languages/Language.php

   1 <?php
   2 /**
   3  * Internationalisation code.
   4  * See https://www.mediawiki.org/wiki/Special:MyLanguage/Localisation for more information.
   5  *
   6  * This program is free software; you can redistribute it and/or modify
   7  * it under the terms of the GNU General Public License as published by
   8  * the Free Software Foundation; either version 2 of the License, or
   9  * (at your option) any later version.
  10  *
  11  * This program is distributed in the hope that it will be useful,
  12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  14  * GNU General Public License for more details.
  15  *
  16  * You should have received a copy of the GNU General Public License along
  17  * with this program; if not, write to the Free Software Foundation, Inc.,
  18  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  19  * http://www.gnu.org/copyleft/gpl.html
  20  *
  21  * @file
  22  * @ingroup Language
  23  */
  24
  25 /**
  26  * @defgroup Language Language
  27  */
  28
  29 use CLDRPluralRuleParser\Evaluator;
  30 use MediaWiki\Languages\LanguageNameUtils;
  31 use MediaWiki\MediaWikiServices;
  32
  33 /**
  34  * Internationalisation code
  35  * @ingroup Language
  36  */
  37 class Language {
  38         /**
  39          * Return autonyms in fetchLanguageName(s).
  40          * @since 1.32
  41          * @deprecated since 1.34, LanguageNameUtils::AUTONYMS
  42          */
  43         const AS_AUTONYMS = LanguageNameUtils::AUTONYMS;
  44
  45         /**
  46          * Return all known languages in fetchLanguageName(s).
  47          * @since 1.32
  48          * @deprecated since 1.34, use LanguageNameUtils::ALL
  49          */
  50         const ALL = LanguageNameUtils::ALL;
  51
  52         /**
  53          * Return in fetchLanguageName(s) only the languages for which we have at
  54          * least some localisation.
  55          * @since 1.32
  56          * @deprecated since 1.34, use LanguageNameUtils::SUPPORTED
  57          */
  58         const SUPPORTED = LanguageNameUtils::SUPPORTED;
  59
  60         /**
  61          * @var LanguageConverter
  62          */
  63         public $mConverter;
  64
  65         public $mVariants, $mCode, $mLoaded = false;
  66         public $mMagicExtensions = [];
  67         private $mHtmlCode = null, $mParentLanguage = false;
  68
  69         public $dateFormatStrings = [];
  70         public $mExtendedSpecialPageAliases;
  71
  72         /** @var array|null */
  73         protected $namespaceNames;
  74         protected $mNamespaceIds, $namespaceAliases;
  75
  76         /**
  77          * ReplacementArray object caches
  78          */
  79         public $transformData = [];
  80
  81         /** @var LocalisationCache */
  82         private $localisationCache;
  83
  84         /** @var LanguageNameUtils */
  85         private $langNameUtils;
  86
  87         public static $mLangObjCache = [];
  88
  89         /**
  90          * Return a fallback chain for messages in getFallbacksFor
  91          * @since 1.32
  92          */
  93         const MESSAGES_FALLBACKS = 0;
  94
  95         /**
  96          * Return a strict fallback chain in getFallbacksFor
  97          * @since 1.32
  98          */
  99         const STRICT_FALLBACKS = 1;
 100
 101         // TODO Make these const once we drop HHVM support (T192166)
 102         public static $mWeekdayMsgs = [
 103                 'sunday', 'monday', 'tuesday', 'wednesday', 'thursday',
 104                 'friday', 'saturday'
 105         ];
 106
 107         public static $mWeekdayAbbrevMsgs = [
 108                 'sun', 'mon', 'tue', 'wed', 'thu', 'fri', 'sat'
 109         ];
 110
 111         public static $mMonthMsgs = [
 112                 'january', 'february', 'march', 'april', 'may_long', 'june',
 113                 'july', 'august', 'september', 'october', 'november',
 114                 'december'
 115         ];
 116         public static $mMonthGenMsgs = [
 117                 'january-gen', 'february-gen', 'march-gen', 'april-gen', 'may-gen', 'june-gen',
 118                 'july-gen', 'august-gen', 'september-gen', 'october-gen', 'november-gen',
 119                 'december-gen'
 120         ];
 121         public static $mMonthAbbrevMsgs = [
 122                 'jan', 'feb', 'mar', 'apr', 'may', 'jun', 'jul', 'aug',
 123                 'sep', 'oct', 'nov', 'dec'
 124         ];
 125
 126         public static $mIranianCalendarMonthMsgs = [
 127                 'iranian-calendar-m1', 'iranian-calendar-m2', 'iranian-calendar-m3',
 128                 'iranian-calendar-m4', 'iranian-calendar-m5', 'iranian-calendar-m6',
 129                 'iranian-calendar-m7', 'iranian-calendar-m8', 'iranian-calendar-m9',
 130                 'iranian-calendar-m10', 'iranian-calendar-m11', 'iranian-calendar-m12'
 131         ];
 132
 133         public static $mHebrewCalendarMonthMsgs = [
 134                 'hebrew-calendar-m1', 'hebrew-calendar-m2', 'hebrew-calendar-m3',
 135                 'hebrew-calendar-m4', 'hebrew-calendar-m5', 'hebrew-calendar-m6',
 136                 'hebrew-calendar-m7', 'hebrew-calendar-m8', 'hebrew-calendar-m9',
 137                 'hebrew-calendar-m10', 'hebrew-calendar-m11', 'hebrew-calendar-m12',
 138                 'hebrew-calendar-m6a', 'hebrew-calendar-m6b'
 139         ];
 140
 141         public static $mHebrewCalendarMonthGenMsgs = [
 142                 'hebrew-calendar-m1-gen', 'hebrew-calendar-m2-gen', 'hebrew-calendar-m3-gen',
 143                 'hebrew-calendar-m4-gen', 'hebrew-calendar-m5-gen', 'hebrew-calendar-m6-gen',
 144                 'hebrew-calendar-m7-gen', 'hebrew-calendar-m8-gen', 'hebrew-calendar-m9-gen',
 145                 'hebrew-calendar-m10-gen', 'hebrew-calendar-m11-gen', 'hebrew-calendar-m12-gen',
 146                 'hebrew-calendar-m6a-gen', 'hebrew-calendar-m6b-gen'
 147         ];
 148
 149         public static $mHijriCalendarMonthMsgs = [
 150                 'hijri-calendar-m1', 'hijri-calendar-m2', 'hijri-calendar-m3',
 151                 'hijri-calendar-m4', 'hijri-calendar-m5', 'hijri-calendar-m6',
 152                 'hijri-calendar-m7', 'hijri-calendar-m8', 'hijri-calendar-m9',
 153                 'hijri-calendar-m10', 'hijri-calendar-m11', 'hijri-calendar-m12'
 154         ];
 155
 156         /**
 157          * @since 1.20
 158          * @var array
 159          */
 160         public static $durationIntervals = [
 161                 'millennia' => 31556952000,
 162                 'centuries' => 3155695200,
 163                 'decades' => 315569520,
 164                 'years' => 31556952, // 86400 * ( 365 + ( 24 * 3 + 25 ) / 400 )
 165                 'weeks' => 604800,
 166                 'days' => 86400,
 167                 'hours' => 3600,
 168                 'minutes' => 60,
 169                 'seconds' => 1,
 170         ];
 171
 172         /**
 173          * Cache for language fallbacks.
 174          * @see Language::getFallbacksIncludingSiteLanguage
 175          * @since 1.21
 176          * @var array
 177          */
 178         private static $fallbackLanguageCache = [];
 179
 180         /**
 181          * Cache for grammar rules data
 182          * @var MapCacheLRU|null
 183          */
 184         private static $grammarTransformations;
 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                 $langNameUtils = MediaWikiServices::getInstance()->getLanguageNameUtils();
 242                 if ( !$langNameUtils->isValidCode( $code ) ) {
 243                         throw new MWException( "Invalid language code \"$code\"" );
 244                 }
 245
 246                 if ( !$langNameUtils->isValidBuiltInCode( $code ) ) {
 247                         // It's not possible to customise this code with class files, so
 248                         // just return a Language object. This is to support uselang= hacks.
 249                         $lang = new Language;
 250                         $lang->mCode = $code;
 251                         return $lang;
 252                 }
 253
 254                 // Check if there is a language class for the code
 255                 $class = self::classFromCode( $code, $fallback );
 256                 // LanguageCode does not inherit Language
 257                 if ( class_exists( $class ) && is_a( $class, 'Language', true ) ) {
 258                         $lang = new $class;
 259                         return $lang;
 260                 }
 261
 262                 // Keep trying the fallback list until we find an existing class
 263                 $fallbacks = self::getFallbacksFor( $code );
 264                 foreach ( $fallbacks as $fallbackCode ) {
 265                         if ( !$langNameUtils->isValidBuiltInCode( $fallbackCode ) ) {
 266                                 throw new MWException( "Invalid fallback '$fallbackCode' in fallback sequence for '$code'" );
 267                         }
 268
 269                         $class = self::classFromCode( $fallbackCode );
 270                         if ( class_exists( $class ) ) {
 271                                 $lang = new $class;
 272                                 $lang->mCode = $code;
 273                                 return $lang;
 274                         }
 275                 }
 276
 277                 throw new MWException( "Invalid fallback sequence for language '$code'" );
 278         }
 279
 280         /**
 281          * Intended for tests that may change configuration in a way that invalidates caches.
 282          *
 283          * @since 1.32
 284          */
 285         public static function clearCaches() {
 286                 if ( !defined( 'MW_PHPUNIT_TEST' ) && !defined( 'MEDIAWIKI_INSTALL' ) ) {
 287                         throw new MWException( __METHOD__ . ' must not be used outside tests/installer' );
 288                 }
 289                 if ( !defined( 'MEDIAWIKI_INSTALL' ) ) {
 290                         MediaWikiServices::getInstance()->resetServiceForTesting( 'LocalisationCache' );
 291                         MediaWikiServices::getInstance()->resetServiceForTesting( 'LanguageNameUtils' );
 292                 }
 293                 self::$mLangObjCache = [];
 294                 self::$fallbackLanguageCache = [];
 295                 self::$grammarTransformations = null;
 296         }
 297
 298         /**
 299          * Checks whether any localisation is available for that language tag
 300          * in MediaWiki (MessagesXx.php exists).
 301          *
 302          * @deprecated since 1.34, use LanguageNameUtils
 303          * @param string $code Language tag (in lower case)
 304          * @return bool Whether language is supported
 305          * @since 1.21
 306          */
 307         public static function isSupportedLanguage( $code ) {
 308                 return MediaWikiServices::getInstance()->getLanguageNameUtils()
 309                         ->isSupportedLanguage( $code );
 310         }
 311
 312         /**
 313          * Returns true if a language code string is a well-formed language tag
 314          * according to RFC 5646.
 315          * This function only checks well-formedness; it doesn't check that
 316          * language, script or variant codes actually exist in the repositories.
 317          *
 318          * Based on regexes by Mark Davis of the Unicode Consortium:
 319          * https://www.unicode.org/repos/cldr/trunk/tools/java/org/unicode/cldr/util/data/langtagRegex.txt
 320          *
 321          * @param string $code
 322          * @param bool $lenient Whether to allow '_' as separator. The default is only '-'.
 323          *
 324          * @return bool
 325          * @since 1.21
 326          */
 327         public static function isWellFormedLanguageTag( $code, $lenient = false ) {
 328                 $alpha = '[a-z]';
 329                 $digit = '[0-9]';
 330                 $alphanum = '[a-z0-9]';
 331                 $x = 'x'; # private use singleton
 332                 $singleton = '[a-wy-z]'; # other singleton
 333                 $s = $lenient ? '[-_]' : '-';
 334
 335                 $language = "$alpha{2,8}|$alpha{2,3}$s$alpha{3}";
 336                 $script = "$alpha{4}"; # ISO 15924
 337                 $region = "(?:$alpha{2}|$digit{3})"; # ISO 3166-1 alpha-2 or UN M.49
 338                 $variant = "(?:$alphanum{5,8}|$digit$alphanum{3})";
 339                 $extension = "$singleton(?:$s$alphanum{2,8})+";
 340                 $privateUse = "$x(?:$s$alphanum{1,8})+";
 341
 342                 # Define certain grandfathered codes, since otherwise the regex is pretty useless.
 343                 # Since these are limited, this is safe even later changes to the registry --
 344                 # the only oddity is that it might change the type of the tag, and thus
 345                 # the results from the capturing groups.
 346                 # https://www.iana.org/assignments/language-subtag-registry
 347
 348                 $grandfathered = "en{$s}GB{$s}oed"
 349                         . "|i{$s}(?:ami|bnn|default|enochian|hak|klingon|lux|mingo|navajo|pwn|tao|tay|tsu)"
 350                         . "|no{$s}(?:bok|nyn)"
 351                         . "|sgn{$s}(?:BE{$s}(?:fr|nl)|CH{$s}de)"
 352                         . "|zh{$s}min{$s}nan";
 353
 354                 $variantList = "$variant(?:$s$variant)*";
 355                 $extensionList = "$extension(?:$s$extension)*";
 356
 357                 $langtag = "(?:($language)"
 358                         . "(?:$s$script)?"
 359                         . "(?:$s$region)?"
 360                         . "(?:$s$variantList)?"
 361                         . "(?:$s$extensionList)?"
 362                         . "(?:$s$privateUse)?)";
 363
 364                 # The final breakdown, with capturing groups for each of these components
 365                 # The variants, extensions, grandfathered, and private-use may have interior '-'
 366
 367                 $root = "^(?:$langtag|$privateUse|$grandfathered)$";
 368
 369                 return (bool)preg_match( "/$root/", strtolower( $code ) );
 370         }
 371
 372         /**
 373          * Returns true if a language code string is of a valid form, whether or
 374          * not it exists. This includes codes which are used solely for
 375          * customisation via the MediaWiki namespace.
 376          *
 377          * @deprecated since 1.34, use LanguageNameUtils
 378          *
 379          * @param string $code
 380          *
 381          * @return bool
 382          */
 383         public static function isValidCode( $code ) {
 384                 return MediaWikiServices::getInstance()->getLanguageNameUtils()->isValidCode( $code );
 385         }
 386
 387         /**
 388          * Returns true if a language code is of a valid form for the purposes of
 389          * internal customisation of MediaWiki, via Messages*.php or *.json.
 390          *
 391          * @deprecated since 1.34, use LanguageNameUtils
 392          *
 393          * @param string $code
 394          *
 395          * @since 1.18
 396          * @return bool
 397          */
 398         public static function isValidBuiltInCode( $code ) {
 399                 return MediaWikiServices::getInstance()->getLanguageNameUtils()
 400                         ->isValidBuiltInCode( $code );
 401         }
 402
 403         /**
 404          * Returns true if a language code is an IETF tag known to MediaWiki.
 405          *
 406          * @deprecated since 1.34, use LanguageNameUtils
 407          *
 408          * @param string $tag
 409          *
 410          * @since 1.21
 411          * @return bool
 412          */
 413         public static function isKnownLanguageTag( $tag ) {
 414                 return MediaWikiServices::getInstance()->getLanguageNameUtils()
 415                         ->isKnownLanguageTag( $tag );
 416         }
 417
 418         /**
 419          * Get the LocalisationCache instance
 420          *
 421          * @deprecated since 1.34, use MediaWikiServices
 422          * @return LocalisationCache
 423          */
 424         public static function getLocalisationCache() {
 425                 return MediaWikiServices::getInstance()->getLocalisationCache();
 426         }
 427
 428         function __construct() {
 429                 $this->mConverter = new FakeConverter( $this );
 430                 // Set the code to the name of the descendant
 431                 if ( static::class === 'Language' ) {
 432                         $this->mCode = 'en';
 433                 } else {
 434                         $this->mCode = str_replace( '_', '-', strtolower( substr( static::class, 8 ) ) );
 435                 }
 436                 $services = MediaWikiServices::getInstance();
 437                 $this->localisationCache = $services->getLocalisationCache();
 438                 $this->langNameUtils = $services->getLanguageNameUtils();
 439         }
 440
 441         /**
 442          * Reduce memory usage
 443          */
 444         function __destruct() {
 445                 foreach ( $this as $name => $value ) {
 446                         unset( $this->$name );
 447                 }
 448         }
 449
 450         /**
 451          * Hook which will be called if this is the content language.
 452          * Descendants can use this to register hook functions or modify globals
 453          */
 454         function initContLang() {
 455         }
 456
 457         /**
 458          * @return array
 459          * @since 1.19
 460          */
 461         public function getFallbackLanguages() {
 462                 return self::getFallbacksFor( $this->mCode );
 463         }
 464
 465         /**
 466          * Exports $wgBookstoreListEn
 467          * @return array
 468          */
 469         public function getBookstoreList() {
 470                 return $this->localisationCache->getItem( $this->mCode, 'bookstoreList' );
 471         }
 472
 473         /**
 474          * Returns an array of localised namespaces indexed by their numbers. If the namespace is not
 475          * available in localised form, it will be included in English.
 476          *
 477          * @return string[] List of localized namespace names, indexed by numeric namespace ID.
 478          */
 479         public function getNamespaces() {
 480                 if ( is_null( $this->namespaceNames ) ) {
 481                         global $wgMetaNamespace, $wgMetaNamespaceTalk, $wgExtraNamespaces;
 482
 483                         $validNamespaces = MediaWikiServices::getInstance()->getNamespaceInfo()->
 484                                 getCanonicalNamespaces();
 485
 486                         $this->namespaceNames = $wgExtraNamespaces +
 487                                 $this->localisationCache->getItem( $this->mCode, 'namespaceNames' );
 488                         $this->namespaceNames += $validNamespaces;
 489
 490                         $this->namespaceNames[NS_PROJECT] = $wgMetaNamespace;
 491                         if ( $wgMetaNamespaceTalk ) {
 492                                 $this->namespaceNames[NS_PROJECT_TALK] = $wgMetaNamespaceTalk;
 493                         } else {
 494                                 $talk = $this->namespaceNames[NS_PROJECT_TALK];
 495                                 $this->namespaceNames[NS_PROJECT_TALK] =
 496                                         $this->fixVariableInNamespace( $talk );
 497                         }
 498
 499                         # Sometimes a language will be localised but not actually exist on this wiki.
 500                         foreach ( $this->namespaceNames as $key => $text ) {
 501                                 if ( !isset( $validNamespaces[$key] ) ) {
 502                                         unset( $this->namespaceNames[$key] );
 503                                 }
 504                         }
 505
 506                         # The above mixing may leave namespaces out of canonical order.
 507                         # Re-order by namespace ID number...
 508                         ksort( $this->namespaceNames );
 509
 510                         Hooks::run( 'LanguageGetNamespaces', [ &$this->namespaceNames ] );
 511                 }
 512
 513                 return $this->namespaceNames;
 514         }
 515
 516         /**
 517          * Arbitrarily set all of the namespace names at once. Mainly used for testing
 518          * @param array $namespaces Array of namespaces (id => name)
 519          */
 520         public function setNamespaces( array $namespaces ) {
 521                 $this->namespaceNames = $namespaces;
 522                 $this->mNamespaceIds = null;
 523         }
 524
 525         /**
 526          * Resets all of the namespace caches. Mainly used for testing
 527          */
 528         public function resetNamespaces() {
 529                 $this->namespaceNames = null;
 530                 $this->mNamespaceIds = null;
 531                 $this->namespaceAliases = null;
 532         }
 533
 534         /**
 535          * A convenience function that returns getNamespaces() with spaces instead of underscores
 536          * in values. Useful for producing output to be displayed e.g. in `<select>` forms.
 537          *
 538          * @return array
 539          */
 540         public function getFormattedNamespaces() {
 541                 $ns = $this->getNamespaces();
 542                 foreach ( $ns as $k => $v ) {
 543                         $ns[$k] = strtr( $v, '_', ' ' );
 544                 }
 545                 return $ns;
 546         }
 547
 548         /**
 549          * Get a namespace value by key
 550          *
 551          * <code>
 552          * $mw_ns = $lang->getNsText( NS_MEDIAWIKI );
 553          * echo $mw_ns; // prints 'MediaWiki'
 554          * </code>
 555          *
 556          * @param int $index The array key of the namespace to return
 557          * @return string|bool String if the namespace value exists, otherwise false
 558          */
 559         public function getNsText( $index ) {
 560                 $ns = $this->getNamespaces();
 561                 return $ns[$index] ?? false;
 562         }
 563
 564         /**
 565          * A convenience function that returns the same thing as
 566          * getNsText() except with '_' changed to ' ', useful for
 567          * producing output.
 568          *
 569          * <code>
 570          * $mw_ns = $lang->getFormattedNsText( NS_MEDIAWIKI_TALK );
 571          * echo $mw_ns; // prints 'MediaWiki talk'
 572          * </code>
 573          *
 574          * @param int $index The array key of the namespace to return
 575          * @return string Namespace name without underscores (empty string if namespace does not exist)
 576          */
 577         public function getFormattedNsText( $index ) {
 578                 $ns = $this->getNsText( $index );
 579                 return strtr( $ns, '_', ' ' );
 580         }
 581
 582         /**
 583          * Returns gender-dependent namespace alias if available.
 584          * See https://www.mediawiki.org/wiki/Manual:$wgExtraGenderNamespaces
 585          * @param int $index Namespace index
 586          * @param string $gender Gender key (male, female... )
 587          * @return string
 588          * @since 1.18
 589          */
 590         public function getGenderNsText( $index, $gender ) {
 591                 global $wgExtraGenderNamespaces;
 592
 593                 $ns = $wgExtraGenderNamespaces +
 594                         (array)$this->localisationCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 595
 596                 return $ns[$index][$gender] ?? $this->getNsText( $index );
 597         }
 598
 599         /**
 600          * Whether this language uses gender-dependent namespace aliases.
 601          * See https://www.mediawiki.org/wiki/Manual:$wgExtraGenderNamespaces
 602          * @return bool
 603          * @since 1.18
 604          */
 605         public function needsGenderDistinction() {
 606                 global $wgExtraGenderNamespaces, $wgExtraNamespaces;
 607                 if ( count( $wgExtraGenderNamespaces ) > 0 ) {
 608                         // $wgExtraGenderNamespaces overrides everything
 609                         return true;
 610                 } elseif ( isset( $wgExtraNamespaces[NS_USER] ) && isset( $wgExtraNamespaces[NS_USER_TALK] ) ) {
 611                         /// @todo There may be other gender namespace than NS_USER & NS_USER_TALK in the future
 612                         // $wgExtraNamespaces overrides any gender aliases specified in i18n files
 613                         return false;
 614                 } else {
 615                         // Check what is in i18n files
 616                         $aliases = $this->localisationCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 617                         return count( $aliases ) > 0;
 618                 }
 619         }
 620
 621         /**
 622          * Get a namespace key by value, case insensitive.
 623          * Only matches namespace names for the current language, not the
 624          * canonical ones defined in Namespace.php.
