languages/Language.php

   1 <?php
   2 /**
   3  * Internationalisation code.
   4  * See https://www.mediawiki.org/wiki/Special:MyLanguage/Localisation for more information.
   5  *
   6  * This program is free software; you can redistribute it and/or modify
   7  * it under the terms of the GNU General Public License as published by
   8  * the Free Software Foundation; either version 2 of the License, or
   9  * (at your option) any later version.
  10  *
  11  * This program is distributed in the hope that it will be useful,
  12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  14  * GNU General Public License for more details.
  15  *
  16  * You should have received a copy of the GNU General Public License along
  17  * with this program; if not, write to the Free Software Foundation, Inc.,
  18  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  19  * http://www.gnu.org/copyleft/gpl.html
  20  *
  21  * @file
  22  * @ingroup Language
  23  */
  24
  25 /**
  26  * @defgroup Language Language
  27  */
  28
  29 use CLDRPluralRuleParser\Evaluator;
  30 use MediaWiki\MediaWikiServices;
  31 use Wikimedia\Assert\Assert;
  32
  33 /**
  34  * Internationalisation code
  35  * @ingroup Language
  36  */
  37 class Language {
  38         /**
  39          * Return autonyms in fetchLanguageName(s).
  40          * @since 1.32
  41          */
  42         const AS_AUTONYMS = null;
  43
  44         /**
  45          * Return all known languages in fetchLanguageName(s).
  46          * @since 1.32
  47          */
  48         const ALL = 'all';
  49
  50         /**
  51          * Return in fetchLanguageName(s) only the languages for which we have at
  52          * least some localisation.
  53          * @since 1.32
  54          */
  55         const SUPPORTED = 'mwfile';
  56
  57         /**
  58          * @var LanguageConverter
  59          */
  60         public $mConverter;
  61
  62         public $mVariants, $mCode, $mLoaded = false;
  63         public $mMagicExtensions = [];
  64         private $mHtmlCode = null;
  65         /** @var Language|false */
  66         private $mParentLanguage = false;
  67
  68         public $dateFormatStrings = [];
  69         public $mExtendedSpecialPageAliases;
  70
  71         /** @var array|null */
  72         protected $namespaceNames;
  73         protected $mNamespaceIds, $namespaceAliases;
  74
  75         /**
  76          * ReplacementArray object caches
  77          */
  78         public $transformData = [];
  79
  80         /**
  81          * @var LocalisationCache
  82          */
  83         public static $dataCache;
  84
  85         public static $mLangObjCache = [];
  86
  87         /**
  88          * Return a fallback chain for messages in getFallbacksFor
  89          * @since 1.32
  90          */
  91         const MESSAGES_FALLBACKS = 0;
  92
  93         /**
  94          * Return a strict fallback chain in getFallbacksFor
  95          * @since 1.32
  96          */
  97         const STRICT_FALLBACKS = 1;
  98
  99         public static $mWeekdayMsgs = [
 100                 'sunday', 'monday', 'tuesday', 'wednesday', 'thursday',
 101                 'friday', 'saturday'
 102         ];
 103
 104         public static $mWeekdayAbbrevMsgs = [
 105                 'sun', 'mon', 'tue', 'wed', 'thu', 'fri', 'sat'
 106         ];
 107
 108         public static $mMonthMsgs = [
 109                 'january', 'february', 'march', 'april', 'may_long', 'june',
 110                 'july', 'august', 'september', 'october', 'november',
 111                 'december'
 112         ];
 113         public static $mMonthGenMsgs = [
 114                 'january-gen', 'february-gen', 'march-gen', 'april-gen', 'may-gen', 'june-gen',
 115                 'july-gen', 'august-gen', 'september-gen', 'october-gen', 'november-gen',
 116                 'december-gen'
 117         ];
 118         public static $mMonthAbbrevMsgs = [
 119                 'jan', 'feb', 'mar', 'apr', 'may', 'jun', 'jul', 'aug',
 120                 'sep', 'oct', 'nov', 'dec'
 121         ];
 122
 123         public static $mIranianCalendarMonthMsgs = [
 124                 'iranian-calendar-m1', 'iranian-calendar-m2', 'iranian-calendar-m3',
 125                 'iranian-calendar-m4', 'iranian-calendar-m5', 'iranian-calendar-m6',
 126                 'iranian-calendar-m7', 'iranian-calendar-m8', 'iranian-calendar-m9',
 127                 'iranian-calendar-m10', 'iranian-calendar-m11', 'iranian-calendar-m12'
 128         ];
 129
 130         public static $mHebrewCalendarMonthMsgs = [
 131                 'hebrew-calendar-m1', 'hebrew-calendar-m2', 'hebrew-calendar-m3',
 132                 'hebrew-calendar-m4', 'hebrew-calendar-m5', 'hebrew-calendar-m6',
 133                 'hebrew-calendar-m7', 'hebrew-calendar-m8', 'hebrew-calendar-m9',
 134                 'hebrew-calendar-m10', 'hebrew-calendar-m11', 'hebrew-calendar-m12',
 135                 'hebrew-calendar-m6a', 'hebrew-calendar-m6b'
 136         ];
 137
 138         public static $mHebrewCalendarMonthGenMsgs = [
 139                 'hebrew-calendar-m1-gen', 'hebrew-calendar-m2-gen', 'hebrew-calendar-m3-gen',
 140                 'hebrew-calendar-m4-gen', 'hebrew-calendar-m5-gen', 'hebrew-calendar-m6-gen',
 141                 'hebrew-calendar-m7-gen', 'hebrew-calendar-m8-gen', 'hebrew-calendar-m9-gen',
 142                 'hebrew-calendar-m10-gen', 'hebrew-calendar-m11-gen', 'hebrew-calendar-m12-gen',
 143                 'hebrew-calendar-m6a-gen', 'hebrew-calendar-m6b-gen'
 144         ];
 145
 146         public static $mHijriCalendarMonthMsgs = [
 147                 'hijri-calendar-m1', 'hijri-calendar-m2', 'hijri-calendar-m3',
 148                 'hijri-calendar-m4', 'hijri-calendar-m5', 'hijri-calendar-m6',
 149                 'hijri-calendar-m7', 'hijri-calendar-m8', 'hijri-calendar-m9',
 150                 'hijri-calendar-m10', 'hijri-calendar-m11', 'hijri-calendar-m12'
 151         ];
 152
 153         /**
 154          * @since 1.20
 155          * @var array
 156          */
 157         public static $durationIntervals = [
 158                 'millennia' => 31556952000,
 159                 'centuries' => 3155695200,
 160                 'decades' => 315569520,
 161                 'years' => 31556952, // 86400 * ( 365 + ( 24 * 3 + 25 ) / 400 )
 162                 'weeks' => 604800,
 163                 'days' => 86400,
 164                 'hours' => 3600,
 165                 'minutes' => 60,
 166                 'seconds' => 1,
 167         ];
 168
 169         /**
 170          * Cache for language fallbacks.
 171          * @see Language::getFallbacksIncludingSiteLanguage
 172          * @since 1.21
 173          * @var array
 174          */
 175         private static $fallbackLanguageCache = [];
 176
 177         /**
 178          * Cache for grammar rules data
 179          * @var MapCacheLRU|null
 180          */
 181         private static $grammarTransformations;
 182
 183         /**
 184          * Cache for language names
 185          * @var HashBagOStuff|null
 186          */
 187         private static $languageNameCache;
 188
 189         /**
 190          * Unicode directional formatting characters, for embedBidi()
 191          */
 192         private static $lre = "\u{202A}"; // U+202A LEFT-TO-RIGHT EMBEDDING
 193         private static $rle = "\u{202B}"; // U+202B RIGHT-TO-LEFT EMBEDDING
 194         private static $pdf = "\u{202C}"; // U+202C POP DIRECTIONAL FORMATTING
 195
 196         /**
 197          * Directionality test regex for embedBidi(). Matches the first strong directionality codepoint:
 198          * - in group 1 if it is LTR
 199          * - in group 2 if it is RTL
 200          * Does not match if there is no strong directionality codepoint.
 201          *
 202          * The form is '/(?:([strong ltr codepoint])|([strong rtl codepoint]))/u' .
 203          *
 204          * Generated by UnicodeJS (see tools/strongDir) from the UCD; see
 205          * https://phabricator.wikimedia.org/diffusion/GUJS/ .
 206          */
 207         // @codeCoverageIgnoreStart
 208         // phpcs:ignore Generic.Files.LineLength
 209         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';
 210         // @codeCoverageIgnoreEnd
 211
 212         /**
 213          * Get a cached or new language object for a given language code
 214          * @param string $code
 215          * @throws MWException
 216          * @return Language
 217          */
 218         static function factory( $code ) {
 219                 global $wgDummyLanguageCodes, $wgLangObjCacheSize;
 220
 221                 if ( isset( $wgDummyLanguageCodes[$code] ) ) {
 222                         $code = $wgDummyLanguageCodes[$code];
 223                 }
 224
 225                 // get the language object to process
 226                 $langObj = self::$mLangObjCache[$code] ?? self::newFromCode( $code );
 227
 228                 // merge the language object in to get it up front in the cache
 229                 self::$mLangObjCache = array_merge( [ $code => $langObj ], self::$mLangObjCache );
 230                 // get rid of the oldest ones in case we have an overflow
 231                 self::$mLangObjCache = array_slice( self::$mLangObjCache, 0, $wgLangObjCacheSize, true );
 232
 233                 return $langObj;
 234         }
 235
 236         /**
 237          * Create a language object for a given language code
 238          * @param string $code
 239          * @param bool $fallback Whether we're going through language fallback chain
 240          * @throws MWException
 241          * @return Language
 242          */
 243         protected static function newFromCode( $code, $fallback = false ) {
 244                 if ( !self::isValidCode( $code ) ) {
 245                         throw new MWException( "Invalid language code \"$code\"" );
 246                 }
 247
 248                 if ( !self::isValidBuiltInCode( $code ) ) {
 249                         // It's not possible to customise this code with class files, so
 250                         // just return a Language object. This is to support uselang= hacks.
 251                         $lang = new Language;
 252                         $lang->mCode = $code;
 253                         return $lang;
 254                 }
 255
 256                 // Check if there is a language class for the code
 257                 $class = self::classFromCode( $code, $fallback );
 258                 // LanguageCode does not inherit Language
 259                 if ( class_exists( $class ) && is_a( $class, 'Language', true ) ) {
 260                         $lang = new $class;
 261                         return $lang;
 262                 }
 263
 264                 // Keep trying the fallback list until we find an existing class
 265                 $fallbacks = self::getFallbacksFor( $code );
 266                 foreach ( $fallbacks as $fallbackCode ) {
 267                         if ( !self::isValidBuiltInCode( $fallbackCode ) ) {
 268                                 throw new MWException( "Invalid fallback '$fallbackCode' in fallback sequence for '$code'" );
 269                         }
 270
 271                         $class = self::classFromCode( $fallbackCode );
 272                         if ( class_exists( $class ) ) {
 273                                 $lang = new $class;
 274                                 $lang->mCode = $code;
 275                                 return $lang;
 276                         }
 277                 }
 278
 279                 throw new MWException( "Invalid fallback sequence for language '$code'" );
 280         }
 281
 282         /**
 283          * Intended for tests that may change configuration in a way that invalidates caches.
 284          *
 285          * @since 1.32
 286          */
 287         public static function clearCaches() {
 288                 if ( !defined( 'MW_PHPUNIT_TEST' ) ) {
 289                         throw new MWException( __METHOD__ . ' must not be used outside tests' );
 290                 }
 291                 self::$dataCache = null;
 292                 // Reinitialize $dataCache, since it's expected to always be available
 293                 self::getLocalisationCache();
 294                 self::$mLangObjCache = [];
 295                 self::$fallbackLanguageCache = [];
 296                 self::$grammarTransformations = null;
 297                 self::$languageNameCache = null;
 298         }
 299
 300         /**
 301          * Checks whether any localisation is available for that language tag
 302          * in MediaWiki (MessagesXx.php exists).
 303          *
 304          * @param string $code Language tag (in lower case)
 305          * @return bool Whether language is supported
 306          * @since 1.21
 307          */
 308         public static function isSupportedLanguage( $code ) {
 309                 if ( !self::isValidBuiltInCode( $code ) ) {
 310                         return false;
 311                 }
 312
 313                 if ( $code === 'qqq' ) {
 314                         return false;
 315                 }
 316
 317                 return is_readable( self::getMessagesFileName( $code ) ) ||
 318                         is_readable( self::getJsonMessagesFileName( $code ) );
 319         }
 320
 321         /**
 322          * Returns true if a language code string is a well-formed language tag
 323          * according to RFC 5646.
 324          * This function only checks well-formedness; it doesn't check that
 325          * language, script or variant codes actually exist in the repositories.
 326          *
 327          * Based on regexes by Mark Davis of the Unicode Consortium:
 328          * https://www.unicode.org/repos/cldr/trunk/tools/java/org/unicode/cldr/util/data/langtagRegex.txt
 329          *
 330          * @param string $code
 331          * @param bool $lenient Whether to allow '_' as separator. The default is only '-'.
 332          *
 333          * @return bool
 334          * @since 1.21
 335          */
 336         public static function isWellFormedLanguageTag( $code, $lenient = false ) {
 337                 $alpha = '[a-z]';
 338                 $digit = '[0-9]';
 339                 $alphanum = '[a-z0-9]';
 340                 $x = 'x'; # private use singleton
 341                 $singleton = '[a-wy-z]'; # other singleton
 342                 $s = $lenient ? '[-_]' : '-';
 343
 344                 $language = "$alpha{2,8}|$alpha{2,3}$s$alpha{3}";
 345                 $script = "$alpha{4}"; # ISO 15924
 346                 $region = "(?:$alpha{2}|$digit{3})"; # ISO 3166-1 alpha-2 or UN M.49
 347                 $variant = "(?:$alphanum{5,8}|$digit$alphanum{3})";
 348                 $extension = "$singleton(?:$s$alphanum{2,8})+";
 349                 $privateUse = "$x(?:$s$alphanum{1,8})+";
 350
 351                 # Define certain grandfathered codes, since otherwise the regex is pretty useless.
 352                 # Since these are limited, this is safe even later changes to the registry --
 353                 # the only oddity is that it might change the type of the tag, and thus
 354                 # the results from the capturing groups.
 355                 # https://www.iana.org/assignments/language-subtag-registry
 356
 357                 $grandfathered = "en{$s}GB{$s}oed"
 358                         . "|i{$s}(?:ami|bnn|default|enochian|hak|klingon|lux|mingo|navajo|pwn|tao|tay|tsu)"
 359                         . "|no{$s}(?:bok|nyn)"
 360                         . "|sgn{$s}(?:BE{$s}(?:fr|nl)|CH{$s}de)"
 361                         . "|zh{$s}min{$s}nan";
 362
 363                 $variantList = "$variant(?:$s$variant)*";
 364                 $extensionList = "$extension(?:$s$extension)*";
 365
 366                 $langtag = "(?:($language)"
 367                         . "(?:$s$script)?"
 368                         . "(?:$s$region)?"
 369                         . "(?:$s$variantList)?"
 370                         . "(?:$s$extensionList)?"
 371                         . "(?:$s$privateUse)?)";
 372
 373                 # The final breakdown, with capturing groups for each of these components
 374                 # The variants, extensions, grandfathered, and private-use may have interior '-'
 375
 376                 $root = "^(?:$langtag|$privateUse|$grandfathered)$";
 377
 378                 return (bool)preg_match( "/$root/", strtolower( $code ) );
 379         }
 380
 381         /**
 382          * Returns true if a language code string is of a valid form, whether or
 383          * not it exists. This includes codes which are used solely for
 384          * customisation via the MediaWiki namespace.
 385          *
 386          * @param string $code
 387          *
 388          * @return bool
 389          */
 390         public static function isValidCode( $code ) {
 391                 static $cache = [];
 392                 Assert::parameterType( 'string', $code, '$code' );
 393                 if ( !isset( $cache[$code] ) ) {
 394                         // People think language codes are html safe, so enforce it.
 395                         // Ideally we should only allow a-zA-Z0-9-
 396                         // but, .+ and other chars are often used for {{int:}} hacks
 397                         // see bugs T39564, T39587, T38938
 398                         $cache[$code] =
 399                                 // Protect against path traversal
 400                                 strcspn( $code, ":/\\\000&<>'\"" ) === strlen( $code )
 401                                 && !preg_match( MediaWikiTitleCodec::getTitleInvalidRegex(), $code );
 402                 }
 403                 return $cache[$code];
 404         }
 405
 406         /**
 407          * Returns true if a language code is of a valid form for the purposes of
 408          * internal customisation of MediaWiki, via Messages*.php or *.json.
 409          *
 410          * @param string $code
 411          *
 412          * @since 1.18
 413          * @return bool
 414          */
 415         public static function isValidBuiltInCode( $code ) {
 416                 Assert::parameterType( 'string', $code, '$code' );
 417
 418                 return (bool)preg_match( '/^[a-z0-9-]{2,}$/', $code );
 419         }
 420
 421         /**
 422          * Returns true if a language code is an IETF tag known to MediaWiki.
 423          *
 424          * @param string $tag
 425          *
 426          * @since 1.21
 427          * @return bool
 428          */
 429         public static function isKnownLanguageTag( $tag ) {
 430                 // Quick escape for invalid input to avoid exceptions down the line
 431                 // when code tries to process tags which are not valid at all.
 432                 if ( !self::isValidBuiltInCode( $tag ) ) {
 433                         return false;
 434                 }
 435
 436                 if ( isset( MediaWiki\Languages\Data\Names::$names[$tag] )
 437                         || self::fetchLanguageName( $tag, $tag ) !== ''
 438                 ) {
 439                         return true;
 440                 }
 441
 442                 return false;
 443         }
 444
 445         /**
 446          * Get the LocalisationCache instance
 447          *
 448          * @return LocalisationCache
 449          */
 450         public static function getLocalisationCache() {
 451                 if ( is_null( self::$dataCache ) ) {
 452                         global $wgLocalisationCacheConf;
 453                         $class = $wgLocalisationCacheConf['class'];
 454                         self::$dataCache = new $class( $wgLocalisationCacheConf );
 455                 }
 456                 return self::$dataCache;
 457         }
 458
 459         function __construct() {
 460                 // @phan-suppress-next-line PhanTypeMismatchProperty
 461                 $this->mConverter = new FakeConverter( $this );
 462                 // Set the code to the name of the descendant
 463                 if ( static::class === 'Language' ) {
 464                         $this->mCode = 'en';
 465                 } else {
 466                         $this->mCode = str_replace( '_', '-', strtolower( substr( static::class, 8 ) ) );
 467                 }
 468                 self::getLocalisationCache();
 469         }
 470
 471         /**
 472          * Reduce memory usage
 473          * @suppress PhanTypeSuspiciousNonTraversableForeach
 474          */
 475         function __destruct() {
 476                 foreach ( $this as $name => $value ) {
 477                         unset( $this->$name );
 478                 }
 479         }
 480
 481         /**
 482          * Hook which will be called if this is the content language.
 483          * Descendants can use this to register hook functions or modify globals
 484          */
 485         function initContLang() {
 486         }
 487
 488         /**
 489          * @return array
 490          * @since 1.19
 491          */
 492         public function getFallbackLanguages() {
 493                 return self::getFallbacksFor( $this->mCode );
 494         }
 495
 496         /**
 497          * Exports $wgBookstoreListEn
 498          * @return array
 499          */
 500         public function getBookstoreList() {
 501                 return self::$dataCache->getItem( $this->mCode, 'bookstoreList' );
 502         }
 503
 504         /**
 505          * Returns an array of localised namespaces indexed by their numbers. If the namespace is not
 506          * available in localised form, it will be included in English.
 507          *
 508          * @return string[] List of localized namespace names, indexed by numeric namespace ID.
 509          */
 510         public function getNamespaces() {
 511                 if ( is_null( $this->namespaceNames ) ) {
 512                         global $wgMetaNamespace, $wgMetaNamespaceTalk, $wgExtraNamespaces;
 513
 514                         $validNamespaces = MediaWikiServices::getInstance()->getNamespaceInfo()->
 515                                 getCanonicalNamespaces();
 516
 517                         $this->namespaceNames = $wgExtraNamespaces +
 518                                 self::$dataCache->getItem( $this->mCode, 'namespaceNames' );
 519                         $this->namespaceNames += $validNamespaces;
 520
 521                         $this->namespaceNames[NS_PROJECT] = $wgMetaNamespace;
 522                         if ( $wgMetaNamespaceTalk ) {
 523                                 $this->namespaceNames[NS_PROJECT_TALK] = $wgMetaNamespaceTalk;
 524                         } else {
 525                                 $talk = $this->namespaceNames[NS_PROJECT_TALK];
 526                                 $this->namespaceNames[NS_PROJECT_TALK] =
 527                                         $this->fixVariableInNamespace( $talk );
 528                         }
 529
 530                         # Sometimes a language will be localised but not actually exist on this wiki.
 531                         // @phan-suppress-next-line PhanTypeMismatchForeach
 532                         foreach ( $this->namespaceNames as $key => $text ) {
 533                                 if ( !isset( $validNamespaces[$key] ) ) {
 534                                         unset( $this->namespaceNames[$key] );
 535                                 }
 536                         }
 537
 538                         # The above mixing may leave namespaces out of canonical order.
 539                         # Re-order by namespace ID number...
 540                         // @phan-suppress-next-line PhanTypeMismatchArgumentInternal
 541                         ksort( $this->namespaceNames );
 542
 543                         Hooks::run( 'LanguageGetNamespaces', [ &$this->namespaceNames ] );
 544                 }
 545
 546                 return $this->namespaceNames;
 547         }
 548
 549         /**
 550          * Arbitrarily set all of the namespace names at once. Mainly used for testing
 551          * @param array $namespaces Array of namespaces (id => name)
 552          */
 553         public function setNamespaces( array $namespaces ) {
 554                 $this->namespaceNames = $namespaces;
 555                 $this->mNamespaceIds = null;
 556         }
 557
 558         /**
 559          * Resets all of the namespace caches. Mainly used for testing
 560          */
 561         public function resetNamespaces() {
 562                 $this->namespaceNames = null;
 563                 $this->mNamespaceIds = null;
 564                 $this->namespaceAliases = null;
 565         }
 566
 567         /**
 568          * A convenience function that returns getNamespaces() with spaces instead of underscores
 569          * in values. Useful for producing output to be displayed e.g. in `<select>` forms.
 570          *
 571          * @return array
 572          */
 573         public function getFormattedNamespaces() {
 574                 $ns = $this->getNamespaces();
 575                 foreach ( $ns as $k => $v ) {
 576                         $ns[$k] = strtr( $v, '_', ' ' );
 577                 }
 578                 return $ns;
 579         }
 580
 581         /**
 582          * Get a namespace value by key
 583          *
 584          * <code>
 585          * $mw_ns = $lang->getNsText( NS_MEDIAWIKI );
 586          * echo $mw_ns; // prints 'MediaWiki'
 587          * </code>
 588          *
 589          * @param int $index The array key of the namespace to return
 590          * @return string|bool String if the namespace value exists, otherwise false
 591          */
 592         public function getNsText( $index ) {
 593                 $ns = $this->getNamespaces();
 594                 return $ns[$index] ?? false;
 595         }
 596
 597         /**
 598          * A convenience function that returns the same thing as
 599          * getNsText() except with '_' changed to ' ', useful for
 600          * producing output.
 601          *
 602          * <code>
 603          * $mw_ns = $lang->getFormattedNsText( NS_MEDIAWIKI_TALK );
 604          * echo $mw_ns; // prints 'MediaWiki talk'
 605          * </code>
 606          *
 607          * @param int $index The array key of the namespace to return
 608          * @return string Namespace name without underscores (empty string if namespace does not exist)
 609          */
 610         public function getFormattedNsText( $index ) {
 611                 $ns = $this->getNsText( $index );
 612                 return strtr( $ns, '_', ' ' );
 613         }
 614
 615         /**
 616          * Returns gender-dependent namespace alias if available.
