param array $options Array, can contain the following keys:
2508          *   - 'timecorrection': time correction, can have the following values:
2509          *     - true: use user's preference
2510          *     - false: don't use time correction
2511          *     - int: value of time correction in minutes
2512          *   - 'format': format to use, can have the following values:
2513          *     - true: use user's preference
2514          *     - false: use default preference
2515          *     - string: format to use
2516          * @since 1.19
2517          * @return string
2518          */
2519         public function userTimeAndDate( $ts, User $user, array $options = [] ) {
2520                 return $this->internalUserTimeAndDate( 'both', $ts, $user, $options );
2521         }
2522
2523         /**
2524          * Get the timestamp in a human-friendly relative format, e.g., "3 days ago".
2525          *
2526          * Determine the difference between the timestamp and the current time, and
2527          * generate a readable timestamp by returning "<N> <units> ago", where the
2528          * largest possible unit is used.
2529          *
2530          * @since 1.26 (Prior to 1.26 method existed but was not meant to be used directly)
2531          *
2532          * @param MWTimestamp $time
2533          * @param MWTimestamp|null $relativeTo The base timestamp to compare to (defaults to now)
2534          * @param User|null $user User the timestamp is being generated for
2535          *  (or null to use main context's user)
2536          * @return string Formatted timestamp
2537          */
2538         public function getHumanTimestamp(
2539                 MWTimestamp $time, MWTimestamp $relativeTo = null, User $user = null
2540         ) {
2541                 if ( $relativeTo === null ) {
2542                         $relativeTo = new MWTimestamp();
2543                 }
2544                 if ( $user === null ) {
2545                         $user = RequestContext::getMain()->getUser();
2546                 }
2547
2548                 // Adjust for the user's timezone.
2549                 $offsetThis = $time->offsetForUser( $user );
2550                 $offsetRel = $relativeTo->offsetForUser( $user );
2551
2552                 $ts = '';
2553                 if ( Hooks::run( 'GetHumanTimestamp', [ &$ts, $time, $relativeTo, $user, $this ] ) ) {
2554                         $ts = $this->getHumanTimestampInternal( $time, $relativeTo, $user );
2555                 }
2556
2557                 // Reset the timezone on the objects.
2558                 $time->timestamp->sub( $offsetThis );
2559                 $relativeTo->timestamp->sub( $offsetRel );
2560
2561                 return $ts;
2562         }
2563
2564         /**
2565          * Convert an MWTimestamp into a pretty human-readable timestamp using
2566          * the given user preferences and relative base time.
2567          *
2568          * @see Language::getHumanTimestamp
2569          * @param MWTimestamp $ts Timestamp to prettify
2570          * @param MWTimestamp $relativeTo Base timestamp
2571          * @param User $user User preferences to use
2572          * @return string Human timestamp
2573          * @since 1.26
2574          */
2575         private function getHumanTimestampInternal(
2576                 MWTimestamp $ts, MWTimestamp $relativeTo, User $user
2577         ) {
2578                 $diff = $ts->diff( $relativeTo );
2579                 $diffDay = (bool)( (int)$ts->timestamp->format( 'w' ) -
2580                         (int)$relativeTo->timestamp->format( 'w' ) );
2581                 $days = $diff->days ?: (int)$diffDay;
2582                 if ( $diff->invert || $days > 5
2583                         && $ts->timestamp->format( 'Y' ) !== $relativeTo->timestamp->format( 'Y' )
2584                 ) {
2585                         // Timestamps are in different years: use full timestamp
2586                         // Also do full timestamp for future dates
2587                         /**
2588                          * @todo FIXME: Add better handling of future timestamps.
2589                          */
2590                         $format = $this->getDateFormatString( 'both', $user->getDatePreference() ?: 'default' );
2591                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2592                 } elseif ( $days > 5 ) {
2593                         // Timestamps are in same year,  but more than 5 days ago: show day and month only.
2594                         $format = $this->getDateFormatString( 'pretty', $user->getDatePreference() ?: 'default' );
2595                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2596                 } elseif ( $days > 1 ) {
2597                         // Timestamp within the past week: show the day of the week and time
2598                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2599                         $weekday = self::$mWeekdayMsgs[$ts->timestamp->format( 'w' )];
2600                         // Messages:
2601                         // sunday-at, monday-at, tuesday-at, wednesday-at, thursday-at, friday-at, saturday-at
2602                         $ts = wfMessage( "$weekday-at" )
2603                                 ->inLanguage( $this )
2604                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2605                                 ->text();
2606                 } elseif ( $days == 1 ) {
2607                         // Timestamp was yesterday: say 'yesterday' and the time.
2608                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2609                         $ts = wfMessage( 'yesterday-at' )
2610                                 ->inLanguage( $this )
2611                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2612                                 ->text();
2613                 } elseif ( $diff->h > 1 || $diff->h == 1 && $diff->i > 30 ) {
2614                         // Timestamp was today, but more than 90 minutes ago: say 'today' and the time.
2615                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2616                         $ts = wfMessage( 'today-at' )
2617                                 ->inLanguage( $this )
2618                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2619                                 ->text();
2620
2621                 // From here on in, the timestamp was soon enough ago so that we can simply say
2622                 // XX units ago, e.g., "2 hours ago" or "5 minutes ago"
2623                 } elseif ( $diff->h == 1 ) {
2624                         // Less than 90 minutes, but more than an hour ago.
2625                         $ts = wfMessage( 'hours-ago' )->inLanguage( $this )->numParams( 1 )->text();
2626                 } elseif ( $diff->i >= 1 ) {
2627                         // A few minutes ago.
2628                         $ts = wfMessage( 'minutes-ago' )->inLanguage( $this )->numParams( $diff->i )->text();
2629                 } elseif ( $diff->s >= 30 ) {
2630                         // Less than a minute, but more than 30 sec ago.
2631                         $ts = wfMessage( 'seconds-ago' )->inLanguage( $this )->numParams( $diff->s )->text();
2632                 } else {
2633                         // Less than 30 seconds ago.
2634                         $ts = wfMessage( 'just-now' )->text();
2635                 }
2636
2637                 return $ts;
2638         }
2639
2640         /**
2641          * @param string $key
2642          * @return string|null
2643          */
2644         public function getMessage( $key ) {
2645                 return self::$dataCache->getSubitem( $this->mCode, 'messages', $key );
2646         }
2647
2648         /**
2649          * @return array
2650          */
2651         function getAllMessages() {
2652                 return self::$dataCache->getItem( $this->mCode, 'messages' );
2653         }
2654
2655         /**
2656          * @param string $in
2657          * @param string $out
2658          * @param string $string
2659          * @return string
2660          */
2661         public function iconv( $in, $out, $string ) {
2662                 # Even with //IGNORE iconv can whine about illegal characters in
2663                 # *input* string. We just ignore those too.
2664                 # REF: https://bugs.php.net/bug.php?id=37166
2665                 # REF: https://phabricator.wikimedia.org/T18885
2666                 Wikimedia\suppressWarnings();
2667                 $text = iconv( $in, $out . '//IGNORE', $string );
2668                 Wikimedia\restoreWarnings();
2669                 return $text;
2670         }
2671
2672         // callback functions for ucwords(), ucwordbreaks()
2673
2674         /**
2675          * @param array $matches
2676          * @return mixed|string
2677          */
2678         function ucwordbreaksCallbackAscii( $matches ) {
2679                 return $this->ucfirst( $matches[1] );
2680         }
2681
2682         /**
2683          * @param array $matches
2684          * @return string
2685          */
2686         function ucwordbreaksCallbackMB( $matches ) {
2687                 return mb_strtoupper( $matches[0] );
2688         }
2689
2690         /**
2691          * @param array $matches
2692          * @return string
2693          */
2694         function ucwordsCallbackMB( $matches ) {
2695                 return mb_strtoupper( $matches[0] );
2696         }
2697
2698         /**
2699          * Make a string's first character uppercase
2700          *
2701          * @param string $str
2702          *
2703          * @return string
2704          */
2705         public function ucfirst( $str ) {
2706                 $o = ord( $str );
2707                 if ( $o < 96 ) { // if already uppercase...
2708                         return $str;
2709                 } elseif ( $o < 128 ) {
2710                         return ucfirst( $str ); // use PHP's ucfirst()
2711                 } else {
2712                         // fall back to more complex logic in case of multibyte strings
2713                         return $this->uc( $str, true );
2714                 }
2715         }
2716
2717         /**
2718          * Convert a string to uppercase
2719          *
2720          * @param string $str
2721          * @param bool $first
2722          *
2723          * @return string
2724          */
2725         public function uc( $str, $first = false ) {
2726                 if ( $first ) {
2727                         if ( $this->isMultibyte( $str ) ) {
2728                                 return $this->mbUpperChar( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2729                         } else {
2730                                 return ucfirst( $str );
2731                         }
2732                 } else {
2733                         return $this->isMultibyte( $str ) ? mb_strtoupper( $str ) : strtoupper( $str );
2734                 }
2735         }
2736
2737         /**
2738          * Convert character to uppercase, allowing overrides of the default mb_upper
2739          * behaviour, which is buggy in many ways. Having a conversion table can be
2740          * useful during transitions between PHP versions where unicode changes happen.
2741          * This can make some resources unreachable on-wiki, see discussion at T219279.
2742          * Providing such a conversion table can allow to manage the transition period.
2743          *
2744          * @since 1.34
2745          *
2746          * @param string $char
2747          *
2748          * @return string
2749          */
2750         protected function mbUpperChar( $char ) {
2751                 global $wgOverrideUcfirstCharacters;
2752                 if ( array_key_exists( $char, $wgOverrideUcfirstCharacters ) ) {
2753                         return $wgOverrideUcfirstCharacters[$char];
2754                 } else {
2755                         return mb_strtoupper( $char );
2756                 }
2757         }
2758
2759         /**
2760          * @param string $str
2761          * @return mixed|string
2762          */
2763         function lcfirst( $str ) {
2764                 $o = ord( $str );
2765                 if ( !$o ) {
2766                         return strval( $str );
2767                 } elseif ( $o >= 128 ) {
2768                         return $this->lc( $str, true );
2769                 } elseif ( $o > 96 ) {
2770                         return $str;
2771                 } else {
2772                         $str[0] = strtolower( $str[0] );
2773                         return $str;
2774                 }
2775         }
2776
2777         /**
2778          * @param string $str
2779          * @param bool $first
2780          * @return mixed|string
2781          */
2782         function lc( $str, $first = false ) {
2783                 if ( $first ) {
2784                         if ( $this->isMultibyte( $str ) ) {
2785                                 return mb_strtolower( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2786                         } else {
2787                                 return strtolower( substr( $str, 0, 1 ) ) . substr( $str, 1 );
2788                         }
2789                 } else {
2790                         return $this->isMultibyte( $str ) ? mb_strtolower( $str ) : strtolower( $str );
2791                 }
2792         }
2793
2794         /**
2795          * @param string $str
2796          * @return bool
2797          */
2798         function isMultibyte( $str ) {
2799                 return strlen( $str ) !== mb_strlen( $str );
2800         }
2801
2802         /**
2803          * @param string $str
2804          * @return mixed|string
2805          */
2806         function ucwords( $str ) {
2807                 if ( $this->isMultibyte( $str ) ) {
2808                         $str = $this->lc( $str );
2809
2810                         // regexp to find first letter in each word (i.e. after each space)
2811                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)| ([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2812
2813                         // function to use to capitalize a single char
2814                         return preg_replace_callback(
2815                                 $replaceRegexp,
2816                                 [ $this, 'ucwordsCallbackMB' ],
2817                                 $str
2818                         );
2819                 } else {
2820                         return ucwords( strtolower( $str ) );
2821                 }
2822         }
2823
2824         /**
2825          * capitalize words at word breaks
2826          *
2827          * @param string $str
2828          * @return mixed
2829          */
2830         function ucwordbreaks( $str ) {
2831                 if ( $this->isMultibyte( $str ) ) {
2832                         $str = $this->lc( $str );
2833
2834                         // since \b doesn't work for UTF-8, we explicitely define word break chars
2835                         $breaks = "[ \-\(\)\}\{\.,\?!]";
2836
2837                         // find first letter after word break
2838                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)|" .