 625          *
 626          * @param string $text
 627          * @return int|bool An integer if $text is a valid value otherwise false
 628          */
 629         function getLocalNsIndex( $text ) {
 630                 $lctext = $this->lc( $text );
 631                 $ids = $this->getNamespaceIds();
 632                 return $ids[$lctext] ?? false;
 633         }
 634
 635         /**
 636          * @return array
 637          */
 638         public function getNamespaceAliases() {
 639                 if ( is_null( $this->namespaceAliases ) ) {
 640                         $aliases = $this->localisationCache->getItem( $this->mCode, 'namespaceAliases' );
 641                         if ( !$aliases ) {
 642                                 $aliases = [];
 643                         } else {
 644                                 foreach ( $aliases as $name => $index ) {
 645                                         if ( $index === NS_PROJECT_TALK ) {
 646                                                 unset( $aliases[$name] );
 647                                                 $name = $this->fixVariableInNamespace( $name );
 648                                                 $aliases[$name] = $index;
 649                                         }
 650                                 }
 651                         }
 652
 653                         global $wgExtraGenderNamespaces;
 654                         $genders = $wgExtraGenderNamespaces + (array)$this->localisationCache
 655                                 ->getItem( $this->mCode, 'namespaceGenderAliases' );
 656                         foreach ( $genders as $index => $forms ) {
 657                                 foreach ( $forms as $alias ) {
 658                                         $aliases[$alias] = $index;
 659                                 }
 660                         }
 661
 662                         # Also add converted namespace names as aliases, to avoid confusion.
 663                         $convertedNames = [];
 664                         foreach ( $this->getVariants() as $variant ) {
 665                                 if ( $variant === $this->mCode ) {
 666                                         continue;
 667                                 }
 668                                 foreach ( $this->getNamespaces() as $ns => $_ ) {
 669                                         $convertedNames[$this->getConverter()->convertNamespace( $ns, $variant )] = $ns;
 670                                 }
 671                         }
 672
 673                         $this->namespaceAliases = $aliases + $convertedNames;
 674
 675                         # Filter out aliases to namespaces that don't exist, e.g. from extensions
 676                         # that aren't loaded here but are included in the l10n cache.
 677                         # (array_intersect preserves keys from its first argument)
 678                         $this->namespaceAliases = array_intersect(
 679                                 $this->namespaceAliases,
 680                                 array_keys( $this->getNamespaces() )
 681                         );
 682                 }
 683
 684                 return $this->namespaceAliases;
 685         }
 686
 687         /**
 688          * @return array
 689          */
 690         public function getNamespaceIds() {
 691                 if ( is_null( $this->mNamespaceIds ) ) {
 692                         global $wgNamespaceAliases;
 693                         # Put namespace names and aliases into a hashtable.
 694                         # If this is too slow, then we should arrange it so that it is done
 695                         # before caching. The catch is that at pre-cache time, the above
 696                         # class-specific fixup hasn't been done.
 697                         $this->mNamespaceIds = [];
 698                         foreach ( $this->getNamespaces() as $index => $name ) {
 699                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 700                         }
 701                         foreach ( $this->getNamespaceAliases() as $name => $index ) {
 702                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 703                         }
 704                         if ( $wgNamespaceAliases ) {
 705                                 foreach ( $wgNamespaceAliases as $name => $index ) {
 706                                         $this->mNamespaceIds[$this->lc( $name )] = $index;
 707                                 }
 708                         }
 709                 }
 710                 return $this->mNamespaceIds;
 711         }
 712
 713         /**
 714          * Get a namespace key by value, case insensitive.  Canonical namespace
 715          * names override custom ones defined for the current language.
 716          *
 717          * @param string $text
 718          * @return int|bool An integer if $text is a valid value otherwise false
 719          */
 720         public function getNsIndex( $text ) {
 721                 $lctext = $this->lc( $text );
 722                 $ns = MediaWikiServices::getInstance()->getNamespaceInfo()->
 723                         getCanonicalIndex( $lctext );
 724                 if ( $ns !== null ) {
 725                         return $ns;
 726                 }
 727                 $ids = $this->getNamespaceIds();
 728                 return $ids[$lctext] ?? false;
 729         }
 730
 731         /**
 732          * short names for language variants used for language conversion links.
 733          *
 734          * @param string $code
 735          * @param bool $usemsg Use the "variantname-xyz" message if it exists
 736          * @return string
 737          */
 738         public function getVariantname( $code, $usemsg = true ) {
 739                 $msg = "variantname-$code";
 740                 if ( $usemsg && wfMessage( $msg )->exists() ) {
 741                         return $this->getMessageFromDB( $msg );
 742                 }
 743                 $name = $this->langNameUtils->getLanguageName( $code );
 744                 if ( $name ) {
 745                         return $name; # if it's defined as a language name, show that
 746                 } else {
 747                         # otherwise, output the language code
 748                         return $code;
 749                 }
 750         }
 751
 752         /**
 753          * @return string[]|bool List of date format preference keys, or false if disabled.
 754          */
 755         public function getDatePreferences() {
 756                 return $this->localisationCache->getItem( $this->mCode, 'datePreferences' );
 757         }
 758
 759         /**
 760          * @return array
 761          */
 762         function getDateFormats() {
 763                 return $this->localisationCache->getItem( $this->mCode, 'dateFormats' );
 764         }
 765
 766         /**
 767          * @return array|string
 768          */
 769         public function getDefaultDateFormat() {
 770                 $df = $this->localisationCache->getItem( $this->mCode, 'defaultDateFormat' );
 771                 if ( $df === 'dmy or mdy' ) {
 772                         global $wgAmericanDates;
 773                         return $wgAmericanDates ? 'mdy' : 'dmy';
 774                 } else {
 775                         return $df;
 776                 }
 777         }
 778
 779         /**
 780          * @return array
 781          */
 782         public function getDatePreferenceMigrationMap() {
 783                 return $this->localisationCache->getItem( $this->mCode, 'datePreferenceMigrationMap' );
 784         }
 785
 786         /**
 787          * @return array
 788          */
 789         public function getExtraUserToggles() {
 790                 wfDeprecated( __METHOD__, '1.34' );
 791                 return [];
 792         }
 793
 794         /**
 795          * @param string $tog
 796          * @return string
 797          */
 798         function getUserToggle( $tog ) {
 799                 return $this->getMessageFromDB( "tog-$tog" );
 800         }
 801
 802         /**
 803          * Get an array of language names, indexed by code.
 804          *
 805          * @deprecated since 1.34, use LanguageNameUtils::getLanguageNames
 806          * @param null|string $inLanguage Code of language in which to return the names
 807          *              Use self::AS_AUTONYMS for autonyms (native names)
 808          * @param string $include One of:
 809          *              self::ALL all available languages
 810          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 811          *              self::SUPPORTED only if the language is in 'mw' *and* has a message file
 812          * @return array Language code => language name (sorted by key)
 813          * @since 1.20
 814          */
 815         public static function fetchLanguageNames( $inLanguage = self::AS_AUTONYMS, $include = 'mw' ) {
 816                 return MediaWikiServices::getInstance()->getLanguageNameUtils()
 817                         ->getLanguageNames( $inLanguage, $include );
 818         }
 819
 820         /**
 821          * @deprecated since 1.34, use LanguageNameUtils::getLanguageName
 822          * @param string $code The code of the language for which to get the name
 823          * @param null|string $inLanguage Code of language in which to return the name
 824          *   (SELF::AS_AUTONYMS for autonyms)
 825          * @param string $include See fetchLanguageNames()
 826          * @return string Language name or empty
 827          * @since 1.20
 828          */
 829         public static function fetchLanguageName(
 830                 $code,
 831                 $inLanguage = self::AS_AUTONYMS,
 832                 $include = self::ALL
 833         ) {
 834                 return MediaWikiServices::getInstance()->getLanguageNameUtils()
 835                         ->getLanguageName( $code, $inLanguage, $include );
 836         }
 837
 838         /**
 839          * Get a message from the MediaWiki namespace.
 840          *
 841          * @param string $msg Message name
 842          * @return string
 843          */
 844         public function getMessageFromDB( $msg ) {
 845                 return $this->msg( $msg )->text();
 846         }
 847
 848         /**
 849          * Get message object in this language. Only for use inside this class.
 850          *
 851          * @param string $msg Message name
 852          * @return Message
 853          */
 854         protected function msg( $msg ) {
 855                 return wfMessage( $msg )->inLanguage( $this );
 856         }
 857
 858         /**
 859          * @param string $key
 860          * @return string
 861          */
 862         public function getMonthName( $key ) {
 863                 return $this->getMessageFromDB( self::$mMonthMsgs[$key - 1] );
 864         }
 865
 866         /**
 867          * @return array
 868          */
 869         public function getMonthNamesArray() {
 870                 $monthNames = [ '' ];
 871                 for ( $i = 1; $i < 13; $i++ ) {
 872                         $monthNames[] = $this->getMonthName( $i );
 873                 }
 874                 return $monthNames;
 875         }
 876
 877         /**
 878          * @param string $key
 879          * @return string
 880          */
 881         public function getMonthNameGen( $key ) {
 882                 return $this->getMessageFromDB( self::$mMonthGenMsgs[$key - 1] );
 883         }
 884
 885         /**
 886          * @param string $key
 887          * @return string
 888          */
 889         public function getMonthAbbreviation( $key ) {
 890                 return $this->getMessageFromDB( self::$mMonthAbbrevMsgs[$key - 1] );
 891         }
 892
 893         /**
 894          * @return array
 895          */
 896         public function getMonthAbbreviationsArray() {
 897                 $monthNames = [ '' ];
 898                 for ( $i = 1; $i < 13; $i++ ) {
 899                         $monthNames[] = $this->getMonthAbbreviation( $i );
 900                 }
 901                 return $monthNames;
 902         }
 903
 904         /**
 905          * @param string $key
 906          * @return string
 907          */
 908         public function getWeekdayName( $key ) {
 909                 return $this->getMessageFromDB( self::$mWeekdayMsgs[$key - 1] );
 910         }
 911
 912         /**
 913          * @param string $key
 914          * @return string
 915          */
 916         function getWeekdayAbbreviation( $key ) {
 917                 return $this->getMessageFromDB( self::$mWeekdayAbbrevMsgs[$key - 1] );
 918         }
 919
 920         /**
 921          * @param string $key
 922          * @return string
 923          */
 924         function getIranianCalendarMonthName( $key ) {
 925                 return $this->getMessageFromDB( self::$mIranianCalendarMonthMsgs[$key - 1] );
 926         }
 927
 928         /**
 929          * @param string $key
 930          * @return string
 931          */
 932         function getHebrewCalendarMonthName( $key ) {
 933                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthMsgs[$key - 1] );
 934         }
 935
 936         /**
 937          * @param string $key
 938          * @return string
 939          */
 940         function getHebrewCalendarMonthNameGen( $key ) {
 941                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthGenMsgs[$key - 1] );
 942         }
 943
 944         /**
 945          * @param string $key
 946          * @return string
 947          */
 948         function getHijriCalendarMonthName( $key ) {
 949                 return $this->getMessageFromDB( self::$mHijriCalendarMonthMsgs[$key - 1] );
 950         }
 951
 952         /**
 953          * Pass through result from $dateTimeObj->format()
 954          * @param DateTime|bool|null &$dateTimeObj
 955          * @param string $ts
 956          * @param DateTimeZone|bool|null $zone
 957          * @param string $code
 958          * @return string
 959          */
 960         private static function dateTimeObjFormat( &$dateTimeObj, $ts, $zone, $code ) {
 961                 if ( !$dateTimeObj ) {
 962                         $dateTimeObj = DateTime::createFromFormat(
 963                                 'YmdHis', $ts, $zone ?: new DateTimeZone( 'UTC' )
 964                         );
 965                 }
 966                 return $dateTimeObj->format( $code );
 967         }
 968
 969         /**
 970          * This is a workalike of PHP's date() function, but with better
 971          * internationalisation, a reduced set of format characters, and a better
 972          * escaping format.
 973          *
 974          * Supported format characters are dDjlNwzWFmMntLoYyaAgGhHiscrUeIOPTZ. See
 975          * the PHP manual for definitions. There are a number of extensions, which
 976          * start with "x":
 977          *
 978          *    xn   Do not translate digits of the next numeric format character
 979          *    xN   Toggle raw digit (xn) flag, stays set until explicitly unset
 980          *    xr   Use roman numerals for the next numeric format character
 981          *    xh   Use hebrew numerals for the next numeric format character
 982          *    xx   Literal x
 983          *    xg   Genitive month name
 984          *
 985          *    xij  j (day number) in Iranian calendar
 986          *    xiF  F (month name) in Iranian calendar
 987          *    xin  n (month number) in Iranian calendar
 988          *    xiy  y (two digit year) in Iranian calendar
 989          *    xiY  Y (full year) in Iranian calendar
 990          *    xit  t (days in month) in Iranian calendar
 991          *    xiz  z (day of the year) in Iranian calendar
 992          *
 993          *    xjj  j (day number) in Hebrew calendar
 994          *    xjF  F (month name) in Hebrew calendar
 995          *    xjt  t (days in month) in Hebrew calendar
 996          *    xjx  xg (genitive month name) in Hebrew calendar
 997          *    xjn  n (month number) in Hebrew calendar
 998          *    xjY  Y (full year) in Hebrew calendar
 999          *
1000          *    xmj  j (day number) in Hijri calendar
1001          *    xmF  F (month name) in Hijri calendar
1002          *    xmn  n (month number) in Hijri calendar
1003          *    xmY  Y (full year) in Hijri calendar
1004          *
1005          *    xkY  Y (full year) in Thai solar calendar. Months and days are
1006          *                       identical to the Gregorian calendar
1007          *    xoY  Y (full year) in Minguo calendar or Juche year.
1008          *                       Months and days are identical to the
1009          *                       Gregorian calendar
1010          *    xtY  Y (full year) in Japanese nengo. Months and days are
1011          *                       identical to the Gregorian calendar
1012          *
1013          * Characters enclosed in double quotes will be considered literal (with
1014          * the quotes themselves removed). Unmatched quotes will be considered
1015          * literal quotes. Example:
1016          *
1017          * "The month is" F       => The month is January
1018          * i's"                   => 20'11"
1019          *
1020          * Backslash escaping is also supported.
1021          *
1022          * Input timestamp is assumed to be pre-normalized to the desired local
1023          * time zone, if any. Note that the format characters crUeIOPTZ will assume
1024          * $ts is UTC if $zone is not given.
1025          *
1026          * @param string $format
1027          * @param string $ts 14-character timestamp
1028          *      YYYYMMDDHHMMSS
1029          *      01234567890123
1030          * @param DateTimeZone|null $zone Timezone of $ts
1031          * @param int &$ttl The amount of time (in seconds) the output may be cached for.
1032          * Only makes sense if $ts is the current time.
1033          * @todo handling of "o" format character for Iranian, Hebrew, Hijri & Thai?
1034          *
1035          * @throws MWException
1036          * @return string
1037          */
1038         public function sprintfDate( $format, $ts, DateTimeZone $zone = null, &$ttl = 'unused' ) {
1039                 $s = '';
1040                 $raw = false;
1041                 $roman = false;
1042                 $hebrewNum = false;
1043                 $dateTimeObj = false;
1044                 $rawToggle = false;
1045                 $iranian = false;
1046                 $hebrew = false;
1047                 $hijri = false;
1048                 $thai = false;
1049                 $minguo = false;
1050                 $tenno = false;
1051
1052                 $usedSecond = false;
1053                 $usedMinute = false;
1054                 $usedHour = false;
1055                 $usedAMPM = false;
1056                 $usedDay = false;
1057                 $usedWeek = false;
1058                 $usedMonth = false;
1059                 $usedYear = false;
1060                 $usedISOYear = false;
1061                 $usedIsLeapYear = false;
1062
1063                 $usedHebrewMonth = false;
1064                 $usedIranianMonth = false;
1065                 $usedHijriMonth = false;
1066                 $usedHebrewYear = false;
1067                 $usedIranianYear = false;
1068                 $usedHijriYear = false;
1069                 $usedTennoYear = false;
1070
1071                 if ( strlen( $ts ) !== 14 ) {
1072                         throw new MWException( __METHOD__ . ": The timestamp $ts should have 14 characters" );
1073                 }
1074
1075                 if ( !ctype_digit( $ts ) ) {
1076                         throw new MWException( __METHOD__ . ": The timestamp $ts should be a number" );
1077                 }
1078
1079                 $formatLength = strlen( $format );
1080                 for ( $p = 0; $p < $formatLength; $p++ ) {
1081                         $num = false;
1082                         $code = $format[$p];
1083                         if ( $code == 'x' && $p < $formatLength - 1 ) {
1084                                 $code .= $format[++$p];
1085                         }
1086
1087                         if ( ( $code === 'xi'
1088                                         || $code === 'xj'
1089                                         || $code === 'xk'
1090                                         || $code === 'xm'
1091                                         || $code === 'xo'
1092                                         || $code === 'xt' )
1093                                 && $p < $formatLength - 1 ) {
1094                                 $code .= $format[++$p];
1095                         }
1096
1097                         switch ( $code ) {
1098                                 case 'xx':
1099                                         $s .= 'x';
1100                                         break;
1101                                 case 'xn':
1102                                         $raw = true;
1103                                         break;
1104                                 case 'xN':
1105                                         $rawToggle = !$rawToggle;
1106                                         break;
1107                                 case 'xr':
1108                                         $roman = true;
1109                                         break;
1110                                 case 'xh':
1111                                         $hebrewNum = true;
1112                                         break;
1113                                 case 'xg':
1114                                         $usedMonth = true;
1115                                         $s .= $this->getMonthNameGen( substr( $ts, 4, 2 ) );
1116                                         break;
1117                                 case 'xjx':
1118                                         $usedHebrewMonth = true;
1119                                         if ( !$hebrew ) {
1120                                                 $hebrew = self::tsToHebrew( $ts );
1121                                         }
1122                                         $s .= $this->getHebrewCalendarMonthNameGen( $hebrew[1] );
1123                                         break;
1124                                 case 'd':
1125                                         $usedDay = true;
1126                                         $num = substr( $ts, 6, 2 );
1127                                         break;
1128                                 case 'D':
1129                                         $usedDay = true;
1130                                         $s .= $this->getWeekdayAbbreviation(
1131                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1132                                         );
1133                                         break;
1134                                 case 'j':
1135                                         $usedDay = true;
1136                                         $num = intval( substr( $ts, 6, 2 ) );
1137                                         break;
1138                                 case 'xij':
1139                                         $usedDay = true;
1140                                         if ( !$iranian ) {
1141                                                 $iranian = self::tsToIranian( $ts );
1142                                         }
1143                                         $num = $iranian[2];
1144                                         break;
1145                                 case 'xmj':
1146                                         $usedDay = true;
1147                                         if ( !$hijri ) {
1148                                                 $hijri = self::tsToHijri( $ts );
1149                                         }
1150                                         $num = $hijri[2];
1151                                         break;
1152                                 case 'xjj':
1153                                         $usedDay = true;
1154                                         if ( !$hebrew ) {
1155                                                 $hebrew = self::tsToHebrew( $ts );
1156                                         }
1157                                         $num = $hebrew[2];
1158                                         break;
1159                                 case 'l':
1160                                         $usedDay = true;
1161                                         $s .= $this->getWeekdayName(
1162                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1163                                         );
1164                                         break;
1165                                 case 'F':
1166                                         $usedMonth = true;
1167                                         $s .= $this->getMonthName( substr( $ts, 4, 2 ) );
1168                                         break;
1169                                 case 'xiF':
1170                                         $usedIranianMonth = true;
1171                                         if ( !$iranian ) {
1172                                                 $iranian = self::tsToIranian( $ts );
1173                                         }
1174                                         $s .= $this->getIranianCalendarMonthName( $iranian[1] );
1175                                         break;
1176                                 case 'xmF':
1177                                         $usedHijriMonth = true;
1178                                         if ( !$hijri ) {
1179                                                 $hijri = self::tsToHijri( $ts );
1180                                         }
1181                                         $s .= $this->getHijriCalendarMonthName( $hijri[1] );
1182                                         break;
1183                                 case 'xjF':
1184                                         $usedHebrewMonth = true;
1185                                         if ( !$hebrew ) {
1186                                                 $hebrew = self::tsToHebrew( $ts );
1187                                         }
1188                                         $s .= $this->getHebrewCalendarMonthName( $hebrew[1] );
1189                                         break;
1190                                 case 'm':
1191                                         $usedMonth = true;
1192                                         $num = substr( $ts, 4, 2 );
1193                                         break;
1194                                 case 'M':
1195                                         $usedMonth = true;
1196                                         $s .= $this->getMonthAbbreviation( substr( $ts, 4, 2 ) );
1197                                         break;
1198                                 case 'n':
1199                                         $usedMonth = true;
1200                                         $num = intval( substr( $ts, 4, 2 ) );
1201                                         break;
1202                                 case 'xin':
1203                                         $usedIranianMonth = true;
1204                                         if ( !$iranian ) {
1205                                                 $iranian = self::tsToIranian( $ts );
1206                                         }
1207                                         $num = $iranian[1];
1208                                         break;
1209                                 case 'xmn':
1210                                         $usedHijriMonth = true;
1211                                         if ( !$hijri ) {
1212                                                 $hijri = self::tsToHijri( $ts );
1213                                         }
1214                                         $num = $hijri[1];
1215                                         break;
1216                                 case 'xjn':
1217                                         $usedHebrewMonth = true;
1218                                         if ( !$hebrew ) {
1219                                                 $hebrew = self::tsToHebrew( $ts );
1220                                         }
1221                                         $num = $hebrew[1];
1222                                         break;
1223                                 case 'xjt':
1224                                         $usedHebrewMonth = true;
1225                                         if ( !$hebrew ) {
1226                                                 $hebrew = self::tsToHebrew( $ts );
1227                                         }
1228                                         $num = $hebrew[3];
1229                                         break;
1230                                 case 'Y':
1231                                         $usedYear = true;
1232                                         $num = substr( $ts, 0, 4 );
1233                                         break;
1234                                 case 'xiY':
1235                                         $usedIranianYear = true;
1236                                         if ( !$iranian ) {
1237                                                 $iranian = self::tsToIranian( $ts );
1238                                         }
1239                                         $num = $iranian[0];
1240                                         break;
1241                                 case 'xmY':
1242                                         $usedHijriYear = true;
1243                                         if ( !$hijri ) {
1244                                                 $hijri = self::tsToHijri( $ts );
1245                                         }
1246                                         $num = $hijri[0];
1247                                         break;
1248                                 case 'xjY':
1249                                         $usedHebrewYear = true;
1250                                         if ( !$hebrew ) {
1251                                                 $hebrew = self::tsToHebrew( $ts );
1252                                         }
1253                                         $num = $hebrew[0];
1254                                         break;
1255                                 case 'xkY':
1256                                         $usedYear = true;
1257                                         if ( !$thai ) {
1258                                                 $thai = self::tsToYear( $ts, 'thai' );
1259                                         }
1260                                         $num = $thai[0];
1261                                         break;
1262                                 case 'xoY':
1263                                         $usedYear = true;
1264                                         if ( !$minguo ) {
1265                                                 $minguo = self::tsToYear( $ts, 'minguo' );
1266                                         }
1267                                         $num = $minguo[0];
1268                                         break;
1269                                 case 'xtY':
1270                                         $usedTennoYear = true;
1271                                         if ( !$tenno ) {
1272                                                 $tenno = self::tsToYear( $ts, 'tenno' );
1273                                         }
1274                                         $num = $tenno[0];
1275                                         break;
1276                                 case 'y':
1277                                         $usedYear = true;
1278                                         $num = substr( $ts, 2, 2 );
1279                                         break;
1280                                 case 'xiy':
1281                                         $usedIranianYear = true;
1282                                         if ( !$iranian ) {
1283                                                 $iranian = self::tsToIranian( $ts );
1284                                         }
1285                                         $num = substr( $iranian[0], -2 );
1286                                         break;
1287                                 case 'xit':
1288                                         $usedIranianYear = true;
1289                                         if ( !$iranian ) {
1290                                                 $iranian = self::tsToIranian( $ts );
1291                                         }
1292                                         $num = self::$IRANIAN_DAYS[$iranian[1] - 1];
1293                                         break;
1294                                 case 'xiz':
1295                                         $usedIranianYear = true;
1296                                         if ( !$iranian ) {
1297                                                 $iranian = self::tsToIranian( $ts );
1298                                         }
1299                                         $num = $iranian[3];
1300                                         break;
1301                                 case 'a':
1302                                         $usedAMPM = true;
1303                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'am' : 'pm';
1304                                         break;
1305                                 case 'A':
1306                                         $usedAMPM = true;
1307                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'AM' : 'PM';
1308                                         break;
1309                                 case 'g':
1310                                         $usedHour = true;
1311                                         $h = substr( $ts, 8, 2 );
1312                                         $num = $h % 12 ?: 12;
1313                                         break;
1314                                 case 'G':
1315                                         $usedHour = true;
1316                                         $num = intval( substr( $ts, 8, 2 ) );
1317                                         break;
1318                                 case 'h':
1319                                         $usedHour = true;
1320                                         $h = substr( $ts, 8, 2 );
1321                                         $num = sprintf( '%02d', $h % 12 ?: 12 );
1322                                         break;
1323                                 case 'H':
1324                                         $usedHour = true;
1325                                         $num = substr( $ts, 8, 2 );
1326                                         break;
1327                                 case 'i':
1328                                         $usedMinute = true;
1329                                         $num = substr( $ts, 10, 2 );
1330                                         break;
1331                                 case 's':
1332                                         $usedSecond = true;
1333                                         $num = substr( $ts, 12, 2 );
1334                                         break;
1335                                 case 'c':
1336                                 case 'r':
1337                                         $usedSecond = true;
1338                                         // fall through
1339                                 case 'e':
1340                                 case 'O':
1341                                 case 'P':
1342                                 case 'T':
1343                                         $s .= self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1344                                         break;
1345                                 case 'w':
1346                                 case 'N':
1347                                 case 'z':
1348                                         $usedDay = true;
1349                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1350                                         break;
1351                                 case 'W':
1352                                         $usedWeek = true;
1353                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1354                                         break;
1355                                 case 't':
1356                                         $usedMonth = true;
1357                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1358                                         break;
1359                                 case 'L':
1360                                         $usedIsLeapYear = true;
1361                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1362                                         break;
1363                                 case 'o':
1364                                         $usedISOYear = true;
1365                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1366                                         break;
1367                                 case 'U':
1368                                         $usedSecond = true;
1369                                         // fall through
1370                                 case 'I':
1371                                 case 'Z':
1372                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1373                                         break;
1374                                 case '\\':
1375                                         # Backslash escaping
1376                                         if ( $p < $formatLength - 1 ) {
1377                                                 $s .= $format[++$p];
1378                                         } else {
1379                                                 $s .= '\\';
1380                                         }
1381                                         break;
1382                                 case '"':
1383                                         # Quoted literal
1384                                         if ( $p < $formatLength - 1 ) {
1385                                                 $endQuote = strpos( $format, '"', $p + 1 );
1386                                                 if ( $endQuote === false ) {
1387                                                         # No terminating quote, assume literal "
1388                                                         $s .= '"';
1389                                                 } else {
1390                                                         $s .= substr( $format, $p + 1, $endQuote - $p - 1 );
1391                                                         $p = $endQuote;
1392                                                 }
1393                                         } else {
1394                                                 # Quote at end of string, assume literal "
1395                                                 $s .= '"';
1396                                         }
1397                                         break;
1398                                 default:
1399                                         $s .= $format[$p];
1400                         }
1401                         if ( $num !== false ) {
1402                                 if ( $rawToggle || $raw ) {
1403                                         $s .= $num;
1404                                         $raw = false;
1405                                 } elseif ( $roman ) {
1406                                         $s .= self::romanNumeral( $num );
1407                                         $roman = false;
1408                                 } elseif ( $hebrewNum ) {
1409                                         $s .= self::hebrewNumeral( $num );
1410                                         $hebrewNum = false;
1411                                 } else {
1412                                         $s .= $this->formatNum( $num, true );
1413                                 }
1414                         }
1415                 }
1416
1417                 if ( $ttl === 'unused' ) {
1418                         // No need to calculate the TTL, the caller wont use it anyway.