 617          * See https://www.mediawiki.org/wiki/Manual:$wgExtraGenderNamespaces
 618          * @param int $index Namespace index
 619          * @param string $gender Gender key (male, female... )
 620          * @return string
 621          * @since 1.18
 622          */
 623         public function getGenderNsText( $index, $gender ) {
 624                 global $wgExtraGenderNamespaces;
 625
 626                 $ns = $wgExtraGenderNamespaces +
 627                         (array)self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 628
 629                 return $ns[$index][$gender] ?? $this->getNsText( $index );
 630         }
 631
 632         /**
 633          * Whether this language uses gender-dependent namespace aliases.
 634          * See https://www.mediawiki.org/wiki/Manual:$wgExtraGenderNamespaces
 635          * @return bool
 636          * @since 1.18
 637          */
 638         public function needsGenderDistinction() {
 639                 global $wgExtraGenderNamespaces, $wgExtraNamespaces;
 640                 if ( count( $wgExtraGenderNamespaces ) > 0 ) {
 641                         // $wgExtraGenderNamespaces overrides everything
 642                         return true;
 643                 } elseif ( isset( $wgExtraNamespaces[NS_USER] ) && isset( $wgExtraNamespaces[NS_USER_TALK] ) ) {
 644                         /// @todo There may be other gender namespace than NS_USER & NS_USER_TALK in the future
 645                         // $wgExtraNamespaces overrides any gender aliases specified in i18n files
 646                         return false;
 647                 } else {
 648                         // Check what is in i18n files
 649                         $aliases = self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 650                         return count( $aliases ) > 0;
 651                 }
 652         }
 653
 654         /**
 655          * Get a namespace key by value, case insensitive.
 656          * Only matches namespace names for the current language, not the
 657          * canonical ones defined in Namespace.php.
 658          *
 659          * @param string $text
 660          * @return int|bool An integer if $text is a valid value otherwise false
 661          */
 662         function getLocalNsIndex( $text ) {
 663                 $lctext = $this->lc( $text );
 664                 $ids = $this->getNamespaceIds();
 665                 return $ids[$lctext] ?? false;
 666         }
 667
 668         /**
 669          * @return array
 670          */
 671         public function getNamespaceAliases() {
 672                 if ( is_null( $this->namespaceAliases ) ) {
 673                         $aliases = self::$dataCache->getItem( $this->mCode, 'namespaceAliases' );
 674                         if ( !$aliases ) {
 675                                 $aliases = [];
 676                         } else {
 677                                 foreach ( $aliases as $name => $index ) {
 678                                         if ( $index === NS_PROJECT_TALK ) {
 679                                                 unset( $aliases[$name] );
 680                                                 $name = $this->fixVariableInNamespace( $name );
 681                                                 $aliases[$name] = $index;
 682                                         }
 683                                 }
 684                         }
 685
 686                         global $wgExtraGenderNamespaces;
 687                         $genders = $wgExtraGenderNamespaces +
 688                                 (array)self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 689                         foreach ( $genders as $index => $forms ) {
 690                                 foreach ( $forms as $alias ) {
 691                                         $aliases[$alias] = $index;
 692                                 }
 693                         }
 694
 695                         # Also add converted namespace names as aliases, to avoid confusion.
 696                         $convertedNames = [];
 697                         foreach ( $this->getVariants() as $variant ) {
 698                                 if ( $variant === $this->mCode ) {
 699                                         continue;
 700                                 }
 701                                 foreach ( $this->getNamespaces() as $ns => $_ ) {
 702                                         $convertedNames[$this->getConverter()->convertNamespace( $ns, $variant )] = $ns;
 703                                 }
 704                         }
 705
 706                         $this->namespaceAliases = $aliases + $convertedNames;
 707
 708                         # Filter out aliases to namespaces that don't exist, e.g. from extensions
 709                         # that aren't loaded here but are included in the l10n cache.
 710                         # (array_intersect preserves keys from its first argument)
 711                         $this->namespaceAliases = array_intersect(
 712                                 $this->namespaceAliases,
 713                                 array_keys( $this->getNamespaces() )
 714                         );
 715                 }
 716
 717                 return $this->namespaceAliases;
 718         }
 719
 720         /**
 721          * @return array
 722          */
 723         public function getNamespaceIds() {
 724                 if ( is_null( $this->mNamespaceIds ) ) {
 725                         global $wgNamespaceAliases;
 726                         # Put namespace names and aliases into a hashtable.
 727                         # If this is too slow, then we should arrange it so that it is done
 728                         # before caching. The catch is that at pre-cache time, the above
 729                         # class-specific fixup hasn't been done.
 730                         $this->mNamespaceIds = [];
 731                         foreach ( $this->getNamespaces() as $index => $name ) {
 732                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 733                         }
 734                         foreach ( $this->getNamespaceAliases() as $name => $index ) {
 735                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 736                         }
 737                         if ( $wgNamespaceAliases ) {
 738                                 foreach ( $wgNamespaceAliases as $name => $index ) {
 739                                         $this->mNamespaceIds[$this->lc( $name )] = $index;
 740                                 }
 741                         }
 742                 }
 743                 return $this->mNamespaceIds;
 744         }
 745
 746         /**
 747          * Get a namespace key by value, case insensitive.  Canonical namespace
 748          * names override custom ones defined for the current language.
 749          *
 750          * @param string $text
 751          * @return int|bool An integer if $text is a valid value otherwise false
 752          */
 753         public function getNsIndex( $text ) {
 754                 $lctext = $this->lc( $text );
 755                 $ns = MediaWikiServices::getInstance()->getNamespaceInfo()->
 756                         getCanonicalIndex( $lctext );
 757                 if ( $ns !== null ) {
 758                         return $ns;
 759                 }
 760                 $ids = $this->getNamespaceIds();
 761                 return $ids[$lctext] ?? false;
 762         }
 763
 764         /**
 765          * short names for language variants used for language conversion links.
 766          *
 767          * @param string $code
 768          * @param bool $usemsg Use the "variantname-xyz" message if it exists
 769          * @return string
 770          */
 771         public function getVariantname( $code, $usemsg = true ) {
 772                 $msg = "variantname-$code";
 773                 if ( $usemsg && wfMessage( $msg )->exists() ) {
 774                         return $this->getMessageFromDB( $msg );
 775                 }
 776                 $name = self::fetchLanguageName( $code );
 777                 if ( $name ) {
 778                         return $name; # if it's defined as a language name, show that
 779                 } else {
 780                         # otherwise, output the language code
 781                         return $code;
 782                 }
 783         }
 784
 785         /**
 786          * @return string[]|bool List of date format preference keys, or false if disabled.
 787          */
 788         public function getDatePreferences() {
 789                 return self::$dataCache->getItem( $this->mCode, 'datePreferences' );
 790         }
 791
 792         /**
 793          * @return array
 794          */
 795         function getDateFormats() {
 796                 return self::$dataCache->getItem( $this->mCode, 'dateFormats' );
 797         }
 798
 799         /**
 800          * @return array|string
 801          */
 802         public function getDefaultDateFormat() {
 803                 $df = self::$dataCache->getItem( $this->mCode, 'defaultDateFormat' );
 804                 if ( $df === 'dmy or mdy' ) {
 805                         global $wgAmericanDates;
 806                         return $wgAmericanDates ? 'mdy' : 'dmy';
 807                 } else {
 808                         return $df;
 809                 }
 810         }
 811
 812         /**
 813          * @return array
 814          */
 815         public function getDatePreferenceMigrationMap() {
 816                 return self::$dataCache->getItem( $this->mCode, 'datePreferenceMigrationMap' );
 817         }
 818
 819         /**
 820          * @return array
 821          */
 822         public function getExtraUserToggles() {
 823                 wfDeprecated( __METHOD__, '1.34' );
 824                 return [];
 825         }
 826
 827         /**
 828          * @param string $tog
 829          * @return string
 830          */
 831         function getUserToggle( $tog ) {
 832                 return $this->getMessageFromDB( "tog-$tog" );
 833         }
 834
 835         /**
 836          * Get an array of language names, indexed by code.
 837          * @param null|string $inLanguage Code of language in which to return the names
 838          *              Use self::AS_AUTONYMS for autonyms (native names)
 839          * @param string $include One of:
 840          *              self::ALL all available languages
 841          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 842          *              self::SUPPORTED only if the language is in 'mw' *and* has a message file
 843          * @return array Language code => language name (sorted by key)
 844          * @since 1.20
 845          */
 846         public static function fetchLanguageNames( $inLanguage = self::AS_AUTONYMS, $include = 'mw' ) {
 847                 $cacheKey = $inLanguage === self::AS_AUTONYMS ? 'null' : $inLanguage;
 848                 $cacheKey .= ":$include";
 849                 if ( self::$languageNameCache === null ) {
 850                         self::$languageNameCache = new HashBagOStuff( [ 'maxKeys' => 20 ] );
 851                 }
 852
 853                 $ret = self::$languageNameCache->get( $cacheKey );
 854                 if ( !$ret ) {
 855                         $ret = self::fetchLanguageNamesUncached( $inLanguage, $include );
 856                         self::$languageNameCache->set( $cacheKey, $ret );
 857                 }
 858                 return $ret;
 859         }
 860
 861         /**
 862          * Uncached helper for fetchLanguageNames
 863          * @param null|string $inLanguage Code of language in which to return the names
 864          *              Use self::AS_AUTONYMS for autonyms (native names)
 865          * @param string $include One of:
 866          *              self::ALL all available languages
 867          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 868          *              self::SUPPORTED only if the language is in 'mw' *and* has a message file
 869          * @return array Language code => language name (sorted by key)
 870          */
 871         private static function fetchLanguageNamesUncached(
 872                 $inLanguage = self::AS_AUTONYMS,
 873                 $include = 'mw'
 874         ) {
 875                 global $wgExtraLanguageNames, $wgUsePigLatinVariant;
 876
 877                 // If passed an invalid language code to use, fallback to en
 878                 if ( $inLanguage !== self::AS_AUTONYMS && !self::isValidCode( $inLanguage ) ) {
 879                         $inLanguage = 'en';
 880                 }
 881
 882                 $names = [];
 883
 884                 if ( $inLanguage ) {
 885                         # TODO: also include when $inLanguage is null, when this code is more efficient
 886                         Hooks::run( 'LanguageGetTranslatedLanguageNames', [ &$names, $inLanguage ] );
 887                 }
 888
 889                 $mwNames = $wgExtraLanguageNames + MediaWiki\Languages\Data\Names::$names;
 890                 if ( $wgUsePigLatinVariant ) {
 891                         // Pig Latin (for variant development)
 892                         $mwNames['en-x-piglatin'] = 'Igpay Atinlay';
 893                 }
 894
 895                 foreach ( $mwNames as $mwCode => $mwName ) {
 896                         # - Prefer own MediaWiki native name when not using the hook
 897                         # - For other names just add if not added through the hook
 898                         if ( $mwCode === $inLanguage || !isset( $names[$mwCode] ) ) {
 899                                 $names[$mwCode] = $mwName;
 900                         }
 901                 }
 902
 903                 if ( $include === self::ALL ) {
 904                         ksort( $names );
 905                         return $names;
 906                 }
 907
 908                 $returnMw = [];
 909                 $coreCodes = array_keys( $mwNames );
 910                 foreach ( $coreCodes as $coreCode ) {
 911                         $returnMw[$coreCode] = $names[$coreCode];
 912                 }
 913
 914                 if ( $include === self::SUPPORTED ) {
 915                         $namesMwFile = [];
 916                         # We do this using a foreach over the codes instead of a directory
 917                         # loop so that messages files in extensions will work correctly.
 918                         foreach ( $returnMw as $code => $value ) {
 919                                 if ( is_readable( self::getMessagesFileName( $code ) )
 920                                         || is_readable( self::getJsonMessagesFileName( $code ) )
 921                                 ) {
 922                                         $namesMwFile[$code] = $names[$code];
 923                                 }
 924                         }
 925
 926                         ksort( $namesMwFile );
 927                         return $namesMwFile;
 928                 }
 929
 930                 ksort( $returnMw );
 931                 # 'mw' option; default if it's not one of the other two options (all/mwfile)
 932                 return $returnMw;
 933         }
 934
 935         /**
 936          * @param string $code The code of the language for which to get the name
 937          * @param null|string $inLanguage Code of language in which to return the name
 938          *   (SELF::AS_AUTONYMS for autonyms)
 939          * @param string $include See fetchLanguageNames()
 940          * @return string Language name or empty
 941          * @since 1.20
 942          */
 943         public static function fetchLanguageName(
 944                 $code,
 945                 $inLanguage = self::AS_AUTONYMS,
 946                 $include = self::ALL
 947         ) {
 948                 $code = strtolower( $code );
 949                 $array = self::fetchLanguageNames( $inLanguage, $include );
 950                 return !array_key_exists( $code, $array ) ? '' : $array[$code];
 951         }
 952
 953         /**
 954          * Get a message from the MediaWiki namespace.
 955          *
 956          * @param string $msg Message name
 957          * @return string
 958          */
 959         public function getMessageFromDB( $msg ) {
 960                 return $this->msg( $msg )->text();
 961         }
 962
 963         /**
 964          * Get message object in this language. Only for use inside this class.
 965          *
 966          * @param string $msg Message name
 967          * @return Message
 968          */
 969         protected function msg( $msg ) {
 970                 return wfMessage( $msg )->inLanguage( $this );
 971         }
 972
 973         /**
 974          * @param string $key
 975          * @return string
 976          */
 977         public function getMonthName( $key ) {
 978                 return $this->getMessageFromDB( self::$mMonthMsgs[$key - 1] );
 979         }
 980
 981         /**
 982          * @return array
 983          */
 984         public function getMonthNamesArray() {
 985                 $monthNames = [ '' ];
 986                 for ( $i = 1; $i < 13; $i++ ) {
 987                         $monthNames[] = $this->getMonthName( $i );
 988                 }
 989                 return $monthNames;
 990         }
 991
 992         /**
 993          * @param string $key
 994          * @return string
 995          */
 996         public function getMonthNameGen( $key ) {
 997                 return $this->getMessageFromDB( self::$mMonthGenMsgs[$key - 1] );
 998         }
 999
1000         /**
1001          * @param string $key
1002          * @return string
1003          */
1004         public function getMonthAbbreviation( $key ) {
1005                 return $this->getMessageFromDB( self::$mMonthAbbrevMsgs[$key - 1] );
1006         }
1007
1008         /**
1009          * @return array
1010          */
1011         public function getMonthAbbreviationsArray() {
1012                 $monthNames = [ '' ];
1013                 for ( $i = 1; $i < 13; $i++ ) {
1014                         $monthNames[] = $this->getMonthAbbreviation( $i );
1015                 }
1016                 return $monthNames;
1017         }
1018
1019         /**
1020          * @param string $key
1021          * @return string
1022          */
1023         public function getWeekdayName( $key ) {
1024                 return $this->getMessageFromDB( self::$mWeekdayMsgs[$key - 1] );
1025         }
1026
1027         /**
1028          * @param string $key
1029          * @return string
1030          */
1031         function getWeekdayAbbreviation( $key ) {
1032                 return $this->getMessageFromDB( self::$mWeekdayAbbrevMsgs[$key - 1] );
1033         }
1034
1035         /**
1036          * @param string $key
1037          * @return string
1038          */
1039         function getIranianCalendarMonthName( $key ) {
1040                 return $this->getMessageFromDB( self::$mIranianCalendarMonthMsgs[$key - 1] );
1041         }
1042
1043         /**
1044          * @param string $key
1045          * @return string
1046          */
1047         function getHebrewCalendarMonthName( $key ) {
1048                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthMsgs[$key - 1] );
1049         }
1050
1051         /**
1052          * @param string $key
1053          * @return string
1054          */
1055         function getHebrewCalendarMonthNameGen( $key ) {
1056                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthGenMsgs[$key - 1] );
1057         }
1058
1059         /**
1060          * @param string $key
1061          * @return string
1062          */
1063         function getHijriCalendarMonthName( $key ) {
1064                 return $this->getMessageFromDB( self::$mHijriCalendarMonthMsgs[$key - 1] );
1065         }
1066
1067         /**
1068          * Pass through result from $dateTimeObj->format()
1069          * @param DateTime|bool|null &$dateTimeObj
1070          * @param string $ts
1071          * @param DateTimeZone|bool|null $zone
1072          * @param string $code
1073          * @return string
1074          */
1075         private static function dateTimeObjFormat( &$dateTimeObj, $ts, $zone, $code ) {
1076                 if ( !$dateTimeObj ) {
1077                         $dateTimeObj = DateTime::createFromFormat(
1078                                 'YmdHis', $ts, $zone ?: new DateTimeZone( 'UTC' )
1079                         );
1080                 }
1081                 return $dateTimeObj->format( $code );
1082         }
1083
1084         /**
1085          * This is a workalike of PHP's date() function, but with better
1086          * internationalisation, a reduced set of format characters, and a better
1087          * escaping format.
1088          *
1089          * Supported format characters are dDjlNwzWFmMntLoYyaAgGhHiscrUeIOPTZ. See
1090          * the PHP manual for definitions. There are a number of extensions, which
1091          * start with "x":
1092          *
1093          *    xn   Do not translate digits of the next numeric format character
1094          *    xN   Toggle raw digit (xn) flag, stays set until explicitly unset
1095          *    xr   Use roman numerals for the next numeric format character
1096          *    xh   Use hebrew numerals for the next numeric format character
1097          *    xx   Literal x
1098          *    xg   Genitive month name
1099          *
1100          *    xij  j (day number) in Iranian calendar
1101          *    xiF  F (month name) in Iranian calendar
1102          *    xin  n (month number) in Iranian calendar
1103          *    xiy  y (two digit year) in Iranian calendar
1104          *    xiY  Y (full year) in Iranian calendar
1105          *    xit  t (days in month) in Iranian calendar
1106          *    xiz  z (day of the year) in Iranian calendar
1107          *
1108          *    xjj  j (day number) in Hebrew calendar
1109          *    xjF  F (month name) in Hebrew calendar
1110          *    xjt  t (days in month) in Hebrew calendar
1111          *    xjx  xg (genitive month name) in Hebrew calendar
1112          *    xjn  n (month number) in Hebrew calendar
1113          *    xjY  Y (full year) in Hebrew calendar
1114          *
1115          *    xmj  j (day number) in Hijri calendar
1116          *    xmF  F (month name) in Hijri calendar
1117          *    xmn  n (month number) in Hijri calendar
1118          *    xmY  Y (full year) in Hijri calendar
1119          *
1120          *    xkY  Y (full year) in Thai solar calendar. Months and days are
1121          *                       identical to the Gregorian calendar
1122          *    xoY  Y (full year) in Minguo calendar or Juche year.
1123          *                       Months and days are identical to the
1124          *                       Gregorian calendar
1125          *    xtY  Y (full year) in Japanese nengo. Months and days are
1126          *                       identical to the Gregorian calendar
1127          *
1128          * Characters enclosed in double quotes will be considered literal (with
1129          * the quotes themselves removed). Unmatched quotes will be considered
1130          * literal quotes. Example:
1131          *
1132          * "The month is" F       => The month is January
1133          * i's"                   => 20'11"
1134          *
1135          * Backslash escaping is also supported.
1136          *
1137          * Input timestamp is assumed to be pre-normalized to the desired local
1138          * time zone, if any. Note that the format characters crUeIOPTZ will assume
1139          * $ts is UTC if $zone is not given.
1140          *
1141          * @param string $format
1142          * @param string $ts 14-character timestamp
1143          *      YYYYMMDDHHMMSS
1144          *      01234567890123
1145          * @param DateTimeZone|null $zone Timezone of $ts
1146          * @param int &$ttl The amount of time (in seconds) the output may be cached for.
1147          * Only makes sense if $ts is the current time.
1148          * @todo handling of "o" format character for Iranian, Hebrew, Hijri & Thai?