2839                                 "$breaks([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2840
2841                         return preg_replace_callback(
2842                                 $replaceRegexp,
2843                                 [ $this, 'ucwordbreaksCallbackMB' ],
2844                                 $str
2845                         );
2846                 } else {
2847                         return preg_replace_callback(
2848                                 '/\b([\w\x80-\xff]+)\b/',
2849                                 [ $this, 'ucwordbreaksCallbackAscii' ],
2850                                 $str
2851                         );
2852                 }
2853         }
2854
2855         /**
2856          * Return a case-folded representation of $s
2857          *
2858          * This is a representation such that caseFold($s1)==caseFold($s2) if $s1
2859          * and $s2 are the same except for the case of their characters. It is not
2860          * necessary for the value returned to make sense when displayed.
2861          *
2862          * Do *not* perform any other normalisation in this function. If a caller
2863          * uses this function when it should be using a more general normalisation
2864          * function, then fix the caller.
2865          *
2866          * @param string $s
2867          *
2868          * @return string
2869          */
2870         function caseFold( $s ) {
2871                 return $this->uc( $s );
2872         }
2873
2874         /**
2875          * TODO: $s is not always a string per T218883
2876          * @param string $s
2877          * @return string
2878          */
2879         function checkTitleEncoding( $s ) {
2880                 if ( is_array( $s ) ) {
2881                         throw new MWException( 'Given array to checkTitleEncoding.' );
2882                 }
2883                 if ( StringUtils::isUtf8( $s ) ) {
2884                         return $s;
2885                 }
2886
2887                 return $this->iconv( $this->fallback8bitEncoding(), 'utf-8', $s );
2888         }
2889
2890         /**
2891          * @return string
2892          */
2893         function fallback8bitEncoding() {
2894                 return self::$dataCache->getItem( $this->mCode, 'fallback8bitEncoding' );
2895         }
2896
2897         /**
2898          * Most writing systems use whitespace to break up words.
2899          * Some languages such as Chinese don't conventionally do this,
2900          * which requires special handling when breaking up words for
2901          * searching etc.
2902          *
2903          * @return bool
2904          */
2905         function hasWordBreaks() {
2906                 return true;
2907         }
2908
2909         /**
2910          * Some languages such as Chinese require word segmentation,
2911          * Specify such segmentation when overridden in derived class.
2912          *
2913          * @param string $string
2914          * @return string
2915          */
2916         function segmentByWord( $string ) {
2917                 return $string;
2918         }
2919
2920         /**
2921          * Some languages have special punctuation need to be normalized.
2922          * Make such changes here.
2923          *
2924          * @param string $string
2925          * @return string
2926          */
2927         function normalizeForSearch( $string ) {
2928                 return self::convertDoubleWidth( $string );
2929         }
2930
2931         /**
2932          * convert double-width roman characters to single-width.
2933          * range: ff00-ff5f ~= 0020-007f
2934          *
2935          * @param string $string
2936          *
2937          * @return string
2938          */
2939         protected static function convertDoubleWidth( $string ) {
2940                 static $full = null;
2941                 static $half = null;
2942
2943                 if ( $full === null ) {
2944                         $fullWidth = "０１２３４５６７８９ＡＢＣＤＥＦＧＨＩＪＫＬＭＮＯＰＱＲＳＴＵＶＷＸＹＺａｂｃｄｅｆｇｈｉｊｋｌｍｎｏｐｑｒｓｔｕｖｗｘｙｚ";
2945                         $halfWidth = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz";
2946                         $full = str_split( $fullWidth, 3 );
2947                         $half = str_split( $halfWidth );
2948                 }
2949
2950                 $string = str_replace( $full, $half, $string );
2951                 return $string;
2952         }
2953
2954         /**
2955          * @param string $string
2956          * @param string $pattern
2957          * @return string
2958          */
2959         protected static function insertSpace( $string, $pattern ) {
2960                 $string = preg_replace( $pattern, " $1 ", $string );
2961                 $string = preg_replace( '/ +/', ' ', $string );
2962                 return $string;
2963         }
2964
2965         /**
2966          * @param array $termsArray
2967          * @return array
2968          */
2969         function convertForSearchResult( $termsArray ) {
2970                 # some languages, e.g. Chinese, need to do a conversion
2971                 # in order for search results to be displayed correctly
2972                 return $termsArray;
2973         }
2974
2975         /**
2976          * Get the first character of a string.
2977          *
2978          * @param string $s
2979          * @return string
2980          */
2981         function firstChar( $s ) {
2982                 $matches = [];
2983                 preg_match(
2984                         '/^([\x00-\x7f]|[\xc0-\xdf][\x80-\xbf]|' .
2985                                 '[\xe0-\xef][\x80-\xbf]{2}|[\xf0-\xf7][\x80-\xbf]{3})/',
2986                         $s,
2987                         $matches
2988                 );
2989
2990                 if ( isset( $matches[1] ) ) {
2991                         if ( strlen( $matches[1] ) != 3 ) {
2992                                 return $matches[1];
2993                         }
2994
2995                         // Break down Hangul syllables to grab the first jamo
2996                         $code = UtfNormal\Utils::utf8ToCodepoint( $matches[1] );
2997                         if ( $code < 0xac00 || 0xd7a4 <= $code ) {
2998                                 return $matches[1];
2999                         } elseif ( $code < 0xb098 ) {
3000                                 return "\u{3131}";
3001                         } elseif ( $code < 0xb2e4 ) {
3002                                 return "\u{3134}";
3003                         } elseif ( $code < 0xb77c ) {
3004                                 return "\u{3137}";
3005                         } elseif ( $code < 0xb9c8 ) {
3006                                 return "\u{3139}";
3007                         } elseif ( $code < 0xbc14 ) {
3008                                 return "\u{3141}";
3009                         } elseif ( $code < 0xc0ac ) {
3010                                 return "\u{3142}";
3011                         } elseif ( $code < 0xc544 ) {
3012                                 return "\u{3145}";
3013                         } elseif ( $code < 0xc790 ) {
3014                                 return "\u{3147}";
3015                         } elseif ( $code < 0xcc28 ) {
3016                                 return "\u{3148}";
3017                         } elseif ( $code < 0xce74 ) {
3018                                 return "\u{314A}";
3019                         } elseif ( $code < 0xd0c0 ) {
3020                                 return "\u{314B}";
3021                         } elseif ( $code < 0xd30c ) {
3022                                 return "\u{314C}";
3023                         } elseif ( $code < 0xd558 ) {
3024                                 return "\u{314D}";
3025                         } else {
3026                                 return "\u{314E}";
3027                         }
3028                 } else {
3029                         return '';
3030                 }
3031         }
3032
3033         /**
3034          * @deprecated No-op since 1.28
3035          */
3036         function initEncoding() {
3037                 wfDeprecated( __METHOD__, '1.28' );
3038                 // No-op.
3039         }
3040
3041         /**
3042          * @param string $s
3043          * @return string
3044          * @deprecated No-op since 1.28
3045          */
3046         function recodeForEdit( $s ) {
3047                 wfDeprecated( __METHOD__, '1.28' );
3048                 return $s;
3049         }
3050
3051         /**
3052          * @param string $s
3053          * @return string
3054          * @deprecated No-op since 1.28
3055          */
3056         function recodeInput( $s ) {
3057                 wfDeprecated( __METHOD__, '1.28' );
3058                 return $s;
3059         }
3060
3061         /**
3062          * Convert a UTF-8 string to normal form C. In Malayalam and Arabic, this
3063          * also cleans up certain backwards-compatible sequences, converting them
3064          * to the modern Unicode equivalent.
3065          *
3066          * This is language-specific for performance reasons only.
3067          *
3068          * @param string $s
3069          *
3070          * @return string
3071          */
3072         public function normalize( $s ) {
3073                 global $wgAllUnicodeFixes;
3074                 $s = UtfNormal\Validator::cleanUp( $s );
3075                 if ( $wgAllUnicodeFixes ) {
3076                         $s = $this->transformUsingPairFile( 'normalize-ar.php', $s );
3077                         $s = $this->transformUsingPairFile( 'normalize-ml.php', $s );
3078                 }
3079
3080                 return $s;
3081         }
3082
3083         /**
3084          * Transform a string using serialized data stored in the given file (which
3085          * must be in the serialized subdirectory of $IP). The file contains pairs
3086          * mapping source characters to destination characters.
3087          *
3088          * The data is cached in process memory. This will go faster if you have the
3089          * FastStringSearch extension.
3090          *
3091          * @param string $file
3092          * @param string $string
3093          *
3094          * @throws MWException
3095          * @return string
3096          */
3097         protected function transformUsingPairFile( $file, $string ) {
3098                 if ( !isset( $this->transformData[$file] ) ) {
3099                         global $IP;
3100                         $data = require "$IP/languages/data/{$file}";
3101                         $this->transformData[$file] = new ReplacementArray( $data );
3102                 }
3103                 return $this->transformData[$file]->replace( $string );
3104         }
3105
3106         /**
3107          * For right-to-left language support
3108          *
3109          * @return bool
3110          */
3111         function isRTL() {
3112                 return self::$dataCache->getItem( $this->mCode, 'rtl' );
3113         }
3114
3115         /**
3116          * Return the correct HTML 'dir' attribute value for this language.
3117          * @return string
3118          */
3119         function getDir() {
3120                 return $this->isRTL() ? 'rtl' : 'ltr';
3121         }
3122
3123         /**
3124          * Return 'left' or 'right' as appropriate alignment for line-start
3125          * for this language's text direction.
3126          *
3127          * Should be equivalent to CSS3 'start' text-align value....
3128          *
3129          * @return string
3130          */
3131         function alignStart() {
3132                 return $this->isRTL() ? 'right' : 'left';
3133         }
3134
3135         /**
3136          * Return 'right' or 'left' as appropriate alignment for line-end
3137          * for this language's text direction.
3138          *
3139          * Should be equivalent to CSS3 'end' text-align value....
3140          *
3141          * @return string
3142          */
3143         function alignEnd() {
3144                 return $this->isRTL() ? 'left' : 'right';
3145         }
3146
3147         /**
3148          * A hidden direction mark (LRM or RLM), depending on the language direction.
3149          * Unlike getDirMark(), this function returns the character as an HTML entity.
3150          * This function should be used when the output is guaranteed to be HTML,
3151          * because it makes the output HTML source code more readable. When
3152          * the output is plain text or can be escaped, getDirMark() should be used.
3153          *
3154          * @param bool $opposite Get the direction mark opposite to your language
3155          * @return string
3156          * @since 1.20
3157          */
3158         function getDirMarkEntity( $opposite = false ) {
3159                 if ( $opposite ) {
3160                         return $this->isRTL() ? '&lrm;' : '&rlm;';
3161                 }
3162                 return $this->isRTL() ? '&rlm;' : '&lrm;';
3163         }
3164
3165         /**
3166          * A hidden direction mark (LRM or RLM), depending on the language direction.
3167          * This function produces them as invisible Unicode characters and
3168          * the output may be hard to read and debug, so it should only be used
3169          * when the output is plain text or can be escaped. When the output is
3170          * HTML, use getDirMarkEntity() instead.