1419                 } elseif ( $usedSecond ) {
1420                         $ttl = 1;
1421                 } elseif ( $usedMinute ) {
1422                         $ttl = 60 - substr( $ts, 12, 2 );
1423                 } elseif ( $usedHour ) {
1424                         $ttl = 3600 - substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1425                 } elseif ( $usedAMPM ) {
1426                         $ttl = 43200 - ( substr( $ts, 8, 2 ) % 12 ) * 3600 -
1427                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1428                 } elseif (
1429                         $usedDay ||
1430                         $usedHebrewMonth ||
1431                         $usedIranianMonth ||
1432                         $usedHijriMonth ||
1433                         $usedHebrewYear ||
1434                         $usedIranianYear ||
1435                         $usedHijriYear ||
1436                         $usedTennoYear
1437                 ) {
1438                         // @todo Someone who understands the non-Gregorian calendars
1439                         // should write proper logic for them so that they don't need purged every day.
1440                         $ttl = 86400 - substr( $ts, 8, 2 ) * 3600 -
1441                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1442                 } else {
1443                         $possibleTtls = [];
1444                         $timeRemainingInDay = 86400 - substr( $ts, 8, 2 ) * 3600 -
1445                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1446                         if ( $usedWeek ) {
1447                                 $possibleTtls[] =
1448                                         ( 7 - self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400 +
1449                                         $timeRemainingInDay;
1450                         } elseif ( $usedISOYear ) {
1451                                 // December 28th falls on the last ISO week of the year, every year.
1452                                 // The last ISO week of a year can be 52 or 53.
1453                                 $lastWeekOfISOYear = DateTime::createFromFormat(
1454                                         'Ymd',
1455                                         substr( $ts, 0, 4 ) . '1228',
1456                                         $zone ?: new DateTimeZone( 'UTC' )
1457                                 )->format( 'W' );
1458                                 $currentISOWeek = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'W' );
1459                                 $weeksRemaining = $lastWeekOfISOYear - $currentISOWeek;
1460                                 $timeRemainingInWeek =
1461                                         ( 7 - self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400
1462                                         + $timeRemainingInDay;
1463                                 $possibleTtls[] = $weeksRemaining * 604800 + $timeRemainingInWeek;
1464                         }
1465
1466                         if ( $usedMonth ) {
1467                                 $possibleTtls[] =
1468                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 't' ) -
1469                                                 substr( $ts, 6, 2 ) ) * 86400
1470                                         + $timeRemainingInDay;
1471                         } elseif ( $usedYear ) {
1472                                 $possibleTtls[] =
1473                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1474                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1475                                         + $timeRemainingInDay;
1476                         } elseif ( $usedIsLeapYear ) {
1477                                 $year = substr( $ts, 0, 4 );
1478                                 $timeRemainingInYear =
1479                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1480                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1481                                         + $timeRemainingInDay;
1482                                 $mod = $year % 4;
1483                                 if ( $mod || ( !( $year % 100 ) && $year % 400 ) ) {
1484                                         // this isn't a leap year. see when the next one starts
1485                                         $nextCandidate = $year - $mod + 4;
1486                                         if ( $nextCandidate % 100 || !( $nextCandidate % 400 ) ) {
1487                                                 $possibleTtls[] = ( $nextCandidate - $year - 1 ) * 365 * 86400 +
1488                                                         $timeRemainingInYear;
1489                                         } else {
1490                                                 $possibleTtls[] = ( $nextCandidate - $year + 3 ) * 365 * 86400 +
1491                                                         $timeRemainingInYear;
1492                                         }
1493                                 } else {
1494                                         // this is a leap year, so the next year isn't
1495                                         $possibleTtls[] = $timeRemainingInYear;
1496                                 }
1497                         }
1498
1499                         if ( $possibleTtls ) {
1500                                 $ttl = min( $possibleTtls );
1501                         }
1502                 }
1503
1504                 return $s;
1505         }
1506
1507         private static $GREG_DAYS = [ 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 ];
1508         private static $IRANIAN_DAYS = [ 31, 31, 31, 31, 31, 31, 30, 30, 30, 30, 30, 29 ];
1509
1510         /**
1511          * Algorithm by Roozbeh Pournader and Mohammad Toossi to convert
1512          * Gregorian dates to Iranian dates. Originally written in C, it
1513          * is released under the terms of GNU Lesser General Public
1514          * License. Conversion to PHP was performed by Niklas Laxström.
1515          *
1516          * Link: http://www.farsiweb.info/jalali/jalali.c
1517          *
1518          * @param string $ts
1519          *
1520          * @return int[]
1521          */
1522         private static function tsToIranian( $ts ) {
1523                 $gy = substr( $ts, 0, 4 ) - 1600;
1524                 $gm = substr( $ts, 4, 2 ) - 1;
1525                 $gd = substr( $ts, 6, 2 ) - 1;
1526
1527                 # Days passed from the beginning (including leap years)
1528                 $gDayNo = 365 * $gy
1529                         + floor( ( $gy + 3 ) / 4 )
1530                         - floor( ( $gy + 99 ) / 100 )
1531                         + floor( ( $gy + 399 ) / 400 );
1532
1533                 // Add days of the past months of this year
1534                 for ( $i = 0; $i < $gm; $i++ ) {
1535                         $gDayNo += self::$GREG_DAYS[$i];
1536                 }
1537
1538                 // Leap years
1539                 if ( $gm > 1 && ( ( $gy % 4 === 0 && $gy % 100 !== 0 || ( $gy % 400 == 0 ) ) ) ) {
1540                         $gDayNo++;
1541                 }
1542
1543                 // Days passed in current month
1544                 $gDayNo += (int)$gd;
1545
1546                 $jDayNo = $gDayNo - 79;
1547
1548                 $jNp = floor( $jDayNo / 12053 );
1549                 $jDayNo %= 12053;
1550
1551                 $jy = 979 + 33 * $jNp + 4 * floor( $jDayNo / 1461 );
1552                 $jDayNo %= 1461;
1553
1554                 if ( $jDayNo >= 366 ) {
1555                         $jy += floor( ( $jDayNo - 1 ) / 365 );
1556                         $jDayNo = floor( ( $jDayNo - 1 ) % 365 );
1557                 }
1558
1559                 $jz = $jDayNo;
1560
1561                 for ( $i = 0; $i < 11 && $jDayNo >= self::$IRANIAN_DAYS[$i]; $i++ ) {
1562                         $jDayNo -= self::$IRANIAN_DAYS[$i];
1563                 }
1564
1565                 $jm = $i + 1;
1566                 $jd = $jDayNo + 1;
1567
1568                 return [ $jy, $jm, $jd, $jz ];
1569         }
1570
1571         /**
1572          * Converting Gregorian dates to Hijri dates.
1573          *
1574          * Based on a PHP-Nuke block by Sharjeel which is released under GNU/GPL license
1575          *
1576          * @see https://phpnuke.org/modules.php?name=News&file=article&sid=8234&mode=thread&order=0&thold=0
1577          *
1578          * @param string $ts
1579          *
1580          * @return int[]
1581          */
1582         private static function tsToHijri( $ts ) {
1583                 $year = substr( $ts, 0, 4 );
1584                 $month = substr( $ts, 4, 2 );
1585                 $day = substr( $ts, 6, 2 );
1586
1587                 $zyr = $year;
1588                 $zd = $day;
1589                 $zm = $month;
1590                 $zy = $zyr;
1591
1592                 if (
1593                         ( $zy > 1582 ) || ( ( $zy == 1582 ) && ( $zm > 10 ) ) ||
1594                         ( ( $zy == 1582 ) && ( $zm == 10 ) && ( $zd > 14 ) )
1595                 ) {
1596                         $zjd = (int)( ( 1461 * ( $zy + 4800 + (int)( ( $zm - 14 ) / 12 ) ) ) / 4 ) +
1597                                         (int)( ( 367 * ( $zm - 2 - 12 * ( (int)( ( $zm - 14 ) / 12 ) ) ) ) / 12 ) -
1598                                         (int)( ( 3 * (int)( ( ( $zy + 4900 + (int)( ( $zm - 14 ) / 12 ) ) / 100 ) ) ) / 4 ) +
1599                                         $zd - 32075;
1600                 } else {
1601                         $zjd = 367 * $zy - (int)( ( 7 * ( $zy + 5001 + (int)( ( $zm - 9 ) / 7 ) ) ) / 4 ) +
1602                                                                 (int)( ( 275 * $zm ) / 9 ) + $zd + 1729777;
1603                 }
1604
1605                 $zl = $zjd - 1948440 + 10632;
1606                 $zn = (int)( ( $zl - 1 ) / 10631 );
1607                 $zl = $zl - 10631 * $zn + 354;
1608                 $zj = ( (int)( ( 10985 - $zl ) / 5316 ) ) * ( (int)( ( 50 * $zl ) / 17719 ) ) +
1609                         ( (int)( $zl / 5670 ) ) * ( (int)( ( 43 * $zl ) / 15238 ) );
1610                 $zl = $zl - ( (int)( ( 30 - $zj ) / 15 ) ) * ( (int)( ( 17719 * $zj ) / 50 ) ) -
1611                         ( (int)( $zj / 16 ) ) * ( (int)( ( 15238 * $zj ) / 43 ) ) + 29;
1612                 $zm = (int)( ( 24 * $zl ) / 709 );
1613                 $zd = $zl - (int)( ( 709 * $zm ) / 24 );
1614                 $zy = 30 * $zn + $zj - 30;
1615
1616                 return [ $zy, $zm, $zd ];
1617         }
1618
1619         /**
1620          * Converting Gregorian dates to Hebrew dates.
1621          *
1622          * Based on a JavaScript code by Abu Mami and Yisrael Hersch
1623          * (abu-mami@kaluach.net, http://www.kaluach.net), who permitted
1624          * to translate the relevant functions into PHP and release them under
1625          * GNU GPL.
1626          *
1627          * The months are counted from Tishrei = 1. In a leap year, Adar I is 13
1628          * and Adar II is 14. In a non-leap year, Adar is 6.
1629          *
1630          * @param string $ts
1631          *
1632          * @return int[]
1633          */
1634         private static function tsToHebrew( $ts ) {
1635                 # Parse date
1636                 $year = substr( $ts, 0, 4 );
1637                 $month = substr( $ts, 4, 2 );
1638                 $day = substr( $ts, 6, 2 );
1639
1640                 # Calculate Hebrew year
1641                 $hebrewYear = $year + 3760;
1642
1643                 # Month number when September = 1, August = 12
1644                 $month += 4;
1645                 if ( $month > 12 ) {
1646                         # Next year
1647                         $month -= 12;
1648                         $year++;
1649                         $hebrewYear++;
1650                 }
1651
1652                 # Calculate day of year from 1 September
1653                 $dayOfYear = $day;
1654                 for ( $i = 1; $i < $month; $i++ ) {
1655                         if ( $i == 6 ) {
1656                                 # February
1657                                 $dayOfYear += 28;
1658                                 # Check if the year is leap
1659                                 if ( $year % 400 == 0 || ( $year % 4 == 0 && $year % 100 > 0 ) ) {
1660                                         $dayOfYear++;
1661                                 }
1662                         } elseif ( $i == 8 || $i == 10 || $i == 1 || $i == 3 ) {
1663                                 $dayOfYear += 30;
1664                         } else {
1665                                 $dayOfYear += 31;
1666                         }
1667                 }
1668
1669                 # Calculate the start of the Hebrew year
1670                 $start = self::hebrewYearStart( $hebrewYear );
1671
1672                 # Calculate next year's start
1673                 if ( $dayOfYear <= $start ) {
1674                         # Day is before the start of the year - it is the previous year
1675                         # Next year's start
1676                         $nextStart = $start;
1677                         # Previous year
1678                         $year--;
1679                         $hebrewYear--;
1680                         # Add days since previous year's 1 September
1681                         $dayOfYear += 365;
1682                         if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1683                                 # Leap year
1684                                 $dayOfYear++;
1685                         }
1686                         # Start of the new (previous) year
1687                         $start = self::hebrewYearStart( $hebrewYear );
1688                 } else {
1689                         # Next year's start
1690                         $nextStart = self::hebrewYearStart( $hebrewYear + 1 );
1691                 }
1692
1693                 # Calculate Hebrew day of year
1694                 $hebrewDayOfYear = $dayOfYear - $start;
1695
1696                 # Difference between year's days
1697                 $diff = $nextStart - $start;
1698                 # Add 12 (or 13 for leap years) days to ignore the difference between
1699                 # Hebrew and Gregorian year (353 at least vs. 365/6) - now the
1700                 # difference is only about the year type
1701                 if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1702                         $diff += 13;
1703                 } else {
1704                         $diff += 12;
1705                 }
1706
1707                 # Check the year pattern, and is leap year
1708                 # 0 means an incomplete year, 1 means a regular year, 2 means a complete year
1709                 # This is mod 30, to work on both leap years (which add 30 days of Adar I)
1710                 # and non-leap years
1711                 $yearPattern = $diff % 30;
1712                 # Check if leap year
1713                 $isLeap = $diff >= 30;
1714
1715                 # Calculate day in the month from number of day in the Hebrew year
1716                 # Don't check Adar - if the day is not in Adar, we will stop before;
1717                 # if it is in Adar, we will use it to check if it is Adar I or Adar II
1718                 $hebrewDay = $hebrewDayOfYear;
1719                 $hebrewMonth = 1;
1720                 $days = 0;
1721                 while ( $hebrewMonth <= 12 ) {
1722                         # Calculate days in this month
1723                         if ( $isLeap && $hebrewMonth == 6 ) {
1724                                 # Leap year - has Adar I, with 30 days, and Adar II, with 29 days
1725                                 $days = 30;
1726                                 if ( $hebrewDay <= $days ) {
1727                                         # Day in Adar I
1728                                         $hebrewMonth = 13;
1729                                 } else {
1730                                         # Subtract the days of Adar I
1731                                         $hebrewDay -= $days;
1732                                         # Try Adar II
1733                                         $days = 29;
1734                                         if ( $hebrewDay <= $days ) {
1735                                                 # Day in Adar II
1736                                                 $hebrewMonth = 14;
1737                                         }
1738                                 }
1739                         } elseif ( $hebrewMonth == 2 && $yearPattern == 2 ) {
1740                                 # Cheshvan in a complete year (otherwise as the rule below)
1741                                 $days = 30;
1742                         } elseif ( $hebrewMonth == 3 && $yearPattern == 0 ) {
1743                                 # Kislev in an incomplete year (otherwise as the rule below)
1744                                 $days = 29;
1745                         } else {
1746                                 # Odd months have 30 days, even have 29
1747                                 $days = 30 - ( $hebrewMonth - 1 ) % 2;
1748                         }
1749                         if ( $hebrewDay <= $days ) {
1750                                 # In the current month
1751                                 break;
1752                         } else {
1753                                 # Subtract the days of the current month
1754                                 $hebrewDay -= $days;
1755                                 # Try in the next month
1756                                 $hebrewMonth++;
1757                         }
1758                 }
1759
1760                 return [ $hebrewYear, $hebrewMonth, $hebrewDay, $days ];
1761         }
1762
1763         /**
1764          * This calculates the Hebrew year start, as days since 1 September.
1765          * Based on Carl Friedrich Gauss algorithm for finding Easter date.
1766          * Used for Hebrew date.
1767          *
1768          * @param int $year
1769          *
1770          * @return string
1771          */
1772         private static function hebrewYearStart( $year ) {
1773                 $a = intval( ( 12 * ( $year - 1 ) + 17 ) % 19 );
1774                 $b = intval( ( $year - 1 ) % 4 );
1775                 $m = 32.044093161144 + 1.5542417966212 * $a + $b / 4.0 - 0.0031777940220923 * ( $year - 1 );
1776                 if ( $m < 0 ) {
1777                         $m--;
1778                 }
1779                 $Mar = intval( $m );
1780                 if ( $m < 0 ) {
1781                         $m++;
1782                 }
1783                 $m -= $Mar;
1784
1785                 $c = intval( ( $Mar + 3 * ( $year - 1 ) + 5 * $b + 5 ) % 7 );
1786                 if ( $c == 0 && $a > 11 && $m >= 0.89772376543210 ) {
1787                         $Mar++;
1788                 } elseif ( $c == 1 && $a > 6 && $m >= 0.63287037037037 ) {
1789                         $Mar += 2;
1790                 } elseif ( $c == 2 || $c == 4 || $c == 6 ) {
1791                         $Mar++;
1792                 }
1793
1794                 $Mar += intval( ( $year - 3761 ) / 100 ) - intval( ( $year - 3761 ) / 400 ) - 24;
1795                 return $Mar;
1796         }
1797
1798         /**
1799          * Algorithm to convert Gregorian dates to Thai solar dates,
1800          * Minguo dates or Minguo dates.