1149          *
1150          * @throws MWException
1151          * @return string
1152          */
1153         public function sprintfDate( $format, $ts, DateTimeZone $zone = null, &$ttl = 'unused' ) {
1154                 $s = '';
1155                 $raw = false;
1156                 $roman = false;
1157                 $hebrewNum = false;
1158                 $dateTimeObj = false;
1159                 $rawToggle = false;
1160                 $iranian = false;
1161                 $hebrew = false;
1162                 $hijri = false;
1163                 $thai = false;
1164                 $minguo = false;
1165                 $tenno = false;
1166
1167                 $usedSecond = false;
1168                 $usedMinute = false;
1169                 $usedHour = false;
1170                 $usedAMPM = false;
1171                 $usedDay = false;
1172                 $usedWeek = false;
1173                 $usedMonth = false;
1174                 $usedYear = false;
1175                 $usedISOYear = false;
1176                 $usedIsLeapYear = false;
1177
1178                 $usedHebrewMonth = false;
1179                 $usedIranianMonth = false;
1180                 $usedHijriMonth = false;
1181                 $usedHebrewYear = false;
1182                 $usedIranianYear = false;
1183                 $usedHijriYear = false;
1184                 $usedTennoYear = false;
1185
1186                 if ( strlen( $ts ) !== 14 ) {
1187                         throw new MWException( __METHOD__ . ": The timestamp $ts should have 14 characters" );
1188                 }
1189
1190                 if ( !ctype_digit( $ts ) ) {
1191                         throw new MWException( __METHOD__ . ": The timestamp $ts should be a number" );
1192                 }
1193
1194                 $formatLength = strlen( $format );
1195                 for ( $p = 0; $p < $formatLength; $p++ ) {
1196                         $num = false;
1197                         $code = $format[$p];
1198                         if ( $code == 'x' && $p < $formatLength - 1 ) {
1199                                 $code .= $format[++$p];
1200                         }
1201
1202                         if ( ( $code === 'xi'
1203                                         || $code === 'xj'
1204                                         || $code === 'xk'
1205                                         || $code === 'xm'
1206                                         || $code === 'xo'
1207                                         || $code === 'xt' )
1208                                 && $p < $formatLength - 1 ) {
1209                                 $code .= $format[++$p];
1210                         }
1211
1212                         switch ( $code ) {
1213                                 case 'xx':
1214                                         $s .= 'x';
1215                                         break;
1216                                 case 'xn':
1217                                         $raw = true;
1218                                         break;
1219                                 case 'xN':
1220                                         $rawToggle = !$rawToggle;
1221                                         break;
1222                                 case 'xr':
1223                                         $roman = true;
1224                                         break;
1225                                 case 'xh':
1226                                         $hebrewNum = true;
1227                                         break;
1228                                 case 'xg':
1229                                         $usedMonth = true;
1230                                         $s .= $this->getMonthNameGen( substr( $ts, 4, 2 ) );
1231                                         break;
1232                                 case 'xjx':
1233                                         $usedHebrewMonth = true;
1234                                         if ( !$hebrew ) {
1235                                                 $hebrew = self::tsToHebrew( $ts );
1236                                         }
1237                                         $s .= $this->getHebrewCalendarMonthNameGen( $hebrew[1] );
1238                                         break;
1239                                 case 'd':
1240                                         $usedDay = true;
1241                                         $num = substr( $ts, 6, 2 );
1242                                         break;
1243                                 case 'D':
1244                                         $usedDay = true;
1245                                         $s .= $this->getWeekdayAbbreviation(
1246                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1247                                         );
1248                                         break;
1249                                 case 'j':
1250                                         $usedDay = true;
1251                                         $num = intval( substr( $ts, 6, 2 ) );
1252                                         break;
1253                                 case 'xij':
1254                                         $usedDay = true;
1255                                         if ( !$iranian ) {
1256                                                 $iranian = self::tsToIranian( $ts );
1257                                         }
1258                                         $num = $iranian[2];
1259                                         break;
1260                                 case 'xmj':
1261                                         $usedDay = true;
1262                                         if ( !$hijri ) {
1263                                                 $hijri = self::tsToHijri( $ts );
1264                                         }
1265                                         $num = $hijri[2];
1266                                         break;
1267                                 case 'xjj':
1268                                         $usedDay = true;
1269                                         if ( !$hebrew ) {
1270                                                 $hebrew = self::tsToHebrew( $ts );
1271                                         }
1272                                         $num = $hebrew[2];
1273                                         break;
1274                                 case 'l':
1275                                         $usedDay = true;
1276                                         $s .= $this->getWeekdayName(
1277                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1278                                         );
1279                                         break;
1280                                 case 'F':
1281                                         $usedMonth = true;
1282                                         $s .= $this->getMonthName( substr( $ts, 4, 2 ) );
1283                                         break;
1284                                 case 'xiF':
1285                                         $usedIranianMonth = true;
1286                                         if ( !$iranian ) {
1287                                                 $iranian = self::tsToIranian( $ts );
1288                                         }
1289                                         $s .= $this->getIranianCalendarMonthName( $iranian[1] );
1290                                         break;
1291                                 case 'xmF':
1292                                         $usedHijriMonth = true;
1293                                         if ( !$hijri ) {
1294                                                 $hijri = self::tsToHijri( $ts );
1295                                         }
1296                                         $s .= $this->getHijriCalendarMonthName( $hijri[1] );
1297                                         break;
1298                                 case 'xjF':
1299                                         $usedHebrewMonth = true;
1300                                         if ( !$hebrew ) {
1301                                                 $hebrew = self::tsToHebrew( $ts );
1302                                         }
1303                                         $s .= $this->getHebrewCalendarMonthName( $hebrew[1] );
1304                                         break;
1305                                 case 'm':
1306                                         $usedMonth = true;
1307                                         $num = substr( $ts, 4, 2 );
1308                                         break;
1309                                 case 'M':
1310                                         $usedMonth = true;
1311                                         $s .= $this->getMonthAbbreviation( substr( $ts, 4, 2 ) );
1312                                         break;
1313                                 case 'n':
1314                                         $usedMonth = true;
1315                                         $num = intval( substr( $ts, 4, 2 ) );
1316                                         break;
1317                                 case 'xin':
1318                                         $usedIranianMonth = true;
1319                                         if ( !$iranian ) {
1320                                                 $iranian = self::tsToIranian( $ts );
1321                                         }
1322                                         $num = $iranian[1];
1323                                         break;
1324                                 case 'xmn':
1325                                         $usedHijriMonth = true;
1326                                         if ( !$hijri ) {
1327                                                 $hijri = self::tsToHijri( $ts );
1328                                         }
1329                                         $num = $hijri[1];
1330                                         break;
1331                                 case 'xjn':
1332                                         $usedHebrewMonth = true;
1333                                         if ( !$hebrew ) {
1334                                                 $hebrew = self::tsToHebrew( $ts );
1335                                         }
1336                                         $num = $hebrew[1];
1337                                         break;
1338                                 case 'xjt':
1339                                         $usedHebrewMonth = true;
1340                                         if ( !$hebrew ) {
1341                                                 $hebrew = self::tsToHebrew( $ts );
1342                                         }
1343                                         $num = $hebrew[3];
1344                                         break;
1345                                 case 'Y':
1346                                         $usedYear = true;
1347                                         $num = substr( $ts, 0, 4 );
1348                                         break;
1349                                 case 'xiY':
1350                                         $usedIranianYear = true;
1351                                         if ( !$iranian ) {
1352                                                 $iranian = self::tsToIranian( $ts );
1353                                         }
1354                                         $num = $iranian[0];
1355                                         break;
1356                                 case 'xmY':
1357                                         $usedHijriYear = true;
1358                                         if ( !$hijri ) {
1359                                                 $hijri = self::tsToHijri( $ts );
1360                                         }
1361                                         $num = $hijri[0];
1362                                         break;
1363                                 case 'xjY':
1364                                         $usedHebrewYear = true;
1365                                         if ( !$hebrew ) {
1366                                                 $hebrew = self::tsToHebrew( $ts );
1367                                         }
1368                                         $num = $hebrew[0];
1369                                         break;
1370                                 case 'xkY':
1371                                         $usedYear = true;
1372                                         if ( !$thai ) {
1373                                                 $thai = self::tsToYear( $ts, 'thai' );
1374                                         }
1375                                         $num = $thai[0];
1376                                         break;
1377                                 case 'xoY':
1378                                         $usedYear = true;
1379                                         if ( !$minguo ) {
1380                                                 $minguo = self::tsToYear( $ts, 'minguo' );
1381                                         }
1382                                         $num = $minguo[0];
1383                                         break;
1384                                 case 'xtY':
1385                                         $usedTennoYear = true;
1386                                         if ( !$tenno ) {
1387                                                 $tenno = self::tsToYear( $ts, 'tenno' );
1388                                         }
1389                                         $num = $tenno[0];
1390                                         break;
1391                                 case 'y':
1392                                         $usedYear = true;
1393                                         $num = substr( $ts, 2, 2 );
1394                                         break;
1395                                 case 'xiy':
1396                                         $usedIranianYear = true;
1397                                         if ( !$iranian ) {
1398                                                 $iranian = self::tsToIranian( $ts );
1399                                         }
1400                                         $num = substr( $iranian[0], -2 );
1401                                         break;
1402                                 case 'xit':
1403                                         $usedIranianYear = true;
1404                                         if ( !$iranian ) {
1405                                                 $iranian = self::tsToIranian( $ts );
1406                                         }
1407                                         $num = self::$IRANIAN_DAYS[$iranian[1] - 1];
1408                                         break;
1409                                 case 'xiz':
1410                                         $usedIranianYear = true;
1411                                         if ( !$iranian ) {
1412                                                 $iranian = self::tsToIranian( $ts );
1413                                         }
1414                                         $num = $iranian[3];
1415                                         break;
1416                                 case 'a':
1417                                         $usedAMPM = true;
1418                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'am' : 'pm';
1419                                         break;
1420                                 case 'A':
1421                                         $usedAMPM = true;
1422                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'AM' : 'PM';
1423                                         break;
1424                                 case 'g':
1425                                         $usedHour = true;
1426                                         $h = substr( $ts, 8, 2 );
1427                                         $num = $h % 12 ?: 12;
1428                                         break;
1429                                 case 'G':
1430                                         $usedHour = true;
1431                                         $num = intval( substr( $ts, 8, 2 ) );
1432                                         break;
1433                                 case 'h':
1434                                         $usedHour = true;
1435                                         $h = substr( $ts, 8, 2 );
1436                                         $num = sprintf( '%02d', $h % 12 ?: 12 );
1437                                         break;
1438                                 case 'H':
1439                                         $usedHour = true;
1440                                         $num = substr( $ts, 8, 2 );
1441                                         break;
1442                                 case 'i':
1443                                         $usedMinute = true;
1444                                         $num = substr( $ts, 10, 2 );
1445                                         break;
1446                                 case 's':
1447                                         $usedSecond = true;
1448                                         $num = substr( $ts, 12, 2 );
1449                                         break;
1450                                 case 'c':
1451                                 case 'r':
1452                                         $usedSecond = true;
1453                                         // fall through
1454                                 case 'e':
1455                                 case 'O':
1456                                 case 'P':
1457                                 case 'T':
1458                                         $s .= self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1459                                         break;
1460                                 case 'w':
1461                                 case 'N':
1462                                 case 'z':
1463                                         $usedDay = true;
1464                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1465                                         break;
1466                                 case 'W':
1467                                         $usedWeek = true;
1468                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1469                                         break;
1470                                 case 't':
1471                                         $usedMonth = true;
1472                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1473                                         break;
1474                                 case 'L':
1475                                         $usedIsLeapYear = true;
1476                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1477                                         break;
1478                                 case 'o':
1479                                         $usedISOYear = true;
1480                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1481                                         break;
1482                                 case 'U':
1483                                         $usedSecond = true;
1484                                         // fall through
1485                                 case 'I':
1486                                 case 'Z':
1487                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1488                                         break;
1489                                 case '\\':
1490                                         # Backslash escaping
1491                                         if ( $p < $formatLength - 1 ) {
1492                                                 $s .= $format[++$p];
1493                                         } else {
1494                                                 $s .= '\\';
1495                                         }
1496                                         break;
1497                                 case '"':
1498                                         # Quoted literal
1499                                         if ( $p < $formatLength - 1 ) {
1500                                                 $endQuote = strpos( $format, '"', $p + 1 );
1501                                                 if ( $endQuote === false ) {
1502                                                         # No terminating quote, assume literal "
1503                                                         $s .= '"';
1504                                                 } else {
1505                                                         $s .= substr( $format, $p + 1, $endQuote - $p - 1 );
1506                                                         $p = $endQuote;
1507                                                 }
1508                                         } else {
1509                                                 # Quote at end of string, assume literal "
1510                                                 $s .= '"';
1511                                         }
1512                                         break;
1513                                 default:
1514                                         $s .= $format[$p];
1515                         }
1516                         if ( $num !== false ) {
1517                                 if ( $rawToggle || $raw ) {
1518                                         $s .= $num;
1519                                         $raw = false;
1520                                 } elseif ( $roman ) {
1521                                         $s .= self::romanNumeral( $num );
1522                                         $roman = false;
1523                                 } elseif ( $hebrewNum ) {
1524                                         $s .= self::hebrewNumeral( $num );
1525                                         $hebrewNum = false;
1526                                 } else {
1527                                         $s .= $this->formatNum( $num, true );
1528                                 }
1529                         }
1530                 }
1531
1532                 if ( $ttl === 'unused' ) {
1533                         // No need to calculate the TTL, the caller wont use it anyway.
1534                 } elseif ( $usedSecond ) {
1535                         $ttl = 1;
1536                 } elseif ( $usedMinute ) {
1537                         $ttl = 60 - substr( $ts, 12, 2 );
1538                 } elseif ( $usedHour ) {
1539                         $ttl = 3600 - substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1540                 } elseif ( $usedAMPM ) {
1541                         $ttl = 43200 - ( substr( $ts, 8, 2 ) % 12 ) * 3600 -
1542                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1543                 } elseif (
1544                         $usedDay ||
1545                         $usedHebrewMonth ||
1546                         $usedIranianMonth ||
1547                         $usedHijriMonth ||
1548                         $usedHebrewYear ||
1549                         $usedIranianYear ||
1550                         $usedHijriYear ||
1551                         $usedTennoYear
1552                 ) {
1553                         // @todo Someone who understands the non-Gregorian calendars
1554                         // should write proper logic for them so that they don't need purged every day.
1555                         $ttl = 86400 - substr( $ts, 8, 2 ) * 3600 -
1556                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1557                 } else {
1558                         $possibleTtls = [];
1559                         $timeRemainingInDay = 86400 - substr( $ts, 8, 2 ) * 3600 -
1560                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1561                         if ( $usedWeek ) {
1562                                 $possibleTtls[] =
1563                                         ( 7 - self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400 +
1564                                         $timeRemainingInDay;
1565                         } elseif ( $usedISOYear ) {
1566                                 // December 28th falls on the last ISO week of the year, every year.
1567                                 // The last ISO week of a year can be 52 or 53.
1568                                 $lastWeekOfISOYear = DateTime::createFromFormat(
1569                                         'Ymd',
1570                                         substr( $ts, 0, 4 ) . '1228',
1571                                         $zone ?: new DateTimeZone( 'UTC' )
1572                                 )->format( 'W' );
1573                                 $currentISOWeek = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'W' );
1574                                 $weeksRemaining = $lastWeekOfISOYear - $currentISOWeek;
1575                                 $timeRemainingInWeek =
1576                                         ( 7 - self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400
1577                                         + $timeRemainingInDay;
1578                                 $possibleTtls[] = $weeksRemaining * 604800 + $timeRemainingInWeek;
1579                         }
1580
1581                         if ( $usedMonth ) {
1582                                 $possibleTtls[] =
1583                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 't' ) -
1584                                                 substr( $ts, 6, 2 ) ) * 86400
1585                                         + $timeRemainingInDay;
1586                         } elseif ( $usedYear ) {
1587                                 $possibleTtls[] =
1588                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1589                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1590                                         + $timeRemainingInDay;
1591                         } elseif ( $usedIsLeapYear ) {
1592                                 $year = substr( $ts, 0, 4 );
1593                                 $timeRemainingInYear =
1594                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1595                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1596                                         + $timeRemainingInDay;
1597                                 $mod = $year % 4;
1598                                 if ( $mod || ( !( $year % 100 ) && $year % 400 ) ) {
1599                                         // this isn't a leap year. see when the next one starts
1600                                         $nextCandidate = $year - $mod + 4;
1601                                         if ( $nextCandidate % 100 || !( $nextCandidate % 400 ) ) {
1602                                                 $possibleTtls[] = ( $nextCandidate - $year - 1 ) * 365 * 86400 +
1603                                                         $timeRemainingInYear;
1604                                         } else {
1605                                                 $possibleTtls[] = ( $nextCandidate - $year + 3 ) * 365 * 86400 +
1606                                                         $timeRemainingInYear;
1607                                         }
1608                                 } else {
1609                                         // this is a leap year, so the next year isn't
1610                                         $possibleTtls[] = $timeRemainingInYear;
1611                                 }
1612                         }
1613
1614                         if ( $possibleTtls ) {
1615                                 $ttl = min( $possibleTtls );
1616                         }
1617                 }
1618
1619                 return $s;
1620         }
1621
1622         private static $GREG_DAYS = [ 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 ];
1623         private static $IRANIAN_DAYS = [ 31, 31, 31, 31, 31, 31, 30, 30, 30, 30, 30, 29 ];
1624
1625         /**
1626          * Algorithm by Roozbeh Pournader and Mohammad Toossi to convert
1627          * Gregorian dates to Iranian dates. Originally written in C, it
1628          * is released under the terms of GNU Lesser General Public
1629          * License. Conversion to PHP was performed by Niklas Laxström.
1630          *
1631          * Link: http://www.farsiweb.info/jalali/jalali.c
1632          *
1633          * @param string $ts
1634          *
1635          * @return int[]
1636          */
1637         private static function tsToIranian( $ts ) {
1638                 $gy = substr( $ts, 0, 4 ) - 1600;
1639                 $gm = substr( $ts, 4, 2 ) - 1;
1640                 $gd = substr( $ts, 6, 2 ) - 1;
1641
1642                 # Days passed from the beginning (including leap years)
1643                 $gDayNo = 365 * $gy
1644                         + floor( ( $gy + 3 ) / 4 )
1645                         - floor( ( $gy + 99 ) / 100 )
1646                         + floor( ( $gy + 399 ) / 400 );
1647
1648                 // Add days of the past months of this year
1649                 for ( $i = 0; $i < $gm; $i++ ) {
1650                         $gDayNo += self::$GREG_DAYS[$i];
1651                 }
1652
1653                 // Leap years
1654                 if ( $gm > 1 && ( ( $gy % 4 === 0 && $gy % 100 !== 0 || ( $gy % 400 == 0 ) ) ) ) {
1655                         $gDayNo++;
1656                 }
1657
1658                 // Days passed in current month
1659                 $gDayNo += (int)$gd;
1660
1661                 $jDayNo = $gDayNo - 79;
1662
1663                 $jNp = floor( $jDayNo / 12053 );
1664                 $jDayNo %= 12053;
1665
1666                 $jy = 979 + 33 * $jNp + 4 * floor( $jDayNo / 1461 );
1667                 $jDayNo %= 1461;
1668
1669                 if ( $jDayNo >= 366 ) {
1670                         $jy += floor( ( $jDayNo - 1 ) / 365 );
1671                         $jDayNo = floor( ( $jDayNo - 1 ) % 365 );
1672                 }
1673
1674                 $jz = $jDayNo;
1675
1676                 for ( $i = 0; $i < 11 && $jDayNo >= self::$IRANIAN_DAYS[$i]; $i++ ) {
1677                         $jDayNo -= self::$IRANIAN_DAYS[$i];
1678                 }
1679
1680                 $jm = $i + 1;
1681                 $jd = $jDayNo + 1;
1682
1683                 return [ $jy, $jm, $jd, $jz ];
1684         }
1685
1686         /**
1687          * Converting Gregorian dates to Hijri dates.
1688          *
1689          * Based on a PHP-Nuke block by Sharjeel which is released under GNU/GPL license
1690          *
1691          * @see https://phpnuke.org/modules.php?name=News&file=article&sid=8234&mode=thread&order=0&thold=0
1692          *
1693          * @param string $ts
1694          *
1695          * @return int[]
1696          */
1697         private static function tsToHijri( $ts ) {
1698                 $year = substr( $ts, 0, 4 );
1699                 $month = substr( $ts, 4, 2 );
1700                 $day = substr( $ts, 6, 2 );
1701
1702                 $zyr = $year;
1703                 $zd = $day;
1704                 $zm = $month;
1705                 $zy = $zyr;
1706
1707                 if (
1708                         ( $zy > 1582 ) || ( ( $zy == 1582 ) && ( $zm > 10 ) ) ||
1709                         ( ( $zy == 1582 ) && ( $zm == 10 ) && ( $zd > 14 ) )
1710                 ) {
1711                         $zjd = (int)( ( 1461 * ( $zy + 4800 + (int)( ( $zm - 14 ) / 12 ) ) ) / 4 ) +
1712                                         (int)( ( 367 * ( $zm - 2 - 12 * ( (int)( ( $zm - 14 ) / 12 ) ) ) ) / 12 ) -
1713                                         (int)( ( 3 * (int)( ( ( $zy + 4900 + (int)( ( $zm - 14 ) / 12 ) ) / 100 ) ) ) / 4 ) +
1714                                         $zd - 32075;
1715                 } else {
1716                         $zjd = 367 * $zy - (int)( ( 7 * ( $zy + 5001 + (int)( ( $zm - 9 ) / 7 ) ) ) / 4 ) +
1717                                                                 (int)( ( 275 * $zm ) / 9 ) + $zd + 1729777;
1718                 }
1719
1720                 $zl = $zjd - 1948440 + 10632;
1721                 $zn = (int)( ( $zl - 1 ) / 10631 );
1722                 $zl = $zl - 10631 * $zn + 354;
1723                 $zj = ( (int)( ( 10985 - $zl ) / 5316 ) ) * ( (int)( ( 50 * $zl ) / 17719 ) ) +
1724                         ( (int)( $zl / 5670 ) ) * ( (int)( ( 43 * $zl ) / 15238 ) );
1725                 $zl = $zl - ( (int)( ( 30 - $zj ) / 15 ) ) * ( (int)( ( 17719 * $zj ) / 50 ) ) -
1726                         ( (int)( $zj / 16 ) ) * ( (int)( ( 15238 * $zj ) / 43 ) ) + 29;
1727                 $zm = (int)( ( 24 * $zl ) / 709 );
1728                 $zd = $zl - (int)( ( 709 * $zm ) / 24 );
1729                 $zy = 30 * $zn + $zj - 30;
1730
1731                 return [ $zy, $zm, $zd ];
1732         }
1733
1734         /**
1735          * Converting Gregorian dates to Hebrew dates.
1736          *
1737          * Based on a JavaScript code by Abu Mami and Yisrael Hersch
1738          * (abu-mami@kaluach.net, http://www.kaluach.net), who permitted
1739          * to translate the relevant functions into PHP and release them under
1740          * GNU GPL.
1741          *
1742          * The months are counted from Tishrei = 1. In a leap year, Adar I is 13
1743          * and Adar II is 14. In a non-leap year, Adar is 6.
1744          *
1745          * @param string $ts
1746          *
1747          * @return int[]
1748          */
1749         private static function tsToHebrew( $ts ) {
1750                 # Parse date
1751                 $year = substr( $ts, 0, 4 );
1752                 $month = substr( $ts, 4, 2 );
1753                 $day = substr( $ts, 6, 2 );
1754
1755                 # Calculate Hebrew year
1756                 $hebrewYear = $year + 3760;
1757
1758                 # Month number when September = 1, August = 12
1759                 $month += 4;
1760                 if ( $month > 12 ) {
1761                         # Next year
1762                         $month -= 12;
1763                         $year++;
1764                         $hebrewYear++;
1765                 }
1766
1767                 # Calculate day of year from 1 September
1768                 $dayOfYear = $day;
1769                 for ( $i = 1; $i < $month; $i++ ) {
1770                         if ( $i == 6 ) {
1771                                 # February
1772                                 $dayOfYear += 28;
1773                                 # Check if the year is leap
1774                                 if ( $year % 400 == 0 || ( $year % 4 == 0 && $year % 100 > 0 ) ) {
1775                                         $dayOfYear++;
1776                                 }
1777                         } elseif ( $i == 8 || $i == 10 || $i == 1 || $i == 3 ) {
1778                                 $dayOfYear += 30;
1779                         } else {
1780                                 $dayOfYear += 31;
1781                         }
1782                 }
1783
1784                 # Calculate the start of the Hebrew year
1785                 $start = self::hebrewYearStart( $hebrewYear );
1786
1787                 # Calculate next year's start
1788                 if ( $dayOfYear <= $start ) {
1789                         # Day is before the start of the year - it is the previous year
1790                         # Next year's start
1791                         $nextStart = $start;
1792                         # Previous year
1793                         $year--;
1794                         $hebrewYear--;
1795                         # Add days since previous year's 1 September
1796                         $dayOfYear += 365;
1797                         if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1798                                 # Leap year
1799                                 $dayOfYear++;
1800                         }
1801                         # Start of the new (previous) year
1802                         $start = self::hebrewYearStart( $hebrewYear );
1803                 } else {
1804                         # Next year's start
1805                         $nextStart = self::hebrewYearStart( $hebrewYear + 1 );
1806                 }
1807
1808                 # Calculate Hebrew day of year
1809                 $hebrewDayOfYear = $dayOfYear - $start;
1810
1811                 # Difference between year's days
1812                 $diff = $nextStart - $start;
1813                 # Add 12 (or 13 for leap years) days to ignore the difference between
1814                 # Hebrew and Gregorian year (353 at least vs. 365/6) - now the
1815                 # difference is only about the year type
1816                 if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1817                         $diff += 13;
1818                 } else {
1819                         $diff += 12;
1820                 }
1821
1822                 # Check the year pattern, and is leap year
1823                 # 0 means an incomplete year, 1 means a regular year, 2 means a complete year
1824                 # This is mod 30, to work on both leap years (which add 30 days of Adar I)
1825                 # and non-leap years
1826                 $yearPattern = $diff % 30;
1827                 # Check if leap year
1828                 $isLeap = $diff >= 30;
1829
1830                 # Calculate day in the month from number of day in the Hebrew year
1831                 # Don't check Adar - if the day is not in Adar, we will stop before;
1832                 # if it is in Adar, we will use it to check if it is Adar I or Adar II
1833                 $hebrewDay = $hebrewDayOfYear;
1834                 $hebrewMonth = 1;
1835                 $days = 0;
1836                 while ( $hebrewMonth <= 12 ) {
1837                         # Calculate days in this month
1838                         if ( $isLeap && $hebrewMonth == 6 ) {
1839                                 # Leap year - has Adar I, with 30 days, and Adar II, with 29 days
1840                                 $days = 30;
1841                                 if ( $hebrewDay <= $days ) {
1842                                         # Day in Adar I
1843                                         $hebrewMonth = 13;
1844                                 } else {
1845                                         # Subtract the days of Adar I
1846                                         $hebrewDay -= $days;
1847                                         # Try Adar II
1848                                         $days = 29;
1849                                         if ( $hebrewDay <= $days ) {
1850                                                 # Day in Adar II
1851                                                 $hebrewMonth = 14;
1852                                         }
1853                                 }
1854                         } elseif ( $hebrewMonth == 2 && $yearPattern == 2 ) {
1855                                 # Cheshvan in a complete year (otherwise as the rule below)
1856                                 $days = 30;
1857                         } elseif ( $hebrewMonth == 3 && $yearPattern == 0 ) {
1858                                 # Kislev in an incomplete year (otherwise as the rule below)
1859                                 $days = 29;
1860                         } else {
1861                                 # Odd months have 30 days, even have 29
1862                                 $days = 30 - ( $hebrewMonth - 1 ) % 2;
1863                         }
1864                         if ( $hebrewDay <= $days ) {
1865                                 # In the current month
1866                                 break;
1867                         } else {
1868                                 # Subtract the days of the current month
1869                                 $hebrewDay -= $days;
1870                                 # Try in the next month
1871                                 $hebrewMonth++;
1872                         }
1873                 }
1874
1875                 return [ $hebrewYear, $hebrewMonth, $hebrewDay, $days ];
1876         }
1877
1878         /**
1879          * This calculates the Hebrew year start, as days since 1 September.