3171          *
3172          * @param bool $opposite Get the direction mark opposite to your language
3173          * @return string
3174          */
3175         function getDirMark( $opposite = false ) {
3176                 $lrm = "\u{200E}"; # LEFT-TO-RIGHT MARK, commonly abbreviated LRM
3177                 $rlm = "\u{200F}"; # RIGHT-TO-LEFT MARK, commonly abbreviated RLM
3178                 if ( $opposite ) {
3179                         return $this->isRTL() ? $lrm : $rlm;
3180                 }
3181                 return $this->isRTL() ? $rlm : $lrm;
3182         }
3183
3184         /**
3185          * @return array
3186          */
3187         function capitalizeAllNouns() {
3188                 return self::$dataCache->getItem( $this->mCode, 'capitalizeAllNouns' );
3189         }
3190
3191         /**
3192          * An arrow, depending on the language direction.
3193          *
3194          * @param string $direction The direction of the arrow: forwards (default),
3195          *   backwards, left, right, up, down.
3196          * @return string
3197          */
3198         function getArrow( $direction = 'forwards' ) {
3199                 switch ( $direction ) {
3200                         case 'forwards':
3201                                 return $this->isRTL() ? '←' : '→';
3202                         case 'backwards':
3203                                 return $this->isRTL() ? '→' : '←';
3204                         case 'left':
3205                                 return '←';
3206                         case 'right':
3207                                 return '→';
3208                         case 'up':
3209                                 return '↑';
3210                         case 'down':
3211                                 return '↓';
3212                 }
3213         }
3214
3215         /**
3216          * To allow "foo[[bar]]" to extend the link over the whole word "foobar"
3217          *
3218          * @return bool
3219          */
3220         function linkPrefixExtension() {
3221                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixExtension' );
3222         }
3223
3224         /**
3225          * Get all magic words from cache.
3226          * @return array
3227          */
3228         function getMagicWords() {
3229                 return self::$dataCache->getItem( $this->mCode, 'magicWords' );
3230         }
3231
3232         /**
3233          * Fill a MagicWord object with data from here
3234          *
3235          * @param MagicWord $mw
3236          */
3237         function getMagic( $mw ) {
3238                 $rawEntry = $this->mMagicExtensions[$mw->mId] ??
3239                         self::$dataCache->getSubitem( $this->mCode, 'magicWords', $mw->mId );
3240
3241                 if ( !is_array( $rawEntry ) ) {
3242                         wfWarn( "\"$rawEntry\" is not a valid magic word for \"$mw->mId\"" );
3243                 } else {
3244                         $mw->mCaseSensitive = $rawEntry[0];
3245                         $mw->mSynonyms = array_slice( $rawEntry, 1 );
3246                 }
3247         }
3248
3249         /**
3250          * Add magic words to the extension array
3251          *
3252          * @param array $newWords
3253          */
3254         function addMagicWordsByLang( $newWords ) {
3255                 $fallbackChain = $this->getFallbackLanguages();
3256                 $fallbackChain = array_reverse( $fallbackChain );
3257                 foreach ( $fallbackChain as $code ) {
3258                         if ( isset( $newWords[$code] ) ) {
3259                                 $this->mMagicExtensions = $newWords[$code] + $this->mMagicExtensions;
3260                         }
3261                 }
3262         }
3263
3264         /**
3265          * Get special page names, as an associative array
3266          *   canonical name => array of valid names, including aliases
3267          * @return array
3268          */
3269         function getSpecialPageAliases() {
3270                 // Cache aliases because it may be slow to load them
3271                 if ( is_null( $this->mExtendedSpecialPageAliases ) ) {
3272                         // Initialise array
3273                         $this->mExtendedSpecialPageAliases =
3274                                 self::$dataCache->getItem( $this->mCode, 'specialPageAliases' );
3275                 }
3276
3277                 return $this->mExtendedSpecialPageAliases;
3278         }
3279
3280         /**
3281          * Italic is unsuitable for some languages
3282          *
3283          * @param string $text The text to be emphasized.
3284          * @return string
3285          */
3286         function emphasize( $text ) {
3287                 return "<em>$text</em>";
3288         }
3289
3290         /**
3291          * Normally we output all numbers in plain en_US style, that is
3292          * 293,291.235 for twohundredninetythreethousand-twohundredninetyone
3293          * point twohundredthirtyfive. However this is not suitable for all
3294          * languages, some such as Bengali (bn) want ২,৯৩,২৯১.২৩৫ and others such as
3295          * Icelandic just want to use commas instead of dots, and dots instead
3296          * of commas like "293.291,235".
3297          *
3298          * An example of this function being called:
3299          * <code>
3300          * wfMessage( 'message' )->numParams( $num )->text()
3301          * </code>
3302          *
3303          * See $separatorTransformTable on MessageIs.php for
3304          * the , => . and . => , implementation.
3305          *
3306          * @todo check if it's viable to use localeconv() for the decimal separator thing.
3307          * @param int|float $number The string to be formatted, should be an integer
3308          *   or a floating point number.
3309          * @param bool $nocommafy Set to true for special numbers like dates
3310          * @return string
3311          */
3312         public function formatNum( $number, $nocommafy = false ) {
3313                 global $wgTranslateNumerals;
3314                 if ( !$nocommafy ) {
3315                         $number = $this->commafy( $number );
3316                         $s = $this->separatorTransformTable();
3317                         if ( $s ) {
3318                                 $number = strtr( $number, $s );
3319                         }
3320                 }
3321
3322                 if ( $wgTranslateNumerals ) {
3323                         $s = $this->digitTransformTable();
3324                         if ( $s ) {
3325                                 $number = strtr( $number, $s );
3326                         }
3327                 }
3328
3329                 return (string)$number;
3330         }
3331
3332         /**
3333          * Front-end for non-commafied formatNum
3334          *
3335          * @param int|float $number The string to be formatted, should be an integer
3336          *        or a floating point number.
3337          * @since 1.21
3338          * @return string
3339          */
3340         public function formatNumNoSeparators( $number ) {
3341                 return $this->formatNum( $number, true );
3342         }
3343
3344         /**
3345          * @param string $number
3346          * @return string
3347          */
3348         public function parseFormattedNumber( $number ) {
3349                 $s = $this->digitTransformTable();
3350                 if ( $s ) {
3351                         // eliminate empty array values such as ''. (T66347)
3352                         $s = array_filter( $s );
3353                         $number = strtr( $number, array_flip( $s ) );
3354                 }
3355
3356                 $s = $this->separatorTransformTable();
3357                 if ( $s ) {
3358                         // eliminate empty array values such as ''. (T66347)
3359                         $s = array_filter( $s );
3360                         $number = strtr( $number, array_flip( $s ) );
3361                 }
3362
3363                 $number = strtr( $number, [ ',' => '' ] );
3364                 return $number;
3365         }
3366
3367         /**
3368          * Adds commas to a given number
3369          * @since 1.19
3370          * @param mixed $number
3371          * @return string
3372          */
3373         function commafy( $number ) {
3374                 $digitGroupingPattern = $this->digitGroupingPattern();
3375                 $minimumGroupingDigits = $this->minimumGroupingDigits();
3376                 if ( $number === null ) {
3377                         return '';
3378                 }
3379
3380                 if ( !$digitGroupingPattern || $digitGroupingPattern === "###,###,###" ) {
3381                         // Default grouping is at thousands, use the same for ###,###,### pattern too.
3382                         // In some languages it's conventional not to insert a thousands separator
3383                         // in numbers that are four digits long (1000-9999).
3384                         if ( $minimumGroupingDigits ) {
3385                                 // Number of '#' characters after last comma in the grouping pattern.
3386                                 // The pattern is hardcoded here, but this would vary for different patterns.
3387                                 $primaryGroupingSize = 3;
3388                                 // Maximum length of a number to suppress digit grouping for.
3389                                 $maximumLength = $minimumGroupingDigits + $primaryGroupingSize - 1;
3390                                 if ( preg_match( '/^\-?\d{1,' . $maximumLength . '}(\.\d+)?$/', $number ) ) {
3391                                         return $number;
3392                                 }
3393                         }
3394                         return strrev( (string)preg_replace( '/(\d{3})(?=\d)(?!\d*\.)/', '$1,', strrev( $number ) ) );
3395                 } else {
3396                         // Ref: http://cldr.unicode.org/translation/number-patterns
3397                         $sign = "";
3398                         if ( intval( $number ) < 0 ) {
3399                                 // For negative numbers apply the algorithm like positive number and add sign.
3400                                 $sign = "-";
3401                                 $number = substr( $number, 1 );
3402                         }
3403                         $integerPart = [];
3404                         $decimalPart = [];
3405                         $numMatches = preg_match_all( "/(#+)/", $digitGroupingPattern, $matches );
3406                         preg_match( "/\d+/", $number, $integerPart );
3407                         preg_match( "/\.\d*/", $number, $decimalPart );
3408                         $groupedNumber = ( count( $decimalPart ) > 0 ) ? $decimalPart[0] : "";
3409                         if ( $groupedNumber === $number ) {
3410                                 // the string does not have any number part. Eg: .12345
3411                                 return $sign . $groupedNumber;
3412                         }
3413                         $start = $end = ( $integerPart ) ? strlen( $integerPart[0] ) : 0;
3414                         while ( $start > 0 ) {
3415                                 $match = $matches[0][$numMatches - 1];
3416                                 $matchLen = strlen( $match );
3417                                 $start = $end - $matchLen;
3418                                 if ( $start < 0 ) {
3419                                         $start = 0;
3420                                 }
3421                                 $groupedNumber = substr( $number, $start, $end - $start ) . $groupedNumber;
3422                                 $end = $start;
3423                                 if ( $numMatches > 1 ) {
3424                                         // use the last pattern for the rest of the number
3425                                         $numMatches--;
3426                                 }
3427                                 if ( $start > 0 ) {
3428                                         $groupedNumber = "," . $groupedNumber;
3429                                 }
3430                         }
3431                         return $sign . $groupedNumber;
3432                 }
3433         }
3434
3435         /**
3436          * @return string
3437          */
3438         function digitGroupingPattern() {
3439                 return self::$dataCache->getItem( $this->mCode, 'digitGroupingPattern' );
3440         }
3441
3442         /**
3443          * @return array
3444          */
3445         function digitTransformTable() {
3446                 return self::$dataCache->getItem( $this->mCode, 'digitTransformTable' );
3447         }
3448
3449         /**
3450          * @return array
3451          */
3452         function separatorTransformTable() {
3453                 return self::$dataCache->getItem( $this->mCode, 'separatorTransformTable' );
3454         }
3455
3456         /**
3457          * @return int|null
3458          */
3459         function minimumGroupingDigits() {
3460                 return self::$dataCache->getItem( $this->mCode, 'minimumGroupingDigits' );
3461         }
3462
3463         /**
3464          * Take a list of strings and build a locale-friendly comma-separated
3465          * list, using the local comma-separator message.
3466          * The last two strings are chained with an "and".
3467          *
3468          * @param string[] $list
3469          * @return string
3470          */
3471         public function listToText( array $list ) {
3472                 $itemCount = count( $list );
3473                 if ( $itemCount < 1 ) {
3474                         return '';
3475                 }
3476                 $text = array_pop( $list );
3477                 if ( $itemCount > 1 ) {
3478                         $and = $this->msg( 'and' )->escaped();
3479                         $space = $this->msg( 'word-separator' )->escaped();
3480                         $comma = '';
3481                         if ( $itemCount > 2 ) {
3482                                 $comma = $this->msg( 'comma-separator' )->escaped();
3483                         }
3484                         $text = implode( $comma, $list ) . $and . $space . $text;
3485                 }
3486                 return $text;
3487         }
3488
3489         /**
3490          * Take a list of strings and build a locale-friendly comma-separated
3491          * list, using the local comma-separator message.