1801          *
1802          * Link: https://en.wikipedia.org/wiki/Thai_solar_calendar
1803          *       https://en.wikipedia.org/wiki/Minguo_calendar
1804          *       https://en.wikipedia.org/wiki/Japanese_era_name
1805          *
1806          * @param string $ts 14-character timestamp
1807          * @param string $cName Calender name
1808          * @return array Converted year, month, day
1809          */
1810         private static function tsToYear( $ts, $cName ) {
1811                 $gy = substr( $ts, 0, 4 );
1812                 $gm = substr( $ts, 4, 2 );
1813                 $gd = substr( $ts, 6, 2 );
1814
1815                 if ( !strcmp( $cName, 'thai' ) ) {
1816                         # Thai solar dates
1817                         # Add 543 years to the Gregorian calendar
1818                         # Months and days are identical
1819                         $gy_offset = $gy + 543;
1820                         # fix for dates between 1912 and 1941
1821                         # https://en.wikipedia.org/?oldid=836596673#New_year
1822                         if ( $gy >= 1912 && $gy <= 1940 ) {
1823                                 if ( $gm <= 3 ) {
1824                                         $gy_offset--;
1825                                 }
1826                                 $gm = ( $gm - 3 ) % 12;
1827                         }
1828                 } elseif ( ( !strcmp( $cName, 'minguo' ) ) || !strcmp( $cName, 'juche' ) ) {
1829                         # Minguo dates
1830                         # Deduct 1911 years from the Gregorian calendar
1831                         # Months and days are identical
1832                         $gy_offset = $gy - 1911;
1833                 } elseif ( !strcmp( $cName, 'tenno' ) ) {
1834                         # Nengō dates up to Meiji period
1835                         # Deduct years from the Gregorian calendar
1836                         # depending on the nengo periods
1837                         # Months and days are identical
1838                         if ( ( $gy < 1912 )
1839                                 || ( ( $gy == 1912 ) && ( $gm < 7 ) )
1840                                 || ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd < 31 ) )
1841                         ) {
1842                                 # Meiji period
1843                                 $gy_gannen = $gy - 1868 + 1;
1844                                 $gy_offset = $gy_gannen;
1845                                 if ( $gy_gannen == 1 ) {
1846                                         $gy_offset = '元';
1847                                 }
1848                                 $gy_offset = '明治' . $gy_offset;
1849                         } elseif (
1850                                 ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd == 31 ) ) ||
1851                                 ( ( $gy == 1912 ) && ( $gm >= 8 ) ) ||
1852                                 ( ( $gy > 1912 ) && ( $gy < 1926 ) ) ||
1853                                 ( ( $gy == 1926 ) && ( $gm < 12 ) ) ||
1854                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd < 26 ) )
1855                         ) {
1856                                 # Taishō period
1857                                 $gy_gannen = $gy - 1912 + 1;
1858                                 $gy_offset = $gy_gannen;
1859                                 if ( $gy_gannen == 1 ) {
1860                                         $gy_offset = '元';
1861                                 }
1862                                 $gy_offset = '大正' . $gy_offset;
1863                         } elseif (
1864                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd >= 26 ) ) ||
1865                                 ( ( $gy > 1926 ) && ( $gy < 1989 ) ) ||
1866                                 ( ( $gy == 1989 ) && ( $gm == 1 ) && ( $gd < 8 ) )
1867                         ) {
1868                                 # Shōwa period
1869                                 $gy_gannen = $gy - 1926 + 1;
1870                                 $gy_offset = $gy_gannen;
1871                                 if ( $gy_gannen == 1 ) {
1872                                         $gy_offset = '元';
1873                                 }
1874                                 $gy_offset = '昭和' . $gy_offset;
1875                         } elseif (
1876                                 ( ( $gy == 1989 ) && ( $gm == 1 ) && ( $gd >= 8 ) ) ||
1877                                 ( ( $gy > 1989 ) && ( $gy < 2019 ) ) ||
1878                                 ( ( $gy == 2019 ) && ( $gm < 5 ) )
1879                         ) {
1880                                 # Heisei period
1881                                 $gy_gannen = $gy - 1989 + 1;
1882                                 $gy_offset = $gy_gannen;
1883                                 if ( $gy_gannen == 1 ) {
1884                                         $gy_offset = '元';
1885                                 }
1886                                 $gy_offset = '平成' . $gy_offset;
1887                         } else {
1888                                 # Reiwa period
1889                                 $gy_gannen = $gy - 2019 + 1;
1890                                 $gy_offset = $gy_gannen;
1891                                 if ( $gy_gannen == 1 ) {
1892                                         $gy_offset = '元';
1893                                 }
1894                                 $gy_offset = '令和' . $gy_offset;
1895                         }
1896                 } else {
1897                         $gy_offset = $gy;
1898                 }
1899
1900                 return [ $gy_offset, $gm, $gd ];
1901         }
1902
1903         /**
1904          * Gets directionality of the first strongly directional codepoint, for embedBidi()
1905          *
1906          * This is the rule the BIDI algorithm uses to determine the directionality of
1907          * paragraphs ( https://www.unicode.org/reports/tr9/#The_Paragraph_Level ) and
1908          * FSI isolates ( https://www.unicode.org/reports/tr9/#Explicit_Directional_Isolates ).
1909          *
1910          * TODO: Does not handle BIDI control characters inside the text.
1911          * TODO: Does not handle unallocated characters.
1912          *
1913          * @param string $text Text to test
1914          * @return null|string Directionality ('ltr' or 'rtl') or null
1915          */
1916         private static function strongDirFromContent( $text = '' ) {
1917                 if ( !preg_match( self::$strongDirRegex, $text, $matches ) ) {
1918                         return null;
1919                 }
1920                 if ( $matches[1] === '' ) {
1921                         return 'rtl';
1922                 }
1923                 return 'ltr';
1924         }
1925
1926         /**
1927          * Roman number formatting up to 10000
1928          *
1929          * @param int $num
1930          *
1931          * @return string
1932          */
1933         static function romanNumeral( $num ) {
1934                 static $table = [
1935                         [ '', 'I', 'II', 'III', 'IV', 'V', 'VI', 'VII', 'VIII', 'IX', 'X' ],
1936                         [ '', 'X', 'XX', 'XXX', 'XL', 'L', 'LX', 'LXX', 'LXXX', 'XC', 'C' ],
1937                         [ '', 'C', 'CC', 'CCC', 'CD', 'D', 'DC', 'DCC', 'DCCC', 'CM', 'M' ],
1938                         [ '', 'M', 'MM', 'MMM', 'MMMM', 'MMMMM', 'MMMMMM', 'MMMMMMM',
1939                                 'MMMMMMMM', 'MMMMMMMMM', 'MMMMMMMMMM' ]
1940                 ];
1941
1942                 $num = intval( $num );
1943                 if ( $num > 10000 || $num <= 0 ) {
1944                         return $num;
1945                 }
1946
1947                 $s = '';
1948                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
1949                         if ( $num >= $pow10 ) {
1950                                 $s .= $table[$i][(int)floor( $num / $pow10 )];
1951                         }
1952                         $num = $num % $pow10;
1953                 }
1954                 return $s;
1955         }
1956
1957         /**
1958          * Hebrew Gematria number formatting up to 9999
1959          *
1960          * @param int $num
1961          *
1962          * @return string
1963          */
1964         static function hebrewNumeral( $num ) {
1965                 static $table = [
1966                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ],
1967                         [ '', 'י', 'כ', 'ל', 'מ', 'נ', 'ס', 'ע', 'פ', 'צ', 'ק' ],
1968                         [ '',
1969                                 [ 'ק' ],
1970                                 [ 'ר' ],
1971                                 [ 'ש' ],
1972                                 [ 'ת' ],
1973                                 [ 'ת', 'ק' ],
1974                                 [ 'ת', 'ר' ],
1975                                 [ 'ת', 'ש' ],
1976                                 [ 'ת', 'ת' ],
1977                                 [ 'ת', 'ת', 'ק' ],
1978                                 [ 'ת', 'ת', 'ר' ],
1979                         ],
1980                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ]
1981                 ];
1982
1983                 $num = intval( $num );
1984                 if ( $num > 9999 || $num <= 0 ) {
1985                         return $num;
1986                 }
1987
1988                 // Round thousands have special notations
1989                 if ( $num === 1000 ) {
1990                         return "א' אלף";
1991                 } elseif ( $num % 1000 === 0 ) {
1992                         return $table[0][$num / 1000] . "' אלפים";
1993                 }
1994
1995                 $letters = [];
1996
1997                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
1998                         if ( $num >= $pow10 ) {
1999                                 if ( $num === 15 || $num === 16 ) {
2000                                         $letters[] = $table[0][9];
2001                                         $letters[] = $table[0][$num - 9];
2002                                         $num = 0;
2003                                 } else {
2004                                         $letters = array_merge(
2005                                                 $letters,
2006                                                 (array)$table[$i][intval( $num / $pow10 )]
2007                                         );
2008
2009                                         if ( $pow10 === 1000 ) {
2010                                                 $letters[] = "'";
2011                                         }
2012                                 }
2013                         }
2014
2015                         $num = $num % $pow10;
2016                 }
2017
2018                 $preTransformLength = count( $letters );
2019                 if ( $preTransformLength === 1 ) {
2020                         // Add geresh (single quote) to one-letter numbers
2021                         $letters[] = "'";
2022                 } else {
2023                         $lastIndex = $preTransformLength - 1;
2024                         $letters[$lastIndex] = str_replace(
2025                                 [ 'כ', 'מ', 'נ', 'פ', 'צ' ],
2026                                 [ 'ך', 'ם', 'ן', 'ף', 'ץ' ],
2027                                 $letters[$lastIndex]
2028                         );
2029
2030                         // Add gershayim (double quote) to multiple-letter numbers,
2031                         // but exclude numbers with only one letter after the thousands
2032                         // (1001-1009, 1020, 1030, 2001-2009, etc.)
2033                         if ( $letters[1] === "'" && $preTransformLength === 3 ) {
2034                                 $letters[] = "'";
2035                         } else {
2036                                 array_splice( $letters, -1, 0, '"' );
2037                         }
2038                 }
2039
2040                 return implode( $letters );
2041         }
2042
2043         /**
2044          * Used by date() and time() to adjust the time output.
2045          *
2046          * @param string $ts The time in date('YmdHis') format
2047          * @param mixed $tz Adjust the time by this amount (default false, mean we
2048          *   get user timecorrection setting)
2049          * @return int
2050          */
2051         public function userAdjust( $ts, $tz = false ) {
2052                 global $wgUser, $wgLocalTZoffset;
2053
2054                 if ( $tz === false ) {
2055                         $tz = $wgUser->getOption( 'timecorrection' );
2056                 }
2057
2058                 $data = explode( '|', $tz, 3 );
2059
2060                 if ( $data[0] == 'ZoneInfo' ) {
2061                         try {
2062                                 $userTZ = new DateTimeZone( $data[2] );
2063                                 $date = new DateTime( $ts, new DateTimeZone( 'UTC' ) );
2064                                 $date->setTimezone( $userTZ );
2065                                 return $date->format( 'YmdHis' );
2066                         } catch ( Exception $e ) {
2067                                 // Unrecognized timezone, default to 'Offset' with the stored offset.
2068                                 $data[0] = 'Offset';
2069                         }
2070                 }
2071
2072                 if ( $data[0] == 'System' || $tz == '' ) {
2073                         # Global offset in minutes.
2074                         $minDiff = $wgLocalTZoffset;
2075                 } elseif ( $data[0] == 'Offset' ) {
2076                         $minDiff = intval( $data[1] );
2077                 } else {
2078                         $data = explode( ':', $tz );
2079                         if ( count( $data ) == 2 ) {
2080                                 $data[0] = intval( $data[0] );
2081                                 $data[1] = intval( $data[1] );
2082                                 $minDiff = abs( $data[0] ) * 60 + $data[1];
2083                                 if ( $data[0] < 0 ) {
2084                                         $minDiff = -$minDiff;
2085                                 }
2086                         } else {
2087                                 $minDiff = intval( $data[0] ) * 60;
2088                         }
2089                 }
2090
2091                 # No difference ? Return time unchanged
2092                 if ( $minDiff == 0 ) {
2093                         return $ts;
2094                 }
2095
2096                 Wikimedia\suppressWarnings(); // E_STRICT system time bitching
2097                 # Generate an adjusted date; take advantage of the fact that mktime
2098                 # will normalize out-of-range values so we don't have to split $minDiff
2099                 # into hours and minutes.
2100                 $t = mktime( (
2101                         (int)substr( $ts, 8, 2 ) ), # Hours
2102                         (int)substr( $ts, 10, 2 ) + $minDiff, # Minutes
2103                         (int)substr( $ts, 12, 2 ), # Seconds
2104                         (int)substr( $ts, 4, 2 ), # Month
2105                         (int)substr( $ts, 6, 2 ), # Day
2106                         (int)substr( $ts, 0, 4 ) ); # Year
2107
2108                 $date = date( 'YmdHis', $t );
2109                 Wikimedia\restoreWarnings();
2110
2111                 return $date;
2112         }
2113
2114         /**
2115          * This is meant to be used by time(), date(), and timeanddate() to get
2116          * the date preference they're supposed to use, it should be used in
2117          * all children.
2118          *
2119          *     function timeanddate([...], $format = true) {
2120          *       $datePreference = $this->dateFormat($format);
2121          *       [...]
2122          *     }
2123          *
2124          * @param int|string|bool $usePrefs If true, the user's preference is used
2125          *   if false, the site/language default is used
2126          *   if int/string, assumed to be a format.
2127          * @return string
2128          */
2129         function dateFormat( $usePrefs = true ) {
2130                 global $wgUser;
2131
2132                 if ( is_bool( $usePrefs ) ) {
2133                         if ( $usePrefs ) {
2134                                 $datePreference = $wgUser->getDatePreference();
2135                         } else {
2136                                 $datePreference = (string)User::getDefaultOption( 'date' );
2137                         }
2138                 } else {
2139                         $datePreference = (string)$usePrefs;
2140                 }
2141
2142                 // return int
2143                 if ( $datePreference == '' ) {
2144                         return 'default';
2145                 }
2146
2147                 return $datePreference;
2148         }
2149
2150         /**
2151          * Get a format string for a given type and preference
2152          * @param string $type May be 'date', 'time', 'both', or 'pretty'.
2153          * @param string $pref The format name as it appears in Messages*.php under
2154          *  $datePreferences.
2155          *
2156          * @since 1.22 New type 'pretty' that provides a more readable timestamp format
2157          *
2158          * @return string
2159          */
2160         function getDateFormatString( $type, $pref ) {
2161                 $wasDefault = false;
2162                 if ( $pref == 'default' ) {
2163                         $wasDefault = true;
2164                         $pref = $this->getDefaultDateFormat();
2165                 }
2166
2167                 if ( !isset( $this->dateFormatStrings[$type][$pref] ) ) {
2168                         $df =
2169                                 $this->localisationCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2170
2171                         if ( $type === 'pretty' && $df === null ) {
2172                                 $df = $this->getDateFormatString( 'date', $pref );
2173                         }
2174
2175                         if ( !$wasDefault && $df === null ) {
2176                                 $pref = $this->getDefaultDateFormat();
2177                                 $df = $this->getLocalisationCache()
2178                                         ->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2179                         }
2180
2181                         $this->dateFormatStrings[$type][$pref] = $df;
2182                 }
2183                 return $this->dateFormatStrings[$type][$pref];
2184         }
2185
2186         /**
2187          * @param string $ts The time format which needs to be turned into a
2188          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2189          * @param bool $adj Whether to adjust the time output according to the
2190          *   user configured offset ($timecorrection)
2191          * @param mixed $format True to use user's date format preference
2192          * @param string|bool $timecorrection The time offset as returned by
2193          *   validateTimeZone() in Special:Preferences
2194          * @return string
2195          */
2196         public function date( $ts, $adj = false, $format = true, $timecorrection = false ) {
2197                 $ts = wfTimestamp( TS_MW, $ts );
2198                 if ( $adj ) {
2199                         $ts = $this->userAdjust( $ts, $timecorrection );
2200                 }
2201                 $df = $this->getDateFormatString( 'date', $this->dateFormat( $format ) );
2202                 return $this->sprintfDate( $df, $ts );
2203         }
2204
2205         /**
2206          * @param string $ts The time format which needs to be turned into a
2207          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2208          * @param bool $adj Whether to adjust the time output according to the
2209          *   user configured offset ($timecorrection)
2210          * @param mixed $format True to use user's date format preference
2211          * @param string|bool $timecorrection The time offset as returned by
2212          *   validateTimeZone() in Special:Preferences
2213          * @return string
2214          */
2215         public function time( $ts, $adj = false, $format = true, $timecorrection = false ) {
2216                 $ts = wfTimestamp( TS_MW, $ts );
2217                 if ( $adj ) {
2218                         $ts = $this->userAdjust( $ts, $timecorrection );
2219                 }
2220                 $df = $this->getDateFormatString( 'time', $this->dateFormat( $format ) );
2221                 return $this->sprintfDate( $df, $ts );
2222         }
2223
2224         /**
2225          * @param string $ts The time format which needs to be turned into a
2226          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2227          * @param bool $adj Whether to adjust the time output according to the
2228          *   user configured offset ($timecorrection)
2229          * @param mixed $format What format to return, if it's false output the
2230          *   default one (default true)
2231          * @param string|bool $timecorrection The time offset as returned by
2232          *   validateTimeZone() in Special:Preferences
2233          * @return string
2234          */
2235         public function timeanddate( $ts, $adj = false, $format = true, $timecorrection = false ) {
2236                 $ts = wfTimestamp( TS_MW, $ts );
2237                 if ( $adj ) {
2238                         $ts = $this->userAdjust( $ts, $timecorrection );
2239                 }
2240                 $df = $this->getDateFormatString( 'both', $this->dateFormat( $format ) );
2241                 return $this->sprintfDate( $df, $ts );
2242         }
2243
2244         /**
2245          * Takes a number of seconds and turns it into a text using values such as hours and minutes.
2246          *
2247          * @since 1.20
2248          *
2249          * @param int $seconds The amount of seconds.
2250          * @param array $chosenIntervals The intervals to enable.
2251          *
2252          * @return string
2253          */
2254         public function formatDuration( $seconds, array $chosenIntervals = [] ) {
2255                 $intervals = $this->getDurationIntervals( $seconds, $chosenIntervals );
2256
2257                 $segments = [];
2258
2259                 foreach ( $intervals as $intervalName => $intervalValue ) {
2260                         // Messages: duration-seconds, duration-minutes, duration-hours, duration-days, duration-weeks,
2261                         // duration-years, duration-decades, duration-centuries, duration-millennia
2262                         $message = wfMessage( 'duration-' . $intervalName )->numParams( $intervalValue );
2263                         $segments[] = $message->inLanguage( $this )->escaped();
2264                 }
2265
2266                 return $this->listToText( $segments );
2267         }
2268
2269         /**
2270          * Takes a number of seconds and returns an array with a set of corresponding intervals.
2271          * For example 65 will be turned into [ minutes => 1, seconds => 5 ].
2272          *
2273          * @since 1.20
2274          *
2275          * @param int $seconds The amount of seconds.
2276          * @param array $chosenIntervals The intervals to enable.
2277          *
2278          * @return array
2279          */
2280         public function getDurationIntervals( $seconds, array $chosenIntervals = [] ) {
2281                 if ( empty( $chosenIntervals ) ) {
2282                         $chosenIntervals = [
2283                                 'millennia',
2284                                 'centuries',
2285                                 'decades',
2286                                 'years',
2287                                 'days',
2288                                 'hours',
2289                                 'minutes',
2290                                 'seconds'
2291                         ];
2292                 }
2293
2294                 $intervals = array_intersect_key( self::$durationIntervals, array_flip( $chosenIntervals ) );
2295                 $sortedNames = array_keys( $intervals );
2296                 $smallestInterval = array_pop( $sortedNames );
2297
2298                 $segments = [];
2299
2300                 foreach ( $intervals as $name => $length ) {
2301                         $value = floor( $seconds / $length );
2302
2303                         if ( $value > 0 || ( $name == $smallestInterval && empty( $segments ) ) ) {
2304                                 $seconds -= $value * $length;
2305                                 $segments[$name] = $value;
2306                         }
2307                 }
2308
2309                 return $segments;
2310         }
2311
2312         /**
2313          * Internal helper function for userDate(), userTime() and userTimeAndDate()
2314          *
2315          * @param string $type Can be 'date', 'time' or 'both'
2316          * @param string $ts The time format which needs to be turned into a
2317          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2318          * @param User $user User object used to get preferences for timezone and format
2319          * @param array $options Array, can contain the following keys:
2320          *   - 'timecorrection': time correction, can have the following values:
2321          *     - true: use user's preference
2322          *     - false: don't use time correction
2323          *     - int: value of time correction in minutes
2324          *   - 'format': format to use, can have the following values:
2325          *     - true: use user's preference
2326          *     - false: use default preference
2327          *     - string: format to use
2328          * @since 1.19
2329          * @return string
2330          */
2331         private function internalUserTimeAndDate( $type, $ts, User $user, array $options ) {
2332                 $ts = wfTimestamp( TS_MW, $ts );
2333                 $options += [ 'timecorrection' => true, 'format' => true ];
2334                 if ( $options['timecorrection'] !== false ) {
2335                         if ( $options['timecorrection'] === true ) {
2336                                 $offset = $user->getOption( 'timecorrection' );
2337                         } else {
2338                                 $offset = $options['timecorrection'];
2339                         }
2340                         $ts = $this->userAdjust( $ts, $offset );
2341                 }
2342                 if ( $options['format'] === true ) {
2343                         $format = $user->getDatePreference();
2344                 } else {
2345                         $format = $options['format'];
2346                 }
2347                 $df = $this->getDateFormatString( $type, $this->dateFormat( $format ) );
2348                 return $this->sprintfDate( $df, $ts );
2349         }
2350
2351         /**
2352          * Get the formatted date for the given timestamp and formatted for
2353          * the given user.
2354          *
2355          * @param mixed $ts Mixed: the time format which needs to be turned into a
2356          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2357          * @param User $user User object used to get preferences for timezone and format
2358          * @param array $options Array, can contain the following keys:
2359          *   - 'timecorrection': time correction, can have the following values:
2360          *     - true: use user's preference
2361          *     - false: don't use time correction
2362          *     - int: value of time correction in minutes
2363          *   - 'format': format to use, can have the following values:
2364          *     - true: use user's preference
2365          *     - false: use default preference
2366          *     - string: format to use
2367          * @since 1.19
2368          * @return string
2369          */
2370         public function userDate( $ts, User $user, array $options = [] ) {
2371                 return $this->internalUserTimeAndDate( 'date', $ts, $user, $options );
2372         }
2373
2374         /**
2375          * Get the formatted time for the given timestamp and formatted for
2376          * the given user.
2377          *
2378          * @param mixed $ts The time format which needs to be turned into a
2379          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2380          * @param User $user User object used to get preferences for timezone and format
2381          * @param array $options Array, can contain the following keys:
2382          *   - 'timecorrection': time correction, can have the following values:
2383          *     - true: use user's preference
2384          *     - false: don't use time correction
2385          *     - int: value of time correction in minutes
2386          *   - 'format': format to use, can have the following values:
2387          *     - true: use user's preference
2388          *     - false: use default preference
2389          *     - string: format to use
2390          * @since 1.19
2391          * @return string
2392          */
2393         public function userTime( $ts, User $user, array $options = [] ) {
2394                 return $this->internalUserTimeAndDate( 'time', $ts, $user, $options );
2395         }
2396
2397         /**
2398          * Get the formatted date and time for the given timestamp and formatted for
2399          * the given user.
2400          *
2401          * @param mixed $ts The time format which needs to be turned into a
2402          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2403          * @param User $user User object used to get preferences for timezone and format
2404          * @param array $options Array, can contain the following keys:
2405          *   - 'timecorrection': time correction, can have the following values:
2406          *     - true: use user's preference
2407          *     - false: don't use time correction
2408          *     - int: value of time correction in minutes
2409          *   - 'format': format to use, can have the following values:
2410          *     - true: use user's preference
2411          *     - false: use default preference
2412          *     - string: format to use
2413          * @since 1.19
2414          * @return string
2415          */
2416         public function userTimeAndDate( $ts, User $user, array $options = [] ) {
2417                 return $this->internalUserTimeAndDate( 'both', $ts, $user, $options );
2418         }
2419
2420         /**
2421          * Get the timestamp in a human-friendly relative format, e.g., "3 days ago".