1880          * Based on Carl Friedrich Gauss algorithm for finding Easter date.
1881          * Used for Hebrew date.
1882          *
1883          * @param int $year
1884          *
1885          * @return string
1886          */
1887         private static function hebrewYearStart( $year ) {
1888                 $a = intval( ( 12 * ( $year - 1 ) + 17 ) % 19 );
1889                 $b = intval( ( $year - 1 ) % 4 );
1890                 $m = 32.044093161144 + 1.5542417966212 * $a + $b / 4.0 - 0.0031777940220923 * ( $year - 1 );
1891                 if ( $m < 0 ) {
1892                         $m--;
1893                 }
1894                 $Mar = intval( $m );
1895                 if ( $m < 0 ) {
1896                         $m++;
1897                 }
1898                 $m -= $Mar;
1899
1900                 $c = intval( ( $Mar + 3 * ( $year - 1 ) + 5 * $b + 5 ) % 7 );
1901                 if ( $c == 0 && $a > 11 && $m >= 0.89772376543210 ) {
1902                         $Mar++;
1903                 } elseif ( $c == 1 && $a > 6 && $m >= 0.63287037037037 ) {
1904                         $Mar += 2;
1905                 } elseif ( $c == 2 || $c == 4 || $c == 6 ) {
1906                         $Mar++;
1907                 }
1908
1909                 $Mar += intval( ( $year - 3761 ) / 100 ) - intval( ( $year - 3761 ) / 400 ) - 24;
1910                 return $Mar;
1911         }
1912
1913         /**
1914          * Algorithm to convert Gregorian dates to Thai solar dates,
1915          * Minguo dates or Minguo dates.
1916          *
1917          * Link: https://en.wikipedia.org/wiki/Thai_solar_calendar
1918          *       https://en.wikipedia.org/wiki/Minguo_calendar
1919          *       https://en.wikipedia.org/wiki/Japanese_era_name
1920          *
1921          * @param string $ts 14-character timestamp
1922          * @param string $cName Calender name
1923          * @return array Converted year, month, day
1924          */
1925         private static function tsToYear( $ts, $cName ) {
1926                 $gy = substr( $ts, 0, 4 );
1927                 $gm = substr( $ts, 4, 2 );
1928                 $gd = substr( $ts, 6, 2 );
1929
1930                 if ( !strcmp( $cName, 'thai' ) ) {
1931                         # Thai solar dates
1932                         # Add 543 years to the Gregorian calendar
1933                         # Months and days are identical
1934                         $gy_offset = $gy + 543;
1935                         # fix for dates between 1912 and 1941
1936                         # https://en.wikipedia.org/?oldid=836596673#New_year
1937                         if ( $gy >= 1912 && $gy <= 1940 ) {
1938                                 if ( $gm <= 3 ) {
1939                                         $gy_offset--;
1940                                 }
1941                                 $gm = ( $gm - 3 ) % 12;
1942                         }
1943                 } elseif ( ( !strcmp( $cName, 'minguo' ) ) || !strcmp( $cName, 'juche' ) ) {
1944                         # Minguo dates
1945                         # Deduct 1911 years from the Gregorian calendar
1946                         # Months and days are identical
1947                         $gy_offset = $gy - 1911;
1948                 } elseif ( !strcmp( $cName, 'tenno' ) ) {
1949                         # Nengō dates up to Meiji period
1950                         # Deduct years from the Gregorian calendar
1951                         # depending on the nengo periods
1952                         # Months and days are identical
1953                         if ( ( $gy < 1912 )
1954                                 || ( ( $gy == 1912 ) && ( $gm < 7 ) )
1955                                 || ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd < 31 ) )
1956                         ) {
1957                                 # Meiji period
1958                                 $gy_gannen = $gy - 1868 + 1;
1959                                 $gy_offset = $gy_gannen;
1960                                 if ( $gy_gannen == 1 ) {
1961                                         $gy_offset = '元';
1962                                 }
1963                                 $gy_offset = '明治' . $gy_offset;
1964                         } elseif (
1965                                 ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd == 31 ) ) ||
1966                                 ( ( $gy == 1912 ) && ( $gm >= 8 ) ) ||
1967                                 ( ( $gy > 1912 ) && ( $gy < 1926 ) ) ||
1968                                 ( ( $gy == 1926 ) && ( $gm < 12 ) ) ||
1969                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd < 26 ) )
1970                         ) {
1971                                 # Taishō period
1972                                 $gy_gannen = $gy - 1912 + 1;
1973                                 $gy_offset = $gy_gannen;
1974                                 if ( $gy_gannen == 1 ) {
1975                                         $gy_offset = '元';
1976                                 }
1977                                 $gy_offset = '大正' . $gy_offset;
1978                         } elseif (
1979                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd >= 26 ) ) ||
1980                                 ( ( $gy > 1926 ) && ( $gy < 1989 ) ) ||
1981                                 ( ( $gy == 1989 ) && ( $gm == 1 ) && ( $gd < 8 ) )
1982                         ) {
1983                                 # Shōwa period
1984                                 $gy_gannen = $gy - 1926 + 1;
1985                                 $gy_offset = $gy_gannen;
1986                                 if ( $gy_gannen == 1 ) {
1987                                         $gy_offset = '元';
1988                                 }
1989                                 $gy_offset = '昭和' . $gy_offset;
1990                         } elseif (
1991                                 ( ( $gy == 1989 ) && ( $gm == 1 ) && ( $gd >= 8 ) ) ||
1992                                 ( ( $gy > 1989 ) && ( $gy < 2019 ) ) ||
1993                                 ( ( $gy == 2019 ) && ( $gm < 5 ) )
1994                         ) {
1995                                 # Heisei period
1996                                 $gy_gannen = $gy - 1989 + 1;
1997                                 $gy_offset = $gy_gannen;
1998                                 if ( $gy_gannen == 1 ) {
1999                                         $gy_offset = '元';
2000                                 }
2001                                 $gy_offset = '平成' . $gy_offset;
2002                         } else {
2003                                 # Reiwa period
2004                                 $gy_gannen = $gy - 2019 + 1;
2005                                 $gy_offset = $gy_gannen;
2006                                 if ( $gy_gannen == 1 ) {
2007                                         $gy_offset = '元';
2008                                 }
2009                                 $gy_offset = '令和' . $gy_offset;
2010                         }
2011                 } else {
2012                         $gy_offset = $gy;
2013                 }
2014
2015                 return [ $gy_offset, $gm, $gd ];
2016         }
2017
2018         /**
2019          * Gets directionality of the first strongly directional codepoint, for embedBidi()
2020          *
2021          * This is the rule the BIDI algorithm uses to determine the directionality of
2022          * paragraphs ( https://www.unicode.org/reports/tr9/#The_Paragraph_Level ) and
2023          * FSI isolates ( https://www.unicode.org/reports/tr9/#Explicit_Directional_Isolates ).
2024          *
2025          * TODO: Does not handle BIDI control characters inside the text.
2026          * TODO: Does not handle unallocated characters.
2027          *
2028          * @param string $text Text to test
2029          * @return null|string Directionality ('ltr' or 'rtl') or null
2030          */
2031         private static function strongDirFromContent( $text = '' ) {
2032                 if ( !preg_match( self::$strongDirRegex, $text, $matches ) ) {
2033                         return null;
2034                 }
2035                 if ( $matches[1] === '' ) {
2036                         return 'rtl';
2037                 }
2038                 return 'ltr';
2039         }
2040
2041         /**
2042          * Roman number formatting up to 10000
2043          *
2044          * @param int $num
2045          *
2046          * @return string
2047          */
2048         static function romanNumeral( $num ) {
2049                 static $table = [
2050                         [ '', 'I', 'II', 'III', 'IV', 'V', 'VI', 'VII', 'VIII', 'IX', 'X' ],
2051                         [ '', 'X', 'XX', 'XXX', 'XL', 'L', 'LX', 'LXX', 'LXXX', 'XC', 'C' ],
2052                         [ '', 'C', 'CC', 'CCC', 'CD', 'D', 'DC', 'DCC', 'DCCC', 'CM', 'M' ],
2053                         [ '', 'M', 'MM', 'MMM', 'MMMM', 'MMMMM', 'MMMMMM', 'MMMMMMM',
2054                                 'MMMMMMMM', 'MMMMMMMMM', 'MMMMMMMMMM' ]
2055                 ];
2056
2057                 $num = intval( $num );
2058                 if ( $num > 10000 || $num <= 0 ) {
2059                         return $num;
2060                 }
2061
2062                 $s = '';
2063                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
2064                         if ( $num >= $pow10 ) {
2065                                 $s .= $table[$i][(int)floor( $num / $pow10 )];
2066                         }
2067                         $num = $num % $pow10;
2068                 }
2069                 return $s;
2070         }
2071
2072         /**
2073          * Hebrew Gematria number formatting up to 9999
2074          *
2075          * @param int $num
2076          *
2077          * @return string
2078          */
2079         static function hebrewNumeral( $num ) {
2080                 static $table = [
2081                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ],
2082                         [ '', 'י', 'כ', 'ל', 'מ', 'נ', 'ס', 'ע', 'פ', 'צ', 'ק' ],
2083                         [ '',
2084                                 [ 'ק' ],
2085                                 [ 'ר' ],
2086                                 [ 'ש' ],
2087                                 [ 'ת' ],
2088                                 [ 'ת', 'ק' ],
2089                                 [ 'ת', 'ר' ],
2090                                 [ 'ת', 'ש' ],
2091                                 [ 'ת', 'ת' ],
2092                                 [ 'ת', 'ת', 'ק' ],
2093                                 [ 'ת', 'ת', 'ר' ],
2094                         ],
2095                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ]
2096                 ];
2097
2098                 $num = intval( $num );
2099                 if ( $num > 9999 || $num <= 0 ) {
2100                         return $num;
2101                 }
2102
2103                 // Round thousands have special notations
2104                 if ( $num === 1000 ) {
2105                         return "א' אלף";
2106                 } elseif ( $num % 1000 === 0 ) {
2107                         return $table[0][$num / 1000] . "' אלפים";
2108                 }
2109
2110                 $letters = [];
2111
2112                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
2113                         if ( $num >= $pow10 ) {
2114                                 if ( $num === 15 || $num === 16 ) {
2115                                         $letters[] = $table[0][9];
2116                                         $letters[] = $table[0][$num - 9];
2117                                         $num = 0;
2118                                 } else {
2119                                         $letters = array_merge(
2120                                                 $letters,
2121                                                 (array)$table[$i][intval( $num / $pow10 )]
2122                                         );
2123
2124                                         if ( $pow10 === 1000 ) {
2125                                                 $letters[] = "'";
2126                                         }
2127                                 }
2128                         }
2129
2130                         $num = $num % $pow10;
2131                 }
2132
2133                 $preTransformLength = count( $letters );
2134                 if ( $preTransformLength === 1 ) {
2135                         // Add geresh (single quote) to one-letter numbers
2136                         $letters[] = "'";
2137                 } else {
2138                         $lastIndex = $preTransformLength - 1;
2139                         $letters[$lastIndex] = str_replace(
2140                                 [ 'כ', 'מ', 'נ', 'פ', 'צ' ],
2141                                 [ 'ך', 'ם', 'ן', 'ף', 'ץ' ],
2142                                 $letters[$lastIndex]
2143                         );
2144
2145                         // Add gershayim (double quote) to multiple-letter numbers,
2146                         // but exclude numbers with only one letter after the thousands
2147                         // (1001-1009, 1020, 1030, 2001-2009, etc.)
2148                         if ( $letters[1] === "'" && $preTransformLength === 3 ) {
2149                                 $letters[] = "'";
2150                         } else {
2151                                 array_splice( $letters, -1, 0, '"' );
2152                         }
2153                 }
2154
2155                 return implode( $letters );
2156         }
2157
2158         /**
2159          * Used by date() and time() to adjust the time output.
2160          *
2161          * @param string $ts The time in date('YmdHis') format
2162          * @param mixed $tz Adjust the time by this amount (default false, mean we
2163          *   get user timecorrection setting)
2164          * @return int
2165          */
2166         public function userAdjust( $ts, $tz = false ) {
2167                 global $wgUser, $wgLocalTZoffset;
2168
2169                 if ( $tz === false ) {
2170                         $tz = $wgUser->getOption( 'timecorrection' );
2171                 }
2172
2173                 $data = explode( '|', $tz, 3 );
2174
2175                 if ( $data[0] == 'ZoneInfo' ) {
2176                         try {
2177                                 $userTZ = new DateTimeZone( $data[2] );
2178                                 $date = new DateTime( $ts, new DateTimeZone( 'UTC' ) );
2179                                 $date->setTimezone( $userTZ );
2180                                 return $date->format( 'YmdHis' );
2181                         } catch ( Exception $e ) {
2182                                 // Unrecognized timezone, default to 'Offset' with the stored offset.
2183                                 $data[0] = 'Offset';
2184                         }
2185                 }
2186
2187                 if ( $data[0] == 'System' || $tz == '' ) {
2188                         # Global offset in minutes.
2189                         $minDiff = $wgLocalTZoffset;
2190                 } elseif ( $data[0] == 'Offset' ) {
2191                         $minDiff = intval( $data[1] );
2192                 } else {
2193                         $data = explode( ':', $tz );
2194                         if ( count( $data ) == 2 ) {
2195                                 $data[0] = intval( $data[0] );
2196                                 $data[1] = intval( $data[1] );
2197                                 $minDiff = abs( $data[0] ) * 60 + $data[1];
2198                                 if ( $data[0] < 0 ) {
2199                                         $minDiff = -$minDiff;
2200                                 }
2201                         } else {
2202                                 $minDiff = intval( $data[0] ) * 60;
2203                         }
2204                 }
2205
2206                 # No difference ? Return time unchanged
2207                 if ( $minDiff == 0 ) {
2208                         return $ts;
2209                 }
2210
2211                 Wikimedia\suppressWarnings(); // E_STRICT system time bitching
2212                 # Generate an adjusted date; take advantage of the fact that mktime
2213                 # will normalize out-of-range values so we don't have to split $minDiff
2214                 # into hours and minutes.
2215                 $t = mktime( (
2216                         (int)substr( $ts, 8, 2 ) ), # Hours
2217                         (int)substr( $ts, 10, 2 ) + $minDiff, # Minutes
2218                         (int)substr( $ts, 12, 2 ), # Seconds
2219                         (int)substr( $ts, 4, 2 ), # Month
2220                         (int)substr( $ts, 6, 2 ), # Day
2221                         (int)substr( $ts, 0, 4 ) ); # Year
2222
2223                 $date = date( 'YmdHis', $t );
2224                 Wikimedia\restoreWarnings();
2225
2226                 return $date;
2227         }
2228
2229         /**
2230          * This is meant to be used by time(), date(), and timeanddate() to get
2231          * the date preference they're supposed to use, it should be used in
2232          * all children.
2233          *
2234          *     function timeanddate([...], $format = true) {
2235          *       $datePreference = $this->dateFormat($format);
2236          *       [...]
2237          *     }
2238          *
2239          * @param int|string|bool $usePrefs If true, the user's preference is used
2240          *   if false, the site/language default is used
2241          *   if int/string, assumed to be a format.
2242          * @return string
2243          */
2244         function dateFormat( $usePrefs = true ) {
2245                 global $wgUser;
2246
2247                 if ( is_bool( $usePrefs ) ) {
2248                         if ( $usePrefs ) {
2249                                 $datePreference = $wgUser->getDatePreference();
2250                         } else {
2251                                 $datePreference = (string)User::getDefaultOption( 'date' );
2252                         }
2253                 } else {
2254                         $datePreference = (string)$usePrefs;
2255                 }
2256
2257                 // return int
2258                 if ( $datePreference == '' ) {
2259                         return 'default';
2260                 }
2261
2262                 return $datePreference;
2263         }
2264
2265         /**
2266          * Get a format string for a given type and preference
2267          * @param string $type May be 'date', 'time', 'both', or 'pretty'.
2268          * @param string $pref The format name as it appears in Messages*.php under
2269          *  $datePreferences.
2270          *
2271          * @since 1.22 New type 'pretty' that provides a more readable timestamp format
2272          *
2273          * @return string
2274          */
2275         function getDateFormatString( $type, $pref ) {
2276                 $wasDefault = false;
2277                 if ( $pref == 'default' ) {
2278                         $wasDefault = true;
2279                         $pref = $this->getDefaultDateFormat();
2280                 }
2281
2282                 if ( !isset( $this->dateFormatStrings[$type][$pref] ) ) {
2283                         $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2284
2285                         if ( $type === 'pretty' && $df === null ) {
2286                                 $df = $this->getDateFormatString( 'date', $pref );
2287                         }
2288
2289                         if ( !$wasDefault && $df === null ) {
2290                                 $pref = $this->getDefaultDateFormat();
2291                                 $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2292                         }
2293
2294                         $this->dateFormatStrings[$type][$pref] = $df;
2295                 }
2296                 return $this->dateFormatStrings[$type][$pref];
2297         }
2298
2299         /**
2300          * @param string $ts The time format which needs to be turned into a
2301          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2302          * @param bool $adj Whether to adjust the time output according to the
2303          *   user configured offset ($timecorrection)
2304          * @param mixed $format True to use user's date format preference
2305          * @param string|bool $timecorrection The time offset as returned by
2306          *   validateTimeZone() in Special:Preferences
2307          * @return string
2308          */
2309         public function date( $ts, $adj = false, $format = true, $timecorrection = false ) {
2310                 $ts = wfTimestamp( TS_MW, $ts );
2311                 if ( $adj ) {
2312                         $ts = $this->userAdjust( $ts, $timecorrection );
2313                 }
2314                 $df = $this->getDateFormatString( 'date', $this->dateFormat( $format ) );
2315                 return $this->sprintfDate( $df, $ts );
2316         }
2317
2318         /**
2319          * @param string $ts The time format which needs to be turned into a
2320          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2321          * @param bool $adj Whether to adjust the time output according to the
2322          *   user configured offset ($timecorrection)
2323          * @param mixed $format True to use user's date format preference
2324          * @param string|bool $timecorrection The time offset as returned by
2325          *   validateTimeZone() in Special:Preferences
2326          * @return string
2327          */
2328         public function time( $ts, $adj = false, $format = true, $timecorrection = false ) {
2329                 $ts = wfTimestamp( TS_MW, $ts );
2330                 if ( $adj ) {
2331                         $ts = $this->userAdjust( $ts, $timecorrection );
2332                 }
2333                 $df = $this->getDateFormatString( 'time', $this->dateFormat( $format ) );
2334                 return $this->sprintfDate( $df, $ts );
2335         }
2336
2337         /**
2338          * @param string $ts The time format which needs to be turned into a
2339          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2340          * @param bool $adj Whether to adjust the time output according to the
2341          *   user configured offset ($timecorrection)
2342          * @param mixed $format What format to return, if it's false output the
2343          *   default one (default true)
2344          * @param string|bool $timecorrection The time offset as returned by
2345          *   validateTimeZone() in Special:Preferences
2346          * @return string
2347          */
2348         public function timeanddate( $ts, $adj = false, $format = true, $timecorrection = false ) {
2349                 $ts = wfTimestamp( TS_MW, $ts );
2350                 if ( $adj ) {
2351                         $ts = $this->userAdjust( $ts, $timecorrection );
2352                 }
2353                 $df = $this->getDateFormatString( 'both', $this->dateFormat( $format ) );
2354                 return $this->sprintfDate( $df, $ts );
2355         }
2356
2357         /**
2358          * Takes a number of seconds and turns it into a text using values such as hours and minutes.
2359          *
2360          * @since 1.20
2361          *
2362          * @param int $seconds The amount of seconds.
2363          * @param array $chosenIntervals The intervals to enable.
2364          *
2365          * @return string
2366          */
2367         public function formatDuration( $seconds, array $chosenIntervals = [] ) {
2368                 $intervals = $this->getDurationIntervals( $seconds, $chosenIntervals );
2369
2370                 $segments = [];
2371
2372                 foreach ( $intervals as $intervalName => $intervalValue ) {
2373                         // Messages: duration-seconds, duration-minutes, duration-hours, duration-days, duration-weeks,
2374                         // duration-years, duration-decades, duration-centuries, duration-millennia
2375                         $message = wfMessage( 'duration-' . $intervalName )->numParams( $intervalValue );
2376                         $segments[] = $message->inLanguage( $this )->escaped();
2377                 }
2378
2379                 return $this->listToText( $segments );
2380         }
2381
2382         /**
2383          * Takes a number of seconds and returns an array with a set of corresponding intervals.
2384          * For example 65 will be turned into [ minutes => 1, seconds => 5 ].
2385          *
2386          * @since 1.20
2387          *
2388          * @param int $seconds The amount of seconds.
2389          * @param array $chosenIntervals The intervals to enable.
2390          *
2391          * @return array
2392          */
2393         public function getDurationIntervals( $seconds, array $chosenIntervals = [] ) {
2394                 if ( empty( $chosenIntervals ) ) {
2395                         $chosenIntervals = [
2396                                 'millennia',
2397                                 'centuries',
2398                                 'decades',
2399                                 'years',
2400                                 'days',
2401                                 'hours',
2402                                 'minutes',
2403                                 'seconds'
2404                         ];
2405                 }
2406
2407                 $intervals = array_intersect_key( self::$durationIntervals, array_flip( $chosenIntervals ) );
2408                 $sortedNames = array_keys( $intervals );
2409                 $smallestInterval = array_pop( $sortedNames );
2410
2411                 $segments = [];
2412
2413                 foreach ( $intervals as $name => $length ) {
2414                         $value = floor( $seconds / $length );
2415
2416                         if ( $value > 0 || ( $name == $smallestInterval && empty( $segments ) ) ) {
2417                                 $seconds -= $value * $length;
2418                                 $segments[$name] = $value;
2419                         }
2420                 }
2421
2422                 return $segments;
2423         }
2424
2425         /**
2426          * Internal helper function for userDate(), userTime() and userTimeAndDate()
2427          *
2428          * @param string $type Can be 'date', 'time' or 'both'
2429          * @param string $ts The time format which needs to be turned into a
2430          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2431          * @param User $user User object used to get preferences for timezone and format
2432          * @param array $options Array, can contain the following keys:
2433          *   - 'timecorrection': time correction, can have the following values:
2434          *     - true: use user's preference
2435          *     - false: don't use time correction
2436          *     - int: value of time correction in minutes
2437          *   - 'format': format to use, can have the following values:
2438          *     - true: use user's preference
2439          *     - false: use default preference
2440          *     - string: format to use
2441          * @since 1.19
2442          * @return string
2443          */
2444         private function internalUserTimeAndDate( $type, $ts, User $user, array $options ) {
2445                 $ts = wfTimestamp( TS_MW, $ts );
2446                 $options += [ 'timecorrection' => true, 'format' => true ];
2447                 if ( $options['timecorrection'] !== false ) {
2448                         if ( $options['timecorrection'] === true ) {
2449                                 $offset = $user->getOption( 'timecorrection' );
2450                         } else {
2451                                 $offset = $options['timecorrection'];
2452                         }
2453                         $ts = $this->userAdjust( $ts, $offset );
2454                 }
2455                 if ( $options['format'] === true ) {
2456                         $format = $user->getDatePreference();
2457                 } else {
2458                         $format = $options['format'];
2459                 }
2460                 $df = $this->getDateFormatString( $type, $this->dateFormat( $format ) );
2461                 return $this->sprintfDate( $df, $ts );
2462         }
2463
2464         /**
2465          * Get the formatted date for the given timestamp and formatted for
2466          * the given user.