3492          * @param string[] $list Array of strings to put in a comma list
3493          * @return string
3494          */
3495         function commaList( array $list ) {
3496                 return implode(
3497                         wfMessage( 'comma-separator' )->inLanguage( $this )->escaped(),
3498                         $list
3499                 );
3500         }
3501
3502         /**
3503          * Take a list of strings and build a locale-friendly semicolon-separated
3504          * list, using the local semicolon-separator message.
3505          * @param string[] $list Array of strings to put in a semicolon list
3506          * @return string
3507          */
3508         function semicolonList( array $list ) {
3509                 return implode(
3510                         wfMessage( 'semicolon-separator' )->inLanguage( $this )->escaped(),
3511                         $list
3512                 );
3513         }
3514
3515         /**
3516          * Same as commaList, but separate it with the pipe instead.
3517          * @param string[] $list Array of strings to put in a pipe list
3518          * @return string
3519          */
3520         function pipeList( array $list ) {
3521                 return implode(
3522                         wfMessage( 'pipe-separator' )->inLanguage( $this )->escaped(),
3523                         $list
3524                 );
3525         }
3526
3527         /**
3528          * Truncate a string to a specified length in bytes, appending an optional
3529          * string (e.g. for ellipsis)
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 length in bytes
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 truncateForDatabase( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3543                 return $this->truncateInternal(
3544                         $string, $length, $ellipsis, $adjustLength, 'strlen', 'substr'
3545                 );
3546         }
3547
3548         /**
3549          * Truncate a string to a specified number of characters, appending an optional
3550          * string (e.g. for ellipsis).
3551          *
3552          * This provides multibyte version of truncateForDatabase() method of this class,
3553          * suitable for truncation based on number of characters, instead of number of bytes.
3554          *
3555          * If $length is negative, the string will be truncated from the beginning.
3556          *
3557          * @since 1.31
3558          *
3559          * @param string $string String to truncate
3560          * @param int $length Maximum number of characters
3561          * @param string $ellipsis String to append to the end of truncated text
3562          * @param bool $adjustLength Subtract length of ellipsis from $length
3563          *
3564          * @return string
3565          */
3566         function truncateForVisual( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3567                 // Passing encoding to mb_strlen and mb_substr is optional.
3568                 // Encoding defaults to mb_internal_encoding(), which is set to UTF-8 in Setup.php, so
3569                 // explicit specification of encoding is skipped.
3570                 // Note: Both multibyte methods are callables invoked in truncateInternal.
3571                 return $this->truncateInternal(
3572                         $string, $length, $ellipsis, $adjustLength, 'mb_strlen', 'mb_substr'
3573                 );
3574         }
3575
3576         /**
3577          * Internal method used for truncation. This method abstracts text truncation into
3578          * one common method, allowing users to provide length measurement function and
3579          * function for finding substring.
3580          *
3581          * For usages, see truncateForDatabase and truncateForVisual.
3582          *
3583          * @param string $string String to truncate
3584          * @param int $length Maximum length of final text
3585          * @param string $ellipsis String to append to the end of truncated text
3586          * @param bool $adjustLength Subtract length of ellipsis from $length
3587          * @param callable $measureLength Callable function used for determining the length of text
3588          * @param callable $getSubstring Callable function used for getting the substrings
3589          *
3590          * @return string
3591          */
3592         private function truncateInternal(
3593                 $string, $length, $ellipsis, $adjustLength, callable $measureLength, callable $getSubstring
3594         ) {
3595                 # Check if there is no need to truncate
3596                 if ( $measureLength( $string ) <= abs( $length ) ) {
3597                         return $string; // no need to truncate
3598                 }
3599
3600                 # Use the localized ellipsis character
3601                 if ( $ellipsis == '...' ) {
3602                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3603                 }
3604                 if ( $length == 0 ) {
3605                         return $ellipsis; // convention
3606                 }
3607
3608                 $stringOriginal = $string;
3609                 # If ellipsis length is >= $length then we can't apply $adjustLength
3610                 if ( $adjustLength && $measureLength( $ellipsis ) >= abs( $length ) ) {
3611                         $string = $ellipsis; // this can be slightly unexpected
3612                 # Otherwise, truncate and add ellipsis...
3613                 } else {
3614                         $ellipsisLength = $adjustLength ? $measureLength( $ellipsis ) : 0;
3615                         if ( $length > 0 ) {
3616                                 $length -= $ellipsisLength;
3617                                 $string = $getSubstring( $string, 0, $length ); // xyz...
3618                                 $string = $this->removeBadCharLast( $string );
3619                                 $string = rtrim( $string ) . $ellipsis;
3620                         } else {
3621                                 $length += $ellipsisLength;
3622                                 $string = $getSubstring( $string, $length ); // ...xyz
3623                                 $string = $this->removeBadCharFirst( $string );
3624                                 $string = $ellipsis . ltrim( $string );
3625                         }
3626                 }
3627
3628                 # Do not truncate if the ellipsis makes the string longer/equal (T24181).
3629                 # This check is *not* redundant if $adjustLength, due to the single case where
3630                 # LEN($ellipsis) > ABS($limit arg); $stringOriginal could be shorter than $string.
3631                 if ( $measureLength( $string ) < $measureLength( $stringOriginal ) ) {
3632                         return $string;
3633                 } else {
3634                         return $stringOriginal;
3635                 }
3636         }
3637
3638         /**
3639          * Remove bytes that represent an incomplete Unicode character
3640          * at the end of string (e.g. bytes of the char are missing)
3641          *
3642          * @param string $string
3643          * @return string
3644          */
3645         protected function removeBadCharLast( $string ) {
3646                 if ( $string != '' ) {
3647                         $char = ord( $string[strlen( $string ) - 1] );
3648                         $m = [];
3649                         if ( $char >= 0xc0 ) {
3650                                 # We got the first byte only of a multibyte char; remove it.
3651                                 $string = substr( $string, 0, -1 );
3652                         } elseif ( $char >= 0x80 &&
3653                                 // Use the /s modifier (PCRE_DOTALL) so (.*) also matches newlines
3654                                 preg_match( '/^(.*)(?:[\xe0-\xef][\x80-\xbf]|' .
3655                                         '[\xf0-\xf7][\x80-\xbf]{1,2})$/s', $string, $m )
3656                         ) {
3657                                 # We chopped in the middle of a character; remove it
3658                                 $string = $m[1];
3659                         }
3660                 }
3661                 return $string;
3662         }
3663
3664         /**
3665          * Remove bytes that represent an incomplete Unicode character
3666          * at the start of string (e.g. bytes of the char are missing)
3667          *
3668          * @param string $string
3669          * @return string
3670          */
3671         protected function removeBadCharFirst( $string ) {
3672                 if ( $string != '' ) {
3673                         $char = ord( $string[0] );
3674                         if ( $char >= 0x80 && $char < 0xc0 ) {
3675                                 # We chopped in the middle of a character; remove the whole thing
3676                                 $string = preg_replace( '/^[\x80-\xbf]+/', '', $string );
3677                         }
3678                 }
3679                 return $string;
3680         }
3681
3682         /**
3683          * Truncate a string of valid HTML to a specified length in bytes,
3684          * appending an optional string (e.g. for ellipses), and return valid HTML
3685          *
3686          * This is only intended for styled/linked text, such as HTML with
3687          * tags like <span> and <a>, were the tags are self-contained (valid HTML).
3688          * Also, this will not detect things like "display:none" CSS.
3689          *
3690          * Note: since 1.18 you do not need to leave extra room in $length for ellipses.
3691          *
3692          * @param string $text HTML string to truncate
3693          * @param int $length (zero/positive) Maximum length (including ellipses)
3694          * @param string $ellipsis String to append to the truncated text
3695          * @return string
3696          */
3697         function truncateHtml( $text, $length, $ellipsis = '...' ) {
3698                 # Use the localized ellipsis character
3699                 if ( $ellipsis == '...' ) {
3700                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3701                 }
3702                 # Check if there is clearly no need to truncate
3703                 if ( $length <= 0 ) {
3704                         return $ellipsis; // no text shown, nothing to format (convention)
3705                 } elseif ( strlen( $text ) <= $length ) {
3706                         return $text; // string short enough even *with* HTML (short-circuit)
3707                 }
3708
3709                 $dispLen = 0; // innerHTML legth so far
3710                 $testingEllipsis = false; // checking if ellipses will make string longer/equal?
3711                 $tagType = 0; // 0-open, 1-close
3712                 $bracketState = 0; // 1-tag start, 2-tag name, 0-neither
3713                 $entityState = 0; // 0-not entity, 1-entity
3714                 $tag = $ret = ''; // accumulated tag name, accumulated result string
3715                 $openTags = []; // open tag stack
3716                 $maybeState = null; // possible truncation state
3717
3718                 $textLen = strlen( $text );
3719                 $neLength = max( 0, $length - strlen( $ellipsis ) ); // non-ellipsis len if truncated
3720                 for ( $pos = 0; true; ++$pos ) {
3721                         # Consider truncation once the display length has reached the maximim.
3722                         # We check if $dispLen > 0 to grab tags for the $neLength = 0 case.
3723                         # Check that we're not in the middle of a bracket/entity...
3724                         if ( $dispLen && $dispLen >= $neLength && $bracketState == 0 && !$entityState ) {
3725                                 if ( !$testingEllipsis ) {
3726                                         $testingEllipsis = true;
3727                                         # Save where we are; we will truncate here unless there turn out to
3728                                         # be so few remaining characters that truncation is not necessary.
3729                                         if ( !$maybeState ) { // already saved? ($neLength = 0 case)
3730                                                 $maybeState = [ $ret, $openTags ]; // save state
3731                                         }
3732                                 } elseif ( $dispLen > $length && $dispLen > strlen( $ellipsis ) ) {
3733                                         # String in fact does need truncation, the truncation point was OK.
3734                                         // @phan-suppress-next-line PhanTypeInvalidExpressionArrayDestructuring
3735                                         list( $ret, $openTags ) = $maybeState; // reload state
3736                                         $ret = $this->removeBadCharLast( $ret ); // multi-byte char fix
3737                                         $ret .= $ellipsis; // add ellipsis
3738                                         break;
3739                                 }
3740                         }
3741                         if ( $pos >= $textLen ) {
3742                                 break; // extra iteration just for above checks
3743                         }
3744
3745                         # Read the next char...
3746                         $ch = $text[$pos];
3747                         $lastCh = $pos ? $text[$pos - 1] : '';
3748                         $ret .= $ch; // add to result string
3749                         if ( $ch == '<' ) {
3750                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags ); // for bad HTML
3751                                 $entityState = 0; // for bad HTML
3752                                 $bracketState = 1; // tag started (checking for backslash)
3753                         } elseif ( $ch == '>' ) {
3754                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags );
3755                                 $entityState = 0; // for bad HTML
3756                                 $bracketState = 0; // out of brackets
3757                         } elseif ( $bracketState == 1 ) {
3758                                 if ( $ch == '/' ) {
3759                                         $tagType = 1; // close tag (e.g. "</span>")
3760                                 } else {
3761                                         $tagType = 0; // open tag (e.g. "<span>")
3762                                         $tag .= $ch;
3763                                 }
3764                                 $bracketState = 2; // building tag name
3765                         } elseif ( $bracketState == 2 ) {
3766                                 if ( $ch != ' ' ) {
3767                                         $tag .= $ch;
3768                                 } else {
3769                                         // Name found (e.g. "<a href=..."), add on tag attributes...