2422          *
2423          * Determine the difference between the timestamp and the current time, and
2424          * generate a readable timestamp by returning "<N> <units> ago", where the
2425          * largest possible unit is used.
2426          *
2427          * @since 1.26 (Prior to 1.26 method existed but was not meant to be used directly)
2428          *
2429          * @param MWTimestamp $time
2430          * @param MWTimestamp|null $relativeTo The base timestamp to compare to (defaults to now)
2431          * @param User|null $user User the timestamp is being generated for
2432          *  (or null to use main context's user)
2433          * @return string Formatted timestamp
2434          */
2435         public function getHumanTimestamp(
2436                 MWTimestamp $time, MWTimestamp $relativeTo = null, User $user = null
2437         ) {
2438                 if ( $relativeTo === null ) {
2439                         $relativeTo = new MWTimestamp();
2440                 }
2441                 if ( $user === null ) {
2442                         $user = RequestContext::getMain()->getUser();
2443                 }
2444
2445                 // Adjust for the user's timezone.
2446                 $offsetThis = $time->offsetForUser( $user );
2447                 $offsetRel = $relativeTo->offsetForUser( $user );
2448
2449                 $ts = '';
2450                 if ( Hooks::run( 'GetHumanTimestamp', [ &$ts, $time, $relativeTo, $user, $this ] ) ) {
2451                         $ts = $this->getHumanTimestampInternal( $time, $relativeTo, $user );
2452                 }
2453
2454                 // Reset the timezone on the objects.
2455                 $time->timestamp->sub( $offsetThis );
2456                 $relativeTo->timestamp->sub( $offsetRel );
2457
2458                 return $ts;
2459         }
2460
2461         /**
2462          * Convert an MWTimestamp into a pretty human-readable timestamp using
2463          * the given user preferences and relative base time.
2464          *
2465          * @see Language::getHumanTimestamp
2466          * @param MWTimestamp $ts Timestamp to prettify
2467          * @param MWTimestamp $relativeTo Base timestamp
2468          * @param User $user User preferences to use
2469          * @return string Human timestamp
2470          * @since 1.26
2471          */
2472         private function getHumanTimestampInternal(
2473                 MWTimestamp $ts, MWTimestamp $relativeTo, User $user
2474         ) {
2475                 $diff = $ts->diff( $relativeTo );
2476                 $diffDay = (bool)( (int)$ts->timestamp->format( 'w' ) -
2477                         (int)$relativeTo->timestamp->format( 'w' ) );
2478                 $days = $diff->days ?: (int)$diffDay;
2479                 if ( $diff->invert || $days > 5
2480                         && $ts->timestamp->format( 'Y' ) !== $relativeTo->timestamp->format( 'Y' )
2481                 ) {
2482                         // Timestamps are in different years: use full timestamp
2483                         // Also do full timestamp for future dates
2484                         /**
2485                          * @todo FIXME: Add better handling of future timestamps.
2486                          */
2487                         $format = $this->getDateFormatString( 'both', $user->getDatePreference() ?: 'default' );
2488                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2489                 } elseif ( $days > 5 ) {
2490                         // Timestamps are in same year,  but more than 5 days ago: show day and month only.
2491                         $format = $this->getDateFormatString( 'pretty', $user->getDatePreference() ?: 'default' );
2492                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2493                 } elseif ( $days > 1 ) {
2494                         // Timestamp within the past week: show the day of the week and time
2495                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2496                         $weekday = self::$mWeekdayMsgs[$ts->timestamp->format( 'w' )];
2497                         // Messages:
2498                         // sunday-at, monday-at, tuesday-at, wednesday-at, thursday-at, friday-at, saturday-at
2499                         $ts = wfMessage( "$weekday-at" )
2500                                 ->inLanguage( $this )
2501                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2502                                 ->text();
2503                 } elseif ( $days == 1 ) {
2504                         // Timestamp was yesterday: say 'yesterday' and the time.
2505                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2506                         $ts = wfMessage( 'yesterday-at' )
2507                                 ->inLanguage( $this )
2508                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2509                                 ->text();
2510                 } elseif ( $diff->h > 1 || $diff->h == 1 && $diff->i > 30 ) {
2511                         // Timestamp was today, but more than 90 minutes ago: say 'today' and the time.
2512                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2513                         $ts = wfMessage( 'today-at' )
2514                                 ->inLanguage( $this )
2515                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2516                                 ->text();
2517
2518                 // From here on in, the timestamp was soon enough ago so that we can simply say
2519                 // XX units ago, e.g., "2 hours ago" or "5 minutes ago"
2520                 } elseif ( $diff->h == 1 ) {
2521                         // Less than 90 minutes, but more than an hour ago.
2522                         $ts = wfMessage( 'hours-ago' )->inLanguage( $this )->numParams( 1 )->text();
2523                 } elseif ( $diff->i >= 1 ) {
2524                         // A few minutes ago.
2525                         $ts = wfMessage( 'minutes-ago' )->inLanguage( $this )->numParams( $diff->i )->text();
2526                 } elseif ( $diff->s >= 30 ) {
2527                         // Less than a minute, but more than 30 sec ago.
2528                         $ts = wfMessage( 'seconds-ago' )->inLanguage( $this )->numParams( $diff->s )->text();
2529                 } else {
2530                         // Less than 30 seconds ago.
2531                         $ts = wfMessage( 'just-now' )->text();
2532                 }
2533
2534                 return $ts;
2535         }
2536
2537         /**
2538          * @param string $key
2539          * @return string|null
2540          */
2541         public function getMessage( $key ) {
2542                 return $this->localisationCache->getSubitem( $this->mCode, 'messages', $key );
2543         }
2544
2545         /**
2546          * @return array
2547          */
2548         function getAllMessages() {
2549                 return $this->localisationCache->getItem( $this->mCode, 'messages' );
2550         }
2551
2552         /**
2553          * @param string $in
2554          * @param string $out
2555          * @param string $string
2556          * @return string
2557          */
2558         public function iconv( $in, $out, $string ) {
2559                 # Even with //IGNORE iconv can whine about illegal characters in
2560                 # *input* string. We just ignore those too.
2561                 # REF: https://bugs.php.net/bug.php?id=37166
2562                 # REF: https://phabricator.wikimedia.org/T18885
2563                 Wikimedia\suppressWarnings();
2564                 $text = iconv( $in, $out . '//IGNORE', $string );
2565                 Wikimedia\restoreWarnings();
2566                 return $text;
2567         }
2568
2569         // callback functions for ucwords(), ucwordbreaks()
2570
2571         /**
2572          * @param array $matches
2573          * @return mixed|string
2574          */
2575         function ucwordbreaksCallbackAscii( $matches ) {
2576                 return $this->ucfirst( $matches[1] );
2577         }
2578
2579         /**
2580          * @param array $matches
2581          * @return string
2582          */
2583         function ucwordbreaksCallbackMB( $matches ) {
2584                 return mb_strtoupper( $matches[0] );
2585         }
2586
2587         /**
2588          * @param array $matches
2589          * @return string
2590          */
2591         function ucwordsCallbackMB( $matches ) {
2592                 return mb_strtoupper( $matches[0] );
2593         }
2594
2595         /**
2596          * Make a string's first character uppercase
2597          *
2598          * @param string $str
2599          *
2600          * @return string
2601          */
2602         public function ucfirst( $str ) {
2603                 $o = ord( $str );
2604                 if ( $o < 96 ) { // if already uppercase...
2605                         return $str;
2606                 } elseif ( $o < 128 ) {
2607                         return ucfirst( $str ); // use PHP's ucfirst()
2608                 } else {
2609                         // fall back to more complex logic in case of multibyte strings
2610                         return $this->uc( $str, true );
2611                 }
2612         }
2613
2614         /**
2615          * Convert a string to uppercase
2616          *
2617          * @param string $str
2618          * @param bool $first
2619          *
2620          * @return string
2621          */
2622         public function uc( $str, $first = false ) {
2623                 if ( $first ) {
2624                         if ( $this->isMultibyte( $str ) ) {
2625                                 return $this->mbUpperChar( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2626                         } else {
2627                                 return ucfirst( $str );
2628                         }
2629                 } else {
2630                         return $this->isMultibyte( $str ) ? mb_strtoupper( $str ) : strtoupper( $str );
2631                 }
2632         }
2633
2634         /**
2635          * Convert character to uppercase, allowing overrides of the default mb_upper
2636          * behaviour, which is buggy in many ways. Having a conversion table can be
2637          * useful during transitions between PHP versions where unicode changes happen.
2638          * This can make some resources unreachable on-wiki, see discussion at T219279.
2639          * Providing such a conversion table can allow to manage the transition period.
2640          *
2641          * @since 1.34
2642          *
2643          * @param string $char
2644          *
2645          * @return string
2646          */
2647         protected function mbUpperChar( $char ) {
2648                 global $wgOverrideUcfirstCharacters;
2649                 if ( array_key_exists( $char, $wgOverrideUcfirstCharacters ) ) {
2650                         return $wgOverrideUcfirstCharacters[$char];
2651                 } else {
2652                         return mb_strtoupper( $char );
2653                 }
2654         }
2655
2656         /**
2657          * @param string $str
2658          * @return mixed|string
2659          */
2660         function lcfirst( $str ) {
2661                 $o = ord( $str );
2662                 if ( !$o ) {
2663                         return strval( $str );
2664                 } elseif ( $o >= 128 ) {
2665                         return $this->lc( $str, true );
2666                 } elseif ( $o > 96 ) {
2667                         return $str;
2668                 } else {
2669                         $str[0] = strtolower( $str[0] );
2670                         return $str;
2671                 }
2672         }
2673
2674         /**
2675          * @param string $str
2676          * @param bool $first
2677          * @return mixed|string
2678          */
2679         function lc( $str, $first = false ) {
2680                 if ( $first ) {
2681                         if ( $this->isMultibyte( $str ) ) {
2682                                 return mb_strtolower( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2683                         } else {
2684                                 return strtolower( substr( $str, 0, 1 ) ) . substr( $str, 1 );
2685                         }
2686                 } else {
2687                         return $this->isMultibyte( $str ) ? mb_strtolower( $str ) : strtolower( $str );
2688                 }
2689         }
2690
2691         /**
2692          * @param string $str
2693          * @return bool
2694          */
2695         function isMultibyte( $str ) {
2696                 return strlen( $str ) !== mb_strlen( $str );
2697         }
2698
2699         /**
2700          * @param string $str
2701          * @return mixed|string
2702          */
2703         function ucwords( $str ) {
2704                 if ( $this->isMultibyte( $str ) ) {
2705                         $str = $this->lc( $str );
2706
2707                         // regexp to find first letter in each word (i.e. after each space)
2708                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)| ([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2709
2710                         // function to use to capitalize a single char
2711                         return preg_replace_callback(
2712                                 $replaceRegexp,
2713                                 [ $this, 'ucwordsCallbackMB' ],
2714                                 $str
2715                         );
2716                 } else {
2717                         return ucwords( strtolower( $str ) );
2718                 }
2719         }
2720
2721         /**
2722          * capitalize words at word breaks
2723          *
2724          * @param string $str
2725          * @return mixed
2726          */
2727         function ucwordbreaks( $str ) {
2728                 if ( $this->isMultibyte( $str ) ) {
2729                         $str = $this->lc( $str );
2730
2731                         // since \b doesn't work for UTF-8, we explicitely define word break chars
2732                         $breaks = "[ \-\(\)\}\{\.,\?!]";
2733
2734                         // find first letter after word break
2735                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)|" .
2736                                 "$breaks([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2737
2738                         return preg_replace_callback(
2739                                 $replaceRegexp,
2740                                 [ $this, 'ucwordbreaksCallbackMB' ],
2741                                 $str
2742                         );
2743                 } else {
2744                         return preg_replace_callback(
2745                                 '/\b([\w\x80-\xff]+)\b/',
2746                                 [ $this, 'ucwordbreaksCallbackAscii' ],
2747                                 $str
2748                         );
2749                 }
2750         }
2751
2752         /**
2753          * Return a case-folded representation of $s
2754          *
2755          * This is a representation such that caseFold($s1)==caseFold($s2) if $s1
2756          * and $s2 are the same except for the case of their characters. It is not
2757          * necessary for the value returned to make sense when displayed.
2758          *
2759          * Do *not* perform any other normalisation in this function. If a caller
2760          * uses this function when it should be using a more general normalisation
2761          * function, then fix the caller.
2762          *
2763          * @param string $s
2764          *
2765          * @return string
2766          */
2767         function caseFold( $s ) {
2768                 return $this->uc( $s );
2769         }
2770
2771         /**
2772          * TODO: $s is not always a string per T218883
2773          * @param string $s
2774          * @return string
2775          */
2776         function checkTitleEncoding( $s ) {
2777                 if ( is_array( $s ) ) {
2778                         throw new MWException( 'Given array to checkTitleEncoding.' );
2779                 }
2780                 if ( StringUtils::isUtf8( $s ) ) {
2781                         return $s;
2782                 }
2783
2784                 return $this->iconv( $this->fallback8bitEncoding(), 'utf-8', $s );
2785         }
2786
2787         /**
2788          * @return string
2789          */
2790         function fallback8bitEncoding() {
2791                 return $this->localisationCache->getItem( $this->mCode, 'fallback8bitEncoding' );
2792         }
2793
2794         /**
2795          * Most writing systems use whitespace to break up words.
2796          * Some languages such as Chinese don't conventionally do this,
2797          * which requires special handling when breaking up words for
2798          * searching etc.
2799          *
2800          * @return bool
2801          */
2802         function hasWordBreaks() {
2803                 return true;
2804         }
2805
2806         /**
2807          * Some languages such as Chinese require word segmentation,
2808          * Specify such segmentation when overridden in derived class.
2809          *
2810          * @param string $string
2811          * @return string
2812          */
2813         function segmentByWord( $string ) {
2814                 return $string;
2815         }
2816
2817         /**
2818          * Some languages have special punctuation need to be normalized.
2819          * Make such changes here.
2820          *
2821          * @param string $string
2822          * @return string
2823          */
2824         function normalizeForSearch( $string ) {
2825                 return self::convertDoubleWidth( $string );
2826         }
2827
2828         /**
2829          * convert double-width roman characters to single-width.
2830          * range: ff00-ff5f ~= 0020-007f
2831          *
2832          * @param string $string
2833          *
2834          * @return string
2835          */
2836         protected static function convertDoubleWidth( $string ) {
2837                 static $full = null;
2838                 static $half = null;
2839
2840                 if ( $full === null ) {
2841                         $fullWidth = "０１２３４５６７８９ＡＢＣＤＥＦＧＨＩＪＫＬＭＮＯＰＱＲＳＴＵＶＷＸＹＺａｂｃｄｅｆｇｈｉｊｋｌｍｎｏｐｑｒｓｔｕｖｗｘｙｚ";
2842                         $halfWidth = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz";
2843                         $full = str_split( $fullWidth, 3 );
2844                         $half = str_split( $halfWidth );
2845                 }
2846
2847                 $string = str_replace( $full, $half, $string );
2848                 return $string;
2849         }
2850
2851         /**
2852          * @param string $string
2853          * @param string $pattern
2854          * @return string
2855          */
2856         protected static function insertSpace( $string, $pattern ) {
2857                 $string = preg_replace( $pattern, " $1 ", $string );
2858                 $string = preg_replace( '/ +/', ' ', $string );
2859                 return $string;
2860         }
2861
2862         /**
2863          * @param string[] $termsArray
2864          * @return string[]
2865          */
2866         function convertForSearchResult( $termsArray ) {
2867                 # some languages, e.g. Chinese, need to do a conversion
2868                 # in order for search results to be displayed correctly
2869                 return $termsArray;
2870         }
2871
2872         /**
2873          * Get the first character of a string.
2874          *
2875          * @param string $s
2876          * @return string
2877          */
2878         function firstChar( $s ) {
2879                 $matches = [];
2880                 preg_match(
2881                         '/^([\x00-\x7f]|[\xc0-\xdf][\x80-\xbf]|' .
2882                                 '[\xe0-\xef][\x80-\xbf]{2}|[\xf0-\xf7][\x80-\xbf]{3})/',
2883                         $s,
2884                         $matches
2885                 );
2886
2887                 if ( isset( $matches[1] ) ) {
2888                         if ( strlen( $matches[1] ) != 3 ) {
2889                                 return $matches[1];
2890                         }
2891
2892                         // Break down Hangul syllables to grab the first jamo
2893                         $code = UtfNormal\Utils::utf8ToCodepoint( $matches[1] );
2894                         if ( $code < 0xac00 || 0xd7a4 <= $code ) {
2895                                 return $matches[1];
2896                         } elseif ( $code < 0xb098 ) {
2897                                 return "\u{3131}";
2898                         } elseif ( $code < 0xb2e4 ) {
2899                                 return "\u{3134}";
2900                         } elseif ( $code < 0xb77c ) {
2901                                 return "\u{3137}";
2902                         } elseif ( $code < 0xb9c8 ) {
2903                                 return "\u{3139}";
2904                         } elseif ( $code < 0xbc14 ) {
2905                                 return "\u{3141}";
2906                         } elseif ( $code < 0xc0ac ) {
2907                                 return "\u{3142}";
2908                         } elseif ( $code < 0xc544 ) {
2909                                 return "\u{3145}";
2910                         } elseif ( $code < 0xc790 ) {
2911                                 return "\u{3147}";
2912                         } elseif ( $code < 0xcc28 ) {
2913                                 return "\u{3148}";
2914                         } elseif ( $code < 0xce74 ) {
2915                                 return "\u{314A}";
2916                         } elseif ( $code < 0xd0c0 ) {
2917                                 return "\u{314B}";
2918                         } elseif ( $code < 0xd30c ) {
2919                                 return "\u{314C}";
2920                         } elseif ( $code < 0xd558 ) {
2921                                 return "\u{314D}";
2922                         } else {
2923                                 return "\u{314E}";
2924                         }
2925                 } else {
2926                         return '';
2927                 }
2928         }
2929
2930         /**
2931          * Convert a UTF-8 string to normal form C. In Malayalam and Arabic, this
2932          * also cleans up certain backwards-compatible sequences, converting them
2933          * to the modern Unicode equivalent.
2934          *
2935          * This is language-specific for performance reasons only.
2936          *
2937          * @param string $s
2938          *
2939          * @return string
2940          */
2941         public function normalize( $s ) {
2942                 global $wgAllUnicodeFixes;
2943                 $s = UtfNormal\Validator::cleanUp( $s );
2944                 if ( $wgAllUnicodeFixes ) {
2945                         $s = $this->transformUsingPairFile( 'normalize-ar.php', $s );
2946                         $s = $this->transformUsingPairFile( 'normalize-ml.php', $s );
2947                 }
2948
2949                 return $s;
2950         }
2951
2952         /**
2953          * Transform a string using serialized data stored in the given file (which
2954          * must be in the serialized subdirectory of $IP). The file contains pairs
2955          * mapping source characters to destination characters.
2956          *
2957          * The data is cached in process memory. This will go faster if you have the
2958          * FastStringSearch extension.
2959          *
2960          * @param string $file
2961          * @param string $string
2962          *
2963          * @throws MWException
2964          * @return string
2965          */
2966         protected function transformUsingPairFile( $file, $string ) {
2967                 if ( !isset( $this->transformData[$file] ) ) {
2968                         global $IP;
2969                         $data = require "$IP/languages/data/{$file}";
2970                         $this->transformData[$file] = new ReplacementArray( $data );
2971                 }
2972                 return $this->transformData[$file]->replace( $string );
2973         }
2974
2975         /**
2976          * For right-to-left language support
2977          *
2978          * @return bool
2979          */
2980         function isRTL() {
2981                 return $this->localisationCache->getItem( $this->mCode, 'rtl' );
2982         }
2983
2984         /**
2985          * Return the correct HTML 'dir' attribute value for this language.
2986          * @return string
2987          */
2988         function getDir() {
2989                 return $this->isRTL() ? 'rtl' : 'ltr';
2990         }
2991
2992         /**
2993          * Return 'left' or 'right' as appropriate alignment for line-start
2994          * for this language's text direction.
2995          *
2996          * Should be equivalent to CSS3 'start' text-align value....
2997          *
2998          * @return string
2999          */
3000         function alignStart() {
3001                 return $this->isRTL() ? 'right' : 'left';
3002         }
3003
3004         /**
3005          * Return 'right' or 'left' as appropriate alignment for line-end
3006          * for this language's text direction.
3007          *
3008          * Should be equivalent to CSS3 'end' text-align value....
3009          *
3010          * @return string
3011          */
3012         function alignEnd() {
3013                 return $this->isRTL() ? 'left' : 'right';
3014         }
3015
3016         /**
3017          * A hidden direction mark (LRM or RLM), depending on the language direction.
3018          * Unlike getDirMark(), this function returns the character as an HTML entity.
3019          * This function should be used when the output is guaranteed to be HTML,
3020          * because it makes the output HTML source code more readable. When
3021          * the output is plain text or can be escaped, getDirMark() should be used.
3022          *
3023          * @param bool $opposite Get the direction mark opposite to your language
3024          * @return string
3025          * @since 1.20
3026          */
3027         function getDirMarkEntity( $opposite = false ) {
3028                 if ( $opposite ) {
3029                         return $this->isRTL() ? '&lrm;' : '&rlm;';
3030                 }
3031                 return $this->isRTL() ? '&rlm;' : '&lrm;';
3032         }
3033
3034         /**
3035          * A hidden direction mark (LRM or RLM), depending on the language direction.
3036          * This function produces them as invisible Unicode characters and
3037          * the output may be hard to read and debug, so it should only be used
3038          * when the output is plain text or can be escaped. When the output is
3039          * HTML, use getDirMarkEntity() instead.
3040          *
3041          * @param bool $opposite Get the direction mark opposite to your language
3042          * @return string
3043          */
3044         function getDirMark( $opposite = false ) {
3045                 $lrm = "\u{200E}"; # LEFT-TO-RIGHT MARK, commonly abbreviated LRM
3046                 $rlm = "\u{200F}"; # RIGHT-TO-LEFT MARK, commonly abbreviated RLM
3047                 if ( $opposite ) {
3048                         return $this->isRTL() ? $lrm : $rlm;
3049                 }
3050                 return $this->isRTL() ? $rlm : $lrm;
3051         }
3052
3053         /**
3054          * @return array
3055          */
3056         function capitalizeAllNouns() {
3057                 return $this->localisationCache->getItem( $this->mCode, 'capitalizeAllNouns' );
3058         }
3059
3060         /**
3061          * An arrow, depending on the language direction.
3062          *
3063          * @param string $direction The direction of the arrow: forwards (default),
3064          *   backwards, left, right, up, down.
3065          * @return string
3066          */
3067         function getArrow( $direction = 'forwards' ) {
3068                 switch ( $direction ) {
3069                         case 'forwards':
3070                                 return $this->isRTL() ? '←' : '→';
3071                         case 'backwards':
3072                                 return $this->isRTL() ? '→' : '←';
3073                         case 'left':
3074                                 return '←';
3075                         case 'right':
3076                                 return '→';
3077                         case 'up':
3078                                 return '↑';
3079                         case 'down':
3080                                 return '↓';
3081                 }
3082         }
3083
3084         /**
3085          * To allow "foo[[bar]]" to extend the link over the whole word "foobar"
3086          *
3087          * @return bool
3088          */
3089         function linkPrefixExtension() {
3090                 return $this->localisationCache->getItem( $this->mCode, 'linkPrefixExtension' );
3091         }
3092
3093         /**
3094          * Get all magic words from cache.