2467          *
2468          * @param mixed $ts Mixed: the time format which needs to be turned into a
2469          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2470          * @param User $user User object used to get preferences for timezone and format
2471          * @param array $options Array, can contain the following keys:
2472          *   - 'timecorrection': time correction, can have the following values:
2473          *     - true: use user's preference
2474          *     - false: don't use time correction
2475          *     - int: value of time correction in minutes
2476          *   - 'format': format to use, can have the following values:
2477          *     - true: use user's preference
2478          *     - false: use default preference
2479          *     - string: format to use
2480          * @since 1.19
2481          * @return string
2482          */
2483         public function userDate( $ts, User $user, array $options = [] ) {
2484                 return $this->internalUserTimeAndDate( 'date', $ts, $user, $options );
2485         }
2486
2487         /**
2488          * Get the formatted time for the given timestamp and formatted for
2489          * the given user.
2490          *
2491          * @param mixed $ts The time format which needs to be turned into a
2492          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2493          * @param User $user User object used to get preferences for timezone and format
2494          * @param array $options Array, can contain the following keys:
2495          *   - 'timecorrection': time correction, can have the following values:
2496          *     - true: use user's preference
2497          *     - false: don't use time correction
2498          *     - int: value of time correction in minutes
2499          *   - 'format': format to use, can have the following values:
2500          *     - true: use user's preference
2501          *     - false: use default preference
2502          *     - string: format to use
2503          * @since 1.19
2504          * @return string
2505          */
2506         public function userTime( $ts, User $user, array $options = [] ) {
2507                 return $this->internalUserTimeAndDate( 'time', $ts, $user, $options );
2508         }
2509
2510         /**
2511          * Get the formatted date and time for the given timestamp and formatted for
2512          * the given user.
2513          *
2514          * @param mixed $ts The time format which needs to be turned into a
2515          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2516          * @param User $user User object used to get preferences for timezone and format
2517          * @param array $options Array, can contain the following keys:
2518          *   - 'timecorrection': time correction, can have the following values:
2519          *     - true: use user's preference
2520          *     - false: don't use time correction
2521          *     - int: value of time correction in minutes
2522          *   - 'format': format to use, can have the following values:
2523          *     - true: use user's preference
2524          *     - false: use default preference
2525          *     - string: format to use
2526          * @since 1.19
2527          * @return string
2528          */
2529         public function userTimeAndDate( $ts, User $user, array $options = [] ) {
2530                 return $this->internalUserTimeAndDate( 'both', $ts, $user, $options );
2531         }
2532
2533         /**
2534          * Get the timestamp in a human-friendly relative format, e.g., "3 days ago".
2535          *
2536          * Determine the difference between the timestamp and the current time, and
2537          * generate a readable timestamp by returning "<N> <units> ago", where the
2538          * largest possible unit is used.
2539          *
2540          * @since 1.26 (Prior to 1.26 method existed but was not meant to be used directly)
2541          *
2542          * @param MWTimestamp $time
2543          * @param MWTimestamp|null $relativeTo The base timestamp to compare to (defaults to now)
2544          * @param User|null $user User the timestamp is being generated for
2545          *  (or null to use main context's user)
2546          * @return string Formatted timestamp
2547          */
2548         public function getHumanTimestamp(
2549                 MWTimestamp $time, MWTimestamp $relativeTo = null, User $user = null
2550         ) {
2551                 if ( $relativeTo === null ) {
2552                         $relativeTo = new MWTimestamp();
2553                 }
2554                 if ( $user === null ) {
2555                         $user = RequestContext::getMain()->getUser();
2556                 }
2557
2558                 // Adjust for the user's timezone.
2559                 $offsetThis = $time->offsetForUser( $user );
2560                 $offsetRel = $relativeTo->offsetForUser( $user );
2561
2562                 $ts = '';
2563                 if ( Hooks::run( 'GetHumanTimestamp', [ &$ts, $time, $relativeTo, $user, $this ] ) ) {
2564                         $ts = $this->getHumanTimestampInternal( $time, $relativeTo, $user );
2565                 }
2566
2567                 // Reset the timezone on the objects.
2568                 $time->timestamp->sub( $offsetThis );
2569                 $relativeTo->timestamp->sub( $offsetRel );
2570
2571                 return $ts;
2572         }
2573
2574         /**
2575          * Convert an MWTimestamp into a pretty human-readable timestamp using
2576          * the given user preferences and relative base time.
2577          *
2578          * @see Language::getHumanTimestamp
2579          * @param MWTimestamp $ts Timestamp to prettify
2580          * @param MWTimestamp $relativeTo Base timestamp
2581          * @param User $user User preferences to use
2582          * @return string Human timestamp
2583          * @since 1.26
2584          */
2585         private function getHumanTimestampInternal(
2586                 MWTimestamp $ts, MWTimestamp $relativeTo, User $user
2587         ) {
2588                 $diff = $ts->diff( $relativeTo );
2589                 $diffDay = (bool)( (int)$ts->timestamp->format( 'w' ) -
2590                         (int)$relativeTo->timestamp->format( 'w' ) );
2591                 $days = $diff->days ?: (int)$diffDay;
2592                 if ( $diff->invert || $days > 5
2593                         && $ts->timestamp->format( 'Y' ) !== $relativeTo->timestamp->format( 'Y' )
2594                 ) {
2595                         // Timestamps are in different years: use full timestamp
2596                         // Also do full timestamp for future dates
2597                         /**
2598                          * @todo FIXME: Add better handling of future timestamps.
2599                          */
2600                         $format = $this->getDateFormatString( 'both', $user->getDatePreference() ?: 'default' );
2601                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2602                 } elseif ( $days > 5 ) {
2603                         // Timestamps are in same year,  but more than 5 days ago: show day and month only.
2604                         $format = $this->getDateFormatString( 'pretty', $user->getDatePreference() ?: 'default' );
2605                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2606                 } elseif ( $days > 1 ) {
2607                         // Timestamp within the past week: show the day of the week and time
2608                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2609                         $weekday = self::$mWeekdayMsgs[$ts->timestamp->format( 'w' )];
2610                         // Messages:
2611                         // sunday-at, monday-at, tuesday-at, wednesday-at, thursday-at, friday-at, saturday-at
2612                         $ts = wfMessage( "$weekday-at" )
2613                                 ->inLanguage( $this )
2614                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2615                                 ->text();
2616                 } elseif ( $days == 1 ) {
2617                         // Timestamp was yesterday: say 'yesterday' and the time.
2618                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2619                         $ts = wfMessage( 'yesterday-at' )
2620                                 ->inLanguage( $this )
2621                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2622                                 ->text();
2623                 } elseif ( $diff->h > 1 || $diff->h == 1 && $diff->i > 30 ) {
2624                         // Timestamp was today, but more than 90 minutes ago: say 'today' and the time.
2625                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2626                         $ts = wfMessage( 'today-at' )
2627                                 ->inLanguage( $this )
2628                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2629                                 ->text();
2630
2631                 // From here on in, the timestamp was soon enough ago so that we can simply say
2632                 // XX units ago, e.g., "2 hours ago" or "5 minutes ago"
2633                 } elseif ( $diff->h == 1 ) {
2634                         // Less than 90 minutes, but more than an hour ago.
2635                         $ts = wfMessage( 'hours-ago' )->inLanguage( $this )->numParams( 1 )->text();
2636                 } elseif ( $diff->i >= 1 ) {
2637                         // A few minutes ago.
2638                         $ts = wfMessage( 'minutes-ago' )->inLanguage( $this )->numParams( $diff->i )->text();
2639                 } elseif ( $diff->s >= 30 ) {
2640                         // Less than a minute, but more than 30 sec ago.
2641                         $ts = wfMessage( 'seconds-ago' )->inLanguage( $this )->numParams( $diff->s )->text();
2642                 } else {
2643                         // Less than 30 seconds ago.
2644                         $ts = wfMessage( 'just-now' )->text();
2645                 }
2646
2647                 return $ts;
2648         }
2649
2650         /**
2651          * @param string $key
2652          * @return string|null
2653          */
2654         public function getMessage( $key ) {
2655                 return self::$dataCache->getSubitem( $this->mCode, 'messages', $key );
2656         }
2657
2658         /**
2659          * @return array
2660          */
2661         function getAllMessages() {
2662                 return self::$dataCache->getItem( $this->mCode, 'messages' );
2663         }
2664
2665         /**
2666          * @param string $in
2667          * @param string $out
2668          * @param string $string
2669          * @return string
2670          */
2671         public function iconv( $in, $out, $string ) {
2672                 # Even with //IGNORE iconv can whine about illegal characters in
2673                 # *input* string. We just ignore those too.
2674                 # REF: https://bugs.php.net/bug.php?id=37166
2675                 # REF: https://phabricator.wikimedia.org/T18885
2676                 Wikimedia\suppressWarnings();
2677                 $text = iconv( $in, $out . '//IGNORE', $string );
2678                 Wikimedia\restoreWarnings();
2679                 return $text;
2680         }
2681
2682         // callback functions for ucwords(), ucwordbreaks()
2683
2684         /**
2685          * @param array $matches
2686          * @return mixed|string
2687          */
2688         function ucwordbreaksCallbackAscii( $matches ) {
2689                 return $this->ucfirst( $matches[1] );
2690         }
2691
2692         /**
2693          * @param array $matches
2694          * @return string
2695          */
2696         function ucwordbreaksCallbackMB( $matches ) {
2697                 return mb_strtoupper( $matches[0] );
2698         }
2699
2700         /**
2701          * @param array $matches
2702          * @return string
2703          */
2704         function ucwordsCallbackMB( $matches ) {
2705                 return mb_strtoupper( $matches[0] );
2706         }
2707
2708         /**
2709          * Make a string's first character uppercase
2710          *
2711          * @param string $str
2712          *
2713          * @return string
2714          */
2715         public function ucfirst( $str ) {
2716                 $o = ord( $str );
2717                 if ( $o < 96 ) { // if already uppercase...
2718                         return $str;
2719                 } elseif ( $o < 128 ) {
2720                         return ucfirst( $str ); // use PHP's ucfirst()
2721                 } else {
2722                         // fall back to more complex logic in case of multibyte strings
2723                         return $this->uc( $str, true );
2724                 }
2725         }
2726
2727         /**
2728          * Convert a string to uppercase
2729          *
2730          * @param string $str
2731          * @param bool $first
2732          *
2733          * @return string
2734          */
2735         public function uc( $str, $first = false ) {
2736                 if ( $first ) {
2737                         if ( $this->isMultibyte( $str ) ) {
2738                                 return $this->mbUpperChar( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2739                         } else {
2740                                 return ucfirst( $str );
2741                         }
2742                 } else {
2743                         return $this->isMultibyte( $str ) ? mb_strtoupper( $str ) : strtoupper( $str );
2744                 }
2745         }
2746
2747         /**
2748          * Convert character to uppercase, allowing overrides of the default mb_upper
2749          * behaviour, which is buggy in many ways. Having a conversion table can be
2750          * useful during transitions between PHP versions where unicode changes happen.
2751          * This can make some resources unreachable on-wiki, see discussion at T219279.
2752          * Providing such a conversion table can allow to manage the transition period.
2753          *
2754          * @since 1.34
2755          *
2756          * @param string $char
2757          *
2758          * @return string
2759          */
2760         protected function mbUpperChar( $char ) {
2761                 global $wgOverrideUcfirstCharacters;
2762                 if ( array_key_exists( $char, $wgOverrideUcfirstCharacters ) ) {
2763                         return $wgOverrideUcfirstCharacters[$char];
2764                 } else {
2765                         return mb_strtoupper( $char );
2766                 }
2767         }
2768
2769         /**
2770          * @param string $str
2771          * @return mixed|string
2772          */
2773         function lcfirst( $str ) {
2774                 $o = ord( $str );
2775                 if ( !$o ) {
2776                         return strval( $str );
2777                 } elseif ( $o >= 128 ) {
2778                         return $this->lc( $str, true );
2779                 } elseif ( $o > 96 ) {
2780                         return $str;
2781                 } else {
2782                         $str[0] = strtolower( $str[0] );
2783                         return $str;
2784                 }
2785         }
2786
2787         /**
2788          * @param string $str
2789          * @param bool $first
2790          * @return mixed|string
2791          */
2792         function lc( $str, $first = false ) {
2793                 if ( $first ) {
2794                         if ( $this->isMultibyte( $str ) ) {
2795                                 return mb_strtolower( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2796                         } else {
2797                                 return strtolower( substr( $str, 0, 1 ) ) . substr( $str, 1 );
2798                         }
2799                 } else {
2800                         return $this->isMultibyte( $str ) ? mb_strtolower( $str ) : strtolower( $str );
2801                 }
2802         }
2803
2804         /**
2805          * @param string $str
2806          * @return bool
2807          */
2808         function isMultibyte( $str ) {
2809                 return strlen( $str ) !== mb_strlen( $str );
2810         }
2811
2812         /**
2813          * @param string $str
2814          * @return mixed|string
2815          */
2816         function ucwords( $str ) {
2817                 if ( $this->isMultibyte( $str ) ) {
2818                         $str = $this->lc( $str );
2819
2820                         // regexp to find first letter in each word (i.e. after each space)
2821                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)| ([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2822
2823                         // function to use to capitalize a single char
2824                         return preg_replace_callback(
2825                                 $replaceRegexp,
2826                                 [ $this, 'ucwordsCallbackMB' ],
2827                                 $str
2828                         );
2829                 } else {
2830                         return ucwords( strtolower( $str ) );
2831                 }
2832         }
2833
2834         /**
2835          * capitalize words at word breaks
2836          *
2837          * @param string $str
2838          * @return mixed
2839          */
2840         function ucwordbreaks( $str ) {
2841                 if ( $this->isMultibyte( $str ) ) {
2842                         $str = $this->lc( $str );
2843
2844                         // since \b doesn't work for UTF-8, we explicitely define word break chars
2845                         $breaks = "[ \-\(\)\}\{\.,\?!]";
2846
2847                         // find first letter after word break
2848                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)|" .
2849                                 "$breaks([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2850
2851                         return preg_replace_callback(
2852                                 $replaceRegexp,
2853                                 [ $this, 'ucwordbreaksCallbackMB' ],
2854                                 $str
2855                         );
2856                 } else {
2857                         return preg_replace_callback(
2858                                 '/\b([\w\x80-\xff]+)\b/',
2859                                 [ $this, 'ucwordbreaksCallbackAscii' ],
2860                                 $str
2861                         );
2862                 }
2863         }
2864
2865         /**
2866          * Return a case-folded representation of $s
2867          *
2868          * This is a representation such that caseFold($s1)==caseFold($s2) if $s1
2869          * and $s2 are the same except for the case of their characters. It is not
2870          * necessary for the value returned to make sense when displayed.
2871          *
2872          * Do *not* perform any other normalisation in this function. If a caller
2873          * uses this function when it should be using a more general normalisation
2874          * function, then fix the caller.
2875          *
2876          * @param string $s
2877          *
2878          * @return string
2879          */
2880         function caseFold( $s ) {
2881                 return $this->uc( $s );
2882         }
2883
2884         /**
2885          * TODO: $s is not always a string per T218883
2886          * @param string $s
2887          * @return string
2888          */
2889         function checkTitleEncoding( $s ) {
2890                 if ( is_array( $s ) ) {
2891                         throw new MWException( 'Given array to checkTitleEncoding.' );
2892                 }
2893                 if ( StringUtils::isUtf8( $s ) ) {
2894                         return $s;
2895                 }
2896
2897                 return $this->iconv( $this->fallback8bitEncoding(), 'utf-8', $s );
2898         }
2899
2900         /**
2901          * @return string
2902          */
2903         function fallback8bitEncoding() {
2904                 return self::$dataCache->getItem( $this->mCode, 'fallback8bitEncoding' );
2905         }
2906
2907         /**
2908          * Most writing systems use whitespace to break up words.
2909          * Some languages such as Chinese don't conventionally do this,
2910          * which requires special handling when breaking up words for
2911          * searching etc.
2912          *
2913          * @return bool
2914          */
2915         function hasWordBreaks() {
2916                 return true;
2917         }
2918
2919         /**
2920          * Some languages such as Chinese require word segmentation,
2921          * Specify such segmentation when overridden in derived class.
2922          *
2923          * @param string $string
2924          * @return string
2925          */
2926         function segmentByWord( $string ) {
2927                 return $string;
2928         }
2929
2930         /**
2931          * Some languages have special punctuation need to be normalized.
2932          * Make such changes here.
2933          *
2934          * @param string $string
2935          * @return string
2936          */
2937         function normalizeForSearch( $string ) {
2938                 return self::convertDoubleWidth( $string );
2939         }
2940
2941         /**
2942          * convert double-width roman characters to single-width.
2943          * range: ff00-ff5f ~= 0020-007f
2944          *
2945          * @param string $string
2946          *
2947          * @return string
2948          */
2949         protected static function convertDoubleWidth( $string ) {
2950                 static $full = null;
2951                 static $half = null;
2952
2953                 if ( $full === null ) {
2954                         $fullWidth = "０１２３４５６７８９ＡＢＣＤＥＦＧＨＩＪＫＬＭＮＯＰＱＲＳＴＵＶＷＸＹＺａｂｃｄｅｆｇｈｉｊｋｌｍｎｏｐｑｒｓｔｕｖｗｘｙｚ";
2955                         $halfWidth = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz";
2956                         $full = str_split( $fullWidth, 3 );
2957                         $half = str_split( $halfWidth );
2958                 }
2959
2960                 $string = str_replace( $full, $half, $string );
2961                 return $string;
2962         }
2963
2964         /**
2965          * @param string $string
2966          * @param string $pattern
2967          * @return string
2968          */
2969         protected static function insertSpace( $string, $pattern ) {
2970                 $string = preg_replace( $pattern, " $1 ", $string );
2971                 $string = preg_replace( '/ +/', ' ', $string );
2972                 return $string;
2973         }
2974
2975         /**
2976          * @param string[] $termsArray
2977          * @return string[]
2978          */
2979         function convertForSearchResult( $termsArray ) {
2980                 # some languages, e.g. Chinese, need to do a conversion
2981                 # in order for search results to be displayed correctly
2982                 return $termsArray;
2983         }
2984
2985         /**
2986          * Get the first character of a string.
2987          *
2988          * @param string $s
2989          * @return string
2990          */
2991         function firstChar( $s ) {
2992                 $matches = [];
2993                 preg_match(
2994                         '/^([\x00-\x7f]|[\xc0-\xdf][\x80-\xbf]|' .
2995                                 '[\xe0-\xef][\x80-\xbf]{2}|[\xf0-\xf7][\x80-\xbf]{3})/',
2996                         $s,
2997                         $matches
2998                 );
2999
3000                 if ( isset( $matches[1] ) ) {
3001                         if ( strlen( $matches[1] ) != 3 ) {
3002                                 return $matches[1];
3003                         }
3004
3005                         // Break down Hangul syllables to grab the first jamo
3006                         $code = UtfNormal\Utils::utf8ToCodepoint( $matches[1] );
3007                         if ( $code < 0xac00 || 0xd7a4 <= $code ) {
3008                                 return $matches[1];
3009                         } elseif ( $code < 0xb098 ) {
3010                                 return "\u{3131}";
3011                         } elseif ( $code < 0xb2e4 ) {
3012                                 return "\u{3134}";
3013                         } elseif ( $code < 0xb77c ) {
3014                                 return "\u{3137}";
3015                         } elseif ( $code < 0xb9c8 ) {
3016                                 return "\u{3139}";
3017                         } elseif ( $code < 0xbc14 ) {
3018                                 return "\u{3141}";
3019                         } elseif ( $code < 0xc0ac ) {
3020                                 return "\u{3142}";
3021                         } elseif ( $code < 0xc544 ) {
3022                                 return "\u{3145}";
3023                         } elseif ( $code < 0xc790 ) {
3024                                 return "\u{3147}";
3025                         } elseif ( $code < 0xcc28 ) {
3026                                 return "\u{3148}";
3027                         } elseif ( $code < 0xce74 ) {
3028                                 return "\u{314A}";
3029                         } elseif ( $code < 0xd0c0 ) {
3030                                 return "\u{314B}";
3031                         } elseif ( $code < 0xd30c ) {
3032                                 return "\u{314C}";
3033                         } elseif ( $code < 0xd558 ) {
3034                                 return "\u{314D}";
3035                         } else {
3036                                 return "\u{314E}";
3037                         }
3038                 } else {
3039                         return '';
3040                 }
3041         }
3042
3043         /**
3044          * Convert a UTF-8 string to normal form C. In Malayalam and Arabic, this
3045          * also cleans up certain backwards-compatible sequences, converting them
3046          * to the modern Unicode equivalent.
3047          *
3048          * This is language-specific for performance reasons only.
3049          *
3050          * @param string $s
3051          *
3052          * @return string
3053          */
3054         public function normalize( $s ) {
3055                 global $wgAllUnicodeFixes;
3056                 $s = UtfNormal\Validator::cleanUp( $s );
3057                 if ( $wgAllUnicodeFixes ) {
3058                         $s = $this->transformUsingPairFile( 'normalize-ar.php', $s );
3059                         $s = $this->transformUsingPairFile( 'normalize-ml.php', $s );
3060                 }
3061
3062                 return $s;
3063         }
3064
3065         /**
3066          * Transform a string using serialized data stored in the given file (which
3067          * must be in the serialized subdirectory of $IP). The file contains pairs
3068          * mapping source characters to destination characters.
3069          *
3070          * The data is cached in process memory. This will go faster if you have the
3071          * FastStringSearch extension.
3072          *
3073          * @param string $file
3074          * @param string $string
3075          *
3076          * @throws MWException
3077          * @return string
3078          */
3079         protected function transformUsingPairFile( $file, $string ) {
3080                 if ( !isset( $this->transformData[$file] ) ) {
3081                         global $IP;
3082                         $data = require "$IP/languages/data/{$file}";
3083                         $this->transformData[$file] = new ReplacementArray( $data );
3084                 }
3085                 return $this->transformData[$file]->replace( $string );
3086         }
3087
3088         /**
3089          * For right-to-left language support
3090          *
3091          * @return bool
3092          */
3093         function isRTL() {
3094                 return self::$dataCache->getItem( $this->mCode, 'rtl' );
3095         }
3096
3097         /**
3098          * Return the correct HTML 'dir' attribute value for this language.
3099          * @return string
3100          */
3101         function getDir() {
3102                 return $this->isRTL() ? 'rtl' : 'ltr';
3103         }
3104
3105         /**
3106          * Return 'left' or 'right' as appropriate alignment for line-start
3107          * for this language's text direction.
3108          *
3109          * Should be equivalent to CSS3 'start' text-align value....
3110          *
3111          * @return string
3112          */
3113         function alignStart() {
3114                 return $this->isRTL() ? 'right' : 'left';
3115         }
3116
3117         /**
3118          * Return 'right' or 'left' as appropriate alignment for line-end
3119          * for this language's text direction.
3120          *
3121          * Should be equivalent to CSS3 'end' text-align value....