3770                                         $pos += $this->truncate_skip( $ret, $text, "<>", $pos + 1 );
3771                                 }
3772                         } elseif ( $bracketState == 0 ) {
3773                                 if ( $entityState ) {
3774                                         if ( $ch == ';' ) {
3775                                                 $entityState = 0;
3776                                                 $dispLen++; // entity is one displayed char
3777                                         }
3778                                 } else {
3779                                         if ( $neLength == 0 && !$maybeState ) {
3780                                                 // Save state without $ch. We want to *hit* the first
3781                                                 // display char (to get tags) but not *use* it if truncating.
3782                                                 $maybeState = [ substr( $ret, 0, -1 ), $openTags ];
3783                                         }
3784                                         if ( $ch == '&' ) {
3785                                                 $entityState = 1; // entity found, (e.g. "&#160;")
3786                                         } else {
3787                                                 $dispLen++; // this char is displayed
3788                                                 // Add the next $max display text chars after this in one swoop...
3789                                                 $max = ( $testingEllipsis ? $length : $neLength ) - $dispLen;
3790                                                 $skipped = $this->truncate_skip( $ret, $text, "<>&", $pos + 1, $max );
3791                                                 $dispLen += $skipped;
3792                                                 $pos += $skipped;
3793                                         }
3794                                 }
3795                         }
3796                 }
3797                 // Close the last tag if left unclosed by bad HTML
3798                 $this->truncate_endBracket( $tag, $text[$textLen - 1], $tagType, $openTags );
3799                 while ( count( $openTags ) > 0 ) {
3800                         $ret .= '</' . array_pop( $openTags ) . '>'; // close open tags
3801                 }
3802                 return $ret;
3803         }
3804
3805         /**
3806          * truncateHtml() helper function
3807          * like strcspn() but adds the skipped chars to $ret
3808          *
3809          * @param string $ret
3810          * @param string $text
3811          * @param string $search
3812          * @param int $start
3813          * @param null|int $len
3814          * @return int
3815          */
3816         private function truncate_skip( &$ret, $text, $search, $start, $len = null ) {
3817                 if ( $len === null ) {
3818                         $len = -1; // -1 means "no limit" for strcspn
3819                 } elseif ( $len < 0 ) {
3820                         $len = 0; // sanity
3821                 }
3822                 $skipCount = 0;
3823                 if ( $start < strlen( $text ) ) {
3824                         $skipCount = strcspn( $text, $search, $start, $len );
3825                         $ret .= substr( $text, $start, $skipCount );
3826                 }
3827                 return $skipCount;
3828         }
3829
3830         /**
3831          * truncateHtml() helper function
3832          * (a) push or pop $tag from $openTags as needed
3833          * (b) clear $tag value
3834          * @param string &$tag Current HTML tag name we are looking at
3835          * @param int $tagType (0-open tag, 1-close tag)
3836          * @param string $lastCh Character before the '>' that ended this tag
3837          * @param array &$openTags Open tag stack (not accounting for $tag)
3838          */
3839         private function truncate_endBracket( &$tag, $tagType, $lastCh, &$openTags ) {
3840                 $tag = ltrim( $tag );
3841                 if ( $tag != '' ) {
3842                         if ( $tagType == 0 && $lastCh != '/' ) {
3843                                 $openTags[] = $tag; // tag opened (didn't close itself)
3844                         } elseif ( $tagType == 1 ) {
3845                                 if ( $openTags && $tag == $openTags[count( $openTags ) - 1] ) {
3846                                         array_pop( $openTags ); // tag closed
3847                                 }
3848                         }
3849                         $tag = '';
3850                 }
3851         }
3852
3853         /**
3854          * Grammatical transformations, needed for inflected languages
3855          * Invoked by putting {{grammar:case|word}} in a message
3856          *
3857          * @param string $word
3858          * @param string $case
3859          * @return string
3860          */
3861         function convertGrammar( $word, $case ) {
3862                 global $wgGrammarForms;
3863                 if ( isset( $wgGrammarForms[$this->getCode()][$case][$word] ) ) {
3864                         return $wgGrammarForms[$this->getCode()][$case][$word];
3865                 }
3866
3867                 $grammarTransformations = $this->getGrammarTransformations();
3868
3869                 if ( isset( $grammarTransformations[$case] ) ) {
3870                         $forms = $grammarTransformations[$case];
3871
3872                         // Some names of grammar rules are aliases for other rules.
3873                         // In such cases the value is a string rather than object,
3874                         // so load the actual rules.
3875                         if ( is_string( $forms ) ) {
3876                                 $forms = $grammarTransformations[$forms];
3877                         }
3878
3879                         foreach ( array_values( $forms ) as $rule ) {
3880                                 $form = $rule[0];
3881
3882                                 if ( $form === '@metadata' ) {
3883                                         continue;
3884                                 }
3885
3886                                 $replacement = $rule[1];
3887
3888                                 $regex = '/' . addcslashes( $form, '/' ) . '/u';
3889                                 $patternMatches = preg_match( $regex, $word );
3890
3891                                 if ( $patternMatches === false ) {
3892                                         wfLogWarning(
3893                                                 'An error occurred while processing grammar. ' .
3894                                                 "Word: '$word'. Regex: /$form/."
3895                                         );
3896                                 } elseif ( $patternMatches === 1 ) {
3897                                         $word = preg_replace( $regex, $replacement, $word );
3898
3899                                         break;
3900                                 }
3901                         }
3902                 }
3903
3904                 return $word;
3905         }
3906
3907         /**
3908          * Get the grammar forms for the content language
3909          * @return array Array of grammar forms
3910          * @since 1.20
3911          */
3912         function getGrammarForms() {
3913                 global $wgGrammarForms;
3914                 if ( isset( $wgGrammarForms[$this->getCode()] )
3915                         && is_array( $wgGrammarForms[$this->getCode()] )
3916                 ) {
3917                         return $wgGrammarForms[$this->getCode()];
3918                 }
3919
3920                 return [];
3921         }
3922
3923         /**
3924          * Get the grammar transformations data for the language.
3925          * Used like grammar forms, with {{GRAMMAR}} and cases,
3926          * but uses pairs of regexes and replacements instead of code.
3927          *
3928          * @return array[] Array of grammar transformations.
3929          * @throws MWException
3930          * @since 1.28
3931          */
3932         public function getGrammarTransformations() {
3933                 $languageCode = $this->getCode();
3934
3935                 if ( self::$grammarTransformations === null ) {
3936                         self::$grammarTransformations = new MapCacheLRU( 10 );
3937                 }
3938
3939                 if ( self::$grammarTransformations->has( $languageCode ) ) {
3940                         return self::$grammarTransformations->get( $languageCode );
3941                 }
3942
3943                 $data = [];
3944
3945                 $grammarDataFile = __DIR__ . "/data/grammarTransformations/$languageCode.json";
3946                 if ( is_readable( $grammarDataFile ) ) {
3947                         $data = FormatJson::decode(
3948                                 file_get_contents( $grammarDataFile ),
3949                                 true
3950                         );
3951
3952                         if ( $data === null ) {
3953                                 throw new MWException( "Invalid grammar data for \"$languageCode\"." );
3954                         }
3955
3956                         self::$grammarTransformations->set( $languageCode, $data );
3957                 }
3958
3959                 return $data;
3960         }
3961
3962         /**
3963          * Provides an alternative text depending on specified gender.
3964          * Usage {{gender:username|masculine|feminine|unknown}}.
3965          * username is optional, in which case the gender of current user is used,
3966          * but only in (some) interface messages; otherwise default gender is used.
3967          *
3968          * If no forms are given, an empty string is returned. If only one form is
3969          * given, it will be returned unconditionally. These details are implied by
3970          * the caller and cannot be overridden in subclasses.
3971          *
3972          * If three forms are given, the default is to use the third (unknown) form.
3973          * If fewer than three forms are given, the default is to use the first (masculine) form.
3974          * These details can be overridden in subclasses.
3975          *
3976          * @param string $gender
3977          * @param array $forms
3978          *
3979          * @return string
3980          */
3981         function gender( $gender, $forms ) {
3982                 if ( !count( $forms ) ) {
3983                         return '';
3984                 }
3985                 $forms = $this->preConvertPlural( $forms, 2 );
3986                 if ( $gender === 'male' ) {
3987                         return $forms[0];
3988                 }
3989                 if ( $gender === 'female' ) {
3990                         return $forms[1];
3991                 }
3992                 return $forms[2] ?? $forms[0];
3993         }
3994
3995         /**
3996          * Plural form transformations, needed for some languages.
3997          * For example, there are 3 form of plural in Russian and Polish,
3998          * depending on "count mod 10". See [[w:Plural]]
3999          * For English it is pretty simple.
4000          *
4001          * Invoked by putting {{plural:count|wordform1|wordform2}}
4002          * or {{plural:count|wordform1|wordform2|wordform3}}
4003          *
4004          * Example: {{plural:{{NUMBEROFARTICLES}}|article|articles}}
4005          *
4006          * @param int $count Non-localized number
4007          * @param array $forms Different plural forms
4008          * @return string Correct form of plural for $count in this language
4009          */
4010         function convertPlural( $count, $forms ) {
4011                 // Handle explicit n=pluralform cases
4012                 $forms = $this->handleExplicitPluralForms( $count, $forms );
4013                 if ( is_string( $forms ) ) {
4014                         return $forms;
4015                 }
4016                 if ( !count( $forms ) ) {
4017                         return '';
4018                 }
4019
4020                 $pluralForm = $this->getPluralRuleIndexNumber( $count );
4021                 $pluralForm = min( $pluralForm, count( $forms ) - 1 );
4022                 return $forms[$pluralForm];
4023         }
4024
4025         /**
4026          * Handles explicit plural forms for Language::convertPlural()
4027          *
4028          * In {{PLURAL:$1|0=nothing|one|many}}, 0=nothing will be returned if $1 equals zero.
4029          * If an explicitly defined plural form matches the $count, then
4030          * string value returned, otherwise array returned for further consideration
4031          * by CLDR rules or overridden convertPlural().
4032          *
4033          * @since 1.23
4034          *
4035          * @param int $count Non-localized number
4036          * @param array $forms Different plural forms
4037          *
4038          * @return array|string
4039          */
4040         protected function handleExplicitPluralForms( $count, array $forms ) {
4041                 foreach ( $forms as $index => $form ) {
4042                         if ( preg_match( '/\d+=/i', $form ) ) {
4043                                 $pos = strpos( $form, '=' );
4044                                 if ( substr( $form, 0, $pos ) === (string)$count ) {
4045                                         return substr( $form, $pos + 1 );
4046                                 }
4047                                 unset( $forms[$index] );
4048                         }
4049                 }
4050                 return array_values( $forms );
4051         }
4052
4053         /**
4054          * Checks that convertPlural was given an array and pads it to requested
4055          * amount of forms by copying the last one.
4056          *
4057          * @param array $forms
4058          * @param int $count How many forms should there be at least
4059          * @return array Padded array of forms
4060          */
4061         protected function preConvertPlural( /* Array */ $forms, $count ) {
4062                 return array_pad( $forms, $count, end( $forms ) );
4063         }
4064
4065         /**
4066          * Wraps argument with unicode control characters for directionality safety
4067          *
4068          * This solves the problem where directionality-neutral characters at the edge of
4069          * the argument string get interpreted with the wrong directionality from the
4070          * enclosing context, giving renderings that look corrupted like "(Ben_(WMF".
4071          *
4072          * The wrapping is LRE...PDF or RLE...PDF, depending on the detected
4073          * directionality of the argument string, using the BIDI algorithm's own "First
4074          * strong directional codepoint" rule. Essentially, this works round the fact that
4075          * there is no embedding equivalent of U+2068 FSI (isolation with heuristic
4076          * direction inference). The latter is cleaner but still not widely supported.