3095          * @return array
3096          */
3097         function getMagicWords() {
3098                 return $this->localisationCache->getItem( $this->mCode, 'magicWords' );
3099         }
3100
3101         /**
3102          * Fill a MagicWord object with data from here
3103          *
3104          * @param MagicWord $mw
3105          */
3106         function getMagic( $mw ) {
3107                 $rawEntry = $this->mMagicExtensions[$mw->mId] ??
3108                         $this->localisationCache->getSubitem( $this->mCode, 'magicWords', $mw->mId );
3109
3110                 if ( !is_array( $rawEntry ) ) {
3111                         wfWarn( "\"$rawEntry\" is not a valid magic word for \"$mw->mId\"" );
3112                 } else {
3113                         $mw->mCaseSensitive = $rawEntry[0];
3114                         $mw->mSynonyms = array_slice( $rawEntry, 1 );
3115                 }
3116         }
3117
3118         /**
3119          * Add magic words to the extension array
3120          *
3121          * @param array $newWords
3122          */
3123         function addMagicWordsByLang( $newWords ) {
3124                 $fallbackChain = $this->getFallbackLanguages();
3125                 $fallbackChain = array_reverse( $fallbackChain );
3126                 foreach ( $fallbackChain as $code ) {
3127                         if ( isset( $newWords[$code] ) ) {
3128                                 $this->mMagicExtensions = $newWords[$code] + $this->mMagicExtensions;
3129                         }
3130                 }
3131         }
3132
3133         /**
3134          * Get special page names, as an associative array
3135          *   canonical name => array of valid names, including aliases
3136          * @return array
3137          */
3138         function getSpecialPageAliases() {
3139                 // Cache aliases because it may be slow to load them
3140                 if ( is_null( $this->mExtendedSpecialPageAliases ) ) {
3141                         // Initialise array
3142                         $this->mExtendedSpecialPageAliases =
3143                                 $this->localisationCache->getItem( $this->mCode, 'specialPageAliases' );
3144                 }
3145
3146                 return $this->mExtendedSpecialPageAliases;
3147         }
3148
3149         /**
3150          * Italic is unsuitable for some languages
3151          *
3152          * @param string $text The text to be emphasized.
3153          * @return string
3154          */
3155         function emphasize( $text ) {
3156                 return "<em>$text</em>";
3157         }
3158
3159         /**
3160          * Normally we output all numbers in plain en_US style, that is
3161          * 293,291.235 for twohundredninetythreethousand-twohundredninetyone
3162          * point twohundredthirtyfive. However this is not suitable for all
3163          * languages, some such as Bengali (bn) want ২,৯৩,২৯১.২৩৫ and others such as
3164          * Icelandic just want to use commas instead of dots, and dots instead
3165          * of commas like "293.291,235".
3166          *
3167          * An example of this function being called:
3168          * <code>
3169          * wfMessage( 'message' )->numParams( $num )->text()
3170          * </code>
3171          *
3172          * See $separatorTransformTable on MessageIs.php for
3173          * the , => . and . => , implementation.
3174          *
3175          * @todo check if it's viable to use localeconv() for the decimal separator thing.
3176          * @param int|float $number The string to be formatted, should be an integer
3177          *   or a floating point number.
3178          * @param bool $nocommafy Set to true for special numbers like dates
3179          * @return string
3180          */
3181         public function formatNum( $number, $nocommafy = false ) {
3182                 global $wgTranslateNumerals;
3183                 if ( !$nocommafy ) {
3184                         $number = $this->commafy( $number );
3185                         $s = $this->separatorTransformTable();
3186                         if ( $s ) {
3187                                 $number = strtr( $number, $s );
3188                         }
3189                 }
3190
3191                 if ( $wgTranslateNumerals ) {
3192                         $s = $this->digitTransformTable();
3193                         if ( $s ) {
3194                                 $number = strtr( $number, $s );
3195                         }
3196                 }
3197
3198                 return (string)$number;
3199         }
3200
3201         /**
3202          * Front-end for non-commafied formatNum
3203          *
3204          * @param int|float $number The string to be formatted, should be an integer
3205          *        or a floating point number.
3206          * @since 1.21
3207          * @return string
3208          */
3209         public function formatNumNoSeparators( $number ) {
3210                 return $this->formatNum( $number, true );
3211         }
3212
3213         /**
3214          * @param string $number
3215          * @return string
3216          */
3217         public function parseFormattedNumber( $number ) {
3218                 $s = $this->digitTransformTable();
3219                 if ( $s ) {
3220                         // eliminate empty array values such as ''. (T66347)
3221                         $s = array_filter( $s );
3222                         $number = strtr( $number, array_flip( $s ) );
3223                 }
3224
3225                 $s = $this->separatorTransformTable();
3226                 if ( $s ) {
3227                         // eliminate empty array values such as ''. (T66347)
3228                         $s = array_filter( $s );
3229                         $number = strtr( $number, array_flip( $s ) );
3230                 }
3231
3232                 $number = strtr( $number, [ ',' => '' ] );
3233                 return $number;
3234         }
3235
3236         /**
3237          * Adds commas to a given number
3238          * @since 1.19
3239          * @param mixed $number
3240          * @return string
3241          */
3242         function commafy( $number ) {
3243                 $digitGroupingPattern = $this->digitGroupingPattern();
3244                 $minimumGroupingDigits = $this->minimumGroupingDigits();
3245                 if ( $number === null ) {
3246                         return '';
3247                 }
3248
3249                 if ( !$digitGroupingPattern || $digitGroupingPattern === "###,###,###" ) {
3250                         // Default grouping is at thousands, use the same for ###,###,### pattern too.
3251                         // In some languages it's conventional not to insert a thousands separator
3252                         // in numbers that are four digits long (1000-9999).
3253                         if ( $minimumGroupingDigits ) {
3254                                 // Number of '#' characters after last comma in the grouping pattern.
3255                                 // The pattern is hardcoded here, but this would vary for different patterns.
3256                                 $primaryGroupingSize = 3;
3257                                 // Maximum length of a number to suppress digit grouping for.
3258                                 $maximumLength = $minimumGroupingDigits + $primaryGroupingSize - 1;
3259                                 if ( preg_match( '/^\-?\d{1,' . $maximumLength . '}(\.\d+)?$/', $number ) ) {
3260                                         return $number;
3261                                 }
3262                         }
3263                         return strrev( (string)preg_replace( '/(\d{3})(?=\d)(?!\d*\.)/', '$1,', strrev( $number ) ) );
3264                 } else {
3265                         // Ref: http://cldr.unicode.org/translation/number-patterns
3266                         $sign = "";
3267                         if ( intval( $number ) < 0 ) {
3268                                 // For negative numbers apply the algorithm like positive number and add sign.
3269                                 $sign = "-";
3270                                 $number = substr( $number, 1 );
3271                         }
3272                         $integerPart = [];
3273                         $decimalPart = [];
3274                         $numMatches = preg_match_all( "/(#+)/", $digitGroupingPattern, $matches );
3275                         preg_match( "/\d+/", $number, $integerPart );
3276                         preg_match( "/\.\d*/", $number, $decimalPart );
3277                         $groupedNumber = ( count( $decimalPart ) > 0 ) ? $decimalPart[0] : "";
3278                         if ( $groupedNumber === $number ) {
3279                                 // the string does not have any number part. Eg: .12345
3280                                 return $sign . $groupedNumber;
3281                         }
3282                         $start = $end = ( $integerPart ) ? strlen( $integerPart[0] ) : 0;
3283                         while ( $start > 0 ) {
3284                                 $match = $matches[0][$numMatches - 1];
3285                                 $matchLen = strlen( $match );
3286                                 $start = $end - $matchLen;
3287                                 if ( $start < 0 ) {
3288                                         $start = 0;
3289                                 }
3290                                 $groupedNumber = substr( $number, $start, $end - $start ) . $groupedNumber;
3291                                 $end = $start;
3292                                 if ( $numMatches > 1 ) {
3293                                         // use the last pattern for the rest of the number
3294                                         $numMatches--;
3295                                 }
3296                                 if ( $start > 0 ) {
3297                                         $groupedNumber = "," . $groupedNumber;
3298                                 }
3299                         }
3300                         return $sign . $groupedNumber;
3301                 }
3302         }
3303
3304         /**
3305          * @return string
3306          */
3307         function digitGroupingPattern() {
3308                 return $this->localisationCache->getItem( $this->mCode, 'digitGroupingPattern' );
3309         }
3310
3311         /**
3312          * @return array
3313          */
3314         function digitTransformTable() {
3315                 return $this->localisationCache->getItem( $this->mCode, 'digitTransformTable' );
3316         }
3317
3318         /**
3319          * @return array
3320          */
3321         function separatorTransformTable() {
3322                 return $this->localisationCache->getItem( $this->mCode, 'separatorTransformTable' );
3323         }
3324
3325         /**
3326          * @return int|null
3327          */
3328         function minimumGroupingDigits() {
3329                 return $this->localisationCache->getItem( $this->mCode, 'minimumGroupingDigits' );
3330         }
3331
3332         /**
3333          * Take a list of strings and build a locale-friendly comma-separated
3334          * list, using the local comma-separator message.
3335          * The last two strings are chained with an "and".
3336          *
3337          * @param string[] $list
3338          * @return string
3339          */
3340         public function listToText( array $list ) {
3341                 $itemCount = count( $list );
3342                 if ( $itemCount < 1 ) {
3343                         return '';
3344                 }
3345                 $text = array_pop( $list );
3346                 if ( $itemCount > 1 ) {
3347                         $and = $this->msg( 'and' )->escaped();
3348                         $space = $this->msg( 'word-separator' )->escaped();
3349                         $comma = '';
3350                         if ( $itemCount > 2 ) {
3351                                 $comma = $this->msg( 'comma-separator' )->escaped();
3352                         }
3353                         $text = implode( $comma, $list ) . $and . $space . $text;
3354                 }
3355                 return $text;
3356         }
3357
3358         /**
3359          * Take a list of strings and build a locale-friendly comma-separated
3360          * list, using the local comma-separator message.
3361          * @param string[] $list Array of strings to put in a comma list
3362          * @return string
3363          */
3364         function commaList( array $list ) {
3365                 return implode(
3366                         wfMessage( 'comma-separator' )->inLanguage( $this )->escaped(),
3367                         $list
3368                 );
3369         }
3370
3371         /**
3372          * Take a list of strings and build a locale-friendly semicolon-separated
3373          * list, using the local semicolon-separator message.
3374          * @param string[] $list Array of strings to put in a semicolon list
3375          * @return string
3376          */
3377         function semicolonList( array $list ) {
3378                 return implode(
3379                         wfMessage( 'semicolon-separator' )->inLanguage( $this )->escaped(),
3380                         $list
3381                 );
3382         }
3383
3384         /**
3385          * Same as commaList, but separate it with the pipe instead.
3386          * @param string[] $list Array of strings to put in a pipe list
3387          * @return string
3388          */
3389         function pipeList( array $list ) {
3390                 return implode(
3391                         wfMessage( 'pipe-separator' )->inLanguage( $this )->escaped(),
3392                         $list
3393                 );
3394         }
3395
3396         /**
3397          * Truncate a string to a specified length in bytes, appending an optional
3398          * string (e.g. for ellipsis)
3399          *
3400          * If $length is negative, the string will be truncated from the beginning
3401          *
3402          * @since 1.31
3403          *
3404          * @param string $string String to truncate
3405          * @param int $length Maximum length in bytes
3406          * @param string $ellipsis String to append to the end of truncated text
3407          * @param bool $adjustLength Subtract length of ellipsis from $length
3408          *
3409          * @return string
3410          */
3411         function truncateForDatabase( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3412                 return $this->truncateInternal(
3413                         $string, $length, $ellipsis, $adjustLength, 'strlen', 'substr'
3414                 );
3415         }
3416
3417         /**
3418          * Truncate a string to a specified number of characters, appending an optional
3419          * string (e.g. for ellipsis).
3420          *
3421          * This provides multibyte version of truncateForDatabase() method of this class,
3422          * suitable for truncation based on number of characters, instead of number of bytes.
3423          *
3424          * If $length is negative, the string will be truncated from the beginning.
3425          *
3426          * @since 1.31
3427          *
3428          * @param string $string String to truncate
3429          * @param int $length Maximum number of characters
3430          * @param string $ellipsis String to append to the end of truncated text
3431          * @param bool $adjustLength Subtract length of ellipsis from $length
3432          *
3433          * @return string
3434          */
3435         function truncateForVisual( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3436                 // Passing encoding to mb_strlen and mb_substr is optional.
3437                 // Encoding defaults to mb_internal_encoding(), which is set to UTF-8 in Setup.php, so
3438                 // explicit specification of encoding is skipped.
3439                 // Note: Both multibyte methods are callables invoked in truncateInternal.
3440                 return $this->truncateInternal(
3441                         $string, $length, $ellipsis, $adjustLength, 'mb_strlen', 'mb_substr'
3442                 );
3443         }
3444
3445         /**
3446          * Internal method used for truncation. This method abstracts text truncation into
3447          * one common method, allowing users to provide length measurement function and
3448          * function for finding substring.
3449          *
3450          * For usages, see truncateForDatabase and truncateForVisual.
3451          *
3452          * @param string $string String to truncate
3453          * @param int $length Maximum length of final text
3454          * @param string $ellipsis String to append to the end of truncated text
3455          * @param bool $adjustLength Subtract length of ellipsis from $length
3456          * @param callable $measureLength Callable function used for determining the length of text
3457          * @param callable $getSubstring Callable function used for getting the substrings
3458          *
3459          * @return string
3460          */
3461         private function truncateInternal(
3462                 $string, $length, $ellipsis, $adjustLength, callable $measureLength, callable $getSubstring
3463         ) {
3464                 # Check if there is no need to truncate
3465                 if ( $measureLength( $string ) <= abs( $length ) ) {
3466                         return $string; // no need to truncate
3467                 }
3468
3469                 # Use the localized ellipsis character
3470                 if ( $ellipsis == '...' ) {
3471                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3472                 }
3473                 if ( $length == 0 ) {
3474                         return $ellipsis; // convention
3475                 }
3476
3477                 $stringOriginal = $string;
3478                 # If ellipsis length is >= $length then we can't apply $adjustLength
3479                 if ( $adjustLength && $measureLength( $ellipsis ) >= abs( $length ) ) {
3480                         $string = $ellipsis; // this can be slightly unexpected
3481                 # Otherwise, truncate and add ellipsis...
3482                 } else {
3483                         $ellipsisLength = $adjustLength ? $measureLength( $ellipsis ) : 0;
3484                         if ( $length > 0 ) {
3485                                 $length -= $ellipsisLength;
3486                                 $string = $getSubstring( $string, 0, $length ); // xyz...
3487                                 $string = $this->removeBadCharLast( $string );
3488                                 $string = rtrim( $string ) . $ellipsis;
3489                         } else {
3490                                 $length += $ellipsisLength;
3491                                 $string = $getSubstring( $string, $length ); // ...xyz
3492                                 $string = $this->removeBadCharFirst( $string );
3493                                 $string = $ellipsis . ltrim( $string );
3494                         }
3495                 }
3496
3497                 # Do not truncate if the ellipsis makes the string longer/equal (T24181).
3498                 # This check is *not* redundant if $adjustLength, due to the single case where
3499                 # LEN($ellipsis) > ABS($limit arg); $stringOriginal could be shorter than $string.
3500                 if ( $measureLength( $string ) < $measureLength( $stringOriginal ) ) {
3501                         return $string;
3502                 } else {
3503                         return $stringOriginal;
3504                 }
3505         }
3506
3507         /**
3508          * Remove bytes that represent an incomplete Unicode character
3509          * at the end of string (e.g. bytes of the char are missing)
3510          *
3511          * @param string $string
3512          * @return string
3513          */
3514         protected function removeBadCharLast( $string ) {
3515                 if ( $string != '' ) {
3516                         $char = ord( $string[strlen( $string ) - 1] );
3517                         $m = [];
3518                         if ( $char >= 0xc0 ) {
3519                                 # We got the first byte only of a multibyte char; remove it.
3520                                 $string = substr( $string, 0, -1 );
3521                         } elseif ( $char >= 0x80 &&
3522                                 // Use the /s modifier (PCRE_DOTALL) so (.*) also matches newlines
3523                                 preg_match( '/^(.*)(?:[\xe0-\xef][\x80-\xbf]|' .
3524                                         '[\xf0-\xf7][\x80-\xbf]{1,2})$/s', $string, $m )
3525                         ) {
3526                                 # We chopped in the middle of a character; remove it
3527                                 $string = $m[1];
3528                         }
3529                 }
3530                 return $string;
3531         }
3532
3533         /**
3534          * Remove bytes that represent an incomplete Unicode character
3535          * at the start of string (e.g. bytes of the char are missing)
3536          *
3537          * @param string $string
3538          * @return string
3539          */
3540         protected function removeBadCharFirst( $string ) {
3541                 if ( $string != '' ) {
3542                         $char = ord( $string[0] );
3543                         if ( $char >= 0x80 && $char < 0xc0 ) {
3544                                 # We chopped in the middle of a character; remove the whole thing
3545                                 $string = preg_replace( '/^[\x80-\xbf]+/', '', $string );
3546                         }
3547                 }
3548                 return $string;
3549         }
3550
3551         /**
3552          * Truncate a string of valid HTML to a specified length in bytes,
3553          * appending an optional string (e.g. for ellipses), and return valid HTML
3554          *
3555          * This is only intended for styled/linked text, such as HTML with
3556          * tags like <span> and <a>, were the tags are self-contained (valid HTML).
3557          * Also, this will not detect things like "display:none" CSS.
3558          *
3559          * Note: since 1.18 you do not need to leave extra room in $length for ellipses.
3560          *
3561          * @param string $text HTML string to truncate
3562          * @param int $length (zero/positive) Maximum length (including ellipses)
3563          * @param string $ellipsis String to append to the truncated text
3564          * @return string
3565          */
3566         function truncateHtml( $text, $length, $ellipsis = '...' ) {
3567                 # Use the localized ellipsis character
3568                 if ( $ellipsis == '...' ) {
3569                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3570                 }
3571                 # Check if there is clearly no need to truncate
3572                 if ( $length <= 0 ) {
3573                         return $ellipsis; // no text shown, nothing to format (convention)
3574                 } elseif ( strlen( $text ) <= $length ) {
3575                         return $text; // string short enough even *with* HTML (short-circuit)
3576                 }
3577
3578                 $dispLen = 0; // innerHTML legth so far
3579                 $testingEllipsis = false; // checking if ellipses will make string longer/equal?
3580                 $tagType = 0; // 0-open, 1-close
3581                 $bracketState = 0; // 1-tag start, 2-tag name, 0-neither
3582                 $entityState = 0; // 0-not entity, 1-entity
3583                 $tag = $ret = ''; // accumulated tag name, accumulated result string
3584                 $openTags = []; // open tag stack
3585                 $maybeState = null; // possible truncation state
3586
3587                 $textLen = strlen( $text );
3588                 $neLength = max( 0, $length - strlen( $ellipsis ) ); // non-ellipsis len if truncated
3589                 for ( $pos = 0; true; ++$pos ) {
3590                         # Consider truncation once the display length has reached the maximim.
3591                         # We check if $dispLen > 0 to grab tags for the $neLength = 0 case.
3592                         # Check that we're not in the middle of a bracket/entity...
3593                         if ( $dispLen && $dispLen >= $neLength && $bracketState == 0 && !$entityState ) {
3594                                 if ( !$testingEllipsis ) {
3595                                         $testingEllipsis = true;
3596                                         # Save where we are; we will truncate here unless there turn out to
3597                                         # be so few remaining characters that truncation is not necessary.
3598                                         if ( !$maybeState ) { // already saved? ($neLength = 0 case)
3599                                                 $maybeState = [ $ret, $openTags ]; // save state
3600                                         }
3601                                 } elseif ( $dispLen > $length && $dispLen > strlen( $ellipsis ) ) {
3602                                         # String in fact does need truncation, the truncation point was OK.
3603                                         // @phan-suppress-next-line PhanTypeInvalidExpressionArrayDestructuring
3604                                         list( $ret, $openTags ) = $maybeState; // reload state
3605                                         $ret = $this->removeBadCharLast( $ret ); // multi-byte char fix
3606                                         $ret .= $ellipsis; // add ellipsis
3607                                         break;
3608                                 }
3609                         }
3610                         if ( $pos >= $textLen ) {
3611                                 break; // extra iteration just for above checks
3612                         }
3613
3614                         # Read the next char...
3615                         $ch = $text[$pos];
3616                         $lastCh = $pos ? $text[$pos - 1] : '';
3617                         $ret .= $ch; // add to result string
3618                         if ( $ch == '<' ) {
3619                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags ); // for bad HTML
3620                                 $entityState = 0; // for bad HTML
3621                                 $bracketState = 1; // tag started (checking for backslash)
3622                         } elseif ( $ch == '>' ) {
3623                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags );
3624                                 $entityState = 0; // for bad HTML
3625                                 $bracketState = 0; // out of brackets
3626                         } elseif ( $bracketState == 1 ) {
3627                                 if ( $ch == '/' ) {
3628                                         $tagType = 1; // close tag (e.g. "</span>")
3629                                 } else {
3630                                         $tagType = 0; // open tag (e.g. "<span>")
3631                                         $tag .= $ch;
3632                                 }
3633                                 $bracketState = 2; // building tag name
3634                         } elseif ( $bracketState == 2 ) {
3635                                 if ( $ch != ' ' ) {
3636                                         $tag .= $ch;
3637                                 } else {
3638                                         // Name found (e.g. "<a href=..."), add on tag attributes...
3639                                         $pos += $this->truncate_skip( $ret, $text, "<>", $pos + 1 );
3640                                 }
3641                         } elseif ( $bracketState == 0 ) {
3642                                 if ( $entityState ) {
3643                                         if ( $ch == ';' ) {
3644                                                 $entityState = 0;
3645                                                 $dispLen++; // entity is one displayed char
3646                                         }
3647                                 } else {
3648                                         if ( $neLength == 0 && !$maybeState ) {
3649                                                 // Save state without $ch. We want to *hit* the first
3650                                                 // display char (to get tags) but not *use* it if truncating.
3651                                                 $maybeState = [ substr( $ret, 0, -1 ), $openTags ];
3652                                         }
3653                                         if ( $ch == '&' ) {
3654                                                 $entityState = 1; // entity found, (e.g. "&#160;")
3655                                         } else {
3656                                                 $dispLen++; // this char is displayed
3657                                                 // Add the next $max display text chars after this in one swoop...