3122          *
3123          * @return string
3124          */
3125         function alignEnd() {
3126                 return $this->isRTL() ? 'left' : 'right';
3127         }
3128
3129         /**
3130          * A hidden direction mark (LRM or RLM), depending on the language direction.
3131          * Unlike getDirMark(), this function returns the character as an HTML entity.
3132          * This function should be used when the output is guaranteed to be HTML,
3133          * because it makes the output HTML source code more readable. When
3134          * the output is plain text or can be escaped, getDirMark() should be used.
3135          *
3136          * @param bool $opposite Get the direction mark opposite to your language
3137          * @return string
3138          * @since 1.20
3139          */
3140         function getDirMarkEntity( $opposite = false ) {
3141                 if ( $opposite ) {
3142                         return $this->isRTL() ? '&lrm;' : '&rlm;';
3143                 }
3144                 return $this->isRTL() ? '&rlm;' : '&lrm;';
3145         }
3146
3147         /**
3148          * A hidden direction mark (LRM or RLM), depending on the language direction.
3149          * This function produces them as invisible Unicode characters and
3150          * the output may be hard to read and debug, so it should only be used
3151          * when the output is plain text or can be escaped. When the output is
3152          * HTML, use getDirMarkEntity() instead.
3153          *
3154          * @param bool $opposite Get the direction mark opposite to your language
3155          * @return string
3156          */
3157         function getDirMark( $opposite = false ) {
3158                 $lrm = "\u{200E}"; # LEFT-TO-RIGHT MARK, commonly abbreviated LRM
3159                 $rlm = "\u{200F}"; # RIGHT-TO-LEFT MARK, commonly abbreviated RLM
3160                 if ( $opposite ) {
3161                         return $this->isRTL() ? $lrm : $rlm;
3162                 }
3163                 return $this->isRTL() ? $rlm : $lrm;
3164         }
3165
3166         /**
3167          * @return array
3168          */
3169         function capitalizeAllNouns() {
3170                 return self::$dataCache->getItem( $this->mCode, 'capitalizeAllNouns' );
3171         }
3172
3173         /**
3174          * An arrow, depending on the language direction.
3175          *
3176          * @param string $direction The direction of the arrow: forwards (default),
3177          *   backwards, left, right, up, down.
3178          * @return string
3179          */
3180         function getArrow( $direction = 'forwards' ) {
3181                 switch ( $direction ) {
3182                         case 'forwards':
3183                                 return $this->isRTL() ? '←' : '→';
3184                         case 'backwards':
3185                                 return $this->isRTL() ? '→' : '←';
3186                         case 'left':
3187                                 return '←';
3188                         case 'right':
3189                                 return '→';
3190                         case 'up':
3191                                 return '↑';
3192                         case 'down':
3193                                 return '↓';
3194                 }
3195         }
3196
3197         /**
3198          * To allow "foo[[bar]]" to extend the link over the whole word "foobar"
3199          *
3200          * @return bool
3201          */
3202         function linkPrefixExtension() {
3203                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixExtension' );
3204         }
3205
3206         /**
3207          * Get all magic words from cache.
3208          * @return array
3209          */
3210         function getMagicWords() {
3211                 return self::$dataCache->getItem( $this->mCode, 'magicWords' );
3212         }
3213
3214         /**
3215          * Fill a MagicWord object with data from here
3216          *
3217          * @param MagicWord $mw
3218          */
3219         function getMagic( $mw ) {
3220                 $rawEntry = $this->mMagicExtensions[$mw->mId] ??
3221                         self::$dataCache->getSubitem( $this->mCode, 'magicWords', $mw->mId );
3222
3223                 if ( !is_array( $rawEntry ) ) {
3224                         wfWarn( "\"$rawEntry\" is not a valid magic word for \"$mw->mId\"" );
3225                 } else {
3226                         $mw->mCaseSensitive = $rawEntry[0];
3227                         $mw->mSynonyms = array_slice( $rawEntry, 1 );
3228                 }
3229         }
3230
3231         /**
3232          * Add magic words to the extension array
3233          *
3234          * @param array $newWords
3235          */
3236         function addMagicWordsByLang( $newWords ) {
3237                 $fallbackChain = $this->getFallbackLanguages();
3238                 $fallbackChain = array_reverse( $fallbackChain );
3239                 foreach ( $fallbackChain as $code ) {
3240                         if ( isset( $newWords[$code] ) ) {
3241                                 // @phan-suppress-next-line PhanTypeMismatchProperty
3242                                 $this->mMagicExtensions = $newWords[$code] + $this->mMagicExtensions;
3243                         }
3244                 }
3245         }
3246
3247         /**
3248          * Get special page names, as an associative array
3249          *   canonical name => array of valid names, including aliases
3250          * @return array
3251          */
3252         function getSpecialPageAliases() {
3253                 // Cache aliases because it may be slow to load them
3254                 if ( is_null( $this->mExtendedSpecialPageAliases ) ) {
3255                         // Initialise array
3256                         $this->mExtendedSpecialPageAliases =
3257                                 self::$dataCache->getItem( $this->mCode, 'specialPageAliases' );
3258                 }
3259
3260                 return $this->mExtendedSpecialPageAliases;
3261         }
3262
3263         /**
3264          * Italic is unsuitable for some languages
3265          *
3266          * @param string $text The text to be emphasized.
3267          * @return string
3268          */
3269         function emphasize( $text ) {
3270                 return "<em>$text</em>";
3271         }
3272
3273         /**
3274          * Normally we output all numbers in plain en_US style, that is
3275          * 293,291.235 for twohundredninetythreethousand-twohundredninetyone
3276          * point twohundredthirtyfive. However this is not suitable for all
3277          * languages, some such as Bengali (bn) want ২,৯৩,২৯১.২৩৫ and others such as
3278          * Icelandic just want to use commas instead of dots, and dots instead
3279          * of commas like "293.291,235".
3280          *
3281          * An example of this function being called:
3282          * <code>
3283          * wfMessage( 'message' )->numParams( $num )->text()
3284          * </code>
3285          *
3286          * See $separatorTransformTable on MessageIs.php for
3287          * the , => . and . => , implementation.
3288          *
3289          * @todo check if it's viable to use localeconv() for the decimal separator thing.
3290          * @param int|float $number The string to be formatted, should be an integer
3291          *   or a floating point number.
3292          * @param bool $nocommafy Set to true for special numbers like dates
3293          * @return string
3294          */
3295         public function formatNum( $number, $nocommafy = false ) {
3296                 global $wgTranslateNumerals;
3297                 if ( !$nocommafy ) {
3298                         $number = $this->commafy( $number );
3299                         $s = $this->separatorTransformTable();
3300                         if ( $s ) {
3301                                 $number = strtr( $number, $s );
3302                         }
3303                 }
3304
3305                 if ( $wgTranslateNumerals ) {
3306                         $s = $this->digitTransformTable();
3307                         if ( $s ) {
3308                                 $number = strtr( $number, $s );
3309                         }
3310                 }
3311
3312                 return (string)$number;
3313         }
3314
3315         /**
3316          * Front-end for non-commafied formatNum
3317          *
3318          * @param int|float $number The string to be formatted, should be an integer
3319          *        or a floating point number.
3320          * @since 1.21
3321          * @return string
3322          */
3323         public function formatNumNoSeparators( $number ) {
3324                 return $this->formatNum( $number, true );
3325         }
3326
3327         /**
3328          * @param string $number
3329          * @return string
3330          */
3331         public function parseFormattedNumber( $number ) {
3332                 $s = $this->digitTransformTable();
3333                 if ( $s ) {
3334                         // eliminate empty array values such as ''. (T66347)
3335                         $s = array_filter( $s );
3336                         $number = strtr( $number, array_flip( $s ) );
3337                 }
3338
3339                 $s = $this->separatorTransformTable();
3340                 if ( $s ) {
3341                         // eliminate empty array values such as ''. (T66347)
3342                         $s = array_filter( $s );
3343                         $number = strtr( $number, array_flip( $s ) );
3344                 }
3345
3346                 $number = strtr( $number, [ ',' => '' ] );
3347                 return $number;
3348         }
3349
3350         /**
3351          * Adds commas to a given number
3352          * @since 1.19
3353          * @param mixed $number
3354          * @return string
3355          */
3356         function commafy( $number ) {
3357                 $digitGroupingPattern = $this->digitGroupingPattern();
3358                 $minimumGroupingDigits = $this->minimumGroupingDigits();
3359                 if ( $number === null ) {
3360                         return '';
3361                 }
3362
3363                 if ( !$digitGroupingPattern || $digitGroupingPattern === "###,###,###" ) {
3364                         // Default grouping is at thousands, use the same for ###,###,### pattern too.
3365                         // In some languages it's conventional not to insert a thousands separator
3366                         // in numbers that are four digits long (1000-9999).
3367                         if ( $minimumGroupingDigits ) {
3368                                 // Number of '#' characters after last comma in the grouping pattern.
3369                                 // The pattern is hardcoded here, but this would vary for different patterns.
3370                                 $primaryGroupingSize = 3;
3371                                 // Maximum length of a number to suppress digit grouping for.
3372                                 $maximumLength = $minimumGroupingDigits + $primaryGroupingSize - 1;
3373                                 if ( preg_match( '/^\-?\d{1,' . $maximumLength . '}(\.\d+)?$/', $number ) ) {
3374                                         return $number;
3375                                 }
3376                         }
3377                         return strrev( (string)preg_replace( '/(\d{3})(?=\d)(?!\d*\.)/', '$1,', strrev( $number ) ) );
3378                 } else {
3379                         // Ref: http://cldr.unicode.org/translation/number-patterns
3380                         $sign = "";
3381                         if ( intval( $number ) < 0 ) {
3382                                 // For negative numbers apply the algorithm like positive number and add sign.
3383                                 $sign = "-";
3384                                 $number = substr( $number, 1 );
3385                         }
3386                         $integerPart = [];
3387                         $decimalPart = [];
3388                         $numMatches = preg_match_all( "/(#+)/", $digitGroupingPattern, $matches );
3389                         preg_match( "/\d+/", $number, $integerPart );
3390                         preg_match( "/\.\d*/", $number, $decimalPart );
3391                         $groupedNumber = ( count( $decimalPart ) > 0 ) ? $decimalPart[0] : "";
3392                         if ( $groupedNumber === $number ) {
3393                                 // the string does not have any number part. Eg: .12345
3394                                 return $sign . $groupedNumber;
3395                         }
3396                         $start = $end = ( $integerPart ) ? strlen( $integerPart[0] ) : 0;
3397                         while ( $start > 0 ) {
3398                                 $match = $matches[0][$numMatches - 1];
3399                                 $matchLen = strlen( $match );
3400                                 $start = $end - $matchLen;
3401                                 if ( $start < 0 ) {
3402                                         $start = 0;
3403                                 }
3404                                 $groupedNumber = substr( $number, $start, $end - $start ) . $groupedNumber;
3405                                 $end = $start;
3406                                 if ( $numMatches > 1 ) {
3407                                         // use the last pattern for the rest of the number
3408                                         $numMatches--;
3409                                 }
3410                                 if ( $start > 0 ) {
3411                                         $groupedNumber = "," . $groupedNumber;
3412                                 }
3413                         }
3414                         return $sign . $groupedNumber;
3415                 }
3416         }
3417
3418         /**
3419          * @return string
3420          */
3421         function digitGroupingPattern() {
3422                 return self::$dataCache->getItem( $this->mCode, 'digitGroupingPattern' );
3423         }
3424
3425         /**
3426          * @return array
3427          */
3428         function digitTransformTable() {
3429                 return self::$dataCache->getItem( $this->mCode, 'digitTransformTable' );
3430         }
3431
3432         /**
3433          * @return array
3434          */
3435         function separatorTransformTable() {
3436                 return self::$dataCache->getItem( $this->mCode, 'separatorTransformTable' );
3437         }
3438
3439         /**
3440          * @return int|null
3441          */
3442         function minimumGroupingDigits() {
3443                 return self::$dataCache->getItem( $this->mCode, 'minimumGroupingDigits' );
3444         }
3445
3446         /**
3447          * Take a list of strings and build a locale-friendly comma-separated
3448          * list, using the local comma-separator message.
3449          * The last two strings are chained with an "and".
3450          *
3451          * @param string[] $list
3452          * @return string
3453          */
3454         public function listToText( array $list ) {
3455                 $itemCount = count( $list );
3456                 if ( $itemCount < 1 ) {
3457                         return '';
3458                 }
3459                 $text = array_pop( $list );
3460                 if ( $itemCount > 1 ) {
3461                         $and = $this->msg( 'and' )->escaped();
3462                         $space = $this->msg( 'word-separator' )->escaped();
3463                         $comma = '';
3464                         if ( $itemCount > 2 ) {
3465                                 $comma = $this->msg( 'comma-separator' )->escaped();
3466                         }
3467                         $text = implode( $comma, $list ) . $and . $space . $text;
3468                 }
3469                 return $text;
3470         }
3471
3472         /**
3473          * Take a list of strings and build a locale-friendly comma-separated
3474          * list, using the local comma-separator message.
3475          * @param string[] $list Array of strings to put in a comma list
3476          * @return string
3477          */
3478         function commaList( array $list ) {
3479                 return implode(
3480                         wfMessage( 'comma-separator' )->inLanguage( $this )->escaped(),
3481                         $list
3482                 );
3483         }
3484
3485         /**
3486          * Take a list of strings and build a locale-friendly semicolon-separated
3487          * list, using the local semicolon-separator message.
3488          * @param string[] $list Array of strings to put in a semicolon list
3489          * @return string
3490          */
3491         function semicolonList( array $list ) {
3492                 return implode(
3493                         wfMessage( 'semicolon-separator' )->inLanguage( $this )->escaped(),
3494                         $list
3495                 );
3496         }
3497
3498         /**
3499          * Same as commaList, but separate it with the pipe instead.
3500          * @param string[] $list Array of strings to put in a pipe list
3501          * @return string
3502          */
3503         function pipeList( array $list ) {
3504                 return implode(
3505                         wfMessage( 'pipe-separator' )->inLanguage( $this )->escaped(),
3506                         $list
3507                 );
3508         }
3509
3510         /**
3511          * Truncate a string to a specified length in bytes, appending an optional
3512          * string (e.g. for ellipsis)
3513          *
3514          * If $length is negative, the string will be truncated from the beginning
3515          *
3516          * @since 1.31
3517          *
3518          * @param string $string String to truncate
3519          * @param int $length Maximum length in bytes
3520          * @param string $ellipsis String to append to the end of truncated text
3521          * @param bool $adjustLength Subtract length of ellipsis from $length
3522          *
3523          * @return string
3524          */
3525         function truncateForDatabase( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3526                 return $this->truncateInternal(
3527                         $string, $length, $ellipsis, $adjustLength, 'strlen', 'substr'
3528                 );
3529         }
3530
3531         /**
3532          * Truncate a string to a specified number of characters, appending an optional
3533          * string (e.g. for ellipsis).
3534          *
3535          * This provides multibyte version of truncateForDatabase() method of this class,
3536          * suitable for truncation based on number of characters, instead of number of bytes.
3537          *
3538          * If $length is negative, the string will be truncated from the beginning.
3539          *
3540          * @since 1.31
3541          *
3542          * @param string $string String to truncate
3543          * @param int $length Maximum number of characters
3544          * @param string $ellipsis String to append to the end of truncated text
3545          * @param bool $adjustLength Subtract length of ellipsis from $length
3546          *
3547          * @return string
3548          */
3549         function truncateForVisual( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3550                 // Passing encoding to mb_strlen and mb_substr is optional.
3551                 // Encoding defaults to mb_internal_encoding(), which is set to UTF-8 in Setup.php, so
3552                 // explicit specification of encoding is skipped.
3553                 // Note: Both multibyte methods are callables invoked in truncateInternal.
3554                 return $this->truncateInternal(
3555                         $string, $length, $ellipsis, $adjustLength, 'mb_strlen', 'mb_substr'
3556                 );
3557         }
3558
3559         /**
3560          * Internal method used for truncation. This method abstracts text truncation into
3561          * one common method, allowing users to provide length measurement function and
3562          * function for finding substring.
3563          *
3564          * For usages, see truncateForDatabase and truncateForVisual.
3565          *
3566          * @param string $string String to truncate
3567          * @param int $length Maximum length of final text
3568          * @param string $ellipsis String to append to the end of truncated text
3569          * @param bool $adjustLength Subtract length of ellipsis from $length
3570          * @param callable $measureLength Callable function used for determining the length of text
3571          * @param callable $getSubstring Callable function used for getting the substrings
3572          *
3573          * @return string
3574          */
3575         private function truncateInternal(
3576                 $string, $length, $ellipsis, $adjustLength, callable $measureLength, callable $getSubstring
3577         ) {
3578                 # Check if there is no need to truncate
3579                 if ( $measureLength( $string ) <= abs( $length ) ) {
3580                         return $string; // no need to truncate
3581                 }
3582
3583                 # Use the localized ellipsis character
3584                 if ( $ellipsis == '...' ) {
3585                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3586                 }
3587                 if ( $length == 0 ) {
3588                         return $ellipsis; // convention
3589                 }
3590
3591                 $stringOriginal = $string;
3592                 # If ellipsis length is >= $length then we can't apply $adjustLength
3593                 if ( $adjustLength && $measureLength( $ellipsis ) >= abs( $length ) ) {
3594                         $string = $ellipsis; // this can be slightly unexpected
3595                 # Otherwise, truncate and add ellipsis...
3596                 } else {
3597                         $ellipsisLength = $adjustLength ? $measureLength( $ellipsis ) : 0;
3598                         if ( $length > 0 ) {
3599                                 $length -= $ellipsisLength;
3600                                 $string = $getSubstring( $string, 0, $length ); // xyz...
3601                                 $string = $this->removeBadCharLast( $string );
3602                                 $string = rtrim( $string ) . $ellipsis;
3603                         } else {
3604                                 $length += $ellipsisLength;
3605                                 $string = $getSubstring( $string, $length ); // ...xyz
3606                                 $string = $this->removeBadCharFirst( $string );
3607                                 $string = $ellipsis . ltrim( $string );
3608                         }
3609                 }
3610
3611                 # Do not truncate if the ellipsis makes the string longer/equal (T24181).
3612                 # This check is *not* redundant if $adjustLength, due to the single case where
3613                 # LEN($ellipsis) > ABS($limit arg); $stringOriginal could be shorter than $string.
3614                 if ( $measureLength( $string ) < $measureLength( $stringOriginal ) ) {
3615                         return $string;
3616                 } else {
3617                         return $stringOriginal;
3618                 }
3619         }
3620
3621         /**
3622          * Remove bytes that represent an incomplete Unicode character
3623          * at the end of string (e.g. bytes of the char are missing)
3624          *
3625          * @param string $string
3626          * @return string
3627          */
3628         protected function removeBadCharLast( $string ) {
3629                 if ( $string != '' ) {
3630                         $char = ord( $string[strlen( $string ) - 1] );
3631                         $m = [];
3632                         if ( $char >= 0xc0 ) {
3633                                 # We got the first byte only of a multibyte char; remove it.
3634                                 $string = substr( $string, 0, -1 );
3635                         } elseif ( $char >= 0x80 &&
3636                                 // Use the /s modifier (PCRE_DOTALL) so (.*) also matches newlines
3637                                 preg_match( '/^(.*)(?:[\xe0-\xef][\x80-\xbf]|' .
3638                                         '[\xf0-\xf7][\x80-\xbf]{1,2})$/s', $string, $m )
3639                         ) {
3640                                 # We chopped in the middle of a character; remove it
3641                                 $string = $m[1];
3642                         }
3643                 }
3644                 return $string;
3645         }
3646
3647         /**
3648          * Remove bytes that represent an incomplete Unicode character
3649          * at the start of string (e.g. bytes of the char are missing)
3650          *
3651          * @param string $string
3652          * @return string
3653          */
3654         protected function removeBadCharFirst( $string ) {
3655                 if ( $string != '' ) {
3656                         $char = ord( $string[0] );
3657                         if ( $char >= 0x80 && $char < 0xc0 ) {
3658                                 # We chopped in the middle of a character; remove the whole thing
3659                                 $string = preg_replace( '/^[\x80-\xbf]+/', '', $string );
3660                         }
3661                 }
3662                 return $string;
3663         }
3664
3665         /**
3666          * Truncate a string of valid HTML to a specified length in bytes,
3667          * appending an optional string (e.g. for ellipses), and return valid HTML
3668          *
3669          * This is only intended for styled/linked text, such as HTML with
3670          * tags like <span> and <a>, were the tags are self-contained (valid HTML).
3671          * Also, this will not detect things like "display:none" CSS.
3672          *
3673          * Note: since 1.18 you do not need to leave extra room in $length for ellipses.
3674          *
3675          * @param string $text HTML string to truncate
3676          * @param int $length (zero/positive) Maximum length (including ellipses)
3677          * @param string $ellipsis String to append to the truncated text
3678          * @return string
3679          */
3680         function truncateHtml( $text, $length, $ellipsis = '...' ) {
3681                 # Use the localized ellipsis character
3682                 if ( $ellipsis == '...' ) {
3683                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3684                 }
3685                 # Check if there is clearly no need to truncate
3686                 if ( $length <= 0 ) {
3687                         return $ellipsis; // no text shown, nothing to format (convention)
3688                 } elseif ( strlen( $text ) <= $length ) {
3689                         return $text; // string short enough even *with* HTML (short-circuit)
3690                 }
3691
3692                 $dispLen = 0; // innerHTML legth so far
3693                 $testingEllipsis = false; // checking if ellipses will make string longer/equal?
3694                 $tagType = 0; // 0-open, 1-close
3695                 $bracketState = 0; // 1-tag start, 2-tag name, 0-neither
3696                 $entityState = 0; // 0-not entity, 1-entity
3697                 $tag = $ret = ''; // accumulated tag name, accumulated result string
3698                 $openTags = []; // open tag stack
3699                 $maybeState = null; // possible truncation state
3700
3701                 $textLen = strlen( $text );
3702                 $neLength = max( 0, $length - strlen( $ellipsis ) ); // non-ellipsis len if truncated
3703                 for ( $pos = 0; true; ++$pos ) {
3704                         # Consider truncation once the display length has reached the maximim.
3705                         # We check if $dispLen > 0 to grab tags for the $neLength = 0 case.
3706                         # Check that we're not in the middle of a bracket/entity...
3707                         if ( $dispLen && $dispLen >= $neLength && $bracketState == 0 && !$entityState ) {
3708                                 if ( !$testingEllipsis ) {
3709                                         $testingEllipsis = true;
3710                                         # Save where we are; we will truncate here unless there turn out to
3711                                         # be so few remaining characters that truncation is not necessary.
3712                                         if ( !$maybeState ) { // already saved? ($neLength = 0 case)
3713                                                 $maybeState = [ $ret, $openTags ]; // save state
3714                                         }
3715                                 } elseif ( $dispLen > $length && $dispLen > strlen( $ellipsis ) ) {
3716                                         # String in fact does need truncation, the truncation point was OK.
3717                                         // @phan-suppress-next-line PhanTypeInvalidExpressionArrayDestructuring
3718                                         list( $ret, $openTags ) = $maybeState; // reload state
3719                                         $ret = $this->removeBadCharLast( $ret ); // multi-byte char fix
3720                                         $ret .= $ellipsis; // add ellipsis
3721                                         break;
3722                                 }
3723                         }
3724                         if ( $pos >= $textLen ) {
3725                                 break; // extra iteration just for above checks
3726                         }
3727
3728                         # Read the next char...