4077          *
4078          * @param string $text Text to wrap
4079          * @return string Text, wrapped in LRE...PDF or RLE...PDF or nothing
4080          */
4081         public function embedBidi( $text = '' ) {
4082                 $dir = self::strongDirFromContent( $text );
4083                 if ( $dir === 'ltr' ) {
4084                         // Wrap in LEFT-TO-RIGHT EMBEDDING ... POP DIRECTIONAL FORMATTING
4085                         return self::$lre . $text . self::$pdf;
4086                 }
4087                 if ( $dir === 'rtl' ) {
4088                         // Wrap in RIGHT-TO-LEFT EMBEDDING ... POP DIRECTIONAL FORMATTING
4089                         return self::$rle . $text . self::$pdf;
4090                 }
4091                 // No strong directionality: do not wrap
4092                 return $text;
4093         }
4094
4095         /**
4096          * @todo Maybe translate block durations.  Note that this function is somewhat misnamed: it
4097          * deals with translating the *duration* ("1 week", "4 days", etc), not the expiry time
4098          * (which is an absolute timestamp). Please note: do NOT add this blindly, as it is used
4099          * on old expiry lengths recorded in log entries. You'd need to provide the start date to
4100          * match up with it.
4101          *
4102          * @param string $str The validated block duration in English
4103          * @param User|null $user User object to use timezone from or null for $wgUser
4104          * @param int $now Current timestamp, for formatting relative block durations
4105          * @return string Somehow translated block duration
4106          * @see LanguageFi.php for example implementation
4107          */
4108         function translateBlockExpiry( $str, User $user = null, $now = 0 ) {
4109                 $duration = SpecialBlock::getSuggestedDurations( $this );
4110                 foreach ( $duration as $show => $value ) {
4111                         if ( strcmp( $str, $value ) == 0 ) {
4112                                 return htmlspecialchars( trim( $show ) );
4113                         }
4114                 }
4115
4116                 if ( wfIsInfinity( $str ) ) {
4117                         foreach ( $duration as $show => $value ) {
4118                                 if ( wfIsInfinity( $value ) ) {
4119                                         return htmlspecialchars( trim( $show ) );
4120                                 }
4121                         }
4122                 }
4123
4124                 // If all else fails, return a standard duration or timestamp description.
4125                 $time = strtotime( $str, $now );
4126                 if ( $time === false ) { // Unknown format. Return it as-is in case.
4127                         return $str;
4128                 } elseif ( $time !== strtotime( $str, $now + 1 ) ) { // It's a relative timestamp.
4129                         // The result differs based on current time, so the difference
4130                         // is a fixed duration length.
4131                         return $this->formatDuration( $time - $now );
4132                 } else { // It's an absolute timestamp.
4133                         if ( $time === 0 ) {
4134                                 // wfTimestamp() handles 0 as current time instead of epoch.
4135                                 $time = '19700101000000';
4136                         }
4137                         if ( $user ) {
4138                                 return $this->userTimeAndDate( $time, $user );
4139                         }
4140                         return $this->timeanddate( $time );
4141                 }
4142         }
4143
4144         /**
4145          * languages like Chinese need to be segmented in order for the diff
4146          * to be of any use
4147          *
4148          * @param string $text
4149          * @return string
4150          */
4151         public function segmentForDiff( $text ) {
4152                 return $text;
4153         }
4154
4155         /**
4156          * and unsegment to show the result
4157          *
4158          * @param string $text
4159          * @return string
4160          */
4161         public function unsegmentForDiff( $text ) {
4162                 return $text;
4163         }
4164
4165         /**
4166          * Return the LanguageConverter used in the Language
4167          *
4168          * @since 1.19
4169          * @return LanguageConverter
4170          */
4171         public function getConverter() {
4172                 return $this->mConverter;
4173         }
4174
4175         /**
4176          * convert text to a variant
4177          *
4178          * @param string $text text to convert
4179          * @param string|bool $variant variant to convert to, or false to use the user's preferred
4180          *      variant (if logged in), or the project default variant
4181          * @return string the converted string
4182          */
4183         public function autoConvert( $text, $variant = false ) {
4184                 return $this->mConverter->autoConvert( $text, $variant );
4185         }
4186
4187         /**
4188          * convert text to all supported variants
4189          *
4190          * @param string $text
4191          * @return array
4192          */
4193         public function autoConvertToAllVariants( $text ) {
4194                 return $this->mConverter->autoConvertToAllVariants( $text );
4195         }
4196
4197         /**
4198          * convert text to different variants of a language.
4199          *
4200          * @warning Glossary state is maintained between calls. This means
4201          *  if you pass unescaped text to this method it can cause an XSS
4202          *  in later calls to this method, even if the later calls have properly
4203          *  escaped the input. Never feed this method user controlled text that
4204          *  is not properly escaped!
4205          * @param string $text Content that has been already escaped for use in HTML
4206          * @return string HTML
4207          */
4208         public function convert( $text ) {
4209                 return $this->mConverter->convert( $text );
4210         }
4211
4212         /**
4213          * Convert a Title object to a string in the preferred variant
4214          *
4215          * @param Title $title
4216          * @return string
4217          */
4218         public function convertTitle( $title ) {
4219                 return $this->mConverter->convertTitle( $title );
4220         }
4221
4222         /**
4223          * Convert a namespace index to a string in the preferred variant
4224          *
4225          * @param int $ns namespace index (https://www.mediawiki.org/wiki/Manual:Namespace)
4226          * @param string|null $variant variant to convert to, or null to use the user's preferred
4227          *      variant (if logged in), or the project default variant
4228          * @return string a string representation of the namespace
4229          */
4230         public function convertNamespace( $ns, $variant = null ) {
4231                 return $this->mConverter->convertNamespace( $ns, $variant );
4232         }
4233
4234         /**
4235          * Check if this is a language with variants
4236          *
4237          * @return bool
4238          */
4239         public function hasVariants() {
4240                 return count( $this->getVariants() ) > 1;
4241         }
4242
4243         /**
4244          * Strict check if the language has the specific variant.
4245          *
4246          * Compare to LanguageConverter::validateVariant() which does a more
4247          * lenient check and attempts to coerce the given code to a valid one.
4248          *
4249          * @since 1.19
4250          * @param string $variant
4251          * @return bool
4252          */
4253         public function hasVariant( $variant ) {
4254                 return $variant && ( $variant === $this->mConverter->validateVariant( $variant ) );
4255         }
4256
4257         /**
4258          * Perform output conversion on a string, and encode for safe HTML output.
4259          * @param string $text Text to be converted
4260          * @return string
4261          * @todo this should get integrated somewhere sane
4262          */
4263         public function convertHtml( $text ) {
4264                 return htmlspecialchars( $this->convert( $text ) );
4265         }
4266
4267         /**
4268          * @param string $key
4269          * @return string
4270          */
4271         public function convertCategoryKey( $key ) {
4272                 return $this->mConverter->convertCategoryKey( $key );
4273         }
4274
4275         /**
4276          * Get the list of variants supported by this language
4277          * see sample implementation in LanguageZh.php
4278          *
4279          * @return string[] An array of language codes
4280          */
4281         public function getVariants() {
4282                 return $this->mConverter->getVariants();
4283         }
4284
4285         /**
4286          * @return string
4287          */
4288         public function getPreferredVariant() {
4289                 return $this->mConverter->getPreferredVariant();
4290         }
4291
4292         /**
4293          * @return string
4294          */
4295         public function getDefaultVariant() {
4296                 return $this->mConverter->getDefaultVariant();
4297         }
4298
4299         /**
4300          * @return string
4301          */
4302         public function getURLVariant() {
4303                 return $this->mConverter->getURLVariant();
4304         }
4305
4306         /**
4307          * If a language supports multiple variants, it is
4308          * possible that non-existing link in one variant
4309          * actually exists in another variant. this function
4310          * tries to find it. See e.g. LanguageZh.php
4311          * The input parameters may be modified upon return
4312          *
4313          * @param string &$link The name of the link
4314          * @param Title &$nt The title object of the link
4315          * @param bool $ignoreOtherCond To disable other conditions when
4316          *   we need to transclude a template or update a category's link
4317          */
4318         public function findVariantLink( &$link, &$nt, $ignoreOtherCond = false ) {
4319                 $this->mConverter->findVariantLink( $link, $nt, $ignoreOtherCond );
4320         }
4321
4322         /**
4323          * returns language specific options used by User::getPageRenderHash()
4324          * for example, the preferred language variant
4325          *
4326          * @return string
4327          */
4328         function getExtraHashOptions() {
4329                 return $this->mConverter->getExtraHashOptions();
4330         }
4331
4332         /**
4333          * For languages that support multiple variants, the title of an
4334          * article may be displayed differently in different variants. this
4335          * function returns the apporiate title defined in the body of the article.
4336          *
4337          * @return string
4338          */
4339         public function getParsedTitle() {
4340                 return $this->mConverter->getParsedTitle();
4341         }
4342
4343         /**
4344          * Refresh the cache of conversion tables when
4345          * MediaWiki:Conversiontable* is updated.
4346          *
4347          * @param Title $title The Title of the page being updated
4348          */
4349         public function updateConversionTable( Title $title ) {
4350                 $this->mConverter->updateConversionTable( $title );
4351         }
4352
4353         /**
4354          * A regular expression to match legal word-trailing characters
4355          * which should be merged onto a link of the form [[foo]]bar.
4356          *
4357          * @return string
4358          */
4359         public function linkTrail() {
4360                 return self::$dataCache->getItem( $this->mCode, 'linkTrail' );
4361         }
4362
4363         /**
4364          * A regular expression character set to match legal word-prefixing
4365          * characters which should be merged onto a link of the form foo[[bar]].
4366          *
4367          * @return string
4368          */
4369         public function linkPrefixCharset() {
4370                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixCharset' );
4371         }
4372
4373         /**
4374          * Get the "parent" language which has a converter to convert a "compatible" language
4375          * (in another variant) to this language (eg. zh for zh-cn, but not en for en-gb).
4376          *
4377          * @return Language|null
4378          * @since 1.22
4379          */
4380         public function getParentLanguage() {
4381                 if ( $this->mParentLanguage !== false ) {
4382                         return $this->mParentLanguage;
4383                 }
4384
4385                 $code = explode( '-', $this->getCode() )[0];
4386                 if ( !in_array( $code, LanguageConverter::$languagesWithVariants ) ) {
4387                         $this->mParentLanguage = null;
4388                         return null;
4389                 }
4390                 $lang = self::factory( $code );
4391                 if ( !$lang->hasVariant( $this->getCode() ) ) {
4392                         $this->mParentLanguage = null;
4393                         return null;
4394                 }
4395
4396                 $this->mParentLanguage = $lang;
4397                 return $lang;
4398         }
4399
4400         /**
4401          * Compare with an other language object
4402          *
4403          * @since 1.28
4404          * @param Language $lang
4405          * @return bool
4406          */
4407         public function equals( Language $lang ) {
4408                 return $lang === $this || $lang->getCode() === $this->mCode;
4409         }
4410
4411         /**
4412          * Get the internal language code for this language object
4413          *
4414          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4415          * htmlspecialchars() or similar
4416          *
4417          * @return string
4418          */
4419         public function getCode() {
4420                 return $this->mCode;
4421         }
4422
4423         /**
4424          * Get the code in BCP 47 format which we can use
4425          * inside of html lang="" tags.
4426          *
4427          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4428          * htmlspecialchars() or similar.