3658                                                 $max = ( $testingEllipsis ? $length : $neLength ) - $dispLen;
3659                                                 $skipped = $this->truncate_skip( $ret, $text, "<>&", $pos + 1, $max );
3660                                                 $dispLen += $skipped;
3661                                                 $pos += $skipped;
3662                                         }
3663                                 }
3664                         }
3665                 }
3666                 // Close the last tag if left unclosed by bad HTML
3667                 $this->truncate_endBracket( $tag, $text[$textLen - 1], $tagType, $openTags );
3668                 while ( count( $openTags ) > 0 ) {
3669                         $ret .= '</' . array_pop( $openTags ) . '>'; // close open tags
3670                 }
3671                 return $ret;
3672         }
3673
3674         /**
3675          * truncateHtml() helper function
3676          * like strcspn() but adds the skipped chars to $ret
3677          *
3678          * @param string $ret
3679          * @param string $text
3680          * @param string $search
3681          * @param int $start
3682          * @param null|int $len
3683          * @return int
3684          */
3685         private function truncate_skip( &$ret, $text, $search, $start, $len = null ) {
3686                 if ( $len === null ) {
3687                         $len = -1; // -1 means "no limit" for strcspn
3688                 } elseif ( $len < 0 ) {
3689                         $len = 0; // sanity
3690                 }
3691                 $skipCount = 0;
3692                 if ( $start < strlen( $text ) ) {
3693                         $skipCount = strcspn( $text, $search, $start, $len );
3694                         $ret .= substr( $text, $start, $skipCount );
3695                 }
3696                 return $skipCount;
3697         }
3698
3699         /**
3700          * truncateHtml() helper function
3701          * (a) push or pop $tag from $openTags as needed
3702          * (b) clear $tag value
3703          * @param string &$tag Current HTML tag name we are looking at
3704          * @param int $tagType (0-open tag, 1-close tag)
3705          * @param string $lastCh Character before the '>' that ended this tag
3706          * @param array &$openTags Open tag stack (not accounting for $tag)
3707          */
3708         private function truncate_endBracket( &$tag, $tagType, $lastCh, &$openTags ) {
3709                 $tag = ltrim( $tag );
3710                 if ( $tag != '' ) {
3711                         if ( $tagType == 0 && $lastCh != '/' ) {
3712                                 $openTags[] = $tag; // tag opened (didn't close itself)
3713                         } elseif ( $tagType == 1 ) {
3714                                 if ( $openTags && $tag == $openTags[count( $openTags ) - 1] ) {
3715                                         array_pop( $openTags ); // tag closed
3716                                 }
3717                         }
3718                         $tag = '';
3719                 }
3720         }
3721
3722         /**
3723          * Grammatical transformations, needed for inflected languages
3724          * Invoked by putting {{grammar:case|word}} in a message
3725          *
3726          * @param string $word
3727          * @param string $case
3728          * @return string
3729          */
3730         function convertGrammar( $word, $case ) {
3731                 global $wgGrammarForms;
3732                 if ( isset( $wgGrammarForms[$this->getCode()][$case][$word] ) ) {
3733                         return $wgGrammarForms[$this->getCode()][$case][$word];
3734                 }
3735
3736                 $grammarTransformations = $this->getGrammarTransformations();
3737
3738                 if ( isset( $grammarTransformations[$case] ) ) {
3739                         $forms = $grammarTransformations[$case];
3740
3741                         // Some names of grammar rules are aliases for other rules.
3742                         // In such cases the value is a string rather than object,
3743                         // so load the actual rules.
3744                         if ( is_string( $forms ) ) {
3745                                 $forms = $grammarTransformations[$forms];
3746                         }
3747
3748                         foreach ( array_values( $forms ) as $rule ) {
3749                                 $form = $rule[0];
3750
3751                                 if ( $form === '@metadata' ) {
3752                                         continue;
3753                                 }
3754
3755                                 $replacement = $rule[1];
3756
3757                                 $regex = '/' . addcslashes( $form, '/' ) . '/u';
3758                                 $patternMatches = preg_match( $regex, $word );
3759
3760                                 if ( $patternMatches === false ) {
3761                                         wfLogWarning(
3762                                                 'An error occurred while processing grammar. ' .
3763                                                 "Word: '$word'. Regex: /$form/."
3764                                         );
3765                                 } elseif ( $patternMatches === 1 ) {
3766                                         $word = preg_replace( $regex, $replacement, $word );
3767
3768                                         break;
3769                                 }
3770                         }
3771                 }
3772
3773                 return $word;
3774         }
3775
3776         /**
3777          * Get the grammar forms for the content language
3778          * @return array Array of grammar forms
3779          * @since 1.20
3780          */
3781         function getGrammarForms() {
3782                 global $wgGrammarForms;
3783                 if ( isset( $wgGrammarForms[$this->getCode()] )
3784                         && is_array( $wgGrammarForms[$this->getCode()] )
3785                 ) {
3786                         return $wgGrammarForms[$this->getCode()];
3787                 }
3788
3789                 return [];
3790         }
3791
3792         /**
3793          * Get the grammar transformations data for the language.
3794          * Used like grammar forms, with {{GRAMMAR}} and cases,
3795          * but uses pairs of regexes and replacements instead of code.
3796          *
3797          * @return array[] Array of grammar transformations.
3798          * @throws MWException
3799          * @since 1.28
3800          */
3801         public function getGrammarTransformations() {
3802                 $languageCode = $this->getCode();
3803
3804                 if ( self::$grammarTransformations === null ) {
3805                         self::$grammarTransformations = new MapCacheLRU( 10 );
3806                 }
3807
3808                 if ( self::$grammarTransformations->has( $languageCode ) ) {
3809                         return self::$grammarTransformations->get( $languageCode );
3810                 }
3811
3812                 $data = [];
3813
3814                 $grammarDataFile = __DIR__ . "/data/grammarTransformations/$languageCode.json";
3815                 if ( is_readable( $grammarDataFile ) ) {
3816                         $data = FormatJson::decode(
3817                                 file_get_contents( $grammarDataFile ),
3818                                 true
3819                         );
3820
3821                         if ( $data === null ) {
3822                                 throw new MWException( "Invalid grammar data for \"$languageCode\"." );
3823                         }
3824
3825                         self::$grammarTransformations->set( $languageCode, $data );
3826                 }
3827
3828                 return $data;
3829         }
3830
3831         /**
3832          * Provides an alternative text depending on specified gender.
3833          * Usage {{gender:username|masculine|feminine|unknown}}.
3834          * username is optional, in which case the gender of current user is used,
3835          * but only in (some) interface messages; otherwise default gender is used.
3836          *
3837          * If no forms are given, an empty string is returned. If only one form is
3838          * given, it will be returned unconditionally. These details are implied by
3839          * the caller and cannot be overridden in subclasses.
3840          *
3841          * If three forms are given, the default is to use the third (unknown) form.
3842          * If fewer than three forms are given, the default is to use the first (masculine) form.
3843          * These details can be overridden in subclasses.
3844          *
3845          * @param string $gender
3846          * @param array $forms
3847          *
3848          * @return string
3849          */
3850         function gender( $gender, $forms ) {
3851                 if ( !count( $forms ) ) {
3852                         return '';
3853                 }
3854                 $forms = $this->preConvertPlural( $forms, 2 );
3855                 if ( $gender === 'male' ) {
3856                         return $forms[0];
3857                 }
3858                 if ( $gender === 'female' ) {
3859                         return $forms[1];
3860                 }
3861                 return $forms[2] ?? $forms[0];
3862         }
3863
3864         /**
3865          * Plural form transformations, needed for some languages.
3866          * For example, there are 3 form of plural in Russian and Polish,
3867          * depending on "count mod 10". See [[w:Plural]]
3868          * For English it is pretty simple.
3869          *
3870          * Invoked by putting {{plural:count|wordform1|wordform2}}
3871          * or {{plural:count|wordform1|wordform2|wordform3}}
3872          *
3873          * Example: {{plural:{{NUMBEROFARTICLES}}|article|articles}}
3874          *
3875          * @param int $count Non-localized number
3876          * @param array $forms Different plural forms
3877          * @return string Correct form of plural for $count in this language
3878          */
3879         function convertPlural( $count, $forms ) {
3880                 // Handle explicit n=pluralform cases
3881                 $forms = $this->handleExplicitPluralForms( $count, $forms );
3882                 if ( is_string( $forms ) ) {
3883                         return $forms;
3884                 }
3885                 if ( !count( $forms ) ) {
3886                         return '';
3887                 }
3888
3889                 $pluralForm = $this->getPluralRuleIndexNumber( $count );
3890                 $pluralForm = min( $pluralForm, count( $forms ) - 1 );
3891                 return $forms[$pluralForm];
3892         }
3893
3894         /**
3895          * Handles explicit plural forms for Language::convertPlural()
3896          *
3897          * In {{PLURAL:$1|0=nothing|one|many}}, 0=nothing will be returned if $1 equals zero.
3898          * If an explicitly defined plural form matches the $count, then
3899          * string value returned, otherwise array returned for further consideration
3900          * by CLDR rules or overridden convertPlural().
3901          *
3902          * @since 1.23
3903          *
3904          * @param int $count Non-localized number
3905          * @param array $forms Different plural forms
3906          *
3907          * @return array|string
3908          */
3909         protected function handleExplicitPluralForms( $count, array $forms ) {
3910                 foreach ( $forms as $index => $form ) {
3911                         if ( preg_match( '/\d+=/i', $form ) ) {
3912                                 $pos = strpos( $form, '=' );
3913                                 if ( substr( $form, 0, $pos ) === (string)$count ) {
3914                                         return substr( $form, $pos + 1 );
3915                                 }
3916                                 unset( $forms[$index] );
3917                         }
3918                 }
3919                 return array_values( $forms );
3920         }
3921
3922         /**
3923          * Checks that convertPlural was given an array and pads it to requested
3924          * amount of forms by copying the last one.
3925          *
3926          * @param array $forms
3927          * @param int $count How many forms should there be at least
3928          * @return array Padded array of forms
3929          */
3930         protected function preConvertPlural( /* Array */ $forms, $count ) {
3931                 return array_pad( $forms, $count, end( $forms ) );
3932         }
3933
3934         /**
3935          * Wraps argument with unicode control characters for directionality safety
3936          *
3937          * This solves the problem where directionality-neutral characters at the edge of
3938          * the argument string get interpreted with the wrong directionality from the
3939          * enclosing context, giving renderings that look corrupted like "(Ben_(WMF".
3940          *
3941          * The wrapping is LRE...PDF or RLE...PDF, depending on the detected
3942          * directionality of the argument string, using the BIDI algorithm's own "First
3943          * strong directional codepoint" rule. Essentially, this works round the fact that
3944          * there is no embedding equivalent of U+2068 FSI (isolation with heuristic
3945          * direction inference). The latter is cleaner but still not widely supported.
3946          *
3947          * @param string $text Text to wrap
3948          * @return string Text, wrapped in LRE...PDF or RLE...PDF or nothing
3949          */
3950         public function embedBidi( $text = '' ) {
3951                 $dir = self::strongDirFromContent( $text );
3952                 if ( $dir === 'ltr' ) {
3953                         // Wrap in LEFT-TO-RIGHT EMBEDDING ... POP DIRECTIONAL FORMATTING
3954                         return self::$lre . $text . self::$pdf;
3955                 }
3956                 if ( $dir === 'rtl' ) {
3957                         // Wrap in RIGHT-TO-LEFT EMBEDDING ... POP DIRECTIONAL FORMATTING
3958                         return self::$rle . $text . self::$pdf;
3959                 }
3960                 // No strong directionality: do not wrap
3961                 return $text;
3962         }
3963
3964         /**
3965          * @todo Maybe translate block durations.  Note that this function is somewhat misnamed: it
3966          * deals with translating the *duration* ("1 week", "4 days", etc), not the expiry time
3967          * (which is an absolute timestamp). Please note: do NOT add this blindly, as it is used
3968          * on old expiry lengths recorded in log entries. You'd need to provide the start date to
3969          * match up with it.
3970          *
3971          * @param string $str The validated block duration in English
3972          * @param User|null $user User object to use timezone from or null for $wgUser
3973          * @param int $now Current timestamp, for formatting relative block durations
3974          * @return string Somehow translated block duration
3975          * @see LanguageFi.php for example implementation
3976          */
3977         function translateBlockExpiry( $str, User $user = null, $now = 0 ) {
3978                 $duration = SpecialBlock::getSuggestedDurations( $this );
3979                 foreach ( $duration as $show => $value ) {
3980                         if ( strcmp( $str, $value ) == 0 ) {
3981                                 return htmlspecialchars( trim( $show ) );
3982                         }
3983                 }
3984
3985                 if ( wfIsInfinity( $str ) ) {
3986                         foreach ( $duration as $show => $value ) {
3987                                 if ( wfIsInfinity( $value ) ) {
3988                                         return htmlspecialchars( trim( $show ) );
3989                                 }
3990                         }
3991                 }
3992
3993                 // If all else fails, return a standard duration or timestamp description.
3994                 $time = strtotime( $str, $now );
3995                 if ( $time === false ) { // Unknown format. Return it as-is in case.
3996                         return $str;
3997                 } elseif ( $time !== strtotime( $str, $now + 1 ) ) { // It's a relative timestamp.
3998                         // The result differs based on current time, so the difference
3999                         // is a fixed duration length.
4000                         return $this->formatDuration( $time - $now );
4001                 } else { // It's an absolute timestamp.
4002                         if ( $time === 0 ) {
4003                                 // wfTimestamp() handles 0 as current time instead of epoch.
4004                                 $time = '19700101000000';
4005                         }
4006                         if ( $user ) {
4007                                 return $this->userTimeAndDate( $time, $user );
4008                         }
4009                         return $this->timeanddate( $time );
4010                 }
4011         }
4012
4013         /**
4014          * languages like Chinese need to be segmented in order for the diff
4015          * to be of any use
4016          *
4017          * @param string $text
4018          * @return string
4019          */
4020         public function segmentForDiff( $text ) {
4021                 return $text;
4022         }
4023
4024         /**
4025          * and unsegment to show the result
4026          *
4027          * @param string $text
4028          * @return string
4029          */
4030         public function unsegmentForDiff( $text ) {
4031                 return $text;
4032         }
4033
4034         /**
4035          * Return the LanguageConverter used in the Language
4036          *
4037          * @since 1.19
4038          * @return LanguageConverter
4039          */
4040         public function getConverter() {
4041                 return $this->mConverter;
4042         }
4043
4044         /**
4045          * convert text to a variant
4046          *
4047          * @param string $text text to convert
4048          * @param string|bool $variant variant to convert to, or false to use the user's preferred
4049          *      variant (if logged in), or the project default variant
4050          * @return string the converted string
4051          */
4052         public function autoConvert( $text, $variant = false ) {
4053                 return $this->mConverter->autoConvert( $text, $variant );
4054         }
4055
4056         /**
4057          * convert text to all supported variants
4058          *
4059          * @param string $text
4060          * @return array
4061          */
4062         public function autoConvertToAllVariants( $text ) {
4063                 return $this->mConverter->autoConvertToAllVariants( $text );
4064         }
4065
4066         /**
4067          * convert text to different variants of a language.
4068          *
4069          * @warning Glossary state is maintained between calls. This means
4070          *  if you pass unescaped text to this method it can cause an XSS
4071          *  in later calls to this method, even if the later calls have properly
4072          *  escaped the input. Never feed this method user controlled text that
4073          *  is not properly escaped!
4074          * @param string $text Content that has been already escaped for use in HTML
4075          * @return string HTML
4076          */
4077         public function convert( $text ) {
4078                 return $this->mConverter->convert( $text );
4079         }
4080
4081         /**
4082          * Convert a Title object to a string in the preferred variant
4083          *
4084          * @param Title $title
4085          * @return string
4086          */
4087         public function convertTitle( $title ) {
4088                 return $this->mConverter->convertTitle( $title );
4089         }
4090
4091         /**
4092          * Convert a namespace index to a string in the preferred variant
4093          *
4094          * @param int $ns namespace index (https://www.mediawiki.org/wiki/Manual:Namespace)
4095          * @param string|null $variant variant to convert to, or null to use the user's preferred
4096          *      variant (if logged in), or the project default variant
4097          * @return string a string representation of the namespace
4098          */
4099         public function convertNamespace( $ns, $variant = null ) {
4100                 return $this->mConverter->convertNamespace( $ns, $variant );
4101         }
4102
4103         /**
4104          * Check if this is a language with variants
4105          *
4106          * @return bool
4107          */
4108         public function hasVariants() {
4109                 return count( $this->getVariants() ) > 1;
4110         }
4111
4112         /**
4113          * Strict check if the language has the specific variant.
4114          *
4115          * Compare to LanguageConverter::validateVariant() which does a more
4116          * lenient check and attempts to coerce the given code to a valid one.
4117          *
4118          * @since 1.19
4119          * @param string $variant
4120          * @return bool
4121          */
4122         public function hasVariant( $variant ) {
4123                 return $variant && ( $variant === $this->mConverter->validateVariant( $variant ) );
4124         }
4125
4126         /**
4127          * Perform output conversion on a string, and encode for safe HTML output.
4128          * @param string $text Text to be converted
4129          * @return string
4130          * @todo this should get integrated somewhere sane
4131          */
4132         public function convertHtml( $text ) {
4133                 return htmlspecialchars( $this->convert( $text ) );
4134         }
4135
4136         /**
4137          * @param string $key
4138          * @return string
4139          */
4140         public function convertCategoryKey( $key ) {
4141                 return $this->mConverter->convertCategoryKey( $key );
4142         }
4143
4144         /**
4145          * Get the list of variants supported by this language
4146          * see sample implementation in LanguageZh.php
4147          *
4148          * @return string[] An array of language codes
4149          */
4150         public function getVariants() {
4151                 return $this->mConverter->getVariants();
4152         }
4153
4154         /**
4155          * @return string
4156          */
4157         public function getPreferredVariant() {
4158                 return $this->mConverter->getPreferredVariant();
4159         }
4160
4161         /**
4162          * @return string
4163          */
4164         public function getDefaultVariant() {
4165                 return $this->mConverter->getDefaultVariant();
4166         }
4167
4168         /**
4169          * @return string
4170          */
4171         public function getURLVariant() {
4172                 return $this->mConverter->getURLVariant();
4173         }
4174
4175         /**
4176          * If a language supports multiple variants, it is
4177          * possible that non-existing link in one variant
4178          * actually exists in another variant. this function
4179          * tries to find it. See e.g. LanguageZh.php
4180          * The input parameters may be modified upon return
4181          *
4182          * @param string &$link The name of the link
4183          * @param Title &$nt The title object of the link
4184          * @param bool $ignoreOtherCond To disable other conditions when
4185          *   we need to transclude a template or update a category's link
4186          */
4187         public function findVariantLink( &$link, &$nt, $ignoreOtherCond = false ) {
4188                 $this->mConverter->findVariantLink( $link, $nt, $ignoreOtherCond );
4189         }
4190
4191         /**
4192          * returns language specific options used by User::getPageRenderHash()
4193          * for example, the preferred language variant
4194          *
4195          * @return string
4196          */
4197         function getExtraHashOptions() {
4198                 return $this->mConverter->getExtraHashOptions();
4199         }
4200
4201         /**
4202          * For languages that support multiple variants, the title of an
4203          * article may be displayed differently in different variants. this
4204          * function returns the apporiate title defined in the body of the article.
4205          *
4206          * @return string
4207          */
4208         public function getParsedTitle() {
4209                 return $this->mConverter->getParsedTitle();
4210         }
4211
4212         /**
4213          * Refresh the cache of conversion tables when
4214          * MediaWiki:Conversiontable* is updated.
4215          *
4216          * @param Title $title The Title of the page being updated
4217          */
4218         public function updateConversionTable( Title $title ) {
4219                 $this->mConverter->updateConversionTable( $title );
4220         }
4221
4222         /**
4223          * A regular expression to match legal word-trailing characters
4224          * which should be merged onto a link of the form [[foo]]bar.
4225          *
4226          * @return string
4227          */
4228         public function linkTrail() {
4229                 return $this->localisationCache->getItem( $this->mCode, 'linkTrail' );
4230         }
4231
4232         /**
4233          * A regular expression character set to match legal word-prefixing
4234          * characters which should be merged onto a link of the form foo[[bar]].
4235          *
4236          * @return string
4237          */
4238         public function linkPrefixCharset() {
4239                 return $this->localisationCache->getItem( $this->mCode, 'linkPrefixCharset' );
4240         }
4241
4242         /**
4243          * Get the "parent" language which has a converter to convert a "compatible" language
4244          * (in another variant) to this language (eg. zh for zh-cn, but not en for en-gb).
4245          *
4246          * @return Language|null
4247          * @since 1.22
4248          */
4249         public function getParentLanguage() {
4250                 if ( $this->mParentLanguage !== false ) {
4251                         return $this->mParentLanguage;
4252                 }
4253
4254                 $code = explode( '-', $this->getCode() )[0];
4255                 if ( !in_array( $code, LanguageConverter::$languagesWithVariants ) ) {
4256                         $this->mParentLanguage = null;
4257                         return null;
4258                 }
4259                 $lang = self::factory( $code );
4260                 if ( !$lang->hasVariant( $this->getCode() ) ) {
4261                         $this->mParentLanguage = null;
4262                         return null;
4263                 }
4264
4265                 $this->mParentLanguage = $lang;
4266                 return $lang;
4267         }
4268
4269         /**
4270          * Compare with an other language object
4271          *
4272          * @since 1.28
4273          * @param Language $lang
4274          * @return bool
4275          */
4276         public function equals( Language $lang ) {
4277                 return $lang === $this || $lang->getCode() === $this->mCode;
4278         }
4279
4280         /**
4281          * Get the internal language code for this language object
4282          *
4283          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4284          * htmlspecialchars() or similar
4285          *
4286          * @return string
4287          */
4288         public function getCode() {
4289                 return $this->mCode;
4290         }
4291
4292         /**
4293          * Get the code in BCP 47 format which we can use
4294          * inside of html lang="" tags.
4295          *
4296          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4297          * htmlspecialchars() or similar.
4298          *
4299          * @since 1.19
4300          * @return string
4301          */
4302         public function getHtmlCode() {
4303                 if ( is_null( $this->mHtmlCode ) ) {
4304                         $this->mHtmlCode = LanguageCode::bcp47( $this->getCode() );
4305                 }
4306                 return $this->mHtmlCode;
4307         }
4308
4309         /**
4310          * Get the language code from a file name. Inverse of getFileName()
4311          * @param string $filename $prefix . $languageCode . $suffix
4312          * @param string $prefix Prefix before the language code
4313          * @param string $suffix Suffix after the language code
4314          * @return string Language code, or false if $prefix or $suffix isn't found
4315          */
4316         public static function getCodeFromFileName( $filename, $prefix = 'Language', $suffix = '.php' ) {
4317                 $m = null;
4318                 preg_match( '/' . preg_quote( $prefix, '/' ) . '([A-Z][a-z_]+)' .