3729                         $ch = $text[$pos];
3730                         $lastCh = $pos ? $text[$pos - 1] : '';
3731                         $ret .= $ch; // add to result string
3732                         if ( $ch == '<' ) {
3733                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags ); // for bad HTML
3734                                 $entityState = 0; // for bad HTML
3735                                 $bracketState = 1; // tag started (checking for backslash)
3736                         } elseif ( $ch == '>' ) {
3737                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags );
3738                                 $entityState = 0; // for bad HTML
3739                                 $bracketState = 0; // out of brackets
3740                         } elseif ( $bracketState == 1 ) {
3741                                 if ( $ch == '/' ) {
3742                                         $tagType = 1; // close tag (e.g. "</span>")
3743                                 } else {
3744                                         $tagType = 0; // open tag (e.g. "<span>")
3745                                         $tag .= $ch;
3746                                 }
3747                                 $bracketState = 2; // building tag name
3748                         } elseif ( $bracketState == 2 ) {
3749                                 if ( $ch != ' ' ) {
3750                                         $tag .= $ch;
3751                                 } else {
3752                                         // Name found (e.g. "<a href=..."), add on tag attributes...
3753                                         $pos += $this->truncate_skip( $ret, $text, "<>", $pos + 1 );
3754                                 }
3755                         } elseif ( $bracketState == 0 ) {
3756                                 if ( $entityState ) {
3757                                         if ( $ch == ';' ) {
3758                                                 $entityState = 0;
3759                                                 $dispLen++; // entity is one displayed char
3760                                         }
3761                                 } else {
3762                                         if ( $neLength == 0 && !$maybeState ) {
3763                                                 // Save state without $ch. We want to *hit* the first
3764                                                 // display char (to get tags) but not *use* it if truncating.
3765                                                 $maybeState = [ substr( $ret, 0, -1 ), $openTags ];
3766                                         }
3767                                         if ( $ch == '&' ) {
3768                                                 $entityState = 1; // entity found, (e.g. "&#160;")
3769                                         } else {
3770                                                 $dispLen++; // this char is displayed
3771                                                 // Add the next $max display text chars after this in one swoop...
3772                                                 $max = ( $testingEllipsis ? $length : $neLength ) - $dispLen;
3773                                                 $skipped = $this->truncate_skip( $ret, $text, "<>&", $pos + 1, $max );
3774                                                 $dispLen += $skipped;
3775                                                 $pos += $skipped;
3776                                         }
3777                                 }
3778                         }
3779                 }
3780                 // Close the last tag if left unclosed by bad HTML
3781                 $this->truncate_endBracket( $tag, $text[$textLen - 1], $tagType, $openTags );
3782                 while ( count( $openTags ) > 0 ) {
3783                         $ret .= '</' . array_pop( $openTags ) . '>'; // close open tags
3784                 }
3785                 return $ret;
3786         }
3787
3788         /**
3789          * truncateHtml() helper function
3790          * like strcspn() but adds the skipped chars to $ret
3791          *
3792          * @param string $ret
3793          * @param string $text
3794          * @param string $search
3795          * @param int $start
3796          * @param null|int $len
3797          * @return int
3798          */
3799         private function truncate_skip( &$ret, $text, $search, $start, $len = null ) {
3800                 if ( $len === null ) {
3801                         $len = -1; // -1 means "no limit" for strcspn
3802                 } elseif ( $len < 0 ) {
3803                         $len = 0; // sanity
3804                 }
3805                 $skipCount = 0;
3806                 if ( $start < strlen( $text ) ) {
3807                         $skipCount = strcspn( $text, $search, $start, $len );
3808                         $ret .= substr( $text, $start, $skipCount );
3809                 }
3810                 return $skipCount;
3811         }
3812
3813         /**
3814          * truncateHtml() helper function
3815          * (a) push or pop $tag from $openTags as needed
3816          * (b) clear $tag value
3817          * @param string &$tag Current HTML tag name we are looking at
3818          * @param int $tagType (0-open tag, 1-close tag)
3819          * @param string $lastCh Character before the '>' that ended this tag
3820          * @param array &$openTags Open tag stack (not accounting for $tag)
3821          */
3822         private function truncate_endBracket( &$tag, $tagType, $lastCh, &$openTags ) {
3823                 $tag = ltrim( $tag );
3824                 if ( $tag != '' ) {
3825                         if ( $tagType == 0 && $lastCh != '/' ) {
3826                                 $openTags[] = $tag; // tag opened (didn't close itself)
3827                         } elseif ( $tagType == 1 ) {
3828                                 if ( $openTags && $tag == $openTags[count( $openTags ) - 1] ) {
3829                                         array_pop( $openTags ); // tag closed
3830                                 }
3831                         }
3832                         $tag = '';
3833                 }
3834         }
3835
3836         /**
3837          * Grammatical transformations, needed for inflected languages
3838          * Invoked by putting {{grammar:case|word}} in a message
3839          *
3840          * @param string $word
3841          * @param string $case
3842          * @return string
3843          */
3844         function convertGrammar( $word, $case ) {
3845                 global $wgGrammarForms;
3846                 if ( isset( $wgGrammarForms[$this->getCode()][$case][$word] ) ) {
3847                         return $wgGrammarForms[$this->getCode()][$case][$word];
3848                 }
3849
3850                 $grammarTransformations = $this->getGrammarTransformations();
3851
3852                 if ( isset( $grammarTransformations[$case] ) ) {
3853                         $forms = $grammarTransformations[$case];
3854
3855                         // Some names of grammar rules are aliases for other rules.
3856                         // In such cases the value is a string rather than object,
3857                         // so load the actual rules.
3858                         if ( is_string( $forms ) ) {
3859                                 $forms = $grammarTransformations[$forms];
3860                         }
3861
3862                         foreach ( array_values( $forms ) as $rule ) {
3863                                 $form = $rule[0];
3864
3865                                 if ( $form === '@metadata' ) {
3866                                         continue;
3867                                 }
3868
3869                                 $replacement = $rule[1];
3870
3871                                 $regex = '/' . addcslashes( $form, '/' ) . '/u';
3872                                 $patternMatches = preg_match( $regex, $word );
3873
3874                                 if ( $patternMatches === false ) {
3875                                         wfLogWarning(
3876                                                 'An error occurred while processing grammar. ' .
3877                                                 "Word: '$word'. Regex: /$form/."
3878                                         );
3879                                 } elseif ( $patternMatches === 1 ) {
3880                                         $word = preg_replace( $regex, $replacement, $word );
3881
3882                                         break;
3883                                 }
3884                         }
3885                 }
3886
3887                 return $word;
3888         }
3889
3890         /**
3891          * Get the grammar forms for the content language
3892          * @return array Array of grammar forms
3893          * @since 1.20
3894          */
3895         function getGrammarForms() {
3896                 global $wgGrammarForms;
3897                 if ( isset( $wgGrammarForms[$this->getCode()] )
3898                         && is_array( $wgGrammarForms[$this->getCode()] )
3899                 ) {
3900                         return $wgGrammarForms[$this->getCode()];
3901                 }
3902
3903                 return [];
3904         }
3905
3906         /**
3907          * Get the grammar transformations data for the language.
3908          * Used like grammar forms, with {{GRAMMAR}} and cases,
3909          * but uses pairs of regexes and replacements instead of code.
3910          *
3911          * @return array[] Array of grammar transformations.
3912          * @throws MWException
3913          * @since 1.28
3914          */
3915         public function getGrammarTransformations() {
3916                 $languageCode = $this->getCode();
3917
3918                 if ( self::$grammarTransformations === null ) {
3919                         self::$grammarTransformations = new MapCacheLRU( 10 );
3920                 }
3921
3922                 if ( self::$grammarTransformations->has( $languageCode ) ) {
3923                         return self::$grammarTransformations->get( $languageCode );
3924                 }
3925
3926                 $data = [];
3927
3928                 $grammarDataFile = __DIR__ . "/data/grammarTransformations/$languageCode.json";
3929                 if ( is_readable( $grammarDataFile ) ) {
3930                         $data = FormatJson::decode(
3931                                 file_get_contents( $grammarDataFile ),
3932                                 true
3933                         );
3934
3935                         if ( $data === null ) {
3936                                 throw new MWException( "Invalid grammar data for \"$languageCode\"." );
3937                         }
3938
3939                         self::$grammarTransformations->set( $languageCode, $data );
3940                 }
3941
3942                 return $data;
3943         }
3944
3945         /**
3946          * Provides an alternative text depending on specified gender.
3947          * Usage {{gender:username|masculine|feminine|unknown}}.
3948          * username is optional, in which case the gender of current user is used,
3949          * but only in (some) interface messages; otherwise default gender is used.
3950          *
3951          * If no forms are given, an empty string is returned. If only one form is
3952          * given, it will be returned unconditionally. These details are implied by
3953          * the caller and cannot be overridden in subclasses.
3954          *
3955          * If three forms are given, the default is to use the third (unknown) form.
3956          * If fewer than three forms are given, the default is to use the first (masculine) form.
3957          * These details can be overridden in subclasses.
3958          *
3959          * @param string $gender
3960          * @param array $forms
3961          *
3962          * @return string
3963          */
3964         function gender( $gender, $forms ) {
3965                 if ( !count( $forms ) ) {
3966                         return '';
3967                 }
3968                 $forms = $this->preConvertPlural( $forms, 2 );
3969                 if ( $gender === 'male' ) {
3970                         return $forms[0];
3971                 }
3972                 if ( $gender === 'female' ) {
3973                         return $forms[1];
3974                 }
3975                 return $forms[2] ?? $forms[0];
3976         }
3977
3978         /**
3979          * Plural form transformations, needed for some languages.
3980          * For example, there are 3 form of plural in Russian and Polish,
3981          * depending on "count mod 10". See [[w:Plural]]
3982          * For English it is pretty simple.
3983          *
3984          * Invoked by putting {{plural:count|wordform1|wordform2}}
3985          * or {{plural:count|wordform1|wordform2|wordform3}}
3986          *
3987          * Example: {{plural:{{NUMBEROFARTICLES}}|article|articles}}
3988          *
3989          * @param int $count Non-localized number
3990          * @param array $forms Different plural forms
3991          * @return string Correct form of plural for $count in this language
3992          */
3993         function convertPlural( $count, $forms ) {
3994                 // Handle explicit n=pluralform cases
3995                 $forms = $this->handleExplicitPluralForms( $count, $forms );
3996                 if ( is_string( $forms ) ) {
3997                         return $forms;
3998                 }
3999                 if ( !count( $forms ) ) {
4000                         return '';
4001                 }
4002
4003                 $pluralForm = $this->getPluralRuleIndexNumber( $count );
4004                 $pluralForm = min( $pluralForm, count( $forms ) - 1 );
4005                 return $forms[$pluralForm];
4006         }
4007
4008         /**
4009          * Handles explicit plural forms for Language::convertPlural()
4010          *
4011          * In {{PLURAL:$1|0=nothing|one|many}}, 0=nothing will be returned if $1 equals zero.
4012          * If an explicitly defined plural form matches the $count, then
4013          * string value returned, otherwise array returned for further consideration
4014          * by CLDR rules or overridden convertPlural().
4015          *
4016          * @since 1.23
4017          *
4018          * @param int $count Non-localized number
4019          * @param array $forms Different plural forms
4020          *
4021          * @return array|string
4022          */
4023         protected function handleExplicitPluralForms( $count, array $forms ) {
4024                 foreach ( $forms as $index => $form ) {
4025                         if ( preg_match( '/\d+=/i', $form ) ) {
4026                                 $pos = strpos( $form, '=' );
4027                                 if ( substr( $form, 0, $pos ) === (string)$count ) {
4028                                         return substr( $form, $pos + 1 );
4029                                 }
4030                                 unset( $forms[$index] );
4031                         }
4032                 }
4033                 return array_values( $forms );
4034         }
4035
4036         /**
4037          * Checks that convertPlural was given an array and pads it to requested
4038          * amount of forms by copying the last one.
4039          *
4040          * @param array $forms
4041          * @param int $count How many forms should there be at least
4042          * @return array Padded array of forms
4043          */
4044         protected function preConvertPlural( /* Array */ $forms, $count ) {
4045                 return array_pad( $forms, $count, end( $forms ) );
4046         }
4047
4048         /**
4049          * Wraps argument with unicode control characters for directionality safety
4050          *
4051          * This solves the problem where directionality-neutral characters at the edge of
4052          * the argument string get interpreted with the wrong directionality from the
4053          * enclosing context, giving renderings that look corrupted like "(Ben_(WMF".
4054          *
4055          * The wrapping is LRE...PDF or RLE...PDF, depending on the detected
4056          * directionality of the argument string, using the BIDI algorithm's own "First
4057          * strong directional codepoint" rule. Essentially, this works round the fact that
4058          * there is no embedding equivalent of U+2068 FSI (isolation with heuristic
4059          * direction inference). The latter is cleaner but still not widely supported.
4060          *
4061          * @param string $text Text to wrap
4062          * @return string Text, wrapped in LRE...PDF or RLE...PDF or nothing
4063          */
4064         public function embedBidi( $text = '' ) {
4065                 $dir = self::strongDirFromContent( $text );
4066                 if ( $dir === 'ltr' ) {
4067                         // Wrap in LEFT-TO-RIGHT EMBEDDING ... POP DIRECTIONAL FORMATTING
4068                         return self::$lre . $text . self::$pdf;
4069                 }
4070                 if ( $dir === 'rtl' ) {
4071                         // Wrap in RIGHT-TO-LEFT EMBEDDING ... POP DIRECTIONAL FORMATTING
4072                         return self::$rle . $text . self::$pdf;
4073                 }
4074                 // No strong directionality: do not wrap
4075                 return $text;
4076         }
4077
4078         /**
4079          * @todo Maybe translate block durations.  Note that this function is somewhat misnamed: it
4080          * deals with translating the *duration* ("1 week", "4 days", etc), not the expiry time
4081          * (which is an absolute timestamp). Please note: do NOT add this blindly, as it is used
4082          * on old expiry lengths recorded in log entries. You'd need to provide the start date to
4083          * match up with it.
4084          *
4085          * @param string $str The validated block duration in English
4086          * @param User|null $user User object to use timezone from or null for $wgUser
4087          * @param int $now Current timestamp, for formatting relative block durations
4088          * @return string Somehow translated block duration
4089          * @see LanguageFi.php for example implementation
4090          */
4091         function translateBlockExpiry( $str, User $user = null, $now = 0 ) {
4092                 $duration = SpecialBlock::getSuggestedDurations( $this );
4093                 foreach ( $duration as $show => $value ) {
4094                         if ( strcmp( $str, $value ) == 0 ) {
4095                                 return htmlspecialchars( trim( $show ) );
4096                         }
4097                 }
4098
4099                 if ( wfIsInfinity( $str ) ) {
4100                         foreach ( $duration as $show => $value ) {
4101                                 if ( wfIsInfinity( $value ) ) {
4102                                         return htmlspecialchars( trim( $show ) );
4103                                 }
4104                         }
4105                 }
4106
4107                 // If all else fails, return a standard duration or timestamp description.
4108                 $time = strtotime( $str, $now );
4109                 if ( $time === false ) { // Unknown format. Return it as-is in case.
4110                         return $str;
4111                 } elseif ( $time !== strtotime( $str, $now + 1 ) ) { // It's a relative timestamp.
4112                         // The result differs based on current time, so the difference
4113                         // is a fixed duration length.
4114                         return $this->formatDuration( $time - $now );
4115                 } else { // It's an absolute timestamp.
4116                         if ( $time === 0 ) {
4117                                 // wfTimestamp() handles 0 as current time instead of epoch.
4118                                 $time = '19700101000000';
4119                         }
4120                         if ( $user ) {
4121                                 return $this->userTimeAndDate( $time, $user );
4122                         }
4123                         return $this->timeanddate( $time );
4124                 }
4125         }
4126
4127         /**
4128          * languages like Chinese need to be segmented in order for the diff
4129          * to be of any use
4130          *
4131          * @param string $text
4132          * @return string
4133          */
4134         public function segmentForDiff( $text ) {
4135                 return $text;
4136         }
4137
4138         /**
4139          * and unsegment to show the result
4140          *
4141          * @param string $text
4142          * @return string
4143          */
4144         public function unsegmentForDiff( $text ) {
4145                 return $text;
4146         }
4147
4148         /**
4149          * Return the LanguageConverter used in the Language
4150          *
4151          * @since 1.19
4152          * @return LanguageConverter
4153          */
4154         public function getConverter() {
4155                 return $this->mConverter;
4156         }
4157
4158         /**
4159          * convert text to a variant
4160          *
4161          * @param string $text text to convert
4162          * @param string|bool $variant variant to convert to, or false to use the user's preferred
4163          *      variant (if logged in), or the project default variant
4164          * @return string the converted string
4165          */
4166         public function autoConvert( $text, $variant = false ) {
4167                 return $this->mConverter->autoConvert( $text, $variant );
4168         }
4169
4170         /**
4171          * convert text to all supported variants
4172          *
4173          * @param string $text
4174          * @return array
4175          */
4176         public function autoConvertToAllVariants( $text ) {
4177                 return $this->mConverter->autoConvertToAllVariants( $text );
4178         }
4179
4180         /**
4181          * convert text to different variants of a language.
4182          *
4183          * @warning Glossary state is maintained between calls. This means
4184          *  if you pass unescaped text to this method it can cause an XSS
4185          *  in later calls to this method, even if the later calls have properly
4186          *  escaped the input. Never feed this method user controlled text that
4187          *  is not properly escaped!
4188          * @param string $text Content that has been already escaped for use in HTML
4189          * @return string HTML
4190          */
4191         public function convert( $text ) {
4192                 return $this->mConverter->convert( $text );
4193         }
4194
4195         /**
4196          * Convert a Title object to a string in the preferred variant
4197          *
4198          * @param Title $title
4199          * @return string
4200          */
4201         public function convertTitle( $title ) {
4202                 return $this->mConverter->convertTitle( $title );
4203         }
4204
4205         /**
4206          * Convert a namespace index to a string in the preferred variant
4207          *
4208          * @param int $ns namespace index (https://www.mediawiki.org/wiki/Manual:Namespace)
4209          * @param string|null $variant variant to convert to, or null to use the user's preferred
4210          *      variant (if logged in), or the project default variant
4211          * @return string a string representation of the namespace
4212          */
4213         public function convertNamespace( $ns, $variant = null ) {
4214                 return $this->mConverter->convertNamespace( $ns, $variant );
4215         }
4216
4217         /**
4218          * Check if this is a language with variants
4219          *
4220          * @return bool
4221          */
4222         public function hasVariants() {
4223                 return count( $this->getVariants() ) > 1;
4224         }
4225
4226         /**
4227          * Strict check if the language has the specific variant.
4228          *
4229          * Compare to LanguageConverter::validateVariant() which does a more
4230          * lenient check and attempts to coerce the given code to a valid one.
4231          *
4232          * @since 1.19
4233          * @param string $variant
4234          * @return bool
4235          */
4236         public function hasVariant( $variant ) {
4237                 return $variant && ( $variant === $this->mConverter->validateVariant( $variant ) );
4238         }
4239
4240         /**
4241          * Perform output conversion on a string, and encode for safe HTML output.
4242          * @param string $text Text to be converted
4243          * @return string
4244          * @todo this should get integrated somewhere sane
4245          */
4246         public function convertHtml( $text ) {
4247                 return htmlspecialchars( $this->convert( $text ) );
4248         }
4249
4250         /**
4251          * @param string $key
4252          * @return string
4253          */
4254         public function convertCategoryKey( $key ) {
4255                 return $this->mConverter->convertCategoryKey( $key );
4256         }
4257
4258         /**
4259          * Get the list of variants supported by this language
4260          * see sample implementation in LanguageZh.php
4261          *
4262          * @return string[] An array of language codes
4263          */
4264         public function getVariants() {
4265                 return $this->mConverter->getVariants();
4266         }
4267
4268         /**
4269          * @return string
4270          */
4271         public function getPreferredVariant() {
4272                 return $this->mConverter->getPreferredVariant();
4273         }
4274
4275         /**
4276          * @return string
4277          */
4278         public function getDefaultVariant() {
4279                 return $this->mConverter->getDefaultVariant();
4280         }
4281
4282         /**
4283          * @return string
4284          */
4285         public function getURLVariant() {
4286                 return $this->mConverter->getURLVariant();
4287         }
4288
4289         /**
4290          * If a language supports multiple variants, it is
4291          * possible that non-existing link in one variant
4292          * actually exists in another variant. this function
4293          * tries to find it. See e.g. LanguageZh.php
4294          * The input parameters may be modified upon return
4295          *
4296          * @param string &$link The name of the link
4297          * @param Title &$nt The title object of the link
4298          * @param bool $ignoreOtherCond To disable other conditions when
4299          *   we need to transclude a template or update a category's link
4300          */
4301         public function findVariantLink( &$link, &$nt, $ignoreOtherCond = false ) {
4302                 $this->mConverter->findVariantLink( $link, $nt, $ignoreOtherCond );
4303         }
4304
4305         /**
4306          * returns language specific options used by User::getPageRenderHash()
4307          * for example, the preferred language variant
4308          *
4309          * @return string
4310          */
4311         function getExtraHashOptions() {
4312                 return $this->mConverter->getExtraHashOptions();
4313         }
4314
4315         /**
4316          * For languages that support multiple variants, the title of an
4317          * article may be displayed differently in different variants. this
4318          * function returns the apporiate title defined in the body of the article.
4319          *
4320          * @return string
4321          */
4322         public function getParsedTitle() {
4323                 return $this->mConverter->getParsedTitle();
4324         }
4325
4326         /**
4327          * Refresh the cache of conversion tables when
4328          * MediaWiki:Conversiontable* is updated.
4329          *
4330          * @param Title $title The Title of the page being updated
4331          */
4332         public function updateConversionTable( Title $title ) {
4333                 $this->mConverter->updateConversionTable( $title );
4334         }
4335
4336         /**
4337          * A regular expression to match legal word-trailing characters
4338          * which should be merged onto a link of the form [[foo]]bar.
4339          *
4340          * @return string
4341          */
4342         public function linkTrail() {
4343                 return self::$dataCache->getItem( $this->mCode, 'linkTrail' );
4344         }
4345
4346         /**
4347          * A regular expression character set to match legal word-prefixing
4348          * characters which should be merged onto a link of the form foo[[bar]].
4349          *
4350          * @return string
4351          */
4352         public function linkPrefixCharset() {
4353                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixCharset' );
4354         }
4355
4356         /**
4357          * Get the "parent" language which has a converter to convert a "compatible" language
4358          * (in another variant) to this language (eg. zh for zh-cn, but not en for en-gb).
4359          *
4360          * @return Language|null
4361          * @since 1.22
4362          */
4363         public function getParentLanguage() {
4364                 if ( $this->mParentLanguage !== false ) {
4365                         return $this->mParentLanguage;
4366                 }
4367
4368                 $code = explode( '-', $this->getCode() )[0];
4369                 if ( !in_array( $code, LanguageConverter::$languagesWithVariants ) ) {
4370                         $this->mParentLanguage = null;
4371                         return null;
4372                 }
4373                 $lang = self::factory( $code );
4374                 if ( !$lang->hasVariant( $this->getCode() ) ) {
4375                         $this->mParentLanguage = null;
4376                         return null;
4377                 }
4378
4379                 $this->mParentLanguage = $lang;
4380                 return $lang;
4381         }
4382
4383         /**
4384          * Compare with an other language object
4385          *
4386          * @since 1.28
4387          * @param Language $lang
4388          * @return bool
4389          */
4390         public function equals( Language $lang ) {
4391                 return $lang === $this || $lang->getCode() === $this->mCode;
4392         }
4393
4394         /**
4395          * Get the internal language code for this language object
4396          *
4397          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4398          * htmlspecialchars() or similar
4399          *
4400          * @return string
4401          */
4402         public function getCode() {
4403                 return $this->mCode;
4404         }
4405
4406         /**
4407          * Get the code in BCP 47 format which we can use
4408          * inside of html lang="" tags.
4409          *
4410          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4411          * htmlspecialchars() or similar.