4429          *
4430          * @since 1.19
4431          * @return string
4432          */
4433         public function getHtmlCode() {
4434                 if ( is_null( $this->mHtmlCode ) ) {
4435                         $this->mHtmlCode = LanguageCode::bcp47( $this->getCode() );
4436                 }
4437                 return $this->mHtmlCode;
4438         }
4439
4440         /**
4441          * Get the language code from a file name. Inverse of getFileName()
4442          * @param string $filename $prefix . $languageCode . $suffix
4443          * @param string $prefix Prefix before the language code
4444          * @param string $suffix Suffix after the language code
4445          * @return string Language code, or false if $prefix or $suffix isn't found
4446          */
4447         public static function getCodeFromFileName( $filename, $prefix = 'Language', $suffix = '.php' ) {
4448                 $m = null;
4449                 preg_match( '/' . preg_quote( $prefix, '/' ) . '([A-Z][a-z_]+)' .
4450                         preg_quote( $suffix, '/' ) . '/', $filename, $m );
4451                 if ( !count( $m ) ) {
4452                         return false;
4453                 }
4454                 return str_replace( '_', '-', strtolower( $m[1] ) );
4455         }
4456
4457         /**
4458          * @param string $code
4459          * @param bool $fallback Whether we're going through language fallback chain
4460          * @return string Name of the language class
4461          */
4462         public static function classFromCode( $code, $fallback = true ) {
4463                 if ( $fallback && $code == 'en' ) {
4464                         return 'Language';
4465                 } else {
4466                         return 'Language' . str_replace( '-', '_', ucfirst( $code ) );
4467                 }
4468         }
4469
4470         /**
4471          * Get the name of a file for a certain language code
4472          * @param string $prefix Prepend this to the filename
4473          * @param string $code Language code
4474          * @param string $suffix Append this to the filename
4475          * @throws MWException
4476          * @return string $prefix . $mangledCode . $suffix
4477          */
4478         public static function getFileName( $prefix, $code, $suffix = '.php' ) {
4479                 if ( !self::isValidBuiltInCode( $code ) ) {
4480                         throw new MWException( "Invalid language code \"$code\"" );
4481                 }
4482
4483                 return $prefix . str_replace( '-', '_', ucfirst( $code ) ) . $suffix;
4484         }
4485
4486         /**
4487          * @param string $code
4488          * @return string
4489          */
4490         public static function getMessagesFileName( $code ) {
4491                 global $IP;
4492                 $file = self::getFileName( "$IP/languages/messages/Messages", $code, '.php' );
4493                 Hooks::run( 'Language::getMessagesFileName', [ $code, &$file ] );
4494                 return $file;
4495         }
4496
4497         /**
4498          * @param string $code
4499          * @return string
4500          * @throws MWException
4501          * @since 1.23
4502          */
4503         public static function getJsonMessagesFileName( $code ) {
4504                 global $IP;
4505
4506                 if ( !self::isValidBuiltInCode( $code ) ) {
4507                         throw new MWException( "Invalid language code \"$code\"" );
4508                 }
4509
4510                 return "$IP/languages/i18n/$code.json";
4511         }
4512
4513         /**
4514          * Get the first fallback for a given language.
4515          *
4516          * @param string $code
4517          *
4518          * @return bool|string
4519          */
4520         public static function getFallbackFor( $code ) {
4521                 $fallbacks = self::getFallbacksFor( $code );
4522                 if ( $fallbacks ) {
4523                         return $fallbacks[0];
4524                 }
4525                 return false;
4526         }
4527
4528         /**
4529          * Get the ordered list of fallback languages.
4530          *
4531          * @since 1.19
4532          * @param string $code Language code
4533          * @param int $mode Fallback mode, either MESSAGES_FALLBACKS (which always falls back to 'en'),
4534          * or STRICT_FALLBACKS (whic honly falls back to 'en' when explicitly defined)
4535          * @throws MWException
4536          * @return array List of language codes
4537          */
4538         public static function getFallbacksFor( $code, $mode = self::MESSAGES_FALLBACKS ) {
4539                 if ( $code === 'en' || !self::isValidBuiltInCode( $code ) ) {
4540                         return [];
4541                 }
4542                 switch ( $mode ) {
4543                         case self::MESSAGES_FALLBACKS:
4544                                 // For unknown languages, fallbackSequence returns an empty array,
4545                                 // hardcode fallback to 'en' in that case as English messages are
4546                                 // always defined.
4547                                 return self::getLocalisationCache()->getItem( $code, 'fallbackSequence' ) ?: [ 'en' ];
4548                         case self::STRICT_FALLBACKS:
4549                                 // Use this mode when you don't want to fallback to English unless
4550                                 // explicitly defined, for example when you have language-variant icons
4551                                 // and an international language-independent fallback.
4552                                 return self::getLocalisationCache()->getItem( $code, 'originalFallbackSequence' );
4553                         default:
4554                                 throw new MWException( "Invalid fallback mode \"$mode\"" );
4555                 }
4556         }
4557
4558         /**
4559          * Get the ordered list of fallback languages, ending with the fallback
4560          * language chain for the site language.
4561          *
4562          * @since 1.22
4563          * @param string $code Language code
4564          * @return array Array( fallbacks, site fallbacks )
4565          */
4566         public static function getFallbacksIncludingSiteLanguage( $code ) {
4567                 global $wgLanguageCode;
4568
4569                 // Usually, we will only store a tiny number of fallback chains, so we
4570                 // keep them in static memory.
4571                 $cacheKey = "{$code}-{$wgLanguageCode}";
4572
4573                 if ( !array_key_exists( $cacheKey, self::$fallbackLanguageCache ) ) {
4574                         $fallbacks = self::getFallbacksFor( $code );
4575
4576                         // Append the site's fallback chain, including the site language itself
4577                         $siteFallbacks = self::getFallbacksFor( $wgLanguageCode );
4578                         array_unshift( $siteFallbacks, $wgLanguageCode );
4579
4580                         // Eliminate any languages already included in the chain
4581                         $siteFallbacks = array_diff( $siteFallbacks, $fallbacks );
4582
4583                         self::$fallbackLanguageCache[$cacheKey] = [ $fallbacks, $siteFallbacks ];
4584                 }
4585                 return self::$fallbackLanguageCache[$cacheKey];
4586         }
4587
4588         /**
4589          * Get all messages for a given language
4590          * WARNING: this may take a long time. If you just need all message *keys*
4591          * but need the *contents* of only a few messages, consider using getMessageKeysFor().
4592          *
4593          * @param string $code
4594          *
4595          * @return array
4596          */
4597         public static function getMessagesFor( $code ) {
4598                 return self::getLocalisationCache()->getItem( $code, 'messages' );
4599         }
4600
4601         /**
4602          * Get a message for a given language
4603          *
4604          * @param string $key
4605          * @param string $code
4606          *
4607          * @return string
4608          */
4609         public static function getMessageFor( $key, $code ) {
4610                 return self::getLocalisationCache()->getSubitem( $code, 'messages', $key );
4611         }
4612
4613         /**
4614          * Get all message keys for a given language. This is a faster alternative to
4615          * array_keys( Language::getMessagesFor( $code ) )
4616          *
4617          * @since 1.19
4618          * @param string $code Language code
4619          * @return array Array of message keys (strings)
4620          */
4621         public static function getMessageKeysFor( $code ) {
4622                 return self::getLocalisationCache()->getSubitemList( $code, 'messages' );
4623         }
4624
4625         /**
4626          * @param string $talk
4627          * @return string
4628          */
4629         function fixVariableInNamespace( $talk ) {
4630                 if ( strpos( $talk, '$1' ) === false ) {
4631                         return $talk;
4632                 }
4633
4634                 global $wgMetaNamespace;
4635                 $talk = str_replace( '$1', $wgMetaNamespace, $talk );
4636
4637                 # Allow grammar transformations
4638                 # Allowing full message-style parsing would make simple requests
4639                 # such as action=raw much more expensive than they need to be.
4640                 # This will hopefully cover most cases.
4641                 $talk = preg_replace_callback( '/{{grammar:(.*?)\|(.*?)}}/i',
4642                         [ $this, 'replaceGrammarInNamespace' ], $talk );
4643                 return str_replace( ' ', '_', $talk );
4644         }
4645
4646         /**
4647          * @param string $m
4648          * @return string
4649          */
4650         function replaceGrammarInNamespace( $m ) {
4651                 return $this->convertGrammar( trim( $m[2] ), trim( $m[1] ) );
4652         }
4653
4654         /**
4655          * Decode an expiry (block, protection, etc) which has come from the DB
4656          *
4657          * @param string $expiry Database expiry String
4658          * @param bool|int $format True to process using language functions, or TS_ constant
4659          *     to return the expiry in a given timestamp
4660          * @param string $infinity If $format is not true, use this string for infinite expiry
4661          * @return string
4662          * @since 1.18
4663          */
4664         public function formatExpiry( $expiry, $format = true, $infinity = 'infinity' ) {
4665                 static $dbInfinity;
4666                 if ( $dbInfinity === null ) {
4667                         $dbInfinity = wfGetDB( DB_REPLICA )->getInfinity();
4668                 }
4669
4670                 if ( $expiry == '' || $expiry === 'infinity' || $expiry == $dbInfinity ) {
4671                         return $format === true
4672                                 ? $this->getMessageFromDB( 'infiniteblock' )
4673                                 : $infinity;
4674                 } else {
4675                         return $format === true
4676                                 ? $this->timeanddate( $expiry, /* User preference timezone */ true )
4677                                 : wfTimestamp( $format, $expiry );
4678                 }
4679         }
4680
4681         /**
4682          * Formats a time given in seconds into a string representation of that time.
4683          *
4684          * @param int|float $seconds
4685          * @param array $format An optional argument that formats the returned string in different ways:
4686          *   If $format['avoid'] === 'avoidseconds': don't show seconds if $seconds >= 1 hour,
4687          *   If $format['avoid'] === 'avoidminutes': don't show seconds/minutes if $seconds > 48 hours,
4688          *   If $format['noabbrevs'] is true: use 'seconds' and friends instead of 'seconds-abbrev'
4689          *     and friends.
4690          * @note For backwards compatibility, $format may also be one of the strings 'avoidseconds'
4691          *     or 'avoidminutes'.