4319                         preg_quote( $suffix, '/' ) . '/', $filename, $m );
4320                 if ( !count( $m ) ) {
4321                         return false;
4322                 }
4323                 return str_replace( '_', '-', strtolower( $m[1] ) );
4324         }
4325
4326         /**
4327          * @param string $code
4328          * @param bool $fallback Whether we're going through language fallback chain
4329          * @return string Name of the language class
4330          */
4331         public static function classFromCode( $code, $fallback = true ) {
4332                 if ( $fallback && $code == 'en' ) {
4333                         return 'Language';
4334                 } else {
4335                         return 'Language' . str_replace( '-', '_', ucfirst( $code ) );
4336                 }
4337         }
4338
4339         /**
4340          * Get the name of a file for a certain language code
4341          *
4342          * @deprecated since 1.34, use LanguageNameUtils
4343          * @param string $prefix Prepend this to the filename
4344          * @param string $code Language code
4345          * @param string $suffix Append this to the filename
4346          * @throws MWException
4347          * @return string $prefix . $mangledCode . $suffix
4348          */
4349         public static function getFileName( $prefix, $code, $suffix = '.php' ) {
4350                 return MediaWikiServices::getInstance()->getLanguageNameUtils()
4351                         ->getFileName( $prefix, $code, $suffix );
4352         }
4353
4354         /**
4355          * @deprecated since 1.34, use LanguageNameUtils
4356          * @param string $code
4357          * @return string
4358          */
4359         public static function getMessagesFileName( $code ) {
4360                 return MediaWikiServices::getInstance()->getLanguageNameUtils()
4361                         ->getMessagesFileName( $code );
4362         }
4363
4364         /**
4365          * @deprecated since 1.34, use LanguageNameUtils
4366          * @param string $code
4367          * @return string
4368          * @throws MWException
4369          * @since 1.23
4370          */
4371         public static function getJsonMessagesFileName( $code ) {
4372                 return MediaWikiServices::getInstance()->getLanguageNameUtils()
4373                         ->getJsonMessagesFileName( $code );
4374         }
4375
4376         /**
4377          * Get the first fallback for a given language.
4378          *
4379          * @param string $code
4380          *
4381          * @return bool|string
4382          */
4383         public static function getFallbackFor( $code ) {
4384                 $fallbacks = self::getFallbacksFor( $code );
4385                 if ( $fallbacks ) {
4386                         return $fallbacks[0];
4387                 }
4388                 return false;
4389         }
4390
4391         /**
4392          * Get the ordered list of fallback languages.
4393          *
4394          * @since 1.19
4395          * @param string $code Language code
4396          * @param int $mode Fallback mode, either MESSAGES_FALLBACKS (which always falls back to 'en'),
4397          * or STRICT_FALLBACKS (whic honly falls back to 'en' when explicitly defined)
4398          * @throws MWException
4399          * @return array List of language codes
4400          */
4401         public static function getFallbacksFor( $code, $mode = self::MESSAGES_FALLBACKS ) {
4402                 if ( $code === 'en' || !self::isValidBuiltInCode( $code ) ) {
4403                         return [];
4404                 }
4405                 switch ( $mode ) {
4406                         case self::MESSAGES_FALLBACKS:
4407                                 // For unknown languages, fallbackSequence returns an empty array,
4408                                 // hardcode fallback to 'en' in that case as English messages are
4409                                 // always defined.
4410                                 return self::getLocalisationCache()->getItem( $code, 'fallbackSequence' ) ?: [ 'en' ];
4411                         case self::STRICT_FALLBACKS:
4412                                 // Use this mode when you don't want to fallback to English unless
4413                                 // explicitly defined, for example when you have language-variant icons
4414                                 // and an international language-independent fallback.
4415                                 return self::getLocalisationCache()->getItem( $code, 'originalFallbackSequence' );
4416                         default:
4417                                 throw new MWException( "Invalid fallback mode \"$mode\"" );
4418                 }
4419         }
4420
4421         /**
4422          * Get the ordered list of fallback languages, ending with the fallback
4423          * language chain for the site language.
4424          *
4425          * @since 1.22
4426          * @param string $code Language code
4427          * @return array [ fallbacks, site fallbacks ]
4428          */
4429         public static function getFallbacksIncludingSiteLanguage( $code ) {
4430                 global $wgLanguageCode;
4431
4432                 // Usually, we will only store a tiny number of fallback chains, so we
4433                 // keep them in static memory.
4434                 $cacheKey = "{$code}-{$wgLanguageCode}";
4435
4436                 if ( !array_key_exists( $cacheKey, self::$fallbackLanguageCache ) ) {
4437                         $fallbacks = self::getFallbacksFor( $code );
4438
4439                         // Append the site's fallback chain, including the site language itself
4440                         $siteFallbacks = self::getFallbacksFor( $wgLanguageCode );
4441                         array_unshift( $siteFallbacks, $wgLanguageCode );
4442
4443                         // Eliminate any languages already included in the chain
4444                         $siteFallbacks = array_diff( $siteFallbacks, $fallbacks );
4445
4446                         self::$fallbackLanguageCache[$cacheKey] = [ $fallbacks, $siteFallbacks ];
4447                 }
4448                 return self::$fallbackLanguageCache[$cacheKey];
4449         }
4450
4451         /**
4452          * Get all messages for a given language
4453          * WARNING: this may take a long time. If you just need all message *keys*
4454          * but need the *contents* of only a few messages, consider using getMessageKeysFor().
4455          *
4456          * @param string $code
4457          *
4458          * @return array
4459          */
4460         public static function getMessagesFor( $code ) {
4461                 return self::getLocalisationCache()->getItem( $code, 'messages' );
4462         }
4463
4464         /**
4465          * Get a message for a given language
4466          *
4467          * @param string $key
4468          * @param string $code
4469          *
4470          * @return string
4471          */
4472         public static function getMessageFor( $key, $code ) {
4473                 return self::getLocalisationCache()->getSubitem( $code, 'messages', $key );
4474         }
4475
4476         /**
4477          * Get all message keys for a given language. This is a faster alternative to
4478          * array_keys( Language::getMessagesFor( $code ) )
4479          *
4480          * @since 1.19
4481          * @param string $code Language code
4482          * @return array Array of message keys (strings)
4483          */
4484         public static function getMessageKeysFor( $code ) {
4485                 return self::getLocalisationCache()->getSubitemList( $code, 'messages' );
4486         }
4487
4488         /**
4489          * @param string $talk
4490          * @return string
4491          */
4492         function fixVariableInNamespace( $talk ) {
4493                 if ( strpos( $talk, '$1' ) === false ) {
4494                         return $talk;
4495                 }
4496
4497                 global $wgMetaNamespace;
4498                 $talk = str_replace( '$1', $wgMetaNamespace, $talk );
4499
4500                 # Allow grammar transformations
4501                 # Allowing full message-style parsing would make simple requests
4502                 # such as action=raw much more expensive than they need to be.
4503                 # This will hopefully cover most cases.
4504                 $talk = preg_replace_callback( '/{{grammar:(.*?)\|(.*?)}}/i',
4505                         [ $this, 'replaceGrammarInNamespace' ], $talk );
4506                 return str_replace( ' ', '_', $talk );
4507         }
4508
4509         /**
4510          * @param string $m
4511          * @return string
4512          */
4513         function replaceGrammarInNamespace( $m ) {
4514                 return $this->convertGrammar( trim( $m[2] ), trim( $m[1] ) );
4515         }
4516
4517         /**
4518          * Decode an expiry (block, protection, etc) which has come from the DB
4519          *
4520          * @param string $expiry Database expiry String
4521          * @param bool|int $format True to process using language functions, or TS_ constant
4522          *     to return the expiry in a given timestamp
4523          * @param string $infinity If $format is not true, use this string for infinite expiry
4524          * @return string
4525          * @since 1.18
4526          */
4527         public function formatExpiry( $expiry, $format = true, $infinity = 'infinity' ) {
4528                 static $dbInfinity;
4529                 if ( $dbInfinity === null ) {
4530                         $dbInfinity = wfGetDB( DB_REPLICA )->getInfinity();
4531                 }
4532
4533                 if ( $expiry == '' || $expiry === 'infinity' || $expiry == $dbInfinity ) {
4534                         return $format === true
4535                                 ? $this->getMessageFromDB( 'infiniteblock' )
4536                                 : $infinity;
4537                 } else {
4538                         return $format === true
4539                                 ? $this->timeanddate( $expiry, /* User preference timezone */ true )
4540                                 : wfTimestamp( $format, $expiry );
4541                 }
4542         }
4543
4544         /**
4545          * Formats a time given in seconds into a string representation of that time.
4546          *
4547          * @param int|float $seconds
4548          * @param array $format An optional argument that formats the returned string in different ways:
4549          *   If $format['avoid'] === 'avoidhours': don't show hours, just show days
4550          *   If $format['avoid'] === 'avoidseconds': don't show seconds if $seconds >= 1 hour,
4551          *   If $format['avoid'] === 'avoidminutes': don't show seconds/minutes if $seconds > 48 hours,
4552          *   If $format['noabbrevs'] is true: use 'seconds' and friends instead of 'seconds-abbrev'
4553          *     and friends.
4554          * @note For backwards compatibility, $format may also be one of the strings 'avoidseconds'
4555          *     or 'avoidminutes'.
4556          * @return string
4557          */
4558         function formatTimePeriod( $seconds, $format = [] ) {
4559                 if ( !is_array( $format ) ) {
4560                         $format = [ 'avoid' => $format ]; // For backwards compatibility
4561                 }
4562                 if ( !isset( $format['avoid'] ) ) {
4563                         $format['avoid'] = false;
4564                 }
4565                 if ( !isset( $format['noabbrevs'] ) ) {
4566                         $format['noabbrevs'] = false;
4567                 }
4568                 $secondsMsg = wfMessage(
4569                         $format['noabbrevs'] ? 'seconds' : 'seconds-abbrev' )->inLanguage( $this );
4570                 $minutesMsg = wfMessage(
4571                         $format['noabbrevs'] ? 'minutes' : 'minutes-abbrev' )->inLanguage( $this );
4572                 $hoursMsg = wfMessage(
4573                         $format['noabbrevs'] ? 'hours' : 'hours-abbrev' )->inLanguage( $this );
4574                 $daysMsg = wfMessage(
4575                         $format['noabbrevs'] ? 'days' : 'days-abbrev' )->inLanguage( $this );
4576
4577                 if ( round( $seconds * 10 ) < 100 ) {
4578                         $s = $this->formatNum( sprintf( "%.1f", round( $seconds * 10 ) / 10 ) );
4579                         $s = $secondsMsg->params( $s )->text();
4580                 } elseif ( round( $seconds ) < 60 ) {
4581                         $s = $this->formatNum( round( $seconds ) );
4582                         $s = $secondsMsg->params( $s )->text();
4583                 } elseif ( round( $seconds ) < 3600 ) {
4584                         $minutes = floor( $seconds / 60 );
4585                         $secondsPart = round( fmod( $seconds, 60 ) );
4586                         if ( $secondsPart == 60 ) {
4587                                 $secondsPart = 0;
4588                                 $minutes++;
4589                         }
4590                         $s = $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4591                         $s .= ' ';
4592                         $s .= $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4593                 } elseif ( round( $seconds ) <= 2 * 86400 ) {
4594                         $hours = floor( $seconds / 3600 );
4595                         $minutes = floor( ( $seconds - $hours * 3600 ) / 60 );
4596                         $secondsPart = round( $seconds - $hours * 3600 - $minutes * 60 );
4597                         if ( $secondsPart == 60 ) {
4598                                 $secondsPart = 0;
4599                                 $minutes++;
4600                         }
4601                         if ( $minutes == 60 ) {
4602                                 $minutes = 0;
4603                                 $hours++;
4604                         }
4605                         $s = $hoursMsg->params( $this->formatNum( $hours ) )->text();
4606                         $s .= ' ';
4607                         $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4608                         if ( !in_array( $format['avoid'], [ 'avoidseconds', 'avoidminutes', 'avoidhours' ] ) ) {
4609                                 $s .= ' ' . $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4610                         }
4611                 } else {
4612                         $days = floor( $seconds / 86400 );
4613                         if ( $format['avoid'] === 'avoidhours' ) {
4614                                 $hours = round( ( $seconds - $days * 86400 ) / 3600 );
4615                                 if ( $hours == 24 ) {
4616                                         $hours = 0;
4617                                         $days++;
4618                                 }
4619                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4620                         } elseif ( $format['avoid'] === 'avoidminutes' ) {
4621                                 $hours = round( ( $seconds - $days * 86400 ) / 3600 );
4622                                 if ( $hours == 24 ) {
4623                                         $hours = 0;
4624                                         $days++;
4625                                 }
4626                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4627                                 $s .= ' ';
4628                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4629                         } elseif ( $format['avoid'] === 'avoidseconds' ) {
4630                                 $hours = floor( ( $seconds - $days * 86400 ) / 3600 );
4631                                 $minutes = round( ( $seconds - $days * 86400 - $hours * 3600 ) / 60 );
4632                                 if ( $minutes == 60 ) {
4633                                         $minutes = 0;
4634                                         $hours++;
4635                                 }
4636                                 if ( $hours == 24 ) {
4637                                         $hours = 0;
4638                                         $days++;
4639                                 }
4640                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4641                                 $s .= ' ';
4642                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4643                                 $s .= ' ';
4644                                 $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4645                         } else {
4646                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4647                                 $s .= ' ';
4648                                 $s .= $this->formatTimePeriod( $seconds - $days * 86400, $format );
4649                         }
4650                 }
4651                 return $s;
4652         }
4653
4654         /**
4655          * Format a bitrate for output, using an appropriate
4656          * unit (bps, kbps, Mbps, Gbps, Tbps, Pbps, Ebps, Zbps or Ybps) according to
4657          *   the magnitude in question.
4658          *
4659          * This use base 1000. For base 1024 use formatSize(), for another base
4660          * see formatComputingNumbers().
4661          *
4662          * @param int $bps
4663          * @return string
4664          */
4665         function formatBitrate( $bps ) {
4666                 return $this->formatComputingNumbers( $bps, 1000, "bitrate-$1bits" );
4667         }
4668
4669         /**
4670          * @param int $size Size of the unit
4671          * @param int $boundary Size boundary (1000, or 1024 in most cases)
4672          * @param string $messageKey Message key to be uesd
4673          * @return string
4674          */
4675         function formatComputingNumbers( $size, $boundary, $messageKey ) {
4676                 if ( $size <= 0 ) {
4677                         return str_replace( '$1', $this->formatNum( $size ),
4678                                 $this->getMessageFromDB( str_replace( '$1', '', $messageKey ) )
4679                         );
4680                 }
4681                 $sizes = [ '', 'kilo', 'mega', 'giga', 'tera', 'peta', 'exa', 'zeta', 'yotta' ];
4682                 $index = 0;
4683
4684                 $maxIndex = count( $sizes ) - 1;
4685                 while ( $size >= $boundary && $index < $maxIndex ) {
4686                         $index++;
4687                         $size /= $boundary;
4688                 }
4689
4690                 // For small sizes no decimal places necessary
4691                 $round = 0;
4692                 if ( $index > 1 ) {
4693                         // For MB and bigger two decimal places are smarter
4694                         $round = 2;
4695                 }
4696                 $msg = str_replace( '$1', $sizes[$index], $messageKey );
4697
4698                 $size = round( $size, $round );
4699                 $text = $this->getMessageFromDB( $msg );
4700                 return str_replace( '$1', $this->formatNum( $size ), $text );
4701         }
4702
4703         /**
4704          * Format a size in bytes for output, using an appropriate
4705          * unit (B, KB, MB, GB, TB, PB, EB, ZB or YB) according to the magnitude in question
4706          *
4707          * This method use base 1024. For base 1000 use formatBitrate(), for
4708          * another base see formatComputingNumbers()
4709          *
4710          * @param int $size Size to format
4711          * @return string Plain text (not HTML)
4712          */
4713         function formatSize( $size ) {
4714                 return $this->formatComputingNumbers( $size, 1024, "size-$1bytes" );
4715         }
4716
4717         /**
4718          * Make a list item, used by various special pages
4719          *
4720          * @param string $page Page link
4721          * @param string $details HTML safe text between brackets
4722          * @param bool $oppositedm Add the direction mark opposite to your
4723          *   language, to display text properly
4724          * @return string HTML escaped
4725          */
4726         function specialList( $page, $details, $oppositedm = true ) {
4727                 if ( !$details ) {
4728                         return $page;
4729                 }
4730
4731                 $dirmark = ( $oppositedm ? $this->getDirMark( true ) : '' ) . $this->getDirMark();
4732                 return $page .
4733                         $dirmark .
4734                         $this->msg( 'word-separator' )->escaped() .
4735                         $this->msg( 'parentheses' )->rawParams( $details )->escaped();
4736         }
4737
4738         /**
4739          * Generate (prev x| next x) (20|50|100...) type links for paging
4740          *
4741          * @param Title $title Title object to link
4742          * @param int $offset
4743          * @param int $limit
4744          * @param array $query Optional URL query parameter string
4745          * @param bool $atend Optional param for specified if this is the last page
4746          * @return string
4747          * @deprecated since 1.34, use PrevNextNavigationRenderer::buildPrevNextNavigation()
4748          *  instead.
4749          */
4750         public function viewPrevNext( Title $title, $offset, $limit,
4751                 array $query = [], $atend = false
4752         ) {
4753                 wfDeprecated( __METHOD__, '1.34' );
4754                 // @todo FIXME: Why on earth this needs one message for the text and another one for tooltip?
4755
4756                 # Make 'previous' link
4757                 $prev = wfMessage( 'prevn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4758                 if ( $offset > 0 ) {
4759                         $plink = $this->numLink( $title, max( $offset - $limit, 0 ), $limit,
4760                                 $query, $prev, 'prevn-title', 'mw-prevlink' );
4761                 } else {
4762                         $plink = htmlspecialchars( $prev );
4763                 }
4764
4765                 # Make 'next' link
4766                 $next = wfMessage( 'nextn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4767                 if ( $atend ) {
4768                         $nlink = htmlspecialchars( $next );
4769                 } else {
4770                         $nlink = $this->numLink( $title, $offset + $limit, $limit,
4771                                 $query, $next, 'nextn-title', 'mw-nextlink' );
4772                 }
4773
4774                 # Make links to set number of items per page
4775                 $numLinks = [];
4776                 foreach ( [ 20, 50, 100, 250, 500 ] as $num ) {
4777                         $numLinks[] = $this->numLink( $title, $offset, $num,
4778                                 $query, $this->formatNum( $num ), 'shown-title', 'mw-numlink' );
4779                 }
4780
4781                 return wfMessage( 'viewprevnext' )->inLanguage( $this )->title( $title
4782                         )->rawParams( $plink, $nlink, $this->pipeList( $numLinks ) )->escaped();
4783         }
4784
4785         /**
4786          * Helper function for viewPrevNext() that generates links
4787          *
4788          * @param Title $title Title object to link
4789          * @param int $offset
4790          * @param int $limit
4791          * @param array $query Extra query parameters
4792          * @param string $link Text to use for the link; will be escaped
4793          * @param string $tooltipMsg Name of the message to use as tooltip
4794          * @param string $class Value of the "class" attribute of the link
4795          * @return string HTML fragment
4796          */
4797         private function numLink( Title $title, $offset, $limit, array $query, $link,
4798                 $tooltipMsg, $class
4799         ) {
4800                 $query = [ 'limit' => $limit, 'offset' => $offset ] + $query;
4801                 $tooltip = wfMessage( $tooltipMsg )->inLanguage( $this )->title( $title )
4802                         ->numParams( $limit )->text();
4803
4804                 return Html::element( 'a', [ 'href' => $title->getLocalURL( $query ),
4805                         'title' => $tooltip, 'class' => $class ], $link );
4806         }
4807
4808         /**
4809          * Get the conversion rule title, if any.
4810          *
4811          * @return string
4812          */
4813         public function getConvRuleTitle() {
4814                 return $this->mConverter->getConvRuleTitle();
4815         }
4816
4817         /**
4818          * Get the compiled plural rules for the language
4819          * @since 1.20
4820          * @return array Associative array with plural form, and plural rule as key-value pairs
4821          */
4822         public function getCompiledPluralRules() {
4823                 $pluralRules =
4824                         $this->localisationCache->getItem( strtolower( $this->mCode ), 'compiledPluralRules' );
4825                 $fallbacks = self::getFallbacksFor( $this->mCode );
4826                 if ( !$pluralRules ) {
4827                         foreach ( $fallbacks as $fallbackCode ) {
4828                                 $pluralRules = $this->localisationCache
4829                                         ->getItem( strtolower( $fallbackCode ), 'compiledPluralRules' );
4830                                 if ( $pluralRules ) {
4831                                         break;
4832                                 }
4833                         }
4834                 }
4835                 return $pluralRules;
4836         }
4837
4838         /**
4839          * Get the plural rules for the language
4840          * @since 1.20
4841          * @return array Associative array with plural form number and plural rule as key-value pairs
4842          */
4843         public function getPluralRules() {
4844                 $pluralRules =
4845                         $this->localisationCache->getItem( strtolower( $this->mCode ), 'pluralRules' );
4846                 $fallbacks = self::getFallbacksFor( $this->mCode );
4847                 if ( !$pluralRules ) {
4848                         foreach ( $fallbacks as $fallbackCode ) {
4849                                 $pluralRules = $this->localisationCache
4850                                         ->getItem( strtolower( $fallbackCode ), 'pluralRules' );
4851                                 if ( $pluralRules ) {
4852                                         break;
4853                                 }
4854                         }
4855                 }
4856                 return $pluralRules;
4857         }
4858
4859         /**
4860          * Get the plural rule types for the language
4861          * @since 1.22
4862          * @return array Associative array with plural form number and plural rule type as key-value pairs
4863          */
4864         public function getPluralRuleTypes() {
4865                 $pluralRuleTypes =
4866                         $this->localisationCache->getItem( strtolower( $this->mCode ), 'pluralRuleTypes' );
4867                 $fallbacks = self::getFallbacksFor( $this->mCode );
4868                 if ( !$pluralRuleTypes ) {
4869                         foreach ( $fallbacks as $fallbackCode ) {
4870                                 $pluralRuleTypes = $this->localisationCache
4871                                         ->getItem( strtolower( $fallbackCode ), 'pluralRuleTypes' );
4872                                 if ( $pluralRuleTypes ) {
4873                                         break;
4874                                 }
4875                         }
4876                 }
4877                 return $pluralRuleTypes;
4878         }
4879
4880         /**
4881          * Find the index number of the plural rule appropriate for the given number
4882          * @param int $number
4883          * @return int The index number of the plural rule
4884          */
4885         public function getPluralRuleIndexNumber( $number ) {
4886                 $pluralRules = $this->getCompiledPluralRules();
4887                 $form = Evaluator::evaluateCompiled( $number, $pluralRules );
4888                 return $form;
4889         }
4890
4891         /**
4892          * Find the plural rule type appropriate for the given number
4893          * For example, if the language is set to Arabic, getPluralType(5) should
4894          * return 'few'.
4895          * @since 1.22
4896          * @param int $number
4897          * @return string The name of the plural rule type, e.g. one, two, few, many
4898          */
4899         public function getPluralRuleType( $number ) {
4900                 $index = $this->getPluralRuleIndexNumber( $number );
4901                 $pluralRuleTypes = $this->getPluralRuleTypes();
4902                 return $pluralRuleTypes[$index] ?? 'other';
4903         }
4904 }