4412          *
4413          * @since 1.19
4414          * @return string
4415          */
4416         public function getHtmlCode() {
4417                 if ( is_null( $this->mHtmlCode ) ) {
4418                         $this->mHtmlCode = LanguageCode::bcp47( $this->getCode() );
4419                 }
4420                 return $this->mHtmlCode;
4421         }
4422
4423         /**
4424          * Get the language code from a file name. Inverse of getFileName()
4425          * @param string $filename $prefix . $languageCode . $suffix
4426          * @param string $prefix Prefix before the language code
4427          * @param string $suffix Suffix after the language code
4428          * @return string Language code, or false if $prefix or $suffix isn't found
4429          */
4430         public static function getCodeFromFileName( $filename, $prefix = 'Language', $suffix = '.php' ) {
4431                 $m = null;
4432                 preg_match( '/' . preg_quote( $prefix, '/' ) . '([A-Z][a-z_]+)' .
4433                         preg_quote( $suffix, '/' ) . '/', $filename, $m );
4434                 if ( !count( $m ) ) {
4435                         return false;
4436                 }
4437                 return str_replace( '_', '-', strtolower( $m[1] ) );
4438         }
4439
4440         /**
4441          * @param string $code
4442          * @param bool $fallback Whether we're going through language fallback chain
4443          * @return string Name of the language class
4444          */
4445         public static function classFromCode( $code, $fallback = true ) {
4446                 if ( $fallback && $code == 'en' ) {
4447                         return 'Language';
4448                 } else {
4449                         return 'Language' . str_replace( '-', '_', ucfirst( $code ) );
4450                 }
4451         }
4452
4453         /**
4454          * Get the name of a file for a certain language code
4455          * @param string $prefix Prepend this to the filename
4456          * @param string $code Language code
4457          * @param string $suffix Append this to the filename
4458          * @throws MWException
4459          * @return string $prefix . $mangledCode . $suffix
4460          */
4461         public static function getFileName( $prefix, $code, $suffix = '.php' ) {
4462                 if ( !self::isValidBuiltInCode( $code ) ) {
4463                         throw new MWException( "Invalid language code \"$code\"" );
4464                 }
4465
4466                 return $prefix . str_replace( '-', '_', ucfirst( $code ) ) . $suffix;
4467         }
4468
4469         /**
4470          * @param string $code
4471          * @return string
4472          */
4473         public static function getMessagesFileName( $code ) {
4474                 global $IP;
4475                 $file = self::getFileName( "$IP/languages/messages/Messages", $code, '.php' );
4476                 Hooks::run( 'Language::getMessagesFileName', [ $code, &$file ] );
4477                 return $file;
4478         }
4479
4480         /**
4481          * @param string $code
4482          * @return string
4483          * @throws MWException
4484          * @since 1.23
4485          */
4486         public static function getJsonMessagesFileName( $code ) {
4487                 global $IP;
4488
4489                 if ( !self::isValidBuiltInCode( $code ) ) {
4490                         throw new MWException( "Invalid language code \"$code\"" );
4491                 }
4492
4493                 return "$IP/languages/i18n/$code.json";
4494         }
4495
4496         /**
4497          * Get the first fallback for a given language.
4498          *
4499          * @param string $code
4500          *
4501          * @return bool|string
4502          */
4503         public static function getFallbackFor( $code ) {
4504                 $fallbacks = self::getFallbacksFor( $code );
4505                 if ( $fallbacks ) {
4506                         return $fallbacks[0];
4507                 }
4508                 return false;
4509         }
4510
4511         /**
4512          * Get the ordered list of fallback languages.
4513          *
4514          * @since 1.19
4515          * @param string $code Language code
4516          * @param int $mode Fallback mode, either MESSAGES_FALLBACKS (which always falls back to 'en'),
4517          * or STRICT_FALLBACKS (whic honly falls back to 'en' when explicitly defined)
4518          * @throws MWException
4519          * @return array List of language codes
4520          */
4521         public static function getFallbacksFor( $code, $mode = self::MESSAGES_FALLBACKS ) {
4522                 if ( $code === 'en' || !self::isValidBuiltInCode( $code ) ) {
4523                         return [];
4524                 }
4525                 switch ( $mode ) {
4526                         case self::MESSAGES_FALLBACKS:
4527                                 // For unknown languages, fallbackSequence returns an empty array,
4528                                 // hardcode fallback to 'en' in that case as English messages are
4529                                 // always defined.
4530                                 return self::getLocalisationCache()->getItem( $code, 'fallbackSequence' ) ?: [ 'en' ];
4531                         case self::STRICT_FALLBACKS:
4532                                 // Use this mode when you don't want to fallback to English unless
4533                                 // explicitly defined, for example when you have language-variant icons
4534                                 // and an international language-independent fallback.
4535                                 return self::getLocalisationCache()->getItem( $code, 'originalFallbackSequence' );
4536                         default:
4537                                 throw new MWException( "Invalid fallback mode \"$mode\"" );
4538                 }
4539         }
4540
4541         /**
4542          * Get the ordered list of fallback languages, ending with the fallback
4543          * language chain for the site language.
4544          *
4545          * @since 1.22
4546          * @param string $code Language code
4547          * @return array [ fallbacks, site fallbacks ]
4548          */
4549         public static function getFallbacksIncludingSiteLanguage( $code ) {
4550                 global $wgLanguageCode;
4551
4552                 // Usually, we will only store a tiny number of fallback chains, so we
4553                 // keep them in static memory.
4554                 $cacheKey = "{$code}-{$wgLanguageCode}";
4555
4556                 if ( !array_key_exists( $cacheKey, self::$fallbackLanguageCache ) ) {
4557                         $fallbacks = self::getFallbacksFor( $code );
4558
4559                         // Append the site's fallback chain, including the site language itself
4560                         $siteFallbacks = self::getFallbacksFor( $wgLanguageCode );
4561                         array_unshift( $siteFallbacks, $wgLanguageCode );
4562
4563                         // Eliminate any languages already included in the chain
4564                         $siteFallbacks = array_diff( $siteFallbacks, $fallbacks );
4565
4566                         self::$fallbackLanguageCache[$cacheKey] = [ $fallbacks, $siteFallbacks ];
4567                 }
4568                 return self::$fallbackLanguageCache[$cacheKey];
4569         }
4570
4571         /**
4572          * Get all messages for a given language
4573          * WARNING: this may take a long time. If you just need all message *keys*
4574          * but need the *contents* of only a few messages, consider using getMessageKeysFor().
4575          *
4576          * @param string $code
4577          *
4578          * @return array
4579          */
4580         public static function getMessagesFor( $code ) {
4581                 return self::getLocalisationCache()->getItem( $code, 'messages' );
4582         }
4583
4584         /**
4585          * Get a message for a given language
4586          *
4587          * @param string $key
4588          * @param string $code
4589          *
4590          * @return string
4591          */
4592         public static function getMessageFor( $key, $code ) {
4593                 return self::getLocalisationCache()->getSubitem( $code, 'messages', $key );
4594         }
4595
4596         /**
4597          * Get all message keys for a given language. This is a faster alternative to
4598          * array_keys( Language::getMessagesFor( $code ) )
4599          *
4600          * @since 1.19
4601          * @param string $code Language code
4602          * @return array Array of message keys (strings)
4603          */
4604         public static function getMessageKeysFor( $code ) {
4605                 return self::getLocalisationCache()->getSubitemList( $code, 'messages' );
4606         }
4607
4608         /**
4609          * @param string $talk
4610          * @return string
4611          */
4612         function fixVariableInNamespace( $talk ) {
4613                 if ( strpos( $talk, '$1' ) === false ) {
4614                         return $talk;
4615                 }
4616
4617                 global $wgMetaNamespace;
4618                 $talk = str_replace( '$1', $wgMetaNamespace, $talk );
4619
4620                 # Allow grammar transformations
4621                 # Allowing full message-style parsing would make simple requests
4622                 # such as action=raw much more expensive than they need to be.
4623                 # This will hopefully cover most cases.
4624                 $talk = preg_replace_callback( '/{{grammar:(.*?)\|(.*?)}}/i',
4625                         [ $this, 'replaceGrammarInNamespace' ], $talk );
4626                 return str_replace( ' ', '_', $talk );
4627         }
4628
4629         /**
4630          * @param string $m
4631          * @return string
4632          */
4633         function replaceGrammarInNamespace( $m ) {
4634                 return $this->convertGrammar( trim( $m[2] ), trim( $m[1] ) );
4635         }
4636
4637         /**
4638          * Decode an expiry (block, protection, etc) which has come from the DB
4639          *
4640          * @param string $expiry Database expiry String
4641          * @param bool|int $format True to process using language functions, or TS_ constant
4642          *     to return the expiry in a given timestamp
4643          * @param string $infinity If $format is not true, use this string for infinite expiry
4644          * @return string
4645          * @since 1.18
4646          */
4647         public function formatExpiry( $expiry, $format = true, $infinity = 'infinity' ) {
4648                 static $dbInfinity;
4649                 if ( $dbInfinity === null ) {
4650                         $dbInfinity = wfGetDB( DB_REPLICA )->getInfinity();
4651                 }
4652
4653                 if ( $expiry == '' || $expiry === 'infinity' || $expiry == $dbInfinity ) {
4654                         return $format === true
4655                                 ? $this->getMessageFromDB( 'infiniteblock' )
4656                                 : $infinity;
4657                 } else {
4658                         return $format === true
4659                                 ? $this->timeanddate( $expiry, /* User preference timezone */ true )
4660                                 : wfTimestamp( $format, $expiry );
4661                 }
4662         }
4663
4664         /**
4665          * Formats a time given in seconds into a string representation of that time.
4666          *
4667          * @param int|float $seconds
4668          * @param array $format An optional argument that formats the returned string in different ways:
4669          *   If $format['avoid'] === 'avoidhours': don't show hours, just show days
4670          *   If $format['avoid'] === 'avoidseconds': don't show seconds if $seconds >= 1 hour,
4671          *   If $format['avoid'] === 'avoidminutes': don't show seconds/minutes if $seconds > 48 hours,
4672          *   If $format['noabbrevs'] is true: use 'seconds' and friends instead of 'seconds-abbrev'
4673          *     and friends.
4674          * @note For backwards compatibility, $format may also be one of the strings 'avoidseconds'
4675          *     or 'avoidminutes'.
4676          * @return string
4677          */
4678         function formatTimePeriod( $seconds, $format = [] ) {
4679                 if ( !is_array( $format ) ) {
4680                         $format = [ 'avoid' => $format ]; // For backwards compatibility
4681                 }
4682                 if ( !isset( $format['avoid'] ) ) {
4683                         $format['avoid'] = false;
4684                 }
4685                 if ( !isset( $format['noabbrevs'] ) ) {
4686                         $format['noabbrevs'] = false;
4687                 }
4688                 $secondsMsg = wfMessage(
4689                         $format['noabbrevs'] ? 'seconds' : 'seconds-abbrev' )->inLanguage( $this );
4690                 $minutesMsg = wfMessage(
4691                         $format['noabbrevs'] ? 'minutes' : 'minutes-abbrev' )->inLanguage( $this );
4692                 $hoursMsg = wfMessage(
4693                         $format['noabbrevs'] ? 'hours' : 'hours-abbrev' )->inLanguage( $this );
4694                 $daysMsg = wfMessage(
4695                         $format['noabbrevs'] ? 'days' : 'days-abbrev' )->inLanguage( $this );
4696
4697                 if ( round( $seconds * 10 ) < 100 ) {
4698                         $s = $this->formatNum( sprintf( "%.1f", round( $seconds * 10 ) / 10 ) );
4699                         $s = $secondsMsg->params( $s )->text();
4700                 } elseif ( round( $seconds ) < 60 ) {
4701                         $s = $this->formatNum( round( $seconds ) );
4702                         $s = $secondsMsg->params( $s )->text();
4703                 } elseif ( round( $seconds ) < 3600 ) {
4704                         $minutes = floor( $seconds / 60 );
4705                         $secondsPart = round( fmod( $seconds, 60 ) );
4706                         if ( $secondsPart == 60 ) {
4707                                 $secondsPart = 0;
4708                                 $minutes++;
4709                         }
4710                         $s = $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4711                         $s .= ' ';
4712                         $s .= $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4713                 } elseif ( round( $seconds ) <= 2 * 86400 ) {
4714                         $hours = floor( $seconds / 3600 );
4715                         $minutes = floor( ( $seconds - $hours * 3600 ) / 60 );
4716                         $secondsPart = round( $seconds - $hours * 3600 - $minutes * 60 );
4717                         if ( $secondsPart == 60 ) {
4718                                 $secondsPart = 0;
4719                                 $minutes++;
4720                         }
4721                         if ( $minutes == 60 ) {
4722                                 $minutes = 0;
4723                                 $hours++;
4724                         }
4725                         $s = $hoursMsg->params( $this->formatNum( $hours ) )->text();
4726                         $s .= ' ';
4727                         $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4728                         if ( !in_array( $format['avoid'], [ 'avoidseconds', 'avoidminutes', 'avoidhours' ] ) ) {
4729                                 $s .= ' ' . $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4730                         }
4731                 } else {
4732                         $days = floor( $seconds / 86400 );
4733                         if ( $format['avoid'] === 'avoidhours' ) {
4734                                 $hours = round( ( $seconds - $days * 86400 ) / 3600 );
4735                                 if ( $hours == 24 ) {
4736                                         $hours = 0;
4737                                         $days++;
4738                                 }
4739                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4740                         } elseif ( $format['avoid'] === 'avoidminutes' ) {
4741                                 $hours = round( ( $seconds - $days * 86400 ) / 3600 );
4742                                 if ( $hours == 24 ) {
4743                                         $hours = 0;
4744                                         $days++;
4745                                 }
4746                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4747                                 $s .= ' ';
4748                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4749                         } elseif ( $format['avoid'] === 'avoidseconds' ) {
4750                                 $hours = floor( ( $seconds - $days * 86400 ) / 3600 );
4751                                 $minutes = round( ( $seconds - $days * 86400 - $hours * 3600 ) / 60 );
4752                                 if ( $minutes == 60 ) {
4753                                         $minutes = 0;
4754                                         $hours++;
4755                                 }
4756                                 if ( $hours == 24 ) {
4757                                         $hours = 0;
4758                                         $days++;
4759                                 }
4760                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4761                                 $s .= ' ';
4762                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4763                                 $s .= ' ';
4764                                 $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4765                         } else {
4766                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4767                                 $s .= ' ';
4768                                 $s .= $this->formatTimePeriod( $seconds - $days * 86400, $format );
4769                         }
4770                 }
4771                 return $s;
4772         }
4773
4774         /**
4775          * Format a bitrate for output, using an appropriate
4776          * unit (bps, kbps, Mbps, Gbps, Tbps, Pbps, Ebps, Zbps or Ybps) according to
4777          *   the magnitude in question.
4778          *
4779          * This use base 1000. For base 1024 use formatSize(), for another base
4780          * see formatComputingNumbers().
4781          *
4782          * @param int $bps
4783          * @return string
4784          */
4785         function formatBitrate( $bps ) {
4786                 return $this->formatComputingNumbers( $bps, 1000, "bitrate-$1bits" );
4787         }
4788
4789         /**
4790          * @param int $size Size of the unit
4791          * @param int $boundary Size boundary (1000, or 1024 in most cases)
4792          * @param string $messageKey Message key to be uesd
4793          * @return string
4794          */
4795         function formatComputingNumbers( $size, $boundary, $messageKey ) {
4796                 if ( $size <= 0 ) {
4797                         return str_replace( '$1', $this->formatNum( $size ),
4798                                 $this->getMessageFromDB( str_replace( '$1', '', $messageKey ) )
4799                         );
4800                 }
4801                 $sizes = [ '', 'kilo', 'mega', 'giga', 'tera', 'peta', 'exa', 'zeta', 'yotta' ];
4802                 $index = 0;
4803
4804                 $maxIndex = count( $sizes ) - 1;
4805                 while ( $size >= $boundary && $index < $maxIndex ) {
4806                         $index++;
4807                         $size /= $boundary;
4808                 }
4809
4810                 // For small sizes no decimal places necessary
4811                 $round = 0;
4812                 if ( $index > 1 ) {
4813                         // For MB and bigger two decimal places are smarter
4814                         $round = 2;
4815                 }
4816                 $msg = str_replace( '$1', $sizes[$index], $messageKey );
4817
4818                 $size = round( $size, $round );
4819                 $text = $this->getMessageFromDB( $msg );
4820                 return str_replace( '$1', $this->formatNum( $size ), $text );
4821         }
4822
4823         /**
4824          * Format a size in bytes for output, using an appropriate
4825          * unit (B, KB, MB, GB, TB, PB, EB, ZB or YB) according to the magnitude in question
4826          *
4827          * This method use base 1024. For base 1000 use formatBitrate(), for
4828          * another base see formatComputingNumbers()
4829          *
4830          * @param int $size Size to format
4831          * @return string Plain text (not HTML)
4832          */
4833         function formatSize( $size ) {
4834                 return $this->formatComputingNumbers( $size, 1024, "size-$1bytes" );
4835         }
4836
4837         /**
4838          * Make a list item, used by various special pages
4839          *
4840          * @param string $page Page link
4841          * @param string $details HTML safe text between brackets
4842          * @param bool $oppositedm Add the direction mark opposite to your
4843          *   language, to display text properly
4844          * @return string HTML escaped
4845          */
4846         function specialList( $page, $details, $oppositedm = true ) {
4847                 if ( !$details ) {
4848                         return $page;
4849                 }
4850
4851                 $dirmark = ( $oppositedm ? $this->getDirMark( true ) : '' ) . $this->getDirMark();
4852                 return $page .
4853                         $dirmark .
4854                         $this->msg( 'word-separator' )->escaped() .
4855                         $this->msg( 'parentheses' )->rawParams( $details )->escaped();
4856         }
4857
4858         /**
4859          * Generate (prev x| next x) (20|50|100...) type links for paging
4860          *
4861          * @param Title $title Title object to link
4862          * @param int $offset
4863          * @param int $limit
4864          * @param array $query Optional URL query parameter string
4865          * @param bool $atend Optional param for specified if this is the last page
4866          * @return string
4867          * @deprecated since 1.34, use PrevNextNavigationRenderer::buildPrevNextNavigation()
4868          *  instead.
4869          */
4870         public function viewPrevNext( Title $title, $offset, $limit,
4871                 array $query = [], $atend = false
4872         ) {
4873                 wfDeprecated( __METHOD__, '1.34' );
4874                 // @todo FIXME: Why on earth this needs one message for the text and another one for tooltip?
4875
4876                 # Make 'previous' link
4877                 $prev = wfMessage( 'prevn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4878                 if ( $offset > 0 ) {
4879                         $plink = $this->numLink( $title, max( $offset - $limit, 0 ), $limit,
4880                                 $query, $prev, 'prevn-title', 'mw-prevlink' );
4881                 } else {
4882                         $plink = htmlspecialchars( $prev );
4883                 }
4884
4885                 # Make 'next' link
4886                 $next = wfMessage( 'nextn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4887                 if ( $atend ) {
4888                         $nlink = htmlspecialchars( $next );
4889                 } else {
4890                         $nlink = $this->numLink( $title, $offset + $limit, $limit,
4891                                 $query, $next, 'nextn-title', 'mw-nextlink' );
4892                 }
4893
4894                 # Make links to set number of items per page
4895                 $numLinks = [];
4896                 foreach ( [ 20, 50, 100, 250, 500 ] as $num ) {
4897                         $numLinks[] = $this->numLink( $title, $offset, $num,
4898                                 $query, $this->formatNum( $num ), 'shown-title', 'mw-numlink' );
4899                 }
4900
4901                 return wfMessage( 'viewprevnext' )->inLanguage( $this )->title( $title
4902                         )->rawParams( $plink, $nlink, $this->pipeList( $numLinks ) )->escaped();
4903         }
4904
4905         /**
4906          * Helper function for viewPrevNext() that generates links
4907          *
4908          * @param Title $title Title object to link
4909          * @param int $offset
4910          * @param int $limit
4911          * @param array $query Extra query parameters
4912          * @param string $link Text to use for the link; will be escaped
4913          * @param string $tooltipMsg Name of the message to use as tooltip
4914          * @param string $class Value of the "class" attribute of the link
4915          * @return string HTML fragment
4916          */
4917         private function numLink( Title $title, $offset, $limit, array $query, $link,
4918                 $tooltipMsg, $class
4919         ) {
4920                 $query = [ 'limit' => $limit, 'offset' => $offset ] + $query;
4921                 $tooltip = wfMessage( $tooltipMsg )->inLanguage( $this )->title( $title )
4922                         ->numParams( $limit )->text();
4923
4924                 return Html::element( 'a', [ 'href' => $title->getLocalURL( $query ),
4925                         'title' => $tooltip, 'class' => $class ], $link );
4926         }
4927
4928         /**
4929          * Get the conversion rule title, if any.
4930          *
4931          * @return string
4932          */
4933         public function getConvRuleTitle() {
4934                 return $this->mConverter->getConvRuleTitle();
4935         }
4936
4937         /**
4938          * Get the compiled plural rules for the language
4939          * @since 1.20
4940          * @return array Associative array with plural form, and plural rule as key-value pairs
4941          */
4942         public function getCompiledPluralRules() {
4943                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'compiledPluralRules' );
4944                 $fallbacks = self::getFallbacksFor( $this->mCode );
4945                 if ( !$pluralRules ) {
4946                         foreach ( $fallbacks as $fallbackCode ) {
4947                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'compiledPluralRules' );
4948                                 if ( $pluralRules ) {
4949                                         break;
4950                                 }
4951                         }
4952                 }
4953                 return $pluralRules;
4954         }
4955
4956         /**
4957          * Get the plural rules for the language
4958          * @since 1.20
4959          * @return array Associative array with plural form number and plural rule as key-value pairs
4960          */
4961         public function getPluralRules() {
4962                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRules' );
4963                 $fallbacks = self::getFallbacksFor( $this->mCode );
4964                 if ( !$pluralRules ) {
4965                         foreach ( $fallbacks as $fallbackCode ) {
4966                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRules' );
4967                                 if ( $pluralRules ) {
4968                                         break;
4969                                 }
4970                         }
4971                 }
4972                 return $pluralRules;
4973         }
4974
4975         /**
4976          * Get the plural rule types for the language
4977          * @since 1.22
4978          * @return array Associative array with plural form number and plural rule type as key-value pairs
4979          */
4980         public function getPluralRuleTypes() {
4981                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRuleTypes' );
4982                 $fallbacks = self::getFallbacksFor( $this->mCode );
4983                 if ( !$pluralRuleTypes ) {
4984                         foreach ( $fallbacks as $fallbackCode ) {
4985                                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRuleTypes' );
4986                                 if ( $pluralRuleTypes ) {
4987                                         break;
4988                                 }
4989                         }
4990                 }
4991                 return $pluralRuleTypes;
4992         }
4993
4994         /**
4995          * Find the index number of the plural rule appropriate for the given number
4996          * @param int $number
4997          * @return int The index number of the plural rule
4998          */
4999         public function getPluralRuleIndexNumber( $number ) {
5000                 $pluralRules = $this->getCompiledPluralRules();
5001                 $form = Evaluator::evaluateCompiled( $number, $pluralRules );
5002                 return $form;
5003         }
5004
5005         /**
5006          * Find the plural rule type appropriate for the given number
5007          * For example, if the language is set to Arabic, getPluralType(5) should
5008          * return 'few'.
5009          * @since 1.22
5010          * @param int $number
5011          * @return string The name of the plural rule type, e.g. one, two, few, many
5012          */
5013         public function getPluralRuleType( $number ) {
5014                 $index = $this->getPluralRuleIndexNumber( $number );
5015                 $pluralRuleTypes = $this->getPluralRuleTypes();
5016                 return $pluralRuleTypes[$index] ?? 'other';
5017         }
5018 }