4692          * @return string
4693          */
4694         function formatTimePeriod( $seconds, $format = [] ) {
4695                 if ( !is_array( $format ) ) {
4696                         $format = [ 'avoid' => $format ]; // For backwards compatibility
4697                 }
4698                 if ( !isset( $format['avoid'] ) ) {
4699                         $format['avoid'] = false;
4700                 }
4701                 if ( !isset( $format['noabbrevs'] ) ) {
4702                         $format['noabbrevs'] = false;
4703                 }
4704                 $secondsMsg = wfMessage(
4705                         $format['noabbrevs'] ? 'seconds' : 'seconds-abbrev' )->inLanguage( $this );
4706                 $minutesMsg = wfMessage(
4707                         $format['noabbrevs'] ? 'minutes' : 'minutes-abbrev' )->inLanguage( $this );
4708                 $hoursMsg = wfMessage(
4709                         $format['noabbrevs'] ? 'hours' : 'hours-abbrev' )->inLanguage( $this );
4710                 $daysMsg = wfMessage(
4711                         $format['noabbrevs'] ? 'days' : 'days-abbrev' )->inLanguage( $this );
4712
4713                 if ( round( $seconds * 10 ) < 100 ) {
4714                         $s = $this->formatNum( sprintf( "%.1f", round( $seconds * 10 ) / 10 ) );
4715                         $s = $secondsMsg->params( $s )->text();
4716                 } elseif ( round( $seconds ) < 60 ) {
4717                         $s = $this->formatNum( round( $seconds ) );
4718                         $s = $secondsMsg->params( $s )->text();
4719                 } elseif ( round( $seconds ) < 3600 ) {
4720                         $minutes = floor( $seconds / 60 );
4721                         $secondsPart = round( fmod( $seconds, 60 ) );
4722                         if ( $secondsPart == 60 ) {
4723                                 $secondsPart = 0;
4724                                 $minutes++;
4725                         }
4726                         $s = $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4727                         $s .= ' ';
4728                         $s .= $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4729                 } elseif ( round( $seconds ) <= 2 * 86400 ) {
4730                         $hours = floor( $seconds / 3600 );
4731                         $minutes = floor( ( $seconds - $hours * 3600 ) / 60 );
4732                         $secondsPart = round( $seconds - $hours * 3600 - $minutes * 60 );
4733                         if ( $secondsPart == 60 ) {
4734                                 $secondsPart = 0;
4735                                 $minutes++;
4736                         }
4737                         if ( $minutes == 60 ) {
4738                                 $minutes = 0;
4739                                 $hours++;
4740                         }
4741                         $s = $hoursMsg->params( $this->formatNum( $hours ) )->text();
4742                         $s .= ' ';
4743                         $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4744                         if ( !in_array( $format['avoid'], [ 'avoidseconds', 'avoidminutes' ] ) ) {
4745                                 $s .= ' ' . $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4746                         }
4747                 } else {
4748                         $days = floor( $seconds / 86400 );
4749                         if ( $format['avoid'] === 'avoidminutes' ) {
4750                                 $hours = round( ( $seconds - $days * 86400 ) / 3600 );
4751                                 if ( $hours == 24 ) {
4752                                         $hours = 0;
4753                                         $days++;
4754                                 }
4755                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4756                                 $s .= ' ';
4757                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4758                         } elseif ( $format['avoid'] === 'avoidseconds' ) {
4759                                 $hours = floor( ( $seconds - $days * 86400 ) / 3600 );
4760                                 $minutes = round( ( $seconds - $days * 86400 - $hours * 3600 ) / 60 );
4761                                 if ( $minutes == 60 ) {
4762                                         $minutes = 0;
4763                                         $hours++;
4764                                 }
4765                                 if ( $hours == 24 ) {
4766                                         $hours = 0;
4767                                         $days++;
4768                                 }
4769                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4770                                 $s .= ' ';
4771                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4772                                 $s .= ' ';
4773                                 $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4774                         } else {
4775                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4776                                 $s .= ' ';
4777                                 $s .= $this->formatTimePeriod( $seconds - $days * 86400, $format );
4778                         }
4779                 }
4780                 return $s;
4781         }
4782
4783         /**
4784          * Format a bitrate for output, using an appropriate
4785          * unit (bps, kbps, Mbps, Gbps, Tbps, Pbps, Ebps, Zbps or Ybps) according to
4786          *   the magnitude in question.
4787          *
4788          * This use base 1000. For base 1024 use formatSize(), for another base
4789          * see formatComputingNumbers().
4790          *
4791          * @param int $bps
4792          * @return string
4793          */
4794         function formatBitrate( $bps ) {
4795                 return $this->formatComputingNumbers( $bps, 1000, "bitrate-$1bits" );
4796         }
4797
4798         /**
4799          * @param int $size Size of the unit
4800          * @param int $boundary Size boundary (1000, or 1024 in most cases)
4801          * @param string $messageKey Message key to be uesd
4802          * @return string
4803          */
4804         function formatComputingNumbers( $size, $boundary, $messageKey ) {
4805                 if ( $size <= 0 ) {
4806                         return str_replace( '$1', $this->formatNum( $size ),
4807                                 $this->getMessageFromDB( str_replace( '$1', '', $messageKey ) )
4808                         );
4809                 }
4810                 $sizes = [ '', 'kilo', 'mega', 'giga', 'tera', 'peta', 'exa', 'zeta', 'yotta' ];
4811                 $index = 0;
4812
4813                 $maxIndex = count( $sizes ) - 1;
4814                 while ( $size >= $boundary && $index < $maxIndex ) {
4815                         $index++;
4816                         $size /= $boundary;
4817                 }
4818
4819                 // For small sizes no decimal places necessary
4820                 $round = 0;
4821                 if ( $index > 1 ) {
4822                         // For MB and bigger two decimal places are smarter
4823                         $round = 2;
4824                 }
4825                 $msg = str_replace( '$1', $sizes[$index], $messageKey );
4826
4827                 $size = round( $size, $round );
4828                 $text = $this->getMessageFromDB( $msg );
4829                 return str_replace( '$1', $this->formatNum( $size ), $text );
4830         }
4831
4832         /**
4833          * Format a size in bytes for output, using an appropriate
4834          * unit (B, KB, MB, GB, TB, PB, EB, ZB or YB) according to the magnitude in question
4835          *
4836          * This method use base 1024. For base 1000 use formatBitrate(), for
4837          * another base see formatComputingNumbers()
4838          *
4839          * @param int $size Size to format
4840          * @return string Plain text (not HTML)
4841          */
4842         function formatSize( $size ) {
4843                 return $this->formatComputingNumbers( $size, 1024, "size-$1bytes" );
4844         }
4845
4846         /**
4847          * Make a list item, used by various special pages
4848          *
4849          * @param string $page Page link
4850          * @param string $details HTML safe text between brackets
4851          * @param bool $oppositedm Add the direction mark opposite to your
4852          *   language, to display text properly
4853          * @return string HTML escaped
4854          */
4855         function specialList( $page, $details, $oppositedm = true ) {
4856                 if ( !$details ) {
4857                         return $page;
4858                 }
4859
4860                 $dirmark = ( $oppositedm ? $this->getDirMark( true ) : '' ) . $this->getDirMark();
4861                 return $page .
4862                         $dirmark .
4863                         $this->msg( 'word-separator' )->escaped() .
4864                         $this->msg( 'parentheses' )->rawParams( $details )->escaped();
4865         }
4866
4867         /**
4868          * Generate (prev x| next x) (20|50|100...) type links for paging
4869          *
4870          * @param Title $title Title object to link
4871          * @param int $offset
4872          * @param int $limit
4873          * @param array $query Optional URL query parameter string
4874          * @param bool $atend Optional param for specified if this is the last page
4875          * @return string
4876          * @deprecated since 1.33, use SpecialPage::viewPrevNext()
4877          *  instead.
4878          */
4879         public function viewPrevNext( Title $title, $offset, $limit,
4880                 array $query = [], $atend = false
4881         ) {
4882                 // @todo FIXME: Why on earth this needs one message for the text and another one for tooltip?
4883
4884                 # Make 'previous' link
4885                 $prev = wfMessage( 'prevn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4886                 if ( $offset > 0 ) {
4887                         $plink = $this->numLink( $title, max( $offset - $limit, 0 ), $limit,
4888                                 $query, $prev, 'prevn-title', 'mw-prevlink' );
4889                 } else {
4890                         $plink = htmlspecialchars( $prev );
4891                 }
4892
4893                 # Make 'next' link
4894                 $next = wfMessage( 'nextn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4895                 if ( $atend ) {
4896                         $nlink = htmlspecialchars( $next );
4897                 } else {
4898                         $nlink = $this->numLink( $title, $offset + $limit, $limit,
4899                                 $query, $next, 'nextn-title', 'mw-nextlink' );
4900                 }
4901
4902                 # Make links to set number of items per page
4903                 $numLinks = [];
4904                 foreach ( [ 20, 50, 100, 250, 500 ] as $num ) {
4905                         $numLinks[] = $this->numLink( $title, $offset, $num,
4906                                 $query, $this->formatNum( $num ), 'shown-title', 'mw-numlink' );
4907                 }
4908
4909                 return wfMessage( 'viewprevnext' )->inLanguage( $this )->title( $title
4910                         )->rawParams( $plink, $nlink, $this->pipeList( $numLinks ) )->escaped();
4911         }
4912
4913         /**
4914          * Helper function for viewPrevNext() that generates links
4915          *
4916          * @param Title $title Title object to link
4917          * @param int $offset
4918          * @param int $limit
4919          * @param array $query Extra query parameters
4920          * @param string $link Text to use for the link; will be escaped
4921          * @param string $tooltipMsg Name of the message to use as tooltip
4922          * @param string $class Value of the "class" attribute of the link
4923          * @return string HTML fragment
4924          */
4925         private function numLink( Title $title, $offset, $limit, array $query, $link,
4926                 $tooltipMsg, $class
4927         ) {
4928                 $query = [ 'limit' => $limit, 'offset' => $offset ] + $query;
4929                 $tooltip = wfMessage( $tooltipMsg )->inLanguage( $this )->title( $title )
4930                         ->numParams( $limit )->text();
4931
4932                 return Html::element( 'a', [ 'href' => $title->getLocalURL( $query ),
4933                         'title' => $tooltip, 'class' => $class ], $link );
4934         }
4935
4936         /**
4937          * Get the conversion rule title, if any.
4938          *
4939          * @return string
4940          */
4941         public function getConvRuleTitle() {
4942                 return $this->mConverter->getConvRuleTitle();
4943         }
4944
4945         /**
4946          * Get the compiled plural rules for the language
4947          * @since 1.20
4948          * @return array Associative array with plural form, and plural rule as key-value pairs
4949          */
4950         public function getCompiledPluralRules() {
4951                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'compiledPluralRules' );
4952                 $fallbacks = self::getFallbacksFor( $this->mCode );
4953                 if ( !$pluralRules ) {
4954                         foreach ( $fallbacks as $fallbackCode ) {
4955                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'compiledPluralRules' );
4956                                 if ( $pluralRules ) {
4957                                         break;
4958                                 }
4959                         }
4960                 }
4961                 return $pluralRules;
4962         }
4963
4964         /**
4965          * Get the plural rules for the language
4966          * @since 1.20
4967          * @return array Associative array with plural form number and plural rule as key-value pairs
4968          */
4969         public function getPluralRules() {
4970                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRules' );
4971                 $fallbacks = self::getFallbacksFor( $this->mCode );
4972                 if ( !$pluralRules ) {
4973                         foreach ( $fallbacks as $fallbackCode ) {
4974                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRules' );
4975                                 if ( $pluralRules ) {
4976                                         break;
4977                                 }
4978                         }
4979                 }
4980                 return $pluralRules;
4981         }
4982
4983         /**
4984          * Get the plural rule types for the language
4985          * @since 1.22
4986          * @return array Associative array with plural form number and plural rule type as key-value pairs
4987          */
4988         public function getPluralRuleTypes() {
4989                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRuleTypes' );
4990                 $fallbacks = self::getFallbacksFor( $this->mCode );
4991                 if ( !$pluralRuleTypes ) {
4992                         foreach ( $fallbacks as $fallbackCode ) {
4993                                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRuleTypes' );
4994                                 if ( $pluralRuleTypes ) {
4995                                         break;
4996                                 }
4997                         }
4998                 }
4999                 return $pluralRuleTypes;
5000         }
5001
5002         /**
5003          * Find the index number of the plural rule appropriate for the given number
5004          * @param int $number
5005          * @return int The index number of the plural rule
5006          */
5007         public function getPluralRuleIndexNumber( $number ) {
5008                 $pluralRules = $this->getCompiledPluralRules();
5009                 $form = Evaluator::evaluateCompiled( $number, $pluralRules );
5010                 return $form;
5011         }
5012
5013         /**
5014          * Find the plural rule type appropriate for the given number
5015          * For example, if the language is set to Arabic, getPluralType(5) should
5016          * return 'few'.
5017          * @since 1.22
5018          * @param int $number
5019          * @return string The name of the plural rule type, e.g. one, two, few, many
5020          */
5021         public function getPluralRuleType( $number ) {
5022                 $index = $this->getPluralRuleIndexNumber( $number );
5023                 $pluralRuleTypes = $this->getPluralRuleTypes();
5024                 return $pluralRuleTypes[$index] ?? 'other';
5025         }
5026 }